LIBRARY NEW YORK STATE VETERINARY COLLEGE ITHACA, N.Y. RC 1 92 G2T e " Universi,y Ubrar y W™imiffiS?J? d ,yphoid imm "nity, 3 1924 000 242 218 Date Due r~ '- ' ~ fyrtHh -fa^Mf* ; ip'i/'V & ^M Library Bureau Cat. No. 1137 Cornell University Library The original of this book is in the Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924000242218 MONOGRAPHS OF THE ROCKEFELLER INSTITUTE FOR MEDICAL RESEARCH No. 16 - May 10, 1922 TYPHOID CARRIERS AND TYPHOID IMMUNITY Omnis Typhus ex Typho By ABRAHAM L. GARBAT, M.D. Gat NEW YORK The rockefeller Institute for Medical Research 1922 1 Made in the United Slates of America en ^ monograph of the rockefeller institute for medical research, no. 16, may 10, 1922. TYPHOID CARRIERS AND TYPHOID IMMUNITY.* Omnis Typhus ex Typho. By ABRAHAM L. GARBAT, M.D. (From U. S. A. General Hospital No. 12, Biltmore, N. C, and the Lenox Hill Hospital, New York.) (Received for publication, July 7, 1919.) CONTENTS. I. Introduction and General Considerations. Importance of the carrier problem in civil life and in the army ... 3 Origin of the epidemic 4 Endo's medium: its preparation and use 5 Division of typhoid illness into two stages 6 II. Duodenal Cultures Versus Feces Cultures as a Means or Detect- ing Carriers. History of the "typhoid carrier problem" 8 Technique for obtaining bile 9 Characteristics of bile specimens 10 Technique for bacteriologic examination 11 Unreliability of feces examinations for detection of typhoid bacilli; accuracy of bile cultures 11 Constancy of positive bile cultures 14 Use of stool cultures for detection of certain carriers 21 Types of bacteria other than typhoid found in the bile 22 Typhoid Feces Carriers, with Special Reference to a Classifi- cation of the Various Types and Their Surgical Treatment. Classification of types of feces carriers 24 Frequency of various types 25 Diagnosis of intestinal carriers 27 Diagnosis of bile carriers 30 Ascending infection of the gall bladder and bile ducts 37 Surgical treatment of carriers 41 IV. Predisposing Factors to the Typhodd Carrier State with Special Reference to Cholecystitis and Cholelithiasis. Cholecystitis 47 Pathology of the gall bladder in carriers 51 Cholelithiasis , 51 * This essay was awarded the Cartwright Prize by the College of Physicians and Surgeons, Columbia University, June 4, 1919. 1 2 TYPHOID CARRIERS AND TYPHODJ IMMUNITY V. Typhoid Urine Cashiers. A. Irregularity and Intermittency of Bacilluria; New Method for Detecting Urine Carriers 55 Accurate method for detecting typhoid bacilli in urine. . . 70 B. Character of Urine; Origin and Frequency of Carriers. Importance of typhoid urine carriers 72 Chemical properties of urines containing typhoid bacilli. . 73 Source and origin of typhoid bacilluria 76 Classification of urine carriers 80 VI. The Complement Fixation Test for Typhoid Fever. Statistical, Clinical, and Experimental Studies. Complement fixation test during typhoid convalescence 82 Factors accounting for the persistence of a positive test 84 The complement fixation test in carriers 87 Experimental studies to prove the origin of the complement fixa- tion bodies 89 Comparison between Widal test and complement fixation test during convalescence 92 The complement fixation reaction as a diagnostic test during convalescence 93 Technique of the, complement fixation tests 100 Conclusions 104 Bibliography 108 I. INTRODUCTION AND GENERAL CONSIDERATIONS. Importance of the Carrier Problem in Civil Life and hi the Army. In spite of our increased knowledge of typhoid prophylaxis, typhoid fever is still a disease to be reckoned with. In the army the generalized use of antityphoid inoculations has undoubtedly prevented the occur- rence of large and serious epidemics; but isolated instances and small outbreaks (1) are still frequent. While official reports with the accu- rate numbers of cases of typhoid infection among the 2,000,000 men in the American Expeditionary Forces are not as yet available, the total will be below 2,000 cases. It is important for patients to realize that antityphoid inoculation per se is not an absolute safeguard against typhoid infection and that the observation of the other prophylactic measures is therefore necessary. Failures of vaccination to protect against the disease are in some instances explained by "mass infection;" i.e., the patients imbibe so large a number of bacteria at one time that the antibodies obtained by previous prophylactic injections are not sufficient to overcome the bacilli. In other patients, the disease may arise because the prophylactic inoculations do not stimulate the pro- duction of immune bodies. This failure of the tissue cells in certain individuals to react, is well known, and can usually be proven by the absence of agglutinating antibodies in the blood. A Widal test should therefore be made on every person receiving antityphoid inoculations. Whereas if antibodies