Center for Population Research Monograph THE WALNUT CREEK CONTRACEPTIV DRUG STUDY A PROSPECTIVE STUDY OF THE SIDE EFFECTS OF ORAL CONTRACEPTIVES Volume I Findings in Oral Contraceptive Users and Nonusers at Entry Into the Study Edited by SAVITRI RAMCHARAN, M.D., Ph.D. Research Director, Contraceptive Drug Study Kaiser-Permanente Medical Center Walnut Creek, California DHEW Publication No. (NIH) 74-562 Supported by the Center for Population Research, National Institute of Child Health and Human Development under Contract Number NO1-HD 38-2710 WN WON XN aN WR NN NN Na) AON) RN NX ™N ~, WX J BON |) 2 A ) YY U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE \ Public Health Service National Institutes of Health National Institute of Child Health and Human Development Center for Population Research Bethesda, Maryland 20014 For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, D.C. 20402 - Price $6.05 Stock Number 1746-00021 Fe /37 i. WV3 V. / PUBLIC HEALTH LIBRARY Foreword The mission of the Contraceptive Evaluation Branch (former- ly the Fertility Regulating Methods Evaluation Branch) of the Center for Population Research, National Institute of Child Health and Human Development, is to support studies to assess the medical effects of contraceptive agents in present use. Development and support of The Walnut Creek Contraceptive Drug Study by the Branch represents one of the earliest efforts to detect and quanti- tate metabolic and adverse effects of oral contraceptive drugs in a large population. The availability of a multiphasic screening pro- gram linked with a population using a comprehensive medical care program suggested the feasibility of the endeavor. The data pre- sented in this and two subsequent volumes are unique in describing (1) patterns of contraceptive use, (2) effects of steroid contracep- tive drugs on a generally healthy, free-living population, and (3) population-based laboratory and clinical data of value for refer- ence purposes. PHILIP A. CORFMAN, M.D. Director Center for Population Research National Institute of Child Health and Human Development iii APR 19 1976 iE CD heen er AT he HG a pL SE CoE ae ae Introduction This contraceptive drug study monograph brings together published reports which have originated with The Walnut Creek Contraceptive Drug Study. It also elaborates on the background data supporting the published results and communicates some unpublished findings. The monograph, to appear in three volumes, is based on studies drawn from data collected on approximately 18,000 women recruited into the study during the period December 1868 through February 1972. This first volume describes, in part A, the study, the popula- tion and the sample, and laboratory procedures used and, in part B, the results of cross-sectional analyses of observations made on subjects examined during the early recruitment period, pri- marily 1969. Volume II will contain results of cross-sectional analyses of additional variables studied, with the observations covering the entire cohort of subjects. Some longitudinal analyses will be carried out on the data where they are found to be relevant and feasible. A projected table of contents for volume II follows: 1. The occurrence and course of hypertensive disease in users and nonusers of oral contraceptive drugs Oral contraceptives and pulse rate Oral contraceptives and the red blood cell Oral contraceptives, pregnancy, and the leukocyte count The effect of oral contraceptive use and other factors on the thromboelastogram 6. Effects of oral contraceptives on serum protein fractions obtained by electrophoresis 7. The prevalence of rheumatoid factor: the effects of oral contraceptives 8. Hearing levels according to use of oral contraceptives 9. Intraocular tension and use of oral contraceptive drugs 10. The effect of oral contraceptives on Achilles tendon reflex relaxation time 11. Oral contraceptives, personality, and changes in depression 12. Bacteriuria and oral contraceptives BoD vi INTRODUCTION 13. Some physiological and biochemical measurements over the menstrual cycle 14. Standards for automated spirometric screening 15. Body height, weight, and skinfold thickness 16. Technical problems of data handling in drug surveillance studies Volume III is planned to cover analyses of certain variables not included in volume II, primarily serum chemistry results, mammograms, radiograms of the chest, and certain medical con- ditions and symptoms. The editor would like to express her thanks to Dr. Alan Goldfien, Professor of Medicine and Obstetrics and Gynecology, University of California School of Medicine, for his comments and guidance; Dr. Frederick A. Pellegrin, Chief of Internal Medicine, Kaiser-Permanente Medical Center, Walnut Creek, for his sug- gestions; Dr. Heinz W. Berendes, Chief of the Contraceptive Evaluation Branch; and Dr. John J. Schrogie, formerly Branch Chief, Center for Population Research, National Institute of Child Health and Human Development, for their encouragement and support; and, finally, the publishers and authors for their permission to reproduce published manuscripts. The authors also wish to acknowledge with thanks the contributions made by the many persons who at one time were or currently are associated with the Contraceptive Drug Study. An ongoing large project, of necessity, depends on the collaborative efforts of many persons and organizations. It is impossible to list by name all those who have helped in the project. SAVITRI RAMCHARAN, M.D., Ph.D. Editor Research Director Contraceptive Drug Study Kaiser-Permanente Medical Center Walnut Creek, Calif. Contributors to Volume I NEIL E. ANDERSON, M.D., Department of Internal Medicine, Per- manente Medical Group, Walnut Creek, Calif. WILLARD L. BROWN, M.D., Chief, Department of Obstetrics and Gynecology, Permanente Medical Group, Walnut Creek, Calif. THOMAS J. DUFFY, M.D., Department of Internal Medicine, Per- manente Medical Group, Walnut Creek, Calif. IRWIN R. FiscH, M.D., Medical Director, Contraceptive Drug Study, and Department of Internal Medicine, Permanente Medical Group, Walnut Creek, Calif. SHANNA H. FREEDMAN, Ph.D., Assistant Professor of Mathe- matics, California State College, Sonoma, Rohnert Park, Calif., and Statistician to the Contraceptive Drug Study. S. JEROME KUTNER, Ph.D., Psychologist, Langley-Porter Institute for Medical Research, Valley Medical Center, San Jose, Calif., formerly Psychologist to the Contraceptive Drug Study. ALBERT V. MYATT, M.D., Department of Internal Medicine, Perma- mente Medical Group, Walnut Creek, Calif. FREDERICK A. PELLEGRIN, M.D., Principal Investigator, Contra- ceptive Drug Study, Chief of Internal Medicine, and Assistant Physician-in-Chief, Permanente Medical Group, Walnut Creek, Calif. NANcy R. PHILLIPS, Ph.D., Epidemiologist, Cardiovascular Re- search Institute, University of California School of Medicine, San Francisco, formerly Epidemiologist to the Contraceptive Drug Study. SAVITRI RAMCHARAN, M.D., Ph.D., Research Director, Contracep- tive Drug Study, Kaiser-Permanente Medical Center, Walnut Creek, Calif. IQs a pas a 10 El a i 3 ul oo a fo = I of amy 1 1 wl ERE ATL Ry in SN fi Sig ary Ig h FE ar Thi agiin [7 Bl AEE Contents Page Foreword wns III Introduction _______________________ He Contributors to Volume I _________________ VII PART A. STUDY, POPULATION, PLAN, AND PROCEDURES ______________ 1 Chapter 1. The Noncontraceptive Effects of Oral Contraceptive Drugs: The Kaiser-Permanente Study coe 3 F. A. Pellegrin, S. Ramcharan, I. R. Fisch, and N. R. Phillips A general description, including preliminary results of the Kaiser- Permanente Contraceptive Drug Study, is presented here. This is a prospective study which has been conducted at Walnut Creek, Calif., since 1968 to determine the side effects of oral contracep- tives by comparing the differences in health status between oral contraceptive users and nonusers. Chapter 2. Socioeconomic Characteristics and Methods of Fertility Control of Female Subscribers to the Kaiser Foundation Health Plan in the Walnut Creek Service Area, 1968 ________________________ 21 N. R. Phillips The data obtained from this sample survey indicated that this population of women was predominantly white, middle class, and married. From this population was selected the group of women who participated in the Contraceptive Drug Study. Chapter 3. Changes in the Use of Oral Contraception and Other Meth- ods of Fertility Control Between 1968 and 1971 __________________ 83 N. R. Phillips The methods of fertility control used by women who were admitted into the Contraceptive Drug Study between January 1969 and June 1971 are here compared with those reported by women who responded to the sample survey of 1968. Surgical sterilization and use of intrauterine devices increased between 1968 and 1971; oral contraceptive use decreased; all other methods showed little change. Chapter 4. Automated Multitest Laboratory Phases and Quality- Control ProcolUIes oie esr sion mem sea mi om mr mei 91 I. R. Fisch and S. H. Freedman A brief description is given of each phase of the Automated Multi- test Laboratory examination which was administered to each study subject at entry. Quality-control procedures undertaken and the reliability of the data are discussed. ix X CONTENTS PART B. SOME INITIAL RESULTS htm stint wim sree ee Chapter 1. Oral Contraceptives, Pregnancy, and Blood Pressure ______ I. R. Fisch, S. H. Freedman, and A.V. Myatt Cross-sectional analyses of blood pressure measurements obtained from 7,605 women who were admitted into the study in 1969 showed that oral contraceptives were associated with a modest yet definite increase in blood pressure, more marked for systolic than diastolic. Chapter 2. Sensitivity and Specificity of the 1-hour Glucose-Tolerance BETEONING BEL oii mmm ss min se ssasist soars meni vt N. R. Phillips and T. J. Duffy Glucose-tolerance screening and followup procedures in the Auto- mated Multitest Laboratory are presented here. Assuming a preva- lence rate of glucose intolerance of 1 percent in the study subjects, the sensitivity of the screening test was 48 percent, its specificity 95.4 percent. Chapter 8. Effect of Age, Hours Since Last Food, Time of Day, and N. R. Phillips and T. J. Duffy Ketonuria on 11-hour Glucose Tolerance occa ee, One-hour glucose tolerance was analyzed for 6,692 women at entry into the study in 1969. Mean serum-glucose values increased with increasing age, increasing hours since last food (except for over- night fasts), and prevalence of ketonuria. Time of day had little influence on glucose values. Chapter 4. One-hour Glucose Tolerance in Relation to the Use of Oral COnirateptive Drums ee ne i ——— N. R. Phillips and T. J. Duffy The results of cross-sectional analysis of 1-hr glucose tolerance in 4,815 women who entered the study in 1969 indicated an association between oral contraceptive use and reduced glucose tolerance (high 1-hr serum-glucose levels). Chapter 5. Spirometry and Oral Contraceptives ____________________ S. H. Freedman and N. E. Anderson No relationship was found between oral contraceptive use and spirometric values measured by means of a wedge spirometer in 2,066 women examined at entry into the study in 1969. Chapter 6. Effects of Oral Contraceptives on the Thromboelastogram __ I. R. Fisch, S. H. Freedman, and F. A. Pellegrin Thromboelastography, a method of measuring in vitro coagulation activity of the blood, was performed on 1,133 women between January 1969 and April 1970. Users of oral contraceptives were found to have shorter clotting times and firmer clots than non- users. Chapter 7. Types of Oral Contraceptives, Depression, and Premen- STUAl-SYMDIOIIE oo oi msm oi tie mis Ti tein im bes od BBE Se cl S. J. Kutner and W. L. Brown Analysis of depression in 3,919 women and premenstrual moods in 5,151 women admitted into the study in 1969 casts doubt on the hypothesis that oral contraceptives cause depression. 135 147 169 189 201 213 CONTENTS Chapter 8. History of Depression as a Risk Factor for Depression With Oral Contraceptives and Discontinuance ___________________ S. J. Kutner and W. L. Brown Cross-sectional analysis of 5,151 women admitted into the study in 1969 indicated that women who had discontinued oral contracep- tives had a greater prevalence of history of depression during or after pregnancy than had current users or those who had never used oral contraceptives. Contributors to the Contraceptive Drug Study ——______________ Index TABLES Part A, Chapter 1. The Noncontraceptive Effects of Oral Contracep- tive Drugs: The Kaiser-Permanente Study ______________________ 1. Subjects having their first Automated Multitest Laboratory examination during the period December 1968 through Febru- ary 1972, by age and use of oral contraceptives _____________ Relative frequency of current users having their first Auto- mated Multitest Laboratory examination December 1968 through February 1972 taking various compositions of oral COMLLBCOPLINGS. odin dari ii mimeo sim mm mbm oe nm mp mtr Number and percent of Contraceptive Drug Study cohort remaining in the Health Plan, Walnut Creek Service Area, at the end of 1971, by year of entry into the study and year of DILLH: tim smi arm elim Svs oi i ss ss rin asin aero res Part A, Chapter 2. Socioeconomic Characteristics and Methods of Fertility Control of Female Subscribers to the Kaiser Foundation Health Plan in the Walnut Creek Service Area, 1968 ____________ 4. 5. 6. 7. 10. 11. 12. 13. 14. 15. Age distribution of study population and survey sample ____ Cumulative response rate by number of probes ____________ Cumulative proportion ever exposed to contraceptive drugs by NUMDRY OF DIODES coe mm i mi mim om im mim mom Use of contraceptive drugs by female members of Kaiser Foundation Health Plan, living in the Walnut Creek Service Area, February 1968 ____________________ ________________ Frequency of noncontraceptive users among oral contraceptive TEOIR ron vim mmmm mri ar ee tt Ms ES Accumulated proportion ever exposed to contraceptive drugs by the end of each calendar year, 1960-67, by year of birth__ Duration of exposure—percent of ever users _______________ Duration of exposure—percent of ever users of contraceptive drugs for contraceptive purposes _________________________ Duration of exposure—percent of ever users of contraceptive drugs for noncontraceptive purposes ______________________ Duration of exposure—percent of current users for purpose OF COMPACODUION oe cima in Duration of exposure—percent of past users for purpose of LT Se Duration of exposure—percent of current users for non- contraceptive purposes __________________________________ xi Page 229 241 243 12 21 23 24 24 27 28 29 34 35 35 36 36 37 16. 17. 18. 19. 20. 21. 23. 24. 25. 26. 21. 28. 29. 30. 31. 32. 33. 34. 35. 36. 317. CONTENTS Duration of exposure—percent of past users for noncontra- COPLIVE PULDOSES: omission sims iim sitio sms ime Extent of total population exposure (percent) _____________ Current users of oral contraceptives (percent) in December of each calendar year, 1960-67, among white women, then married, and 20-44 yeors of A486 eee —— Comparison between National Fertility Study and Contra- ceptive Drug Study survey on oral contraceptive use among white married women in 1965 (percent) ___________________ Use of contraceptive drug, December 1965, by white married women, according to whether membership in Health Plan DOSLARICE TOBE cm comm ih mm st iii i ee mai Discontinuation of oral contraception—white married women Past users of oral contraception: reason for discontinuation— white married women (Percent) cee —emsmsi s—— Total months of oral contraceptive use: dropouts—white mar- ried women (percent) ________.___________________________ Total months of oral contraceptive use: discontinuance be- cause of side effects—white married women (percent) ______ Relative frequency of ever users of contraceptive drugs, by number of live births—white married women (percent) _____ Relative frequency of ever users of contraceptive drugs for preventing pregnancy, by number of live births—white mar- ried Women (DErCRnl) vec msi mimi ss mts mi bos Relative frequency of ever users of contraceptive drugs, by education—white married women (percent) _______________ Use of oral contraceptive drugs for purpose of preventing pregnancy, by religion—white married women ______________ Ever users of contraceptive drugs for contraceptive purposes, by frequency of church attendance—white married Catholic WOBI os cant ho comers Si i eaten Ever users of contraceptive drugs for contraceptive purposes, by frequency of church attendance—white married Protestant NOTCH or nts am A a srr sos Eligiblity status of white married women (percent) _______ Use of contraceptive drugs among white married women eligible to use contraception (percent) ____________________ Methods of contraception employed by white married women eligible to use contraception (percent) _____________________ Prior method of birth control employed by current users of contraceptive drugs (percent) ___________________________ Method of birth control used by:white married women eli- gible to use contraception—not taking contraceptive drugs RDBLCENIEY | imme eee sie mr ce ren er re eR EE Method of birth control used by white married women eligible to use contraceptive drugs—past users of contraceptive drugs for contraception (percent) _____.________________________ Method of birth control used by white married women eligible to use contraception—never users of contraceptive drugs (percent) — — Page 38 38 39 42 43 45 45 47 47 48 49 49 50 51 51 52 52 55 56 58 59 60 38. 39. 40. 41. 42, 43. 44. 45. 46. Part A, Other 47. 48. CONTENTS Prevalence of contraceptive operations—white married Health Plath SUDSEYIDELE oo cme membre stmt mi rmirimnda lei pein mei Prevalence of remedial sterilizing operations—white married Health Plan 2ubsenibens oc nvomenarumamn rah fphben dene Prevalence of all sterilizing surgery (percent distribution of COUDIERY woes conse sin urine somtimes ri oes asi demesne asic mioton Comparison of Walnut Creek Kaiser (WCK) population with 1965 National Fertility Study population: Contraceptive op- erations among white couples (percent) ____________ _______ Comparison of Walnut Creek Kaiser (WCK) population with 1965 National Fertility Study population: Remedial sterilizing operations among white women (percent) _________________ Prevalence of contraceptive operations, by number of live BIH oe rere rrr er EE ER SEE ERS Difference between voluntarily sterilized couples and others in number of live births ________ ________________________ » Prevalence of contraceptive operations, by religion of couple and by wife’s edueation __________________________________ Prevalence of contraceptive operations, by wife’s education and religion of couple and by wife’s education and age at first PIEERBICY —omemrimm wer rpms Ear rE eR im a Rm mmm Chapter 3. Changes in the Use of Oral Contraception and Methods of Fertility Control Between 1968 and 1971 ________ used by women 20-44 years of age _______________________ Relative frequency of various methods of fertility control Prevalence of sterilizing female surgery among women 20-44 YES OF B00 nomen re AE ee ee ee cl Part B, Chapter 1. Oral Contraceptives, Pregnancy, and Blood Pressure 49. 50. 51. 52. 53. 54. 55. 56. 617. 58. 59. 60. 61. Blood pressure by age, hormone use, and pregnancy status __ Percent cases with elevated blood pressure by age and oral contraceptive use _______________________________________ Mean blood pressure (mm Hg) of current users by estrogen LYDE ANA QOBO ec ios mei mem mmm ete memes rene Mean blood pressure (mm Hg) by progestin type and dose _ Effect of Deladumone on blood pressure (mm Hg) of 250 DOSLDATLUIN WOIMMEH cc eso i i mim on mm i en ses sp mie em Relative frequency of systolic blood pressure (mm Hg) for all oral contraceptive groups, by age _________________________ Relative frequency of diastolic blood pressure (mm Hg) for all oral contraceptive groups, by age _____________________ Comparison of Contraceptive Drug Study blood-pressure read- ings with National Health Survey data (1960-62) __________ Mean systolic blood pressure (mm Hg) by weight and age for all oral contraceptive groups ____________________________ Mean diastolic blood pressure (mm Hg) by weight and age for all oral contraceptive groups __________________________ Blood pressure by parity for four age groups ______________ Mean blood pressure (mm Hg) by ABO blood type —_______ Systolic blood pressure by day of menstrual cycle for users and nonusers of oral contraceptive drugs 63 64 66 67 69 70 72 83 817 88 105 107 110 113 115 116 118 119 121 123 124 126 128 xiv 62. 63. 64. CONTENTS Diastolic blood pressure by day of menstrual cycle for users and nonusers of oral contraceptive drugs .._______________ Systolic blood pressure by time of day and hours since last 00d om rit rm i RE Tp So en Diastolic blood pressure by time of day and hours since last OO i i we ts fl ER SA etl ef Part B, Chapter 2. Sensitivity and Specificity of the One-hour Glucose-Tolerance Screening Test _________________._____ 65. 66. 67. 68. 69. 70. 71, 72. Recommended glucose 1084 eee emer Mean 1-hr serum glucose level following 75-g challenge in relation to recommended load for body size —_______________ Screening limits in effect from Jan. 1, 1969, to June 15, 1970 _ Screening limits in effect from June 16, 1970, to Oct. 18, 1972 Glucose loading dose by height and weight based on 40 g per square meter of DOAY SUPIBE0 cui ims ms soma Somogyi-Nelson blood test (venous) sensitivity and specificity for diabetes at different screening levels, by hours after test NEE re re Re ET i ae Se ear res Sensitivity and specificity of Automated Multitest Laboratory glucose-tolerance screening test at postulated disease-preva- JEHCE FOR08 . .o. Ll ie rms pm pe ie e a rte slr im raonce Frequency of abnormal standard glucose-tolerance test (SGTT) among women referred for followup testing in rela- tion to the conditions of the screening test —_______________ Part B, Chapter 3. Effect of Age, Hours Since Last Food, Time of Day, and Ketonuria on 1-hour Glucose Tolerance _________________ 73. 74. 75. 76. 7% 78. 79. 80. 81. The mean and standard deviation of 1-hr serum-glucose levels (mg per 100 TY, DY BH0 i... eee ess ior cise solid eine Location of selected percentiles of distribution of 1-hr serum glucose (mg per: 100 M1), DY 880 ..oeimemtomm mmm Sm The mean and standard deviation of 1-hr serum-glucose levels (mg per 100 ml), by hours since last food when challenged __ The mean and standard deviation of 1-hr serum-glucose levels (mg per 100 ml), by time of challenge and hours since last 00d rr ee Be ER es Mr rds pe a Es One-hour serum-glucose levels (mg per 100 ml) by age and by time of challenge since last food: Means and standard devia- HON “(2,585 WOMEN) oo i ommend md it i Prevalence of ketonuria, by time of day mene ineeammims Prevalence of ketonuria (test positive) by time of day speci- men collected and hours since last food ____________________ Prevalence of ketonuria (positive test), by age, among 5,644 WOT oe if ii od se ci i i tH The mean and standard deviation of 1-hr serum-glucose levels (mg per 100 ml), by result of Ketostix test, and the mean difference between ketonurics and nonketonuries (==95- percent confidence interval), by time of challenge and hours since last food Part B, Chapter 4. One-Hour Glucose Tolerance in Relation to the Use of Oral Contraceptive Drugs ___ Page 130 133 133 135 137 138 139 139 140 141 144 145 147 150 153 154 156 158 160 161 162 163 169 CONTENTS 82. Age and oral contraceptive drug-use status of the study SUDJOCEE oo. imi ie omens ot rites ed SES hed eg es ems 83. Time of glucose challenge and interval since last food, by oral contraceptive Arug-use StatUS o.oo mmm Ee me 84. Serum-glucose levels by age and oral contraceptive (OC) drug- use status: Mean and standard deviation (SD) _____________ 85. Distribution of 1-hr serum glucose levels among users of oral contraceptive (OC) drugs and among nonusers of comparable age: Means, standard deviations, and percentiles __________ 86. Mean 1-hr serum glucose (mg/100 ml) by type of oral contra- ceptive (OC) drug—Age adjusted to the overall age distri- bution of current OC users ______________________________ 87. Mean 1-hr serum glucose by estrogen dose for women taking combination-type drugs employing mestranol ______________ 88. 1-hr serum glucose (mg/100 ml) by family history of diabetes for users of oral contraceptive (OC) drugs and for nonusers of comparable age: Mean and standard deviation (SD) ____ 89. 1-hr serum glucose (mg/100 ml) by Ponderal Index for users of oral contraceptive (OC) drugs and for nonusers of com- parable age: Mean and standard deviation (SD) __________ Part B, Chapter 5. Spirometry and Oral Contraceptives ____________ 90. Distribution of oral contraceptives taken by current users admitted to the study in 1969 _____________________________ 91. Mean spirometric values and age-adjusted mean spirometric values, by oral contraceptive drug-user status _____________ 92. Mean smoking index by age and oral contraceptive use for Contraceptive Drug Study population _._.__.____________ ____ 93. Age-adjusted comparison of pregnant women with all women under 40 years of age __________________._________________ Part B, Chapter 6. Effects of Oral Contraceptives on the Thrombo- CIOS OR TAIN mmr mates isa ais sess aie Bt i i AG a i i mi i ri Sacra 94. Thromboelastographic parameters by oral contraceptive use (mean + 1.96 standard deviation of mean) _______________ 95. Thromboelastographic parameters by estrogen type and dose (mean = 1.65 standard deviation of mean) _______________ Part B, Chapter ?. Types of Oral Contraceptives, Depression, and Premenstrus] SYMPIOME: vii mmission 96. Product-moment correlations between depression and demo- graphic characteristies __________________________________ 97. Composition of oral contraceptives taken by current users admitted to the study in 1969 ___________________________ 98. Use of oral contraceptives and depression _________________ 99. Type of regimen and depression _________________________ 100. Two types of oral contraceptives and depression ___________ 101. Dose of progestin and depression _________________________ 102. Months of current oral contraceptive use and depression ____ Part B, Chapter 8. History of Depression as a Risk Factor for De- pression With Oral Contraceptives and Discontinuance ____________ 103. Product-moment intercorrelations among the measures of Te = XV Page 170 171 173 175 178 179 181 183 189 191 195 199 199 201 206 207 213 215 217 218 220 221 223 224 229 232 xvi 104. 105. 106. 107. 108. Part A, CONTENTS Percentage with severe depression during or after pregnancy and use of oral contraceptives oe Mean months of last oral contraceptive use and history of depression among Past SEIS —.... coc —m—————— Current depression and history of severe depression during or after pregnancy among never users of oral contraceptives Use of oral contraceptives and depression among patients with history of depression ooo oe Dose of progestin and depression among patients with history OL depression —o . . . e erE — ral FIGURES Chapter 2. Socioeconomic Characteristics and Methods of Fer- tility Control of Female Subscribers to the Kaiser Foundation Health Plan in the Walnut Creek Service Area, 1968 ___________________ 1, Part A, Other 7. Part B, 10. Least-squares extrapolation of percent ever exposed to contra- ceptive drugs (white married female Health Plan subscribers, Walnut Creek Service AT€8) wee mcm ene Proportion ever exposed to contraceptive drugs by end of each calendar year 1960-67, by year of birth (white married female Health Plan subscribers, Walnut Creek Service Area) ______ Proportion ever users of birth control pills (white married female Health Plan subscribers, Walnut Creek Service Area) Proportion ever exposed to contraceptive drugs at given age among individual birth cohorts 1938-47 (white married female Health Plan subscribers, Walnut Creek Service Area) ______ Proportion of white married women 20-44 years of age taking oral contraceptives in December of each calendar year 1960- 67 (Health Plan subscribers, Walnut Creek Service Area) __ Proportion of white married women 20-44 years of age taking oral contraceptives in December of each calendar year 1960- 67, by 5-year age groups (Health Plan subscribers, Walnut Creek Service Avena) ooo lm a el in Chapter 3. Changes in the Use of Oral Contraception and Methods of Fertility Control Between 1968 and 1971 ________ Proportion taking oral contraceptive drugs among women 20-44 years of age having Automated Multitest Laboratory health examinations, January 1969-June 1971 ______________ Chapter 1. Oral Contraceptives, Pregnancy, and Blood Pres- Mean systolic and diastolic blood pressure by age for past and never users of oral contraceptive drugs __________________ Blood pressure by age, contraceptive use, and pregnancy SO OE i EE Ei ee a a SE me ere Percent of patients with elevated blood pressure by age and oral contraceptive use ___________________________________ [The above tables and figures were published with the original paper in JAMA 222:1507-1510 (December 18), No. 12, 1972.] 1a. Diastolic blood pressure by total months of use for current users and past users of oral contraceptive drugs __________ Page 234 234 235 236 237 21 25 30 31 33 40 41 83 84 105 108 109 110 12. 13. 14. 15. 16. 117. 18. 19. 20. 21. CONTENTS Diastolic blood pressure by months since last use for past users of oral contraceptive drugs _.. Mean blood pressure by estrogen and progesterone doses for current oral contraceplive users... roe Relative frequency distribution of systolic blood pressure for all oral contraceptive groups by age — coin en Relative frequency distribution of diastolic blood pressure for all oral contraceptive groups by age ______________________ Mean blood pressure by age for oral contraceptive users and nonusers and for National Health Survey data on white TOMBIGE | or mmm i ins ti ng ado in mem ee rg AR rt ST Mean blood pressure by weight and age for all oral contracep- HIVE ZIOMPE corre armor somo doi ais iain sarin Systolic and diastolic blood pressure by parity (age specific) Blood pressure by ABO blood type (mean and 95-percent confidence intervals of the means) ______________________ Blood pressure by day of the menstrual cycle for users and nonusers of oral contraceptives ._.____________ Blood pressure by time of day, irrespective of hours since last L000 Lomas i Breer si ale wid Part B, Chapter 2. Sensitivity and Specificity of the 1-hour Glucose- Tolerance Screening Tesh cco covers mmm ws tsi 22. Classification of test results for determining sensitivity. and SOCCIRCHY =. om means ire Fema re Hts lita? Part B, Chapter 3. Effect of Age, Hours Since Last Food, Time of Day, and Ketonuria on 1-hour Glucose Tolerance ________________ 23. 24. 25. 26. 217. 28. 29. 30. 31. 32. One-hour serum-glucose relative frequency distribution _____ Mean 1-hr serum-glucose level by age _____________________ Deciles of 1-hr serum-glucose levels by age _______________ Mean 1-hr serum glucose (with 95 percent confidence inter- val) by hours since last food when challenged ____________ Mean 1-hr serum glucose by time of challenge and hours since dash Food | eam rae a Mean 1-hr serum glucose by age, time of challenge, and hours since last food: 2,599 women, first 6 months of 1969 ______ Prevalence of ketonuria (with 95 percent confidence interval) BY No QRY oes mp Eat imei eee epimers Prevalence of ketonuria by time of day and hours since last DOOR. occas serene otis oo SS i ra lo Mean 1-hr serum glucose by time of challenge _____________ Mean 1-hr serum glucose by time of challenge and hours since last food for women with negative Ketostix tests __________ Part B, Chapter 4. One-hour Glucose Tolerance in Relation to the Use of Oral Contraceptive Drugs _______________________________ 33. 34. Mean 1-hr serum glucose by age and oral contraceptive drug- BEE BEALAR cern cmrmmm im wri i es SAA I Er Relative frequency distributions of 1-hr serum glucose values among oral contraceptive users and nonusers of comparable DIR coe concromrmeiiocss simon a HE ST Re xvii Page 111 nz 117 ar 120 122 125 127 129 131 140 147 149 151 152 155 157 159 160 162 164 165 169 172 175 xviii 35. 36. 37. 38. 39. 40. 41. CONTENTS Cumulative frequency distribution of 1-hr serum-glucose values for current oral contraceptive users and nonusers of COMPALIDIO SATE oo aise siminiiommts meee ee pie sca Sint Mean 1-hr serum glucose by age for current oral contracep- tive users by type of formulation and for nonusers ________ Mean 1-hr serum glucose by number of days since onset of last menses for current oral contraceptive users taking com- bination formulation, and for nonusers of comparable age Mean 1-hr serum glucose (with 95 percent confidence level) of current oral contraceptive users by continuous months of EE cre Totnes Ee Mean 1-hr serum glucose (with 95 percent confidence level) of past oral contraceptive users by months since last dose, age adjusted to the age distribution of current users ________ Mean 1-hr serum glucose by age and by family history of dia- betes, and oral contraceptive drug-use status ______________ Mean 1-hr serum glucose by age, Ponderal Index (PI), and oral contraceptive drug-use status oe ein Part B, Chapter 5. Spirometry and Oral Contraceptives _____________ 42. 43. 44. 45. 46. Age distribution of women tested by spirometry according to oral contraceptive drug-user status ________________________ Forced Vital Capacity (FVC) and Forced Expiratory Volume (OEY 1.0)o DY O08 camel emma ar pm mira Forced Vital Capacity (FVC) and Forced Expiratory Volume (FEV 1.0), by height... oo comin par meme Forced Vital Capacity (FVC) and Forced Expiratory Volume (FPEYV 1.0), bY Weight oo oi wm mie mire Forced Vital Capacity (FVC) and Forced Expiratory Volume (FEV 10), by smoking index moomoo oo Part B, Chapter 6. Effect of Oral Contraceptives on the Thrombo- BIAS. a ne ne En le tn em in Eig eth i rt cee 417, 48. Procedure for measuring the thromboelastogram (TEG) ___ Frequency distributions of TEG parameters between current users and current nonusers of contraceptive drugs __________ Part B, Chapter 7. Types of Oral Contraceptives, Depression, and Premenstrual] SYMDIOMS: ic vie eds aia pm 49. 50. Dose of progestin and four measures of severe depression __ Current premenstrual irritability and number of months since last oral contraceptiveuse oe Page 176 177 179 180 180 182 183 189 193 194 196 197 198 201 203 206 213 222 PART A Study, Population, Plan, and Procedures a v he REE alot ft TEE y 5 a RA - ig Lis i a I~ - = hg 1 A ie fread Chapter 1 The Noncontraceptive Effects of Oral Contraceptive Drugs: The Kaiser- Permanente Study FREDERICK A. PELLEGRIN, SAVITRI RAMCHARAN, IRWIN R. FIsCH, AND NANCY R. PHILLIPS Introduction.—In 1968, a prospective study was started within the Kaiser-Permanente Medical Care Program at Walnut Creek to determine the side effects of oral contraceptives (OC’s), by comparing the differ- ences in health status between women who use the drugs and those who do not use them. Population and sample selection.—The Kaiser Foundation Health Plan membership in this suburban area was predominantly white, middle class, and married. From this membership was selected a cohort of women born 1914 and after who were seeking routine medical and gynecological checkups during the period December 1968 through Feb- ruary 1972. Automated Multitest Laboratory (AML) examinations.—At entry into the study women received, in a specially designed laboratory, an AML examination consisting of a battery of medical questionnaires and biochemical and physiological tests, followed by a gynecological and general medical examinations. Cohort followup.—This was accomplished by annual repeat AML examination, supplemented by mailed questionnaires to those who did not have repeat AML examinations and by information abstracted from in- patient and outpatient medical charts at the Kaiser-Permanente Medical Center. Losses from the study.—Based on termination of membership from the Health Plan during the first 3 years of the study, the attrition rate for women 35-44 years of age at entry was about 4 percent per year. There was no difference in the attrition rate between users and non- users of OC’s. Data processing and analysis.—All data were coded and keypunched for storage on magnetic tape. Cross-sectional analyses and to a lesser extent longitudinal analyses of the data were carried out using tabulation programs and multiple regression techniques. Results.—Reports have been published showing an increase in blood pressure, in coagulability of the blood and in red blood cell indices, and a decrease in glucose tolerance in association with OC use. Unpublished preliminary findings indicated an increase in the risk of developing hyper- tensive disease, an increase in pulse rate and in the leukocyte count in 1 2 STUDY, POPULATION, PLAN, AND PROCEDURES relation to OC use. The relationship between serum cholesterol levels and OC’s varied with age, levels being lower in OC users over 40 years of age, and higher in users under 40 years, compared to nonusers. Total serum protein, serum albumin, and gamma globulin levels were lower in users, while alpha and beta globulin fractions were higher. Preliminary results suggested a lower prevalence of rheumatoid factor and a higher prevalence of asymptomatic bacteriuria in users. Hearing levels appeared better and the Achilles-tendon-reflex relaxation time prolonged in OC users. There was no demonstrable relationship between OC use and spirometric measurement, or depression as measured by scores on the Minnesota Multiphasic Personality Inventory. Discussion.—The differences between users and nonusers were small but statistically significant. Detection of such differences was related to the large sample sizes. The clinical importance of many of these differ- ences remains undetermined and will have to be assessed by long-term followup studies. The Contraceptive Drug Study (CDS) at Kaiser-Permanente is a prospective study to determine the side effects of oral contra- ceptive drugs (OC’s) by comparing the differences in health status between women who use and women who do not use these drugs. The purpose of the project is not only to determine whether the side effects of OC’s might be adverse, beneficial, or trivial, but also to identify the characteristics of women who are likely to develop side effects. The study is being conducted by the Kaiser-Permanente Medi- cal Care Program of Northern California. All study subjects, numbering approximately 18,000 women born 1914 and after, are members of the Kaiser Foundation Health Plan of Northern Cali- fornia and reside in the Walnut Creek Service Area, which is composed of a number of suburban communities within commut- ing distance of Oakland and San Francisco. The Health Plan provides comprehensive prepaid medical care to nearly one-fourth of this suburban population. All physician services are provided in the Kaiser-Permanente Medical Center, which includes a hos- pital, physician offices, and an Automated Multitest Laboratory (AML). The AML was developed specifically for this study, and the collection of data began in December 1968. By early 1969 oral contraceptives were being distributed in the United States at the rate of approximately 8.5 million cycles per month. It is unprecedented to have so many people exposed to such an active pharmacologic agent for reasons other than controlling disease. The reported side effects have been legion.! 2 The efficacy of these drugs as contraceptives has been adequately proven and the resultant social benefits may in fact outweigh some NONCONTRACEPTIVE EFFECTS OF ORAL CONTRACEPTIVES 3 of the reported medical hazards associated with their use. Many of these side effects are minor but some are major and life threat- ening.? * For the most part, the minor effects have been shown to be reversible on discontinuance of the drug,’ ¢ but the possibility exists that long-term sequelae might result from many years of exposure to these pharmacological agents. Oral contraceptives have been shown to be related to exacerbation of existing disease.” It has also been shown that the occurrence of serious disease conditions in association with the ingestion of oral contraceptives might result from the combined effect of oral contraceptives in the presence of other predisposing and/or precipitating factors.®® 1° On the other hand, there is some evidence that beneficial side " effects might be expected to occur since Vessey and coworkers have presented data suggesting that OC’s tend to protect against the development of benign lesions of the breast.’* Many reported findings have been inconclusive because they were based on too few observations or because they could not be related to a population base from which measures of risk or benefit could be derived.’? ** The need for followup studies was empha- sized by the Advisory Committee on Obstetrics and Gynecology of the Food and Drug Administration in the 1969 Report. At that time a few projects, including the present one, were in progress. The continued exposure of millions of women, starting in youth and extending possibly throughout their reproductive life, to phar- macologic agents the effects of which have not yet been evaluated, required that careful studies, both of the case-comparison and cohort types, be maintained in human populations. POPULATION CHARACTERISTICS As part of the feasibility study for this project, in February 1968 a survey was carried out among female members of the Kaiser Foundation Health Plan. The women are 20 through 54 years of age. This survey was confined to the Health Plan service area in which the study was to be conducted and is described in detail in chapter 2. The survey was to provide information about social characteristics and contraceptive practices of the potential study subjects and was carried out by means of a self-administered questionnaire mailed to a 20-percent probability sample number- ing 3,886 women. Nonrespondents received telephone and mail reminders and, finally, a telephone interview to complete the ques- tionnaire. The ultimate response rate was 92 percent. Nearly all (97.6 percent) the women in this population were white. Eighty-nine percent were married and living with their 4 STUDY, POPULATION, PLAN, AND PROCEDURES husbands. Less than 4 percent had never married, and most of these were under 25 years of age. One-fourth of the population was Catholic. Women in the surveyed population were generally well educated, having completed, on the average, 12.8 years of school. Seventeen percent were college graduates, and an addi- tional 24 percent had attended college. The husbands of married women were most often skilled blue-collar or white-collar workers. Twelve percent were of a profession requiring a graduate degree or were holding high-level administrative positions. Less than 7 percent were unskilled laborers or service personnel. The median family income for married women in 1967 was between $10,000 and $11,000. Only 2 percent of married women had a family income of $5,000 or less, and nearly half of these were under 25 years of age. The oral contraceptives were widely used by women in this population. In February 1968, 50 percent had used or were using an oral contraceptive and 26 percent were current users. As would be expected, use of the oral contraceptives was found to be in- versely related to age. Nearly 80 percent of women between 20 and 24 years of age, but only about 25 percent of those between 45 and 54 years, had been exposed to the contraceptive drugs. Forty-seven percent of the younger women were current users, compared with 14 percent among the older women. A more detailed description of use of fertility control methods including OC’s is given in chapter 3. SELECTION AND RECRUITMENT OF THE COHORT OF STUDY SUBJECTS Women born 1914 and after (the minimum age at entry was 18 years), who were members of the Kaiser Foundation Health Plan and living in the specified Health Plan service area, were considered eligible for participation in the study. They became study subjects by having an examination in the AML designed to fulfill the special purposes of this study. Starting in mid-December 1968, and continuing through February 1972, participants came to the AML either by self-referral or physician referral. The self- referral subjeets were those women who met the eligibility require- ments for entry into the study and who requested a routine medical or gynecological examination. These women were encouraged by the appointment clerks to see their physicians in the AML and thereby become study subjects. Physician-referral subjects were those women meeting eligibility requirements who were already TABLE 1.—Subjects having their first Automated Multitest Laboratory examination during the period December 1968 through February 1972, by age and use of oral contraceptives Never Past Current Estrogen Postpartum Status Total users users users users and pregnant unknown Age in i years / Num- | Per- | Num- | Per- | Num- | Per- | Num- | Per- | Num- | Per- | Num- | Per- | Num- | Per- - ber cent ber cent ber cent ber cent ber cent ber cent ber cent Under 20________ 272 |. 5.13 76 | 1.59 166 | 3.83 3] 0.18 108 | 8.61 38 | 5.65 663 3.68 20-24. ......... 291 | 5.49 590 | 12.385 888 | 20.50 18 | 1.07 455 | 36.25 105 | 15.63 | 2,347 | 13.03 25-29 _.._. 276 | 5.21 (1,123 | 23.51 | 1,034 | 23.87 37 | 2.20 453 | 86.10 158 | 22.77 | 3,076 | 17.07 30-84... conwusmns 632 | 11.93 | 1,040 | 21.77 751 | 17.34 59 3.50 181 | 14.42 117 | 17.41 | 2,780 15.43 8589. cnn 952 | 17.97 820 | 17.17 550 | 12.70 93 | 5.52 50 | 38.98 85 | 12.65 | 2,550 | 14.15 40-44___________ 1,144 | 21.59 609 | 12.75 494 | 11.41 258 | 15.32 7 .56 78 | 11.61 | 2,590 | 14.37 45-49___________ 1,003 | 18.93 375 | 7.85 834! 7.711 520 | 30.88 1 .08 55 | 8.18 | 2,288 | 12.70 50-54___________ 668 | 12.61 140 2.93 113 2.61 630 | 37.41 0 0 32 4.76 | 1,583 8.79 Over 65_________ 61 1.15 4 .08 1 .02 66 3.92 0 0 9 1.34 141 .78 Total... 5,299 (100.00 | 4,777 {100.00 | 4,331 {100.00 | 1,684 {100.00 | 1,255 (100.00 672 (100.00 [18,018 | 100.00 Mean age____ 039.17 0 | 33.26 0] 31.64 0 | 46.72 0 | 25.54 032.74 0| 385.31 SHAILdADVILNOD TVI0 J0 SLOTIAHT HAILAIDVILNODNON g 6 STUDY, POPULATION, PLAN, AND PROCEDURES under medical care. They either were advised to have an AML examination at the time of their next annual checkup or were referred directly to the AML for a more thorough current evalua- tion. In addition, some women who obtained their first prescrip- tion for an oral contraceptive (new OC users) were referred to the AML before they started taking the drug. The choice of con- traceptive, oral or other, was not dictated by the Contraceptive Drug Study. Women took a particular contraceptive because of personal preference or physician recommendation. Table 1 gives the distribution by age and OC status of the women who had their first AML examination during the period December 1968 through February 1972. The subjects were char- acterized by use of oral contraceptive drugs at entry into the study as follows: 1. Never users: Women who had never used OC’s. 2. Past users: Women who had taken one or more oral con- traceptive pills 1 month or more prior to their AML examination. 3. Current users: Women who had taken one or more oral contraceptive pills within 30 days of their AML ex- amination. Women who fell into the following two categories were excluded from the previous three: 4. Pregnant: Women who were known to be pregnant at the time of their AML examination. Postpartum: Women who were not currently using either OC’s or estrogens and were postpartum at the time of their AML examination. 5. Estrogen users: Women who were taking estrogens (other than OC’s) at the time of their examination or had taken them within the past year. It was not possible to identify separately these two groups with the information avail- able. However, the plan is to include in this study the effects of estrogens, and information regarding their use will be obtained. As table 1 shows, there were approximately equal numbers of women in the never, past, and current OC user groups. The number of women in the entire cohort with unknown status at entry was 3.7 percent. Most of these were in the year 1971-72. The reason for the lack of information in that period is being clarified. Since the age distribution for the unknown group was NONCONTRACEPTIVE EFFECTS OF ORAL CONTRACEPTIVES 7 similar to that for the entire cohort, it seems unlikely that there was a bias related to the information deficit. Table 2 shows the relative frequency among current OC users of the type, dose, and pill formulation (combination or sequential) in use at the time of entry into the study. A breakdown of estrogen dose by year of use (not shown) showed that there was a decrease in the use of OC’s containing 0.10 mg mestranol over the period December 1968 through February 1972. Among women taking combination-type OC’s with mestranol during the period December 1968 through December 1969, 59 percent used OC’s containing 0.10 mg estrogen, while during the period January 1971 through February 1972, only 27 percent used the higher dose. AUTOMATED MULTITEST LABORATORY EXAMINATION The AML was the means by which a woman entered the Contraceptive Drug Study. The examination consisted of three parts. Laboratory procedures were performed and questionnaires answered at the initial visit. This was followed by a gynecologic examination within 10 days. When all data and laboratory testing were completed satisfactorily, the patient returned for an internist examination, which included a review of all information. Abnormal studies and appropriate followup procedures related to the Con- traceptive Drug Study, such as glucose tolerance testing, were conducted before the visit to the internist. In this manner, con- sistency in the use of followup procedures was insured. All data were entered on standardized machine-readable forms or cards. They were entered into the patient computer medical record sys- tem at the Kaiser-Permanente Medical Center, Oakland, from which were generated printed summary reports of medical history and test results. In an occasional instance, the usual sequence of examination as described above was reversed and a patient might see her gynecologist or internist first. These physicians might then refer the subject to the AML. Laboratory procedures were similar to other Kaiser-Perma- nente automated multiphasic examinations and have been described in detail in chapter 4 and elsewhere.'® *¢ These included measure- ments of visual acuity, intraocular pressure, Achilles-tendon-reflex relaxation time, pulmonary function, hearing, blood pressure, and pulse rate. A 12-lead electrocardiogram, chest X-ray, and anthro- pometric measurements were obtained. Blood studies included a 1-hr oral glucose tolerance test, a battery of serum chemical TABLE 2.—Relative frequency of current users having their first Automated Multitest Laboratory examination December 1968 through February 1972 taking various compositions of oral contraceptives Type of formulation Type and dose of estrogen and progestin (mg/tablet) Combination Sequential Total Number Percent Number Percent Number Percent Mestranol: 0.05 Norethindrone 1.0____________________________ 751 19.46 0 0 751 17.35 .06 Norethindrone 10.0. ___________________________ 3 .08 0 0 3 .07 .075 Norethynodrel 5.0_____________________________ 73 1.89 0 0 73 1.69 .08 Norethindrone 1.0_____________________________ 477 12.36 0 0 477 11.02 .08 Norethindrone 2.0... ccc memmmm ees mmm 0 0 114 24.31 114 2.63 .08 Chlormadinone acetate 2.0_____________________ 0 0 259 55.22 259 5.98 .10 Norethindrone 2.0_____________________________ 833 21.58 0 0 833 19.24 .10 Norethynodrel 2.5_____________________________ 251 6.50 0 0 251 5.80 10 Ethynodiol diacetate 1.0____________ I ero 698 18.08 0 0 698 16.12 .15 Norethynodrel 10.0... ina 11 .28 0 0 11 .25 OVEIAN...c orm ee Sree EE 3,097 80.23 373 79.53 3,470 80.16 Ethinyl estradiol: 0.05 Norethindrone acetate 1.0______________________ wi 4.59 0 0 177 4.09 .05 Norethindrone acetate 2.5_ _____________________ 166 4.30 0 0 166 3.83 .05 NOIZESUIO]l Bich cn com a mn Bm hmmm 352 9.12 0 0 352 8.13 .05 Medroxyprogesterone acetate 10.0_______________ 37 .96 0 0 37 .85 .05 Ethynodiol diacetate 1.0.______________________ 31 .80 0 0 31 .72 .10 Dimethistorone 25.0. & i cae oc niinnmn 0 0 96 20.47 96 2.22 SHYNAFO0Yd ANV ‘NVI ‘NOILVINdOd ‘XANLS Overall SR SS ae aio a om Pe pm or er All estrogens and progestins 763 19.77 96 20.47 859 19.84 3,860 100.00 469 100.00 4,329 100.00 SHALLAIDVALNOD TVYI0 J0 SLOFAIAH HAILJADOVILNODNON 6 10 STUDY, POPULATION, PLAN, AND PROCEDURES determinations, a blood count, and ABO blood typing. In addition, specific tests pertinent to studying contraceptive drugs were administered: urine culture, urinalysis, Pap smear, latex titer, protein electrophoresis, and thromboelastography. The health questionnaire covered medical, gynecological, obstetrical, social, and family histories; personal health practices; current symptoms; use of both prescription and nonprescription drugs; and methods of contraception. For OC users, the brand name and dosage were recorded, as well as the duration of use. Psychological assessment was obtained by use of the Minnesota Multiphasic Personality Inventory.” Subjects who were starting on oral contraceptives—that is, new OC users—were given an AML examination to establish base- line measurements. They were then asked to return in 3 months, at which time the laboratory studies were repeated and a shortened questionnaire was answered. Physician examinations were per- formed if warranted. Quality control reports were kept daily for almost every phase of the AML. The data were keypunched onto cards from which summary statistics were produced, comprising the overall and monthly mean, median, standard deviation, 2.5th percentile, and 97.5th percentile values. COHORT FOLLOWUP REPEAT AML EXAMINATIONS During the period April 1971 through May 1972, study sub- jects born 1925 and after who were still members of the Kaiser Foundation Health Plan and residing in the Walnut Creek Service Area were invited in for repeat AML examinations. The objective was to obtain repeat biochemical and physiological measurements that might demonstrate changes since their first examination. There was no difference in the percent frequency of number of AML visits and in the mean number of examinations between OC users and nonusers, 18-44 years of age at entry into the study in 1968 through 1970. About 27 percent of each group of women did not make a return AML visit during the period 1969-71. The data suggest, therefore, that OC users were not different from nonusers in their use of preventive health services. Starting in October 1972, women born before 1925 have been recalled for repeat examinations. NONCONTRACEPTIVE EFFECTS OF ORAL CONTRACEPTIVES 11 INTERIM QUESTIONNAIRE To obtain ongoing consistent information about changes in health status and in use of OC’s and estrogens from cohort sub- jects who come in for repeat AML examinations as well as from those who do not return, a questionnaire was designed that could be either self-administered or administered over the telephone. Subjects are queried as to whether they have used medical services at Kaiser-Permanente facilities other than Walnut Creek, or at non-Kaiser facilities, and if so, the reason for seeking medical care. Preliminary results from this questionnaire indicated that about 10 percent of the women used non-Kaiser services (inpatient and outpatient) during the preceding year. REVIEW OF INPATIENT AND OUTPATIENT MEDICAL RECORDS All visits, inpatient or outpatient, made by study subjects to Kaiser-Permanente medical facilities were recorded in their medi- cal charts and filed according to their Health Plan member num- ber. Information about all medical conditions for which the study subject sought care at a Kaiser-Permanente facility is therefore available. The medical charts of all subjects are currently being reviewed by trained medical-record technicians to determine the prevalence and incidence of selected diseases and the relationship to use of oral contraceptives or other estrogens. For certain con- ditions thought to be related to OC’s, the course and prognosis in OC users are compared to nonusers. LOSSES FROM THE STUDY Data regarding losses are given only for women born 1925 and after, since the information is not yet complete for women born before that year. A total of about 13,000 women (born 1925-54) were admitted into the study during the period December 1968 through December 1971. By the end of December 1971, about 86 percent of them were still members of the Kaiser Foundation Health Plan of Northern California. Four percent had moved out of the Walnut Creek Serv- ice Area, leaving 82 percent who would be likely to obtain most of their medical services at the Walnut Creek Kaiser-Permanente Medical Center. However, since a unique Health Plan membership number is used to identify an individual member throughout the TABLE 3.