have been stimulated it does not necessarily follow that a patient is sufficiently immune, it is highly significant of lack of immunity if the Widal test remains negative. Other factors, such as impotent vaccine, improper technique of inoculation, or infection by a special strain of organism, may also account for failure of adequate protection. In civil life, typhoid fever continues to merit even greater con- sideration. According to Gay (2), this disease remains the ninth contributing cause to the mortality statistics in this country and ranks fifth among the infectious diseases, being exceeded as a cause of death only by 3 4 TYPHOID CARRIERS AND TYPHOID IMMUNITY tuberculosis, pneumonia, infantile diarrhea, and diphtheria. Even as recently as in 1919, the death rate from typhoid in certain locali- ties was as high as 58.4 per 100,000 of the population (3). Although the average mortality of typhoid is usually given as 8 to 10 per cent, this percentage is much higher in some cities than in others; in one town, out of 281 cases there were 49 deaths, and in another, out of 35 cases there were 7 deaths. About 150,000 typhoid cases still occur each year in the United States, and this causes an annual production of approximately 7,500 carriers. When one realizes that according to some statistics (4) as many as 44 per cent of all typhoid cases are due to carriers, the economic importance of the carrier problem is readily estimated. It is therefore the duty of every physician not to discharge a typhoid patient until he is certain that the patient no longer harbors any typhoid bacteria in the excretions. It is with this object in mind that the study of duodenal cultures was undertaken. Origin of the Epidemic. During the recent war, the internment camp for German civilian prisoners was situated at Hot Springs, N. C, on the west bank of the French Broad River, about 38 miles from Asheville. The internment camp was divided into two areas known as Camp A and Camp B> The water supply proved to be inadequate for both camps so that a separate supply was developed for Camp B by digging a number of shallow wells along the river bank. When these wells also proved inadequate, direct connection was made with the river water (French Broad) and this water was intended for flushing latrines. River water was usually applied once or twice a day. When river water was pumped, chlorination was applied. Instructions were in force not to drink the river water. Examinations of the original well water showed it to be perfect. The water from the French Broad River, however, is polluted and always carries a great deal of suspended matter. The first case of typhoid fever occurred at Camp B about July 20, 1918. On Aug. 5 there were about 50 cases with several suspects. On Aug. 12, 13, and 25 all cases of typhoid and typhoid suspects, a total of 183, were transferred to U. S. A. General Hospital No. 12, at Biltmore, N. C, where the writer was stationed. All these cases came from Camp B, apparently due to A. L. GAKBAT 5 the polluted French Broad River water. Of these 183 cases, there were 16 deaths; only S cases were definitely proved not to be typhoid fever. The study of these cases 1 did not terminate as is usual with the cessation of the acute manifestations. While abundant reports are found in the literature referring to symptoms and laboratory findings during the active stage of typhoid fever, comparatively meager and inaccurate statistics exist pertaining to convalescent typhoid patients. Hardly any large epidemics have been studied in detail, as the work entailed in the care of very many, for example 200, patients under one roof at the same time, has in the past made detailed research studies almost impossible. Conclusions were therefore formerly drawn from collected series of cases. This must lead to somewhat erroneous statistics, as the findings are obtained by different workers, at different times, and with different techniques. It is therefore felt that the report to be given here merits consideration, since all the results were obtained by the same worker under the same circumstances. Through the cooperation of Colonel F. F. Russell the writer was given all the necessary laboratory assistance. As there are several questions which will be constantly referred to throughout this monograph, it is wise to discuss them here, in order to avoid repetition in the various chapters. Medium Used for Typhoid-Colon Differentiation. The Endo medium was chosen for all the differential work, as we found it very accurate and very simple to make up. In its preparation, several details were considered absolutely essential for its proper action. Sterile stock 3 per cent agar (made in the usual way from chopped beef) was kept on hand in quarter liter and liter flasks. The required quantity for the day's work was titrated just before use and the reaction adjusted to 0.2 per cent acid to phenolphthalein. While hot, the agar was treated with chemically pure lactose, 10 gm. per liter. The powdered lactose itself was added, or this quantity was preserved in a solution of sterile water in individual tubes and thus used. Next 1.8 cc. of fuchsin solution per liter were added to the hot agar. A 1 A clinical study will be published by the writer in conjunction with Dr. E. Henes. 6 TYPHOID CARRIERS AND TYPHODD IMMUNITY filtered saturated solution of basic fuchsin in 95 per cent alcohol was employed. The agar became deep red. As the last step, the sodium sulfite solution was gradually added until the hot agar was entirely decolorized. It usually required about 25 cc. to the liter. A 10 per cent solution of pure dry sodium sulfite crystals in sterile water was made up fresh every day. Since various fuchsin solutions and also various preparations of sodium sulfite may differ, the absolute quanti- ties given above may not exactly hit the proper balance in separate lots. These were approximate, however, and the proper balance can easily be obtained by a little preliminary testing in which the sodium sulfite solution is added to definite quantities of the fuchsin-colored agar in a test-tube. The finished product was poured into large sterile Petri dishes. The covers were left off until the agar was hard. While hot, a pale rose color was present in the medium which faded to a very faint pink or became almost colorless on cooling. Under no circumstances were the plates exposed to sunlight or daylight, as this colored the medium red. The Petri dishes were covered with a dark towel or kept in the ice box until used. If all these details are strictly observed, the author feels that Endo's medium will be found to give as reliable results as those obtained by the other differential media proposed. (Krystallviolett of von Drigalski and Conradi, malachite green of Loeffler, brilliant green of Krumwiede.) Recognition of Typhoid Bacilli. — After incubation for 18 to 24 hours the typhoid colonies stood out on the plates as the smaller colorless colonies amidst the deeply fuchsin-stained colon colonies. The suspicious (typhoid) colonies were transplanted upon Russell's double sugar agar slants and further identified by specific serum agglutination tests. In several instances even then the result was doubtful, so that recourse was taken to injection of the doubtful organisms into rabbits and to examination of the obtained specific sera for agglutinins against known typhoid strains. Division of Typhoid Illness into Two Stages. For the elucidation of certain statistics, the author divided the typhoid illness into two stages: the acute stage and the stage of con- valescence. The first day when the temperature reached normal and remained normal was taken as the dividing point between these two A. L. GARBAT 7 periods. The writer has always considered this a much more logical plan than the usual method of dividing the typhoid illness into 4 weeks. Those who have seen a great deal of typhoid fever realize how almost impossible it is to tell during the course of the disease what week of the illness the patient is in at a particular time. It seems much more rational to wait until the fever has subsided and then use the first day of normal temperature as the landmark. If it is desired to know at what stage of the disease a certain phenomenon occurred, it can very readily be told by calculating backward from the date of normal tempera- ture and be designated in terms of days or weeks either before normal temperature or after the patient went to bed with fever. Similarly in the study of typhoid convalescence, it is more practical to state that a certain test was positive or negative 10 or 15 days after "normal tem- perature" than to use the designation in terms of "day of disease," as "the 28th day of the disease." "Convalescence" is no longer "dis- ease" nor "illness." The course of typhoid fever is such a variable one that simply stating the "28th day of the disease," for example, associates with it no fixed idea of symptoms. In one patient, this day may coincide with the height of the fever stage, while in another it may coincide with subnormal temperature. The "duration of the disease" was estimated as the interval between the first and last days of fever. Usually, clinicians include in the "length of the illness" the time of incubation, but this period also is such a varying one that it seems wiser to go by the much more definite symptom of "fever" rather than the usual general guides of malaise or indisposition. The latter may be present a long time before the onset of typhoid fever and still not be dependent upon the infection. II. DUODENAL CULTURES VERSUS FECES CULTURES AS A MEANS OF DETECTING CARRIERS. History of the " Typhoid Carrier Problem. " The part played by human carriers in the spread of typhoid fever is a problem the solution of