— Number and percent of Coniraceptive Drug Study cohort remaining in the Health Plan, Walnut Creek Service Area, at the end of 1971, by year of entry into the study and year of birth Year of entry 1969 1970 1971 Year of birth Remaining at end of Remaining at end of Remaining at end of Number 3 years Number 2 years Number 1 year at at at entry entry entry Number Percent Number Percent Number Percent 3995-29. ic. cea 1,176 1,033 87.8 877 762 86.9 464 451 97.2 YO30-24 . ii daia sii 1,175 1,018 86.6 793 713 89.9 483 470 97.3 1938-39... inant din 1,235 1,024 82.9 789 687 87.1 533 517 97.0 NOLO-AE oe iim mies si ws 1,380 1,114 80.7 840 725 86.3 702 681 97.0 YO4B-49. neal 902 679 75.3 611 501 82.8 595 579 97.3 I050-84......coe veins 85 45 52.9 250 194 77.6 485 466 96.1 Total. une niamsas 5,953 4,913 82.5 4,160 3,582 86.1 3,262 3,164 97.0 [4¢ SHINAIAO0Ud ANV ‘NVId ‘NOILVINGOd ‘XdNLs NONCONTRACEPTIVE EFFECTS OF ORAL CONTRACEPTIVES 13 northern California region, it is possible to access medical records for services obtained by CDS patients at Kaiser-Permanente Medi- cal Centers outside Walnut Creek. Thirteen percent of the cohort might therefore be con- sidered as being lost to followup. These consisted of the following categories: 1. Terminated from the Health Plan: 5 percent 2. Membership in the Health Plan, but outside the northern California region: 2 percent 3. Followup status unknown: 4 percent 4. Adamant AML refusals: 2 percent There was no difference among the never, past, current, and estrogen users by followup status at the end of 1971. This suggests that losses from the study were not related to the use of OC’s. Table 3 gives the number and percent of the CDS cohort born 1925-54 remaining in the Health Plan Walnut Creek Service Area at the end of 1971 by year of entry into the study and year of birth. The percent of women remaining after 2 to 3 years is shown for the 1969 entrants into the study, that remaining after 1 to 2 years is shown for the 1970 entrants, and that remaining 1 year or less is shown for those admitted in 1971. As the table shows, the losses from the study through termination from the Health Plan is inversely related to age. The data suggest a continuing 3- to 4-percent attrition rate per year for the older women. DATA PROCESSING FOR RESEARCH ANALYSES All data were ultimately encoded and punched onto cards for storage on magnetic tape. A proprietary tape data file software package® that allowed for accumulation of added information and repeat visits was used. This data base was in character format. It could be accessed for retrieval through its special programs or directly by user routines if desired. The tape itself had a first section containing dictionary information on all variables, fol- lowed by the patient data in Patient Record Number order. Each patient’s data consisted of two or more physical records. First, a short header record containing the Medical Record Number and the number of visits, then a variable number of fixed-length visit records. Since the dictionary structure allowed the user to refer to variables independently of where they appeared in the visit record, it was relatively simple to add new variables as well as a “Medfile” from Synapse, Inc., Oakland, Calif. 14 STUDY, POPULATION, PLAN, AND PROCEDURES new visits during the periodic data loads. A remote job-entry terminal was used for submitting jobs and receiving printed out- puts by telephone line to an IBM 370/155 system at Kaiser- Permanente Medical Center, Oakland, Calif. DATA ANALYSIS The first phase of the analysis was cross-sectional and attempted to answer the following questions: 1. Do OC users differ from nonusers with regard to the measurement in question? 2. Could the observed difference between OC users and non- users be due to differences between the two groups in other related variables rather than to OC use? 3. Are past OC users similar to or different from never users with regard to the measurement in question? 4. Is there variation among OC users by steroid formulation? 5. Does the difference between OC users and nonusers in- crease with duration of OC use? The second phase of the analysis was longitudinal and com- plemented the cross-sectional analysis. That is, if differences were found cross sectionally between users and nonusers, then women who changed from nonusers to OC users from one AML examina- tion to the next should show the expected differences in the two measurements. Similarly, if the cross-sectional comparison be- tween past users and never users showed no difference, the users who went off OC’s between consecutive AML examinations would provide further evidence that the OC effect was reversible. Two-way frequency tabulations and multiple stepwise regres- sion analyses have been applied to the data analyses. RESULTS The following is a summary of some of the salient findings based on information obtained at AML examinations during the period December 1968 through December 1971. This summary represents only a brief review of the published results and other reports in preparation for publication. Some analyses are pre- liminary in nature. Findings related to the heart and blood vessels and glucose tolerance as determined from multitest laboratory results: NONCONTRACEPTIVE EFFECTS OF ORAL CONTRACEPTIVES 15 Blood pressure.—Blood pressure among current users of OC’s was slightly but significantly higher than that among nonusers at every age level ; the difference increased with age and averaged 5 mm Hg systolic blood pressure and 1.3 mm Hg for diastolic. Weight did not account for the difference. Among women with systolic pressure greater than 140 mm Hg or diastolic greater than 90 mm Hg or both, the proportion of OC users was higher than among women with lower blood pressure measurements. There was no relationship to duration of use or dose of drug.'® Pulse rate.—Current users had a higher mean pulse rate than nonusers, and the difference appeared to be greater among the younger age groups. No relationship to dose or duration of OC use was found. A significant decrease in pulse rate occurred among those women who stopped taking OC’s by the time of a repeat AMI, examination, 1 to 2 years later. This analysis was based on values which were adjusted for variation due to examination date. The post-OC values made the former users similar in pulse rate to the nonusers. The longitudinal data therefore confirmed the cross-sectional finding. Cholesterol.—Among women over 40 years of age, the mean level of serum cholesterol was about 8 percent lower in OC users compared to never users, and the difference increased with age. On the other hand, among women under 40 years, the level of serum cholesterol was about 3 percent higher in OC users com- pared to never users. This difference was inversely related to age. One-hour glucose tolerance.—OC users were found to have 1-hr serum glucose values 11 mg/100 ml higher, on the average, than had nonusers. This difference was independent of dose and duration of drug use. The hyperglycemic effect of OC’s was found to be additive to the effect which age, obesity, and a family history of diabetes had on glucose tolerance. Factors affecting the 1-hr glucose tolerance were also studied and reported. Diseases of heart and blood vessels based on physician diagnoses: Hypertension.—The overall prevalence rate for hypertension among all women in the study was 17 percent. For each of the 5-year age groups from 25 to 50 years, the rate for current users © was between 20 and 50 percent higher than that for never users. This difference did not exist between past and never users. Cases of hypertension that were diagnosed only on a repeat but not on the initial AML examination could be considered newly developed cases. So defined, the incidence rate for current users exceeded that for never users for every 5-year age group from 25-55 years, 16 STUDY, POPULATION, PLAN, AND PROCEDURES being almost eightfold at age 50-54 years. The rate among past users was not higher than that for never users. Tachycardia.—Current OC users had higher prevalence rates . for tachycardia than did never users. Coronary heart disease, myocardial infarction, and angina pectoris.—The number of cases were too few to permit valid con- clusions to be drawn. Findings related to the blood as determined from multitest laboratory results: Thromboelastography.—Current users had shorter clotting times and firmer clots, though few subjects demonstrated extreme thromboelastographic hypercoagulability. These changes were not related to a specific brand, component, or dose of OC.2! ABO blood types.—ABO blood types were not related to OC user status, but subjects with type AB had the lowest coagulability as determined by thromboelastography compared to A, B, and O. Women with blood type A formed clots faster by thromboelastog- raphy than did subjects with other blood types; this tendency was most pronounced among current OC users. Red blood cell indices.—After adjustment for age and men- strual blood loss, OC users compared to nonusers showed a slight but statistically significant elevation in RBC indices, together with a decrease in hemoglobin, hematocrit, and red blood cell count.2? This finding was consistent with folic acid and/or vitamin B-12 deficiency. White blood cell count.—The mean white blood cell count for users of OC was 16 percent higher than that for nonusers, but was less than that for pregnant women, in whom the increase was 32 percent over that for nonusers. The elevation among OC users increased with dose and duration of use as determined by cross- sectional analysis. Serum electrophoretic fractions.—Among OC users compared - with nonusers, mean percent albumin was 5 percent lower and gamma globulin was 2 percent lower; alpha-1 was 29 percent higher, alpha-2 was 13 percent higher, and beta globulin was 12 percent higher. Total serum protein was lower among users; but the amount in grams of globulin represented by the percent of serum electrophoretic fraction was higher than that among non- users, while the amount of albumin was less. Rheumatoid factor.—The prevalence of rheumatoid factor (positive by latex agglutination at 1:20 dilution) was lower among current OC users (1.8 percent) than among past users (2.6 per- NONCONTRACEPTIVE EFFECTS OF ORAL CONTRACEPTIVES 17 cent) or never users (2.4 percent). These data are being analyzed again with a positive test defined at 1:40 dilution. Miscellaneous AML findings: Asymptomatic bacteriuria.—From a preliminary analysis of data, the prevalence of significant bacteriuria, defined as two consecutive positive (over 100,000 colonies per ml) urine cul- tures, was slightly higher among current OC users (1.7 percent) compared to past users and never users (1.3 and 1.5 percent, respectively). Audiometry.—In all 5-year age groups but the youngest, the mean hearing level of current users of OC’s was better than that for never users. The difference was most pronounced at the lower frequencies and was of the order of 0.5 to 3 dB. The means for past users tended to fall between those for the other groups. Achilles-tendon-reflex time.—The mean time (age-adjusted) for both current and past users of OC’s was 5 ms longer than for never users. This difference was statistically significant, but its clinical implication remains obscure. Height and weight—Women who had never used OC’s appeared to be slightly shorter and possibly heavier than current or past users. Such a difference may be construed as possibly characterizing persons who select themselves into the OC groups. Some negative AML findings: Spirometric measurements.—These were unaffected by past or present OC use. Age, height, weight, and smoking habits were taken into consideration.2? Depression.—As measured by T scores obtained from re- sponses to the Minnesota Multiphasic Personality Inventory,” no difference was observed among current, past, and never users of 0OC’s.2¢ 25 DISCUSSION Small but statistically significant differences were found be- tween OC users and nonusers for many of the variables studied. With the use of large sample sizes such as these (numbering in the thousands), it is possible to detect such small differences. When multiple comparisons are made in the same data, one would expect to find by chance some that are significantly different.?¢ Most of these differences are of little or no apparent clinical importance; however, the effect of small physiological or bio- chemical changes maintained over a period of many years as a result of continuous or cumulative exposure to OC’s remains to be assessed. 18 STUDY, POPULATION, PLAN, AND PROCEDURES Such statistical association is not the only indication of the relationship between OC’s and certain biological effects that this study can provide. Longitudinal analysis of results obtained on the same group of women who change from being current users to past users, or from nonusers to current users, will help to clarify the relationship between OC’s and observed changes that might be considered attributable to OC’s. See, for example, the results for pulse rate described earlier in this report. The automated laboratory test bettery was a useful way of surveying the wide spectrum of expected and unexpected side effects of OC’s. In addition, the strength of the AML measure- ments was that they were made without knowledge of the subject’s use of contraceptives, and that they were subject to a high degree of quality control. Furthermore, they could detect effects asso- ciated with OC use that were no clinically observable but that might be precursors of clinical disease. Results concerning OC side effects, which have been obtained by the Contraceptive Drug Study, not only can provide support to previously reported findings, but can furnish confirmatory evidence where reported findings based on small samples are contradictory. On the other hand, systematic and reliable docu- mentation of negligible effects or negative results is an important contribution. It will help allay the fears not only of those women who use oral contraceptives, but also of physicians who prescribe them. REFERENCES 1. Advisory Committee on Obstetrics and Gynecology, Food and Drug Ad- ministration. Second Report on the Oral Contraceptives. Washington, D.C.: U.S. Government Printing Office, Aug. 1, 1969. 2. SALHANICK, H. A, KipN1s, D. M., and VANDE WIELE, R. L., Eds.: Metabolic Effects of Gonadal Hormones and Contraceptive Steroids. New York: Plenum Press, 1969. 3. SARTWELL, P. E., Masi, A. T., ARTHES, F. G., GREENE, G. R., and SMITH, H. E.: Thromboembolism and oral contraceptives: An epidemiologic case-control study. Am J Epidemiol 90 (5) : 365-380, Nov. 1969. 4. VEssEy, M. P., and DoLL, R.: Investigation of the relation between use of oral contraceptives and thromboembolic disease. A further report. Br Med J 2: 651-657, June 14, 1969. 5. WYNN, V,, and Doar, J. W. H.: Longitudinal studies of the effects of oral contraceptive therapy on plasma glucose, non-esterified fatty acid, insulin and blood pyruvate levels during oral and intravenous glucose tolerance tests. In H. A. Salhanick, D. M. Kipnis, and R. L. Vande Wiele, Eds.: Metabolic Effects of Gonadal Hormones and Contraceptive Steroids. New York: Plenum Press, 1969. Pp. 157-171. 6. DoE, R. P., MELLINGER, G. T., SwAiM, W. R., and SEAL, U. S.: Estrogen 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. NONCONTRACEPTIVE EFFECTS OF ORAL CONTRACEPTIVES 19 dosage effects on serum proteins: A longitudinal study. J Clin Endo- crinol Metab 27: 1081-1086, Aug. 1967. Woops, J. W.: Oral contraceptives and hypertension. Lancet 2: 653- 654, Sept. 23, 1967. VEssgy, M. P., Dorr, R., FAIRBAIRN, A. S., and GLOBER, G.: Postoperative thromboembolism and the use of oral contraceptives. Br Med J 3: 123- 126, July 1970. GREENE, G. R., and SARTWELL, P. E.: Oral contraceptive use in patients with thromboembolism following surgery, trauma or infection. Am J Public Health 62: 680-685, May 1972. Jick, H., WESTERHOLM, B., VEssEy, M. P., Lewis, G. P., SLoNE, D,, INMAN, W. H. W., SHAPIRO, S., and WORCESTER, J.: Venous thrombo- embolic disease and ABO blood type. Lancet 1: 539-542, Mar. 15, 1969. VESsSEY, M. P., Dorr, R., and SUTTON, P. M.: Investigation of the possible relationship between oral contraceptives and benign and malignant breast disease. Cancer 28: 1395-1399, Dec. 1971. DriLL, V. A., and CALHOUN, D. W.: Oral contraceptives and thrombo- embolic disease. JAMA 206: 77-84, Sept. 30, 1968. DorL, R., INMAN, W. H. W., and VESSEYy, M. P.: Thromboembolic dis- orders and oral contraceptives. JAMA 207: 1150-1151, Feb. 10, 1969. SCHROGIE, J. J.: Facilities available for surveillance of the side effects of fertility regulating agents in the United States. Contraception 6(4): 401-409, Dec. 1971. COLLEN, M. F.: The multitest laboratory in health care of the future. Hospitals 41: 119-125, May 1, 19617. CoLLEN, M. F.: Automated multiphasic screening. In C. Sharp and H. Keen, Eds.: Presymptomatic Detection and Early Diagnosis. London: Pitman Medical Publ. Co., 1968. Ch. 2, pp. 25-66. HatHAWAY, S. R.,, and McKINLEY, J. C.: Minnesota Multiphasic Per- sonality Inventory. New York: The Psychological Corp., 1967. FiscH, I. R., FREEDMAN, S. H., and MYATT, A. V.: Oral contraceptives, pregnancy and blood pressure. JAMA 222 (12): 1507-1510, Dec. 18, 1972. PHILLIPS, N., and DUFFY, T.: One-hour glucose tolerance in relation to use of oral contraceptive drugs. Am J Obstet Gynecol 116 (1) : 91-100, May 1, 1973. PHILLIPS, N., and DUFFY, T.: Effect of age, hours since last food, time of day, and ketonuria on one-hour glucose tolerance. Health Serv Rep 87(7): 649-657, Aug.—Sept. 1972. FiscH, I. R., FREEDMAN, S. H., and PELLEGRIN, F. A.: Effect of oral contraceptives on the thromboelastogram. Clin Pharmacol Ther 14(2) : 238-244, Mar.—Apr. 1973. FiscH, I. R.,, and FREEDMAN, S. H.: Oral contraceptives and the red blood cell. Clin Pharmacol Ther 14(2) : 245-249, Mar.—Apr. 1973. FREEDMAN, S. H., and ANDERSON, N. E.: Spirometry and oral contra- ceptives. Am J Obstet Gynecol 116 (5) : 682-688, July 1, 1973. KUTNER, S. J., and BRowN, W. L.: Types of oral contraceptives, depres- sion and premenstrual symptoms. J Nerv Ment Dis 155(3) : 153-162, Sept. 1972. KUTNER, S. J., and BRowN, W, L.: History of depression as a risk factor for depression with oral contraceptives and discontinuance. J Nerv Ment Dis 155(3) : 163-169, Sept. 1972. MAINLAND, D.: Elementary Medical Statistics, 2d ed. Philadelphia: W. B. Saunders Co., 1963. Pp. 102-105. as 3 a W i 5 «i ie ot el bal I : = = fa J f i, ) np pt a “ull i n aL J, wi y En . ay 2 ' A + = Lea - i} ol u re, = i A Lal i ~ 3 A ae Chapter 2 Socioeconomic Characteristics and Methods of Fertility Control of Female Subscribers to the Kaiser Foundation Health Plan in the Walnut Creek Service Area, 1968 Nancy R. PHILLIPS A sample survey of female subscribers to the Kaiser Foundation Health Plan living in the Walnut Creek Service Area was done in 1968. The sample comprised 3,886 women born between 1913 and 1947. Self- administered mailed questionnaires, followed by telephone contact if nec- essary, achieved responses from 92 percent of those contacted. Walnut Creek Service Area-Health Plan population.—The Walnut Creek Service Area is composed of several suburban communities located within commuting distance of San Francisco, Calif. The female members of the Health Plan in this area are predominantly white, middle class, and married. Use of the oral contraceptive drugs.—Fifty percent of the women had used oral contraceptives (OC’s). Twenty-six percent were current users. Oral contraceptive use was inversely related to age. After their introduction in 1960, oral contraceptives were diffused rapidly throughout the sample. Women who had taken OC’s for contraception had accumu- lated on the average between 2 and 8 years of exposure. The majority of medical users had less than 12 months’ exposure. Time trend in contraceptive drug use among white married women. —The proportion of current users increased rapidly annually from 1960 through 1965; after 1965 the annual rate of increase declined. Comparison with the 1965 National Fertility Study.—Use of oral contraception was more widespread in the Kaiser Study population (26 percent) than in a comparable age, race, and marital-status group in the 1965 National Fertility Study (16 percent). * The section “Prevalence of Surgical Sterilization Among White Married Women and Their Husbands” is a revision of the following article, published here by permission of the editors of Demography: Phillips, N. R.: The prevalence of surgical sterilization in a suburban population. Demography 8: 261-270, 1971. 21 22 STUDY, POPULATION, PLAN, AND PROCEDURES Discontinuation of oral contraception.—Fifteen percent of white married women were past users of oral contraceptives. Unpleasant side- effects were given as a reason for OC discontinuance by 45 percent. Twenty-four percent stopped to conceive. Only 2 percent discontinued on the advice of a physician. Of those who discontinued OC’s, half had used them longer than 12 months and fewer than 15 percent had used them for less than three cycles. Oral contraceptive use in relation to parity, education, and religion among white married women.—Oral contraceptive use varied little by parity. The relative frequency of ever users of OC’s increased with the amount of education the woman had had. This was true of all age groups. The differences between Catholics and non-Catholics in the use of OC's are small, especially in the younger age groups. Methods of contraception used by white married women at risk of pregnancy.—Of those women at risk of pregnancy, 47 percent were taking OC’s for the purpose of contraception and 3 percent were taking OC’s for medical purposes. Oral contraceptives were the most popular method of contraception followed by the condom and the diaphragm. Intrauterine contraceptive device use was infrequent (less than 4 per- cent). Current OC users’ most common prior method of contraception was the diaphragm (34 percent) for almost all age groups. The dia- phragm and the condom were used equally by former OC users and never OC users, whereas the intrauterine device and spermicidal agents were more commonly used by past OC users than never OC users. Never users employed less effective methods than past OC users. Prevalence of surgical sterilization among white married women and their husbands.—Contraceptive operations, more than two-thirds of which were male vasectomies, were found among 23 percent of the couples. Twelve percent of the wives had had remedial operations, usually a hysterectomy. Some form of sterilizing surgery (contraceptive or re- medial) was found among 31 percent of the couples. Comparison with the 1965 National Fertility Study.—The prevalence of vasectomies in this population is considerably higher compared with that found across the Nation or in the Western States in the 1965 National Fertility Study. The prevalence of tubal ligations is slightly higher. Remedial sterilization operation prevalence was somewhat below the national average. Social variables associated with contraceptive operations.—There is a trend for prevalence of contraceptive operations to increase with parity of the wife but not beyond the fourth birth. The modal number of children of voluntarily sterile couples is two or three, depending on the wife’s age. The prevalence of contraceptive operations is lowest when both husband and wife are Catholic (15 percent) and highest (26 per- cent) when neither is Catholic. Contraceptive operation prevalence, both vasectomies and tubal ligations, varies inversely with the education of the wife. The less education a husband has, the more likely he is to have had a vasectomy. The socioeconomic characteristics of female subscribers to the Kaiser Foundation Health Plan living in the Walnut Creek Service SOCIOECONOMIC CHARACTERISTICS AND METHODS 23 TABLE 4.—Age distribution of study population and survey sample q y Study population Survey sample g ampling Year of birth fonction Number | Percent Number Percent 1948-47. ee. 1/8 2,510 12.6 836 21.5 1988-42. _._. 1/4 2,713 13.6 668 17.2 1983-37. eee 1/5 2,615 13.1 517 18.3 1028-32. . cincernnen 1/6 3,104 15.5 522 13.4 1928-2 mim 1/6 3,431 17.2 568 14.6 1918-22. een 1/6 3,219 16.1 529 13.6 1918-17... cena 1/10 2,385 11.9 246 6.3 AIBZO8. sco mmimem min 19,977 100.0 3,886 100.0 Area and their use of contraceptive drugs, from the introduction of such agents through February 1968, and of other methods of fertility control, are described here on the basis of data collected in a sample survey of women aged 20-54 years carried out between February 10 and April 28, 1968. The survey was based on an age-stratified probability sample of 3,886 women selected from the Kaiser Foundation Health Plan membership list of approximately 20,000 women born in the years 1913 through 1947 and living in the Walnut Creek Service Area. The representation of each 5-year age group in the survey sample was not equal to its proportion in the population (table 4). The sample fraction of each age group was determined by three considerations. First, we wished to estimate the age-specific proportion ever exposed to the contraceptive drugs with 95 per- cent confidence of being within 5 percent of the true proportion. This required a minimum of 400 subjects from each age group. Second, we had reason to believe that young married women were very frequently users of the contraceptive drugs, and we wished to assure that the sample contained an adequate number of ‘married nonusers for comparison with users. Larger numbers, therefore, were selected of women below 30 years of age. Third, we had learned on pretesting the questionnaire that special effort was needed to gain the participation of postmenopausal women, whose interest in contraception was less than that of younger women. To compensate for the extra followup that would be re- quired, the sample size of those 50-54 years of age was halved. A questionnaire (see appendix) was mailed to the subjects 24 STUDY, POPULATION, PLAN, AND PROCEDURES TABLE 5.—Cumulative response rate by number of probes Nn 1 probe 2 probes 3 probes 4 probes er Age in sam- | Num-| Per- | Num-| Per- | Num-| Per- | Num-| Per- ple ber | cent | ber | cent | ber | cent | ber | cent 20-2 ic ne 836 | 455 | 54.4 | 688 |82.3| 731 |87.4| 765| 91.5 25-29. conn 668 | 377 | 56.4 | 561 | 84.0 | 583 | 87.3 | 618 | 92.5 80-34...-----__| 517 | 301 | 58.2 | 432 | 83.6 | 462 | 89.4 | 482 | 93.2 35-39__________ 522 | 261 | 50.0 | 413 [ 79.1 | 451 | 86.4 | 477 | 91.4 40-44. _________ 568 | 278 | 48.9 | 457 | 80.5 | 495 | 87.1 | 519 | 91.4 45-49..." -.- 529 | 248 | 46.9 | 425 | 80.3 | 456 | 86.2 | 481 | 90.9 0-84. cana 246 | 101 | 41.1 | 202 | 82.1 | 217 | 88.2 | 224 | 91.0 All ages__|3,886 (2,021 | 52.0 (3,178 | 81.8 |3,395 | 87.4 (3,566 | 91.8 TABLE 6.—Cumulative proportion ever exposed to contraceptive drugs by number of probes Estimate Age 1 2 3 4 adjusted probe probes probes probes for nonre- spondents ! 20-24. .ueencnnans 0.842 0.808 0.802 0.795 | 20.785 25-80... crime .814 97 L794 .786 2, 783 30-34 ____._____. .678 .662 .645 .641 2, 637 85-8 cummin .475 .441 .450 .451 3.439 40-44 oo eeeieea .396 .389 .388 .395 .390 NS 45-49. . conn .335 .299 .298 .299 2 286 850-04. ccotaenminns .297 .228 212 .205 2,192 1 Estimate adjusted by least-square extrapolation of cumulative proportion ex- posed among cumulative proportion responding. 2 Regression is statistically significant at p<0.05. 3 Regression is statistically significant at p =0.05. NS =Not significant. for self-administration. When a subject did not return the ques- tionnaire within 2 weeks, a telephone call was made to explain the importance of her participation. If no answer was received, the call was followed after 3 to 4 weeks by a reminder letter. If again no response was elicited, a telephone interview to complete the questionnaire was attempted. Ninety-two percent of the sample subjects participated in the survey. Both age and exposure to the contraceptive drug affected SOCIOECONOMIC CHARACTERISTICS AND METHODS 25 100 I | T AGE 20-24 a 80 25.29 atm] [§¥] wv oa. X 60 — > 35390 | Ld = 40 40-44 omen yeni - A549 mesic a 20 0 1 | | | 0 20 40 60 80 100 PERCENT RESPONDING FIGURE 1.—Least-squares extrapolation of percent ever exposed to contra- ceptive drugs (white married female Health Plan subscribers, Walnut Creek Service Area). participation, but after four probes their effects were minimal. The cumulative response rate by age and number of probes is shown in table 5. Age-specific estimates: of the proportion of ever users were adjusted: by regression analysis for bias introduced by nonparticipants. Although statistically significant in all but-one age group, the magnitude of the difference between observed and adjusted proportions was, in all age groups, trivial (table 6 and fig. 1). All population estimates, therefore, were based on the proportions found among the survey participants. WALNUT CREEK SERVICE AREA- HEALTH PLAN POPULATION The Walnut Creek Service Area is composed of a number of suburban communities, all within commuting distance of San Francisco, spread throughout the valley and delta sections of 26 STUDY, POPULATION, PLAN, AND PROCEDURES eastern Contra Costa County, Calif. These communities include Walnut Creek, Concord, Pleasant Hill, Danville, Alamo, Lafayette, Orinda, Moraga, Martinez, Pittsburg, and Antioch. In the past 30 years, the area has undergone vast residential development, mostly in the form of single-family detached homes. What was largely open farm and orchard land in 1940 is now inhabited by more than 400,000 people. Nearly one-fourth of the area’s total population is covered by the Kaiser Foundation Health Plan. The majority of Health Plan members in the Walnut Creek Service Area are group sub- scribers; i.e., they have joined the Health Plan through their union or place of employment, but a sizable minority (20 percent) are individual subscribers. In addition to trade and industrial unions, groups represented in the Health Plan include Federal and State employees, the University of California, local school districts, and a variety of commercial concerns. Ways in which Health Plan members differ from the greater population are not known, but the diversity of the Health Plan membership would suggest that subscribers in the Walnut Creek Service Area are not much dif- ferent from their fellow suburbanites. Nonwhites constituted less than 3 percent of the surveyed population. Five percent were determined, through a computer program developed by R. W. Buechley, to have Spanish surnames. Eighty-nine percent of the women were married and living with their husbands. Fewer than 4 percent had never been married ; most of these were younger than 25 years. Sixty percent of the women were Protestants. One-fourth were Roman Catholics, a representation comparable to the national average. Members of the Church of the Latter-day Saints, who traditionally value large families but have no religious objection to contraception, were thought to be sufficiently numerous in this population to provide a special contrast group to Catholics. They were found, however, to constitute only 3 percent of the popula- tion. Approximately 1 percent were Jews. Seven percent had no religious affiliation. The Walnut Creek Kaiser Health Plan population is above the average of the Nation as a whole in education and income. Eighty-four percent of the women surveyed had completed high school ; 42 percent had attended college for at least 1 full year, and 17 percent were college graduates. The median family income of married women in 1967 was above $10,000; it was $5,000 or less in only 2 percent. Family incomes in excess of $15,000 were reported by one-fifth of wives. SOCIOECONOMIC CHARACTERISTICS AND METHODS 27 TABLE 7.—Use of contraceptive drugs by female members of Kaiser Foundation Health Plan, living in the Walnut Creek Service Area, February 1968 Ever users Current users Age (percent + (percent + standard error) | standard error) R0=2L....... comm tw wa 79.5+1.2 46.9+1.5 D520 nbs SE eg A 78.6+1.4 41.8+1.7 BOSSE 2 ivi imi msm iim i i RR 64.1+2.0 33.2+1.9 Bm i —————— EE 45.1+2.1 19.9+1.7 A0~AE. ... ore EE EE 39.5+2.0 20.2+1.6 BAY i ER AR 29.9+1.9 16.0+1.5 BOSE mp ima 20.5+2.6 10.7+2.0 LE 50.14+0.7 26.3+0.6 The occupational level of this population was not so high as these incomes suggest. Nearly half the husbands were blue-collar workers, but few (7 percent) were unskilled. About two-fifths of husbands were in white-collar occupations: salesmen, insur- ance agents, stockbrokers, office managers, surveyors, and school- teachers. The remaining 12 percent were members of a high-status profession, or held high-level administrative positions. Working wives contributed to the affluence of many families. Twenty-nine percent of married women worked full time, and an additional 14 percent were employed part time. USE OF THE ORAL CONTRACEPTIVE DRUGS As of February 1968, 50 percent of the women surveyed had used contraceptive drugs and 26 percent were current users (table 7). Use of the contraceptive drugs was inversely related to age. The relative frequency of ever users (past and current users com- bined) decreased from 80 percent at ages 20-24 years to 20 percent at 50-54 years. The proportion of current users decreased from 47 percent in the youngest age group to 11 percent in the oldest. The contraceptive drugs can be taken for purposes other than contraception, and may be used for different purposes at different times in an individual’s life. In this analysis, all data on the reason for taking the drug related to the purpose most recently served. Women who used the hormones simultaneously for contraception and as treatment for a gynecological problem were counted as con- traceptive drug users. Subjects termed ‘“noncontraceptive users” were those who denied that a contraceptive effect was intended 28 STUDY, POPULATION, PLAN, AND PROCEDURES TABLE 8.—Frequency of noncontraceptive users among oral contraceptive users Ever users Current users Past users Age Number | Percent | Number | Percent | Number | Percent 0D d iramiad 52 8.6 17 4.7 35 14.1 50 di 53 10.9 7 2.1 46 20.2 80:84: 1 63 20.4 13 8.1 50 33.6 85-99 1 oa 64 29.8 6 6.3 58 48.3 A0-Ad. Ci lsd 81 39.5 28 26.7 53 53.0 ABO Gs rie 84 58.3 33 42.8 51 76.1 BOB tiie 35 76.1 16 66.7 19 86.4 Overall percentage |. .__..___ 20.1 | conuine 1.4 {uaans 39.1 by their most recent use of these agents. Twenty-six percent of ever users stated that when they last took contraceptive drugs they did so solely for noncontraceptive purposes (table 8). Not surprisingly, the relative frequency of noncontraceptive users among ever users was related to age, increasing over the age range from 9 to 76 percent. Noncontraceptive users were also concen- trated among past users (table 8). Thirty-nine percent of past users, and only 14 percent of current users, took the hormones for reasons other than preventing pregnancy. This reflects the relative brevity of the periods during which most users took these hormones for noncontraceptive purposes (see section ‘“Accumu- lated Duration of Exposure”). Diffusion The diffusion of oral contraceptive drug use throughout this population (table 9) is calculated on the basis of calendars filled out by survey subjects to show month-by-month use of the drugs, beginning January 1960. In table 9 the birth years 1913-37 are grouped into 5-year intervals, but the 1938-47 birth cohorts are - shown individually. The diffusion of use in a cohort over time may be seen by reading down the columns. Generations may be com- pared with respect to the proportion ever exposed at various points in time by reading across the rows. The age-specific experience of individual cohorts may be compared by reading diagonally down- ward from left to right. The increase over time in the accumulated proportion of ever users among the 1913-17 through 1938-42 birth cohorts is illus- trated in figure 2. (The 1938-42 cohorts, shown individually in TABLE 9.—Accumulated proportion ever exposed to contraceptive drugs by the end of each calendar year, 1960-67, by year of birth Calendar year Year of birth 1913- | 1918 | 1923 | 1928— | 1933 | 1938 | 1939 | 1940 | 1941 | 1942 | 1943 | 1944 | 1945 | 1946 | 1947 17 | 22 | 27 {22 [| 37 1960. _____________. 0.009 [0.021 [0.012 [0.046 [0.041 [0.028 [0.034 [0.039 [0.009 [0.017 {0.006 |. ____ | ____ |__| _____ 1961 oo. 018 | .048 | .054 | .084 | .114 | .110 | .146 | .124 | .073 | .034 | .045 [0.028 [0.008 [0.007 | _____ 1962... 031] .060 | .092 | .132 | .189 | .214 | .276 | .248 | .218 | .161 | .151 | .078 | .039 | .029 |._____ 1968... 040 | .091 | .125 | .195 | .259 | .879 | .896 | .380 | .400 | .330 | .363 | .240 | .086 | .036 | 0.021 1964... 058 | .112 | .179 | .254 | .394 | .496 | .500 | .473 | .554 | .508 | .559 | .447 | .273 | .204 | .077 1965 oo. 089 | .146 | .248 | .342 | .492 | .586 | .612 | .605 | .682 | .695 | .682 | .620 | .484 | .525 | .232 1966. ooo. 165 | .212 | .318 | .394 | .570 | .676 | .690 | .698 | .736 | .788 | .743 | .721 | .688 | .664 | .430 1967 ooo. 201 | .293 | .387 | .444 | .637 | .759 | .741 | .767 | .827 | .814 | .800 | .844 | .820 | .781 | .627 Number___._____.| 224 | 481 | 519 | 477 | 482 | 145 | 116 | 129 | 110| 118 | 179 | 179 | 128 | 187 | 142 SCOHLIN ANV SOILSIYHLOVAVHD DINONODIOIDO0S 62 30 STUDY, POPULATION, PLAN, AND PROCEDURES 1.0 | I | I = BIRTH COHORT - 1. 1913-1917 bk 2. 1918-1922 © ’ 3. 1923-1927 6 5 4.1928.1932 . wd Le 5.1933-1937 = 0 6. 1938-1942 i 5 ; 5 Mos - oc ., [V§] > Lid ET \ o — 2 0 4. [— 0.4 — o 0 o | 5 3 a. ., Oo J oc : a 4 2 0.2 P p 7 _o] rr 2 6 . 0 : 0 ct —1 q . O a O — | | | 0 ! 1960 1961 1962 1963 1964 1965 1966 1967 CALENDAR YEAR FIGURE 2.—Proportion ever exposed to contraceptive drugs by end of each calendar year 1960-67, by year of birth (white married female Health Plan subscribers, Walnut Creek Service Area). table 9, are shown collectively in fig. 2.) At the time the first oral contraceptive drug was released for general use, these women ranged in age from 18 through 47 years. Until the end of 1965, the annual increase in the proportion of ever users tended to be greater in successively younger cohorts. After 1965, the increase among the younger cohorts dropped off somewhat while that among older women accelerated. Conversion from never user to SOCIOECONOMIC CHARACTERISTICS AND METHODS 31 | | ji | | if | i 0.15 BIRTH COHORT: 1913-1917 ~ 0.10 0.05 0.15 Se o 0.05 PROPORTION EVER USERS BIRTH COHORT: 1923-1927 0.20 CONTRACEPTION »* 0.05 oo” =] o-=="""" MEDICAL PURPOSES - —— - - - oLZrm i | | | | | 1 1960 1961 1962 1963 1964 1965 1966 1967 CALENDAR YEAR FIGURE 3.—Proportion ever users of birth control pills (white married female Health Plan subscribers, Walnut Creek Service Area). 32 STUDY, POPULATION, PLAN, AND PROCEDURES user so accelerated among the 1913-17 and 1918-22 cohorts that half or more of those ever exposed by the end of 1967 had first taken OC’s after 1965. This sudden increase in use among women moving out of the reproductive period probably reflected a trend to prescribe the oral contraceptives, the sequential type in particu- lar, as treatment for the symptoms of menopause. When medical users and contraceptive users are considered separately (fig. 3), it is seen that the acceleration in the accumulated proportion of ever users is due largely to an increase in the proportion of medi- cal users. There is no ready explanation for the deceleration after 1965 in the accumulated proportion of ever users among women born in 1928-42. Although small, the decline is probably real, not an arti- fact of error in recalling the date of initial use, since the older 1923-27 cohort shows a small but consistent acceleration in the number of first-time OC users throughout the 7-year period. When the reason for use is taken into account (fig. 3), we find that the increase in the accumulated proportion of ever users for contra- ception among the 1923-27 cohort also appears to have slowed somewhat after 1965, while the accumulated proportion of ever users for medical purposes accelerated markedly. The post-1965 overall accumulated proportion of ever users among this cohort was apparently sustained by hormonal therapy for premenopausal symptoms and related gynecological problems. In general, it appears that conversion to oral contraception by women in this population who were in the reproductive period when OC’s were first licensed, began to lose momentum around 1965. As we will show later, deceleration of the rise in proportion of current users among married women of childbearing age also began about the same time. The age-specific accumulation of ever users among the 1938 through 1947 cohorts is depicted in figure 4. With the exception of the 1947 cohort, the age-specific proportion of ever users increased in each successively younger cohort. Age-specific differences among the three oldest cohorts were due essentially to differences in the lengths of time over which ever users were accumulated. For example, at age 25 years, 38 percent of the 1938 cohort and 60 percent of the 1940 cohort were ever users. The 1940 cohort, how- ever, had 6 years in which to accumulate this age-specific propor- tion of ever users, while the 1938 cohort had only 4 years. Four years after the oral contraceptives were licensed for general use, the proportion of ever users among the 1940 cohort was precisely the same as that among the 1938 cohort: 38 percent (table 9). Six SOCIOECONOMIC CHARACTERISTICS AND METHODS 33 rrr rT TTT TT TTT = YEAR OF BIRTH - lous 1944 , 1943 1742 51041 1939 0 - — Ld [7] o > w 0.6 1938 oZ Ld > [TT] es: So. Z o 0.4) 0 a. o ox — i] a. 0.21 | 0 fo pg gi 17 19 2] 23 25 27 29 AGE FIGURE 4.—Proportion ever exposed to contraceptive drugs at given age among individual birth cohorts 1938-47 (white married female Health Plan subscribers, Walnut Creek Service Area). years after licensing, the proportion of ever users among the 1938 cohort was 59 percent, only 1 percentage point less than that among the younger 1940 cohort (table 9). Interest in cohort differences with respect to the age-specific proportions of ever users focuses primarily on those women born after 1941. Because none was more than 18 years of age in 1960, age-specific differences among them reflect secular trends in age 34 STUDY, POPULATION, PLAN, AND PROCEDURES TABLE 10.—Duration of exposure—percent of ever users Age in years Months 20-24 | 25-29 | 30-34 | 35-39 | 40-44 | 45-49 | 50-54 BS cas 9.7 8.7 10.7 14.1 15.3] 22.5 26.1 Al tL es 5.9 5.4 7.4 12.2 10.3 14.8 13.0 TY i ess 11.5 11.8 13.6 | 16.4 16.2 10.6 10.9 1=18 sae 15.5 7.6 10.0 4.7 9.4 11.8 6.5 192d ni 15.1 11.6 5.2 7.5 10.3 9.2 10.9 ESB eo ee ie mim ier = 24 .4 19.0 15.5 12.7 8.9 12.0 19.6 SHB eee bien 11.8 17.1 18.1 9.81 10.8 7.0 4.4 49-80. ee 4.9 9.9 7.4 9.4 7471 4.9 0 Overel. -caoeaox 1.2 8.9 12.0 | 13.1 11.3 7.7 8.7 Number of women__| 608 486 309 215 205 144 46 > AE EL me IE 22.8] 29.8| 30.0 28.0 | 25.4| 21.1 19.1 Standard deviation_| 14.8 | 20.3 | 23.7 | 26.0 | 23.6 | 22.8 21.0 at first exposure to the oral contraceptives. This population, how- ever, lives in a suburban area and is underrepresented with never- - married women. It might therefore be conjectured that the trend toward earlier use of OC’s by later-born women (fig. 4) could be explained by selection of earlier-married women among succes- sively younger cohorts. The cumulative distribution of age at first marriage, however, revealed no important differences among cohorts with respect to this measure except in the 1946 and 1947 cohorts: the 1947 cohort had not married as early as the older cohorts; whereas the 1946 cohort was outstanding for the large proportion who married between the 18th and 21st birthdays. This probably explains why the difference in the age-specific proportion of ever users in the 1946 and 1945 cohorts is so very large at those ages. The difference in age at marriage, however, cannot account for all of the observed difference.” With the exception of the 1947 cohort, which married late, these data showed a strong secular trend toward a progressively younger age at first exposure to the oral contraceptives. To appreciate the remarkably rapid diffusion of the oral contraceptives, consider that in a span of only 4 years, ® Sixty-seven percent of the 1946 cohort, but only 53 percent of the 1945 cohort, was first married at or before age 19, a difference of 14 percentage points. But the difference between these two cohorts with respect to the accumulated proportion of ever users at age 19 is 25 percentage points. SOCIOECONOMIC CHARACTERISTICS AND METHODS 35 TABLE 11.—Duration of exposure—percent of ever users of contraceptive drugs for contraceptive purposes Age in years Months 20-24 | 25-29 | 30-34 | 35-39 | 40-44 | 45-49 | 50-54 Fe 6.7 5.1 4.9 4.6 5.7 8.3 91 Rel nn T nt 5.0 4.9 6.5 6.6 8.9 10.0 9.1 B= ince vtmes 11.5 11.1 9.8 17.2 11.4 8.3 9.1 18-18. oo. 16.0 7.2 11.0 6.0 8.9 16.7 0 19-24 ecrcaen 16.2 12.0 4.5 6.6 10.6 12.7 13.2 Bl iim 26.1 20.1 18.3 15.2 9.8 16.7 36.4 37-48 _ oo __. 11.9 19.0 22.0 13.3 15.4 6.7 9.1 49-60_ _ _ ____________ 5.4 10.6 8.5 13.3 12.2 5.0 0 Overel...co.consens 1.3 10.0 14.6 17.2 17.1 16.7 9.1 Number of women. _| 556 433 246 151 124 60 11 Xe 23.9 32.0 34.6 35.3 33.5 30.3 25.3 Standard deviation__| 14.5 20.0 23.4 25.8 24.5 26.5 18.4 TABLE 12.—Duration of exposure—percent of ever users of contraceptive drugs for noncontracepltive purposes Age in years Months 20-24 | 25-29 | 80-34 | 35-39 | 40-44 | 45-49 | 50-54 18 coer 42.3 38.5 33.3 37.1 30.0 32.9 31.4 4B i ce sma cm 15.4 9.6 11.1 25.8 12.5 18.3 14.3 0 11.5 17.3 28.6 14.5 | 23.8 12.2 11.4 13-18. eee. 9.6 11.5 6.4 1.6 10.0 7.3 8.6 10-24... cciennnammaw 3.8 7.7 7.9 9.7 10.0 7.3 8.6 BB is EEE 5.8 9.6 4.8 6.4 7.5 8.5 14.3 37-48 oo... 11.5 1.9 3.2 1.6 3.8 7.3 2.9 49-60_ _ ___________.. 0 3.8 3.2 0 0 4.9 0 Over60_.______....__ 0 0 1.6 3.2 2.5 1.2 8.6 Number of women ._| 52 53 63 64 81 84 35 IX niin 11.4 11.6 12.1 11.0 13.1 14.6 17.2 Standard deviation .| 13.2 13.2 14.8 17.0 15.3 17.1 21.7 36 STUDY, POPULATION, PLAN, AND PROCEDURES TABLE 18.—Duration of exposure—percent of current users for purpose of contraception Age in years Months 20-24 | 25-29 | 30-34 | 35-39 | 40-44 | 45-49 | 50-54 158 cn mana Se 5.3 0.4 1.4 0 1.8 0.0 0 del 2.9 2.4 2.7 3.4 2.6 4.6 0 7 CE 12.0 8.4 5.5 7.9 10.4 6.8 12.5 18-18... nines 18.5 5.6 10.9 6.8 7.81 20.5 0 Y0=24 la 13.2 11.2 5.5 4.5 7.81 11.4 25.0 Bn 29.5 19.9 19.0 16.3] 11.7 i 20.5 37.5 BTR nimi 14.3 24.3] 22.4 18.0 19.5 6.8 12.5 A900 er os 7.3 1241 “12.2 16.8 14.3 6.8 0 Overgl.. . Lo. cio 200 15.5] 204i 25.81 24.7| 22.9 12.5 Number of women__| 342 251 147 89 nm 44 8 Econ 26.4 | 88.1 | 41.2 | 44.9 | 41.4 37.0 29.2 Standard deviation .| 15.8 19.7 23.6 25.8 23.9] 26.9 17.9 TABLE 14.—Duralion of exposure—percent of past users for purpose of contraception Age in years Months 20-24 | 25-29 | 30-34 | 35-39 | 40-44 | 45-49 | 50-54 EeBe enti iain mis 8.9 11.6 10.1 11.3 13.0 | 31.3 ®) Gal oe nme 8.4 8.3 12.1 11.8 19.6] 25.0 M) Bl rr Pre a 10.8 14.9 16.2 | 30.6 13.0 12.5 0 13-18. ed 20.1 9.4 11.1 4.8 10.9 6.2 0 30-24... ain 21.0 13.2 3.0 9.7 15.2 12.5 0 2580. cc irimnin—e 20.6 20.4 17.2 12.9 6.5 6.2 ® ST=AB rica 8.0] 11.6 21.2 6.5 8.7 6.2 0 89-80. eee amis 2.3 8.3 3.0 8.1 8.7 0 0 Over60..... . .cc-i- 0 2.2 6.1 4.8 4.3 0 0 Number of women__| 214 182 99 62 47 16 3 Xe 19.9) 23.6] 24.7 21.4; 20.6] 11.6 14.7 Standard deviation | 12.1 16.9 19.5 18.7 19.8 13.2 18.5 1 Only 1 user for this category. SOCIOECONOMIC CHARACTERISTICS AND METHODS 37 from 1962 through 1966, the proportion of ever-exposed 20-year- olds increased from 16 to 66 percent. Accumulated Duration of Exposure The total months of OC use among all ever users is shown by age in table 10. Ever users aged 25-39 years had accumulated the most exposure—214 years, on the average. The relative frequency of long-term users was greatest among ever users aged 35-39 years: 13 percent had used these drugs for more than 5 years, and an additional 9 percent had accumulated 4 to 5 years’ exposure. Duration of exposure was considerably longer among women who had taken or were taking the hormones for contraception than among medical users. Users for eontraception had accumu- lated, on the average, between 2 and 3 years’ exposure, depending on age (table 11) ; whereas the majority of medical users had less than 12 months’ exposure (table 12), and a third or more, depend- ing on age, had taken the drugs not more than 8 months. This is to be expected because many of the pertinent medical objectives can be achieved only with a few cycles of use. Tables 13-16 show total months of OC use among current and past users for contra- ception, and current and past medical users, respectively. Current users for contraception had accumulated the longest exposure to these drugs, and past medical users the shortest. The proportion of current users for contraception with more than 60 months’ ex- TABLE 15.—Duration of exposure—percent of current users for noncontraceptive purposes Age in years Months 20-24 | 25-29 | 30-34 | 35-39 | 40-44 | 45-49 | 50-54 1-8 eee 5.9 28.6 15.4 16.7 7.2 0 6.2 4-6 oo _._ 29 4 0 0 16.7 7.2 15.2 6.2 TR 5.9 28.6 380.8 0 25.0 | 21.2 18.8 13-18 oo _.. 17.6 | 14.3 | 80.8 | 16.7 | 14.8 12.1 12.5 19-24... 5.9 0 7.7 16.7 10.7 15.2 18.8 25-36. _ ____________ 5.9 14.8 7.7] 83.8 | 21.4] 15.2 18.8 37-48 _ _ _ ______. 29.4 0 0 0 7.2 6.1 0 49-60. _____________ 0 14.3 0 0 0 12.1 0 Over60_____________ 0 0 9 0 7.2 3.0 18.8 Number of women._| 17 7 13 6 28 33 16 BY NE rem ii 19.4 16.9 17.2 13.7 22.6 23.8 27.7 Standard deviation_| 15.6 19.6 19.2 1.7 20.3 19.1 26.7 38 STUDY, POPULATION, PLAN, AND PROCEDURES TABLE 16.—Duration of exposure—npercent of past users for noncontraceptive purposes Age in years Months 20-24 | 25-29 | 30-34 | 35-39 | 40-44 | 45-49 | 50-54 Ao ie mimi 60.0 40.0 | 38.0 | 39.3 42.3 55.1 52.6 BB Tact nas 8.6 111 14.0 26.8 15.4 20.4 21.1 ol Ee 14.3 15.6} 28.0 16.1 23.1 6.1 5.3 18-18 ae 5.7 11.1 0 0 7.7 4.1 5.3 39-24 ein 2.9 8.9 8.0 8.9 9.6 2.0 0 25-36. . «cvs 5.7 8.9 4.0 3.6 0 4.1 10.5 re mn asmnn 2.9 2.2 4.0 1.8 1.9 8.2 5.3 9-60. eve cmp wd 0 2.2 4.0 0 0 0 0 Over60. oC nus 0 0 0 3.6 0 0 0 Number of women__| 35 46 | 50 58 53 51 19 NX hans Beets 7.51 10.8] 10.83] 10.5 8.1 8.6 8.3 Standard deviation _ 9.9 12.0 13.3 17.4 8.6 12.6 10.7 TABLE 17.— Extent of total population exposure (percent) Duration of Age in years exposure (months) 20-24 | 25-29 | 30-34 | 35-39 | 40-44 | 45-49 | 50-54 Over 0... ecceuinn 0.9 7.0 7.7 5.9 4.4 2.3 1.8 49-60... oo... 3.9 7.8 4.8 4.2 2.9 1.4 0 =A ed 9.4 13.4 11.6 4.4 4.2 2.1 0.9 IB i memes 19.8 14.9 10.0 5.7 3.5 3.5 4.0 19-2. coc cnirnmaan 12.0 9.1 3.3 3.4 4.0 2.3 2.2 18=18.. once 12.3 6.0 6.4 2.1 3.7 3.3 1.83 2 er inde ewe 9.2 9.2 8.7 7.3 6.4 3.1 2.2 8-0. omens Bem 4.7 4.3 4.8 5.7 4.2 4.6 oq 1-8. ede 7.9 6.9 6.8 6.5 6.2 6.9 5.4 Never exposed... 20.5 21.4 35.9 54.9 60.5 70.1 79.5 Number of women__| 765 618 482 477 519 481 224 posure exceeded 20 percent in all age groups above 29 years except the oldest (table 13). Relatively few women, however, were long- term users; in no age group did long-term users constitute as much as 8 percent of the population (table 17). SOCIOECONOMIC CHARACTERISTICS AND METHODS 39 TIME TREND IN CONTRACEPTIVE DRUG USE AMONG WHITE MARRIED WOMEN The time trend in contraceptive drug use among white mar- ried women 20-44 years of age, shown in table 18, is based on the calendars filled out from memory ; shown here is the proportion of women who were taking OC’s in December of each calendar year, beginning with 1960, among white survey subjects who, at the stated point in time, were 20-44 years of age and married. The increase in use over time is illustrated in figure 5. From 1960 through 1965, the proportion of current users increased rapidly ; after 1965 the annual rate of increase began to fall off. This was due in large part to a leveling off in the proportion of current users among women in their early 20’s who, by 1965, were using oral contraception almost exclusively. The time trend in use among individual 5-year age groups is shown in figure 6. By the end of 1965, contraceptive drug use among women 20-24 years was beginning to approach a ceiling set essentially by the proportion engaged in reproduction. Fertility is normally highest in this age group; the proportion not using con- traception at any given point in time will be large. In February 1968, the only point for which this information is known, 31 per- cent of white married survey subjects aged 20-24 years were not using contraception because they were pregnant or trying to be- come so. When the small proportion of women who had discon- tinued oral contraception because of some dissatisfaction or prob- lem with the method (8 percent) is added to this figure, it seems unlikely that OC use by women in this age range will ever exceed TABLE 18.—Current users of oral contraceptives (percent) in December of each calendar year, 1960-67, among white women, then married, and 20-44 years of age Age in years Calendar # v year 20-24 | 25-29 | 30-34 | 35-39 | 40-44 | 20-44 Y900: oo insane 3.4 4.4 1.5 1.4 0.2 2.0 1961... 12.8 7.9 5.2 8.2 2.8 5.8 1962... ininmminme 16.8 12.5 8.8 6.1 3.4 8.9 1068... cicacinncnen 27.6 17.8 11.0 9.4 5.4 13.2 1904... cies inmnaane en 87.7 26.0 18.2 13.0 8.3 19.3 085 i man 46.2 32.3 24.5 19.8 12.6 25.5 ABB... cere wri 46.6 38.6 28.8 22.0 17.0 29.2 1967... scenes 49.6 41.3 32.2 20.2 21.1 31.2 40 0.5 0.4 0.3 0.2 0.1 PROPORTION USING ORAL CONTRACEPTIVES 0 STUDY, POPULATION, PLAN, AND PROCEDURES | | | | 1960 1961 1962 1963 1964 1965 CALENDAR YEAR FIGURE 5.—Proportion of white married women 20-44 years of age taking oral contraceptives in December of each calendar year 1960-67 (Health Plan subscribers, Walnut Creek Service Area). 1966 1967 60 percent, at least as long as the fertility of this age group remains high. Assuming a practical ceiling of 60 percent, it is un- derstandable that, having attained 46 percent or so by the end of 1965, the rate of increase that characterized the previous 3 years could not continue. Nonetheless, it is curious that the deceleration in the rate of increase after 1965 was so marked (fig. 6). SOCIOECONOMIC CHARACTERISTICS AND METHODS 41 0.5 AGE 20-24 0.2} 0.1 : i) PROPORTION USING ORAL CONTRACEPTIVES | | | | | | 1960 1961 1962 1963 1964 1965 1966 1967 CALENDAR YEAR FIGURE 6.—Proportion of white married women 20-44 years of age taking oral contraceptives in December of each calendar year 1960-67, by 5-year age groups (Health Plan subscribers, Walnut Creek Service Area). 0 Because the concentration of users was so high in the age group 20-24 years, the behavior of women in their early 20’s dominated the time trend in OC use for women across the age range 20-44 years. All age groups below 40 years, however, showed some tendency toward deceleration beginning around 1965 and increasing through 1967. At ages 35-39 years, the prevalence 42 STUDY, POPULATION, PLAN, AND PROCEDURES of users appears to have actually decreased between 1966 and 1967. This decrease is not understood. In contrast to the younger age groups, the 40-44 year group continued to show a fairly constant increase in OC use after 1964. As noted earlier, conversion from never user to user among women approaching menopause began to accelerate around 1965. It is possible that an increasing tendency for premenopausal women to take these drugs as a source of female hormone offset a decelera- tion of the rise in their use for contraception. The proportion who were nonusers at various points in time is not known; but in Feb- ruary 1968 the drugs were being taken for medical reasons by a much larger proportion of women over 40 years than by younger women. Nearly 6 percent of women aged 40-44 years, and only 1 percent of those aged 35-39 years, were noncontraceptive users. COMPARISON WITH THE 1965 NATIONAL FERTILITY STUDY The 1965 National Fertility Study led to an estimate that 16 percent of white, married, American women younger than 45 years were taking oral contraceptives in October of that year.! The calendars on contraceptive drug use filled out by the subjects of the present survey indicated that at the time of the National Survey, 26 percent of those then married and 18-44 years of age TABLE 19.—Comparison between National Fertility Study and Contraceptive Drug Study survey on oral contraceptive use among white married women in 1965 (percent) National Fertility Contraceptive Drug Study,! Oct. 1965 Study,? Dec. 1965 Age in years Ever Current Ever Current users users users users TINAOL Oc iii sriiiiono eo rere 45 30 76 52 02h ee vues 49 82 76 46 OE EL lS 38 22 59 32 803 ea ead 25 14 44 24 a ee ei rel 15 8 30 20 BOR. oo re emai iim ere wre re 9 5 22 13 Overall Ln on 16 44 26 1 Source: Ryder, N. B., and Westoff, C. F.: Use of oral contraception in the United States, 1965. (Table 6.) Science, 153: 1199-1205, Sept. 1966. 2 Based on calendars of OC use of white women who were less than 45 years of age and had ever married by the end of 1965. SOCIOECONOMIC CHARACTERISTICS AND METHODS 43 TABLE 20.