which was directly evolved from the views set forth by Robert Koch in his famous address of November, 1902, "Die Bekampfung des Typhus" (5). Even at that time, he con- sidered the typhoid patient or convalescent as the most fruitful source of further infection. This view was soon abundantly con- firmed by Frosch in 1903 (6), who had assumed the directorship of an experimental typhoid station at Trier, also by von Drigalski (7) at Saarbrucken, and Donitz (8) in Berlin. To Frosch belongs the credit for suggesting the hypothesis that the typhoid bacillus may lead a saprophytic existence in the intestinal tract, and it was von Drigalski who established on a bacteriological basis the hypothesis put forward by Frosch. Von Drigalski also recorded the first chronic carrier traced from convalescence onwards. In addition, he discovered the first female chronic carrier who gave no history of having passed through an attack of typhoid fever. With the recognition of the carrier state as a definite entity, various statistics appeared proving the validity of Koch's original dictum, and some authors have claimed that as high as 44 per cent of typhoid cases could be traced to typhoid carriers (4). It became of great importance to be able to detect the carrier by means of proper urine and stool examinations. For this purpose all types of special nutritive media were introduced, but, as will be seen presently, radical improvements upon these methods were necessary in order to obtain accurate results. In an article published by the author in 1916 (9) entitled, "Duo- denal Cultures in Typhoid Fever as a Means of Determining Com- plete Convalescence," 12 cases of typhoid fever were studied, in which comparative examinations made from the duodenal contents (bile) and the feces, demonstrated in 2 cases typhoid bacteria in the bile and not in the feces. In the present publication, the study of the bile as the source of typhoid bacteria in the stools of typhoid A. L. GARBAT 9 convalescents was resumed on a much larger and more detailed scale and the solution of a number of problems dealing with typhoid car- riers has been attempted. This paper deals with duodenal (bile) cultures made on 136 cases of typhoid fever. Technique for Obtaining Bile. The Einhorn duodenal tube was sterilized by boiling and given to the patient in the evening at 9 or 10, about 3 hours after the last meal. Various other times of the morning and afternoon were tried, but were given up, because the patient would thus miss a meal, and this often met with great objection. Only in rare instances was there any difficulty in having the patient swallow the tube because of gagging. Usually the tube would go down easily and by the following morning it had passed into the duodenum and bile could be aspirated with a sterile 10 or 20 cc. Luer syringe. It was necessary for the piston of the syringe to be a well fitted one, because at times suction had to be quite strong. The bile was collected in four or five separate sterile test-tubes; the first specimen usually contained mucus or other secretions that had collected in the tube during the night. This technique was found to be perfectly satisfactory and is some- what simpler than the one described by the author in 1916. In about S per cent of the cases, the tube remained in the stomach. This was probably due to gas- tric atony and the tests had to be repeated. When there was difficulty in obtaining sufficient bile, it was found that if the patient sat up in bed and bent his head forward between his knees, at the same time pressing upward on his abdomen with the palms of his hands, a flow of bile would set in. Or, if the patient was given a cold drink, the bile would appear, probably due to the reflex relaxation of the sphincter around the opening of the common bile duct. Occasionally, part of the fluid which the patient had drunk would be expelled through the pylorus into the duodenum so rapidly that on aspiration bile mixed with the fluid swallowed would be obtained. Thus, it is wise to give the patient sterile water to drink when this means of stimula- tion is employed. If secretion is obtained and one is doubtful whether it is bile or not, one can, at the bedside, give the patient some colored liquid to drink, grapejuice for example, and immediately thereafter aspirate. If the tube is in the stomach, the undiluted red-stained fluid is obtained on aspiration. In rare instances, if none of these methods yielded any bile, the author would inject S or 10 cc. of sterile fluid through the tube and immediately thereafter aspirate again. This fluid would return bile-stained from its contact with the duodenal wall. In the vast majority of instances, none of these extra manipulations were necessary, and sufficient bile was obtained without any trouble if only patience was practiced. 