— Use of contraceptive drug, December 1965, by white married women, according to whether membership in Health Plan postdates 1965 Joined Health Plan Joined Health Plan 1965 or earlier after 1965 Age in years, Dec. 1965 Num- Ever | Current| Num- Ever | Current ber user user ber user user Percent | Percent Percent | Percent 18-19 __ 63 82.5 58.7 61 68.8 45.9 20-24. ieee 301 79.8 45.6 250 71.4 46.9 OB en vos 384 62.3 32.9 180 52.0 31.2 80-34... een 306 44.7 24.1 103 40.4 25.6 85-80... iene 371 31.8 21.6 83 22.9 12.0 40-44 ____________ 400 23.5 13.8 85 14.1 7.0 were current users of contraceptive drugs. This markedly more widespread use of the oral contraceptives by women in the Kaiser study population was found in all age groups (table 19), but pre- dominated in the age groups below 20 and above 34 years. Because they are based on recall, the figures on OC use in the Kaiser popu- lation in 1965 are not strictly comparable to the national estimates. But there can be no doubt that diffusion of the oral contraceptives was much more rapid in this select group of women than in the general population. Since the representation of Roman Catholics in the Walnut Creek Kaiser population is the same as the national average, the religious composition of this population cannot account for their greater use of oral contraception. Several factors, however, may have contributed to the earlier acceptance of the oral contraceptives by women in the Walnut Creek Kaiser population. First is the region in which they live. The National Survey found that OC use was greatest in the West, al- though regional variation was not large. Second, and more im- portant, the Kaiser population is relatively well educated. Seven- teen percent of the Kaiser survey subjects were college graduates; only 7 percent of the national sample had achieved this level. None- theless, the overall proportion of Kaiser wives using oral contra- ception in 1965 was 4 percentage points higher than that of college graduates in the national sample (26 versus 22 percent). An additional factor that may have contributed to the fre- quency of oral contraceptive drug use among Kaiser wives was their source of medical care. The gynecological staff of the Walnut Creek Kaiser Foundation Hospital preferred oral contraceptive 44 STUDY, POPULATION, PLAN, AND PROCEDURES drugs over other methods of contraception. Furthermore, the physician services usually required to obtain oral contraceptives, as well as the drugs themselves, were available to Health Plan members at little or no cost. Thus, women who wished to use oral contraception would not be deterred from doing so by financial considerations. Evidence of an effect on OC use associated with Health Plan membership is shown in table 20. This table shows the relative frequency of OC users in 1965 according to whether membership in the Health Plan postdates 1965. Because women who were Health Plan members in 1965 were generally older than those who joined later, the overall percentages of users are not comparable. Within age groups, however, ever users in 1965 tended to be more prevalent among women who were then Health Plan members. The difference between the two groups in current users was confined to the age groups below 20 and above 34 years, where, it will be recalled, the overall deviation from the national estimates was greatest. DISCONTINUATION OF ORAL CONTRACEPTION Fifteen percent of white married women were past users of oral contraceptives. This is 39 percent of all ever users of oral con- traception (table 21). REASONS FOR DISCONTINUATION The reasons given for discontinuation are shown in table 22. Twenty-four percent stopped to conceive. For 5 percent, the loss of fertility at menopause or with hysterectomy removed the need for contraception. Women who discontinued oral contraception for reasons other than these may be said to have rejected oral contra- ception as their method of birth control. These women are called “dropouts.” Seventy-one percent of all past users of oral contra- ception were dropouts (table 22). This is 27 percent of all women who have ever used oral contraception (table 21). The major reason given by 45 percent of past users for stopping OC use was unpleasant side effects. Alleged side effects included nausea, weight gain, breast tenderness, skin discoloration, headache, leg cramps, breakthrough bleeding, and heavy or scanty menstrual flow. Other commonly cited complaints were depression, irritability, fatigue, and disinterest in sex. Multiple rather than single complaints were the rule. General statements such as “I just didn’t feel good,” or “I felt as though I were pregnant all the time,” were not infrequent. Because of the wide variation in the detail, precision, and pre- SOCIOECONOMIC CHARACTERISTICS AND METHODS 45 TABLE 21.— Discontinuation of oral contraception—white married women Ever Past users of oral Dropouts! users of contraception Age in years oH ception | Number | Percent | Number | Percent (number) 20-24 _ _ _ ______ 500 193 38.6 85 17.0 D520. oc cis dnn ww 401 169 42.1 115 28.7 80584 oro meant onnsmenes 231 97 42.0 83 35.9 85-80... cco mim nmin mom 136 53 39.0 47 34.5 40-44 _ 116 42 36.2 36 31.0 45-49 _ 54 13 24.1 10 18.5 BO=B4 nmin me 10 2 20.0 1 10.0 Overall pereentage... {cesses cnc emees 838.8 |..cnunnnn 27.4 1 A dropout is a woman who discontinued oral contraception for reasons other than the wish to become pregnant or the lack of need for contraception. TABLE 22.—Past users of oral contraception: Reason for discontinuation— white married women (percent) 0 0 Con- 0 on- . | cern Num-| Pe [tracep-| Side Meds for Mise Dro Age in years ber | come tion & or long- Voges ik reg- | not ects term D op Rol ders | “Gi™ | ous fects 20-24. iecsnnees 193 | 54.9 | 1.0]22.8| 1.0 4.1 |16.1 | 44.0 25-29... ..cnnnnenimmnnm 169 { 80.81 1.21938.8: 1.8} 7.7)20.1{:680 30-84 _____.______... 97110.3) 4.11629] 0 2.1120.6) 85.6 85-89 o-.. 53| 3.8| 7.5623] 0 5.7.120.7-{ 88.7 40-44... connsennsnnon 42 0 14.8 (589.5; 7.1) 7.1111.9; 85.7 A549... eee 13 0 23.1 | 46.2 7.9 7.71 15.4 76.9 BO-54. lim mina iol irin 21 0 50.0 | 50.0 | ©O 0 0 50.0 Overall percentage|..____ 24.3 5.1 | 45.4 1.8 5.83} 18.1 70.6 1 A dropout is a woman who discontinued oral contraception for reasons other than the wish to become pregnant or the lack of need for contraception. sumably the completeness with which respondents reported their complaints, no attempt was made to quantify the frequency of specific complaints. Only nine past users of OC’s were advised to discontinue the drug by a physician. In one case, discontinuation was ordered be- 46 STUDY, POPULATION, PLAN, AND PROCEDURES cause of chronic cystic mastitis; in another, because of a fibroid adenoma of the breast. A 45-year-old woman was instructed to discontinue OC use when a deep vein thrombosis developed in one leg. Two women with marked varicosities were advised not to continue OC use. A fibroid uterine tumor developed in one woman. The medical problem in one case was not specified. Two women were advised to discontinue OC use temporarily, one to undergo abdominal surgery, the other, a thyroid function test. Five percent of past users stopped taking OC’s simply because they were concerned about possible harmful effects. The bases for their concern varied. Some referred to articles in the lay press about thrombophlebitis and sudden death among users. Others made no mention of magazine articles but expressed the view that too little was known about the drug’s effects. For some women, the fact that the drug alters a normal body process was cause for con- cern. A number of women feared becoming infertile or, conversely, remaining fertile into old age. A family history of disease, such as thrombosis, thought to be associated with use of these hormones, led several women to conclude that continuation would subject them to special risk. One-fourth of those who discontinued OC use because of side effects also expressed concern about its possible long-term effects. The extent to which anxiety precipitates side effects is unknown, but many women may feel uneasy while using oral contraception. Nine percent of past users decided to terminate fertility per- manently by tubal ligation or by vasectomy of the sexual partner. None of these women expressed any difficulty or dissatisfaction with oral contraception. One percent changed to an intrauterine device, giving no explanation for discontinuing OC use. Few (1.5 percent) discontinued oral contraception because of problems with pill taking. Other miscellaneous reasons included infrequent inter- course, accidental pregnancy, religious conflict, and lack of con- cern about becoming pregnant. Women who discontinued OC’s to become pregnant or upon divorce and who did not resume it upon resumption of risk of pregnancy, are also grouped under “mis- cellaneous.” ToTAL MONTHS OF USE The total months of OC use among drop outs is shown in table 23. At least half of those younger than 45 years had used the drugs longer than 12 months. Fewer than 15 percent had used them for only three or fewer cycles. The total months of use among dropouts who discontinued SOCIOECONOMIC CHARACTERISTICS AND METHODS 47 TABLE 23.— Total months of oral contraceptive use: Dropouts— white married women (percent) Age in years Minimum number of months’ exposure 20-24 | 25-29 | 30-34 | 35-39 | 40-44 | 45-49 Bienen ne Ren EE EET 85.7 | 87.0 | 87.6 | 87.2 86.1 60.0 0 creme ime a Resim Bn mrmam 7.0; 75.6] 72.8] 76.6 | 63.9 40.0 12 i nmmm pase we wi m——— 61.9 | 56.5| 55.6 | 48.9 | 52.8 40.0 Dh secre Ee EERE 25.0 | 40.0 | 42.0 31.9 | 22.2 20.0 BB. ccm RE 10.7( 21.7} 24.7 | 17.0] 19.4 10.0 48 oo __ 4.8] 10.4 8.6 12.837 11.1 10.0 BQ se oem iran es were mre 0 9 7.2 4.2 2.8 0 Number of women._.________ 84 115 81 47 36 10 Total months of exposure: Te em 18.3 | 21.9 22.7] 21.1] 19.1 13.8 Standard deviation_______ 13.4 17.0 19.2 18.7 19.0 17.6 TABLE 24.—Total months of oral contraceptive use: Discontinuance because of side effects—white married women (percent) Age in years Minimum number of months’ exposure 20-24 | 25-29 | 30-34 | 35-39 | 40-44 | 45-49 8 serie sree reed 79.5 | 84.6 | 86.4 | 87.9 | 80.0 50.0 ES, 72.7 | 80.0| 78.0 | 81.8 68.0 13.3 YZ. ee trim mmm m————— 61.4 | 56.9 | 59.3 54.5 | 52.0 13.8 DA ee erm o e———— 31.8 36.9 | 42.4 | 36.4 | 28.0 13.3 BB dis domes 11.4] 20.0; 25.4 18.2 1 24.0, 0 AR rE TEEPE 4.5 10.8 6.8 15.2] 16.0 0 80. i simmer 0 1.5 5.1 3.0 4.0 0 Number of women__._______ 44 65 59 33 25 6 Total months of exposure: EE 17.83 | 20.8] 22.3 20.6 19.8 6.2 Standard deviation. ._.._._ 13.2 17.3 19.8 17.6 | 20.9 8.8 oral contraception because of side effects is shown in table 24. Even among these women, few used the drugs for no more than three cycles. Again, more than half were users for 13 or more months, a third for at least 25 months. Total months of use among dropouts gives no information about the dropout rate over time. The probability of discontinuing 48 STUDY, POPULATION, PLAN, AND PROCEDURES oral contraception for reasons other than the wish to conceive or involuntary loss of fertility was calculated by the life-table method using the most recent period of continuous OC use. The results of this analysis are not presented because it is felt that they are dis- torted by not taking prior exposure into account. The probability that a woman will discontinue oral contraception within a speci- fied period of time is likely to be influenced by whether she has ever used these drugs before. To be accurate, such an analysis should be based on the experience of first-time users. The reason for discontinuation was known, however, only for the most recent period of OC use. It was therefore not possible to calculate a drop- out rate for the initial period of continuous use. Although we cannot say when a woman was most likely to discontinue oral contraception because of side effects, it is note- worthy that many women who did so had experienced considerable exposure to these drugs. This suggests that these women may have tolerated what they perceived as an unpleasant or undesirable side effect for a considerable period before discontinuing the drug. It is also possible that certain so-called side effects intensified with time, or developed only after many months of OC use. ORAL CONTRACEPTIVE USE IN RELATION TO PARITY, EDUCATION, AND RELIGION AMONG WHITE MARRIED WOMEN PARITY The relative frequency of ever users of the oral contraceptive drugs is shown by age and number of live births in table 25. Child- TABLE 25.— Relative frequency of ever users of contraceptive drugs, by number of live births—white married women (percent) Age in years Number of live births 20-24 | 25-29 | 30-34 | 35-39 | 40-44 | 45-49 | 50-54 0 = in mi mimi wn 87.81 76.61 066.7] 36.4 22.7 | 27.8 I § Yee in prea 88.7 80.0 | 74.4 | 58.5| 89.0] 32.1 20.0 Be reer ERs Te Sn 92.31 82.5] 76.9] 89.4; 45.0 | 32.4 27.3 Rr SLI 78.4 | 79.0 | 59.1 | 44.4 | 43.4 | 36.0 15.4 ER Se TR Ar SORE 100.0 | 71.1 54.7 | 40.8 | 40.5 | 23.4 24.0 BOT MOTE... comms wm) mms 68.4 55.9 56.7 41.0 27.1 2.2 Overall....._.. 88.9 | 79.1 66.0 | 45.8 | 41.4 | 30.6 20.3 SOCIOECONOMIC CHARACTERISTICS AND METHODS 49 TABLE 26.—Relative frequency of ever users of contraceptive drugs for preventing pregnancy, by number of live births—white married women (percent) Age in years Number of live births 20-24 | 25-29 | 30-34 | 35-39 | 40-44 | 45-49 | 50-54 Ons neem rrmasaomai 81.8 | 65.6 | 42.9 9.1 4.6 2.8 0 Voronin 849! 74.4] 51.3} 31.71 22.0} 11.3 11.4 ise reel 89.1) 75.81 68.7] 27.7] 80.6 | 16.7 5.5 Brn eee ee 73.0 | 68.4 | 47.2 | 86.1 | 26.4 | 15.0 2.6 Re mE EEE 100.0 | 60.0 | 39.1 | 32.4 | 22.8 | 14.1 4.0 BOT MOLE. «cen wms ream 78.4 47.1) 37.3 | 25.6 10.4 11.1 Overall. ____.. 84.4 | 71.1 53.4 | 32.6 25.1 13.2 5.2 TABLE 27.—Relative frequency of ever users of contraceptive drugs, by education— white married women (percent) Education Age in years College Some High school | Less than graduate college graduate high school 20-24 eee 93.8 88.5 88.9 87.8 25-20. «vs weneunensmene 82.6 82.2 80.8 56.1 30-34... cori mrmE 74.7 75.5 59.3 57.4 85-30... .cconenwmmnm— 53.0 48.1 42.0 43.0 40-44 ... 48.2 44.1 87.9 39.2 45-49... ionenanmaionan 44 4 25.9 29.9 26.1 EE 34.4 22.5 18.9 10.9 less women older than 34 years were less likely to have used the drugs, particularly for the purpose of preventing pregnancy (table 26), than their parous peers. Otherwise, use varied little by parity. This result is consistent with the findings of the 1965 National Fertility Study. We do not know the extent to which women in this popula- tion had used oral contraception to plan the growth of their fami- lies; i.e., how many first births, second births, etc., were timed by the use of oral contraception. Nearly all women 35 years of age or older had started their families before oral contraceptive drugs became available in 1960, and it is reasonable to assume that use among contraceptive users in the older cohorts was largely for the purpose of curtailing fertility rather than timing or spacing preg- 50 STUDY, POPULATION, PLAN, AND PROCEDURES TABLE 28.— Use of oral contraceptive drugs for purpose of preventing pregnancy, by religion—white married women Ever users of oral contraception [Current users of oral contraception Age in years Catholic] Non- Difference+ |Catholic| Non- Difference+ Catholic | standard error Catholic | standard error Percent | Percent Percent | Percent 20-24_____ 81.4 85.6 —4.243.5 49.7 52.7 —3.0+4.6 25-29____. 62.9 74.3 | —11.44+4.4! 41.5 41.0 544.6 30-34____. 51.9 54.1 —2.245.2 36.4 28.7 7T.7+£5.0 35-39__.._. 23.7 36.0 | —12.3+4.82 13.2 22.4 —9.244.01 40-44_____ 17.6 27.7 | —10.14+4.21 10.9 17.8 —6.943.51 45-49 _____ 3.6 15.5 | —11.94+2.9? 3.6 11.6 —8.04+2.72 50-54..... 5.0 5.2 —-0.2+3.9 5.0 3.9 1.1+3.8 1 p<0.05. 215 <0.01. nancies. The data available to us suggest that even in the young- est cohorts, many women did not begin to take oral contraceptives until after one or more pregnancies. The age at first use was later than the age at first pregnancy in at least half of ever users aged 20-25 years. EDUCATION Use of OC’s varied with education. In all age groups, the rela- tive frequency of ever users increased with the amount of educa- tion the woman had (table 27). Because this population is so highly educated, the difference in education between ever users and never users was not so great as might appear—on the average, one- half year of school. In the age group 20-24 years, where the vast majority had used the oral contraceptives regardless of education, ever users and never users differed little in the amount of formal school. RELIGION Table 28 shows the relative frequency of use for the purpose of contraception among Roman Catholic and non-Catholic women. Although Catholic women had in general used oral contraception less frequently than non-Catholics (32 versus 40 percent), the dif- ference between the two groups was small below the age of 35 years and, except in the age group 25-29 years, not statistically significant. Statistically significant differences between Catholic and non-Catholic women in the proportion of current contracep- SOCIOECONOMIC CHARACTERISTICS AND METHODS 51 TABLE 29.—Ewver users of contraceptive drugs for contraceptive purposes, by frequency of church attendance—white married Catholic women Church attendance Age in Weekly Monthly | Occasionally Never Over- years all : (per- Num-| Per- |Num-| Per- | Num-| Per- | Num-| Per- | cent) ber | cent | ber | cent | ber | cent | ber | cent 20-24. cnn 40 | 69.0 25 | 86.2 44 | 84.6 25 1100.0 81.4 25-29... eee 44 | 55.7 11 | 55.0 30 | 68.2 14 | 93.3 62.9 30-34__________ 35 | 47.9 9] 52.9 17 | 56.7 6175.0 51.9 85-39. ____.____ 15 | 23.8 1| 8.3 620.0 Bi171.43 23.9 40-44 __________ 11 | 16.4 81]25.0 624.0 0 --___ 17.6 45-49 _____. 2 seme Lena 0 a 0 ocwans 3.6 50-54 _________ deme 0l....con 1 end 0 |..onae 5.0 TABLE 30.—Ewver users of contraceptive drugs for contraceptive purposes, by frequency of church attendance—white married Protestant women Church attendance Age in Weekly Monthly Occasionally Never Over- years all (per- Num-| Per- | Num-| Per- | Num-| Per- | Num-| Per- | cent) ber | cent | ber | cent | ber | cent | ber | cent 20-24... comenen 37 | 84.1 41 | 87.2 | 149 | 83.2 61 | 88.4 | 85.0 25-29. ......... 58 | 72.5 39 | 78.0 | 107 | 77.5 89 | 76.5 | 76.5 80-34_________. 37 | 51.4 19 | 52.8 56 | 52.8 18 | 64.83 | 53.5 35-39. _.._. 32 | 45.1 12 | 29.3 3433.7 928.1] 34.9 40-44 __________ 25] 31.2 5]11.9 34 | 29.8 12 129.3 27.9 45-49.......... 41 5.1 10 | 25.6 16 | 16.2 10 (24.4 | 15.5 50-54 _______._. 1] 8.5 1 3.1 2| 8.8 2112.5 4.5 tive users were found only in the age groups above 34 years (table 28). The relative frequency of ever users of oral contraception among Catholic women increased as the frequency of church at- tendance decreased (table 29). No such trend was found among Protestant women (table 30). The important feature of these data is not the difference be- tween Catholic and non-Catholic women in the use of oral contra- ception, or the difference in OC use among Catholics by frequency 52 STUDY, POPULATION, PLAN, AND PROCEDURES of church attendance, but rather the large extent to which Roman Catholic women active in their church were users of oral contra- ception, particularly in the age groups below 35 years. It should be kept in mind, however, that these data were obtained before Pope Paul VI issued the encyclical in which prohibition of oral contra- ception was made explicit. METHODS OF CONTRACEPTION USED BY WHITE MARRIED WOMEN AT RISK OF PREGNANCY Only half the white married population 20-54 years of age was at risk of pregnancy and eligible to use any form of contra- ception. “Eligible to use contraception” means here that the woman was not pregnant, postpartum, lactating, or trying to con- ceive; that neither she nor her husband had had a sterilizing oper- ation, that she menstruated, and that there was no other reason to believe she could not become pregnant by her husband if she so wished. Eight percent of the women were ineligible to be taking OC’s because they were pregnant, postpartum, lactating, or trying to conceive; 40 percent had had a sterilizing operation, were married to vasectomized men, had ceased to menstruate, or had other rea- son to believe they could not conceive with their husbands (table 31).c Voluntary sterilization is a common practice in this popula- tion; as a consequence, the proportion of women eligible to use contraception varied inversely with age. Normally, one would ex- pect the proportion of eligible women to increase with age as childbearing decreases, through the fourth decade, after which the loss of fertility with advancing age would reverse the trend. USE OF ORAL CONTRACEPTIVES Use of the oral contraceptive drugs by women at risk of pregnancy is shown in table 32. Forty-seven percent of women eligible to do so were taking OC’s for the purpose of contraception, making oral contraception the most common method of birth con- trol. An additional 3 percent were receiving a secondary contra- ceptive effect from use of OC’s for a medical purpose. Thus, one- ¢ Fewer than 2 percent of women had reason to believe that they and their husbands could not have any (more) children. Most of these had con- sulted a physician about their problem and could specify its physicial or biological nature. Some women considered failure to conceive despite contin- uous exposure for at least 3 years to be evidence of impaired fertility. SOCIOECONOMIC CHARACTERISTICS AND METHODS 53 TABLE 31.—Eligibility status of white married women (percent) Pre- 6 weeks : Now | post- Trying Sub- No Samed Status Age in years preg- |partum| total longer risk of |, WO- nant | or lac- : fertile known tating | ceive preg- nancy B04... .. oem 14.4 4.2 12.2 30.8 4.21 64.7 0.3 28-29... cwmcinnm mine 8.9 3.4 8.0 | 20.3 16.0 | 63.7 1 30-34___ ____________ 5.1 ni 4.6 10.4 | 29.9 59.0 7 8889. cesrumame sas 1.9 1.0 J 3.4 44.6 | 49.6 2.4 40-44 enn 0 0 2 21 45.2 52.2 2.4 45-49 _ oe. 0 0 .2 .2 55.6 | 41.0 3.2 50-54... een 0 0 0 0 78.1 19.8 2.1 Overall__._____ 3.8 1.2 3.3 8.3] 89.9] 50.1 1.6 NoTteE.—“Eligibility”’ means that the woman was not pregnant, lactating, or trying to conceive; that neither she nor her husband had had a sterilizing operation, that she menstruated, and that there was no other reason to believe she could not become pregnant by her husband if she so wished. TABLE 382.—Use of contraceptive drugs among white married women eligible to use contraception (percent) Current users Past users : Num- Never Age in years ber | For pre- For For pre- For users vention of other vention of other pregnancy | reason |pregnancy| reason 20-24 _____.._. 383 78.9 0.8 12.5 1.3 6.5 25-29 _______._. 359 64.1 .6 19.2 2.8 13.4 30-34__________._ 255 52.2 2.0 18.8 3.1 23.9 35-39. oo... 207 39.1 1.5 12.1 8.2 39.1 40-44 ________.__ 241 29.9 5.4 10.0 6.6 48.1 45-49___________ 169 23.7 3.6 3.0 73 62.1 50-64. _.___. 38 18.4 21.1 0 18.2 47.4 Overall percentage._.| _.._.. 46.7 3.2 12.4 5.3 32.5 Note.—*“Eligibility’’ means that the woman was not pregnant, lactating, or trying to conceive; that neither she nor her husband had had a sterilizing operation, that she menstruated, and that there was no other reason to believe she could not become pregnant by her husband if she so wished. 54 STUDY, POPULATION, PLAN, AND PROCEDURES half of women presumably at risk of pregnancy were protected from this risk by the oral contraceptive drugs. Although the loss of fertility with advancing age had been removed as a determinant of age variation in OC use, the propor- tion of current oral contraception users among eligible women con- tinued to show a marked inverse relationship with age, decreasing monotonically over the age range from 79 to 18 percent (table 32). As noted earlier, 3 percent of women presumably at risk of pregnancy claimed that they were taking these hormones solely for a noncontraceptive purpose. Three-fourths of these women, how- ever, were 40 years of age or older. Below the age of 40 years, very few women at risk of pregnancy took the drugs without contra- ceptive intent—approximately 1 percent. In the older age groups it is difficult to judge whether or not a woman is at risk of preg- nancy. It will be recalled that the occurrence of menstruation was used as one of the eligibility criteria. Because menopausal women taking these drugs for exogenous female hormones continue to have monthly periods, many older medical users of the oral contra- ceptives are likely to be incorrectly classified. OTHER CURRENTLY USED METHODS OF CONTRACEPTION In all age groups, more women were taking OC’s than were using any other method of contraception (table 83). The condom and the diaphragm were the next most commonly used methods, but their ranks varied among age groups. Users of an intrauterine contraceptive device were infrequent in this population, amounting to fewer than 4 percent of eligible women. Fewer than 6 percent of women presumably at risk of preg- nancy were using no method of birth control. It is our impression, from comments written on the questionnaires, that many women at least 45 years old who were using no method believed that be- cause of their age they could no longer conceive even though they still menstruated. No question was asked about the frequency of intercourse, and it is not known what proportion of “no method” women were not exposed to the risk of pregnancy. The group of women using no method was overrepresented with Roman Cath- olics older than 29 years, and, regardless of age or religion, with never users of oral contraception. PRIOR METHOD OF BIRTH CONTROL USED BY CURRENT OC USERS The methods of birth control that had been used by current users just before starting oral contraception are shown in table TABLE 33.—Methods of contraception employed by white married women eligible to use contraception (percent) Oral Dia- Con- Intra- | Spermi- With- | Multi- Un- Age in years Number | contra- | phragm | dom | uterine cide |Rhythm| Douche | drawal ple None | known ceptive device method 20-24_____________. 383 79.6 1.6 4.7 3.4 4.4 1.8 0.3 1.0 1.0 2.1 0 25-29. nema 359 64.6 4.7 7.0 6.4 6.1 1.9 .6 2.2 1.9 3.9 +6 80-34 _____________ 255 54.1 8.6 12.2 5.9 4.3 3.9 4 2.7 2.9 3.9 1.2 35-39 ______________ 207 40.6 1.7 18.4 3.4 3.9 6.8 4.8 3.4 4.8 6.3 0 40-44 ______________ 241 35.3 17.4 16.2 4 3.3 5.8 al 7.0 5.4 5.0 2X 45-49______________ 169 27.2 1.2 18.9 2.4 3.6 7:1 7:1 5.9 3.0 11.2 2.4 80-B4.....concunenas 38 39.5 21.0 15.8 2.6 2.6 2.6 2.6 0 0 10.6 2.6 Overall percentage |________ 49.9 9.3 12.9 3.6 4.2 4.4 2.4 3.6 3.0 5.5 1.1 Note.—“Eligibility’”’ means that the woman was not pregnant, lactating, or trying to conceive; that neither she nor her husband by her husband if she so wished. had had a sterilizing operation, that she menstruated, and that there was no other reason to believe she could not become pregnant SAOHLIAN ANV SOILSISILOVIVHO DINONODHIOID0S go TABLE 34.—Prior method of birth control employed by current users of contraceptive drugs (percent) Dia- Con- Intra- | Spermi- With- | Multi- Age in years Number |[Unknown| phragm | dom | uterine cide | Rhythm| Douche | drawal ple None device method a mn 305 6.2 8.5 15.4 2.6 7.9 4.9 2.6 3.9 2.0 45.9 A ei im ri — 232 2.2 26.3 17.2 4.3 9.1 11.2 3.9 3.0 3.4 19.4 BO=B4. . co rici drm 138 2.9 40.6 16.7 4.4 8.0 11.6 2.2 +7 3.6 9.4 SB=89. curs im smn wa 84 2.4 40.5 13.1 3.6 8.3 7.1 4.8 6.0 0 14.3 BOL... ccm mmm 85 1.2 64.7 11.8 1.2 1.2 3.5 4.7 2.4 1.2 3.2 45-49 oo. 46 0 58.7 19.6 4.8 6.5 4.3 4.3 2.2 0 0 60-64. oo. 15 0 46.7 6.7 0 20.0 0 0 6.7 0 20.0 Overall percentage ____|._______ 3.3 34.5 15.1 3.4 7.7 7.4 3.4 3.2 2.0 20.1 9% SAINAAI0AL ANV ‘NVId ‘NOILVINIOd ‘XANLS SOCIOECONOMIC CHARACTERISTICS AND METHODS 57 34. Forty-six percent of those 20-24 years of age and 19 percent of those 25-29 years of age were using no method of contraception at the time they started OC use. It should not be assumed, how- ever, that such women had had no prior contraceptive experience. The proportion for whom these drugs were their first contracep- tive experience is unknown. In all but the youngest age group, the method most commonly employed prior to starting oral contraception was the diaphragm; all together, 34 percent of current OC users were using a dia- phragm at the time they started oral contraception. A more accurate assessment of the extent to which OC users would be diaphragm users, were it not for OC’s, may be gained by excluding women in the third decade of life. Most women beyond the third decade had married and established contraceptive practices before these hormones were licensed for general use. Forty-nine percent of these women used diaphragms before starting OC use. In con- trast to this, only 19 percent of eligible non-OC users were using the diaphragm (table 35). These data do not provide a direct meas- ure of the impact of OC’s on existing methods of contraception, but the 1965 National Fertility Study found that, with the ascend- ancy of oral contraceptive drugs, the method of contraception showing the greatest concomitant decline has been the diaphragm.? COMPARISON BETWEEN FORMER USERS AND NEVER USERS OF ORAL CONTRACEPTION The methods of birth control employed by former users of oral contraception are shown in table 36 ; those used by never users of oral contraception appear in table 37. Although the proportions using the diaphragm or condom were the same for both groups, never users generally employed less effective methods of contra- ception than did former users. Whereas the intrauterine device and spermicidal agents were much more commonly used by those who had formerly employed oral contraception, coitus interruptus and vaginal douching were practiced by more never users. Further- more, the frequency of women who used no method was greater among never users than among former users of oral contraception (13 versus 6 percent). PREVALENCE OF SURGICAL STERILIZATION AMONG WHITE MARRIED WOMEN AND THEIR HUSBANDS Sterilizing operations may be classified by intent as either remedial or contraceptive. A contraceptive operation is one done TABLE 35.—Method of birth control used by white married women eligible to use coniraception—not taking contraceptive drugs (percent) Dia- Con- Intra- | Spermi- With- | Multi- Age in years Number [Unknown| phragm | dom | uterine cide | Rhythm| Douche | drawal ple None device method 20-24... 72 0 7.7 23.1 16.7 21.8 9.0 1.3 5.1 5.1 10.3 25-29 ___ 127 1.6 13.4 19.7 18.1 17.3 5.5 1.6 6.3 5.5 11.0 80232 deere 17 2.6 18.8 26.5 12.8 9.4 8.6 .9 6.0 6.0 8.6 BBY in mms me mA 123 0 13.0 30.9 5.7 6.5 11.4 8.1 5.7 8.1 10.6 BOB mn simi mer 156 3.2 26.9 25.0 .6 5.1 9.0 3.2 10.9 8.3 7.7 45-49. oem 123 3.3 15.4 26.0 8.3 4.9 9.8 .9.8 8.1 4.1 15.4 50-54 ___ 23 4.4 34.8 26.1 4.4 4.4 4.4 4.4 0 0 17.4 Overall percentage_____|_._______ 2.2 18.6 25.8 7.2 8.3 8.8 4.8 7.2 6.1 11.0 Note.— “Eligibility” means that the woman was not pregnant, lactating, or trying to conceive; that neither she nor her husband by her husband if she so wished. had had a sterilizing operation, that she menstruated, and that there was no other reason to believe she could not become pregnant 89 STINAAD0Ud ANV ‘NVId ‘NOILVINIOd ‘XANLS TABLE 36.—Method of birth control used by white married women eligible to use contraceptive drugs—mpast users of contraceptive drugs for contraception (percent) Dia- Con- Intra- | Spermi- With- | Multi- Age in years Number |[Unknown| phragm | dom | uterine cide | Rhythm| Douche | drawal ple None device method DO-24......concsvamnmnanasses 48 0 6.2 27.1 20.8 20.8 6.2 0 8.3 2.1 8.3 DEADY.. .... rrp mm min ERR 69 1.4 14.5 18.8 23.2 20.3 2.9 0 5.8 5.8 7.2 80-84 __ eo. 48 4.2 18.8 20.8 27.1 14.6 2.1 0 6.2 4.2 2.1 85-89 25 0 12.0 40.0 16.0 12.0 4.0 0 4.0 4.0 8.0 40-44 ___ 24 4.2 33.3 37.5 0 0 4.2 4.2 4.2 8.0 4.2 A549. ...... censwinsmassens 5 0 40.0 20.0 0 40.0 0 0 0 0 0 BOB... eemismimm 0 lion immmalmmssmimalens mmm ses msran pss ames sms mn) nm mm mmm] =e mms fo Overall percentage _ ___|-_______ 2.0 17.5 26.6 18.6 15.4 3.4 .6 5.6 4.7 5.6 NoTe.—*“Eligibility’’ means that the woman was not pregnant, lactating, or trying to conceive; that neither she nor her husband by her husband if she so wished. had had a sterilizing operation, that she menstruated, and that there was no other reason to believe she could not become pregnant SAOHLAN ANV SOILSISHLOVIVHO OINONO0DIOIO0S 69 TABLE 37.—Method of birth control used by white married women eligible to use coniraception— never users of contraceptive drugs (percent) Dia- Con- | Intra- | Spermi- With- | Multi- Age in years Number |Unknown| phragm | dom | uterine cide | Rhythm| Douche | drawal ple None device method 20-24... eee em 25 0 12.0 16.0 8.0 20.0 12.0 4.0 0 12.0 16.0 BR ei md 48 2.1 12.5 20.8 12.5 14.6 6.2 4.2 6.2 4.2 16.7 30-34. ees 61 0 18.0 31.2 3.3 4.9 13.1 1.6 4.9 8.2 14.8 IE SIE J Ee 81 0 14.8 28.4 8.7 4.9 8.6 11.1 7.4 11.1 9.9 B04. one ms 116 2.06 25.0 24.1 .9 6.0 9.5 3.4 12.1 9.5 6.9 45-49... eee 105 3.8 14.38 27.6 2.9 2.9 11.4 8.6 8.6 3.8 16.2 50-54 ______________________ 18 5.6 33.3 27.8 0 0 5.6 5.6 0 0 22.2 Overall percentage_____|________ 2.2 18.8 26.5 3.2 5.5 9.8 6.2 7.9 7.2 12.8 Note.—“Eligibility” means that the woman was not pregnant, lactating, or trying to conceive; that neither she nor her husband by her husband if she so wished. had had a sterilizing operation, that she menstruated, and that there was no other reason to believe she could not become pregnant 09 STINAAO0U ANY ‘NVId ‘NOILVINIOd ‘AANLS SOCIOECONOMIC CHARACTERISTICS AND METHODS 61 for the express purpose of rendering the person sterile. A remedial operation is one performed to correct a disease condition of the reproductive system. Sterilization is a consequence of the opera- tion, but not its intent. In this report contraceptive operations are equated with fe- male tubal ligations and male vasectomies. With rare exception, these operations are performed only to prevent future conceptions. Since other types of operations, such as hysterectomies, are rarely done for the sole purpose of sterilization, all sterilizing surgeries, other than tubal ligations and vasectomies, are classified as re- medial operations. CONTRACEPTIVE OPERATIONS Contraceptive operations were found among 23 percent of white couples, with wife aged 20-54 years. More than two-thirds of these operations were male vasectomies. Only 7 percent of wives had had a tubal ligation) but 16 percent were married to vasectomized men. As shown in table 38, the prevalence of contraceptive opera- tions (tubal ligation and vasectomy combined) increased steadily with age of wife to a maximum of 35 percent at ages 35-39, and TABLE 38.— Prevalence of contraceptive operations—white married Health Plan subscribers Tubal ligation Vasectomy Either operation Num- Wife’s age | berin in years sam- | Num- | Percent | Num-| Percent+ |Num-| Percent ple ber | standard | ber standard ber | standard error error error 20-24________ 592 12 | 2.0+0.5 9 1.54+0.4 21 3.5+0.6 08-29. occa 564 35|6.2+ .9 45 8.0+1.0 80 | 14.2+1.3 30-34________ 432 34 | 7.9+1.2 72 | 16.7+1.6 | 106 | 24.5+1.9 85-39_______. 417 35 |8.4+1.2 | 114 | 27.3+2.0 | 149 | 35.7+2.2 40-44 _______ 462 31 |6.7+1.1 109 | 23.4+1.8 (1139 | 30.1+2.0 45-49________ 412 35 | 8.5+1.3 74 18.0+1.8 | 2107 | 26.0+2.0 50-54. _.... 192 14 | 7.3+1.8 19 9.9+2.0 383 | 17.2+2.6 Weighted total. | eenniabeeeees 6.9+ 4 |._____ 16.0+ .6 |.___.. 22.84 .7 1 In 1 case, both husband and wife have had a contraceptive operation. 2 In 2 cases, both husband and wife have had a contraceptive operation. 62 STUDY, POPULATION, PLAN, AND PROCEDURES then decreased in each successively older age group. This decrease after age 39 is due to a decrease in the prevalence of vasectomies rather than tubal ligations. In contrast to vasectomies, the preva- lence of tubal ligations showed no essential variation among the age groups above 29 years. The absence of an age gradient in the prevalence of tubal ligations after 29 years might suggest that these operations are rarely done on women beyond the third decade of life. This, how- ever, probably is not true. The age at which these operations were performed is not known, but we suspect that the prevalence of tubal ligations within cohorts continues to increase until the end of the reproductive period. The age plateau in the prevalence of tubal ligations which we see in these cross-sectional data beginning at 30 years was probably created by an abrupt and marked in- crease in the incidence of these operations several years before the survey. During the 5-year period 1960-64, 279 tubal ligations were performed at the Walnut Creek Kaiser Foundation Hospital. In 1965, all restrictions on the performance of tubal ligations were removed from the hospital policy. From 1965 through 1967, 634 tubal ligations were done. Not all these operations were performed on Health Plan members living in the Walnut Creek Service Area and not all ligated survey subjects had the operation at a Kaiser hospital; but we can safely assume that the high prevalence of tubal ligations among women under 40 years of age in this popu- lation was a recent phenomenon resulting from the change in hospital policy. ) Prior to 1969, vasectomies were never done at the Walnut Creek Kaiser facility. Health Plan members who wished a vasec- tomy had to go to an outside physician who was willing to perform the operation for a fee. There is no reason, therefore, to believe that the high prevalence of vasectomies in the Health Plan popu- lation is atypical of the greater community. The age at which these operations were performed is not known, but the decrease in the prevalence of vasectomies as age of wife increased beyond 39 years indicates that the incidence of these operations had been on the increase for a number of years. REMEDIAL OPERATIONS Twelve percent of the wives had had a remedial sterilizing operation, usually a hysterectomy. The prevalence of remedial operations increased with age from less than 1 percent under 30 years to 26 percent at ages 50-54 years (table 39). Remedial SOCIOECONOMIC CHARACTERISTICS AND METHODS 63 TABLE 39.— Prevalence of remedial sterilizing operations—white married Health Plan subscribers Wife’s age in years Number Percent + standard error D024... cco wviinreinion inion HR i wei 2 0.3+0.2 Sl citi ie esse sR ER 6 1.1+ 4 80-34 eee 26 6.0+1.0 mn snr EE EE ER 43 10.3+1.4 BOSAL oo cu rE RA RR 65 14.1+1.5 BBY... oe econ mo sm sem ms sm so mt hm mE 89 21.6+1.9 50-54 eee 51 26.6+3.0 Weighted total... cs anmmsnsmsnsnsssssnmslssrsssrens 11.84 .6 TABLE 40.—Prevalence of all sterilizing surgery (percent distribution of couples) Wife has had remedial sterilizing Husband or wife has had contraceptive opera- surgery tion Yes No Total Xe roe 3.2 19.6 22.8 IN ccm sons ns snes se wm in 8.6 68.6 77-2 Potlcocinnssrnerssmeosrorse veiw 11.8 88.2 100.0 sterilizing surgery is uncommon among men and none was re- ported in this survey. ALL STERILIZING SURGERY Of the 12 percent of wives reporting a remedial operation, one-fourth had had a prior tubal ligation or were married to vasectomized men. Thus, some form of sterilizing surgery was found among 31 percent of couples (table 40). Under 40 years, the prevalence of surgical sterilization was 24 percent. COMPARISON WITH THE 1965 NATIONAL FERTILITY STUDY Table 41 presents data from the National Fertility Study of 1965 on the prevalence of tubal ligations and vasectomies among 64 STUDY, POPULATION, PLAN, AND PROCEDURES TABLE 41.—Comparison of Walnut Creek Kaiser (WCK) population with 1965 National Fertility Study population: Contraceptive operations among white couples (percent) Tubal ligation Vasectomy Either operation - Wife's age 1965 study 1965 study 1965 study in years WCK WCK WCK United| West United| West United| West States States States 20-54. . cnc 49| 6.0| 6.9| 3.0|10.0|16.0| 8.0 16.0 | 22.9 20-39..c.cncen- 4.0] 5.7% 6.4] 8.71 9.7114.4 | 7.7118.4} 20.8 20-24... __ Sip 1.0) 2.01 1.7{ 40} 1.5] 2:61 5.0 3.5 25-29_..___ 8.2| 2.5; 6.2] 3.4) 6.7] 8.0} 6.6] 9.2 14.2 30-34......) 5.8;12.0) 7.9; 4.6)111.2716.7]| 10.1 | 253.2 | 24.8 35-89... 58: 6.21 8.4 5.0718.5{27.8 10.8 21.71 85.7 40-54... _____. 6.21 6.6| 7.5 2.2110.5:18.0} 8.4} 17.1} 25.2 40-44. ___ 5.5) 7.2] 6.7) 4.0] 9.4123.4| 9.5] 16.6 80.1 45-49... .. 6.8 9.1 8.5 1.4 4.5 | 18.0 8.2] 13.6 26.5 50-54... 6.3] 2.6 7.3 Ni.2.6% 9.9; 7.0] 85.2; 17.2 white couples, with wife aged 20-54 years, for the Nation as a whole and for the Western States. (These prevalence data were kindly provided by Charles Westoff and Norman Ryder, directors of the 1965 National Fertility Study.) The prevalence data for the Walnut Creek Kaiser population are repeated in table 41 so that they can be easily compared with the national and regional data. The prevalence of vasectomies in the Walnut Creek Kaiser population is five times greater than the 1965 national average for white couples in the same age range (16 percent compared with 3 percent). Male sterilization, however, is much more common in the West than in other parts of the country. In 1965 the prevalence of vasectomies in the Western States was 10 percent, or three times the national average. This, however, is still considerably below the prevalence found in the Walnut Creek Kaiser population. Although the Contraceptive Drug Study (CDS) survey took place 214 years after the national survey, the difference between the Walnut Creek Kaiser population and the regional sample in the prevalence of vasectomies is too large to be due to an increase in the prevalence of these operations. The prevalence of vasec- tomies among western couples, with wife aged 18-39 years, in- creased by only 2 percentage points between 1960 and 1965 (from SOCIOECONOMIC CHARACTERISTICS AND METHODS 65 7 to 9 percent). Furthermore, the difference in prevalence between the Walnut Creek Kaiser population and the western region as a whole is much less marked in the age groups below 40 years than it is in the older age groups (table 41). Since there is no reason to believe that vasectomies are more prevalent among Health Plan members than among other couples in the Walnut Creek area, the unusually high prevalence of these operations in this population, particularly in the oldest age groups, suggests that male steriliza- tion may have gained acceptance first on the Pacific coast and spread to other parts of the West. Although the prevalence of tubal ligations in the Western States in 1965 was somewhat higher than the national average (6 percent compared with 5 percent), these operations are not unusually common among westerners. In the Northeastern States, where the Catholic Church has its greatest influence, the preva- lence of tubal ligations is low and this pulls down the national average. The prevalence of tubal ligations in the Walnut Creek Kaiser population is, as expected, higher than the 1965 regional average. The overall difference, however, is small (7 percent compared with 6 percent). Because the 1965 change in the policy on tubal ligations at the Walnut Creek Kaiser Foundation Hospital would have had little, if any, effect on the prevalence of these operations among the cohorts which were 40 years of age and over at the time of the CDS survey, we expected the prevalence of female contraceptive operations among Kaiser wives to be higher than the regional average only in the age groups under 40 years. In general, this expectation was borne out. Except for the age group 50-54 years, where the prevalence of tubal ligations among western wives is unusually low, there is no essential difference in prevalence be- tween Kaiser wives 40 years of age and over and their age-group peers in the regional sample (table 41). Under 40 years, the preva- lence of tubal ligations among Kaiser wives is considerably higher than that among western wives in all but one age group, 30-34 years (table 41). Since the proportion of western wives in this particular 5-year age group reporting tubal ligations in 1965 is unusually large (12 percent), the overall difference in prevalence between the Kaiser wives and the regional sample in the age-range 20-39 years is not as marked as it might otherwise be. As shown in table 42, the prevalence of remedial sterilizing surgery in the Western States is comparable to the 1965 national average. The prevalence of remedial operations in the Kaiser population is somewhat below the national average (11.8 percent 66 STUDY, POPULATION, PLAN, AND PROCEDURES TABLE 42.—Comparison of Walnut Creek Kaiser (WCK) population with 1965 National Fertility Study population: Remedial sterilizing operations among white women (percent) 1965 study Wife’s age in years WCK United West States DBAs. coms mane smh EEE EERE 13.1 13.4 11.8 D000 rr ree EE Rl 5.9 5.7 4.8 0h ol eR ——— 1.1 1.0 .3 DBD iin miei wi ie i mi hi 2.5 1.7 1.1 80-84 eee 7.2 8.0 6.0 S800 es wes Re Ee 11.8 10.8 10.3 BEB es martes ie Dem die ie Ew a 22.5 25.3 20.0 BOA 2 ee ci en Re SE 17.3 19.6 14.1 ABA ti ene Pred 25.9 22.7 21.6 CTRL NS 26.0 38.5 26.6 compared with 13.1 percent for women 20-54 years). In the 1960 Growth of American Families Study, remedial operations were found to be more common among less educated women.? The lower prevalence of these operations among Health Plan members may, therefore, be a reflection of their relatively high level of education. SOCIAL VARIABLES ASSOCIATED WITH CONTRACEPTIVE OPERATIONS Women who reported contraceptive operations in the CDS survey were not asked to give the reasons for the operation. Therefore, we do not know how many of these operations were done because pregnancy would have endangered the wife’s health or because of Rh incompatibility, genetic disease or other factors operating against a favorable outcome to pregnancy. The high prevalence of contraceptive operations, vasectomies in particular, indicates, however, that most of these operations were done simply because the individuals involved did not wish to have any more children. The factors that prompted these couples to curtail their fer- tility by surgical sterilization rather than conventional methods are also unknown, but the use of this method of fertility control was found to be associated with the wife's piriey, the couple’s religion, and their education. TABLE 43.—Prevalence of contraceptive operations, by number of live births Number of live births Wife's age in years 0 1 a 3 4 5 or more Number | Percent | Number | Percent | Number | Percent | Number | Percent | Number | Percent | Number | Percent 20-24______________ 2 1.1 1 0.5 9 5.8 7 18.9 2 88.3 |-crvimes|on rau BS DD rin ices 0 0 1 1.1 19 9.0 39 29.3 15 33.3 6 31.6 80-84... .cccinnneian 0 0 2 5.1 20 13.6 43 33.9 30 46.9 11 32.4 85-89... comnts 2 18.2 7 17.1 30 31.9 45 33.8 34 47.9 31 46.3 40-44 _____________ 0 0 5 12.2 28 25.2 45 34.9 35 44.3 26 33.3 45-49______________ 0 0 6 11.3 27 25.0 29 29.0 25 39.1 19 39.6 BO-B4.. conc om irra 1 3.8 4 11.4 10 18.2 9 23.1 7 28.0 2 22.2 SAOHLAN ANV SOILSISILOVIVHD DINONODHOIDOS L9 68 STUDY, POPULATION, PLAN, AND PROCEDURES PARITY As one would expect, the prevalence of contraceptive opera- tions increases with the parity of the wife. This trend does not continue, however, beyond the fourth birth and actually appears to decrease somewhat among couples in which the wife has borne five or more children (table 43). It is unusual to find contraceptive operations when the wife is childless. There were only five such cases in the entire sample. One case was that of a 20-year-old woman with subacute nephritis and hypertension who was sterilized at the age of 19 years, prior to her impending marriage. The other four cases are ones in which the husbands were vasectomized. In two of these cases, it is known that the husbands had children from a previous marriage and were already sterilized when they married their current wives. This may also be true of one or both of the two remaining cases. Because the prevalence of voluntary sterilization increases with parity, couples who use this method of fertility control have had, on the average, more children than others (table 44). The dif- ference in mean parity between the two groups varies by age of wife and is greatest in the age groups under 35 years where family size is still incomplete. At ages 35-44 years, where both the num- ber of births and the prevalence of contraceptive operations are highest, the difference in mean parity is lowest (approximately 0.5 births). The 1960 Growth of American Families Study found that vol- untary sterilization was associated with unplanned or unwanted pregnancies.’ Information on contraceptive failure and desired family size was not requested in the CDS survey and it is not pos- sible to say how many of these couples elected sterilization because they had failed to keep family size down to that desired through conventional methods. Although the incidence of contraceptive failure may have been higher among voluntarily sterile couples than others, this would not be the only factor contributing to their larger number of births. Except in the infrequent case of steriliza- tion for medical reasons, couples who elect sterilization surely must have had all the children they planned to have. This would not necessarily be true of all other couples. Biological impair- ments to fertility will have prevented some of them from achieving their desired family size. Regardless of the difference in parity, the important thing to note is that voluntarily sterile couples do not have excessively large families. The modal number of children born to them is either two TABLE 44.— Difference between voluntarily sterilized couples and others in number of live births Percent distribution by number of live births Mean Difference in number of means + Wife's age, in years, and couple’s status births | standard error 1 2 3 4 5 or more 20-24: Voluntarily sterilized... . convince mmmmmns 5 4.8 42.9 33.3 9.5 2.28 All others. 4| 387.0| 258| 5.3 7 107) 1-02 25-29: Voluntarily sterilized -- - ooo __ 1.2 23.8 48.8 18.8 3.11 1.164 12 All others_______ ee 13.3 18.5 39.8 19.5 6.2 1.95 E : 30-34: Voluntarily sterilized. ...comssesmnmsmanmnmmmn 1.9 18.9 40.6 28.3 3.32 854+ .13 AN ONETS. ce ee eee ecm mmm mmm me mm m———— 4 11.4 37.0 25.8 10.4 2.47 : : 35-39: Voluntarily sterilized. ooo _______ 4 4.7 20.1 30.2 22.8 3.46 49+ 16 AH others ...co- nies onsen snamem mR wn 4 12.7 23.9 32.8 13.8 2.97 : : 40-44: Voluntarily sterilized... _________ 3.6 20.2 32.4 25.2 3.48 554+ 15 AL OLhErS. .... ev cece mice mmm mmm mm 6.9 11.2 25.9 26.2 13.7 2.93 2 : 45-49: Voluntarily sterilized... ___________.____ 5.7 25.5 27.4 23.6 3.30 85+ .16 ANWOLNOIS. oo amin mimi mmm RRR 11.9 15.5 26.7 28 .4 12.9 2.45 : ’ 50-54: Voluntarily sterilized... _____________ .0 12.1 30.3 21.3 2n.2 2.713 60+ .26 AN OthOrS. oc cui cinnimenmmnsaanwecn denen 0 19.9 28.8 19.2 11.5 2.13 E STOHLAW ANV SOLLSINALOVIVHO DINONODE0IO0S 69 70 STUDY, POPULATION, PLAN, AND PROCEDURES or three, depending on the wife’s age (table 44). In all age groups, more than half have had three or fewer children and most have had no more than four. Women with tubal ligations tend to have had more births than those married to vasectomized men. Twenty-three percent of those with tubal ligations have had five or more births compared with 13 percent for women whose husbands have vasectomies. This dif- ference may be a reflection of restrictive hospital policy which has traditionally required “multiparity” for the performance of tubal ligations in the absence of medical indications. RELIGION The prevalence of contraceptive operations varies according to the couple’s religion (table 45). It is lowest when both husband and wife are Catholic (15 percent) and highest when neither is Catholic (26 percent). The variation in the prevalence of tubal ligations by religion is small. The lowest prevalence was found among Catholic couples (5 percent) followed by mixed marriages in which the wife is Catholic (6 percent). The prevalence of tubal ligations among non- Catholic women was approximately 7 percent and showed no dif- ference by husband’s religion. The prevalence of vasectomies among Catholic men married to non-Catholic women was about the same as that among those TABLE 45.—Prevalence of contraceptive operations, by religion of couple and by wife's education Tubal Vasectomy Either ligation (percent + operation Characteristic (percent + standard (percent + standard error) standard error) error) Religion of couple: Both Catholle. .. vee cram cmnmemes 5441.0 9.8+1.2 15.2+1.4 Only wife Catholic. _.___________ 6.4+1.4 14.441.9 20.8+2.2 Only husband Catholic. ___.______ 7.1+1.8 11.1+2.1 18.2+2.4 Neither Catholic....____________. 7.3% 18.3+ .8 25.6+ .9 Wife's education: Not high school graduate. ___..____ 11.56+1.4 | 21.2+1.7 32.6+2.0 High school graduate... __..__.___ 6.04 .6 17.5+1.0 23.5+1.1 Some college _ _________________. 6.74+ .9 13.7+1.2 20.4+1.4 College eradunte. ...cuonveneveens 4.94+1.0 9.44+1.3 14.3+1.5 SOCIOECONOMIC CHARACTERISTICS AND METHODS 71 with Catholic wives (11 percent compared with 10 percent). Vasectomies were more common among non-Catholic men married to Catholic women (14 percent), but not as common as that found in non-Catholic marriages (18 percent). Thus a Catholic woman married to a non-Catholic man may be somewhat more likely to have a tubal ligation or agree to a vasectomy on her husband than one married to a Catholic. A Catholic man married to a non- Catholic woman, on the other hand, appears not to object if his wife has a tubal ligation but may be no more likely to have a vasectomy than one married to a Catholic. We expect both tubal ligations and vasectomies to be less com- mon among Catholic than non-Catholic couples. What is remark- able is that the prevalence of contraceptive operations among Catholic couples is as high as it is. The Catholic Church explicitly prohibits operations whose intent is sterilization and one might assume that Catholic couples who elect sterilization are ones who are no longer active in their church. This, however, is not true. Over 60 percent attend church at least once a month, and 45 percent attend weekly. Although the prevalence of contraceptive operations was higher among Catholic couples who attend church infrequently or not at all (20 percent), 13 percent of those who attend at least once a month reported tubal ligation or vasectomy. EDUCATION The prevalence of contraceptive operations varies inversely with the education of the wife (table 45). Both tubal ligations and vasectomies are more than twice as common when the wife has less than a high school education than when she is a college grad- uate (83 percent compared with 14 percent for both operations combined). Some readers may be surprised to learn that the high- est prevalence is found among the least educated. It might be assumed that, because of lack of understanding, poorly educated couples would be less accepting of sterilization, vasectomy in par- ticular. The opposite seems to occur. Even among Catholic couples, where it might be thought that practices deviating from Church dogma would be more common among the better educated, the prevalence of contraceptive operations is highest among the least educated (table 46). An inverse relationship between the prevalence of voluntary sterilization and wife’s education was also found in the 1960 Growth of American Families Study. This relationship could not be explained on the basis of contraceptive failure or excess preg- 72 STUDY, POPULATION, PLAN, AND PROCEDURES TABLE 46.—Prevalence of contraceptive operations, by wife’s education and religion of couple and by wife's education and age at first pregnancy Religion of couple Wife’s age at first pregnancy (years) Wife’s education Cath- Non- | Un- 25 olic [Mixed| Cath-| der |18-19|20-24| and olic 18 over Not high school graduate. ____ 20.9 128.8 136.9 38.2 |33.9(30.8| 28.3 High school graduate._.______ 15.1 121.0 726.61 32.1 27.7|22.4 21.4 Some college _ _ _____________ 16.5 17.8 122.21 26.71 24.1 22.3 19.4 College graduate..._____._____ 0 0 I Ee Se 15.8 | 13.5 nancies, and the investigators speculated that physicians may be more likely to recommend sterilization to couples with less educa- tion than to others.? The gynecologists at the Walnut Creek Kaiser Hospital may "tend to recommend tubal ligation more often to their less educated patients, but this would not explain the high prevalence of vasec- tomies among the husbands. Although husbands and wives covary in their education, the prevalence of vasectomy by education of husband was examined separately and found to give the same result. The less educated a husband is, the more likely he is to have a vasectomy. Since vasectomy was not offered to Health Plan members before 1969, the higher prevalence of these operations among the poorer educated cannot be ascribed to differential treatment by physicians. There is another factor, however, which may explain in part why contraceptive operations decrease with education, and this is the age at which women begin their families. The age at which a woman begins childbearing and her education are closely related. In general, the less a woman’s education, the younger she was when she began to have children. Approximately one-third of the women in the Contraceptive Drug Study survey with less than a high school education were first pregnant before the age of 18 years. Women with college degrees, in contrast, had usually delayed childbearing until well into the third decade of life or later. Over half had never had a pregnancy before the age of 25 years. Having started their families at an earlier age, women with less education are likely to achieve their desired family size at a younger age than their more educated peers. It is reasonable, therefore, for these young couples with completed families to elect SOCIOECONOMIC CHARACTERISTICS AND METHODS 73 sterilization rather than to continue using methods of contracep- tion better suited for timing pregnancy than curtailing fertility. When education is held constant, the prevalence of contraceptive operations does decrease as age at first pregnancy increases (table 46). The relationship between wife’s education and prevalence, however, is still seen regardless of age at first pregnancy. REFERENCES 1. RypER, N. B., and WEsTOFF, C. F.: Use of oral contraception in the United States, 1965. Science 153: 1199-1205, Sept. 9, 1966. 2. WEestorF, C. F., and RYDER, N. B.: United States: Methods of fertility control, 1955, 1960, and 1965. Studies in Family Planning 17: 1-5, Feb. 1967. 3. WHELPTON, P. K., CAMPBELL, A. A., and PATTERSON, J. E.: Fertility and Family Planning in the United States. Princeton, N.J.: Princeton Uni- versity Press, 1966. | Ly h : E RI ) } fo APPENDIX CONTRACEPTIVE DRUG STUDY SURVEY QUESTIONNAIRE This questionnaire is easily and quickly filled out. Most of the questions can be answered by simply checking the box next to the answer which fits you best. Ignore the small numbers appearing alongside the answer box. They are there to help us count your answers. 1. Birth control pills are often taken for rea ons other than preventing pregnancy. Have you ever taken birth control pills for any reason whatsoever? . . . . . . . . . + 4 4 4 eo. [] YES [] NO 1 2 If NO, go to Juestion 8 Please answer Question 2 only if you have ever taken birth control pills. 2. Are you taking birth control pills at the present time? . . . . . . . . [] YES [] NO 7 1 If you are now taking birth control pills, please answer Questions 3-4. 3. Are you taking the birth control pills to prevent pregnancy? . . . . . . [1] YES [J] NO 8 1 2 4. Did you get your last prescription for the pills from a Kaiser doctor? . . . [1] YES [] NO 9 + 2 If you have stopped taking birth control pills, please answer Questions 5-6. 