10 TYPHOID CARRIERS AND TYPHOID IMMUNITY Characteristics of Bile Specimens. (a) Quantity. — It was surprising to note the large quantities of bile that could be aspirated at the first examination. In only 16 cases was less than 10 cc. obtained. The small quantity does not, however, interfere with the finding of the typhoid bacilli, because a positive duodenal culture is recorded in one patient from whom only 2 cc. of bile and in another from whom only 6 cc. had been collected. In Table I the cases are tabulated according to the various quantities obtained. TABLE I. No. of cases. Per cent of cases. Quantity of bile obtained. cc. 16 14 Less than 10. 43 39 10-20 19 17 21-30 21 19 31-40 7 6 41-50 2 1 51-60 1 1 62 1 1 80 26* * Not estimated. (b) The consistency varied from thin, almost watery secretions, to thick mu- cous fluids which came through the tube with difficulty. The former were usually absolutely clear, while the latter were turbid or entirely opaque with, a varying amount of a dirty granular sediment. (c) A variety of colors was noted; very light yellow, canary yellow, golden yellow, light green, dark green, light brown, dark brown, etc. (d) On microscopical examination, the clear fluids showed only a moderate number of bacteria, or none at all. The clear fluids were usually sterile. The turbid fluids contained a great deal of mucus mixed with a moderate or great number of leucocytes and numerous bacteria, usually bacilli. Frequently myelin drops were noted. A Gram stain showed mostly Gram-negative bacilli mixed with some Gram-positive cocci. No absolutely characteristic picture can be ascribed to those bile specimens which contained typhoid bacteria. While almost all of these fluids were thick, very turbid, and contained a good deal of mucus, the same findings were also present in those specimens which on culture showed colon bacilli. As a rule, however, the latter contained very many more leucocytes. In only one instance out of this entire series were actively motile typhoid bacilli (as proven later by culture) noted in hanging drops of the fresh bile specimens. A. L. GARBAT 11 Technique for Bacteriologic Examination. Endo's medium was selected for typhoid-colon differentiation as it was found least complicated and very satisfactory. The details for its preparation are given under the general discussion. (a) The bile collected in each test-tube was examined separately. A broth culture and an Endo plate from each were prepared, thus making 4 broth tubes and 4 Endo plates from each patient. About lto 2 cc. of bile were added to the broth tube, and about 0.25 to 0.5 cc. was spread over a large Endo plate. No special dilutions were found more advantageous than others. The broth tube was plated after 24 hours incubation. At the Walter Reed Hospital where duodenal cultures were undertaken, the original bile was first incubated for 24 hours and then plated. (6) The stool examination was made by the usual technique; namely, a piece of feces, about the size of a split pea, was rubbed up in about 10 cc. of broth, allowed to stand for about 15 minutes until the heavy parts had settled, and then a large loopful from the upper layer of the fluid was spread carefully over the entire surface of a hardened plate of Endo's medium. In the routine stool examination, two such plates were made from each specimen of feces. The problems which this study attempted to solve are so varied that possibly the best way of expressing the results is by stating the particular question under investigation followed by the answer as interpreted from the findings. 7* the Repeated Examination of the Feces a Fair Index as to the Absence or Presence of Typhoid Bacilli in the Intestines! Almost all clinicians and bacteriologists agree that when a stool is reported as "no typhoid bacilli found," this statement in many instances may mean nothing more than that the laboratory methods were such that the typhoid bacilli were not detected. As evidence that bacteriologic examination of the stool for typhoid bacteria is attended with uncertain results, one has but to note the different methods and the new media constantly devised for such examination, in the hope of improving the results. In our series of cases, we adhered strictly to the army rule of not considering a patient free of typhoid fever until 3 consecutive stools and urines examined at intervals of 6 days proved to be negative. The first specimens were sent to the laboratory when the patient's temperature was nearing normal, or had become normal. In 136 patients, bile cultures were made after 3 consecutive stool examinations had been found negative; i.e., apparently free of typhoid 12 TYPHOID CARRIERS AND TYPHOID IMMUNITY bacteria. Out of these 136, 20, or 15 per cent, still showed typhoid in the bile. Of these 20, 15 had never shown typhoid colonies in the feces, while in 5, typhoid bacteria had been demonstrated in the stool cultures at some time previously but had apparently