5. When you last took the birth control pills, were you taking them to prevent pregnancy? wim wow Ww 3 YES NO 0 0 0 6. In the following list, check the reason which best describes why you stopped taking the birth control pills. n-12 [J] To become pregnant. [[] No longer needed. Please explain: [[] Did not like side effects. What were they? [] Worried that continued use might be harmful. Please explain: [[] Other reason. Please write in: ({] 76 STUDY, POPULATION, PLAN, AND POCEDURES Please answer Question 7 if you have ever taken birth control pills for any reason whatsoever. a. Indicate with an “X” every month of each year during which you took birth control pills. If, for any calendar year, you took the pills during all 12 months, put an “X” in the box labeled “Entire Year” marking every month of that year. b. Indicate with an “0” every month since Janurary 1960 during which you were pregnant. You may find it easier to indicate the months during which you were pregnant before indicating the months during which you took birth control pills. X — stands for birth control pills O — stands for pregnancy JAN | FEB [MAR | APR | MAY | JUN | JUL | AUG | SEPT | OCT | NOV | DEC 7. On the calendar below: instead of REMEMBER: ow 1960 1961 1962 1963 1964 1965 1966 1967 1968 COMMENTS: 13-16 17-20 21-24 25-28 29-32 33-36 37-40 41-44 45-48 49-52 53-56 SOCIOECONOMIC CHARACTERISTICS AND METHODS 8. Have you ever taken female hormones, other than birth control pills? . . [] YES [7] NO ¥ 2 9. Are you now taking female hormones, other than birth control pills? . . . [] YES [] NO 1 2 10. How many times have you been pregnant? (Be sure to count miscarriages, stillbirths and children who have died since birth.) . If you have never been pregnant, please answer Question 11. 11. Have you ever tried to become pregnant? . . . . . . . . . . . . [1 YES [] No. 1 2 If you have been pregnant one or more times, please answer Question 12. 12. How old were you the first time you became pregnant? If you have been pregnant more than once, please answer Question 13. 13. How old were you the last time you became pregnant? . M4. Areyounow pregoattl . . «ox se ww wow om ww www ws YES NO y preg: 0 ni 15. Are you now trying to become pregnant? . . . . . . . . . . . . [] YES [J] NO 1 2 16. Do you intend to become pregnant in the future? . . . . . . . . . [J YES [] NO 1 2 17. Have you delivered a baby or lost a pregnancy within the past six weeks? [1] YES [] NO 1 2 18. Are you now breast-feedinga baby? . . . . . . . . . . LL. [J YES [J] NO 1 1 19. How many live births have you had, altogether? . 20. How many living children do you have, not counting stepchildren or adopted children? . in FABER E LRT BES 21. How many adopted children do you have? 22. How many stepchildren do you have living with you? 7 61 62 63-64 65 68-69 74 75-76 77-78 79 80 78 STUDY, POPULATION, PLAN, AND POCEDURES Have you had a tubal ligation (tubes tied)? . . . . . . . . . . . [] YES [] NO Have you had any operation, other than a tubal ligation, which makes it impossible for you now to become pregnant? . . . . . . . . . . . [] YES [] NO 1 If YES, what kind of operation was it? 25. Have you stopped having menstrual periods? (If you have TEMPORARILY STOPPED menstruating because of pregnancy, breast-, fending or medica- tion, answer NO to this question.) . . oo vow [] YES [] NO 1 2 26. Do you have any other reason to think you may not now be able to conceive (become pregnant)? . . . . . . . . . wHR EEE EE [] YES [] NO 1 2 If YES, what is the reason? If you are currently married and living with your husband, please answer Questions 27-28. 27. Has your husband had a vasectomy (cords tied)? . . . . . . . . . [1 YES [] NO 1 2 28. Is there any other reason to think that your husband may not now be able tofatherchildren? . . . +. . +» « + 4 + 4 4 4 «0 0 0 0 ws [] YES [J NO 1 2 If YES, what is the reason? 29. Some commonly used methods of birth control are listed below. Please read the complete list. METHODS OF BIRTH CONTROL If you are now taking birth control pills, please check the method(s), [[] Rhythm (safe period) if any, which you used just before began to t: ills. You began to take the pills [7] Diaphragm and jelly or cream If you are not taking birth control pills, please check the method (s), [] Condom (rubber, prophylactic) if any, which you are now using. [] Intrauterine device (loop, coil) [[] Douche after intercourse [] Coitus interruptus (withdrawal) O Vaginal jelly, cream, foam or suppository [] None SOCIOECONOMIC CHARACTERISTICS AND METHODS 30. What is your birthdate? . . . . . . . . . MONTH resem Saiemee: YT cessed 81. What is your race? [J] White [[] Negro [] Oriental [[] Other 1 2 3 4 82. What is your religious affiliation? [[] Protestant [J Catholic [] LDS [] Jewish [] Other [[] None 3 x 3 4 5 6 83. Which of the following best describes the frequency of your attendance at religious services? [] Atleast [] At least [] Less than [] Never "once a week 2 once a month * once a month 4 34. What is your marital status? [7] Never Married [J Married [] Separated [7] Divorced [] Widowed 1 2 3 4 5 If you have ever been married, please answer Questions 35-36. 85, Whenwereyoufistmarried? . . . . . « . MONTH YEAR 36. Have you been married more than once? . . . . . . . . . . . [] YES [J NO 1 2 If you are currently married, please answer Questions 37-38. 87. What is your husband's birth date? . . . . . MONTH YEAR 38. What is your husband's religious affiliation? [] Protestant [7] Catholic [] LDS [] Jewish [] Other [[] None 1 2 3 4 5 . 39. Please circle the last year of school completed by the following persons: YOURSELF: Elementary School Grade: 1 2 3 4 5 6 7 8 High School Year: 1 2 3 4 College Year: 1 2 3 4 or more. Degree: Graduate School Year: 1 2 3 4 or more. Degree: YOUR HUSBAND: (If you are currently married) Elementary School Grade: 1 2 3 4 5 6 7 8 High School Year: 1 2 3 4 College Year: 1 2 3 4 or more. Degree: Graduate School Year: 1 2 3 4 or more. Degree: 79 20-23 24 25 26 27 28-31 32 33-36 37 38-39 40-41 80 STUDY, POPULATION, PLAN, AND POCEDURES YOUR FATHER OR GUARDIAN: ) aa ~ Elementary School Grade: 1 2 3 4 5 6 7 8 High School Year: 1 2 3 4 College Year: 1 2 3 4 or more. Degree: Graduate School Year: 1 2 3 4 or more. Degree: 40. Are you currently employed: [] Full time [] Part time [] Not employed #4 i" 2 3 . If you are currently married, please answer Questions 41-44. 41. What kind of work does your husband do, and in what kind of business or industry does he work? 451 Please be as specific as possible. For example: clerk at the department of motor vehicles; electronics engineer for an instrument company; truck driver for a sand and gravel company; salesman for a book publisher. (kind of work) (kind of business or industry) 42. What kind of work did your father or guardian do when you were in your teens? 52-58 (kind of work) (kind of business or industry) 43. For the purpose of this survey, we need a rough indication of your total family income in 1967 from 59-60 all sources before taxes and other deductions. Please check the category in which your family income falls. O $ 5,000 or less 4 $11,000 to $12,000 o or . [J $ 5,000 to $ 7,000 [] $12,000 to $13,000 “2 oo” 0 $ 7,000 to $ 8,000 3 $13,000 to $14,000 o oo [] $ 8,000 to $ 9,000 [[] $14,000 to $15,000 “ 0 [1 $ 9,000 to $10,000 [] $15,000 to $20,000 os n 0 $10,000 to $11,000 O $20,000 or more ” n 44. Counting yourself, how many family members are living in your home sa2 at the present time? . ’ iw RE me . There are a few final questions for EVERYONE on the back side of this page. 47. SOCIOECONOMIC CHARACTERISTICS AND METHODS How many years have you lived at your current address? . How many years have you lived in Contra Costa County? If you do not live in Contra Costa County, please write in the town and county where you now live. (town) (county) How many years have you lived in California? Please use the stamped, addressed envelope which we have provided to return the questionnaire to us. Thank you very much for helping in this study. 81 65-66 67-68 Chapter 3 Changes in the Use of Oral Contraception and Other Methods of Fertility Control Between 1968 and 1971 NANCY R. PHILLIPS Methods of fertility control used by women having an Automated Multitest Laboratory examination between January 1969 and June 1971 are compared with the methods reported by women in the February 1968 sample survey of the Kaiser Foundation Health Plan population. Use of oral contraceptives.—Use of oral contraceptives increased between 1968 and 1969 from 34 percent to 39 percent and then decreased to 29 percent in 1971. All methods of fertility control.—The prevalence of surgical steri- lization increased from 1968 to 1971. The proportion using the intra- uterine device also increased. All other methods of fertility control showed little change. Surgical sterilization.—The major part of the increase in surgical sterilization of 6 percent was due to an increase in male sterilization. The methods of fertility control used by women who volun- tarily presented themselves for multiphasic health checkups in the Automated Multitest Laboratory (AML) at the Kaiser Foundation Medical Center in Walnut Creek between January 1969 and June 1971 are here compared with the methods reported by women who participated in the February 1968 sample survey of the Kaiser Foundation Health Plan population, described in part A, chapter 2. USE OF ORAL CONTRACEPTIVES Among women aged 20-44 years who underwent a health checkup in the AML during the specified period, the proportion who were taking oral contraceptives (OC’s) is shown by month and year of examination in figure 7. None of the women on whom this figure was based was known to be pregnant or less than 84 STUDY, POPULATION, PLAN, AND PROCEDURES 0.5 wv J > 0.4 ed fre oa. J O g ox fre 5 U 0.3 seed -—t << ox o oO “ 3 0.2 2 Z o = oc Oo oa. 0 01l- = ox a. oll | | aT | JAN JULY JAN JULY JAN JUNE 1969 1970 1971 MONTH AND YEAR OF EXAMINATION FIGURE 7.—Proportion taking oral contraceptive drugs among women 20-44 years of age having Automated Multitest Laboratory health examina- tions, January 1969-June 1971. 2 months’ postpartum at the time of the AML examination. Eleven percent were unmarried or separated from their husbands. Approximately 2 percent were nonwhite. Figure 7 shows that the proportion of OC users among these women declined during 1969 and the first part of 1970, from approximately 39 percent of those examined during the first half of 1969, to 33 percent during the second half of that year, to 81 percent during the first 6 months CHANGES IN USE OF ORAL CONTRACEPTION 85 of 1970. From July 1970 through June 1971, the proportion of OC users remained fairly constant, at about 29 percent. The observed decrease in OC use cannot be attributed to a shift in age distribution or other social factors associated with OC use, since the composition of the AML examinees did not change with respect to such characteristics during this 2l4-year period. In February 1968, 31 percent of the women aged 20-44 years in the Health Plan population survey sample were using oral contraceptives. This is not sufficient basis, however, for concluding that the use of OC’s by this population increased between com- pletion of the earlier survey and the starting date for inclusion in the AML sample. Although women having health checkups were comparable to the 1968 survey subjects with respect to age, race, marital status, and religion, they tended to be somewhat better educated. Twenty-two percent of the AML subjects, but only 17 percent of the earlier survey sample, were college graduates. Only 8 percent of the AML sample, but 14 percent of the survey sample, did not have a high school diploma. Because of the gen- erally higher educational attainment of the AML subjects, and because none was pregnant, the prevalence of OC use among the AML examinees would be expected to be higher than among the general Health Plan population from which they came. In order to equalize the two groups for statistical purposes, pregnant women were eliminated from the survey sample and the propor- tion of OC users among the remaining sample subjects was adjusted to the educational distribution of the AML examinees. . The estimated prevalence of OC users in February 1968 among women comparable to the AML examinees was 34 percent, 5 per- centage points lower than during the first 6 months of 1969. It appears, therefore, that the use of oral contraceptives continued to increase in this population after the survey, then began to decrease in 1969. One can only speculate on the factors that may have prompted the observed decrease in the use of contraceptive drugs in this population. It should be noted that most of the decrease occurred before the hearings of Senator Gaylord Nelson’s subcommittee on the safety of oral contraceptives, which received wide coverage in the news media. Those hearings were held in January and February of 1970, 1 year after the onset of the decrease in OC use noted here. Since the decline in OC use among these women was very small during the first half of 1970, and the proportion of OC users had stabilized by midyear, the hearings had no observable effect on the behavior of this population of women. 86 STUDY, POPULATION, PLAN, AND PROCEDURES The first reports of a statistical association between throm- boembolic episodes and OC use were published in April 1968. 2 But, since the gynecologists at the Walnut Creek Kaiser Founda- tion Hospital continued to prefer OC’s for contraception unless contraindicated, the decreased use of OC’s during 1969 cannot be attributed to a change in medical advice. Rather, it appears that there was increased concern among the women themselves over the possible health hazards of long-term OC use. Having accepted oral contraception earlier than most women, these women had, by 1969, acquired considerable exposure to these agents. It is perhaps significant, therefore, that the decrease in OC use was much more marked in the older age groups, where the accumu- lated exposure was greater. Among AML examinees aged 35-44 years, OC use decreased by one-third between 1969 and 1971, from an average monthly proportion of 29 percent for the first half of 1969 to 19 percent in 1971. At ages 20-24 years, however, users of oral contraceptives decreased only 4 percentage points, from 54 to 50 percent. ALL METHODS OF FERTILITY CONTROL Table 47 compares the 1968 survey sample and the AML examinees with respect to the proportions using various methods of birth control, by year of examination. All women represented in table 47 are 20-44 years of age. Women who were unmarried or separated from their husbands are included. Both observed and adjusted proportions are shown for the survey sample. The prevalence of surgical sterilization, which was high in 1968, had increased since that survey. Aside from this rise, and the decreased use of oral contraception, the only method of fertility control showing a marked change in frequency of use is the intrauterine device: the proportion of women using this method increased from 2 percent in 1968 to 6 percent in 1971. It can be assumed that, in the two samples studied, nearly all women who were not using any method of contraception were either unmarried, trying to conceive, postmenopausal, or otherwise not at risk of pregnancy. Only 4 percent of married women sur- veyed in 1968 who were then at risk of an unplanned pregnancy were using no method of contraception. SURGICAL STERILIZATION The term “surgical sterilization” includes vasectomy of the husband of a married woman, tubal ligation of the woman, and remedial operations resulting secondarily in sterilization, such as CHANGES IN USE OF ORAL CONTRACEPTION 87 TABLE 47.—Relative frequency of various methods of fertility conirol used by women 20-44 years of age Percent of pop- ulation survey, Percent! of AML? Feb. 1968 examinees (N =2,861) Method of fertility control Ob- Ad- 1969 1970 1971 served |justed® | (N= (N= (N= 4,873) | 5,683) | 4,034) Oral contraceptive drugs_________ 31.2 34.3 36.9 30.1 29.1 Intrauterine device______________ 2.2 2.3 2.7 4.8 6.2 Diaphragm_____________________ 5.0 5.7 4.1 4.7 4.9 COnAOM cue nn comrnnsmmmasmnm 6.9 7.2 6.4 8.3 5.9 Spermicide alone_.______________ 2.2 2.4 2.8 4.2 4.6 Rhythm_______________________ 2.2 2.6 2.8 2.8 2.9 Posteoital douche_______________ 1.3 1.4 9 .8 +9 Coitus interruptus. ____________ 1.9 1.9 1.2 1.5 1.3 Multiple methods_____._________ 2.2 2.5 1.2 29 .2 Surgical sterilization_____________ 24.8 23.1 21.4 26.0 29.5 No method. _.__._.______________ 17.3 13.6 13.5 15.9 14.4 Method unknown. _____________ 2.7 3.0 6.3 0 0 ! Average monthly percent. ? Automated Multitest Laboratory. 3 Adjusted to educational level of AML examinees after pregnant women were excluded. 4 It is known, however, that none of these women was taking OC’s. In the case of the 1968 survey sample, it is also known that none was surgically sterilized or married to a vasectomized man. hysterectomy. Information about the type of operation was not available for those AML examinees who were seen during 1969. The numbers of women with hysterectomy or tubal ligation among the 1970 and 1971 AML examinees are known, but not the number married to vasectomized men. Since a woman may undergo hyster- ectomy after tubal ligation or after vasectomy of her husband, it is not possible to determine the total number of women married to vasectomized men by simply subtracting the women who had undergone sterilizing operations from the total reporting surgical sterilization. The prevalences of tubal ligation, remedial sterilizing operations, and women married to vasectomized men among the 1968 survey sample are each known and have been described in detail in chapter 2. As shown in table 48, the prevalence of remedial sterilizing operations had not increased in this population since 1968. The 88 STUDY, POPULATION, PLAN, AND PROCEDURES TABLE 48.— Prevalence of sterilizing female surgery among women 20-44 years of age Percent of pop- | Percent of multi- ulation survey, | phasic examinees Feb. 1968 Type of operation Ob- Ad- 1970 1971 served | justed ! TUDE] HEtioN conve vo nme c wma mmmm mmm 6.2 6.2 7.8 8.4 Remedial operation... ___________________ 6.9 6.8 6.3 6.5 Either type of operation.......--c-vccnses 12.5 12.5 13.2 13.8 1 Adjusted to educational level of Automated Multitest Laboratory examinees after excluding pregnant women. prevalence of tubal ligation, however, had increased. Among non- pregnant women with the educational attainment of the AML examinees, the prevalence of tubal ligation increased 2 percentage points, from 6 percent in 1968 to 8 percent in 1971. The prevalence of all surgical sterilization among these women and their hus- bands, however, increased 6 percentage points (table 47). It appears, therefore, that the major part of the increase is attributable to voluntary male sterilization. The apparent decrease in the prevalence of surgical steriliza- tion between 1968 and 1969 is thought to be an artifact of the data, stemming from incompleteness of information about the methods of fertility control used by those 1969 AML examinees who were not taking oral contraceptives (table 47). Assuming that the distribution of methods among those for whom this information is missing was the same as that for other nonusers of OC’s, the prevalence of surgical sterilization may be calculated to have been nearly 24 percent, a figure slightly higher than the 1968 adjusted estimate. (In the 1968 survey, the term “missing information” applies only to those instances in which a nonuser of oral contraceptives gave negative answers to the questions on sterilizing surgical operations and failed to indicate her current method of contraception.) It is known that the prevalence of tubal ligation had been increasing in this population for several years before the 1968 survey. The age-specific prevalence rates observed in 1968 sug- gested that the prevalence of vasectomy had also been increasing for many years. Since the use of oral contraception was increasing concurrently, the recent increase in voluntary sterilization cannot CHANGES IN USE OF ORAL CONTRACEPTION 89 be attributed to a decline in oral contraception. It is possible, however, that the rate of increase accelerated with the decreased use of the oral contraceptives. REFERENCES 1. INMAN, W. H. W., and VEsseEY, M. P.: Investigation of deaths from pulmonary, coronary, and cerebral thrombosis and embolism in women of child-bearing age. Br Med J 2: 193-199, Apr. 27, 1968. 2. VEsSEY, M. P,, and DoLL, R.: Investigation of the relation between use of oral contraceptives and thromboembolic disease. Br Med J 2: 199-205, Apr. 27, 1968. TEL ENTE EE Ee Se I. g + ) = oo hn = Timte== th a So nk i. u Sa Sy Sipe mae Eee E or = i: . Ik "a Ea Argel EE - fie iT Ty wt oo 2 i Th wr Eh a : «A= et - N vr J R Co gn Sapp FR Ps i Eeiraiel She Ale . = Chapter 4 Automated Multitest Laborary Phases and Quality-Control Procedures I. R. FI1sCH AND S. H. FREEDMAN From December 1968 until October 1972 a complete Automated Multitest Laboratory (AML) battery was obtained on every Contracep- tive Drug Study subject. This battery included the following tests: Glucose-tolerance testing Latex agglutination and titer Visual acuity and ocular tension Urinalysis and urine cultures Achilles tendon reflex MMPI Pulmonary function test EKG Audiometry Blood pressure and pulse Protein electrophoresis Chest X-ray Thromboelastography Mammography Serum chemistry Anthropometry Blood count Questionnaire history (patient) Blood typing Findings and diagnoses (physician) This report briefly describes each phase and discusses the $'ligxeonival procedures undertaken and the reliability of the ata. GLUCOSE-TOLERANCE TESTING The challenge consisted of 75 g glucose administered in a 6-0z aqueous solution with artificial lemon-lime flavor to which 2 oz of crushed ice were added. Before October 19, 1972, all Con- traceptive Drug Study (CDS) subjects were challenged between 9 a.m. and 1:30 p.m. Although all subjects had been instructed to fast, 40 percent, in fact, had eaten from 1 to 6 hr before receiv- ing the challenge. After October 1972, all subjects fasted overnight and were prepared with 3 days of adequate carbohydrate intake. In addition, the challenge was offered these subjects only between 8 and 9:30 a.m. Sixty minutes after the dose, a sample of venous blood was drawn. The serum was spun down and separated 30-60 min after venipuncture. The serum glucose concentration was determined by the orthotoluidine method. (This method is based on a color reaction given by aldosaccharides with orthotoluidine in glacial acetic acid containing thiouria.) 91 92 STUDY, POPULATION, PLAN, AND PROCEDURES Quality-control procedures were as follows: Glucose dispenser.—Two standard cups of glucose solution were taken from the dispenser each day. Ice was melted and the volume, total glucose, and glucose concentration were recorded. A 1:200 dilution was made for the glucose determination. Blood sugar.—These readings were standardized daily before samples were analyzed using known standards of 50, 100, 150, 200, 250, and 300 g glucose (Harleco Aqueous Standards).: Three Warner controls (Versatol)® were determined twice daily. On January 17, 1972, there was a change from manual to Auto- Analyzerc determination, still using the orthotoluidine method. After July 1972 the calibration curve was determined using Dow protein-based standards? of 50, 100, 200, and 300 g after each 30 specimens. This change was made so that a blood urea nitrogen (BUN) curve could be calibrated at the same time. Harleco and Dow standards were similar. VISUAL ACUITY AND OCULAR TENSION For visual acuity the Snellen eye chart was used and a code from 0-9 was placed on an IBM card. Ocular tension was measured with a Schiotz tonometer which gives progressively smaller readings for increasing tension. Twice daily the instrument was set on the test block to see if it registered zero; if not, it was returned to the optical company for recalibra- tion. Despite these measures, a variation in monthly means was observed which was too great to ascribe to chance alone, although these means stabilized by 1971. ACHILLES TENDON REFLEX Achilles-tendon-reflex time was measured by a Thyrochron 100,¢ a photoelectric instrument that included a two-digit numeri- cal readout of reflex duration in hundredths of a second. The * Harleco Reagents Manufacturing Co., Philadelphia, Pa. : ® Versatol controls, Warner Lambert Pharmaceutical Co., Morris Plains, NJ. ¢ AutoAnalyzer, Technicon Instruments Corp., 511 Benedict Ave., Tarry- town, N.Y. 10591. ¢Dow BUN Glucose Standard protein base, Dow Diagnostics, Post Office Box 1656, Indianapolis, Ind. 46206. ° Thyrochron 100 (for Achilles tendon reflex test), Harworth Manufac- turing Co., 2881 El Camino Real, Redwood City, Calif. 94061. AUTOMATED MULTITEST LABORATORY PHASES 93 Thyrochron computed the time from stimulus to one-half ampli- tude of relaxation. Repeat measurements were taken on each patient until the same reflex time was obtained at least twice. This value was then recorded. This test was discontinued October 5, 1972. For quality control, three determinations were performed on the first and last patients daily. These data indicated that varia- tion of repeat measurements on a single individual was small compared to the total variation in the population. Furthermore, the month-to-month variation was within acceptable limits. PULMONARY FUNCTION TEST A Hi-Fi spirometer, model 470,! with a model 280 pulmo- digitizer was used in the AML from January 1969. Round plastic disposable mouthpieces and large-bore tubing were used. After application of a noseclip, three maximal expiratory efforts were obtained. The third of these was used for analysis. The following values were recorded: Forced Expiratory Volume (FEV) (total, 14 sec and 1 sec) and Peak Flow. Reading of the expirogram was automated. Results were visually displayed and transferred auto- maticaly to IBM cards. Quality-control procedures were as follows: Three determina- tions were recorded for each variable for the first and last patients seen each day. While the monthly variation was within reasonable limits, the intrasubject variation is not significantly higher than intersubject variation for FEV.5 and Peak Flow. The importance of an enthusiastic aide who understands how to administer the test properly cannot be overemphasized. AUDIOMETRY The audiometry test measured hearing threshold to the nearest 10 dB above audiometric zero in each ear at each of six frequencies: 500, 1000, 2000, 3000, 4000, and 6000 Hz. The audi- ometers were intended to measure thresholds to the 1964 standard reference zero of the International Standardization Organiza- tion (ISO). After use during the early period of AML examina- tions, the audiometers were recalibrated in early 1971. The instru- ments were found to be close to the ISO standard and required . * Hi-Fi spirometer, model 470, and model 280 pulmodigitizer, Med-Science Electronics, Inc., 1455 Page Industrial Blvd., St. Louis, Mo. 63132. 94 STUDY, POPULATION, PLAN, AND PROCEDURES little adjustment. No formal procedures for obtaining duplicate measures were employed. PROTEIN ELECTROPHORESIS Sera obtained for protein electrophoretic determinations were refrigerated at 4°-6° and analyzed 1-3 days after venipuncture. Protein electrophoresis was performed on cellulose acetate membranes utilizing the Beckman Microzone Cell. The barbital buffer solution was kept at pH 8.6 and 0.075 ionic strength. Electrophoresis was continued for 20 min at 4.5 to 9.5 mA. Protein was stained by a Ponceau-S fixative dye solution and read visually. No electrophoreses were done after September 11, 1972. Quality-control procedures were as follows: The effect of storing sera for 1-8 days was investigated and found not to be statistically significant. In addition, repeat densitometer tracings from the same cellulose acetate strip stored up to 4 days were evaluated with no significant variation discovered. Reader varia- tion was measured once weekly by obtaining two independent readings of a single densitometric tracing. This variation was high due to difficulty of nonautomated reading. Variation in densitometer tracings was measured by obtaining duplicate trac- ings on a single cellulose acetate strip each week. Once weekly eight determinations were carried out on each cell for a single patient sample. This provided a measure of intercell and intracell variability. The month-to-month variability of these determinations was not acceptable. Serum protein electrophoresis determinations were discontinued ; it was advised that reading of densitometer output be automated if these determinations were to be reinstituted. THROMBOELASTOGRAPHY The Hartert Thromboelastograph® used for this test proved to be a technically simple, highly reproducible method for studying clot kinetics and clot firmness. V2A stainless steel cups were placed on three sockets mounted on a thermostatic block that was maintained at 37° C. Test speci- mens were placed in each cup. Blood was drawn between the hours of 9 am. and 2 p.m. The sample for thromboelastography was & Beckman Microzone Cell, Beckman Instruments, Inc., Spinco Division, 1117 California Ave., Palo Alto, Calif. 94304. " Hartert Thromboelastograph, Litton Medical Products (U.S. distribu- tors), 825 Nicholas Blvd., Elk Grove Village, Ill. 60007. . AUTOMATED MULTITEST LABORATORY PHASES 95 collected in the second or third of three siliconized vacutainer tubes, minimizing the contamination by tissue juice. The patient-flow pattern allowed the transfer of whole blood to the thromboelastogram within 1 min. The use of whole blood rather than citrated whole blood or plasma lessened the chance of technical error. A siliconized pipette was used to transfer 0.35 cm? of whole blood to the cup, which was previously warmed to 37° C. The piston was lowered into the sample, and mineral oil was placed over the blood to prevent changes at the blood-air interface. Until October 1972 three tests were obtained each hour the AML was in operation. After that time the test was performed for special studies only. For purposes of quality control, one replicate series of three was done each week and values of R, K, and Ma observed to estimate the within-subject variability. Furthermore, nine trac- ings were read independently by two readers to estimate the within-reader variability. From March 1972 every 10th TEG was evaluated by two readers. While these error sources were negligi- ble, there was considerable drift over time in mean values due to several changes in personnel, needles, and metal in cups and pins. BLOOD CHEMISTRIES The following serum chemistries were carried out during some or all of the initial 36 months of AML operation: Sodium SGPT Potassium Total protein Calcium Albumin Phosphorus Total lipids BUN Beta lipoproteins Creatinine Cholesterol Total bilirubin Iron Direct bilirubin Iron-binding capacity Alkaline phosphatase Triglycerides SGOT These blood chemistries were obtained on almost all CDS subjects. Until day 78 of 1969, these tests were carried out on a 12-channel AutoAnalyzer (AA). After that time and until January 1, 1972, a 22-channel AutoChemist (AC),' housed in Kaiser Oakland and not under AML jurisdiction, was used to ! AutoChemist, AGA Corp., 550 County Ave., Secaucus, N.J. 07094. 96 STUDY, POPULATION, PLAN, AND PROCEDURES obtain these results. Until day 219 of 1969, 18 channels were operative. After that time, five of these channels were discontinued because of poor quality. After January 1, 1972, an AutoAnalyzer was obtained by the Walnut Creek laboratory, and six tests of good quality were obtained on each subject. It became apparent, soon after the AC became operative, that the quality of the data was poor. Large numbers of tests were reported “not done” (TND), and the number of these varied widely from day to day. An attempt was made to determine the cause of the reported omissions; suspicions were aroused because previous experience with the AA had led us to expect an almost 100 percent return on the tests. In addition, the Walnut Creek laboratory maintained a record of sera that were hemolyzed or chylous ; these numbers were extremely small in comparison with the numbers reported by the AC laboratory. Because of arbitrary discarding of data by AutoChemist per- sonnel, the quality of the remaining data was also questionable. Daily statistics such as means, medians, and standard errors were not very meaningful since extreme values arbitrarily had been discarded and reported as TND. When a large fraction of tests were out of range or reported as TND for a particular channel, on a given day, all tests on that channel for that day became questionable. Rather than delete any data it was decided to attach a modifier to each test value. This modifier (M) was three-valued; it took on the value 0 for tests rated “acceptable,” 1 for “questionable” tests, and 2 for “un- acceptable” data. Because there was no information on the time of day the test was run, it was necessary for all test results on a given channel on a single day to receive the same modifier. This modifier was a function of both the fraction of TND and the fraction of data “out of range.” After September 11, 1972, only creatinine, SGOT, cholesterol, and triglycerides were obtained on study subjects. BLOOD COUNTS A Model S Coulter Counter! was used to determine hemo- globin, hematocrit, red blood cell count, mean cell volume, mean cell hemoglobin, mean cell hemoglobin concentration, and the white blood cell count. Blood samples were drawn between 9 a.m. ! Coulter Counter (Model S), Coulter Electronics, Inc., 590 West 20th St., Hialeah, Fla. 33010. AUTOMATED MULTITEST LABORATORY PHASES o7 and 2 p.m. in vacutainer tubes containing tri-potassium EDTA as an anticoagulant. Quality control was accomplished as follows: The instrument was calibrated daily each morning using Coulter standards. In addition, a pooled blood sample was used after each 10 patient samples. If a control value deviated markedly from prior readings after rechecking, the machine was recalibrated. Three times daily a random patient sample had a hematocrit determined independ- ently by the microhematocrit method as well as by the Model S Coulter Counter. BLOOD TYPING Other typing sera were used to obtain the ABO blood type on each subject. Since this was a longitudinal study, it was felt that the best gauge of quality control would be to compare determina- tions on all women tested during 2 separate years. Of all study subjects typed two or more times, less than 0.4 percent showed inconsistencies in ABO type. LATEX AGGLUTINATION AND TITER Sera were collected and refrigerated at 4°-6° C and analyzed 1-3 days later. Samples were tested for rheumatoid factor using a latex agglutination slide test. Samples with positive agglutina- tion (1:20) were titered. Latex titers that were reactive at a serum dilution of 1:40 or greater were considered positive. From September 11, 1972, routine testing of study subjects was discontinued. Known positive and negative controls were run daily. In addition, one positive specimen each week was repeated independ- ently by two technicians. URINALYSIS One hour after the glucose challenge, a midstream clean-catch urine specimen was obtained. The following quantities were then determined using Labstix:* pH, protein, glucose, ketones, and occult blood. Labstix were used in great quantity. Therefore, chance of error due to storage was minimal. Two strips of each five lots x Labstix reagent strips, Ames Co., division of Miles Laboratories, Inc., Elkhart, Ind. 46514. 98 STUDY, POPULATION, PLAN, AND PROCEDURES were tested periodically for quality control, as follows: Three pH buffers ranging from 5 to 9 were prepared and Labstix tested. Three solutions for glucose, protein, ketones, and occult blood were prepared corresponding to the color ranges of the Labstix. The 10 Labstix were then tested. Later a commercial control (Teck Chek)! was used to test the reliability of each Labstix lot. URINE CULTURES Until September 11, 1972, urine was cultured and a colony count was made. No attempt at organism identification was made on the initial culture. If, on the initial plating, over 100,000 colonies/ml were found, the subject was recalled and a second specimen obtained. A diagnosis of significant bacteriuria was made when both specimens showed over 100,000 colonies/ml. MINNESOTA MULTIPHASIC PERSONALITY INVENTORY The Minnesota Multiphasic Personality Inventory (MMPI)?! was designed to assess some of the major personality character- istics that affect personal and social adjustment. There are 399 statements in the MMPI which are grouped together in different combinations to form various psychological scales. The following 22 scales were scored : Hypochondriasis Correction Depression Social introversion Hysteria Guilt feeling Psychopathic deviate Conscience strength Masculinity-femininity Psychophysiologic reaction Paranoia Intellectual efficiency Psychasthenia Social desirability Schizophrenia Repression sensitization Hypomania Dominance Lie Dependency Validity Ego strength There are two different formats of the MMPI: the booklet and separate answer sheet, and the card sort. Beginning in December 1968, the patients were given a booklet containing the 399 state- ! Teck Chek, Ames Co., division of Miles Laboratories, Inc., Elkhart, Ind. 46514. AUTOMATED MULTITEST LABORATORY PHASES 99 ments and an IBM 1280 mark sense answer sheet on which they indicated their answers. The punched cards produced from the optically scanned answer sheets were then scored by an MMPI scoring program written by CDS programmers. In May 1971 a change was made to the card-sort form of the MMPI. All MMPI administration was discontinued in January 1972. ELECTROCARDIOGRAMS Twelve-lead electrocardiograms were taken on a custom- designed Hewlett-Packard System.m A specially designed table with German silver rests for arms and legs served as limb elec- trodes. Six leads (limb or precordial) were taken simultaneously. Tracings were produced on photosensitive paper and were xeroxed to facilitate handling and insure long-term stability of the electrocardiogram. Five internists rotated in reading these electrocardiograms. BLOOD PRESSURE AND PULSE At the blood-pressure (BP) station, the subject was placed in a supine position. After approximately 8 min, while an EKG was obtained, the BP and pulse were measured. Almost all read- ings were taken on the left arm, and a special design of the tables allowed for a similar bend at the elbow for each woman. Pressures were recorded automatically by a BP apparatus (Godart)." The principle of this technique has been described in detail by Gruen.? Diastolic pressure was measured at the dis- appearance of sound—XKorotkoff, phase 5. Blood-pressure reading and pulse by this machine method were found to be reliable and reproducible. The only difficulty was cuff application in some sub- jects having an obese, cone-shaped, tapered arm. On these few patients, manual readings were obtained. For quality-control purposes, the machine was calibrated weekly against a mercury manometer as a standard. Technician reproducibility was checked twice daily, using random female subjects, and the results were statistically examined. The great majority of values represented the first BP observations, except in the occasional instance where a cuff problem resulted in a technically poor reading. =m Hewlett-Packard Co., Neely Sales Division, 1101 Embarcadero Rd., Palo Alto, Calif. 94300 ® Godart BP apparatus, Godart N.V., Bilthoven, Holland. 100 STUDY, POPULATION, PLAN, AND PROCEDURES CHEST X-RAY Seventy-millimeter chest X-rays were obtained by a Profexray machine.® These films were read within 5 days by a radiologist. MAMMOGRAPHY Two Continental mammography machines? were employed— one fixed in a vertical and the other in a horizontal position—to obtain AP and lateral views on all women 48 years of age or over. At first each film was read by two radiologists; but as their pro- ficiency improved, only one physician read each mammogram. ANTHROPOMETRY Fourteen anthropometric: measures were recorded by an anthropometer specially built for this study. In addition, weight was obtained on a clinical balance, and two skinfold measurements were obtained with Hemco calipers.¢ These tests were discontinued in October 1972 except for height, weight, and skinfold. The instrument technician checked vertical and horizontal measurements with standard rods. These rods consisted of a 36-in. and a 72-in. rod for vertical and a 10-cm and a 40-cm rod for hori- zontal measurements. Standard weights were used to check and calibrate the clinical scale. A 25-mm standard block was used to check the thickness measurement of the skinfold calipers. Furthermore, duplicate readings were obtained on a randomly selected subject twice a day to control the within-subject vari- ability of these tests. While the quality of the weight measurement was excellent and the height and skinfold adequate, the remaining tests were not satisfactory. In fact, the between-subject variability was not significantly greater than the within-subject variability for these tests. QUESTIONNAIRE HISTORY In addition to the AML phases described, each subject com- pleted an extensive questionnaire history. It included the follow- ing items: ° Profexray, division of Litton Industries, Des Plaines, Ill ? Continental mammography machine, Continental X-Ray Corp., Chicago, 11. 4 Hemco calipers, Hemco Corp., Post Office Box 210, Independence, Mo. 64051. AUTOMATED MULTITEST LABORATORY PHASES 101 Identification information and personal description Disease list (self and family) Symptom list (overall and by systems) Current and past medications Surgical history Cancer or tumor history Contraceptive history and methods Past and recent menstrual history Obstetrical history Until December 1971 the examining physician completed an extensive questionnaire containing the following items: Physical examination findings Medical impressions Dermatological impressions Joint examination report Cancer or tumor impression Central-nervous-system report In addition, the gynecologist completed a two-page checklist of impressions and findings. These forms were difficult to complete and to process and were discontinued after the first 3 years of the study. REFERENCES 1. HATHAWAY, S. R.,, and McKINLEY, J. C.: Minnesota Multiphasic Per- sonality Inventory. New York: The Psychological Corp., 1967. 2. GRUEN, W.: An assessment of present automated methods of indirect blood pressure measurement. Ann NY Acad Sci 147: 109-126, Feb. 9, 1968. 1968. Part B Some Initial Results 535-967 A 2 il Chapter 1 Oral Contraceptives, Pregnancy, and Blood Pressure* IRWIN R. FiSCH, SHANNA H. FREEDMAN, AND A.V, MYATT Blood pressure was determined for 7,605 women aged 18 to 60 who had Automated Multitest Laboratory (AML) examinations during 1969. These subjects included 1,941 current users of oral contraceptives (OC’s), 2,189 never users, and 1,593 past users, as well as 682 women taking other estrogenic hormones, 953 pregnant and 247 postpartum women. Methods.—Blood pressures were recorded automatically by a Godart blood-pressure apparatus with the subjects in a supine position. Oral contraceptives.—Oral contraceptives were associated with a modest yet definite increase in blood pressure. Blood pressure among current OC users is slightly higher than that among nonusers at every age level; the difference increases with age and averages 5 mm Hg for systolic and 1.3 mm Hg for diastolic. By arbitrarily considering “elevated” blood pressure as a systolic pressure greater than 140 mm Hg or diastolic greater than 90 mm Hg, or both, the proportion of OC users among subjects with elevated blood pressure was significantly higher than the proportion of nonusers at every age level. No association was found between blood pressure and OC formula- tion or dose or duration of OC use. Estrogenic hormones.—The older women taking estrogenic hormones had blood pressures higher than older nonusers, but the difference was not significant. Pregnancy.—Blood pressure was low among subjects in the first trimester of pregnancy as compared to nonpregnant subjects of similar age. Other factors.—Mean systolic blood pressure of OC nonusers starts to increase abruptly at about 40 years of age. The rise occurs at an earlier age among current users. Diastolic pressure rises more uniformly with age. The fraction of subjects with high blood pressure (systolic 140 or over, diastolic 90 or over) increases sharply after 35 years of age. An increase in blood pressure with increasing weight was found, but * This report is a revision of the following article, published here by permission of the editors of the Journal of the American Medical Association: Fisch, I. R., Freedman, S. H., and Myatt, A. V.: Oral contraceptives, preg- nancy, and blood pressure. JAMA 222: 1507-1510, Dec. 18, 1972. 105 106 SOME INITIAL RESULTS height, parity, time in the menstrual cycle, and hours since last food do not appear to influence blood pressure. An association between systolic pressure and month of examination was found. The summer months all show mean systolic values higher than winter months; however, this requires further evaluation. Comment.—Oral contraceptives are associated with a slight elevation of blood pressure, which appears to be reversible and not related to OC dose or components. Whether this slight elevation of blood pressure is ultimately detrimental to these women is unknown at this time. The goal of this preliminary paper is to report the relation- ship of oral contraceptives and pregnancy to blood pressure (BP). Data are also analyzed to ascertain if an association exists between contraceptive drug composition and BP. Although minor elevations of systolic and diastolic BP have been found in contraceptive drug users, the clinical implications remain debatable.’ 2 Unlike most previous studies, these cross-sectional analyses were obtained from BP readings on a large, closely observed population. MATERIALS AND METHODS STUDY SUBJECTS Blood-pressure data were obtained from 7,605 women who had their Automated Multitest Laboratory (AML) examinations in 1969. Women taking antihypertensive medication (125 women) were excluded from analyses. The contraceptive drug status of women was not known by the technician obtaining the BP. Table 49 summarizes the age-group distribution of the 7,605 study subjects according to hormone or pregnancy status. The 1,941 women who had taken an oral contraceptive within 30 days of their AML visit were considered current users. For some pur- poses the 2,189 never users and 1,593 past users were combined into nonusers, and this group was contrasted with current users. Considered as a separate group were 682 predominantly older women who were taking other estrogenic hormones. In addition, 953 subjects were examined during the first or early second tri- mester of pregnancy; and 247 postpartum women were studied 1-3 months after delivery. METHODS Subjects arrived at the AML on a staggered schedule from 9 a.m. to 2 p.m. They arrived at the BP phase approximately 2 hr after the start of their examination. ORAL CONTRACEPTIVES, PREGNANCY, AND BLOOD PRESSURE 107 TABLE 49.—Blood pressure by age, hormone use, and pregnancy status [Mean 41.96 standard deviation of mean in mm Hg] Age in years, Never Past Current | Other Post- blood pressure, user user user estrogenic|Pregnant| partum and number hormones 15-24: Systolic. ___| 118+3 11942 1234-2 11543 | 118+1 12042 Diastolic_ _. T3+2 7541 7541 7645 | 70+1 762 Number.____ 88 162 293 5 431 115 25-34: Systolic. ___| 118+1 11941 123 +1 11545 | 115+1 11942 Diastolic. _ _ 761 T5+1 7641 80+3 6941 T6+2 Number____ 361 679 833 33 478 122 35-44: Systolic____| 120+1 120 +1 125+1 12243 | 116+4 119410 Diastolic. _ - 7841 TT+1 8041 8142 | 7343 T6+4 Number.____ 977 544 563 112 43 10 45-60: Systolic. ...[ 125+1 123+2 13142 125+2 HOO coe mciio Diastolic. _ - 80+1 7942 8241 81+1 Lo ) Number____ 763 208 252 532 LE Loeeeieee Age adjusted: ! Systolic__..| 120+1 120 +1 125+1 11941 | 11241 11942 Diastolic. __ TT+0 2 7740 2 78+02| 80+1 6941 T6+1 Number____| 2,189 1,593 1,941 682 953 247 1 Adjusted to distribution of total population. 2 Less than 0.5. The technique for obtaining BP values is explained in part A, chapter 4, “Automated Multitest Laboratory Phases and Quality- Control Procedures” and by Gruen.® RESULTS ORAL CONTRACEPTIVES Because of the close relationship between BP and age, all data were analyzed age specifically. Comparisons were made be- tween past and never users, and since no significant difference was found between the means for systolic and diastolic pressure, these groups were combined into the nonuser category (fig. 8). Mean blood pressures of study patients are summarized in table 49, and the association between contraceptive drug status, BP, and age is also displayed in figure 9. Blood pressure among 108 SOME INITIAL RESULTS 135 — 125 | 120 | Never Users SYSTOLIC BLOOD PRESSURE (mm Hg) — — — Past Users 115 | AGE IN YEARS 85 DIASTOLIC BLOOD PRESSURE (mm Hg) 1 1 1 1 i 1 1 1 1 ceed 15 20 25 30 35 40 45 50 55 60 AGE IN YEARS FIGURE 8.—Mean systolic and diastolic blood pressure by age for past and never users of oral contraceptive drugs. current users of oral contraceptives is slightly higher than that of nonusers at every age level ; the difference increases with age and averages 5 mm Hg for systolic blood pressure and 1.3 mm Hg for diastolic. A sign test produced significance probabilities of p < .01. Figure 10 and table 50 further demonstrate an association between oral contraceptive use and BP. Whether we consider “elevated” BP as a systolic pressure greater than 140 mm Hg or diastolic greater than 90 mm Hg or both, oral contraceptive users are overrepresented among women with higher BP readings. Among subjects with both systolic pressure over 140 and diastolic over 90 mm Hg, the proportion of current users was significantly ORAL CONTRACEPTIVES, PREGNANCY, AND BLOOD PRESSURE 109 1351 130 125 120 115} Systolic Blood Pressure (mm Hg) 110+ — 85} oo I £ ~~ 80 ® 2 8 a 75} © o o © e 70 S 8 2 esl - , Oral Contraceptive User Oral Contraceptive ra Nonuser / / A “o— " cr \ D Pregnant Oral Contraceptive User 7 Oral Contraceptive Nonuser Pregnant x 1 1 1 1 i 156 25 35 45 55 65 Age (yr) FIGURE 9.—Blood pressure by age, contraceptive use, and pregnancy status. higher than the proportion of nonusers at every age level »p < .025). No association could be detected between duration of oral con- traceptive use and BP for either current or past users. More 110 SOME INITIAL RESULTS UT uf sob wl C0 Noss | Systolic Blood Pressure > 140 mm Hg | Diastolic Blood Pressure 90 mm Hg Diastolic Blood Pressure > 90 mm Hg 20 Systolic Blood Pressure > 140 mm Hg 1524 2634 3544 4560 1524 2534 3544 45.60 Age (yn) Age yr) Age (yr) FIGURE 10.—Percent of patients with elevated blood pressure by age and oral contraceptive use. TABLE 50.—Percent cases with elevated blood pressure by age and oral contraceptive use Systolic blood pressure Systolic blood pressure | Diastolic blood pressure! 140 and over and Age in 140 or over 90 or over diastolic blood pressure years 90 and over Users Nonusers Users Nonusers Users Nonusers Percent Percent Percent Percent Percent Percent 15-24. .... 11.6 5.8 3.9 1:5 1.6 0.4 25-34..... 11.4 6.7 4.6 4.0 2.0 1.3 35-44_____ 16.6 10.4 9.1 6.2 5.0 2.8 45-60_____ = 28.6 20.0 12.7 12.7 9.8 7.0 intensive study will be necessary to determine the precise onset of the oral contraceptive effect and point at which the blood pressure reverts to preoral contraceptive levels. Figures 11 and 12 demonstrate the lack of association between diastolic blood pressure and duration of oral contraceptive use or time since last use. A similar lack of association has been found for systolic blood pressure. The relationship of estrogen and progestin content of oral contraceptive drugs to BP was studied. Oral contraceptive users were subdivided according to estrogen type and dose separately for combination and sequential preparations. A similar subdivision was made for the progestin component. At the lower dose levels of oral contraceptive used in this population, there was no evident relationship between BP and these variables. This can be seen in figure 13 as well as in tables 51 and 52. The apparent decrease at high progestin dose levels is not significant because of small numbers. ORAL CONTRACEPTIVES, PREGNANCY, AND BLOOD PRESSURE 111 Current Users - === Past Users 85 Age 15-24 r Age 25-34 DIASTOLIC BLOOD PRESSURE (mm Hg) © g E85 | Age 35-44 } Age 45-60 w g : 0 y ? on ; L fon ny 8 . fe oy 3 g mn ; ! . 2 PN o — ! | \ o . nye ! L A) 3 " Se” 5 2% - 5 + = 1 1. 1 1 1 1 i Ln J 24 48 72 96 24 48 72 96 : MONTHS OF OC USE MONTHS OF OC USE FIGURE 11.—Diastolic blood pressure by total months of use for current users and past users of oral contraceptive drugs. 2 Age 15-24 Age 25-34 E . : bin ey £ g . a 2 “. 2 ei ih > J “e. 2 . ~ 85 Age 35-44 Age 45-60 ¥ = / & go . > \ 2 In eta 2 fen * ~om w= ¥ % 2 ie. oe" Mo, 75 » ns or = 1 1 1 L 1 L I 1 J 24 48 72 96 24 48 72 96 MONTHS SINCE LAST OC USE MONTHS SINCE LAST OC USE FIGURE 12.—Diastolic blood pressure by months since last use for past users of oral contraceptive drugs. 112 SOME INITIAL RESULTS 124 pp : ZN, MEAN] ap Ne i i i lL — 122 SYSTOLIC 126 T T T - 5 N I | 120 — o S fe | l Ll | J l T T 1 MEAN 75 = 751 =] DIASTOLIC DIASTOLIC 79 ~N © BLOOD PRESSURE, mm Hg BLOOD PRESSURE, mm Hg 73 | | 1 | 73 1 i | 1 1 | 0.05 0.06-0.075 0.08 0.10 0.5 1 225 5 10 25 ESTROGEN DOSE, mg PROGESTIN DOSE, mg FIGURE 13.—Mean blood pressure by estrogen and progesterone doses for current oral contraceptive users. ESTROGENIC HORMONES Data from 682 women taking female hormones other than birth control pills (conjugated estrogens, ethinyl estradiol, etc.) are presented in table 49. Eighty percent of these were over 45 years of age. In the older age groups mean blood pressures were higher than for nonusers, but the difference was not significant. Women taking estrogenic hormones were a heterogeneous group about which no conclusion can be drawn without a detailed analy- sis, which was beyond the scope of the report. PREGNANCY Blood-pressure readings from the 953 women in early preg- nancy are contained in table 49 and figure 9. Their diastolic BP was approximately 5 mm Hg below nonusers in every age group and over 6 mm Hg below current users. While pregnant women had lower systolic pressure than either users or nonusers of oral contraceptives, the effect of early pregnancy on systolic BP was not as consistent or as dramatic as on diastolic. The 247 postpartum women are summarized in tables 49 and 53. Among these women, 10 percent were current oral con- traceptive users. These subjects had BP values similar to those ORAL CONTRACEPTIVES, PREGNANCY, AND BLOOD PRESSURE 113 TABLE 51.—Mean blood pressure (mm Hg) of current users by estrogen type and dose [Mean =+1.96 standard deviation of mean] Estrogen dose (ug), Mestranol Ethinyl estradiol Total blood pressure, estrogen and Himba of Combina- | Sequen- | Combina-| Sequen- type tion tial tion tial 50: Systolic....cucee- ULSD |eecocacaea 125:£2 |ovceorcae 125+1 Diastolic... _-__ TBE]l [vemmmonnn TT) [.oimammin 8+1 Number______.__ 818. ememoewnsre 268 |-eenecenn- 586 75: : Systolic.......... BEB [-cunneviss]mammneosnnresnanuse 12345 Diastolic......... ITED Nevimmman a ammeter ain mM+3 Number. _..._.... BB Us lime wie ies solmrede tiv mere se hp oon 43 80: Systolic... .cuswsx 12545 1282 in cums ionlsme anime 12343 Diastolic... 7543 BEY ld itm 78+1 Number. _._____. 42 B08 deri sre errmnn ee 265 100: Systolic. .co. dxwen 12H] eb wmmeimn menses 12145 125+1 Diastolic... TBE] emmmommmns) cnn mmmmmns 7642 78+1 Number. ________ O28 [wren me 46 971 150: Systolic... 127208 |. ccm sper es em fe 127 +23 Diastole.....cmm wns TBLIB crave] iim nm arn) mmeman ees 73413 Number_____._.. Dom msi fmm rier pce wire 2 Unknown: Systolic. _..____.. 124 +2 |e 124 +2 Diastolic... -- TOD |ecicoennen)ensnnssica)inesosnnon 7942 Number. ________ BE re enim tsess mee oI Te Yim emis 143 Total for all dose levels: Systolic... 124 +1 123 +2 12542 12145 125+1 Diastolic... 78+2 79+1 (ES 7642 78+1 Number... _.. 1,478 223 268 46 2,010 of nonpostpartum oral contraceptive users. In addition, 51.6 per- cent of postpartum women received a mixture of estradiol valerate and testosterone enanthate (Deladumone) for inhibition of lac- tation, the fraction decreasing with age. These women had mean BP values significantly higher than postpartum subjects not receiving this drug. However, the effect of Deladumone on BP was not quite as marked as that of oral contraceptives. The highest BP values were found among the small fraction of postpartum women receiving both Deladumone and contraceptive drugs. Slightly less than one-half of postpartum women received neither 114 SOME INITIAL RESULTS Deladumone nor oral contraceptives. These subjects had low systolic pressures comparable to those of women in early preg- nancy, while their diastolic values were not depressed. OTHER FACTORS The effect of aging on both systolic and diastolic BP was quite marked but different (fig. 9). There was almost no increase in mean systolic BP of nonusers until about 40 years of age, at which time there was a sharp rise. The point at which a rise occurred was earlier among current oral contraceptive users. Diastolic pressure seemed to rise more uniformly with age, and no abrupt increase was seen. Not only did the mean BP rise with age, but the fraction of women with systolic pressure of 140 or over, or diastolic of 90 or more, increased sharply after 35 years of age. In addition, the age-specific frequency distributions of BP became increasingly bimodal with age indicating an increasing fraction of hypertensives in the population. This phenomenon has been discussed by Clark et al.* The phenomenon also is shown in figures 14 and 15 and tables 54 and 55 which contain frequency distributions of blood-pressure values for four age groups. Mean blood pressures as a function of age for this group of women were compared to those published by the National Health Survey® (fig. 16 and table 56). The rate of increase with age for the latter was significantly greater. That those data did include pregnant and postpartum subjects, however, may explain the low means in the younger age groups. In addition, different techniques for obtaining blood-pressure readings were employed. The influence of weight on BP is well known. Mean diastolic BP in each of the four age groups as a function of weight had a significant linear relationship with slope nearly independent of age (r = 0.30). There was also an increase in systolic BP with increasing weight, but the fit to a straight line was not as good (r = 0.16). This relationship can be seen in figure 17 and tables 57 and 58. Note that the apparent high degree of variability in the highest weight groups is due to the very small numbers of subjects in those groups. In addition, height was examined but was found to be statistically independent of BP. Because of the interrelation between weight and blood pressure, it was suspected that the increase in BP associated with contraceptive drug use might have been ascribable to higher weights in this group. In fact, the opposite was true; among our subjects oral contracep- tive users were lighter at every age level. TABLE 52.—Mean blood pressure (mm Hg) by progestin type and dose [Mean +1.96 standard deviation of mean] Dimethis- Medroxy- Total Progestin dose (mg), blood pressure, | Norethin- Norethyn- | Ethynodiol | Norgestrel | terone and progester- for al and number of users drone odrel diacetate chlormadi- one progestin noneacetate types 0.5: SYSONC. crn ens sm RRR EE REE fe SN EE Re re B2=td le rr nse s ee 124 +4 DH ABEONC.. om on mmm me sme eis ms mre crs ems fo me mem sm mt ee mm i TE2 onenimn sense mmnnamewmms 762 NUINIDOE.. .... ce eee semis stm me mes es ee Pe ees rem mt HE iesest scene sees 44 1.0: Systolic... __________________ 12441 | _________ 12842 icine assem 124 +1 Diastolie______________________ TTL erimemnes RTE, oem remrelrs seme ere ere cen TT+1 Number_ _____________________ 431 | ___ B28 | encom |e nme 757 2.0: Systolie_______________________ 120E1 [oven cman irons mmm me ps ms om 12868 I crcemcas 12542 Diastolic. _____________________ TBE] lei mmm oie te mcm i ere ae 79:2 locoicommmnse 7841 Number_ _____________________ BPD lemme mmm fr meni ie saison Jom mn meni miei 168 j—ceemmenes 690 2.5: Systolic... ______________ 12743 124 +2 |. femme ee 12543 Diastolie_...__________________ 7942 TRHD: [orc on smn errs erase fem sai osm 7942 Number. _____________________ 114 132 reece lene me ane mm ee a en 246 5.0: Systolie_._____________________ ER 12845 |e eee ee eee 12345 Diastolie_.-.__________________ LA THES. Jeeccicoro oul pes CERES see T7T+3 NUR oe ES er ommrmn e ren 7+ A SE i tr re 43 10.0: SYStoHC. cee cas i Sr 126 +22 ee era et se re 11847 11948 GIT JINSSHY AO0Td ANV ‘AONVNOHIAL ‘SHAILIADOVILNOD TVIO TABLE 52.—Mean blood pressure (mm Hg) by progestin type and dose—Continued [Mean+1.96 standard deviation of mean] Dimesthis- Medroxy- Total Progestin dose (mg), blood pressure, | Norethin- Norethyn- | Ethynodiol | Norgestrel | terone and | progester-. for al and number of users drone odrel diacetate chlormadi- one progestin noneacetate DIAS odd ret) 0 sons isn lie Bl Li ti ii an im | ad T5+4 5745 IN TINIE con 2 ii ri Bl rsorng si Bi 1 Tomes einer rae i Ye hia 25 27 25.0 BV OIC ce cnc ir i i mi i ms em i Bo 121£8 | conor 12145 DAREOUE.. tll rmits mememmimemtirim ey mmm marin mmr mmm eit etn wir learnt TOE2 | cnvimnmaiinn 762 NOIABEE: woe mein ie mim wa ww wie pur SE rr oe esate fe aia 48 = A nae 46 1 Chlormadinone acetate. TABLE 53.—Effect of Deladumone on blood pressure (mm Hg) of 250 postpartum women On Deladumone Not on Deladumone Age in years Oral contraceptive users Oral contraceptive nonusers Oral contraceptive users Oral contraceptive nonusers Systolic Diastolic] Number Systolic |Diastolic| Number | Systolic |Diastolic| Number | Systolic |Diastolic| Number Under 25........... 116.4 76.6 81 121.9 76.2 49% 121.0 71.7 gt 117.3 74.6 36 25-84. cco cnnnniins 132.0 74.7 9 119.3 78.0 58 | 121.8 76.0 41 119.1 75.3 69 S544. eins 128.0 75.0 1} 121.2 75.0 Be ramen ea 0| 116.7 77.6 9 Total, all ages.| 124.8 75.6 18 120.5 77.1 112 121.4 74.1 7 118.4 75.2 114 911 SLINSHY TVILINI ANOS ORAL CONTRACEPTIVES, PREGNANCY, AND BLOOD PRESSURE 117 [ [ L 15 |= Age 15-24 15 g g Sw S10 s 5 g g i « T on T Lisi), 1 ] 1 1 ny 90 100 110 120 130 140 150 160 170 Y90 100 110 120 130 140 150 160 170 SYSTOLIC BLOOD PRESSURE (mm Hg) SYSTOLIC BLOOD PRESSURE (mm Hg) 15 15 r - Age 35-44 - Age 45-60 PERCENT OF CASES T 3 T PERCENT OF CASES T on T 1 1 1 1 1 1 1 1 1 nl 1 1 1 1 1 1 1 1 T7790 100 110 120 130 140 150 160 170 "90 100 110 120 130 140 150, 160 170 SYSTOLIC BLOOD PRESSURE (mm Hg) SYSTOLIC BLOOD PRESSURE (mm Hg) FIGURE 14.—Relative frequency distribution of systolic blood pressure for all oral contraceptive groups, by age. 0 - 30 Age 15-24 Age 25-34 «» Sal D2 3 3 5 5 FRLS "E10 A 1 1 1 A 1 1 I 1 vr Vv 60 80 100 120 60 80 100 120 DIASTOLIC BLOOD PRESSURE (mm Hg) DIASTOLIC BLOOD PRESSURE (mm Hg) 30 [ Age 35-44 30 Age 45-60 w Bal Ba 3 S 5 s = = g g SL &10 A 1 { 1 4 1 1 1 ‘80 100 120 60 80 100 120 DIASTOLIC BLOOD PRESSURE (mm Hg) DIASTOLIC BLOOD PRESSURE (mm Hg) FIGURE 15.—Relative frequency distribution of diastolic blood pressure for all oral contraceptive groups, by age. 118 SOME INITIAL RESULTS TABLE 54.— Relative frequency of systolic blood pressure (mm Hg) for all oral contraceplive groups, by age Age in years Systolic blood a 15-24 25-34 35-44 45-60 (mm Hg) Num- | Per- | Num- | Per- | Num- | Per- | Num- | Per- ber cent ber cent ber cent ber cent 70-74... cu 1 DD) eran) mmm a 1 0.1 TE: oom mami pa arnemloren ine nase) seme ymrabl 1 +X 80-84. _______ 2 4 6 0.3 7 0.3 4 2 85-89........ 3 6 11 6 14 .6 14 TX 90-94________ 6 1.3 9 .5 37 1.9 17 9 95-99________ 9 1.9 76 4.1 82 3.8 60 3.1 100-104 ______ 33 6.9 145 7.8 164 7.6 128 6.7 105-109______ 39 8.2 159 8.6 77 2.2 109 5.7 110-124... 78 16.3 239 12.9 313 14.5 182 9.5 115-119...... 3 15.3 323 17 .4 290 13.4 229 12.0 120-124 ______ 46 9.6 197 10.6 219 10.1 155 8.1 125-129... 60 12.6 213 11.5 217 10.0 195 10.2 130-134______ 59 12.3 192 10.4 230 10.6 194 10.2 185-1389...... 20 4.2 90 4.9 97 4.5 109 5.9 140-144 ______ 20 4:2 80 4.3 137 6.3 170 8.9 145-149_____. 14 2.9 67 3.6 86 4.0 150 7.9 150-154______ 4 .8 21 3:1 36 1.7 70 3.7 155-159... 7 1.5 11 6 23 1.1 38 2.0 160-164 ______ 3 6 6 3 17 8 37 1.9 165-169. _____ 1 2 10 5 16 7 45 2.4 Total. _ 478" | ci mmns 1,855 |....... 2,162 |.connn- 1,008 {cous The association of parity and blood pressure was investigated. When all age groups were combined, there appeared to be an increase in BP with increasing parity. However, an age-specific analysis revealed that this correlation is spurious and due only to the correlation of parity and age; women with higher parity have higher BP because they are older, and also heavier. In figure 18 and table 59 the lack of linear relationship can be seen in each of the four age groups considered. Nor was ABO blood type a significant covariable. As can be seen in figure 19 and table 60, the differences in mean blood pressure for the four ABO blood types were not significant. Several variables relating to possible cyclic fluctuations in BP were examined. Neither the time in the menstrual cycle nor hours since last food appeared to influence the BP. Mean blood ORAL CONTRACEPTIVES, PREGNANCY, AND BLOOD PRESSURE 119 TABLE 55.— Relative frequency of diastolic blood pressure (mm Hg) for all oral contraceptive groups, by age Age in years Diastolic blood pressure 15-24 25-34 35-44 45-60 (mm Hg) Num- | Per- | Num- | Per- | Num- | Per- | Num- | Per- ber cent ber cent ber cent ber cent 49 orunder__ |. ___|-__.___ 3 Ot ee i 1 0.1 50-54 _ ||. 4 "2 1 0.1 3 2 EL 10 2a 41 2.2 29 1.8 18 .9 60-64________ 45 9.4 133 7.2 118 5.5 67 3.5 65-69________ 58 | 12.1 203 10.9 156 7.2 115 6.0 70-74________ 122 | 25.5 452 | 24.4 472 | 21.8 296 15.5 T5-T9_____._. 130 | 27.2 482 | 26.0 525 | 24.8 427 22.4 80-84________ 48 | 10.0 188 | 10.1 266 | 12.3 247 12.9 85-89_______. 50 [ 10.5 216 | 11.6 362 | 16.7 352 18.4 90-94________ 11 2.3 94 5.1 150 6.9 223 11.7 95-99_____.__. 3 .6 25 1.3 46 2.1 83 4.4 100-104_____. 1 2 11 .6 2 1.0 49 2.6 105-100... «cn lrimmmsmmfmsmms we 3 .2 6 .3 12 .6 110-114 eee leer] 4 2 11 .6 5-119. coelomic em. 1 1 4 Demmi] minim mine P2002 coe in ude Vom feeders jms 2 1 4 2 125-129... ees emma me mim) mmm foe mmm fmm) 180-184 | |eeeo oo 1 1 |v snmm decanter nase mmm Total _ 478 |. ____ 1,855 ¢....... 2,162 |......- 1,908 |.______ pressure as a function of day of the menstrual cycle is shown in figure 20 and tables 61 and 62. While an elevation in systolic blood pressure over the time of day was observed, it did not hold for all groups when subdivided by “hours since last food” (fig. 21 and tables 63 and 64). Further- more, it was not found for diastolic blood pressure. The interval since last food ingestion may have accounted for the apparent diurnal variation of systolic blood pressure. A strong association between systolic pressure and month of examination was observed. Testing for independence of BP and Julian examination date yielded a x> = 696 at 190 degrees of freedom, which is highly significant. The summer months all showed mean systolic values over 124 mm Hg, while November, December, and January had mean values under 119. The same association was present for diastolic blood pressure, but the fluc- 120 * SOME INITIAL RESULTS 135 ome | | I SYSTOLIC 7 130 |— 8) 125 " 3 Z I a 120 er” NON.USERS xz a ——— TY oe NSH Lo « x ~~" "NATIONAL HEALTH “a ol SURVEY (WHITE) g nol | | | as 3 | I T o o DIASTOLIC po | S so} A 75 ~ 7 NON-USERS sol. -~" NATIONAL HEALTH ] ~ SURVEY (WHITE) 85h 18-24 25.34 35.44 45.54 AGE, years FIGURE 16.—Mean blood pressure by age for oral contraceptive users and nonusers and for National Health Survey data on white females. tuation was less dramatic. Nevertheless, these factors did not affect our findings since the oral-contraceptive status of subjects was independent of time of day, hours since last food, or season. ORAL CONTRACEPTIVES, PREGNANCY, AND BLOOD PRESSURE 121 TABLE 56.—Comparison of Contraceptive Drug Study blood-pressure readings with National Health Survey data (1960-62) Contraceptive Drug Study oral contraceptive user National Age in years and status Health blood pressure Survey 2 Current Nonusers ! users 15-24: mm Hg mm Hg mm Hg Systolic... 122.9 118.0 111.6 Diastolic... cosscsisisndusrmen ons 75.0 74.2 69.1 25-34: BY SLOHC...c cw mo mein swimmer emis oon 3 se 122.6 118.2 115.2 Diastolic. ooo. 76.2 75.3 72.5 35-44: Stole i cree rrr ee an 124.0 119.7 121.6 DIARIOlC. vim vm mim me mm wim on 79.3 77.8 27.0 45-54: Systolic... 1380.4 124.5 182.2 Diagiolie: ccd sas rsmmen rma 81.7 80.1 81.1 ! Nonpregnant, nonpostpartum subjects, including users of noncontraceptive hormones. 2 Data on white females, including pregnant and postpartum subjects from National Center for Health Statistics: Blood pressure of adults by race and area, United States, 1960-1962, Vital Health Stet. PHS Pub. No. 1000, Series 11, No. 5. Washington: U.S. Government Printing Office, July 1964. Moreover, it is necessary to analyze data from several years in order to determine the fraction of this variation, which is attrib- utable to causes persisting from one year to the next. COMMENT Alteration of the renin-angiotemsin system associated with oral contraceptive use is well established, but the frequency with which these women develop significant hypertension is controver- sial. ¢7 Isolated cases of severe hypertension in oral contracep- tive users have been reported; however, this is probably a rare event.! More convincing evidence that oral contraceptives affect blood pressure has been demonstrated by stopping oral contra- ceptives in hypertensive women and observing a drop in BP.° In our study, after age stratification, current users of con- traceptive drugs had a slight but statistically significant elevation in systolic BP. The diastolic BP rise was less marked. This is in agreement with the findings of Kunin et al.? whose study design is similar to our protocol. 122 160 140 | 120 BLOOD PRESSURE, mm Hg 80 60 100 - SOME INITIAL RESULTS TT 7 T I T 1 Si p SYSTOLIC MN ; iN IN FE 5 SN 8% TH AGE 15-24 - X 25.34 iN | 35.44 FN % FM \ AGE 15.24 \ 25.34 A % DIASTOLIC 35.44 Af ~ a | L 1 | | eo Ll <100 120 160 200 240 250+ WEIGHT, POUNDS FIGURE 17.—Mean blood pressure by weight and age for all oral contraceptive groups. Never users and past users of oral contraceptives had similar BP distributions and were, therefore, combined into a single group of nonusers. This suggests that the elevated blood pressure among contraceptive drug users is reversible. Analysis of data from women at the upper end of the dis- tribution of BP values is more striking, as demonstrated in figure TABLE 57.—Mean systolic blood pressure (mm Hg) by weight and age for all oral coniraceptive groups Age in years Total, all ages Weight (lb) 15-24 25-34 35-44 45-60 Mean | Number | Mean | Number | Mean | Number | Mean | Number | Mean | Number Under 100_________________ 113.36 22 112.19 37 113.73 33 128.55 29 116.74 121 100-119___________________ 118.47 165 118.46 527 118.78 437 123.19 364 119.75 1,493 190-189. ce 121.03 185 120.24 794 120.65 913 125.34 749 121.89 2,641 140-159 _ __ ________________ 123.43 72 121.16 319 112.26 471 127.30 457 123.81 1,319 160-179__________________ 126.96 23 124.81 106 123.37 186 131.21 190 126.78 505 180-199__ ____ _____________ 121.40 5 123.68 41 126.15 62 132.34 68 127.83 176 1 OS 137.00 3 122.67 9 126.67 30 130.57 35 128.38 77 280 oo es EERE ERR eR 126.25 12 125.12 17 134.67 9 127.74 38 240-259__ _________________ 135.50 2 | 147.50 6 | 154.25 4 138.50 2 146.43 14 260-279 ___ |e 122.00 Bc ttn ons 128.33 3 125.17 6 280 and over__________ I 100.00 x 110.00 1 130.44 9 166.50 2 132.08 13 SI 143.22 |... 221.89 licavonces 275.88 l.cdeannen 204.90 |..coo coca 396.50 |......... (144) *(170) * (153) *(190) *(190) *p<0.01 ZT HTYNSSTYd A00Td ANV ‘AONVNDHIY ‘SHAILIAOVILNOD TVIO TABLE 58.—Mean diastolic blood pressure (mm Hg) by weight and age for all oral contraceptive groups Age in years Total, all ages Weight (Ib) 15-24 25-34 35-44 45-60 Mean | Number | Mean | Number | Mean | Number | Mean | Number | Mean | Number Under 100... ... cumin 67.87 22 70.68 37 74.61 33 78.41 29 78.09 121 300-110... coin mmicn 72.74 165 73.51 527 78.97 437 78.02 364 74.22 1,493 120-139. .....ccveeencunn 74.60 185 75.13 794 77.05 913 79.60 749 77.02 2,641 10-180... .cvninvmwmmmnmns 76.72 72 77.67 319 79.01 471 81.74 457 79.50 1,319 60-370... ew mrirmi i 80.57 23 80.52 106 81.57 186 83.89 190 82.18 505 AB0-199. vii mimi 81.20 5 83.37 41 84.23 62 88.10 68 85.44 176 200-210... siti mimes 85.33 3 85.33 9 84.67 30 88.23 35 86.39 7 PUI... consi wigs sree wr fine te ms sine fo re ead 81.33 12 87.06 17 85.11 9 84.79 38 280-289... . ovum nm 98.50 2 95.83 6 100.00 4 94.50 2 97.21 14 BO0-079. oi wie) arm a re eia i] 93.33 Be ra 79.33 3 86.33 6 280 and OVer...um une mmnsw 71.00 1 74.00 1 88.67 9] 101.00 2 88.08 13 RY or ml wali ins meri 418.08 |..cuuu..- TRE BD Niet 759.09 |ocoeeeo-- 884.26 |e _- Lo B7 BT (ewmimmean *(72) *(140) * (126) (140) *(160) val SLTINSHY TVILINI INOS SYSTOLIC BLOOD PRESSURE (mm Hg) ORAL CONTRACEPTIVES, PREGNANCY, AND BLOOD PRESSURE 125 128 = ili oe gl 10 DIASTOLIC BLOOD PRESSURE (mm Hg) PARITY FIGURE 18.—Systolic and diastolic blood pressure by parity (age specific). 10. If a woman is selected at random, she is 1.76 times as likely to have a systolic BP 140 mm Hg or higher and diastolic pressure 90 mm Hg or over if she is an oral contraceptive user than if she is a nonuser. 126 SOME INITIAL RESULTS TABLE 59.—Blood pressure by parity for 4 age groups Parity and Age groups All ages ! blood pressure (um Hg) 15-24 | 25-3¢ | 35-44 | 45-60 0: Sytole. vores 120.98 120.56 121.92 126.39 122.38 Diastolic....c..--- 74.33 76.15 77.21 78.83 76.75 Number......... 269 319 201 237 1,030 1: Systolic. .......__ 119.86 119.04 122.94 126.21 122.07 Diastolic. c.cenc-~ 74.29 73.93 78.95 81.21 7.17 Number. ._...... 119 250 155 214 740 2: Systolic. ecueuea-- 119.98 120.13 120.67 126.86 122.23 Diastolic. ..cccoex 75.43 75.57 78.05 80.73 77.91 Number. ._.._... 82 658 559 583 1,884 3: Systolic cea onno. 129.27 120.83 121.77 126.31 122.83 Diastolic... .._ 7.27 76.08 78.13 80.68 78.34 Number.....««ee- 15 401 626 463 1,508 4: Systolit...cwnwncew 116.50 120.32 120.95 127.32 122.80 Diastolic.ocun..._ 76.00 76.68 78.00 81.02 78.64 Number. ........ 2 184 348 247 781 5: Systolic... covvervifeemeecomes 122.34 120.56 126.22 122.65 Diastolic..euv eer efeceeecennwe 76.15 78.61 81.07 79.20 Number....eeicejeceeeeneee 53 171 105 329 6: SYS cow crv enews cones 119.95 118.92 124.49 121.02 Diastolic. cocvvir ste casos 78.05 78.14 80.72 79.06 NUMBON. «ovo ilo mics 20 139 90 249 1 Includes a few individuals over 60 years of age. After establishing the effect of contraceptive drugs on the BP, the question arises whether the particular brand or dosage is important. Examination of these data demonstrated no evident relationship between these factors and elevated blood pressure. Pregnant women had BP readings slightly lower than the nonusers which is noteworthy in view of the concept that oral contraceptives may simulate a pseudopregnant state. However, these gravid women were in the first or early second trimester, and the BP may be different during later pregnancy. The blood pressures of postpartum women were difficult to ORAL CONTRACEPTIVES, PREGNANCY, AND BLOOD PRESSURE 127 125 SYSTOLIC T 124} = 1231- - T 122 i E E wo 121 0 oc =~ wv AL 2 120 oZ a 80 - oa 0 DIASTOLIC SO 791 _ 0 77 fe - N =1772 1600 430 148 | 76 0 A B AB ABO BLOOD TYPE FIGURE 19.—Blood pressure by ABO blood type (mean and 95-percent con- fidence intervals of the means). evaluate since 10 percent were taking oral contraceptives, and 52 percent received an injection of male and female sex hormones (Deladumone) for cessation of lactation. This has been discussed previously. These findings are of interest since one prospective study has utilized 4-week postpartum subjects for baseline determinations and reported that 18 percent of 56 normotensive women had at least a 10-mm rise in diastolic blood pressure when given oral contraceptives.’ 128 SOME INITIAL RESULTS TABLE 60.—Mean blood pressure (mm Hg) by ABO blood type Blood type and pressure Mean Standard Number deviation Type 0: mm Hg SYSORE... co vse clic wea 121.08 15.18 1.772 Diastolic... cv covnnisnminmmmmmsns 77.56 8.87 | 2 Type A: BYSIOHC cose mmm in smn mn na 120.89 15.46 1.600 DIRS ONE re en meni mmm mmm ees 77.45 9.04 : Type B: BYE OHe. cnn m ne nmm mmm rn mma 122.12 14.37 430 DEO. es cmon 77.35 8.85 Type AB: BYSOHC. cc wwii sm mmm 122.49 12.93 148 Diastolic nic icwn snes 78.16 10.97 Total all ages: SYSOlL. reeves memnenisnsma 121.17 15.12 Diastolic. moomoo 77.52 8.97 3,950 Other factors that might account for the observed blood- pressure differences among our subgroups were analyzed. Most obviously, age affects BP. Accordingly, all data were analyzed age-specifically. The observation that the BP rise associated with oral contraceptives is more marked as a woman grows older is of interest. Weight or body build did not apparently explain our observations. On the contrary, current OC users were lighter than nonusers, even after age grouping. In addition, menstrual cycle, parity, and diurnal or seasonal variation did not influence our findings as discussed previously. The higher prevalence of hypertension in American Negroes is well documented,* and race may exert a strong effect on ele- vated blood pressure associated with oral contraceptive use. Because our study population is predominantly Caucasian, we are unable to comment on this variable. In summary, oral contraceptives are associated with a slight elevation of blood pressure, which appears to be reversible and not apparently related to dose or components of oral contracep- tives. Nevertheless, the fact that severe hypertension is probably an unusual occurrence can be reassuring to the great majority of women taking these compounds. Whether a minimal elevation of blood pressure is ultimately detrimental to these women is unknown at this time. ORAL CONTRACEPTIVES, PREGNANCY, AND BLOOD PRESSURE 129 129 4 T T T T T T T = SYSTOLIC 126 z 123 - € - g \ wi \ : ot 1 hd 120 —o* tenes’ \ Nd pa wv % o° wd = vo - = |] Qe. 4 i. 3 nz u J l | J J | J S 82 T T T T T T T ® iq DIASTOLIC - 79 7] Sn Lo Cag “NON USER | 01 5 9 13 17 29 25 29 33 DAY OF MENSTRUAL CYCLE FIGURE 20.—Blood pressure by day of the menstrual cycle for users and nonusers of oral contraceptives. REFERENCES 1. WEIR, R. J., Briggs, E., MACK, A., TAYLOR, L., BROWNING, J., NAISMITH, L., and WiLsoN, E.: Blood-pressure in women after one year of oral contraception. Lancet 1: 467-471, Mar. 6, 1971. 2. KuNIN, C. M., McCoRMACK, R. C., and ABERNATHY, J. R.: Oral contra- ceptives and blood pressure. Arch Intern Med 123: 363-365, Apr. 1969. 3. GRUEN, W.: An assessment of present automated methods of indirect blood pressure measurements. Ann NY Acad Sci 147: 109-126, Feb. 9, 1968. 130 SOME INITIAL RESULTS TABLE 61.—Systolic blood pressure by day of menstrual cycle for users and nonusers of oral contraceptive drugs Day of menstrual cycle Systolic blood pressure (mm Hg) Oral contraceptive users Oral contraceptive nonusers Stand- Stand- Mean ard |Number| Mean ard | Number error error Orne rere de 118.20 8.31 15 | 125.50 3.46 30 Ih. csc cnnnennnens 128.71 2.81 38 | 121.05 1.65 84 Be cr mr St 124.83 1.19 186 | 120.12 .80 413 =] ii dno ge na 128.63 .84 311 | 120.80 .65 592 18-16. ol ci ces eens 124.68 1.08 262 | 120.36 .65 579 Y= vee en sees 121.43 1.07 170 | 119.78 .T4 428 mR i cma iii, 125.16 1.37 136 | 122.79 .85 391 TE 125.83 2.38 57 | 118.77 1.60 99 2G ena 124.27 1.89 66 | 120.23 1.35 160 TABLE 62.—Diastolic blood pressure by day of menstrual cycle for users and nonusers of oral contraceptive drugs Day of menstrual cycle Diastolic blood pressure (mm Hg) Oral contraceptive users Oral contraceptive nonusers Stand- Stand- Mean ard |Number| Mean ard | Number error error Occ mn ie 76.27 2.22 15 | 76.97 2.14 30 Ist tsi vd asin 81.16 1.57 38 | 77.90 .96 84 BB. conn anemEE 78.06 .67 186 | 76.65 .29 413 Bl acini maa 77.49 .51 811 | 177.33 .38 592 13-10 icin 77.69 .62 262 77.82 41 579 10-00. ic cridanmssns 76.72 .70 170 76.81 .43 428 224 ieee 77.54 .80 186 | 178.21 .45 391 08 vin seen mmm 78.61 1.50 57 | 77.84 .95 99 29-82... nnamamn 78.56 1.04 66 | 78.39 .78 160 4. CLARK, V. A., CHAPMAN, J. M., CouLsoN, A. H., and HASSELBLAD, V.: Dividing the blood pressures from the Los Angeles heart study into two normal distributions. Johns Hopkins Med J 122: 77-83, Feb. 1968. 6. National Center for Health Statistics: Blood pressure of adults by race and area, United States, 1960-1962. Vital Health Stat PHS Pub. No. ORAL CONTRACEPTIVES, PREGNANCY, AND BLOOD PRESSURE 131 125 T I SYSTOLIC 124 —] 1234 — oO ’ E E1221 - wr oc - wv 0 121 . oc a. o 0 120 LI | | | | — or | | | | 791" piasToLIC i 78 — — 77 L | | 10a.m. 1la.m. 12 m. 1p.m. 2p.m. TIME OF DAY FIGURE 21.—Blood pressure by time of day, irrespective of hours since last food. 1000, Series 11, No. 5. Washington: U.S. Government Printing Office, July 1964. 6. SARUTA T., SAADE, G. A., and KAPLAN, N. M.: A possible mechanism for hypertension induced by oral contraceptives. Arch Intern Med 126: 621-626, Oct. 1970. 7. Tyson, J. E. A.: Oral contraception and elevated blood pressure. Am J Obstet Gynecol 100(6) : 875-876, Mar. 15, 1968. 8. HARris, P. W. R.: Malignant hypertension associated with oral contra- -ceptives, Lancet 2: 466-467, Aug. 30, 1969. 152 SOME INITIAL RESULTS 9. WEINBERGER, M. H., CoLLiNS, R. D., Dowpy, A. J., Nokes, G. W., and LUETSCHER, J. A.: Hypertension induced by oral contraceptives con- taining estrogen and gestagen. Ann Intern Med 71(5) : 891-902, Nov. 1969. 10. CRANE, M. G., Harris, J. J., and WINSOR, W.: Hypertension, oral contra- ceptive agents, and conjugated estrogens. Ann Intern Med T4(1): 13-21, Jan. 1971. 11. BoyLE, E., Jr.: Biological patterns in hypertension by race, sex, body weight, and skin color, JAMA 213(10) : 1637-1648, Sept. 7, 1970. TABLE 63.—Systolic blood pressure by time of day and hours since last food Time of day Hours since last food 10 a.m. 11 a.m. 12 p.m. 1 p.m. 2 p.m. Mean | Number | Mean | Number | Mean | Number | Mean | Number | Mean | Number 0-4. icine 121.96 78 | 119.82 “134 | 121.87 101 125.72 98 | 122.66 19 BB. comin mmm n 124.30 40 | 121.78 209 | 122.98 689 | 123.60 943 | 124.64 263 0-12 cen a em —— 120.21 1,134 122.25 1,850 | 122.71 894 | 123.44 457 | 124.58 62 Total... 120.45 1,252 122.00 1,693 | 122.74 1,684 | 123.69 1,498 | 124.52 344 TABLE 64.—Djastolic blood pressure by time of day and hours since last food Time of day Hours since last food 10 a.m. 11 a.m. 12 p.m. 1 p.m. 2 p.m. Mean | Number | Mean | Number | Mean | Number | Mean | Number | Mean | Number 0-0 rrr ame 76.50 78 75.52 134 76.37 101 77.55 98 76.32 19 BB ninco am 79.50 40 78.12 209 78.40 689 78.06 943 77.64 263 1 77.62 1,134 77.55 1,350 78.61 894 79.12 457 77.10 62 NWolal. cients 77.61 1,252 77.46 1,693 78.39 1,684 78.85 1,498 77.47 344 G81 HYNSSHAd 001d ANV ‘XONVNDHITId ‘STAILIAOVILNOD TVIO Chapter 2 Sensitivity and Specificity of the 1-Hour Glucose-Tolerance Screening Test NANCY R. PHILLIPS AND THOMAS J. DUFFY Glucose-tolerance screening procedures—Upon arrival for the Auto- mated Multitest Laboratory (AML) examination, the patient received 75 g glucose. Patients had been asked to not take anything except water for at least 4 hr before their appointment. Blood was drawn 1 hr after the challenge. The serum was separated within 30-60 min. Glucose con- centration was obtained by the orthotoluidine method. A urine specimen was tested for ketone bodies. Followup procedures.—After some refinement, the following criteria for followup testing with a standard glucose-tolerance test was estab- lished. Women challenged before 10:30 a.m. were scheduled for a stand- ard test if their serum glucose screening level was 200 mg/100 ml or more. For those challenged after 10:30 a.m., a screening level of 210 mg/100 ml for fasting, or 220 mg/100 ml for nonfasting patients was established. The standard glucose-tolerance test was scheduled following an overnight fast. Patients were prepared with a 800-g carbohydrate diet for 3 days preceding the test. Blood was drawn at 1-, 2-, and 3-hr inter- vals following the challenge. Initially a uniform 100-g glucose-loading dose with given; subsequently 40 g per square meter of body surface was given. Evaluation of the 1-hr test.—Of the patients referred for a standard glucose-tolerance test, 9.6 percent had abnormal curves as defined by the Wilkerson point system. The predictive value of a positive screening test was then used to evaluate the screening tests’ sensitivity and specificity. Assuming a prevalence rate of glucose intolerance of 1 percent, the sensitivity of the screening test was 48 percent; its specificity was 95.4 percent. This report presents results of an evaluation of the sensi- tivity and specificity of a casual 1-hr glucose-tolerance test ad- ministered in the Automated Multitest Laboratory (AML). Included is a description of the screening and followup procedures. 1856 136 SOME INITIAL RESULTS GLUCOSE-TOLERANCE SCREENING PROCEDURE From December 1968 to November 1972, the testing proce- dure was as follows. The patient registered for her examination, scheduled between 9 a.m. and 1 p.m.; she was immediately given 75 g glucose in a 6-0z aqueous solution of artificial lemon-lime flavor to which 2 oz crushed ice had been added. The time of ingestion and the amount taken were recorded. No preparatory dietary instruction was given, but examinees were asked to take nothing by mouth except water for at least 4 hr before their appointment. When the glucose drink was given, the patient was questioned to determine when she last ate or drank anything of caloric value; the interval since last food was recorded in completed hours. If the subject had eaten within the preceding 4 hr, the type and amount of each food item was noted on the back of the punch card, but this information was not included in the machine-readable data set. Sixty minutes after the challenge, a sample of venous blood was drawn. The serum was separated within 30-60 min of the venipuncture. The glucose concentration was determined by the orthotoluidine method, which is based on a color reaction given by aldosaccharides with orthotoluidine in glacial acetic acid containing thiourea.! The results obtained by this method correlate well with those obtained with the AutoAnalyzer* and Beckman.? On 779 split-half pairs, the correlation between the AutoAnalyzer and orthotoluidine methods was 0.94; the correlation with the Beckman on 116 split-half pairs was 0.97. A urine specimen was collected immediately after the veni- puncture and tested for ketone bodies with a Ketostix reagent strip.c The test was read as positive only if there was a decided color change. An analysis of the first 2,746 nonpregnant women tested showed that, in 95 percent of cases, the uniform 75-g glucose challenge met or exceeded the loading dose of 40 g per square meter of body surface recommended by the American Diabetes Association? (table 65). Age-specific mean serum-glucose levels 1 hr following the 75-g challenge are shown in relation to the recommended load for body size in table 66. Although there was * AutoAnalyzer, Technicon Instruments Corp., 511 Benedict Ave., Tarry- town, N.Y. 10591. * Beckman Glucose Oxidase Method, Beckman Instruments, Inc., Spinco Division, 1117 California Ave., Palo Alto, Calif. 94304. © Ketostix, Ames Co., division of Miles Laboratories, Inc., Elkhart, Ind. 46514. ONE-HOUR GLUCOSE-TOLERANCE SCREENING TEST 137 TABLE 65.—Recommended glucose load Recommended glucose load ? Number Percent UDAEY BB. sev crnnmnssmnnsnsmsasenssmsnsnsenns 48 1.8 BEB. oe mre ms mn mR A ESR 411 15.0 OIE cent crm msm ed i en Sloe 923 33.6 BBD... coms mn ns msm mss EE sre swe wa 865 31.5 AD cs nvm am mmr mmm me STS AE ESS A 856 12.9 Rl 2 nes tren edn mm we Se AS mv SS 98 3.6 BE rus r svar rm ar Ee ie rr 24 .9 86-90. 11 a{%8 BY. and OVeL. ser sunnnmevs sesame eT 11 4 D0 A renee er ers errr angie we wm 2,746 100.0 140 g glucose per square meter of body surface. Source: Committee on Statistics of the American Diabetes Association, June 14, 1968. Report. Standardization of the oral glucose tolerance test. Diabetes 18: 299-310, May 1969. a slight trend toward higher 1-hr values as the difference between the 75-g challenge and the recommended loading dose increased, “overloading” had no appreciable effect on the test result. Higher 1-hr values were found among those women who were large enough to require a loading dose greater than 75 g. These higher values may reflect an association between obesity and glucose intolerance. The time of day when challenged, the interval since last food, ketonuria, and age were all associated with variation in the post- challenge values. The effects of each of these variables on the test results are described in part B, chapter 3. FOLLOWUP PROCEDURES Until June 15, 1970, the criteria for followup testing with a standard glucose-tolerance test were specific for age and the time of day when challenged (table 67). What was originally thought to be the effect of time of day on the postchallenge value was found to be primarily the effect of the interval since last food: the proportion of fasting subjects (9 hr or more without food) decreased as the morning progressed, while the proportion who had eaten 4-6 hr earlier increased. The criteria for followup testing were then changed to take into account the interval since last food (table 68). Since few women challenged before 10:30 a.m. had been without food for less than 9 hr, only one screening limit TABLE 66.—Mean 1-hr serum-glucose level following 756-9 challenge in relation to recommended load for body size Age in years Weighted overall Recommended Under 30 30-39 40-49 50 and over mean glucose load ! serum (2) glucose 2 Mean Mean Mean Mean (mg/100 ml+ Number serum Number serum Number serum Number serum [standard error) glucose glucose glucose glucose ; mg/100 ml mg/100 ml mg/100 ml mg/100 ml 60 and under. _________ 150 130.0 139 154.7 119 156.0 37 171.5 | 150.7+1.9 BBB iis, cen wanes 257 136.4 283 146.9 263 158.8 99 165.1 | 149.9+1.3 0690... inant cna 201 133.8 264 141.8 283 161.0 95 169.0 | 148.7+1.3 08. nisi whom 79 127.9 103 146.4 126 159.2 41 165.7 | 147.7+2.1 76 andover___.________._ 17 175.2 39 148.6 65 163.6 13 147.1 | 160.2+3.5 1 Recommended load is 40 g per square meter of body surface. 2 Weighted to the overall age distribution of the women tested. 861 SLTASAY TVILINI INOS ONE-HOUR GLUCOSE-TOLERANCE SCREENING TEST 139 TABLE 67.—Screening limits in effect from Jan. 1, 1969, to June 15, 1970 Ag= in years and time of challenge Serum glucose Under 40: mg/100 ml D DIY, Y 330 D100 em oo ro i 180 40-49: 9-10:59 am... eeeeeeeeeemmmmm———————————— 190 11 20-12:20 Bill. eevee sess s reer ae me eee 200 12:80-1:80 Pili. «comin mrmismma ese RE REE 210 50 and over: 9-10:59 8M. emcee meme m—mm———————— 210 TL BITE B0 DTN. oo ioe vi soso mons st i i 220 TABLE 68.—Screening limits in effect from June 16, 1970, to Oct. 18, 1972 Time of challenge and hours since last food Serum glucose mg/100 ml OID IO BW ee tsi rs po mim mmo iim it m1 200 10:30 a.m.-1:30 p.m.: 9 hr OF MOT. — oo ee eee eee eee mmm 210 Last Nan 9 BT: - cv imoncve oss mioe ems imams aaa 220 was set for this time of day. Two screening limits were set for women challenged after 10:30 a.m., one for fasting subjects and one for those who had eaten less than 9 hr prior to the challenge. Since the contribution of disease process to the heightened 1-hr response of older persons was unclear, all reference to age was deleted from the revised criteria. The screening limits given in table 68 represent the 95th percentile of the glucosemia distribution—specific for time of day when challenged and the interval since last food—among non- pregnant women tested during the first 5 months of 1969. Nearly all (99 percent) these women were below 55 years of age. The standard glucose-tolerance test was scheduled between 8 and 9:30 a.m. following an overnight fast. The patient was put on a preparatory diet containing 300 g carbohydrate for 3 days preceding the test. Blood samples were drawn while fasting and 1, 2, and 3 hr following the challenge. Until June 1970 a uniform 100-g glucose-loading dose was given. After that date, in con- formity with the recommendation of the American Diabetes Association, a loading dose of 40 g per square meter of body 140 SOME INITIAL RESULTS surface was given. The loading dose was determined from a simplified table based on weight and height (table 69). EVALUATION OF THE 1-HR TEST The precision with which a test distinguishes diseased from TABLE 69.—Glucose loading dose by height and weight based on 40 g per square meter of body surface Height Weight in pounds 4ft8in~ | 5ft2in~ | 51t8in~ 5 ft 1in. 548t7in. 6 ft 1 in. Grams Grams Grams R0E08 cc nse ates 50 55 60 RS sa ER er 55 60 65 N16=T40. _ oo cinns mms 60 65 70 I=100 cece cman nner mem 65 70 75 0B iio nie nnn 70 75 80 106-228. eee eececn mee m——— 75 80 85 LL 80 85 90 BD a rE EEE EERE 185 90 95 B= rs cine 290 195 100 Be oe i —————— 395 1100 105 1 A few patients will get 6 g over. 2 A few patients will get 6-7 g over. 3 A few patients will get 6-8 g over. NoTe.—Patients over 6 ft 1 in. receive an extra 5 g; patients under 4 ft 8 in. receive 5 g less for weight category. With the exceptions given below, use of this table yields a glucose load that is within 5 g of the recommended load calculated on the basis of body surface area: Not 2 * Diseased Diseased* b a ‘‘true positives” + a Se Tost b ‘false positives c ‘false negatives’ — Test c d = a d ‘“true negatives *By all criteria available or by “best” criterion available. FIGURE 22.—Classification of test results for determining sensitivity and specificity. ONE-HOUR GLUCOSE-TOLERANCE SCREENING TEST 141 nondiseased persons is described by its sensitivity and specificity. These terms are clearly defined by the fourfold table in figure 22. Sensitivity is the proportion of positive tests among diseased persons, or a/(a + c¢). Specificity, on the other hand, is the proportion of negative tests among nonaffected individuals, or d/(b + d). The false-negative rate and the false-positive rate are the respective complements of sensitivity and specificity. When a screening test measures a continuous variable, such as serum-glucose concentration, a screening limit is set to dichot- omize the results as either positive or negative. The selection of the screening limit is arbitrary, since there is no point on the continuum which clearly separates diseased and nondiseased per- sons. Regardless of the screening limit selected, some diseased individuals will be missed and some well persons will be referred unnecessarily for followup testing. As the screening limit is varied, sensitivity and specificity are both affected, but in opposite directions. If, for example, the screening limit is lowered, the sensitivity of the test increases, but its specificity decreases. That is, a decrease in the rate of false negatives is accompanied by an increase in the rate of false positives. Table 70 shows the inverse relationship of sensitivity and specificity observed at various screening limits in an evaluative study of the efficiency of a postprandial screening test.’ Both the TABLE 70.—Somogyi-Nelson blood test (venous) sensitivity and specificity for diabetes at different screening levels, by hours after test meal At 1 hr after test meal | At 2 hr after test meal Blood sugar level considered positive (mg/100 ml) Sensitivity,| Specificity, [Sensitivity,| Specificity, percent percent percent percent positive | negative | positive | negative Bc mn SR RR er 100.0 8.2 98.6 8.8 ° Be tn mmm mm mi) 97.1 22.4 97.1 25.5 00 ne He Err vr sree 97.1 39.0 94.3 47.6 EF er Eo a 95.7 57.3 88.6 69.8 10... ovivnvina crm amnmmaneen 92.9 70.6 85.7 84.1 100 co mE 88.6 83.3 71.4 92.5 80 on bm T riser Sn war 78.6 90.6 64.3 96.9 40, ee vpvdves sonar dma 68.6 95.1 57.1 99.4 180... coin mens m sms aa 57.1 97.8 50.0 99.6 X80... cc fm 52.9 99.4 47.1 99.8 Source—Remein, Q. R., and Wilkerson, H. L. C.: The efficiency of screening tests for diabetes. J Chronic Dis 13: 6-21, Jan. 1961. By permission. 142 SOME INITIAL RESULTS 1- and 2-hr postprandial sampling points are shown for com- parison. The final arbiter of disease was a score of two or more Wilkerson points? on a standard glucose-tolerance test. At a screening limit of 100 mg/100 ml, the 1-hr postprandial blood sugar had a sensitivity of 95 percent, but a specificity of only 57 percent. In other words, nearly half of normal individuals would have a positive test result at this screening limit. At a screening limit of 140 mg/100 ml, specificity increased to 95 per- cent, but sensitivity fell to 68 percent. That is, 3 out of every 10 diabetics would be missed. For practical reasons, high sensitivity cannot be achieved at the expense of specificity. Reasonable specificity is important to both patients and physicians. Yet, if the test is to be useful in bringing unknown cases of disease to medical attention, it must also be reasonably sensitive. In practice, the sensitivity and speci- ficity of a screening test are rarely determined; the screening limit is simply set to yield a “comfortable” proportion of positive tests. As noted earlier, the screening limits for the AML glucose- tolerance test were set so that 5 percent of the tests would be positive. Since the prevalence of undiagnosed glucose intolerance is likely to be low in a population having multiphasic health checkups, we can feel confident that the specificity of the test is high. We would be surprised if it were less than the comple- ment of the followup referral rate, or 95 percent. The sensitivity of the test, however, could range from near zero to near 100 percent. As an alternative to getting standard glucose-tolerance tests on a probability sample of AML examinees, we used a known quantity, the predictive value of a positive test, to evaluate the test’s sensitivity. The predictive value of a positive test is defined as the pro- portion of affected individuals found among those referred for followup testing. In terms of figure 22, it is ¢/(a + b). The predictive value of a positive test is a function of both sensitivity and specificity and, as well, the prevalence of disease in the popu- lation screened. The lower the prevalence of disease, the lower the probability that a person with a positive test will have the disease in question. Since even common diseases have low preva- lence rates, the predictive value of a positive test may be sur- prisingly low. With a disease prevalence of 1 percent, and a screening test having sensitivity and specificity both at 95 per- cent, the predictive value of a positive test is only 16 percent.* Of the AML examinees referred for followup testing under ONE-HOUR GLUCOSE-TOLERANCE SCREENING TEST 143 the revised criteria of June 1970, 9.6 percent had abnormal glu- cose-tolerance curves as defined by the Wilkerson point system.? The proportion with abnormal curves by the Fajans-Conn criteria? was 24.9 percent. To illustrate how the predictive value of a positive test was used to evaluate the test’s sensitivity and speci- ficity, let us assume that the prevalence of glucose intolerance as defined by a score of two or more Wilkerson points on a standard glucose-tolerance test is 1 percent. In a population of, say, 10,000, a total of 100 people would be affected. Our referral rate is 5 percent; thus, 500 of the 10,000 people would receive a standard glucose-tolerance test. Since the predictive value of a positive test under the Wilkerson point system is 9.6 percent, 48 of the positive subjects would show abnormal results on followup testing. Now that we know one cell value and its row and column totals, we know all cell values of the fourfold table represented in figure 22. The values are as follows: Diseased Nondiseased Row total Positive test __________________ 48 452 500 Negative test _________________ 52 9,448 9,500 Column total —_—.________. 100 9,900 10,000 Thus, assuming a prevalence rate of 1 percent, the sensitivity of the screening test is 48/100, or 48 percent; its specificity is 9,448/9,900, or 95.4 percent. In the general case, sensitivity is found as RK/P and specificityas [(1 — P) —R(1—K)]/(1 — P), where R is the referral rate, K is the predictive value of a positive test, and P is the prevalence of disease, all expressed as pro- portions. We do not know, of course, the actual prevalence of disease among our subjects, but we have calculated the sensitivity and specificity of the test at various postulated prevalence rates. The results of our calculations are shown in table 71. It will be noted that sensitivity is 100 percent when disease prevalence (P) equals RK, the referral rate times the predictive value of the test. The sensitivity of our screening test would be 100 percent if the prevalence of glucose intolerance, as defined by the Wilkerson point system, were 0.0048, less than 0.5 percent. For the Fajans-Conn criteria, sensitivity would be 100 percent if prevalence were 0.0124. In the Sudbury study, where standard glucose-tolerance tests were done on a random sample of the town’s population 15 years of age and over, the prevalence of previously unrecognized glucose intolerance was 1.2 percent, as defined by the Wilkerson point 144 SOME INITIAL RESULTS TABLE 71.—Sensitivity and specificity of Automated Multitest Laboratory glucose-tolerance screening test at postulated disease-prevalence rates Wilkerson point system Fajans-Conn criteria Postulated preyaiencs of Sens Smit isease ensi- fick ensi- ios (P) tivity APR ka tivity a-ak RK,/P Be RK,/P oli i a 0.0048.......-- 1.00 0.958 te cum ai re a pre 0.008... .96 SN mints insane i im 0.0075..ccuen.- .64 IBA Yo tei mith hig esis lin liseli ii 0.0). .48 JOBE | en mre oS a 0012. ee .40 2OB8 ene n fe nna nie 0012. .39 .954 1.00 0.962 0.015...cucuens .32 .954 .83 .962 0.02; . cc cnmnnus .24 .954 .62 .962 0:03. nnrnmns .16 .953 .41 .961 0.04... ..cninens .12 .953 .81 .961 0.056... cnn .09 .952 .22 .960 NoTteE.—Referral rate: BR =0.05. Predictive value of a positive test: Wilkerson point system: K; =0.096; RK, =0.0048 Fajans-Conn criteria: K2=0.249; RK,=0.0124 system, and 5.6 percent by the Fajans-Conn criteria.’ At these prevalence rates, the sensitivity of our screening test is 40 percent for the Wilkerson point system and only 22 percent for the Fajans-Conn criteria (table 71). Since most AML exam- inees were under 55 years of age, however, the prevalence of glucose intolerance among these women is likely to be lower than that of the Sudbury population. Nonetheless, as shown in table 71, the prevalence of disease would have to be very low for our test to have acceptable sensitivity. When the frequency of abnormal standard glucose-tolerance tests was analyzed by the conditions of the initial screening test, it became clear that the predictive value of a positive test varied widely. The frequency of abnormal standard glucose-tolerance tests is shown in table 72 by two screening conditions—time of day and hours since last food. The likelihood of a positive screen- ing indicating disease, as defined by either the Wilkerson point system or the Fajans-Conn criteria, was greatest in fasting examinees tested in the early morning and least in the late- morning, nonfasting group. Theoretical sensitivities and speci- ficities for various postulated rates of disease are not shown in ONE-HOUR GLUCOSE-TOLERANCE SCREENING TEST 145 TABLE 72.—Frequency of abnormal standard glucose-tolerance test (SGT'T) among women referred for followup testing in relation to the conditions of the screening test Frequency of abnormal SGTT Number Time of challenge and rece’ving Wilkerson point Fajans-Conn hours since last food SGTT system criteria Number | Percent | Number | Percent 9-10:29 a.m.: 0-3 A cco crm cwiememrm 10 1 10.0 1 10.0 CE 15 2 18.8 4 28.7 9 hrormore. ....._._... 49 10 20.4 22 44.9 10:30 a.m.-1:30 p.m.: 0-8 Br... cccmememanmun 9 0 0 3 33.3 4-8 hr _________.._. 79 3 3.8 10 12.6 9hrormore........_.. 47 4 8.5 12 25.5 Overgll...ccocnswsnns 209 20 9.6 52 24.9 the table, but it is clear that both sensitivity and specificity are impaired in the late-morning, nonfasting group of examinees. For this reason, beginning in 1972, all study subjects will be screened in a fasting state between 8 and 9:30 a.m. REFERENCES 1. HyVARINEN, A. and NikkiLA, E. A.: Specific determination of blood glucose with o-toluidine. Clin Chim Acta 7: 140-143, 1962. 2. Committee on Statistics of the American Diabetes Association, June 14, 1968: Report. Standardization of the oral glucose tolerance test. Dia- betes 18: 299-310, May 1969. 3. REMEIN, Q. R., and WILKERSON, H. L. C.: The efficiency of screening test for diabetes. J Chronic Dis 13: 6-21, Jan. 1961. 4. VEccHIO, T. J.: Predictive value of a single diagnostic test in unselected populations. N Engl J Med 274: 171-173, May 26, 1966. 5. O’SULLIVAN, J. B,, and WiLLiAMS, R. F.: Early diabetes mellitus in per- spective—a population study in Sudbury, Massachusetts. JAMA 198: 579-582, Nov. 7, 1966. 2 i gi ci aly ean 1 Le] oh pr BH aks a feds gi ih Ei ia FB Be Br Chapter 3 Effect of Age, Hours Since Last Food, Time of Day, and Ketonuria on 1-Hour Glucose Tolerance’ NANCY R. PHILLIPS AND THOMAS J. DUFFY The data in this report were generated from December 1968-Decem- ber 1969 by the Automated Multitest Laboratory. Study subjects.—The mean age of the 6,692 nonpregnant women studied was 37.9 years. Methods.—Testing procedures are described in part B, chapter 2, “Sensitivity and Specificity of the One-Hour Glucose-Tolerance Screen- ing Test.” Statistical analysis.—Although the distribution of 1-hr glucose toler- ance was skewed slightly to the right, the skewness did not invalidate the use of the arithmetic mean or the standard deviation. The skewness was also unrelated to age or the other variables studied. Age.—The rise with age in mean 1-hr glucose of 1.3 mg per 100 ml per year was accompanied by a linear increase in the location of all deciles of the glucosemia distribution, but the rate of increase was greater at the higher than the lower deciles. Hours since last food.—Marked increases in 1-hr values were found as the interval between the last food and the challenge lengthened from 1-6 hr. Women tested after an overnight fast, however, were far more similar in glucose response to women challenged less than 4 hr after eating than to women who had eaten 4-6 hr earlier. Ketonuria.—Among those tested for ketonuria, 7.5 percent had ketone bodies in their urine. The prevalence of ketonuria increased with time of day, regardless of the interval since the last food. Women with ketonuria had elevated 1-hr serum-glucose levels. The observed overall mean difference between ketonurics and nonketonurics of glucose levels was 29 mg per 100 ml. Time of day.—A small but definite tendency for 1-hr glucose values to increase between early morning and midday was found. This tendency was independent of the interval since last food and the prevalence of ketonuria. * Reproduced (with revisions) from Health Services Reports, Health Services and Mental Health Administration, U.S. Department of Health, Education, and Welfare, Vol. 87, No. 7, pp. 649-657, Aug.—Sept. 1972. 147 148 SOME INITIAL RESULTS Discussion.—This study’s findings on glucose levels and age and hours since last food is consistent with previous studies. The interpre- tation of the increase in prevalence of ketonuria with time of day is unclear, as is the interpretation of the high postchallenge glucose values that accompanied the ketonuria. Conclusion.—Some diminution in glucose tolerance at the 1-hr sampling point may be typical of most people as they age, but the rate of deterioration varies from person to person. These data indicate that a screening test 4-6 hr after a meal is not comparable to one administered in a fasting state. Age, interval since last food, time of day, and ketonuria have been shown to affect the results of the 1-hr glucose-tolerance test.-¢ As part of an effort to establish screening norms, the effect of these variables on 1-hr glucose tolerance was further analyzed during the 13-month period December 1968—-December 1969. MATERIALS AND METHODS STUDY SUBJECTS None of the women included in the analysis was known to be pregnant at the time of the multiphasic examination, and all took and retained the full glucose challenge. Known diabetics were excluded from the analysis. Four previously undiagnosed but overt diabetics were also excluded because their extremely high post- challenge values tended to distort the summary statistics. The mean age of the 6,692 women studied was 37.9 years. Although the women’s ages ranged from 18 through 72 years, all but 1.5 percent were 20-54 years of age. One-third of the women were taking an oral contraceptive at the time of the examination. The 1-hr postchallenge serum-glucose levels of users of oral contraceptives were approximately 10 mg per 100 ml higher, on the average, than those of nonusers of the same age. This difference between users of contraceptive drugs and nonusers, however, had no effect on the relationship of 1-hr glucose tolerance with age, interval since last food, time of day, or ketonuria that is described in this report. The prevalence of ketonuria and the distribution in the times of challenge and the hours since last food were the same for both users and nonusers. Although the proportion of the women who were contraceptive drug users was inversely related to age, the coefficients of the regression of 1-hr serum glucose on age, com- puted separately for users and nonusers, were not discernibly different from that computed for the two groups combined. EFFECTS ON ONE-HOUR GLUCOSE TOLERANCE 149 METHODS The testing procedures for glucose tolerance are described in part B, chapter 2, “The Sensitivity and Specificity of the One-Hour Glucose-Tolerance Screening Test.” STATISTICAL ANALYSIS The distribution of 1-hr glucose tolerance, like many physio- logical variables, was skewed to the right (fig. 23). The skewness was too slight, however, to invalidate the use of the arithmetic mean as a description of central tendency or of the standard deviation to describe dispersion. Furthermore, the degree of skew- ness, as evaluated by the difference between the mean and median, was unrelated to age or the other variables studied. Certain quartiles are also used to describe changes in the location and dispersion of the glucosemia distribution associated with increas- ing age. FINDINGS AGE As noted earlier, the effective range of ages covered by these —— Observed Distribution ---- Normal Distribution PROPOR- TION .20 [ 15 1° 05 [- == ~ ~—t—— Ls, A N N » 1 1 1 130 170 210 250 290 330 ONE-HOUR SERUM GLUCOSE - mg/100 ml. FIGURE 23.—One-hr serum-glucose relative frequency distribution. .00 150 SOME INITIAL RESULTS TABLE 73.—The mean and standard deviation of 1-hr serum-glucose levels (mg per 100 ml), by age [6,692 women examined over a 13-month period] Serum glucose Number Age in years of women Mean Standard deviation 20-24 ii cto sh eke ae 558 127.1 30.9 D500. mn ede Sse 919 132.8 35.1 11 LN a 1,042 140.4 88.1 BBB i] ie See tem sme 1,138 144.5 40.9 BOSAL ric wii ite was 1,156 154.4 41.8 BBA or Se 1,060 160.7 41.1 BO-Bh oo nv i ham 731 165.3 43.7 data was 20-54 years. Within this range, 1-hr serum-glucose values increased linearly with age at the rate of 1.3 mg per 100 ml per year. The age-specific mean values are given in table 73 by b-year age intervals. These values are also plotted in figure 24. Not only the mean, but the entire distribution of values shifted upward with age. Thus, as shown in figure 25, the location of all deciles of the distribution also increased linearly with age, but at different rates. In general, the higher the decile, the greater the rate of increase. The mean 1-hr blood-sugar and regression coefficients of selected percentile locations on age are shown in table 74. The heterogeneity of the decile regression coefficients reflects the increased dispersion of the glucosemia distribution with age. The interquartile range, for example, widened from 40 mg per 100 ml at ages 20-24 to 58 mg per 100 ml at 50-54 years. Thus, the variability among examinees in 1-hr serum glucose also tended to increase with age. HOURS SINCE LAST Foop The mean 1-hr values are shown in table 75 and figure 26 by hours since last food when challenged. As can be seen, the post- challenge levels increased as the interval since the last food lengthened from 1-6 hr, and then they began to decrease. Because the number of examinees who last ate 7 or 8 hr earlier was small, these data are insufficient to describe exactly what happens as the interval lengthens beyond 6 hr. It is clear, however, that the post- EFFECTS ON ONE-HOUR GLUCOSE TOLERANCE 151 170 ~ 160 (- 150 140 SERUM GLUCOSE LEVEL (mg/100 m1) 130 120 - 1 1 1 1 1 1 1 1 A J 20 25 30 35 40 45 50 55 AGE IN YEARS FIGURE 24.—Mean 1-hr serum-glucose level by age. challenge values of women tested after an overnight fast were much more similar to those of women who ate less than 4 hr before the test than to those of women who ate 4, 5, or 6 hr earlier. The mean 1-hr values by hours since last food and the time of day when challenged are shown in table 76. Women who ate less than 4 hr before the challenge were grouped together because their mean 1-hr values were not significantly different from one another. Women challenged 5-6 hr after eating are also grouped together because they, too, are relatively homogeneous in glucose response. As shown in figure 27, 1-hr serum-glucose levels tended to increase as the morning progressed into early afternoon; but, regardless of the time of challenge, the 1-hr values of fasting 152 SOME INITIAL RESULTS 230 r Decile 9th 210 | 8th 190 | 7th E 8 . 6th > 170 3 J 5th o - wi 2 150 | 4th Ss a = 3rd -— oc wl wv 130}. . 2nd 10 1st 90 |- 1 | | 1 1 1 fn | ] J 20 25 30 35 40 45 8) 56 AGE I'N YEARS FIGURE 25.—Deciles of 1-hr serum-glucose levels by age. women were much lower than those of women who had eaten 4 or 5 to 6 hr before the challenge. The postchallenge levels of fasting women averaged 16-20 mg per 100 ml lower than those of women who ate 4 hr before the challenge and were tested at the same time of day. The difference between fasting women and EFFECTS ON ONE-HOUR GLUCOSE TOLERANCE 153 TABLE T4.—Location of selected percentiles of distribution of 1-hr serum glucose (mg per 100 ml), by age Age in years Slope, mg/100 Percentile ml per 20- | 25- | 80- | 35- | 40- | 45- | 50- year of 24 29 34 39 44 49 54 age 2d. eo. 71 73 75 78 82 86 88 win Bh win weno 83 83 83 84 91 97 96 0.5 WN. camans 89 89 94 94 | 104 ( 106 | 109 A 20th. coven 101 | 100 | 108 | 109 | 119 | 125 | 127 1.0 25th....ceeen-. 106 | 105 | 114 | 116 | 125 | 132 | 135 1.2 BOLH. co cncvnm inna 109 | 110 | 118 | 121 | 130 | 138 | 141 1.2 AON. coon mice ans 117 | 121 | 127 | 130 | 140 | 147 | 152 1.2 BON cs ccc vm 125 | 129 | 135 | 139 | 150 | 159 | 165 1.4 60th... _____.. 182 | 141 | 145 | 150 | 162 | 169 | 176 1.5 OEY ec nine 141 | 151 | 157 | 163 | 174 | 179 | 187 1.5 OE. elon 146 | 156 | 164 | 171 | 179 | 188 | 193 1.6 80th. .-onianss 152] 162] 171 | 176 188 | 195] 199 1.6 OLR. rimmmmnan 169 | 175 | 192 | 199 | 209 | 213 | 224 1.8 Oth. minim 179 | 195] 211 | 217 | 229 ( 232 | 238 1.9 OBER... cnn am 190 {1 218 | 232] 239 250 | 255 256 |-w-uuereu- Range: 10th-90th___._| 80 86 981 105 | 105 107 {113 {-ccmmunm=e 25th-75th____| 40 51 50 55 54 56 BS |cccmrenune 20th-80th..