cleared up so far as could be judged by the 3 consecutive negative feces cultures. Even more striking is the fact that 15 of these 20 patients, or 11 per cent of the cases examined, had not only 3 negative stools, but also 3 negative urines to their credit when the positive growth in the bile was obtained. In other words, 11 per cent of the typhoid patients who are usually considered ready to leave the hospital and mingle with the community at large, may still be discharging typhoid bacilli unde- tected by the general method of stool examination. TABLE II. Name. Poelman Bergenthal. Dura- tion of original infec- tion. days 49 44 1st day of constant normal tempera- ture. Sept. 21 " 23 Date of 3rd negative urine and stool cultures. Oct. 14 Sept. 30 Date of positive duodenal culture. Oct. 14 " 3 Bate of . onset of relapse. Oct. 16 " 5 Date of positive blood culture during relapse. Date of normal tempera- ture after relapse. Oct. 22 " 8 Nov. S Oct. 31 Date of negative duodenal culture after relapse. Nov. 8 Jan. 6 The remaining 4 per cent of the cases (5 patients) with positive bile cultures and negative feces cultures had other evidences of the per- sistence of typhoid besides the positive duodenal culture. Two still showed typhoid bacteria in the urine. Another had recurrent rises in temperature with tenderness over the gall bladder but a negative blood culture, probably recurrent attacks of cholecystitis. The other two developed definite relapses with positive blood cultures, after normal temperature for 25 days and 12 days respectively. In spite of 3 negative urines and stools these 2 patients began to run temperatures several days following the time of the positive bile culture (Table II). It is interesting to consider the hypothesis that some typhoid re- lapses may be due to the persistence of bacteria in the gall bladder as the source of a vicious circle. There is, first, the primary bacteremia; next, the settling of the bacteria in the gall bladder; then a reabsorp- tion of the bacteria either from the gall bladder or from the intestines, A. L. GARBAT 13 giving a renewed bacteremia and relapse in those patients in whom a sufficient degree of immunity has not taken place from the original infection. Thus a positive duodenal culture during the early convales- cence may be compared to a strongly positive Wassermann test in a lues asymptomatica — a danger sign of possible trouble. All relapses cannot, however, be explained thus, because there were instances where relapses occurred after a negative bile culture. Further in- vestigation along this line is necessary. The bile cultures were not always made on the same day as the third negative feces. The different periods of time which elapsed between the last negative feces and the positive duodenal culture are shown in Table III. table m. No. of cases. Interval between last negative feces and positive bile cultures. No. of cases. Interval between last negative feces and positive bile cultures. days days 5 3 7 1 1 1 9 2 2 1 11 5 3 1 12 1 6 From this table it has been estimated that had bile cultures not been made and were the patients not required to observe the usual typhoid precautions after they had 3 negative urines and 3 negative stools, IS patients would have been excreting typhoid bacteria in their feces over a total period of 68 days. These figures are comparatively mild when one considers the length of time that these bile cultures continued positive after the third negative stool. The Army rule of keeping a patient at the hospital until 3 consecu- tive negative urines and stools at intervals of 6 days are reported, is a comparatively strict precaution. Many health boards and hospitals in this country do not adhere to such a wise requirement. A few years ago, systematic inquiries disclosed that out of 23 leading hospitals in this country, only 9 required an examination of both the urine and stools of their typhoid patients before discharging them from the hos- pital. One institution examined these excretions only when the pa- tient's occupation brought him into contact with foods. Another exam- 14 TYPHOID CARRIERS AND TYPHOID IMMUNITY ined the urine and not the feces; one even discharged patients with bacteria in the urine if these persistently remained there. 