__| 51 62 63 67 68 70 12 iss nnmmmnn those who ate 5-6 hr earlier was somewhat greater, averaging approximately 25 mg per 100 ml. Examinees challenged less than 4 hr after food, however, were not discernibly different from fasting women in glucose response regardless of the time of challenge. In all age groups, women challenged 4-6 hr after eating had higher 1-hr values than fasting women challenged at the same time of day (table 77 and fig. 28). KETONURIA The prevalence of ketonuria is shown in table 78 by the time of day when the urine specimen was collected. Testing for urine ketones was not started until 2 months after the multiphasic screening program began, so that Ketostix® test data are missing on approximately 1,000 women. Among those tested, 7.5 percent * Ketostix, Ames Co., division of Miles Laboratories, Inc.,, Elkhart, Ind. 46514. 154 SOME INITIAL RESULTS TABLE 75.—The mean and standard deviation of 1-hr serum-glucose levels (mg per 100 ml), by hours since last food when challenged Serum glucose Number Hours since last food of women Mean Standard deviation Les thon Yai oe once ch commana 58 142.8 43.8 Lor it od De a el Sem dE 85 185.5 41.8 LT a i iw 137 136.7 43.3 cme cf i cE de wee) 152 144.7 42.9 EL RL RE 1,377 161.2 41.0 ate rete el i 689 169.8 39.3 Bed Se SE eR re ra 129 168.4 42.0 Be esa ste sEew eEen 27 144.6 40.0 Fasting overnight... ccc cei vonenanan 3,991 139.3 38.6 had ketone bodies in their urine. The prevalence of ketonuria, however, was strongly related to the time of day, increasing from less than 1 percent to more than 16 percent as the morning progressed into early afternoon. This increase in prevalence with time of day appeared to accelerate during the late morning and early afternoon hours (fig. 29). Variation in the prevalence of ketonuria with the time of day has not been described before, to our knowledge. The reason for the observed variation is not immediately apparent, but it cannot be attributed to the lengthening fast of women tested later in the day who had not eaten since the previous evening. The same curvilinear increase in the prevalence of ketonuria with time of day was seen regardless of the interval since last food (table 79 and fig. 30). There was no variation in the prevalence of ketonuria asso- ciated with age in the age groups below 50 years. After age 50, however, its prevalence appeared to decrease (table 80). The mean postchallenge serum glucose levels for women with and without ketonuria by hours since last food and time of chal- lenge are shown in table 81. In all time-of-challenge categories and hours-since-last-food categories, the postchallenge values of women with ketonuria were, on the average, markedly higher than those of women with negative Ketostix tests. The observed overall mean difference was 29 mg per 100 ml. The mean differ- ence in the serum-glucose response of ketonurics and non- ketonuries, however, tended to be less marked in midday tests than in morning tests. EFFECTS ON ONE-HOUR GLUCOSE TOLERANCE 155 180 - 170 | | 160 | I 7 150 140 i SERUM GLUCOSE LEVEL (mg/100 m1) 130 | i 120 |- 0 1 2 3 4 5 6 7 8 9 FASTING OVERNIGHT HOURS SINCE LAST FOOD FIGURE 26.—Mean 1-hr serum glucose (with 95 percent confidence interval) by hours since last food when challenged. TIME OF DAY The proportion of fasting women decreased with time of day, while the proportion challenged 4-6 hr after eating increased. When uncontrolled for the interval since last food and the increas- ing prevalence of ketonuria, 1-hr glucose values showed a marked increase with time of day (fig. 31). The mean 1-hr values of women with negative Ketostix tests are plotted in figure 32 against time of challenge by the interval since last food. Although the effect from the increasing prevalence of ketonuria has been removed, the postchallenge values still tended to increase some- 156 SOME INITIAL RESULTS TABLE 76.—Th> mean and standard d-viation of 1-hr serum-glucose levels (mg per 100 ml), by time of challenge and hours since last food Serum glucose Hours since last food and time of Number challenge of women Mean Standard deviation Less than 4 hr: 9:00-10:29 am. ooo 143 127.9 38.0 10:30=11:89 BM ceive escmcnene 170 142.0 44.0 12:00-1:30. pts cn nnn nnn vn nna 119 151.8 43.7 Overall i came 432 140.0 43.0 4 hr: : 9:00-10:29 2.00. cove converses 82 154.0 46.8 10:30-11:59 8M. ence cennccnnns 651 159.9 41.2 12:00-1:80 Pie cee non nneenamnne 644 163.4 40.0 Overall ci occas snen=ee 1,377 161.2 41.1 5 or 6 hr: 0:00-10:29 aN. ccna cnnccanwnsn 22 161.1 36.0 10:30-11:59 8. cov canna sen 207 169.7 41.4 12:00-1:830 patie cece evn ww 589 169.8 39.2 OVERLAIN. ove sn cma mm 818 169.6 29.7 Fasting overnight: 0:00-10:29- 0.M. cece inns smanmen 1,922 133.9 27.8 10:30-11:59 a.m. .cuceuecnconanns 1,548 143.6 389.0 12:00-1:80. Duan, ce cic cence 521 146.3 38.2 Overalls co eee cdcnrenvsmsns 3,991 139.3 38.6 what between early morning and midday regardless of the interval since last food. The mean difference between the early morning and midday test results, adjusted for hours since last food, was 9 mg per 100 ml. DISCUSSION Reports of population surveys consistently show that glucose levels among examinees tend to increase as the age of the persons tested increases.! 2 °° The accumulated evidence suggests that the age gradient is more marked at the 1-hr sampling point than at any other, particularly when a standard but unphysiological load of glucose is used rather than a test meal. EFFECTS ON ONE-HOUR GLUCOSE TOLERANCE 157 Hours Since 170 0 mt mm» —O Last Food 7 5o0r6 a : 2 . Hours Since ’ —* Last Food 4 or pT 160 IY. a ~~ ~ - 2 . Hours Since = _e Last Food o 150 | +” Less Than 4 2 Fasting 2 Overnight o > wo140 | wl wv o E wd o = 2 130 wl [72] 120 | 1 1 1 1 J 9:00 10:30 12:00 1:00 TIME OF CHALLENGE FIGURE 27.—Mean 1-hr serum glucose by time of challenge and hours since . last food. Two population studies, the National Health Examination Survey! and the Tecumseh Community Health Study, are com- parable to the Walnut Creek data in that both were based on determinations from blood samples collected 1 hr following a standard challenge. The age gradient observed in the Walnut Creek data (1.3 mg per 100 ml per year) is the same as that reported in the Tecumseh study (13 mg per 100 ml per decade) which used a 100-g glucose challenge, but somewhat greater than that noted in the national survey (10 mg per 100 ml per decade), which was based on a 50-g challenge. The rise with age in the location of the lower deciles of the glucosemia distribution, which was demonstrated in the Walnut Creek data and also reported in the Tecumseh study, suggests that some loss of glucose tolerance with age is typical of most persons. TABLE 77.—1-hr serum-glucose levels (mg per 100 ml) by age and by time of challenge since last food: Means and standard deviations (2,669 women) Age in years Time of challenge; hours since last food (HSLF) Before 10:30 a.m.; fasting 10:30 a.m.-1:30 p.m.; fasting 10:30 a.m.-1:30 p.m.; 4-8 HSLF Stand- 180 Stand- 180 Stand- 180 Num- Mean ard mg/100 | Num- Mean ard mg/100 | Num- Mean ard mg/100 ber devia- | ml or ber devia- ml or ber devia- | ml or tion more tion more tion more 20-29. as 210 121.2 31.1 3.8 258 133.8 60.4 3.9 190 150.6 89.8 9.5 80-39... a 265 133.4 57.8 6.0 273 144.0 35.6 11.4 243 164.8 56.2 27.6 A049" J eames 260 142.9 85.8 12.7 263 154.1 39.8 20.5 306 178.4 59.2 38.6 50 and over_______.. 82 152.5 87.9 22.0 87 | 159.6 39.4 28.7 105 182.1 37.4 49.5 Total...covne- 828 134.9 44.4 9.0 884 145.5 46.4 13.7 847 168.6 65.4 30.1 841 SLTNSHY TVILINI ANOS EFFECTS ON ONE-HOUR GLUCOSE TOLERANCE 159 Challenged Between . 10:30-1:30; 4-6 Hours ign % Since Last Food 150 nce Last Foo 170 |» a Challenged Between 160 » 10:30-1:30; > 9 Hours a Since Last Food . Challenged Before » 10:30; > 9 Hours 150 | 5 Since Last Food 140 130 120 1 1 1 3 20-29 30-39 40-49 50-59 AGE FIGURE 28.—Mean 1-hr serum glucose by age, time of challenge, and hours since last food: 2,599 women, first 6 months of 1969. The heterogeneity of the decile regressions observed in the Walnut Creek data further suggests that the age-related changes are not uniform among individual persons. That is, as people age, they differ more and more from one another in their response to a glucose challenge. Whether this is because of differences among persons in the rate of aging or because of an increase in the prevalence of unrecognized disease with age, or both, is specula- tive. That some loss of glucose tolerance with age is to be expected indicates, however, that test norms for young adults may not be appropriate for older persons. A marked increase in the 1-hr values as the interval between 160 SOME INITIAL RESULTS TABLE 78.—Prevalence of ketonuria, by time of day Number Positive Ketostix test Time of day of women Nurbor Percent 10:00=10:29 800. conn cnr ems as 329 5 1.5 10:30-10:59 am. _ ooo. 664 20 3.0 11:00-11:20 am... connec smmn Vid 24 3.3 11:80-11:00 Mute enema TR 25 3.5 12:00-12:29 pec cceniccnencenmnnes 72 45 5.8 12:30-12:59 PM. cocoa 738 64 8.9 1:00=1:29 Dll. cc cman a 704 75 10.6 LB0-189 pu... mc ccs mmmm sm me 632 106 16.8 2:00-2:20 Palle cece incnnmn anna 817 53 16.7 Overall... inns 15,644 422 7.5 ! Includes 44 women for whom time of challenge and hours since last food are unknown. 25.0 - 20.0 | od wo o [ PERCENT 10.0 [~ 5.0 0.0 —1, 10:00 11:00 12:00 TIME OF DAY FIGURE 29.—Prevalence of ketonuria (with 95 percent confidence interval) by time of day. TABLE 79.— Prevalence of ketonuria (test positive) by time of day specimen collected and hours since last food Hours since last food Fasting Less than 4 hr 4 hr 5-6 hr Time of day specimen collected Test positive Test positive Test positive Test positive Number Number Number Number tested tested tested tested Number | Percent Number | Percent Number | Percent Number | Percent Morning: 10:00-10:29____. 14 0 0 4 0 0 2 0 0 309 5 1.6 10:30-10:59_____ 55 0 0 7 0 0 7 0 0 582 20 3.4 11:00-11:29_____ 54 0 0 50 1 2.0 8 0 0 614 23 3.7 11:30-11:59_____ 64 2 8.1 117 2 1.7 16 0 0 520 21 4.6 Afternoon: 12:00-12:29____. 49 2 4.1 198 12 6.1 55 4 7.3 468 27 5.8 12:30-12:59____. 44 0 0 277 24 8.7 108 16 14.8 308 24 7.8 1:00-1:29_____. 41 1 2.5 262 32 12.2 163 21 12.9 235 21 8.9 1:30-1:59______ 50 9 18.0 207 33 15.9 218 39 17.9 154 25 16.2 2:00-2:29______ 19 2 10.5 99 15 15.2 140 22 15.7 52 12 23.1 Overall _______ 390 16 4.1 1,231 119 9.7 717 102 14.2 | 3,242 178 5.5 HON VITTOL AS00NTD 4NOH-ANO NO SLOHJIIH 191 162 SOME INITIAL RESULTS PERCENT 25.0 — — —— HSLF < 4 #5 we meu ne te HSLF = 4 . » — sme SUF = § OP 6 20.0 | Fasting ~ —~—— 15.0 |- . . Fs, Fi 10.0 | SI = tL i 7 ° 2 ” / 5.0 ot 7 ow J / x ‘ / /; ral LE —— 0.0 1 oe | ] i 1 9:00 10:00 11:00 12:00 1:00 TIME OF DAY SET UP BY ONE FULL HOUR FIGURE 30.—Prevalence of ketonuria by time of day and hours since last food. TABLE 80.—Prevalence of ketonuria (positive test), by age among 5,644 women Positive test Age in years Number tested Number Percent Eo OL 473 42 8.9 Bd es id de ene REE 758 53 7.0 80-34 eres men cea EE 903 66 7.8 BBB ieee ——_ 957 77 8.0 BOA Lo etiam dae 967 4 7:7 BBA even snes Sle 895 84 9.4 BRB cena SE ERE 610 25 4.1 SANA OVOP. - vies mms mo am——— 81 1 1.2 Ot con svn sen ema Ea 5,644 422 7.5 TABLE 81.—The mean and standard deviation of 1-hr serum-glucose levels (mg per 100 ml), by result of Ketostiz test, and thz mean difference between ketonurics and nonketonurics (95 percent confidence interval), by time of challenge and hours since last food Positive Ketostix test Negative Ketostix Diffarenc2 batwazn katonuries and nonkatonurics Time of challenge and Serum glucose Serum glucose hours since last food Number Nuriber 0 o : women Mean Standard | women Mean | Standard M=an+95 pareaat deviation deviation coafidzacz interval 9:00-10:29 a.m.: Lessthand4 hr... _____. 0D] rem Sl mrss 98 130.6 8.1 J. cme see Em REL. oe cm wig sei dn ei tN ce rene 70 155.7 ABO. [re iret sim som reo; FOE BH cow mmaie ncn temas 0 Joe ep elas 17 161.4 20.3 {a rmreres mee ama ——— Fating... curcrneossnonmeemmses 48 164.4 41.4 1,457 135.4 38.3 29.0+11.9 10:30-11:59 a.m.: Lessthan4 hr_________________ 4 202.0 45.2 153 141.0 43.9 61.0+44.8 IB i ere i lis wei rial ie 38 190.8 36.2 554 159.2 40.8 31.6 +12.0 BOL CAI cunanmnnnnmnes 20 192.2 40.9 154 171.1 40.8 21.14+19.0 Posting... coveveswuwnennenwens 72 171.4 48.0 1,224 143.8 38.4 27.6+11.3 12:00-1:30 p.m.: Lessthan4 hr. _______._. 12 178.2 48.8 123 148.7 43.3 19.5+28.6 He WPL ric tasmanian 70 181.1 38.4 488 162.4 40.4 18.7+ 9.1 BOL BIL... snr emi mama 82 182.0 37.2 439 168.6 39.8 13.4+ 8.9 Pasting . emanoen sumupemesk mee 58 166.9 34.7 383 144.3 39.0 22.6 + 9.7 All examinees. - - - coccceeenn. 422 176.8 40.8 5,222 147.7 41.2 29.14 4.0 HONVIATOL HS0DNTD 4NO0H-ANO NO SLOHIIH e991 164 SOME INITIAL RESULTS 170 ® 160 |- oo £ 1 o 5 w 150 wv 3S -— 3 = & 140 | sg 130 1 1 | 1 9:00 10:00 11:00 12:00 1:00 TIME OF CHALLENGE FIGURE 31.—Mean 1-hr serum glucose by time of challenge. the last food and the challenge lengthened from 1-6 hr was also observed in the Tecumseh data. Because most of the tests were done in the afternoon or evening hours, the Tecumseh study pro- vided no data on the postchallenge response of fasting examinees. Variation in the postchallenge response by hours since last food was described in the National Health Examination Survey only for persons challenged less than 4 hr after eating. It is known that the incremental response to a glucose chal- lenge is less when the challenge is preceded by a prior one or by a recent meal.’ Because this Staub-Traugott effect is gone within 4 hr, it generally is believed that the response of a patient challenged after this interval without food will be similar to that of one who is fasting. The data in this report indicate that this assumption is not tenable. For whatever reason, the postchallenge values of women tested 4-6 hr after eating were clearly elevated above those of women who had been fasting. Because the post- challenge values of women tested less than 4 hr after eating were much more similar to those obtained from fasting women, these data indicate that it is better to screen patients postprandially EFFECTS ON ONE-HOUR GLUCOSE TOLERANCE 165 180 ~ es Hours Since 170 |- 7 T+ ~—.___ Last Food A ® 50r6 . Zz Hours Since rd Lar Last Food 4 160 —- a — aor Hours Since 150 Last Food _-* Less Than 4 SERUM GLUCOSE LEVEL (mg/100 m1) ; Fasting Overnight 140 | 130 |- o Ts 1 1 ] | 9:00 10:30 Noon 1:30 TIME OF CHALLENGE FIGURE 32.—Mean 1-hr serum glucose by time of challenge and hours since last food for women with negative Ketostix tests. rather than to instruct them to eat nothing for at least 4 hr before the examination. Aside from certain metabolic diseases, such as diabetes, ketonuria has been associated with carbohydrate deprivation.!2 Carbohydrate deprivation has long been recognized as a cause of elevated glucose-tolerance curves. To avoid misdiagnosis from carbohydrate deprivation in glucose-tolerance testing, O’Sullivan has recommended routine testing for urine ketones. Information on dietary regimen was not collected in the questionnaire of the Kaiser Foundation’s Automated Multitest Laboratory (AML), but there is no reason to believe that the diets of AML examinees tested in the early morning differed in carbohydrate content from those of women tested later in the day. The marked increase in the prevalence of ketonuria with time of day that was observed 166 SOME INITIAL RESULTS in these data raises doubt as to what a positive test implies. The meaning of the high postchallenge values that accompanied the ketonuria, therefore, is also unclear. Because the time of day covered by the data in this report was limited to the morning and midday hours, the effect of the time of day on the screening test cannot be fully assessed. In the National Health Examinations, the postchallenge levels were somewhat higher after the midday meal than after the morning or evening meals. No significant differences were observed in the Tecumseh survey between afternoon and evening tests. Although true diurnal periodicity is of great biological interest, it would seem of relatively little importance when screening for glucose intolerance. CONCLUSION One-hour glucose-tolerance tests on 6,692 women were ana- lyzed for variation in relation to age, hours since last food, ketonuria, and time of day. Known diabetics were excluded from the analysis. All subjects were challenged with 75-g glucose be- tween the hours of 9 a.m. and 1:30 p.m. Serum glucose was deter- mined by the orthotoluidine method. The urine specimen was collected immediately after the 1-hr postchallenge blood specimen was drawn, and the presence of urine ketones was evaluated with Ketostix. The age range effectively covered by the data was 20-54 years. Within this range, 1-hr serum glucose increased linearly with age at the rate of 1.3 mg per 100 ml per year. The location of all deciles of the distribution also rose linearly with age, but the rate of increase was greater for the higher than for the lower deciles. These cross-sectional observations suggest that some diminution in glucose tolerance at the 1-hr sampling point may be typical of most people as they age but that the rate of deterioration varies from person to person. The length of time between the last meal and the glucose challenge had a marked effect on the test result. Among women who had eaten on the morning of the examination, the mean 1-hr value rose more than 30 mg per 100 ml as the interval since the last food lengthened from 1 to 6 hr. The 1-hr value of examinees challenged after an overnight fast, however, were much more similar to those of women challenged less than 4 hr after eating than to those obtained from women whose last food had been 4-6 hr earlier, regardless of the time of challenge. Thus, these EFFECTS ON ONE-HOUR GLUCOSE TOLERANCE 167 data indicate that a screening test 4-6 hr after a meal is not comparable to one administered in a fasting state. The prevalence of ketonuria increased curvilinearly from less than 1 percent to more than 16 percent as the morning progressed into early afternoon. The increased prevalence of ketonuria with the time of day was independent of the interval since last food. Regardless of the time of challenge and the interval since the last food, the 1-hr postchallenge serum-glucose values of women with ketonuria were considerably higher than those of women with negative Ketostix tests. The overall average difference was 29 mg per 100 ml. When the effect of hours since last food and of ketonuria were removed, the midday test results tended to be higher than those from the early morning. The difference, however, was small, averaging 9 mg per 100 ml. REFERENCES 1. National Center for Health Statistics: Glucose tolerance of adults: United States, 1960-1962. Vital Health Stat PHS Pub. No. 1000, Series 11, No. 2. Washington: U.S. Government Printing Office, May 1964. 2. HAYNER, N. S., KJELSBERG, M. O., EPSTEIN, F. H., and FrANCIS, T., JR.: Carbohydrate tolerance and diabetes in a total community, Tecumseh, Michigan: 1. Effects of age, sex, and test conditions on one-hour glu- cose tolerance in adults. Diabetes 14: 413-423, July 1965. 8. National Center for Health Statistics: The one-hour glucose tolerance test. Vital Health Stat PHS Pub. No. 1000, Series 2, No. 3. Washing- ton: U.S. Government Printing Office, July 1963. 4. O’SULLIVAN, J. B., GRANT, M. E., and FRANCIS, J. O’S.: Glucose tolerance test standardization simplified by urinary ketone testing. Public Health Rep 80: 220-224, Mar. 1965. 6. JORDE, R.: The Diabetes Survey in Bergen, Norway, 1956. Bergen-Oslo: Norwegian Universities Press, Apr. 1962. 6. NILssoN, S. E., LinpHoLM, H., BuLow, S., FROSTBERG, N., EMILSSON, T., and STENKULA, G.: The Kristianstad survey 1963-1964—studies in a normal adult population for variation and correlation in some clinical, anthropometric, and laboratory values, especially the peroral glucose tolerance test. Acta Med Scand (supp. 428) 177: 1-42, 1964. 7. BoyNs, D. R., CROSSLEY, J. N., ApaMs, M. E., JARRETT, R. J., and KEEN, H.: Oral glucose tolerance and related factors in a normal population sample: I. Blood sugar, plasma insulin, glyceride, and cholesterol measurements and the effects of age and sex. Br Med J 1: 595-598, Mar. 8, 1969. 8. KAUFMAN, B. J., GRANT, D. R., and MOORHOUSE, J. A.: An analysis of blood glucose values in a population screened for diabetes mellitus. Can Med Assoc J 100: 692-698, Apr. 1969. 168 10. 11. 12. SOME INITIAL RESULTS LAUVAUX, J. P., and STAQUET, M.: Oral glucose tolerance test: Study of influence of age on response to standard oral 50-gram glucose load. Diabetologia 6: 414-419, Aug. 1970. HAMMON, L., and HIRSCHMAN, I. I.: Studies of blood sugar: IV. Effects upon the blood sugar of the repeated ingestion of glucose. Bull Johns Hopkins Hosp 30: 306-308, Oct. 1919. METZ, R., and FRIEDENBERG, R.: Effects of repetitive glucose loads on plasma concentrations of glucose, insulin, and free fatty acids: para- doxical insulin responses in subjects with mild glucose intolerance. J Clin Endocrinol Metab 30: 602-608, May 1970. Azar, G. J., and BLooM, W. L.: Similarities of carbohydrate deficiency and fasting. Arch Intern Med 112: 338-343, Sept. 1963. Chapter 4 One-Hour Glucose Tolerance in Relation to the Use of Oral Contraceptive Drugs NaNcy R. PHILLIPS AND THOMAS J. DUFFY From December 1968 through December 1969, 4,815 nonpregnant women of reproductive age were tested for glucose tolerance 1 hr after glucose ingestion. Study subjects.—Thirty-seven percent of the women were current oral contraceptive (OC) users, 31 percent past OC users. Oral contra- ceptive use was inversely related to age. Ninety percent took a combina- tion-type OC; nearly half took a combination of mestranol and norethin- drone. Methods.—Testing procedures are described in Part B, Chapter 2, “Sensitivity and Specificity of the One-hour Glucose-Tolerance Screening Test.” OC-use status.—One-hr serum-glucose levels were found to be higher among current OC users than among OC nonusers. The mean difference between OC users and nonusers of like age was 11 mg/100 ml. Drug components and steroid dose.—There was no observable varia- tion in the glucose response of women taking OC’s associated with the formulation and steroid components of the drug used. Duration of exposure—No correlation was found between 1-hr serum-glucose value and duration of exposure to oral contraceptive drugs. Diabetic risk factors.—The effect of the oral contraceptive drugs on 1-hr glucose tolerance was found to be additive to the effects of age, obesity, and a family history of diabetes. A number of studies have demonstrated that glucose toler- ance diminishes with the use of oral contraceptive drugs (OC) .-* Some observers have suggested that the decrease is greater in women with diabetic risk factors, such as obesity or a family history of diabetes.® 7¢ Evidence has also been presented that the glycometabolic effect of this group of compounds depends on the chemical structure of the component steroids ;’ 2 but interprepara- * Reproduced (with slight modifications) from Am J Obstet Gynecol 116(1) : 91-100, May 1, 1973; Copyrighted by The C. V. Mosby Co., St. Louis, Mo. 169 170 SOME INITIAL RESULTS tion evaluation has been handicapped by the relatively small number of subjects in any one study, and by the variation among studies with respect to methods and case material. Automated Multitest Laboratory (AML) testing was begun in December 1968 at the Walnut Creek facility of the Kaiser- Permanente health care program. From this time through De- cember 1969, 4,815 nonpregnant women of reproductive age, some of whom were users and some nonusers of OC’s, were tested for glucose tolerance 1 hr after glucose ingestion. Presented here are the results of cross-sectional analysis of the findings in relation to use or nonuse of OC’s at the time of testing; to the steroid components and dose of the OC used; to the duration of use, and to two diabetic risk factors—obesity and a family history of diabetes. MATERIAL AND METHODS STUDY SUBJECTS None of the women on whose findings this report is based was known to be diabetic at the time of testing. All took and retained the full challenge dose of glucose. The age and status with respect to OC use of the 4,815 study subjects are shown in table 82. At the time of glucose-tolerance testing, 37 percent of the women were current OC users and 31 percent were past TABLE 82.—Age and oral contraceptive drug-use status of the study subjects Current users Past users Never users Total Age in years 3 Num- | Per- | Num- | Per- | Num- | Per- | Num- | Per- ber | cent ber cent ber cent ber cent 20-24 __..... 303 | 53.6 171 | 30.3 91: 16.1 565 | 100.0 25-29. ......- 449 | 48.8 369 | 40.2 101: 11.0 919 | 100.0 80-34........ 412 | 389.5 352 | 33.8 278 | 26.7 | 1,042 | 100.0 85-39: . -....- 321 | 28.3 328 | 29.0 484 | 42.7 11,133 | 100.0 40-44... __ 287 | 24.8 287 | 24.8 582 | 50.4 | 1,156 ( 100.0 Overall.| 1,772 | 36.7 | 1,507 | 31.3 | 1,536 | 31.9 | 4,815 | 100.0 Mean age..... 31.5 32.6 36.5 33.5 Standard deviation... 6.7 6.5 6.0 6.9 GLUCOSE TOLERANCE AND ORAL CONTRACEPTIVES 171 users. Use of OC’s was inversely related to age; current users were, on the average, 5 years younger than never users and 1 year younger than past users. Since 1-h: glucose levels rise with age, all comparisons shown here between use-status classes are either age specific or otherwise age controlled. The oral contraceptive agents currently taken by these women varied with regard to formulation, component steroids, and dos- age; but 90 percent of current users were taking a combination- type drug; and nearly half took a combination of mestranol and norethindrone. As described in part B, chapter 3, both the interval since last food and the time of day at which the challenge was given affected the 1-hr serum-glucose value. The distribution of times of challenge and intervals since last food within use-status class was the same for all three groups (table 83), as was the variation in test results associated with these two temporal factors. Thus, variation in 1-hr glucose tolerance stemming from variation in test conditions was independent of OC-use status. The cross- sectional analysis of 1-hr glucose tolerance in relation to OC use was therefore carried out on the pooled results of these casual tests. TABLE 83.—T'ime of glucose challenge and interval since last food, by oral con- traceptive drug-use status Time of challenge and interval Current Past Never since last food users users users 9:00-10:29 a.m.: Percent Percent Percent Lessthan a Br... coviconsiasoe 1.9 2.8 2.5 GB BY it renee S RRR 1.3 1.8 1.6 9hrormore. . oon 29.8 28.0 29.0 10:30-11:59 a.m.: Tess than 4 By.....c.cccesmcmecnees 2.6 3.2 2.8 d=B HY. ie cet Joo simian aah wd sia 10.8 12.5 12.8 O IY OF MOTE... ccvinnmuaccanmns 22.5 24.9 24.3 12:00 a.m.-1:30 p.m.: Legs thand hr. oc ee 2.4 1.9 1-3 CB AE oto edn ma 20.0 16.3 17.0 0 AT Or NOTE. «ems was ems we hms 8.49 9.1 8.3 Total. camnss snsencnmenninesnes 100.0 100.0 100.0 NoTe.—x?=21.2 with 16 df; p> 0.10. 172 SOME INITIAL RESULTS 160 i rk /o 150 |— v’ / — 8 7 z of g of = 140 ri g Wala 1 fee 8 7) B ° > ad ° / a 130 | 7 / L 5 SL oS Current Users wi / 2 » ~- ——— Past Users «7 Lc ER ty eee Never Users 120 5 T 1 ] 1 ] ] 20 25 30 35 40 45 AGE FIGURE 33.—Mean 1-hr serum glucose by age and oral contraceptive drug- use status. METHODS The testing procedures for glucose tolerance are described in part B, chapter 2, “Sensitivity and Specificity of the 1l-hour Glucose-Tolerance Screening Test.” RESULTS ORAL CONTRACEPTIVE USE STATUS The 1-hr serum-glucose value tended to be higher in current OC users than in past or never users (fig. 33 and table 84). Since past users and never users did not differ discernibly with respect to their postchallenge values, these two groups were combined into a nonuser group. The regression of 1-hr serum-glucose levels with age in current users was 1.4 mg per 100 ml per year, parallel to that TABLE 84.—Serum-glucose levels by age and oral contraceptive (OC) drug-use status: Mean and standard deviation (SD) Serum glucose (mg/100 ml) Difference between current users Age in years Current users Past users Never users Total nonusers and nonusers (Mean he percent 1 Mean SD Mean SD Mean SD Mean SD Under 25_________ 132.0 29.3 122.7 32.9 119.6 30.8 121.6 32.2 10.4+4.8 25-29. cwerrmemesin 138.4 32.6 126.5 34.9 131.2 42.0 127.5 36.5 10.9+4.5 30-34____________ 146.6 35.2 136.1 38.8 136.4 40.1 136.3 39.4 10.3+4.6 35-39. ________ 154.8 39.5 137.5 36.8 142.4 43.0 140.4 40.7 14.445.1 40-44 ___________ 160.3 36.1 153.5 42.7 151.9 43.8 152.4 -43.4 7.945.2 1 95 percent confidence interval. SHAILJIOVALNOD TVH0 ANV HONVIITOL IS0DNTD 6LT 174 SOME INITIAL RESULTS in nonusers (1.5 mg per 100 ml) ; thus, the magnitude of the difference between the mean 1-hr values of users and nonusers did not increase with age. The relative frequency distribution of the postchallenge values among the 1,772 current users is compared in figure 34 with that among an age-stratified probability sample of 1,747 nonusers having the same age distribution. The means of the two glucosemia distributions differed by approximately 11 mg per 100 ml. It will be noted, however, that the difference in the location of the two distributions was more marked in the lower than in the upper percentiles (table 85). The 25th percentile, for example, was at 122 mg per 100 ml for users and at 105 mg per 100 ml for non- users, a difference of 17 mg per 100 ml. At the 75th percentile, the difference was 10 mg per 100 ml; at the 90th it was 4 mg per 100 ml, and at the 95th percentile the two values did not differ. This relationship is shown in the cumulative frequency distri- butions (fig. 35). Thus, while users of contraceptive drugs tended to have higher 1-hr values than did nonusers, the users did not show a preponderance of unusually high values. DRUG COMPONENTS AND STEROID DOSE Mean 1-hr serum glucose is shown in table 86 and figure 36 by type of drug formulation and component steroids. Although there was little age variation among the users of different prepara- tions, the values shown in table 86 are age adjusted to the overall age distribution of current users. The mean 1-hr value for users of all preparations represented in this sample was higher than the age-adjusted nonuser mean, 135 mg per 100 ml. Except for those preparations used by few women, the lower 95-percent confidence limits on the mean values associated with each preparation were well above the nonuser level (table 86). Of the six combination formulations, no specific estrogen-progestogen combination produced a mean 1-hr value significantly different from any other. The progestogen doses for ethynodiol diacetate, norgestrel, and medroxyprogesterone were invariate, as were the doses of the estrogens used with them. Of the remaining combination formulations, mestranol with norethindrone was the only one of which the number of users was sufficiently large to permit assessment of variation by total steroid dose. The age-adjusted mean l-hr value for women receiving 0.05 mg mestranol and GLUCOSE TOLERANCE AND ORAL CONTRACEPTIVES RELATIVE FREQUENCY +25 .20 .15 .05 ~ I T ¥: 130 170 210 SERUM GLUCOSE (mg/100 m1) 175 Current Users — Nonusers 250 FIGURE 34.—Relative frequency distributions of 1-hr serum-glucose values among oral contraceptive users and nonusers of comparable age. TABLE 85.—Distribution of 1-hr serum-glucose levels among users of oral con- traceptive (OC) drugs and among nonusers of comparable age: Means, standard deviations, and percentiles Serum glucose (mg/100 ml) | Difference between OC Location and dispersion of distribution users and OC users Nonusers nonusers (N=1,772) | (N=1,747) | (mg/100 ml) Mean... cco curinnvine sen enms mn sms 145.7 135.0 10.7 Standard deviation.....ccccevsnonmanas 35.9 39.8 -3.9 Percentiles: BH LL coeds mm erate Rie 91 82 9 JON el verve RS TE Rae 101 89 12 BBN. cress seer rar TR Em 122 105 i BOL... iin atin adm 143 129 14 TBI... ois otesmriimraimratoes mi mirenonconm 169 159 10 OOH. «wi sires rem an sa 194 190 4 1 NE SL 209 209 0 176 SOME INITIAL RESULTS CUMULATIVE PERCENT 100 r —— seme a 2 80 |~ 60 pe 7 ’ / / / J / 40 Jd —— Current Users / # -———- Nonusers / 7 / / . 20 | rs /f ’ ’ » / cai lesssoizt— 1 1 1 1 1 1 40 80 120 160 200 240 280 320 SERUM GLUCOSE (mg/100 m1) FIGURE 35.—Cumulative frequency distribution of 1-hr serum glucose values for current oral contraceptive users and nonusers of comparable age. 1.0 mg norethindrone per tablet, however, was only 1 mg per 100 ml lower than that of women receiving twice the dose of these two steroids (z = 0.4; NS). No progestogen used in combination with mestranol was also used with ethinyl estradiol. The mean 1-hr value for users of all combination formulations employing mestranol, however, was the same as that for combination users receiving ethinyl estradiol (table 86). All combination formulations employing ethinyl estradiol contained 0.05 mg estrogen per tablet. The estrogen content of combination formulations employing mestranol ranged from 0.05 to 0.15 mg per tablet. Within this range, no relation was observed between mestranol dose and 1-hr serum glucose (table 87). Users of sequential formulations received estrogen alone for 14 days of the cycle and progestogen with estrogen for only 6 days. As shown in table 86, the mean 1-hr value for users of sequential formulations was not discernibly different from that of women who took combination formulations and who received both steroids throughout most of the cycle. It is noteworthy that GLUCOSE TOLERANCE AND ORAL CONTRACEPTIVES 177 170 Combination - Mestranol — — Combination - Ethinyl Estradiol ——— Sequential - Mestranol 160 ----—- Nonusers = 150 | E o o < o 2 @ 140 | o g a rH = oc wi “130 |= 120 |- i 1 1 1 1 1 J 20 25 30 35 40 45 AGE FIGURE 36.—Mean 1-hr serum glucose by age for current oral contraceptive users by type of formulation and for nonusers. among users of sequential-type drugs, those tested on days 5-18 of the menstrual cycle (while on estrogen alone) and those tested on days 19-24 (while taking both steroids) had comparable mean 1-hr values (145 mg per 100 ml and 142 mg per 100 ml, respec- tively (z = 0.5; NS). When 1-hr glucose values were analyzed by number of days since onset of last menstrual period, again OC users had con- sistently higher values than had nonusers. Because of the small numbers of subjects representing each day since onset of last period and the difficulty of timing ovulation, the results were highly unstable and no conclusions can be drawn from them (fig. 87). TABLE 86.—Mean 1-hr serum glucose (mg/100 ml) by type of oral contraceptive (OC) drug—Age adjusted to the overall age distribution of current OC users Overall mean Overall mean Type of formulation, estrogen, and progestogen ! Number of Serum glucose serum glucose serum glucose women for type of for type of estrogen formulation Mean+95 mg/100 ml mg/100 ml Combination formulations: percent CI? Mestranol: NOTOTMNALONG. . «wow mim me mm www 879 | 146.44 2.4 MNOTCERYNOAEOL... co er mim irom on som om om ti sp om sm on 176 146.04 5.1 145.9 Ethynodiol diacetate __.__________________________ 301 143.84 3.9 Ethinyl estradiol: 145.9 Norethindrone acetate... _______________________ 166 | 145.7% 5.7 Norgestrel ._ __ ee 43 148.64 9.6 145.7 Medroxyprogesterone._ _ __________________________ 20 138.8+11.6 Sequential formulations: Mestranol: NOTOEIINAIONG ... ci eit mmm tim melee i dene ei 28 | 142.8+13.3 144.5 Chlormadinone acetate .-_______________________.. 117 | 145.0+ 6.5 ’ 144.3 Ethinyl estradiol: Dimethisterone __________________________ 37 | 143.5+11.6 143.5 1 The type of drug taken by 5 current users of OC was unknown. 32 95 percent confidence interval. 8LT SLTNSHY TVILINI HN OS GLUCOSE TOLERANCE AND ORAL CONTRACEPTIVES 179 TABLE 87.—Mean 1-hr serum glucose by estrogen dose for women taking com- bination-type drugs employing mestranol Number of | Mean serum Mestranol (mg/tablet) women glucose (mg/100 ml) QS rs ee cmt Ev = 291 145.1 OBOIO8 rrr cers eEE ETE TEER SEE 70 148 .4 rr mame nr Een TR rE TE EN EES NC Re 852 145.3 TD ev i i ee i SR 5 131.8 UINBAOWIY . «tween smn sommes md Sms mean 138 147.2 Current Users 160 —_———— Nonusers — ow o I 140 + SERUM GLUCOSE (mg/100 m1) 130 = 120 lt tt tet tg 6 8 10 12 14 16 18 20 22 24 26 28 NUMBER OF DAYS SINCE ONSET OF LAST MENSTRUAL PERIOD FIGURE 37.—Mean 1-hr serum glucose by number of days since onset of last menses for current oral contraceptive users taking combination formula- tion, and for nonusers of comparable age. 180 SOME INITIAL RESULTS 165 — 155 | ey [1 — fpemmerettmmenne) ee) ei i UE fms mo lc i i te) ee 8 a = | | | | | 1 1 1 1 | 0 12 24 36 48 60 > 60 SERUM GLUCOSE (mg/100 m1) CONTINUOUS MONTHS OF OC USE FIGURE 38.—Mean 1-hr serum glucose (with 95 percent confidence level) of current oral contraceptive (OC) users by continuous months of use. The solid horizontal line represents the overall mean value of current users; the broken line is the mean value of never users, age adjusted to the age distribution of current users. 155 145 — 7 sd w a ¥ | | | | | SERUM GLUCOSE (mg/100 m1) 125 0 12 24 36 48 60 > 60 MONTHS SINCE LAST OC DOSE FIGURE 39.—Mean 1-hr serum glucose (with 95 percent confidence level) of past oral contraceptive (OC) users by months since last dose, age adjusted to the age distribution of current users. The solid horizontal line represents the overall mean value of current users; the broken line is the mean value of never users, age adjusted to the age distribution of current users. GLUCOSE TOLERANCE AND ORAL CONTRACEPTIVES 181 DURATION OF EXPOSURE No correlation was found between 1-hr serum-glucose value and duration of exposure to oral contraceptive drugs, whether duration was measured as total months of OC use or as length of current period of continuous use. The mean 1-hr values of current users, age adjusted to their overall age distribution, are shown in figure 38 by continuous months of OC use. The number of women who had recently started using oral contraception was too small to permit discernment of the precise point in time of use at which the effect of these steroids on glucose tolerance is manifested; but the mean 1-hr value for women tested during the first three cycles of use was well above the level for never users. The duration of exposure to oral contraceptives was shorter for past users, as a group, than for current users; but, again, no variation in 1-hr serum glucose associated with duration of use was found among past users. The mean 1-hr serum-glucose values for past users, age- adjusted to the age distribution of current users, are shown in figure 39 by months since last dose. In view of the width of the confidence intervals, no judgment can be made as to the length of time required for glucose response to return to a never-user level after discontinuation of oral contraception. The 1-hr values for women tested 4-6 months after discontinuation were, how- TABLE 88.—1-hr serum glucose (mg/100 ml) by family history of diabetes for users of oral contraceptive (OC) drugs and for nonusers of comparable age; Mean and standard deviation (SD) Family history of No family history of diabetes diabetes Difference, OC-use status mean + Numb or Serum glucose Nuhber Serum glucose SE ! o of women | pean SD |WOmen| pean SD User. ___._.._..__ 576 | 149.3 | 85.3 | 1,196 | 144.0 | 36.6 | 5.3+1.8 Nonuser....-c.--- 560 | 137.8 | 42.6 | 1,187 | 133.6 | 38.4 | 4.2+2.1 Mean difference +SE_____ |... 11.5+2.3 |....... 10.4+1.5 1 Standard error. ’ 182 SOME INITIAL RESULTS 170 160 + {Fanity Family 4 History Je —~ 150 +7 4% Negative = 7 /R{Famiy = A Pi History o 4 / 5 ~ 2 E / w 140 7 8 Positive 5 = Ss Family } a History 2 Negative ~ oa 2 10 Family ¥2 raat J History ’ Id / /Z Le Current Users »-7 ---—— Nonusers 120 | o 7 1 1 1 1 1 ER | 20 25 30 35 40 45 AGE FIGURE 40.—Mean 1-hr serum glucose by age and by family history of dia- betes, and oral contraceptive drug-use status. ever, clearly below the OC-user level. It will be noted that the 1-hr values appear to decrease progressively as the interval since last dose lengthens to 5 years. This trend is probably accidental, since it is abruptly reversed when the interval since discontinua- tion exceeds 5 years. Furthermore, the never-user level is covered by the 95-percent confidence intervals on all month-since-last-dose groupings. DIABETIC RISK FACTORS Approximately one-third of study subjects reported diabetes in one or more blood relatives. This proportion was independent of age and OC-use status. The precise relationship of each affected relative to the subjects is not known. But, as a group, women with a family history of diabetes tended to have somewhat higher 1-hr glucose values than those without such a family history (4 mg GLUCOSE TOLERANCE AND ORAL CONTRACEPTIVES 183 TABLE 89.—1-hr serum glucose (mg/100 ml) by Ponderal Index for users of oral contraceptive (OC) drugs and for nonusers of comparable age: Mean and standard deviation (SD) Ponderal Index less than Ponderal Index 11.75 or 11.75 more OC-use status Nuanber Serum glucose Wmiber Serum glucose of of women | wean | SD | VO™eR | Mean | SD HISEE. cosas eames 119 150.6 33.1 1,611 145.5 36.1 Nonuser_______________ 162 | 142.2 46.0 | 1,549 | 133.9 39.0 Mean difference SE |e. 8.4447 |... 11.6+1.3 1 Standard error. 170 160 (— ~ WJ PI<11.75 Es / / ® S 10 7 Seran.zs 5 rf £ i 7 Y 2 ry SQ 7 = / 3 140 | » = i “© P1<11.75 5 ob PIznIs- 7 ’ 7 Current Users ” pe Ca — ——~— Nonusers 120 }- . 7 0 1 | | 1 | 20 25 30 35 40 45 AGE FIGURE 41.—Mean 1-hr serum glucose by age, Ponderal Index (PI), and oral contraceptive drug-user status. 184 SOME INITIAL RESULTS per 100 ml higher, on the average). The magnitude of this difference was the same for OC users as for nonusers (table 88). Thus, the difference between the 1-hr serum-glucose values of OC users and nonusers was no greater among subjects with a positive than among those with a negative family history of dia- betes (fig. 40). In all age groups, the 1-hr serum-glucose values for obese women—defined as those with a Ponderal Index (PI) of less than 11.75—were approximately 4 mg per 100 ml higher than for nonobese women of the same age, regardless of OC-use status (table 89). Obesity was somewhat less frequent among OC users (6.9 percent) than among nonusers (9.5 percent) of comparable age. Among obese women, however, the difference between the 1-hr serum-glucose levels of OC users and nonusers was of the same magnitude as that observed between users and nonusers having a higher PI (fig. 41). DISCUSSION In studies of women given the standard oral glucose-tolerance test before and after the start of contraceptive drug use, the mean increase in serum-glucose level observed at the 1-hr sampling point has varied from 7 to 15 mg per 100 ml.2° 1! The mean difference between the 1-hr serum-glucose levels of OC users and nonusers found in this cross-sectional body of data—11 mg per 100 ml—is consistent with the findings of those longitudinal studies. The hyperglycemic effect of the contraceptive drugs is ap- parently manifested soon after their use is begun. Of critical importance is the question whether there is progressive loss of tolerance with their continued use. Some investigators believe that the frequency of abnormal glucose tolerance may increase with the duration of OC use;¢ ® others have failed to find evidence of progressive deterioration.®’* One followup study has led to the suggestion that the tolerance initially lost may be regained during the course of OC use, through some type of adaptive mechanism.’ Although the hyperglycemic effect may be more marked in the initial than in later cycles of use, the cross-sectional data presented here indicate that, when adjustment is made for normal changes with the subjects’ age, the elevation of the 1-hr serum-glucose level that accompanies the use of oral contraceptive drugs remains constant over time. Improved glucose tolerance following discontinuation of con- traceptive drug use has been demonstrated in longitudinal studies.® GLUCOSE TOLERANCE AND ORAL CONTRACEPTIVES 185 Doubt has been expressed that the restoration will prove to be permanent. Our cross-sectional data suggest that the effect of these drugs on the glucose tolerance of most patients is completely and permanently reversible. On the basis of studies in rhesus monkeys, Beck postulated that contraceptive drugs containing mestranol would be more diabetogenic than those containing ethinyl estradiol, and that mestranol plus norethindrone or norethynodrel might cause more deterioration in glucose tolerance than mestranol alone or in combination with ethynodiol diacetate.’? Although the studies of di Paola and his associates show an increase in the frequency of abnormal responses to prednisone glucose-tolerance tests among subjects treated with mestranol but not among those treated with ethinyl estradiol,” Wynn and Doar found no variation among the hyperglycemic effects of various preparations.®® In view of the size of our study group and the variety of preparations represented, it is especially significant that our data revealed no variation in the glycometabolic effect of contraceptive drugs associated with the chemical structure of their component steroids. It has been suggested that only individuals with compromised pancreatic-islet insulin reserve might show a decrease in glucose tolerance with OC use.'? The upward shift in the lower tail of the glucosemia distribution among OC users observed in our body of data indicates that most women, rather than just a sub- population, have decreased tolerance while taking contraceptive drugs. The longitudinal studies of Wynn and Doar have also shown that the area under the glucose-tolerance curve increases in most women after starting contraceptive drug use.? The magni- tude of the change also appears to be unrelated to the presence of diabetic risk factors. The effect of the contraceptive drugs on 1-hr glucose tolerance was found in our analysis to be merely additive to the effects of age, obesity, or a family history of diabetes. Studies from which greater deterioration in glucose tolerance among women with diabetic risk factors was reported were not controlled for the initially lower glucose tolerance or for the more marked decline of tolerance over time which is to be expected in such women whether or not they are taking oral con- traceptive drugs.27* That the upward shift in the glucosemia distribution of OC users was more marked in the lower than in the upper percentiles is of interest. The presence of the shift suggests that women with the poorest tolerance show the least decrease in tolerance with contraceptive drug use; or, alternatively, that oral contraceptives 186 SOME INITIAL RESULTS may not have been prescribed for women with poor tolerance, or may have been discontinued if the postchallenge values were suh- sequently found to be high. Neither users nor women requesting oral contraceptive drugs, however, were routinely screened fou glucose intolerance before the start of this study. Wynn and Doar also found that the larger the area under the oral glucose-tolerance curve before OC use, the less it increases with OC use.’ Recent work by Spellacy and his associates shows similar results.'* Women who scored zero Wilkerson points!® on the oral glucose-tolerance test before OC use showed a more marked elevation in their postchallenge values after 6 months of OC than did women who initially scored 0.5 to 1.5 points. Al- though the findings in these two longitudinal studies may also be partly artifactual, in that on repeat measurement, extreme values tend to be less extreme, they, too, suggest that the effect of oral contraceptive drugs on glucose tolerance is less marked in women with low tolerance than in those with normal tolerance. The mechanism by which oral contraceptive drugs alter response to a glucose challenge is unknown. But if these agents were diabetogenic, one would expect their effect to be most marked in women who are more likely than others to have subclinical diabetes. That the reverse appears to be true suggests that a sharply limited physiologic rate constant is altered during OC use. Small changes in the saturable membrane carrier systems of red cells—muscle or fat, for example—would have a greater effect on the postchallenge values of women with normally rapid up- take of glucose than on those with slower uptake. Although oral contraceptives may not induce or aggravate beta-cell exhaustion, the slight elevation in blood sugar that accompanies OC use may in itself have consequences on health. While 11 mg per 100 ml is not a clinically significant increase in glucose concentration, it implies that the 1-hr glucose tolerance of women taking oral contraceptives has been lowered, on the average, to that of women 7 to 8 years older. Whether this alteration might accelerate the development of atherosclerosis or other degenerative processes is a question that needs to be investigated. REFERENCES 1. SperLracy, W. N,, and CARLSON, K. L.: Plasma insulin and blood glucose levels in patients taking oral contraceptives: A preliminary report of a prospective study. Am J Obstet Gynecol 95: 474-478, June 15, 1966. 2. BUCHLER, D., and WARREN, J. C.: Effects of estrogen on glucose toler- ance. Am J Obstet Gynecol 95: 479-483, June 15, 1966. 10. 11, 12. 13. 14. 15. GLUCOSE TOLERANCE AND ORAL CONTRACEPTIVES 187 WynNN, V., and DoAR, J. W. H.: Some effects of oral contraceptives on carbohydrate metabolism. Lancet 2: 7156-719, Oct. 1, 1966. POSNER, N. A., SILVERSTONE, F. A., POMERANCE, W., and BAuMGoOLD, D.: Oral contraceptives and intravenous glucose tolerance. 1. Data noted early in treatment. Obstet Gynecol 29: 79-86, Jan. 1967. POSNER, N. A., SILVERSTONE, F. A., POMERANCE, W., and SINGER, N.: Oral contraceptives and intravenous glucose’ tolerance. II. Long-term effect. Obstet Gynecol 29: 87-92, Jan. 1967. JAVIER, Z., GERSHBERG, H., and HULSE, M.: Ovulatory suppressants, estrogens, and carbohydrate metabolism. Metabolism 17: 443-456, May 1968. pI PAora, G., PucHULU, F., RoBIN, M., NICHOLSON, R., and MARTI, M.: Oral contraceptives and carbohydrate metabolism. Am J Obstet Gynecol 101: 206-216, May 15, 1968. SPELLACY, W. N.: The effect of ovarian steroids on glucose, insulin, and growth hormone. In H. A. Salhanick, D. M. Kipnis, and R. L. Vande Wiele, Eds.: Metabolic Effects of Gonadal Hormones and Contraceptive Steroids. New York: Plenum Press, 1969. Pp. 126-143. WYNN, V. and DoAR, J. W. H.: Longitudinal studies of the effects of oral contraceptive therapy on plasma glucose, non-esterified fatty acid, insulin and blood pyruvate levels during oral and intravenous glucose tolerance tests. In H. A. Salhanick, D. M. Kipnis, and R. L. Vande Wiele, Eds.: Metabolic Effects of Gonadal Hormones and Contraceptive Steroids. New York: Plenum Press, 1969. Pp. 157-177. pI PaoLaA, G., ROBIN, M., and NICHOLSON, R.: Estrogen therapy and glu- cose tolerance test. Am J Obstet Gynecol 107: 124-132, May 1, 1970. SpeLLAcY, W. N.,, Bui, W. C,, BIRK, S. A., and McCREARY, S. A.: Studies of ethynodiol diacetate and mestranol on blood glucose and plasma insulin. Int J Fertil 16: 556-65, Apr.—June 1971. Beck, P.: Effects of gonadal hormones and contraceptive steroids on glucose and insulin metabolism. In H. A. Salhanick, D. M. Kipnis, and R. L. Vande Wiele, Eds.: Metabolic Effects of Gonadal Hormones and Contraceptive Steroids. New York: Plenum Press, 1969. Pp. 97-125. PI-SUNYER, F. X., and OSTER, S.: Effect of an ovulatory suppressant on glucose tolerance and insulin secretion. Obstet Gynec 31: 482-484, Apr. 1968. FUERTES-DE LA HABA, A., VEGA-DE RODRIGUEZ, G., and PELEGRINA, I.: Carbohydrate metabolism in long-term oral contraceptive users. Obstet Gynec 37: 220-224, Feb. 1971. REMEIN, Q. R., and WILKERSON, H. L. C.: The efficiency of screening tests for diabetes. J Chronic Dis 13: 6-21, Jan. 1961. 5 wf Bi it V ie pa Ci iE Chapter 5 Spirometry and Oral Contraceptives' SHANNA H. FREEDMAN AND NEIL E. ANDERSON Introduction.—Past research on the effect of oral contraceptives (OC’s) on lung functioning has reached various conclusions. Oral contra- ceptives have been postulated to have an adverse effect, no effect, and a beneficial effect. However, there have been no reports of abnormal spiro- metric values ascribed to oral contraceptives. Materials and methods.—Spirometry data were obtained from 2,066 current, past, and never OC users, women on other estrogens, and preg- nant women having an Automated Multitest Laboratory examination in 1969. Measurements of Peak Flow, Forced Vital Capacity (FVC), Forced Expiratory Volume at 0.5 sec (FEV 0.5), and 1 sec (FEV 1.0) were obtained on a wedge spirometer. Analysis of data.—Spirometric values appear to be unaffected by past or present OC use. In addition, no differences were found between spirometric data for women in early pregnancy and those for the total population under consideration. These conclusions were arrived at after consideration of age, height, weight, and smoking habits. Comment.—Because of the size of the population analyzed it seems safe to assume that if an adverse reaction is present, the prevalence is extremely low in the age group under consideration. INTRODUCTION This report describes the relationship of oral contraceptive (OC) use to spirometric values. An association between oral contraceptives and pulmonary emboli has been reported,’ ? al- though these findings have been challenged.? Conversely, oral contraceptives might have a protective effect on the lung paren- chyma. A deficiency of antitrypsin has been suggested as a con- tributory factor to chronic bronchitis and emphysema.* There is evidence that alpha-l-globulin® and antitrypsin levels® are * Reproduced (with slight modification) from Am J Obstet Gynecol 116(5) : 682-688, July 1, 1973; copyrighted by The C. V. Mosby Co., St. Louis, 189 190 SOME INITIAL RESULTS higher among OC users. This may provide protection against chronic obstructive lung disease, but demonstration of this effect would require observations made on a large group of women over a longer period of their lifespan. Alterations in the lung parenchyma of animals, attributable to ethinyl estradiol, have produced progressive pulmonary lesions and finally resulted in diffuse interstitial pneumonitis.” Were this to occur in humans it might affect spirometry results in turn. Lastly, the possibility that oral contraceptives are asthmogenic has been raised, but not confirmed.? MATERIALS AND METHODS STUDY SUBJECTS Spirometry data were obtained from 2,066 women having an Automated Multitest Laboratory (AML) examination in 1969. Data were excluded for 163 women on whom fewer than three readings were obtained. This group did not differ statistically from the whole with respect to either OC use or pulmonary function. The study subjects were divided according to hormone use: (a) a current user was one who was taking oral contraceptives at the time of the AML visit; (b) a past user had taken one or more oral contraceptives prior to that time; (¢) a never user had no experience with oral contraceptives; (d) a heterogeneous group of older women receiving estrogenic compounds for noncontra- ceptive purposes was considered separately; and (e) 69 women were in the first trimester of pregnancy. The distribution of oral contraceptives by brand and dose is given in table 90. WEDGE SPIROMETER A wedge spirometer, model 470 with a model 280 pulmo- digitizer,” has been in use in the AML since July 12, 1969. Round plastic disposable mouthpieces and large-bore tubings were used. Three maximal expiratory efforts were obtained with the aid of a noseclip. The third of these trials was used for this analysis. While the best of three efforts is used for clinical purposes, the results of the third trial were used for analysis. For statistical purposes it was desirable to select data from either the second * Hi-Fi spirometer, model 470, and model 280 pulmodigitizer, Med-Science Electronics, Inc., 1455 Page Industrial Blvd., St. Louis, Mo. 63132. TABLE 90.— Distribution of oral contraceptives taken by current users admitted to the study in 1969 Estrogen dose, mg/tab Type of formulation, estrogen, and progestin Distribution Progestin ose 0.05 0.06-0.08 0.10 0.15 Percent mg /tab Percent Percent Percent Percent COMDINBLION ..... nic mim mim smi mmm iim mie seem mm smi BOTH... comms nmr rE tre A eer rrr Arr rE MeStIanol.. . coon tn mmm mmm wr ——— ED merrier meme EE in ET ET ert phere re rales Norethindrone. _________________._______ 49.7 1.0 19.0 D2) {creer miomisman eins cme Esti en D0 [ence wn mam) sans 28.4 | ___________ 10.0 iver omatend wh | |r me a Norethynodrel _ _ ______________________. 10.0 0.38. 5... . cmermmmmirms ame 7.0 |. ciceeenemas BAY 1... commmttinm Be Bmsmsersni De es cr FO.10) | creerersomemermsmcmsmiomd Ye veressremnio sss Yee mo mms 0.3 Ethynodiol diacetate____________________ 17.0 YL 0 |code ees wmammomes 7.00 cas Bthinyiaatrofiol. co cocoon enna onn mmm ES Oc fl sR Sa rT Norethindrone acetate___________________ 9.4 1.0 80 or srs rer) sree nee se Ee ee 2.5 BB |. i mms imme nnn sa EE Hy ern EEE Norgestrel _____________________________ 2.4 5 Doh 1... ori iri msi nosso mire | os em i Medroxyprogesterone_ __________________ 10.0 10.0 Lol ecommerce oh BOUMENEIRY. cov min ion me ti ms rs 7 3 AR COC RL, en CS |S La Lite MEE BNI0] oe ~ mim imi minim saree 8 ele sasaE Rees) E Es AE he bere me Me Norethindrone_ ________________________ 1.6 D0 {ee rr Vil [on weenie mms mes a Chlormadinone acetate__________________ 6.6 teers 6.5 |e rein mi) mii Bthinyl estradiol... cb cunsmmmmsnnmsanmnm EE ca er Rr RS ra gc a Dimethisterone_________________________ 2.1 2.0 \..cnnaees errata ANV X4LINO0YIdS SHAILdADVALNOD [61 192 SOME INITIAL RESULTS or third trial because the distribution of such values more closely approximates a Gaussian (normal) distribution than does the distribution of the “best” values. The third trial was selected because a significant increase was observed in mean value of each spirometric variable with successive trials. Values recorded were Forced Vital Capacity (FVC), Forced Expiratory Volume at 14 sec (FEV 0.5) and at 1 sec (FEV 1.0), and Peak Flow. Reading of the expirogram was automated. Re- sults were visually displayed and transferred to punchcards for data processing and analysis. Electronic function of the wedge spirometer was calibrated daily to an accuracy of +1 percent. A comparative study® has shown that FVC, FEV 0.5, and FEV 1.0 for this instrument are highly correlated with those produced by the Collins bell-type spirometer.© ANALYSIS OF THE DATA As figure 42 shows, the age distribution of women tested by spirometry varied with OC-user status. Spirometric values are known to vary with age.® This is shown in figure 43 for the sample of women studied here. The most widely used spirometric variables, Forced Vital Capacity (FVC) and Forced Expiratory Volume at 1 sec (FEV 1.0) have been graphed. The other two variables measured, Forced Ex- piratory Volume at 0.5 sec (FEV 0.5) and Peak Flow, demon- strated similar relationships. As is evident from figure 43, the relationship is nonlinear. (Age was entered as a cubic in a step- wise multiple regression.) Therefore, to make comparisons of mean spirometric values among the user categories, the mean values were age adjusted according to the age distribution of the entire sample. Table 91 contains the observed mean values for FVC, FEV 0.5, FEV 1.0, and Peak Flow according to OC-user status. In addition, the ratio FEV 1.0/FVC, a possible measure of obstructive lung disease, is included. As can be seen from the table, a small consistent difference in the observed values for never, past, and present OC users is almost entirely removed by age adjustment. The derived ratio, FEV 1.0/FVC, was the same for each of the three OC-user categories and was affected little by age adjustment. Spirometric values were found to be related to height and ¢ Collins bell-type spirometer, Warren E. Collins, Inc., 220 Wood Road, Braintree, Mass. 02184. SPIROMETRY AND ORAL CONTRACEPTIVES 193 40 ] 40 = 30 30 | +0 0 wv wv << 2 ] 1 1 1 1 ] ) 56 58 60 62 64 66 68 70 HEIGHT IN INCHES FIGURE 44.—Forced Vital Capacity (FVC) and Forced Expiratory Volume (FEV 1.0), by height. these women who were receiving estrogenic compounds for non- contraceptive purposes. The possibility exists that OC effects may be related to dosage or duration of exposure. Therefore, these data were examined with respect to these variables. No relation between spirometric values and dosage of OC at the levels in use by sample subjects could be detected. Three measures of exposure were used: total months of OC use; length of the last continuous interval of OC use; and, for past users, time since last OC. These were found to be statistically independent of spirometric variables. COMMENT A wide range of adverse reactions have been attributed to oral contraceptives. Abnormalities in spirometric values, however, are not included among these. Not a single report of such abnor- malities ascribable to oral contraceptive use has come to our attention. The effects of gonadal hormones and contraceptive steroids on the pulmonary system have been reviewed elsewhere ;* SPIROMETRY AND ORAL CONTRACEPTIVES 197 3.6 3.41 ~ ~ 32k , LITERS IN 3.0 - SPIROMETRY -— ~, ia - Np ~ =o FEV 1.0 2.6 7 2.4 rd . 1 1 1 1 1 1 1 ef <100 120 140 160 180 200 2220 WEIGHT IN POUNDS FiGURE 45.—Forced Vital Capacity (FVC) and Forced Expiratory Volume (FEV 1.0), by weight. there is some indication that these hormones may have an effect on lung mechanics, gas transfer, and the lung parenchyma, with- out any direct influence on spirometry. Additional reported effects have been measured by techniques other than spirometry such as compliance, airway resistance, and diffusing capacity. It is possi- ble, of course, that such techniques would be more successful in discriminating between OC-user groups. Analysis of both the pregnant women and the group of women receiving other estrogenic hormones was limited because of the size of these groups. Women receiving estrogens for noncontracep- tive purposes comprised a diverse group. Not only were they receiving a number of medications (Premarin, Ogen, ethinyl estradiol), but reasons for treatment varied widely. Larger num- bers are required to perform the necessarily age-specific compari- son of these groups. The results of the present study confirmed the absence of any OC effect after consideration of the relevant covariables. Because of the age and size of the population analyzed, it seemed safe to assume that if an adverse reaction was present, the 198 SOME INITIAL RESULTS 3.4 el ~ ~ ~ ~ is 3.2 gc 2 ~oe_ ow ie FVC x w [a — 3.0 | - ° = \ \ > 2.8 \ x \ fo \ w \ = NN o a x 2.6 | So ° —- ~N a > 7) ~ ~ fe 2.4 | ~=—o¢ FEV 1.0 - | 1 1 1 | 0 1 2 3 4 SMOKING INDEX FIGURE 46.—Forced Vital Capacity (FVC) and Forced Expiratory Volume (FEV 1.0), by smoking index. prevalence was extremely low in the age group under considera- tion. It is possible that continuing long-term followup will reveal reactions, whether favorable or adverse, in the older age groups, after prolonged exposure to oral contraceptives. REFERENCES 1. INMAN, W. H. W., VEssEY, M. P., WESTERHOLM, B., and ENGELUND, A.: Thromboembolic disease and the steroidal content of oral contracep- tives: A report to the Committee on Safety and Drugs. Br Med J 2: 203-209, Apr. 25, 1970. 2. VEssey, M. P., Dory, R., FAIRBAIRN, A. S., and GLOBER, G.: Thrombo- embolism and the use of oral contraceptives. Br Med J 8: 123-126, July 18, 1970. SPIROMETRY AND ORAL CONTRACEPTIVES 199 TABLE 92.—Mean smoking index by age and oral contraceptive use for Contraceptive Drug Study population Age in years Never Past Current Total user user user sample Under 20.....«cuinicnsismesnusns 0.89 1.06 11.71 0.99 0-2 eerie im —————— mM .93 1.13 .96 BH Tn we sm Sone 88 .90 .86 .88 28-81. cvcnnmsr an nena 94 1.00 1.05 1.01 Sl i RE RR 93 1.09 1.07 1.03 80:80. viii —— 1.1 .94 1.00 1.03 BOlB.. cv ovo win msi mms 1.05 1.11 1.29 1.13 BAST... ovo et nme ree Rem .94 1.05 .86 .94 ABB... corr mmm nm ey 1.09 1.10 .83 1.05 52.aNG OVEr- cos vim mmm mimi 1.01 1.23 .68 1.02 All 308... ccna nan 1.00 1.00 1.02 1.00 Age-adjusted mean index___ .96 1.03 1.00 1.00 11 case. TABLE 93.—Age-adjusted comparison of pregnant women with all women under 40 years of age Spirometric measurements Total Pregnant sample women ! Forced Vital Capacity (FVC) _ _ _ ooo... Hers... 3.46 3.53 Forced Expiratory Volume, at 1.0 sec EBV 1:0) connie immrrmseswmmeses liters__._ 2.81 2.85 Forced Expiratory Volume, at 0.5 sec (RIV. 0.8) comic sim iimimiss rimimmm i mis liters... 2.01 1.97 Pook Flow... « osm mmm mmm mio liters per sec.... 5.95 5.85 FEV LO/TVC . «cone msm meinem mmm .81 .81 1 Age-adjusted to total population under 40 years of age. 8. DriLL, V. A., and CALHOUN, D. W.: Oral contraceptives and thrombo- embolic disease. JAMA 206: 77-84, Sept. 30, 1968. 4. ERIKSSON, S.: Studies in alpha 1l-antitrypsin deficiency. Acta Med Scand 177: 1-85 (suppl. 432), 1965. 5. LAURELL, C. B., KULLANDER, S., and THORELL, J.: Effect of administra- tion of a combined estrogen-progestin contraceptive on the level of individual plasma proteins. Scand J Clin Lab Invest 21: 337-343, 1968. 6. LIEBERMAN, J., MIiTTMAN, C.,, and KENT, J. R.: Screening for heter- ozygous alpha-l-antitrypsin deficiency. JAMA 217(9): 1198-1206. Aug. 30, 1971. 200 SOME INITIAL RESULTS 7. Lyons, H. A.: Respiratory effects of gonadal hormones. In H. A. Salhanick, D. M. Kipnis, and R. L. Vande Wiele, Eds.: Metabolic Effects of Gonadal Hormones and Contraceptive Steroids. New York: Plenum Press, 1969. Pp. 394-402. 8. HoraAN, J. D., and LEDERMAN, J. J.: Possible asthmogenic effect of oral contraceptives. Canad Med Ass J 99: 130-131, July 20, 1968. 9. FrReepMAN, S. H., and ANDERSON, N. E.: Standards for automated spirometric screening. (Unpublished paper of the Contraceptive Drug Study, Jan. 1973.) 10. Kory, R. C., CALLAHAN, R., and BOREN. H. G.: The Veterans’ Adminis- tration-Army cooperative study of pulmonary function. Am J Med 30: 243-258, Feb. 1961. 13. PETERS, J. M., and FERRIS, B. G., Jr.: Smoking and morbidity in a eol- lege-age group. Am Rev Resp Dis 95: 783-789, May 1967. 12. STEIN, M., TARABEIH, A. YASUTAKE, T., and Hirose, T.: Effects of gonodal hormones and contraceptive steroids on respiration. In H. A. Salhanick, D. M. Kipnis, and R. L. Vande Wiele, Eds.: Metabolic Effects of Gonadal Hormones and Contraceptive Steroids. New York: Plenum Press, 1969. Pp. 381-393. Chapter 6 Effects of Oral Contraceptives on the Thromboelastogram® IRWIN R. FISCH, SHANNA H. FREEDMAN, AND FREDERICK A. PELLEGRIN Many investigators have reported an association between oral con- traceptives (OC’s) and thromboembolic events. This report concerns the effects of contraceptive drugs on in vitro clot formation. The test used, thromboelastography, quantifies the rate of clot formation and clot firmness. Population.—Thromboelastograms were analyzed from 1,133 cur- rent, past, and never OC users and pregnant women tested in 1969 and 1970. Thromboelastographic apparatus.—The thromboelastographic appa- ratus consisted of a piston suspended by wire into a cup. While the thromboelastogram was being made, the cup was rotated and a tracing made by the fibrin attachment between the piston and the cup wall. A permanent tracing was produced upon which the rate of clot formation and the firmness of the clot could be measured. Measurement and interpretation of thromboelastograms.—Three measurements were evaluated on each thromboelastogram; R and K measured clotting times, Ma firmness of clots. The quantity 100(R+K)/Ma was defined as a thromboelastographic index (TI). Method of blood handling.—Whole blood was used in the Automated Multitest Laboratory. Blood was drawn in a siliconized vacutainer tube and transferred to the thromboelastographic cup with a siliconized pipet. The instrument was started within 90 sec of blood drawing. Thromboelastographic reproducibility.—Reproducibility from a sin- gle blood sample was excellent. Between-reader variability of the same sample was very low. Intrawoman variability was small when compared to variation between women. Results.—Analysis of thromboelastographic parameters associated with contraceptive drug use revealed no differences between never and past users. Current users had shorter clotting times and firmer clots. These changes were not related to a specific brand, component, or dose of the contraceptive drug. * Reproduced (with slight modification) from Clin Pharmacol Ther 14: 238-244, 1973; copyrighted by The C. V. Mosby Co., St. Louis, Mo. 201 202 SOME INITIAL RESULTS Discussion.—A hypercoagulable thromboelastogram implies an in- creased propensity for clotting, not thrombogenesis. Nevertheless, the delicate balance between thrombus formation and the body’s resistance to clotting is upset when a hypercoagulable state exists. The thrombo- elastographic parameters for never and past users were similar, which implies that the increased in vitro clotting activity among OC users is reversible. Longitudinal study is needed to determine whether the minor alterations in the thromboelastograms of women taking OC’s are of clinical significance. INTRODUCTION An association between oral contraceptives and thrombo- embolic events has been epidemiologically documented by most investigators'-? although challenged by others.* Whether the impli- cated thrombotic tendency is related to changes in the blood, stasis, vascular alteration, or a combination of these factors remains unknown. While not intentionally ignoring the other variables, this report relates only to the effect of contraceptive drugs on intravascular clot formation. Thromboelastography, a simple, highly reproducible method of measuring in vitro coagulation activity, was used in this study. This test quantifies the rate of clot formation as well as clot firmness. The primary goal of this report is to present preliminary findings on the thromboelastographic differences between users and nonusers of contraceptive drugs. Suspicion has arisen that the dose and components of oral contraceptives are significant. Sartwell et al.2 reported that sequen- tial preparations were associated with a greater risk of thrombo- embolic phenomena than were combination contraceptive drugs. Mestranol was suspected of greater thrombogenicity than ethinyl estradiol ; Inman et al.’ were unable to verify this hypothesis, but proposed a positive correlation between thromboembolic risk and estrogen dose. Data analyzing the effect of contraceptive drug composition and dosage on the thromboelastogram (TEG) are also presented. MATERIALS AND METHODS STUDY SUBJECTS Thromboelastograms from 1,138 women tested in the Auto- mated Multitest Laboratory (AML) between January 1969 and April 1970 were analyzed. These subjects included 323 current users (oral contraceptive taken within 30 days), 328 past users, 435 never users of contraceptive drugs, and 47 women in the first ORAL CONTRACEPTIVES AND THROMBOELASTOGRAM 203 trimester of pregnancy. Selection for this subsample was deter- mined by the availability of the test apparatus without regard to contraceptive use or other patient attributes. THROMBOELASTOGRAPHIC APPARATUS V2A stainless-steel cups were placed in three sockets mounted on a thermostatic block that was maintained at 37° C, and test specimens were placed in each cup. (See Part A, Chapter 4, “Auto- mated Multitest Laboratory Phases and Quality-Control Proce- dures.”) Whole blood was used for this study. A stainless-steel piston suspended by a torsion wire was lowered into each cup; calibration was such that the piston rested 1 mm from the bottom and sides of the cup. Attached to the torsion wire was a mirror that reflected a light source onto a visually read scale as well as onto photographic paper. These cups were rotated about the longitudinal axis with oscillations of 4°45’. The excursion took 8.5 sec in each direction and paused 1 sec at each end position. (Total time for one cycle was 9 sec.) If no significant fibrin strand attached between cup wall and piston, the latter did not move and the light reflected from the mounted mirror did not deviate, thereby recording a straight line on photographic paper. When fibrin strands con- nected cup wall to piston, the piston followed the cup on its excursion so that the reflected light beam moved, and during its FIGURE 47.—Procedure for measuring the thromboelastogram (TEG). All measurements are in millimeters. BR, K: 2 mm = 1 min. 7; = time blood is drawn; T. = time thromboelastograph started; BR = time from when blood is drawn (T.) until 1-mm separation of graph; K = time from 1- to 20- mm separation of graph; Ma = size in millimeters of maximum graph separation. 204 SOME INITIAL RESULTS 1-sec pause a darker line was registered. As firmness of the clot attachment between these surfaces increased, the excursion of the piston became greater and the lines spread farther apart. The resultant tracing appeared as a tuning fork on its side (see fig. 47). If the procedure was allowed to continue a sufficient period of time, lysis was demonstrated on the TEG. MEASUREMENT AND INTERPRETATION OF THROMBOELASTOGRAMS Three parameters were evaluated on each TEG. The proce- dure for measuring the tracing is summarized in figure 47. The parameter R was measured from the time blood was drawn until the line became 1 mm thick. It probably represented the time required to attain significant attachment of the fibrin strand be- tween the cup wall and piston. The parameter K reflected the time needed after R was attained to produce a clot firm enough so that the light source was deflected 20 mm. The photographic paper moved at 2 mm/min, and we elected to report BR and K in milli- meters rather than minutes. The parameter Ma measured the distance in millimeters between the tracing separation when the clot attachment between cup wall and piston was at maximal firmness. It was desirable to summarize the entire thromboelastogram by a single parameter. One possibility was to use the quantity 100(R+K)/Ma, which we have called the thromboelastographic index (TI). This was useful as a summary value, since R and K are positively correlated with each other and negatively correlated with Ma. No doubt other functions of R, K, and Ma might have been used, but this one had the advantages of simplicity and con- venience. Slightly abnormal R, K, and Ma values, when combined in this way, added to an abnormal TEG, which might not have been detected by any of the parameters individually. METHOD OF BLOOD HANDLING Whole blood was found preferable for use in the AML. The patient-flow pattern allowed blood transference to the thrombo- elastograph within 1 min. Technician time and handling of the material was minimal ; thus, the possibility of error was decreased. Antagonists to clot formation were not diluted; and the roles of red blood cells, platelets, and other factors were allowed to act. Blood was drawn in a siliconized vacutainer tube between 9 a.m. and 2 p.m. Since other tests were being performed, the ORAL CONTRACEPTIVES AND THROMBOELASTOGRAM 205 sample for thromboelastography was collected in the second or third tube, so that contamination by tissue juice was minimized. Faulty venipuncture samples were not used. When the specimen was collected, a stopwatch was started and this time used for the calculated of R. A siliconized pipet transferred 0.35 cm® of whole blood to the cup (previously warmed to 37° C). The piston was lowered into the sample approximately 45 sec from blood drawing, and mineral oil was placed over the blood to prevent changes at the blood-air interface, such changes would distort the TEG. The instrument was then started within 90 sec of the blood drawing. Cups and pistons were cleaned scrupulously to avoid artifacts and attain maximum reproducibility. THROMBOELASTOGRAPHIC REPRODUCIBILITY Based on three TEG’s from a single blood sample, machine variability was studied extensively. The reproducibility was ex- cellent, so that usually only one TEG was obtained on each subject. However, quality control was maintained by frequently producing duplicate or triplicate TEG’s from one specimen. Reader vari- ability was also closely monitored with very satisfactory agree- ment between two observers. An intensive series of TEG’s—five women providing three TEG’s each week for 8 weeks—showed that a woman’s TEG is quite constant; the intrawoman variability was small when com- pared to variation between women. An analysis of variance test for R, K, and Ma showed reasonably constant values within each woman for these parameters (p <0.01). Furthermore, variation between the two TEG’s made from a single blood sample was not much less than between two determinations made a week or more apart. RESULTS The first comparison made was between the 435 never users and 328 past users of contraceptive drugs. For none of the values R, K, Ma, or TI was the difference between the two groups signifi- cant at the 5-percent level; in fact three out of the four compari- sons yielded significance probabilities of over 50 percent. For this reason, these two groups were combined into a single group, the current nonusers. Forty-seven women in early pregnancy were excluded from the group of nonusers. These women had thromboelastographic parameters that fell between those for users and nonusers. 206 : SOME INITIAL RESULTS TABLE 94.—Thromboelastographic parameters by oral contraceptive use (mean £1.96 standard deviation of mean) Current users | Current nonusers Parameter ! (N =323) (N =1763) Be ema and hn ek RP Sa 23.4+0.6 25.7+0.3 IR oi ae Reet ain a SS wR en 8.7T+ .2 10.0 .1 Ma, oon nin nen sins 56.0+ .4 53.6% .3 TL ions ais rim ines ini dhmrivie amb Bim 58.3+ .5 67.0+1.0 !R and K measure clotting times; Ma, firmness of clots. The quantity 100 (R+K)/Ma is defined as a thromboelastographic index TI. x @ 20 ul © Ed Ss ° ®* &* 10 5 40 2 8 30 28 "20 20 w © 2. as ® eA 10 © 0 ® / 5 7 5 2 ~ Ns 737 a a5 a9 53 57 6 65 69 “20 30 40 5 6 70 8 90 100 Ma (mm) TI FIGURE 48.—Frequency distributions of TEG parameters between current users and current nonusers of contraceptive drugs. Current users are represented by a solid line, current nonusers by a dashed line. BR, K: 2 mm = 1 min. ’ Table 94 compares the mean values of R, K, Ma, and TI for oral contraceptive users and nonusers. In addition, the quantity 1.96 X standard error of the mean is included to provide a 95- percent confidence interval for the mean. These data demonstrate that, indeed, current oral contraceptive users exhibited a statisti- cally significant tendency toward thromboelastographic hyper- ORAL CONTRACEPTIVES AND THROMBOELASTOGRAM 207 coagulability. In fact, this held true in each of the four decades of age into which the study subjects were subdivided. Since no significant relationship between age and thromboelastographic variables was found in this population, all age groups have been combined. The frequency distributions of R, K, Ma, and TI for TABLE 95.—Thromboelastographic parameters by estrogen type and dose (mean 41.65 standard deviation of mean) Dose, number of women, Type of estrogen and thromboelasto- Ethinyl estradiol Mestranol graphic parameter ! Sequential | Combination | Sequential | Combination formulation | formulation | formulation | formulation 50 ug: Number... .cm=s- 0 2 40 0 346 BB itn mmm means neg 25.240.9 [ouennemacan= 23.74+0.9 Kl mdm inane ema nea 894+ 4 |e. 8.7+ .3 NB osrsrmmse lines SEER BET Bl iim 56.6+ .8 IT oma mimi nS) am mmm 61.93.11 |oreremnnmmmm 57.3+2.3 60-75 ug: Number____________ 0 0 0 19 Beyer eis raer fe resnser rae EP a Ss 24 541.6 Be RR ER ER SRE) SR A Ne TR 9.4+1.5 A de ie i AY mre tage oA ) tee Emir ey wheter 55.7+1.6 IL or remem wii mrs i i a em atars 62.1+7.1 80 ng: Number____________ 0 0 533 623 Bist em A mR eB EA NR imme eerie 23.1+1.8 23.84+0.9 mimes sigma item missed yee es 8.6+ .4 8.5+ .6 Ma |e. 55.8+1.2 55. 741.7 TL... cnwiwmwmmmneonnalmenmnsmasesa vases muses 55.6+4.6 58.5+4.0 100 pg: Number____________ 78 0 0 8144 Foil etn momma memes 24. 841.4 |e eee 23.54+0.5 Bo... commis 8.9% 2 lauemvnenni)ocinanmimann 8.84 .3 Ma....coanuninanes 58.82: 8 Lone weennnle nme smms anne 56.4+ .6 TL. ovanienmmmesie 60.022 0° inves nrminnl sn mis mm atn 58.3+1.5 ! R and K measure clotting times; Ma, firmness of clots. The quantity 100 (R+K)/Ma is defined as a thromboelastographic index TI. 2 Qvral, Provest, Norlestrin. 3 Norinyl-1, Ortho-Novum-1. 4 Ortho-Novum, Norinyl, Enovid-5. 5 C-Quens, Ortho-Novum SQ, Norquen. 6 Ortho-Novum 1/80, Norinyl 1/80. 7 Oracon. 8 Ovulen, Norinyl-2, Enovid-E, Ortho-Novum-2. 208 SOME INITIAL RESULTS oral contraceptive users and nonusers are shown in figure 48. These distributions indicate that the differences in the means between users and nonusers of oral contraceptive drugs were accounted for by a consistent shift toward thromboelastographic hypercoagulability throughout the entire population and not merely by extreme values among a small group of subjects. The effect of drug components and dosage was examined. A breakdown of products by estrogen dose, whether combination or sequential, and type of estrogen is listed in table 95. There were no women in the study on a 150-ug-dose pill. Eight women with unknown estrogen dose were dropped from this analysis. A dose- and type-specific comparison between the estrogens was possible only at the lowest dose level, at which there were 40 women receiving ethinyl estradiol and 46 mestranol. In each of the four comparisons made, use of combinations containing mestranol was associated with slightly more hypercoagulable TEG’s than ethinyl estradiol (p<0.06). Larger differences were found in clotting times (R) than clot firmness (Ma). A slightly prolonged R was also associated with mestranol containing sequen- tial preparations, although it was possible to compare only 33 women using mestranol at 80 pg with eight women receiving ethinyl estradiol at 100 pg. When R, K, Ma, and TI were checked against estrogen dose for combination drugs containing mestranol, no effect of dose could be found for any of the variables. A similar breakdown of drugs into their progestational component and dose was made. None of the thromboelastographic parameters correlated sig- nificantly with progestin drug or dose. A significant thromboelastographic difference exists between users and nonusers of contraceptive drugs. Any conclusions based on the oral contraceptive composition data, however, can only be considered working hypotheses for further investigation due to the small sample size in each group. Furthermore, it should be recognized that these estrogens are coupled with different types of progestin so that estrogen effect is confounded with progestin effect. DISCUSSION Thromboelastography is a useful tool for studying clot kinetics and firmness which quantifies, in vitro, the intravascular coagu- lation activity of an individual. Nevertheless, since Hartert intro- duced the device in 1947,° there have been limited data on its use outside the research laboratory. This study describes the use ORAL CONTRACEPTIVES AND THROMBOELASTOGRAM 209 of this technique on a large number of subjects in a multiphasic testing environment. The value of the TEG for evaluating hypocoagulable and fibrinolytic states has been established by De Nicola,” von Kaulla,? and Weiner and Weisberg;® however, hypercoagulability has been given little attention. Employing the thromboelastograph, increased in vitro clotting activity associated with pregnancy and progestational agents'* has been reported. Hypercoagulable TEG’s have been found also in patients prone to thromboembolism such as in cases of myocardial infarction, malignant disease, and severe infection.!? Thromboelastographic hypercoagulability is defined as an accelerated rate of clotting and increased firmness of the formed clot. A hypercoagulable thromboelastogram implies an increased propensity for clotting—not thrombogenesis. This has probably been the basis for many investigators’ denying the existence of hypercoagulability. It is likely that only a small percentage of individuals in a hypercoagulable state will actually form a sig- nificant clot. Subjects in a hypercoagulable state may be unable to form a clot in the presence of adequate intravascular antago- nism to thrombus formations, good blood flow, or an undamaged vessel wall. Nevertheless, the delicate balance between thrombus formation and the body’s resistance to clotting is upset when a hypercoagulable state exists. The majority of clotting studies assay for either a single factor or combination of factors, usually in the presence of an added procoagulant. This approach is suitable for diagnosis and treatment of deficiency or hypocoagulable states; however, these methods do not precisely reflect what actually happens to the fibrin clot—the endpoint of thrombus formation. To study hypercoagulability, the rate of clot formation as well as its quality is important; thromboelastography provides this information while yielding a permanent record. This technique quantifies the total effect of all procoagulants as well as inhibitors of clot formation. Thus, thromboelastography can still be normal with 250 percent factor VIII activity and 600 mg-percent fibrin- ogen if the other members of the coagulation mechanism have compensated in the opposite direction. Current users of contraceptive drugs formed clots faster (shorter R and K) and firmer (larger Ma) than nonusers; how- ever, TEG parameters of never and past users were similar. This implies that-the increased in vitro clotting activity among oral contraceptive users is reversible. 210 SOME INITIAL RESULTS It should be emphasized that markedly hypercoagulable TEG’s, as observed in thrombotic states, were rare in this studied population. This is somewhat reassuring to women taking oral contraceptives, since it is possible that the minor, but statistically significant, changes in the TEG are not of sufficient magnitude to initiate thrombogenesis. Certain women may, indeed, exhibit extreme thromboelastographic hypercoagulability, especially shortly after starting contraceptive drugs, but this was not un- covered in this cross-sectional study. Thrombogenetic differences between these preparations have been attributed to estrogen dose.’ Although our data included only 50 ug ethinyl estradiol-combination oral contraceptives, the R time was slightly shorter among women taking 50 ug mestranol brands, which requires a larger study sample for verification. No differences were found, however, as the mestranol dose increased from 50 to 100 pg. Therefore, this study failed to uncover any TEG differences related to estrogen dose as well as progestin composition. Although this report presents thromboelastographic evidence of slightly increased intravascular coagulation activity in vitro associated with current contraceptive drug use, only one small aspect of thrombogenesis was investigated. Nevertheless, data obtained in this manner give direction to those areas in which more detailed epidemiologic studies may be most fruitful. REFERENCES 1. InMAN, W. H. W,, and VESSEY, M. P.: Investigation of deaths from pulmonary, coronary, and cerebral thrombosis and embolism in women of child-bearing age. Br Med J 2: 193-199, Apr. 27, 1968. 2. SARTWELL, P. E., Masi, A. T., ARTHES, F. G., GREENE, G. R., and SMITH, H. E.: Thromboembolism and oral contraceptives: An epidemiologic case-control study. Am J Epidemiol 90 (5) : 365-380, Nov. 1969. 3. VEssEy, M. P., and DorL, R.: Investigation of relation between use of oral contraceptives and thromboembolic disease. A further report. Br Med J 2: 651-657, June 14, 1969. 4. DriLL, V. A., and CALHOUN, D. W.: Oral contraceptives and thrombo- embolic disease. JAMA 206: 77-84, Sept. 30, 1968. 5. INmaAN, W. H. W., VEssEY, M. P., WESTERHOLM, B., and ENGELUND, A.: Thromboembolic disease and the steroidal content of oral contracep- tives: A report to the Committee on Safety and Drugs. Br Med J 2: 203-209, Apr. 25, 1970. 6. HARTERT, H.: Blutgerinnungstudien mit der thromboelastographie, einen neuem untersuchungsverfahren. Klin Wochenschr 26: 577-583, 1948. 7. DE NicoLa, P.: Thromboelastography. Springfield, Ill.: Charles C Thomas, 1957. 10. 11. 12. ORAL CONTRACEPTIVES AND THROMBOEL ASTOGRAM 211 voN KAuLra, K. N.: Continuous automatic recording of fibrin formation and fibrinolysis: A valuable tool for coagulation research. J Lab Clin Med 49: 304-312, Feb. 1957. WEINER, M., and WEISBERG, L. G.: Clot firmness. Blood 12: 1125-1131, Dec. 1957. MARKARIAN, M., and JACKSON, J. J.: Comparison of the kinetics of clot formation, fibrinogen, fibrinolysis, and hematocrit in pregnant women and adults. Am J Obstet Gynecol 101: 593-602, July 1, 1968. MarguLls, R. R., AMBRUS, J. L., MINK, I. B.,, and STRYKER, J. C.: Pro- gestational agents and blood coagulation. Am J Obstet Gynecol 93: 161-166, Sept. 15, 1965. MARCHAL, G., LEROUX, M. E., and SAMAMA, M.: Atlas de Thrombody- namographie. Paris: Service de Propagande Edition, 1962. Chapter 7 Types of Oral Contraceptives, Depression, and Premenstrual Symptoms® S. JEROME KUTNER AND WILLARD LL. BROWN The aim of this study was to establish whether there is an associa- tion between oral contraceptives (OC’s) and depression. Measures of depression.—Two types of measures of depression were used, the Minnesota Multiphasic Personality Inventory (MMPI) De- pression Scale, and self-rating mood scales. Sample—The mood-scales sample consisted of 5,151 never, past, and current OC users who took an Automated Multitest Laboratory examination in 1969. These women had a mean age of 35.6. Of these, 3,919 completed the MMPI Depression Scale. User status.—There was no difference in the MMPI depression scores between OC users and nonusers. A significantly smaller percentage of patients reported symptoms of severe premenstrual depression among OC users as against never users. (Past OC users expressed more severe depression than current OC users and never users.) Duration: Exposure and discontinuance.—No relationship was found between duration of exposure or time since discontinuance and depres- sion. Comment.—The findings cast doubt on the hypothesis that OC use causes depression. Rather, there seems to be considerable menstrual im- provement, especially with combination drugs. One study! of 797 oral contraceptive users found that 40 percent of the patients who discontinued the drugs because of side effects had at least one mood complaint. The question of whether OC’s cause depression has been asked for several years. Some patients seem to improve, and others become more depressed. In a leading article, the British Medical Journal® suggested that “...as many as 6-7 percent of women suffer depressive symptoms while taking oral contraceptives compared to 1-2 percent of controls (p. 127).” Their estimate was based on studies! 5 ¢ that * Reproduced (with revisions) from J Nerv Ment Dis 155: 153-162, Sept. 1972. 213 214 SOME INITIAL RESULTS used different measures of depression. In addition, biochemical speculations have been advanced’ and some blame progestin. Hence, concern about depression has deepened within the past 2 years with the widespread use of more and more progestational pills. : Since the present study is cross-sectional, it cannot answer the question of causation. However, it can inquire about an association between depression and the drug, which would be indispensable for causation. The following issues will be considered. Do OC users differ from never users? Do OC users have depres- sion which varies with duration of exposure, discontinuance, or type of drug? These questions can be answered because of the un- usual combination of attributes possessed by this study : the sample size (thousands of patients, including past users) and the use of objective measures of depression. Discontinuance and depression deserve special attention for the following reasons. First, many past users may have discon- tinued because they became depressed, which would lead to underestimates of depression among current users (the survivor effect). Second, past users may provide data on the reversibility of the depression which occurs on the drug. Third, a reduction in progesterone is concurrent with premenstrual depression and postpartum depression.® METHODS AND SAMPLE MEASURES OF DEPRESSION Two types of measures were used for depression: The Min- nesota Multiphasic Personality Inventory (MMPI)? Depression Scale, and self-rating mood scales. Three mood scales were con- tained in a comprehensive health questionnaire. These scales dealt with three symptoms: moodiness, premenstrual moodiness, and premenstrual irritability. The question for each of these symptoms was: “Since your last multiphasic, or in the past year, did you have . ...” The patients answered on 4-point scales (never, mild, moderate, and severe). Scores were computed by assigning 0 for never, through 3 for severe. Moreover, the percentage ‘“severe” was used to estimate the prevalence of extreme cases. These measures were selected for their comparability to those of Moos.? He reported their strong correlations with other symptoms of ® By S. R. Hathaway and J. C. McKinley, New York, The Psychological Corporation, 1967. DEPRESSION AND PREMENSTRUAL SYMPTOMS 215 depression, as well as validity evidence concerning their rela- tionship with use of oral contraceptives. Thus, the mood scales can be used to measure depression. Further, Kutner et al.® meas- ured before-after changes with the premenstrual scales among OC users. Finally, all three scales were significantly (p < 0.001) related to a question on history of severe depression during or after pregnancy. Thus, this history seems to have been reflected in these measures of current symptoms among 1,458 patients. The two premenstrual measures correlated at 0.70 (N = 5,225), so that they should yield fairly consistent relationships with other variables. They were only moderately correlated with moodiness (r = 0.51, with premenstrual moodiness, and r = 0.39, with premenstrual irritability). The MMPI Depression Scale was selected because of its demonstrated!® ability to separate clinically depressed patients from normals. Its standardization permits identification of clinically depressed patients in terms of a T score >70 (or 2 SD above the mean). This scale correlated only mildly with the mood scales. Accordingly, strong consistency of results between these two types of measures is not to be expected. However, the MMPI measure was also significantly (p < 0.05) related to the previously mentioned question on history of depression (N —= 1,045). In addition, the MMPI Depression Scale has measured a reduction in depression among women who continued in contrast with those who discontinued oral contraceptives. Table 96 shows that these four indices of depression are not correlated with either age or parity. Furthermore, the mean age and parity for the severe cases on each index is nearly identical to the mean age and parity for the whole sample. In addition, TABLE 96.—Product-moment correlations between depression and demographic characteristics Demographic Depression Premen- Premen- characteristics Scale Moodiness strual strual (MMPI)! moodiness | irritability Age? __________.. 0.05 —0.08 —0.01 0.06 Parity 3... —.03 .04 .03 .06 1 Minnesota Multiphasic Personality Inventory. 2 N =3,979 for the Depression Scale; N = 5,225 for the other measures of depres- sion. 3 N =38,427 for the Depression Scale; N =4,517 for the other measures of depres- sion. 216 SOME INITIAL RESULTS previous research? has found the MMPI Depression Scale inde- pendent of education and intelligence. SAMPLE The original population consisted of all 8,083 female patients who took the multiphasic examination in 1969. The number of patients who completed the mood scales was 7,658 (94.7 percent). The following types of patients were then deleted: pregnant, postpartum, menopausal, and postmenopausal. The remaining number was 5,151. The mean age was 35.6 years (SD — 8.38). Their mean parity was 2.7 (SD = 1.41). The sample was pre- dominantly a white, suburban group. Of these, 3,919 (76.1 per- cent) completed the MMPI Depression Scale. Table 97 presents the distribution of types of oral contraceptives in the sample. Table 98 gives the frequency distribution of user status among the patients who completed the mood scales: never users (33.5 percent), past users (27.9 percent), and current users (38.6 per- cent). Likewise, current users constituted 39.6 percent of the patients who completed the MMPI Depression Scale. Thus, the representation of current users in both samples is so similar that - there does not seem to be a relationship between current use and completion of the MMPI measure. Additionally, there were 92 patients who took multiphasic examinations and received their first prescriptions for oral con- traceptives (in at least 2 years) on the same day as their examina- tions. Of these, 78 completed the MMPI Depression Scale and thereby provided pre-OC scores. RESULTS USER STATUS Table 98 shows that all three groups yielded the same mean T score on the MMPI Depression Scale, a rather normal 55. Further- more, the 8 percent severe among the current users is identical to ¢ Generic and trade names of drugs—norethindrone with mestranol: Ortho-Novum, Ortho-Novum 1/80, Ortho-Norvum SQ; norethynodrel with mestranol: Enovid, Enovid-E; ethynodiol diacetate with mestranol: Ovulen; norethindrone acetate with ethinyl estradiol: Norlestrin; norgestrel with ethinyl estradiol: Ovral; medroxyprogesterone with ethinyl estradiol: Provera; chlormadinone acetate with mestranol: C-Quens; and dimethisterone with ethinyl estradiol: Oracon. DEPRESSION AND PREMENSTRUAL SYMPTOMS 217 TABLE 97.—Composition of oral contraceptives taken by current users admitted to the study in 1969 Composition Number | Percent Combination with mestranol: Norethindrone, 1.0 mg, and mestranol, 0.05 mg______ 306 16.7 Norethindrone, 1.0 mg, and mestranol, 0.08 mg______ 39 2.1 Norethindrone, 2.0 mg, and mestranol, 0.10 mg______ 476 26.0 Norethindrone, 10.0 mg, and mestranol, 0.06 mg _____ 1 1 Norethynodrel, 2.5 mg, and mestranol, 0.10 mg______ 125 6.8 Norethynodrel, 5.0 mg, and mestranol, 0.075 mg_____ 42 2.8 Norethynodrel, 10.0 mg, and mestranol, 0.15 mg_____ 5 .3 Ethynodiol diacetate, 1.0 mg, and mestranol, 0.10 mg _ 319 17.4 Combination with ethinyl estradiol: Norethindrone acetate, 1.0 mg, and ethinyl estradiol, OOD ING. ee iit rR A HR HER AE ET 69 3.8 Norethindrone acetate, 2.5 mg, and ethinyl estradiol, 0:08 MT oer en bonnes see mami 115 6.3 Norgestrel, 0.5 mg, and ethinyl estradiol, 0.05 mg____ 42 2.3 Medroxyprogesterone, 10.0 mg, and ethinyl estradiol, 0 0D TID we rn erg sir msmrmim———————————————— 23 1.3 OER) ce iri ee A RD SRE 1,562 85.45 Sequential: Norethindrone, 2.0 mg, and mestranol, 0.08 mg______ 56 3.1 Chlormadinone acetate, 2.0 mg, and mestranol, A IE | OORT |: JOR 166 9.1 Dimethisterone, 25.0 mg, and ethinyl estradiol, 0.10 mg_ 44 2.4 Totter renee a dreim mmm wwes——————— 266 14.55 Grand Tota) .. .. iim sims A A BRE 1,828 100.0 NoTE.—The sample includes only patients whose 1st day of their last menstrual flow began from 1 to 50 days before the examination. Composition data were not available for 165 patients who used combination drugs. that estimate received from 78 pre-OC patients (those who re- ceived prescriptions on the same day that they were tested). The pre-OC patients also had a similar mean T' score of 53. In addition, both current users and pre-OC patients were similar to never users on the MMPI measure. By contrast, current users reported significantly less premen- strual moodiness than past users and never users. Correspond- ingly, current users expressed less premenstrual irritability than never users (z = 5.58, p < 0.001, two-tailed test), as well as past TABLE 98.—Use of oral contraceptives and depression Current users Past users Never users Overall Measure x? tree iia Per- Per- Per- Per- Number| cent | Mean |[Number| cent | Mean [Number| cent | Mean Number| cent | Mean severe severe severe severe Minnesota Multiphasic Personality Inventory Depression Scale! _____ 1,551 8.0 | 55 1,121 10.0 | 55 1,247 8.4 | 55 NS 3,919 8.7 55 Moodiness?_ ___________ 1,989 5.1 .88 | 1,436 7.2 .94 | 1,726 4.6 J18 | 282.54 5,151 5.6 .86 Premenstrual moodiness_| 1,989 5.5 1.07 | 1,436 8.8 1.24] 1,726 8.6 1.21] 247.01 | 5,151 7.5 1.16 Premenstrual irrita- ; BY omen aie 1,989 4.3 1.06 | 1,436 12.3 1.34 | 1,726 9.9 1.30 | 3117.60 | 5,151 8.4 1.22 1 Mean is based on T scores, so that the standard mean =50. Percent severe =T score>70 (beyond 2 SD above the standard mean). 2 Mean for moodiness, premenstrual moodiness, and premenstrual irritability is based on a 4-point scale: 0 =none, 1 =mild, 2 =mod- erate, and 3 =severe. 3 p<0.001, df =6 (two-tailed test). NS = Not significant. 812 SLTNASHY TYILINI HNOS DEPRESSION AND PREMENSTRUAL SYMPTOMS 219 users. On moodiness, current users were not significantly different from never users. The past users showed consistently elevated estimates of se- verity. This reaches significance on the MMPI measure when they are compared to current users (z = 2.00, p < 0.05, two-tailed test). Also significant is the difference between past users and current users with respect to severe moodiness (z = 2.12, p < 0.05, two- tailed test). Likewise, more past users showed severe moodiness than never users (z = 2.61, p < 0.01, two-tailed test). On premen- strual irritability, there were more severe cases among past users than never users (z = 2.42, p < 0.05, two-tailed test). TYPE OF REGIMEN Table 99 shows that patients on combination drugs did not significantly differ from those using a sequential type with regard to the MMPI Depression Scale and moodiness. However, combina- tion users exhibited significantly less premenstrual moodiness and premenstrual irritability than sequential users. Patients on se- quential drugs were more like never users than were combination users in terms of the premenstrual symptoms, as well as the MMPI index. The main difference between the two regimens is that combination drugs contain progestin as well as estrogen throughout, while sequential drugs contain only estrogen for the first 14 days, and then both steroids for the next 6 days. In conse- quence, additional analyses of the data were oriented toward progestin, rather than estrogen. Table 100 displays no significant differences in a comparison between patients using two kinds of drugs, which differed both in terms of types of progestin and in terms of dose of mestranol. It was not possible to equate for either of these variables without dealing in small numbers. Figure 49 depicts an inverse relationship between percentage with severe depression and dose of progestin for all indices except moodiness (for the mood scales, N — 1,433; for the MMPI, N = 1,113). Table 101 statistically confirms that drugs containing higher doses of progestin are associated with less severe depres- sion, including less clinical depression (MMPI). DURATION : EXPOSURE AND DISCONTINUANCE Duration of exposure to the drug was indexed in two ways: months of current use or last continuous interval of use, and total months of use or cumulative use. The mean number of months of TABLE 99.— Type of regimen and depression ‘Combination Sequential Measure i x3 Number Percent Mean Number Percent Mean severe severe Minnesota Multiphasic Personality Inventory Depression Scale! _____ 1,325 7.8 55 214 8.4 55 NS Moodiness ? _ _ _ ___________________ 1,708 5.0 .88 266 5.3 .86 NS Premenstrual moodiness__ _ _________ 1,708 5.4 1.05 266 6.0 1.16 38.78 Premenstrual irritability ___________ 1,708 4.8 1.04 266 4.5 1.21 412.21 1 Mean is based on 7 scores, so that the standard mean =50. Percent severe=T score>70 (beyond 2 SD above the standard mean). 2 Mean for moodiness, premenstrual moodiness, and premenstrual irritability is based on a 4-point scale: 0 =none, 1=mild, 2= moderate, and 3 =severe. 3 p<0.05 (two-tailed, df =3). 1p <0.01 (two-tailed, df =3). NS =Not significant. 033 SLTNSHY TVILINI HINOS TABLE 100.—Two types of oral contraceptives and depression Norethindrone, 1.0 mg, with mestranol, 0.05 and 0.08 mg Ethynodiol diacetate, 1.0 mg, with mestranol, 0.10 mg Measure x? Number Percent Mean Number Percent Mean severe severe Minnesota Multiphasic Personality Inventory Depression Scale! _____ 262 9.6 55 239 7.5 55 NS Moodiness? _______________________ 336 5.7 .93 318 3.2 .90 NS Premenstrual moodiness_ _ __________ 336 7.2 1.10 318 4.7 1.10 NS Premenstrual irritability. ___________ 336 6.8 1.08 318 3.2 1.01 NS 1 Mean is based on T scores, so that the standard mean =50. Percent severe=T score>70 (beyond 2 SD above the standard mean). NS =Not significant. 2 Mean for moodiness, premenstrual moodiness, and premenstrual irritability is based on a 4-point scale: 0 =none, 1=mild, 2= moderate, and 3 =severe. SOLAN XS TVAILSNINITId ANV NOISSTIJAA 122 222 SOME INITIAL RESULTS 9% - 8% 2 Oo 7% 2 i w oe - bi 6% i, 0 ~ a a w - om oc ¢ . Y 5% | "iy td Mari in a get ET RS 2 . E hy » Bas ay Y. = gre & 3%h ¥ I 2% 1% | 1 =i 1 1.0 2.0 2.5 DOSE OF PROGESTIN (mg) FIGURE 49.—Dose of progestin and four measures of severe depression. —, MMPI Depression Scale; , moodiness; ®—e, premenstrual moodiness; ® ¢ o, premenstrual irritability. current use was 37.0 (SD = 27.01, N = 1,708). The mean total months of use was 45.5 (SD = 25.89). Table 102 gives the mean number of months of current use by those with each level of de- pression for each measure. Correlations and x* analyses on these data, as well as with total months of use, resulted in no significant relationships. The mean number of months since last OC use among 1,412 past users was 29.6 (SD = 25.38). Correlations and x* analyses demonstrated that the amount of time since discontinuance is in- dependent of all the measures of depression, except for premen- strual irritability (r = 0.12; y* = 47.67, p < 0.001, df = 21, two- tailed test). Figure 50 illustrates that during the first 3 months after discontinuance, severe premenstrual irritability is almost as uncommon as during use. Then, between 4 and 12 months after TABLE 101.—Dose of progestin and depression Low Medium High (0.5 to 1.0 mg) (2.0 mg) (2.5 to 10.0 mg) Measure Z1 Per- Per- Per- N cent Mean N cent Mean N cent Mean severe severe severe Minnesota Multiphasic Personality Inventory Depression Scale 2 ________ 590 9.3 55 378 6.6 55 236 5.9 55 31.70 Moodiness 4 _ ________________________ 764 5.1 .92 472 4.9 .84 310 5.5 .83 NS Premenstrual moodiness_______________ 764 6.4 1.1 472 5.3 1.0 310 3.6 .94 32.04 Premenstrual irritability .______________ 764 5.5 1.1 472 3.8 1.0 310 2.9 .97 31.84 1 Tests of significance were performed on the proportion with severe depression using a low dose of progestin versus that for a high dose. 2 Mean is based on T scores, so that the standard mean =50. Percent severe=T score>T70 (beyond 2 SD above the standard mean). 3 p <0.05 (one-tailed test). + Mean for moodiness, premenstrual moodiness, and premenstrual irritability is based on a 4-point scale: 0=none, 1=mild, 2= moderate, and 3 =severe. NS = Not significant. SOLAN AS TVNILSNINIId ANV NOISSHIIIA €cc "TABLE 102.— Months of current oral contraceptive use and depression MMPI 2 Depression Scale Moodiness Premenstrual moodiness Premenstrual irritability Current users ! 0-49 | 51-59 [61-69 | 71+ | None | Mild | Mod-| Se- | None | Mild | Mod-| Se- | None | Mild | Mod-| Se- erate | vere erate | vere erate | vere Number... 469 529 ;. 223 104 | 724 553 | 345 86 | 453 | 807 | 355 93 | 479 | 759 | 397 73 Months of use_____ 88.0 |37.8/35.6|36.9|39.3|85.6(34.4|36.9|37.9|37.83|35.4|36.4|36.4|36.4(39.5| 33.4 Standard deviation.| 26.6 | 28.0 { 25.5 28.1 | 28.0 | 25.1 | 27.7 | 26.5 | 25.5 | 28.2 (26.2 | 27.3 | 26.0 | 27.2 | 27.9 | 25.7 1 All patients were on combination drugs. 2 Minnesota Multiphasic Personality Inventory. Va SLINSHY TVILINI HIN OS DEPRESSION AND PREMENSTRUAL SYMPTOMS 225 PERCENT WITH SEVERE PREMENSTRUAL IRRITABILITY 3 + rT ™ T T \ 123 4-12 13-24 25-36 37-48 49+ MONTHS SINCE LAST PILL FIGURE 50.—Current premenstrual irritability and number of months since last oral contraceptive use. Total N = 1,412; mean months since last pill = 29.8; SD = 25.4, discontinuance, the prevalence of this symptom approximately doubles and plateaus for about 2 years. Afterward, there is an other sharp increase in prevalence. COMMENT Oral contraceptive users do not seem more depressed than never users. Nevertheless, three sampling problems may have gen- erated underestimates of the relationship between depression and oral contraceptives. First, the most depressed users may have been excluded from the MMPI estimate either because they did not take the multiphasic examination, or they did not complete this scale, even though they took the examination. All that is known about this limitation is that OC users were proportionately represented on the MMPI measure and the mood scales, which a much larger number of patients finished. Second, the extremely depressed pa- tients may have tended to discontinue the drug and become past users. However, past users are not significantly different from never users on the MMPI measure. Indeed, past users’ premen- strual irritability is rather low for the first few months after dis- continuance. What is more, the mean number of months of current use among severely depressed users for each measure of depres- sion is very similar to the mean for all users. A third source of 226 SOME INITIAL RESULTS distortion might have been the lack of discrimination among OC users with different predispositions to depression. On the contrary, additional research with the same sample has discredited this hypothesis.?? The present results cast doubt on the hypothesis that oral con- traceptives cause depression. The fact that current users are simi- lar to never users and OC choosers on the MMPI Depression Scale agrees with longitudinal studies using a variety of objective meas- ures and interviews. * On the other hand, one investigation that found OC users to be very depressed was based on a sample of patients who tended to be dissatisfied with oral contraceptives. Furthermore, neither the present study, nor the one with extraor- dinarily depressed users* obtained a duration-of-exposure effect. Additionally, the only relationship found between a symptom of depression and time since discontinuance involved an increase in the symptom. Moreover, the finding that current users show less premenstrual depression is supported by both previous before- after research® and cross-sectional data.? This premenstrual effect may be due to a shortening of the amount of time when progesta- tional activity is low among users as opposed to nonusers. An alternative explanation is that use of the drug allays fear of preg- nancy, which in turn alleviates depression.’®* Kutner'” has shown this interpretation to be tenuous in that current OC users were not significantly less afraid of pregnancy than either pre-OC patients who had chosen “the pill,” or any other of nine contracep- tive groups. In fact, the greater the progestational activity, the less the depression. This conclusion is consistent with two observations. First, combination users exhibited less premenstrual depression than did sequential users. This is in accord with past research.? Second, the higher the dose of progestin, the less clinical, as well as premenstrual, depression. However, the dose effect may only be an artifact of the tendency for five of the six gynecologists (who participated in the study) interviewed to reduce progestational activity in prescriptions given to patients who complain of depres- sion. Nonetheless, only one of the six said he switched such pa- tients from combination to sequential drugs. Then too, patients who reported a history of severe depression during or after preg- nancy on the health questionnaire did not tend to receive lower doses of progestin than did other patients. The mean progestin dose for patients with this history (N = 130) was 1.9 mg, and the mean for patients without this history (N = 1,170) was 1.8 mg. This depression-progestational effect is diametrically opposed to DEPRESSION AND PREMENSTRUAL SYMPTOMS 27 a previous study, but the measure of depression was not described. Other studies have demonstrated that progesterone'® and proges- tin?® contribute to the amelioration of psychotic females whose symptoms seemed to be related to a progesterone deficiency. Pro- gesterone?® has a suggested mechanism of action comparable to that suggested for lithium?* (a drug used to prevent depression, as well as to prevent and treat mania). Both may lower norepineph- rine levels at critical receptor sites in the central nervous sys- tem. Just as progestin seems to alleviate premenstrual symptoms, so does lithium.2? Thus, it appears important to do well-controlled longitudinal research on progestational activity and depression. Individual differences in the metabolism of progestin in the con- text of depression should be studied, too. The findings of relatively severe symptons of depression among past users are compatible with other observations of patients who stopped taking the drug.® '* Perhaps this withdrawal reaction and postpartum and premenstrual depression all have the same etiology, a hypersensitivity to diminished progestational activity. Nonetheless, it still remains to rule out the possibility that the past users’ depression on the MMPI originated either before or during use of the drug. Accordingly, the Kaiser- Permanente Contraceptive Drug Study will measure the same patients before, during, and after use of oral contraceptives. From a clinical viewpoint, this study provides no justification for withholding oral contraceptives because of depression. How- ever, caution must be used with patients who complain of depres- sion with OC’s. A substantial secondary advantage of reduced premenstrual depression does seem to occur for many patients, especially with the strongly progestational combination drugs. Consideration should be given to progestin replacement for pa- tients presenting depressive reactions to withdrawal from oral contraceptives. REFERENCES 1. GrANT, E. C. G., and PRYSE-DAVIES, J.: Effect of oral contraceptives on depressive mood changes and on endometrial monoamine oxidase and phosphates. Br Med J 3: 777-780, Sept. 28, 1968. 2. Moos, R. H.: Psychological aspects of oral contraceptives. Arch Gen Psychiatry 19: 87-94, July 1968. 3. Depression and oral contraception. Br Med J 4: 127-128 (leading article), Oct. 17, 1970. 4. HERZBERG, B. N., JOHNSON, A. L., and BROWN, S.: Depressive symptoms and oral contraceptives, Br Med J 4: 142-145, Oct. 17, 1970. 5. LEwis, A., and HOGHUGHI, M.: An evaluation of depression as a side effect of oral contraceptives. Br J Psychiatry 115: 697-701, June 1969. 228 SOME INITIAL RESULTS 6. NILSSON, A., and ALMGREN, P.-E.: Psychiatric symptoms during the post- partum period as related to use of oral contraceptives. Br Med J 2: 453-455, May 25, 1968. 7. KANE, F. J.,, TREADWAY, C. R., and EwING, J. A.: Emotional change associated with oral contraceptives in female psychiatric patients. Compr Psychiatry 10: 16-30, Jan. 1969. 8. KANE, F. J.: Psychiatric reactions to oral contraceptives. Amer J Obstet Gynecol 102: 1053-1063, Dec. 1, 1968. 9. KUTNER, S. J.,, PaiLLips, N. R.,, and Hoag, E. J.: Oral contraceptives, personality, and changes in depression. Contraception 4(5): 327- 336, Nov. 1971. i 10. HATHAWAY, S. R., and McKINLEY, J. C.: A multiphasic personality schedule (Minnesota) : III. The measurement of symptomatic depres- sion. J Psychol 14: 73-84, 1942. 11. ZIEGLER, F. J., RoDGERS, D. A., KRIEGSMAN, S. A,, and MARTIN, P. L.: Ovulation suppressors, psychological functioning, and marital adjust- ment. JAMA 204 (10) : 849-853, June 3, 1968. 12. GYNTHER, M. D., and SHIMKUNAS, A. M.: Age and MMPI performance. J Consult Pyakol 30(2): 118-121, Apr. 1966. 13. KUTNER, S. J., and BRowN, W. L.: History of depression as a risk factor for depression with oral contraceptives and discontinuance. J Nerv Ment Dis 155 (3) : 163-169, Sept. 1972. 14. Levitt, E. E., KooIKER, J. E.,, and NorTON, J. A.: Depression and oral contraception. Curr Top Clin Community Psychol 2: 157-182, 1970. 15. Murawski, B. J., Sapir, P. E.,, SHULMAN, B. A, Ryan, G. M,, and STURGIS, S. H.: An investigation of mood states in women taking oral contraceptives. Fertil Steril 19: 50-63, Jan.