11 institu- tions disregarded such examinations altogether; 5 of these definitely stated that the technical phase associated with the inaccuracies of the findings in stool examinations did not warrant the time spent. The author has often felt the same way about the results of stool examina- tions and it is that sense of inaccuracy which led to the study of bile cultures. It is very discouraging to note that the 1920 New York City sanitary code regulations require that, in typhoid fever, quarantine should be continued until 10 days after the patient's temperature reaches normal and further until 2 specimens of feces collected at an interval of at least 24 hours are found to be free of typhoid bacilli (10) . Let us hope that such a dangerous state of affairs, as evidenced by the cited statistics, will be remedied. Is It Possible that Typhoid Bacilli Are Destroyed on Their Way Down from the Gall Bladder, Thus Actually Giving a Negative Stool and a Positive Bile Culture? This question was answered in 9 cases. In 2, after an interval of 6 days, the fourth specimen of feces came to the laboratory as part of the regular routine. Typhoid colonies were found. No special dilutions of the feces nor an increased number of Endo plates were necessary for their discovery. In 4 other patients with positive duodenal cultures, subsequent stools were especially reexamined, and, when 10 to 15 Endo plates from various dilutions of the feces were made, typhoid bacilli were recovered. Furthermore, it was found that in some of these patients with positive duodenal cultures and negative stools, the longer the typhoid bacteria persisted in the bile, the more likely was one to detect the typhoid colonies in the stools. Thus in 3 patients in whom the typhoid bacilli persisted in the bile for over 3 months after normal temperature, isolated typhoid colonies could be detected in the stool plates with little trouble at the end of this time; whereas, in the early days of convalescence, typhoid colonies were either not detected at all in the regular routine stool examinations or only after repeated and numerous cultures. At the same time, we meet with the diffi- culties of stool examinations even in chronic carriers. In one instance A. L. GARBAT 15 of a permanent typhoid carrier, a heavy growth of pure typhoid bac- teria was isolated from the duodenum on two successive occasions. During the same time it required 60 Endo plates (10 made on each of 6 successive days) to demonstrate the typhoid bacilli in the feces. The first 4 days (40 Endo plates) revealed no typhoid bacteria. On the 5th day 4 out of the 10 plates showed 3 or 4 typhoid colonies per plate. The 6th day revealed no typhoid bacteria. While it is shown that there is a great discrepancy between the findings of duodenal cultures and feces cultures, the typhoid bacilli are not de- stroyed in the intestines but they are either absent in the particular drop or drops of feces that are examined or through the use of the usual technique they are overgrown by the colon bacillus. In the bile, the typhoid bacilli are more concentrated. Most authorities have been of this opinion, although definite proof in vivo was evidenced only by this method of duodenal cultures. At post-mortem, von Drigalski (11) and Jurgens (12) showed by cultural methods that in the intestinal tract from the duodenum down to the rectum, the number of typhoid bacilli decrease. In the duodenum and the upper portion of the jejunum one frequently meets with enormous numbers of typhoid bacilli in nearly or actually pure culture. On What Day of the Disease May One Expect the Duodenal Cultures to Become Negative? As was said in the general discussion, it is wise to designate the first day of normal temperature as the dividing point between acute disease and convalescence, so that it is more accurate to put the question: How soon after normal temperature may one expect the bile cultures to be negative? In order to have deduced any such fixed periods of time, it would have been necessary to do repeated cultures from the bile on the same patients at different intervals, beginning before the temperatures had reached normal. Since the main object of the present study was a determination of the comparative value of duodenal (bile) and stool cultures, the above plan was not adopted. The duodenal cultures were done at various intervals after the onset of normal temperature and it can be said that no definite conclusion was reached as to when the bile becomes free of typhoid bacteria. Table IV gives the number of cases and the 16 TYPHOID CARRIERS AND TYPHOID IMMUNITY time at which the duodenal examinations proved negative for typhoid bacilli. About 40 per cent of typhoid cases had negative duodenal cultures within the first monthafter the temperature reached normal. Had it been possible to take bile cultures on the other patients at a date previous to the times stated above, this percentage would undoubtedly have been higher. In some cases, duodenal examinations were made even before the temperatures had become normal, and in others after only 1 or 2 negative urines and stools; and at these times, the bile was already free of typhoid bacteria in some instances. It can, therefore, be stated that the period of time after normal temperature when the bile becomes free of bacteria is a varied one. Patients may be dis- TABLE IV. No. of cases. Fer cent of cases. Interval after normal temperature when duodenal culture was negative. days 7 6 10 13 11 11-20 34 22 21-30 42 37 31-40 