—Feb. 1968. 16. GLICK, I. D.: Mood and behavioral changes associated with the use of the contraceptive agents. Psychopharmacologia (Berlin) 10: 363-374, 1967. 17. KUTNER, S. J.: A test for fear of pregnancy and its relation to oral contraceptives. J Psychiatry Res 9 (4): 337-344, 1972. 18. Bower, W. H., and ALTSCHULE, M. D.: Use of progesterone in the treat- ment of post-partum psychosis. N Engl J Med 254: 157-160, Jan. 26, 1956. 19. SwaNsoN, D. W., BARRON, A., FLOREN, A., and SMITH, J. A.: The use of norethynodrel in psychotic females. Am J Psychiatry 120: 1101-1103, May 1964. 20. STEFANO, F. J., and Donoso, A. O.: Norepinephrine levels in the rat hypothalamus during the estrous cycle. Endocrinology 81: 1405-1406, Dec. 1967. 21. JArvik, M. E.: Drugs used in the treatment of psychiatric disorders. In L. S. Goodman and A. Gilman, Eds.: The Pharmacological Basis of Therapeutics. New York: Macmillan, 1970. P. 193. 22. SLETTEN, I. W., and GERSHON, S.: The premenstrual syndrome: A dis- cussion of its pathophysiology and treatment with lithium ion. Compr Psychiatry 7: 197-205, June 1966. Chapter 8 History of Depression as a Risk Factor for Depression With Oral Contraceptives and Discontinuance S. JEROME KUTNER AND WILLARD L.. BROWN The objective of this study was to evaluate history of severe depres- sion during or after pregnancy as a predisposition to depressive reac- tions with oral contraceptives (OC’s). Measure of history of depression.—The health questionnaire given in the Automated Multitest Laboratory (AML) asked women if any pregnancies had been complicated by severe depression. Measure of current depression.—Two types of measures of current depression were used in the AML, the Minnesota Multiphasic Per- sonality Inventory (MMPI) and three mood scales contained in the health questionnaire. Sample—The mood-scales sample consisted of 5,151 never, past, and current OC users with a mean age of 35.6. Of these, 4,290 com- pleted the question on history of depression, and 3,919 completed the MMPI Depression Scale. History of depression and discontinuance.— Women who discontinued the drug showed a significantly greater prevalence (11.7 percent) of history of severe depression during or after pregnancy than did current users (9.3 percent) or never users (8.9 percent). On the other hand, those who discontinued the drug revealed no association between number of months of OC use and depressive history. History of depression and current depression.—All four measures of present depression were significantly related to depressive history among current, past, and never users of oral contraceptives. Patients with a history of depression generated no evidence for an association between use of OC’s and abnormal depression. Past OC users reported more cases of severe premenstrual irritability than did current users. The four measures of present depression were independent of dose of progestin. Discussion.—This research found no evidence of oral contraceptives aggravating a depressive history. But a history of depression in relation to pregnancy is related to discontinuance of the drug. The sequence of these events is unknown. * Reproduced (with revisions) from J Nerv Ment Dis 155: 163-169, Sept. 1972. 229 230 SOME INITIAL RESULTS According to a summary of research on oral contraceptives and depression,® depression is suffered by 6-7 percent of women using the drug, as compared to 1-2 percent of controls. However, a recent extensive study? with the Minnesota Multiphasic Person- ality Inventory (MMPI)? Depression Scale found no significant difference between current oral contraceptive users and patients who never used contraceptive drugs. Indeed, patients taking OC’s reported less premenstrual symptoms of moodiness and irritability than never users, and this advantage increased with the dose of progestin. Nevertheless, it would be desirable to identify pre-OC patients who are likely to have depressive reactions to the drug. Such predispositions or risk factors might be able to explain these reactions, while permitting a screening device for the selection of patients. Risk factors for depression with oral contraceptives may consist of properties of the drug or the patient. As mentioned previously, some pharmacological aspects of this problem have been investigated. Education, religion, and family size have not differentiated between users who respond with depression and those who do not.? Nor have various psychological and physio- logical factors made a difference.* Grounds et al.’ found the degree of neuroticism unrelated to this phenomenon. By contrast, Lewis and Hoghughi® reported that previous history of depression is positively related to depression with oral contraceptives. Similarly, Nilsson et al.” discovered that psychiatric reactions to OC’s are directly associated with a pre-OC history of psychiatric symptoms. Likewise, Herzberg and Coppen® showed that users with depres- sive reactions more frequently reported a previous history of premenstrual depression than other OC users. However, none of these studies used a control group of nonusers to evaluate the relationship between history of depression and current depression. Hence, it is not known whether a history of depression is any more predisposing to depression among users than nonusers. Despite the lack of controlled observations, the British Medical Journal advised on the subject: “How should the prescriber try to prevent a depressive reaction from occurring? It should be recognized that some women are at greater risk than others and that they include those who have had previous attacks of depres- sion, those with progestogenic menstrual cycles, and those who normally suffer from premenstrual depression” (pp. 127-128). * By S. R. Hathaway and J. C. McKinley, New York, The Psychological Corp., 1967. HISTORY OF DEPRESSION AS A RISK FACTOR 231 In consequence, the present research compared history of depression to current depression among women who never used oral contraceptives, as well as among those who do. Furthermore, the current depression scores of patients with a history of depres- sion were divided into current, past, and never users of oral contraceptives to ascertain whether this history was relevant to the use of the drug. Additionally, current users with a depressive background were studied in terms of dose of progestin to see if this history represented a sensitivity to progestin as reflected by current depression. Moreover, history of depression was compared to duration of exposure to the drug among past users, since dis- continuance of the drug has been connected with depression.? *° What is more, the frequency of this history was analyzed by user status (i.e., current, past, and never users) to determine if there was a relationship between depressive history and discontinuance. METHODS AND SAMPLE MEASURE OF HISTORY OF DEPRESSION A comprehensive health questionnaire contained the following question which was used to measure history of depression. If you have been pregnant, were any of your pregnancies complicated by severe depression during or after pregnancy? Yes No Pregnancy was used as a reference point for measuring history of depression because any depression which did occur might be based on a sensitivity to the changes in progesterone that coincide with pregnancy and the postpartum period. Since oral contra- ceptives induce changes in progestin, depressive reactions to OC’s might have a source similar to the source of these reactions dur- ing or after pregnancy. Alternatively, the same women might tend to have depressive reactions to both pregnancy and oral contraceptives in terms of the symbolic similarity between these experiences, in that both are germane to childbearing. Whichever explanation is correct, there are grounds for expecting an empiri- cal association between these two depressions. MEASURE OF CURRENT DEPRESSION Two types of measures were used for depression: the MMPI Depression Scale and self-rating mood scales. Three mood scales were contained in a comprehensive health questionnaire. These 232 SOME INITIAL RESULTS scales dealt with three symptoms: moodiness, premenstrual moodi- ness, and premenstrual irritability. The question for each of these symptoms was: “Since your last multiphasic, or in the past year, did you have. ...” The patients answered on 4-point scales (never, mild, moderate, and severe). Scores were computed by assigning 0 for never, through 3 for severe. Moreover, the percent “severe” was used to estimate the prevalence of extreme cases. These measures were selected for their comparability to those of Moos. He reported their strong correlations with other symptoms of depression, as well as validity evidence concerning their relation- ship with use of oral contraceptives. Table 103 presents the intercorrelations among the four measures of depression. The mild to moderate correlations be- tween the nonpremenstrual measures (MMPI and moodiness) and the premenstrual measures (premenstrual moodiness and pre- menstrual irritability) suggest that women with premenstrual symptoms tend to have depression in general. This has been documented previously by Moos et al.,’> Coppen and Kessel,** and Lamb et al.* Indeed, Coppen'® maintained that the premenstrual syndrome represents the commonest form of endogenous depres- sion. Thus, the mood scales in the present study can be used to measure depression. The MMPI Depression Scale was selected because of its demonstrated'¢ ability to separate clinically depressed patients from normals. Its standardization permits identification of clini- cally depressed patients in terms of a 7 score >70 (or 2 SD above the mean). In addition, it has made discriminations relevant to oral contraceptives.? 1° It has been shown? that these four measures of depression are unrelated to age and parity. Moreover, Gynther and Shim- TABLE 103.—Product-moment intercorrelations among the measures of depression Number | MMPI Premen- | Premen- Measure of Depres- | Mood- | strual | strual women sion iness mood- | irrita- Scale iness bility Minnesota Multiphasic Person- ality Inventory (MMPI) De- pression Seale_ _ ______________ 8.979 | al 0.33 0.21 0.16 Moodiness. ......_..cciceesnmnms B25 Io ied |e ne .51 .39 Premenstrual moodiness. ________ BRB]. wt wie i mmr yr .70 Premenstrual irritability ________ BAe ot cm fe bw pe EE HISTORY OF DEPRESSION AS A RISK FACTOR 233 kunas!” found the MMPI Depression Scale independent of educa- tion and intelligence. SAMPLE The original population consisted of all 8,083 female patients who took the multiphasic examination in 1969. The number of patients who completed the mood scales was 7,658 (94.7 percent). The following types of patients were then deleted: pregnant, postpartum, menopausal, and postmenopausal. The remaining number was 5,151. The mean age was 35.6 years (SD = 8.38). Their mean parity was 2.7 (SD = 1.41). The sample was pre- dominantly a white, suburban group. Of these, 3,919 (76.1 per- eent) completed the MMPI Depression Scale, whereas 4,290 completed the question on history of depression. The distribution of types of oral contraceptives in this population of patients who took the multiphasic examination has been previously described. History of depression was independent of dose of progestin: 130 OC users with a history of depression used a mean dose of 1.9 mg, and 1,170 OC users without a history of depression used a mean dose of 1.8 mg. RESULTS HISTORY OF DEPRESSION AND DISCONTINUANCE Table 104 shows that a significantly larger proportion of past users reported a history of depression during or after pregnancy than current users (11.7 versus 9.3 percent). The current users were 1,638 patients on either combination or sequential pills. For past users, the number of mean months since discontinuance was 29.6 (SD = 25.4). However, when the duration of exposure to the drug was compared to history of depression among past users, they were unrelated (y* = 6.28, df — 6). Table 105 presents the mean months of exposure for past users who stated a depressive history and for those who did not. "HISTORY OF DEPRESSION AND CURRENT DEPRESSION All four measures of current depression were significantly related to history of depression in each user status. Table 106 displays this relationship for never users. 234 SOME INITIAL RESULTS TABLE 104.— Percentage with severe depression during or after pregnancy and use of oral contracepiives Women with history of depression Type of oral contraceptive use Total Number Percent Combination Users... ....cceeeeenmneen 133 9.4 1,417 Sequential users_ ____________________ 20 9.0 221 DS BB Se amis nem ma ae 1140 11.7 1,194 NOVOr HOTS. «wo cee mn raw mn 129 8.9 1,458 OVEN. ci sienna Tea 422 9.8 4,290 NoTE.—No subjects were pregnant or postpartum at the time of questioning. The percentage of subjects with a history of pregnancy, before nulliparous women were deleted, was the same for all groups, 83 percent. All subjects included in this table had a history of pregnancy. 1 Past users had a higher prevalence of depressive history than combination and sequential users as a group (Z =2.38, p <0.01, one-tailed test). TABLE 105.—Mean months of last oral contraceptive use and history of depression among past users History Number of Mean Standard women months deviation History of depression .....ocuvuinn-n- 146 13.5 15.6 No history of depression 2_____________ 1,274 16.9 18.3 OVELAll. cot imine sna 1,420 16.5 18.1 1 History of depression was during or after pregnancy. No patients were pregnant or postpartum. 2 Includes nulliparous women. Next, the current symptoms of patients with a history of depression were analyzed in terms of user status. Table 107 summarizes the results, which show that use of oral contraceptives is not associated with depression among patients with this history. For example, the percent severely depressed on the MMPI was not significantly high for current users as opposed to never users (Z = 1.30, N = 225). Nor were current users significantly differ- ent from never users with the three mood scales. For instance, concerning premenstrual irritability, the gap between current HISTORY OF DEPRESSION AS A RISK FACTOR 235 TABLE 106.—Current depression and history of severe depression during or after pregnancy among never users of oral contraceptives Measures of current depression x? pt? Number Minnesota Multiphasic Personality Inventory Depression Scale. ________ 6.85 <0.05 1,045 MOOHRCED. ooo ninmisn wane 69.10 <.001 1,458 Premenstrual moodiness. _____________ 46.45 <.001 1,458 Premenstrual irritability ._____________ 45.43 <.001 1,458 1 df=38 (one-tailed tests). NoOTE.—None was pregnant or postpartum. users and never users did not reach statistical significance (Z = 1.92, N = 295, two-tailed test). Nonetheless, past users more frequently reported severe premenstrual irritability than current users (Z = 2.50, p < 0.05, two-tailed test, N = 311). Progestin-dose data were available for 130 current users with a history of severe depression during or after pregnancy. Table 108 shows how these subjects were divided into a high- and a low-dose group. Their mean T scores on the MMPI Depression Scale were identical. Further, the difference in percent severe was not statistically significant for this measure (Z = 1.09, N = 102). Likewise, none of the mood scales differentiated between the two groups. DISCUSSION History of depression was directly related to current depres- sion not only for current users, but also for never users and past users. Therefore, previous research®® which did not employ a control group of never users to assess the association between history of depression and current depression must be reevaluated. Perhaps these studies merely have the significance of showing that individuals who are prone to reporting depression in the present or in the future are also prone to reporting depression in the past. Indeed, in the present study, among patients with this history, no significant differences between current users and never users were found with the four measures of depression. The negative findings were measured by indices which have detected significant effects among the three user statuses, as well as effects relative to the type of OC. Additionally, the present results were based on adequate statistical power, as defined by Cohen.’* The TABLE 107.— Use of oral contraceptives and depression among patients with history of depression Current users Past users Never users Overall Measure 2 | Per- Per- Per- Per- Number| cent | Mean [Number] cent | Mean [Number| cent | Mean Number| cent | Mean severe severe severe severe Minnesota Multiphasic Personality Inventory Depression Scale! _____ 128 18.8 | 59 109 | 20.2 | 59 97 12.4 | 59 NS 334 17.4 59 Moodiness 2 _____________ 165 17.0 1.47 146 19.2 1.41 130 16.2 1.39 NS 441 17.5 1.43 Premenstrual moodiness___ 165 18.8 1.54 146 22.6 1.68 130 21.5 1.72 NS 441 20.1 1.64 Premenstrual irritability ___ 165 16.4 1.43 146 | 28.1 1.80 130 | 25.4 1.80 [319.20 441 22.9 1.66 NoTe.—History of depression was during or after pregnancy. No patients were pregnant or postpartum. X2 analyses were done on each measure of depression relative to all 8 user statuses. 1 Mean is based on T scores, so that the standard mean =50. Percent severe=T score>70 (beyond 2 SD above the standard mean). 2 Means for moodiness, premenstrual moodiness, and premen- 3 p<0.01, df =6, two-tailed test. strual irritability are based on a 4-point scale: 0 =none, 1=mild, 2 =moderate, and 3 =severe. 962 SLTNSHY TVILINI HINOS TABLE 108.—Dose of progestin and depression among patients with history of depression Low dose (0.5 to 1.0 mg) High dose (2.0 to 10.0 mg) Measure x? Z Number Percent Mean Number Percent Mean severe severe Minnesota Multiphasic Personality Inventory Depression Scale! _________ 51 19.6 57 51 11.8 57 NS NS Moodiness 2 __________________________ 69 14.5 1.39 61 18.0 1.51 NS NS Premenstrual moodiness........-..ccusnn- 69 17.4 1.61 61 21.3 1.48 NS NS Premenstrual irritability _______________ 69 14.5 1.48 61 19.7 1.36 NS NS NoTE.—History of depression was during or after pregnancy. No patients were pregnant or postpartum. 1 Mean is based on T scores, so that the standard mean =50. Percent severe=T score>70 (beyond 2 SD above the standard mean). 2 Means for moodiness, premenstrual moodiness, and premen- strual irritability are based on a 4-point scale: 0 =none, 1 =mild, 2 =moderate, and 3 =severe. NS =Not significant. JOLOVA MSIY V SV NOISSHIJHA J0 AYOLSIH Lg 238 SOME INITIAL RESULTS present statistical power exceeded 85 percent, whereas convention in the behavioral sciences requires only an 80-percent probability of detecting a difference, if it exists. This negative result is buttressed by the independence of current depression from dose of progestin among patients with a history of depression, although any effect that might exist may have been obscured by the tend- ency of physicians to shift patients complaining of depression to a lower dose of progestin.? The lack of difference in current depression between current users and never users may have been an artifact of the tendency of depressed patients to discontinue OC use.? In fact, past users do have a higher rate of history of depression than current users. However, the source of this depression is pregnancy, rather than the OC. Past users were not significantly different from either current users or never users on three of the four measures of current depression. Past users reported more premenstrual irrita- bility than current users, but it has been shown? that this symptom originates after discontinuance. Nevertheless, the past users on the average have been off the drug for about 214 years, so that drug-related symptoms may no longer exist. Perhaps better evi- dence against the possible bias due to discontinuance has been reported :? the mean number of months of current use among severely depressed oral contraceptive users on each of the four measures of current depression is very similar to the mean for all users. In conclusion, it appears that women with a history of depression are not prone to depressive reactions to oral contraceptives. Thus, the connection between depressive history and dis- continuance seems less critical than it would otherwise. Duration of exposure to the drug did not seem to be affected by history of depression among past users. This is compatible with the inde- pendence of current depression from user status among cases with a history of depression. On the other hand, patients who have discontinued OC use do report a greater tendency toward this history than the other groups. This effect is ambiguous, however, since it is unknown whether “history of depression during or after pregnancy” antedated or followed use of oral contraceptives. Accordingly, longitudinal research should be conducted with these variables. If this history is an antecedent of the use of oral contraceptives, a risk factor for discontinuance will have been discovered. If the obverse is true, steps could be taken upon dis- continuance to prevent or lessen the impending pregnancy-related depression. Future research on risk factors for depressive reactions to HISTORY OF DEPRESSION AS A RISK FACTOR 239 contraceptive drugs might be concerned with psychological vari- ables, such as guilt over sex, and physiological variables, such as pyridoxine deficiency.”> Meanwhile, the present research suggests that oral contraceptives should not be withheld from women with a history of depression. SUMMARY Past users of oral contraceptives (N = 1,194) showed a significantly greater prevalence (11.7 percent) of history of severe depression during or after pregnancy than current users (9.3 per- cent of 1,638 patients) or never users (8.9 percent of 1,458 patients). However, 1,420 past users revealed no relationship be- tween duration of exposure and this history. Current depression was significantly related to depressive history in each of the three user statuses. Four hundred and forty-one patients with a history of depression showed no relationship between user status and present depressive symptoms, except for past users being more likely than current users to report premenstrual irritability. Moreover, the four measures of present depression were inde- pendent of dose of progestin used by 130 current users with a depressive history. In consequence, oral contraceptives do not seem to exacerbate this history of depression. However, discon- tinuance is in some way involved with depression associated with pregnancy. REFERENCES 1. Depression and oral contraception. Br Med J 4: 127-128 (leading article), Oct. 17, 1970. 2. KUTNER, S. J., and BRowN, W. L.: Types of oral contraceptives, depres- sion and premenstrual symptoms. J Nerv Ment Dis 155: 153-162, Sept. 1972. 8. Murawskl, B. J, SArIr, P. E., SHULMAN, B. A, Ryan, G. M,, and STURGIS, S. H.: An investigation of mood states in women taking oral contraceptives. Fertil Steril 19: 50-63, Jan.—Feb. 1968. 4. HUFFER, V., LEVIN, L., and ARONSON, H.: Oral contraceptives: Depression and frigidity. J Nerv Ment Dis 151: 35-41, July 1970. 5. GrouNnDs, D., DAVIES, B., and MowBRAY, R.: The contraceptive pill, side effects and personality: Report of a controlled double blind trial. Br J Psychiatry 116: 169-172, Feb. 1970. 6. LEwIs, A., and HOGHUGHI, M.: An evaluation of depression as a side effect of oral contraceptives. Br J Psychiatry 115: 697-701, June 1969. 7. NiLssON, A., JACOBSON, L., and INGEMANSON, C.-A.: Side-effects of an oral contraceptive with particular attention to mental symptoms and sexual adaptation. Acta Obstet Gynecol Scand 46: 537-556, 1967. 8. HERZBERG, B., and COPPEN, A.: Changes in psychological symptoms in women taking oral contraceptives. Br J Psychiatry 116: 161-164, Feb. 1970. 240 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. SOME INITIAL RESULTS GRANT, E. C. G., and PrYSE-DAVIES, J.: Effect of oral contraceptives on depressive mood changes and on endometrial monoamine oxidase and phosphates. Br Med J 3: 777-780, Sept. 28, 1968. ZIEGLER, F. J., RODGERS, D. A., KRIEGSMAN, S. A., and MARTIN, P. L.: Ovulation suppressors, psychological functioning, and marital adjust- ment. JAMA 204 (10) : 849-853, June 3, 1968. Moos, R. H.: Psychological aspects of oral contraceptives. Arch Gen Psychiatry 19: 87-91, July 1968. Moos, R. H., KorELL, B. S., MELGES, F. T., YaLoMm, I. D.,, LUuNDE, D. T,, CLAYTON, R. B., and HAMBURG, D. A.: Fluctuations in symptoms and moods during the menstrual cycle. J Psychosom Res 13: 37-44, Mar. 1969. COoPPEN, A., and KESSEL, N.: Menstruation and personality. Br J Psychiatry 109: 711-721, Nov. 1963. ‘LAMB, W. M., ULErT, G. A., Masters, W. H., and RoBINSON, D. W: Premenstrual tension: EEG, hormonal and psychiatric evaluation. Am J Psychiatry 109: 840-848, May 1953. COPPEN, A.: The biochemistry of affective disorders. Br J Psychiatry 113: 1237-1264, Nov. 1967. HaTtHAWwWAY, S. R.,, and McKINLEY, J. C.: A multiphasic personality schedule (Minnesota): III. The measurement of symptomatic depres- sion. J Psychol 14: 73-83, 1942. GYNTHER, M. D., and SHIMKUNAS, A. M.: Age and MMPI performance. J Consult Psychol 30: 118-121, Apr. 1966. COHEN, J.: Statistical Power Analysis for the Behavioral Sciences. New York and London: Academic Press, 1969. Pp. 209-265. BAuMBLATT, M. J., and WINSTON, F.: Letter to the Editor. Lancet 1: 832-833, Apr. 18, 1970. Contributors to the Contraceptive Drug Study Contact followup of cohort: Iris I. Dingle, Carola F. Nauta, Marian R. Runnings. Multitest laboratory examination of subjects: Barbara K. Barckley, Mary Cochran, Charlotte L. Dyke, R.N., Muriel J. Giles, R.N., Melissa J. Hall, Lorlouise Hassler, J. Darrell Jones, Barbara A. Jurkiewicz, R.N., Aimee von Christierson, Dora A. Wallace. Data processing: Roza R. Bresee, Roberta L. Heintz, M.S., Barbara N. McIntosh, Patricia L. McHenry, Helen M. O’Drain, Suzanne M. Rivera, David A. Roberts. Programming: James O. Baker, Bruce M. Brainard, W. Michael Brodie, Margaret H. Hannigan, Dan S. King, Carolyn S. Scarr. Dissemination of information and administration: Sue Ann Attia, Victor I. Matson, Rita J. Moore, Nancy K. Nagy, M.L.S. Abstracting medical records: Mary Ellen Buckles, Evelyn F. Chan, Ivy Frediani, Michael D. Laughlin, Alma J. Pearson. Biostatistics and data analysis: Robert W. Buechley, Ph.D., Elizabeth J. Hoag, M.A., Donald B. Loveland, M.S., John C. Wingerd, M.A. Clinical evaluation of subjects: Permanente Medical Group physicians. 241 = He pce : } Hh a Feng = L = or inl Br i PE Ep 3 A a oll Mong INDEX A ABO blood types, 16 ABO blood types, blood pressure and, 118, 127, 128 ABO blood typing, 97 Achilles-tendon-reflex relaxation time, 2, 7, 17, 92-93 Age blood pressure, and, 105, 107, 109, 114, 128 contraceptive drug use and, 27, 30, 39, 40, 41 fertility, and, 89, 40, 54 first pregnancy, contraceptive op- erations prevalency, and, 72 glucose levels, and, 156, 157, 185 glucose-tolerance test, and the, 147-168, 170 ketonuria variations and, 154 mean serum-glucose levels by, 151 oral contraceptive use and serum- glucose levels and, 173 parity and blood pressure, and, 118, 125, 126 spirometric values and, 192, 193 Age distribution of study population, 23 Age of initial childbearing, educa- tion and, 72 American Diabetes Association, 136, 139 American Medical Association, Jour- nal of, 105 Anderson, Neil E., vii Angina pectoris, 16 Anthropometric measurements, 7, 100 Artihypertensive medication, 106 Antitrypsin deficiency, 189 Anxiety and oral contraceptive side effects, 46 Atherosclerosis, 186 Audiometry, 17, 93-94 Auto Analyzer, 136 blood-test series, used in, 95, 96 glucose-tolerance tests, used in, 92 Auto Chemist used in blood serum testing, 95 Automated Multitest Laboratory ex- aminations, 1 blood pressure recording in, 105 data processing, and, 13-14 data storage in, 7 dietary regimen data absence and, 165 discussion on, 17-18 disease prevalance rates and glu- cose-tolerance screening, 144 education of examinees, and, 85 February 1968 sample survey and, 83 findings following, 14-17 followup procedures, and, 7 glucose-tolerance screening test in, 135-145, 170 multiphasic health checkups and, 83 phases and quality-control proce- dures, 91-101 phases of, described, 91-100 possible health hazards in oral contraceptive use, and, 86 repeat examinations, 10 spirometric data in, 190 subjects and data tabulated, 5 subjects under medical care, in, 6 three parts of, 7 thromboelastograms in, 202 B Bacteriuria, 98 asymptomatic, 2, 17 Beckman Glucose Oxidase Method, 136 Beckman Microzone Cell, 94 Berendes, Heinz W., vi Beta-cell exhaustion, 186 Biological impairments to fertility, 68 . 248 244 INDEX Birth control, oral contraceptives the most common form of, 52 Birth, live, frequency of oral con- traceptives use and, 49 Blood chemistry tests listed, 95 Blood coagulability See also Thromboelastography increased, 1 oral contraceptive use, and, 206 Blood counts, 96-97 Blood handling in thromoelastog- raphy, 204-205 Blood pressure, 15 ABO blood type, by, 118, 127, 128 age and, 105, 128 age and oral contraceptive use, 120 age-specific frequency distributions of, 114 aging, and, 114 brand and dosage of oral contra- ceptives and, 126 contraceptive drug status and, 107 Deladumone’s effect on, tabulated, 116 dropped in hypertensive women, 121 elevated, description of, 105, 108 estrogen and, 110 estrogenic hormones and, 112 food, hours since last, and, 132 food ingestion and, 118, 119 hormone status and, 106 increased, 1 lactation cessation, and, 127 lactation inhibition, and, 113, 114 menstrual cycle, and the, 118, 129- 130 methods of testing, 106 month of examination, and, 106 months of use of oral contracep- tives, and, 111 National Health Survey compari- sons and, 121 oral contraceptives and, 117 parity and, 118 percent of patients with elevated, 110 postpartum women, in, 112-114, 126-127 pregnancy and, 105, 112-114 pregnancy and oral contracep- tives, 105-133 pregnant women, in, 105 progesterone and, 112 progestin type and dose and, 115- 116 pulse recording, and, 99 reversible when elevated, 122 reversion to preoral contraceptive levels, 110 severe hypertension unusual in oral contraceptive users, 128 study subjects described, 106 table of findings, 107 time of day, illustrated by, 131 weight and, 114 weight and age, and, 122 Blood studies, 7 Blood sugar level and diabetes test- ing, 141 Blood sugar readings, 92 Blood test for glucose tolerance, 91 Blood test, Somogyi-Nelson, 141 Blood typing, 9, 97 Blood urea nitrogen curve, 92 Body size and glucose load, 138 Breast lesions, benign, protection against, 3 British Medical Journal, the, 213, 230 Bronchitis, chronic, 189 Brown, Willard L., vii C Carbohydrate content of diets and glucose-tolerance testing, 165 Carbohydrate deprivation and keto- nuria, 165 Carbohydrate intake and glucose- tolerance testing, 91 Catholic-Protestant ratio of health plan population, 26 Catholic women and oral contracep- tive use, 50 Catholics and contraceptive surgery, 70-71 Center for Population Research (NICHD), iii Chest X-ray procedure, 100 Childless wife and contraceptive op- eration prevalence, 68 Children, number of, voluntary ster- lization and, 68 INDEX 245 Chlormadinone acetate, 8, 115-116, 216 Cholesterol levels, 15 Church attendance and oral contra- ceptive use, 51 Clot formation activity, reversibility of, 209 Clot formation intravascular, 202 rate studies, 209 Clotting, propensity for, 202 Clotting studies, comment on, 209 Coagulability of blood, increased, 1 Cohort differences and ever users of oral contraceptives, 33 Coitus interruptus, 55, 56, 57 Coitus interruptus, frequency of use of, 87 College-educated women and age at pregnancy, 72 Collins spirometer, 192 Complaints and discontinuance of oral contraceptive use, 45 Condom, use of, the, 22, 54, 55, 56, 57, 58, 59, 60 condom, frequency of, 87 Continental mammography machine, 100 Contraception methods used by wom- en at risk of pregnancy, 52-57 Contraceptive drug composition and blood pressure findings, 106 Contraceptive Drug Study See Walnut Creek Contraceptive Drug Study Contraceptive Evaluation Branch (NICHD), mission of, iii Contraceptive failure, 68 Contraceptive methods, 55, 56, 58, 59, 60 Contraceptive operations Catholies, and, 70-71 prevalence, by live births number, 67 prevalence of, 61 social variables associated with, 66-73 Contraceptives, oral See Oral contraceptives Contributors to Contraceptive Drug Study, list of, 241 tabulated, Conversion to oral contraceptives’ use, 32 Corfman, Philip A., iii Coronary heart disease, 16 Coulter Counter, Model S, 96 C-Quens, 207, 216 Cultures, urine, 98 D Data analysis questions, 14 Data processing for research anal- ysis, 13-14 Deceleration in use of oral contra- ceptives, 41 Decile regressions in glucose-toler- ance tests, 159 Decrease in use of oral contracep- tives, 89 Degenerative processes and oral con- traceptive use, 186 Deladumone, 113, 114, 127 effect on blood pressure tabulated, 116 Demographic characteristics and de- pression, 215 Depression, mental, 17, 44, 98 clinically depressed patients and the MMPI Scale, 215 comment on, relative to drug users and nonusers, 225 current, 233 demographic characteristics, and, 215 discussion on history of, 235-239 drug exposure and, 219-225 drug regimen and, 219 extremely depressed patients’ drug discontinuance, 225 history of, a risk factor, 229-240 measure of current, 231 measure of history of, 231 measures of, 213, 214-216 measures of, product-moment, 232 months of drug use and, 224 mood scales, and, 231-232 mood scales, measured by, 215 multiphasic examinations and, 216 oral contraceptive types and, 213- 228 oral contraceptive should not be withheld, 239 246 INDEX Depression, mental—Con. oral contraceptive use and, tabu- lated, 218 oral contraceptives used by sub- jects investigated, 217 physiological variables and, 239 predisposition to, 226, 230 premenstrual, 227 progestational activity and, 226, 227 progestin, and, 214 progestin dose and, 222, 223 research on oral contraceptives and, 230 reversibility of, 214 sequential drug users, and, 219 severe, 215 subjects deleted from study of, 216 survivor effect, and the, 214 symptoms reported on, 213 two types of oral contraceptive and, 221 type and regimen of, 220 Deviation, mean and standard, glu- cose-tolerance testing, in, 150 Diabetes, 141, 165, 169 family history of, 182 risk factors of, 169, 182-184, 185 subclinical, 186 Diaphragm, use of the, 22, 54, 55, 56, 57, 58, 59, 60 Diaphragm, frequency of, 87 Diastolic and systolic blood pressure readings tabulated, 117 Diastolic blood pressure by weight and age tabulated, 124 hours since food and, 132 menstrual cycle, and the, 130 Dimethisterone, 8, 115-116, 178, 191, 216 Discontinuation of use of oral con- traceptives, 44-48, 181 Disease, unrecognized, 159 Dispenser for glucose, 92 Distribution of 1-hr serum-glucose levels, 175 Diurnal periodicity, 166 Diurnal variation and blood pres- sure, 128 . Nouche method use, 55, 56 Drugs, sequential-type, 177 Duffy, Thomas J., vii Duration of exposure to oral contra- ceptives, accumulated, 37-38 Education Depression Scale, and the, 216 oral contraceptive use, and, 22, 48, 49, 50, 85 religion and contraceptive opera- tions prevalence, and, 70 of wife and contraceptive opera- tions prevalence, 71-73 Educational status of women in health plan population, 26 Electrocardiograms, 99 Electrophoresis, protein, 94 Elevated blood pressure described, 108 Eligibility status in the use of oral contraceptives, 53 Emphysema, 189 Enovid-E, 207, 216 Enovid-5, 207 Epidemiologic studies may be fruit- ful, 210 Estradiol valerate, 113 Estrogen, 191 blood pressure, and, 110 dose and blood pressure, 113 effect confounded with progestin effect, 208 glucose levels and, 177, 179 progestin types tabulated, and, 8 type and thromboelastographic parameters, 207 users, AML examination of, 6 users’ statistics, 5 Estrogen-progestogen combination and glucose levels, 174 Estrogenic hormones, blood pres- sure status and, 106, 107, 112 Ethinyl estradiol, 113, 176, 178, 191, 197, 207, 208, 217 lung changes, and, 190 thrombogenicity, and, 202 Ethynodiol diacetate, 8, 174, 178, 191 Examinations, AML, 1 Expirogram, automated reading of, 93, 192 Exposure to oral contraceptives, du- ration of, 34, 35 INDEX 247 F Fajans-Conn criteria for glucose tol- erance, 143, 144 Family planning, 49 Fasting women glucose-tolerance levels in, 152 glucose-tolerance tests in, 153, 154 FDA Advisory Committee on Obstet- rics and Gynecology, 3 Fears regarding oral contraceptive use, 18 Fertility age group of oral contraceptive users, and, 39 control and surgical sterilizations, 83 control methods and socioeconomic characteristics, 2-73 control methods, frequency of vari- ous, 87 curtailment of, 49 impaired, 52 involuntary loss of, 48 Fibroid adenoma of breast, 46 Financial considerations and oral contraceptive use, 44 First pregnancy, woman’s age at, 72, 73 Fisch, Irwin R., vii Food Glucose-tolerance tests following intake of, 154, 155 hours since last blood pressure by, 132 glucose-tolerance tests and, 137, 147-168, 150-153, 156, 157, 166, 171 serum-glucose levels and, 158, 159 ingestion and blood pressure, 118, 119 Forced expiratory volume in spirom- etry, 192, 194, 195, 196, 197 Forced vital capacity in spirometry, 192, 194, 195, 196, 197 Freedman, Shanna H., vii Frequency distribution of glucose values, illustrated, 176 G Gas transfer and hormones, 197 Genetic disease and contraceptive operations, 66 Glucose challenge, urinalysis follow- ing, 97 Glucose intolerance no screening for, 186 obesity, and, 137 Glucose levels diabetes history and, 182 discontinuance of oral contracep- tive use, and, 181 Glucose load instead of test meal, 156 recommended, 137 Glucosemia, 150, 157 Glucosemia distribution, 185 Glucose tolerance abnormal, 184 contraceptive drug use and, 169 decreased, 1 deterioration caused by oral con- traceptives, 185 improved after oral contraceptive use discontinued, 184 loss with age, 157 Glucose-tolerance screening test, 135-145 Glucose-tolerance test, 7, 15, 91-92 abnormal standard type, 144, 145 age and the, 147-168, 151, 170 age range covered in testing, 166 age ranges in study, 149-150 age-specific serum-glucose levels in, 136 body size and glucose level, 138 classification of results of, 140 conclusions on study, 166-167 decile regressions in, 159 deciles of serum-glucose levels by age, 152 deviations in serum-glucose levels, 154 diabetes and, 142 diabetic risk factors and, 182-184 diabetics excluded from study using, 148 estrogen-progestogen combinations, and, 174 evaluation of, 135, 140 Fajans-Conn criteria in, 143, 144 fasting subjects, in, 139, 152, 153 139, 174, 248 INDEX Glucose-tolerance test—Con. followup procedure, 135, 137 food, and hours since See Food, hours since frequency of abnormal standards of, 144, 145 glucose concentration determina- tion methods, 136 glucose load, recommended, 137 glucose response homogeneity, 151 ketonuria, and, 147-168 ketonuria prevalence, and, 153 ketonurics and nonketonuries, in, 154 Ketostix test results tabulated, 163 material and methods used in, 170- 172 materials in study of, 148 mean 1-hr serum glucose by time of challenge, 164 months of oral contraceptive use and, 179 nonpregnant women, in, 136 nonusers and users of oral contra- ceptives, in, 175 obesity and, 184 oral contraceptive use and, 169-187 other health surveys, in, 157 percentiles of distribution of 1-hr serum glucose, 153 Ponderal Index used in, 183, 184 postchallenge values in, 164 regression coefficients, and, 150 screening limits and, 139 screening procedure, 136-137 screening test measuring a con- tinuous variable, 141 sensitivity and specificity of, 140 serum-glucose levels, regression with age, 172 statistical analysis, 149 study findings, 149-156 study methods, 148 subjects in testing program, 170 Sudbury study, in the, 143 test conditions variations and, 171 testing methods, 149 time of day, and the, 147-168 time of glucose challenge in, 171 Wilkerson point system in, 142, 148, 144 young adult norms and older per- sons, 159 Glucose value higher in oral contra- ceptive users, 172 Godart blood pressure apparatus, 99 Goldfein, Alan, vi Growth of American Families Study, 1960, 66, 68, T1 Gynecological checklist discontinued, 101 H Harleco Aqueous Standards, 92 Hartert Thromboelastograph, 94 Health hazards and use of oral con- traceptives, 86 Health plan members, 25-27 use of contraceptive drugs, and, 43 vasectomies, and, 62, 72 Health Services Reports, U.S., 147 Hearing levels, 2 Hearings on safety of oral contra- ceptives, 85 Height and weight and glucose load- ing, 140 Height and weight measurements, 7 Height of subject blood pressure, and, 114 spirometric values, and, 192-193, 196 Hewlett-Packard System, electro- cardiograms taken with, 99 Hi-Fi spirometer, 93 History of depression, measure of, 231 Hormone use and spirometry, 190, 195 Hormones, estrogenic, blood pres- sure, and, 105, 112 Hospital policy and tubal ligations, 70 Husband’s education and vasectomy prevalence, 72 Hypercoagulability, blood, 206-207, 209 Hypercoagulability estrogen dose, and, 210 Hyperglycemic effects of oral con- traceptives, 184 Hypertension, 15, 68, 121, 128 INDEX 249 Hypertensive disease risk, 1 Hypochondriasis, 98 Hysterectomy, 22, 62, 87 Hysteria, 98 I Income, median family, in health plan population, 26 Increase in use of oral contracep- tives, 42 Intelligence and the Depression Scale, 216 International Standardization Orga- nization, 93 Intraocular pressure, 7 Intrauterine devices frequency of use of, 87 increase in use of, 86 use of, 22, 55, 56, 58, 59, 60 Irritability, premenstrual, 215, 217, 218, 220, 221, 222, 223, 224, 225 Irritability, premenstrual, past users’, 225 K Kaiser Foundation Health Plan contraceptive drugs used by women in, 27 membership, 1 oral contraceptive users in, 21 participants in, 21 population information, 25-27 socioeconomic characteristics of participants in, 21 Ketones, urinary, routine testing for, 165 Ketonuria, 153-154, 160 carbohydrate deprivation and, 165 contraceptive drug use and glu- cose-tolerance testing, 148 glucose levels and, 167 glucose-tolerance test, 147-168 . hours since last food, and, 161, 162 prevalence charted, 160, 162 time of day and, 165, 166 Ketonurics and nonketonurics, glu- cose level deviations in, 163 Ketostix reagent strips, 136 Ketostix test data, 153, 154 Ketostix test results tabulated, 163 Ketostix tests, negative, 165 Kutner, Jerome S., vii and the, L Labstix, 97 Lactation cessation and blood pres- sure, 127 Lactation inhibition and blood pres- sure, 113, 114 Latex agglutination slide test, 97 Live births and voluntary steriliza- tion differences, 69 Lung changes and ethinyl estradiol, 190 Lung function tests See Spirometry Lung mechanics and hormones, 197 M Male sterilization See also Vasectomy prevalence in West, 65 voluntary, increase in, 88 Mammography, 100 Marriage age and study statistics, 34 Mastitis, chronic cystic, 46 Medical Care Program, Kaiser- Permanente, 1 Medical care source and contracep- tive drug use, 43 Medical records review, 11 Medical use of oral contraceptives by premenopausal women, 42 Medical users of oral contraceptives, 32 Medroxyprogesterone acetate, 8, 115- 116, 174, 178, 191, 216 Menopausal women taking oral con- traceptives, 54 Menopause, oral contraceptives used as treatment in, 32 Menstrual cycle, blood pressure and, 118, 129-130 Menstrual improvement caused by oral contraceptives, 213 Menstrual period and glucose-toler- ance testing, 177, 179 Mental depression See Depression, mental Mestranol, 7, 113, 174, 176, 178, 179, 191, 202, 207, 208, 216, 217 Microhematoerit method of test, 97 Minnesota Multiphasic Personality Inventory, 98-99 250 INDEX Minnesota Multiphasic Personality Inventory, Depression Scale of, 214, 215 Month of examination and systolic blood pressure, 119 Months of oral contraceptive use and depression, 224 Months of oral contraceptive use and " glucose levels, 179, 181 Mood complaint of oral contraceptive users, 213 Mood scales depression measured by, 215 questionnaire, in, 214 Moodiness, premenstrual, 214, 215 Multiparity and tubal ligation, 70 Multiphasic examination, depression measuring in, 233 Multiple contraceptive methods, fre- quency of use of, 87 Myatt, Albert V., vii Myocardial infarction, 16 N National Fertility Study (1965), 21, 22 comparison with, 42-44, 63-66 estimate resulting from, 42 figures on oral contraceptives used in Kaiser population, 43 Walnut Creek Kaiser population sterilizations and the, 66 National Health Examination Sur- vey, 157, 164, 166 National Health Survey blood pressure comparisons in, 114 mean blood pressure and the, 120, 121 Negroes, American, hypertension in, 128 Nelson, Senator Gaylord, 85 Nephritis, subacute, 68 Noncontraceptive use of contracep- tive drugs, 27-28, 36, 37, 38 Norepinephrine levels, 227 Norethindrone, 8, 115-116, 174, 178, 191 Norethynodrel, 8, 115-116, 178, 191 Norgestrel, 8, 115-116, 174, 191 Norinyl, 207 Norlestrin, 207, 216 Norquen, 207 0 Obesity glucose intolerance, and, 137 glucose-tolerance testing and, 184 Occupational level of health plan population, 27 Ocular tension, 92 Ogen, 197 One-hour glucose-tolerance test, 185- 145 Operations, contraceptive See also Tubal ligation; Vasec- tomy Oracon, 207, 216 Oral contraceptives accumulated proportion of those exposed to, 29 adverse reactions attributed to, 196 AML testing, and, 18 antitrypsin levels, and, 189-190 asthmogenesis, and, 190 blood pressure, and, 107, 108 blood pressure tabulations for users of, 124 body weight, and, 114 brand of, blood pressure and, 126 church attendance and, 51 clot formation time, and, 209 components of, 174 composition of drugs in depres- sion study, 217 composition variations in, 171 current users statistics, 27 data on users and nonusers in study, 5 decreased use of, 85, 89 degenerative processes and, 186 depression and, 213-240 depression causing, doubts about, 226 depression history a risk factor with, 229-240 depressive reactions to withdrawal from, 227 diabetes history and, 182 diabetogenic properties of, 185 diffusion of, 28 discontinuance and depression, 233, 238 discontinuance of use of, 22, 44- 48 INDEX 251 Oral contraceptives—Con. distribution of, U.S., 2 “drop-outs” in use of, 44 duration of use, blood pressure, and, 34, 35, 36-38, 109, 181-182 elevated blood pressure, and, 106 “eligibility to use contraception,” and, 52 first contraceptive method, as a, 57 first users of, 6 former users and never users com- parison, 57 frequency of use of, 87 glucose-tolerance relationships to, 169-187 glucose-tolerance testing and, 170 hearings on safety of, 85 hyperglycemic effects of, 184 hypertension, severe, and, 121 increased blood pressure, and, 105 increase in use over time, 39 last use of and depression, 234 long-term effects of, possible, 46 long-term sequelae of use of, 8 long-term users of, 38 medical advice on use of, 86 medical effects, assessment of, iii medical users and duration of ex- posure to, 37, 38 menstrual improvement, causing, 213 mood scales and users of, 215 past users of, statistics on, 44 Ponderal Index used in glucose level testing, and, 188, 184 pregnancy and blood pressure, 105-1383 pregnancy and depression and prior use of, 238 problems leading to discontinu- ance of use, 39 problems of study participants, 24, 25 pulmonary emboli and, 189 rapid diffusion of, 34-37 regional variation in use of, 43 removal of need for, 44 reninangiotensin system altera- tions and, 121 reversible effects on glucose toler- ance, and their, 185 sequential type, 191, 217 sequential type used in menopause, 32 side effects of, 1, 2 spirometric values and, 193 spirometry and, 189-200 statistics of use of in prestudy sur- vey, 4 status of use, age of subject and glucose levels, 173 status of use and glucose toler- ance, 172 steroid dose, and, 174 summary of research on, 230 thromboelastogram, and the, 201- 211 thromboembolic events, and, 202 time trend in use, 39-42 total months of use of, 46-48 total population exposure tabu- lated, 38 use of, 52-54, 172-174 patterns of, iii purposes other than contracep- tion, for, 27, 54 related to age of subject, 170 statistics on the, 27-38 surveyed, 83-86 Ortho-Novum, 207, 216 Orthotoluidine glucose testing meth- od, 91, 92, 136, 166 Ovral, 207, 216 Ovulen, 207 P Pap smear, 9 Papal encyclical and oral contracep- tion, 52 Paranoia, 98 Parity age and blood pressure, 118 blood pressure and, 118 contraceptive operation lence, and, 68-70 oral contraceptive use and, 48 systolic and diastolic blood pres- sure and, 125, 126 Pellegrin, Frederick A., vi, vii Personality testing, 98-99 Phillips, Nancy R., vii Physical examination questionnaire discontinued, 101 Pneumonitis, 190 preva- 252 INDEX Ponceau-S fixative dye solution, 94 Ponderal Index of glucose levels, 183, 184 Population characteristics of Kaiser Foundation Health Plan, 3-4 Postcoital douche frequency, 87 Postmenopausal women in contracep- tive study, 23 Postpartum women blood pressure in, 112-114, 126- 127 Deladumone and blood pressure in, 116 Postprandial screening test for blood sugar level, 141, 142, 164-165 Prednisone, 185 Pregnancy blood pressure, and, 107, 112-114 depression and, 231, 234 fear of, depression and, 226 oral contraceptives and blood pres- sure, 105-133 oral contraceptive use, and, 50 spirometric values in, 195 unwanted, voluntary sterilization and, 68 Pregnant women low blood pressure in, 105 spirometry in, 190 Premarin, 197 Premenopausal use of oral contra- ceptives, 42 Premenstrual irritability, 232, 234, ; 235, 236, 237 Premenstrual moodiness, 214, 232, 234, 235, 236, 237 measurements of, 218, 220, 221, 2292, 223, 224 Premenstrual scales, 215 Prevalence rates of common dis- eases, 142 Prior method of birth control used by oral contraceptive users, 54— 57 Profexray machine, 100 Progestational activity and depres- sion, 226, 227 Progestational component breakdown of drugs, 208 Progesterone and blood pressure, 112 Progestin, 191 blood pressure, and, 110 depression, and, 214, 233 dose data and depression, 235, 237 estrogen types tabulated, and, 8 severe depression and, 222, 223 type and dose, blood pressure, and, 115-116 Progestogen, 176, 178 Progestogenic menstrual cycles, 230 Protein electrophoresis, 9, 94 Protestant-Catholic ratio of health plan population, 26 Protestant women and oral contra- ceptive use, 51 Provera, 216 Provest, 207 Pseudopregnancy and oral contra- ceptives, 126 Psychiatric reactions to oral contra- ceptives, 230 Psychological assessment of study subjects, 9 Psychological scales in personality testing, 98 Psychopathic deviation, 98 Pulmodigitizer, 93, 190 Pulmonary emboli and oral contra- ceptives, 189 Pulmonary function test, 93 Pulmonary system and hormones, 196 Pulse and blood pressure recording, 99 Pulse rate, 15 Pychasthenia, 98 Q Quality control blood counts, in, 97 blood pressure and pulse record- ing, of, 99 glucose-tolerance testing, in, 92 Questionnaire history, 100-101 Questionnaire response rate, cumula- tive, 24 R Ramcharan, Savitri, vi, vii Reader variation in protein electro- phoresis, 94 Reagent strips used in urinalysis, 97, 98 INDEX 253 Reasons for discontinuation of oral - contraceptive use, 44-46 Red blood cells indices of, 16 membrane carrier systems of, 186 Religion and contraceptive use, 22, 43, 48, 50, 51 Religion and education and contra- ceptive operation prevalence, 70— 71 Religious affiliations of health plan participants, 26 Remedial operations prevalence, 62— 63, 65, 88 Renin-angiotensin system altered by oral contraceptive use, 121 Repression sensitization, 98 Responses to questionnaire in con- traceptive use study, 24 Reversibility of side effects of oral contraceptives, 3 Rh incompatibility, 66 Rhesus monkeys, oral contraceptive studies with, 185 Rheumatoid factor, 16 low prevalence of, 2 sera tested for, 97 Rhythm method use, 55, 56, 58, 59, 60, 87 Ryder, Norman, 64 S Safety of oral contraceptives, 85 Schizophrenia, 98 Schrogie, John J., vi Screening limits and glucose-toler- ance tests, 139, 142 Sera, chylous, 96 Sera, homolyzed, 96 Serum chemistry testing, 95 Serum electrophoretic fractions, 16 Serum glucose testing procedure, 91 Serum protein electrophoresis deter- minations discontinued, 94 Sex disinterest in, 44 guilt over, 239 Side effects causing oral contracep- tive discontinuance, 44 Side effects of oral contraceptive use, 48 Skinfold measurements, 100 Smoking and spirometric values, 194, 198, 199 Social factors associated with oral contraceptive use, 85 Social introversion, 98 Social variables and contraceptive operations, 22, 66 Somogyi-Nelson blood test, 141 Spermicidal agents, use of, 22 Spermicides, 55, 56, 58, 59, 60 frequency of use of, 87 Spirometer, wedge, 190, 192 Spirometry, 189-200 analysis limited in, 197 forced expiratory volume in, 192, 194, 195, 196 forced vital capacity value in, 192, 194, 195, 196, 197 height and weight of subject, and, 192-193 measurements in pregnant and nonpregnant women, 17, 199 oral contraceptive dosage levels, and, 196 oral contraceptives and spiromet- ric values, 196 pregnant subjects, in, 195, 199 smoking, and, 194 smoking index in, 198, 199 values and oral contraceptive use, 189, 192 variables used in, 192 Statements pertinent to discontinu- ance of oral contraceptive use, 44 Staub-Traugott effect, 164 Sterilization See also Contraceptive operations; Tubal ligation; Vasectomy; Vol- untary sterilization increase in male, 83 physician’s recommendations, and, 72 prevalence of, 57, 58, 63, 88 surgical, 22, 83 classification of, 57, 61 described, 86-87 frequency of use of, 87 prevalence in 1968, 86 Steroid components and glucose-tol- erance changes, 185 254 INDEX Steroid dose and drug components, 174-180 Steroid dose and glucose response, 169 Steroids and glycometabolic effect, 169 Steroids use and glucose tolerance, 181 Study subjects’ cohort selection, 4-7 Sudbury study, glucose-tolerance tests in the, 143 Sudden death, 46 Surgery, remedial sterilizing, prev- alence in West, 65, 66 Survey of members of Kaiser Foun- dation Health Plan, 3 Systolic and diastolic blood pressure readings tabulated, 117, 118 Systolic blood pressure hours since food and, 132 menstrual cycle, and the, 130 month of examination, and, 119 by weight and age tabulated, 123 T Tachycardia, 16 Tech Check used to check reagent strips, 98 Tecumseh Community Health Study, 157, 164, 166 Tecumseh Community Health Study, glucose tolerance data in, 164 Testosterone enanthate, 113 Tests, modifier attached to values, 96 Tests not done, 96 Thromboelastogram, oral contracep- tives and, 201-211 Thromboelastogram, whole used in, 95 Thromboelastography, 9. 16, 94 apparatus for, 201, 203 blood handling in, 204-205 clot kinetics studied by, 208 factor VIII activity, and, 209 fibrinolytic states and, 209 frequency distribution of param- eters in, 206 hypercoagulable, 202 index, thromboelastographic, 204 interpretation of thromboelasto- grams, 204 blood machine variability in, 205 measuring procedure in, trated, 203 multiphasic testing environment, in a, 209 parameters by estrogen type and dose, 202, 207 parameters in, 202, 205, 206 pregnant women excluded from tests, 205 reproducibility in, 205 shift toward hypercoagulability, 208 simple procedure, a, 202 study subjects for, 202-203 users and nonusers of oral contra- ceptives, in, 208 variability in, 205 Thromboembolie episodes, 86 Thromboembolic risk and estrogen dose, 202 Thrombogenesis, 210 Thrombophlebitis, 46 Thrombosis, 46 Thrombotic tendency and its causes, 202 Thrombus formation, 202 Thyrochron 100, 92, 93 Time of day blood pressure by, illustrated, 131 glucose-tolerance test, and the, 147-168 ketonuria and, 154-156, 160, 161 Time trend in oral contraceptive use, 39-42 ; Time trend statistics in oral contra- ceptive use, 41 Total months of oral contraceptive use, 46-48 Tubal ligation, 22, 46, 86, 87, 88 hospital policy and, 62 increase in prevalence of, 88 number of children prior to, 70 prevalence in Walnut Creek Kaiser population, 65 prevalence of, 62 Type of oral contraceptive and serum-glucose level, 178 U Urinalysis, 9, 97-98 Urine cultures, 9, 98 illus- INDEX 255 Urine, ketone bodies in, 154 Users’ and nonusers’ data analysis, 14 Uterine tumor, fibroid, 46 v Variations in glucose levels and oral contraceptive use, 184 Varicosities, 46 Vasectomy, 22, 46 age of wife, and, 62 Health Plan, and the, 72 husband’s education and, 72 more common in Western United States, 64 prevalence in Kaiser Health Plan population, 64 prevalence in non-Catholic men, 71 prevalence of, 62 Versatol, 92 Visual acuity, 7, 92 Voluntary sterilization, 52 educational status of couples, and, Jit) family size and, 68 prevalence and parity, 68 w Walnut Creek Contraceptive Drug Study attrition rate in, 13 categorization of subjects in, 5, 6 Caucasian study population in, 128 cohort followup in, 10-11 commencement of, 2 comparison with 1965 National Fertility Study, 42-44 contents of subsequent volumes to be published on, v, vi contributors to,. list of, 241 differences between users and non- users, 2 glucose-tolerance testing in, 91-92 laboratory procedures in, v losses of participants from, 1, 11- 13 number of women in, v observations during recruitment period of, v participants in, 1 period of, v personality testing in, 98-99 purpose of, 2 questionnaires used in, 10, 11 referral of subjects to, 4 subjects in, 2 subjects of unknown status in, 6 vasectomies prevalence and, 64-65 Walnut Creek Service Area de- scribed, 25-26 Wedge spirometer, 190, 192 Weight age, and systolic blood pressure, 123 blood pressure and, 114 height and glucose loading, and, 140 spirometric values, and, 192-193, 197 West, the, tubal ligation prevalence in, 65 Westoff, Charles, 64 White blood cell count, 16 White married women time trend in oral contraceptive use in, 39-42 use of contraceptive drugs by, tabulated, 43 Wilkerson point system used in glucose-tolerance tests, 142, 143, 144, 186 Withdrawal method use, 55, 56, 58, 59, 60 Women, 20-44 years of age, oral con- traceptives used by, 84 Working wives in health plan popu- lation, 27 v¢U.S. GOVERNMENT PRINTING OFFICE: 1974 0—535-967 U.C. BERKELEY LIBRARIES ARR (022321072