17 13 41-50 2 51-60 1 61-70 charged before 3 stools have been examined and found negative as long as the duodenal cultures show no typhoid bacilli. Such direct examination is much more accurate than stool examinations. It is much more important to analyze the time relationship of the positive duodenal cultures. Only 20 of our cases (15 per cent) showed positive bile cultures during convalescence, but it must be remembered that bile cultures were only started after 3 stools from a given patient had already been examined at intervals of 6 days and reported as negative. Had duodenal cultures been instituted at an earlier date, it is more than probable that a greater number of positive cultures would have been obtained. The presence of typhoid bacteria in the bile during the acute stage of typhoid fever is a different problem. Table V shows the number of days between the onset of normal A. L. GARBAT 17 temperature and the first -positive bile culture. Here we see that in 10 cases (7.5 per cent) the duodenal cultures were positive later than 1 month after the beginning of normal temperature. table v. No. of cases. Interval after normal temperature when duo- denal culture was positive. No. of cases. Interval after normal temperature when duo- denal culture was positive. days days 4 1-10 8 31-M) 3 11-20 1 51-60 3 21-30 1 61-70 How Long Did These Patients Continue to Harbor Typhoid Bacteria in the Bile? 4 cases, or 3 per cent of the typhoid patients, showed typhoid bacilli in the bile 4 months after the fever had ceased (Table VI). It is interesting to refer to the findings in the stool cultures of these carriers at the same date (Table VII). It is noted that even in pronounced carriers duodenal cultures are much more reliable than stool cultures. Naturally, once it is ascertained or even suspected that a particular convalescent typhoid patient is excret- ing typhoid bacilli in the feces, continued patience with the stool examinations expressed by large numbers of cultures may TABLE VI. 1st duodenal culture. 2nd duodenal culture. . 3rd duodenal culture. 4th duodenal culture. Name. Interval after normal temper- ature. Result. Interval after normal temper- ature. Result. Interval after normal temper- ature. Result. Interval after normal temper- ature. Result. Ericksen days 29 11 30 14 9 62 8 + + + + + + + days 63 39 50 55 42 74 108 + + + + + + + days 97 69 81 82 72 90 111 + + + + days 128 106 122 111 109 127 143 + + Hoffman + Ziska i Gutte + 18 TYPHOID CARRIERS AND TYPHOn) IMMUNITY demonstrate the bacteria. However, one has no guide as to those who are more likely to become typhoid carriers. And since it is impracticable to undertake a long special feces investigation on every patient, particularly when one is dealing with a typhoid epidemic, main reliance should be placed upon bile cultures; for, as has been seen, the usual routine stool examinations during con- valescence missed IS per cent of carriers, and of these, 3 per cent were carriers for over 4 months, possibly permanent carriers. TABLE VII. Ericksen. Schaefer. Hoffman. Lehman. No. of Endo plates made 8 8 6 60 from feces. No. of plates Feces. < showing ty- phoid colonies. Proportion of ty- 8 4 4 phoid to colon lto75 ltoSO ltoSO colonies on or 100 or 75 or 75 , these plates. No. of typhoid 340 to 450 5 to 10 350 to 500 40 to 50 colonies on En- colonies in colonies in colonies in colonies on Bile. j do plates from pure culture pure culture pure culture each plate. bile of same on each on each on each date. i plate. plate. j plate. How Numerous May the Typhoid Bacteria Be in the Bile? In general, if the typhoid bacteria are present in the bile in great numbers, many typhoid colonies are seen on the original Endo plate made directly from several drops of the bile. On the other hand, if the number is not so great, the growth is obtained principally in the broth and only very few colonies or none at all are noted on the original Endo plate. Thus, 2 patients who showed a heavy growth on the original Endo plate at the first culture showed a growth only in the broth and not on the Endo plate at the reexamination of the bile 1 month later. Further examination after several weeks' interval, gave no typhoid bacilli even in the broth culture. In 3 A. L. GARBAT 19 other patients the growth was just as heavy at the second and third examinations as at the primary culture. It was difficult to get accurate figures concerning the actual number of colonies in the bile because pour plates fromEndo's medium do not distinctly differ- entiate typhoid from other types of organisms. Approximate esti- mations were made by spreading 0.5 cc. of the bile very carefully over the entire surface of an Endo plate. In many instances the colonies were in such great numbers that they were confluent and difficult to count. As many as 32,000 colonies per cc. have been estimated by this method and this in a man (Ziska) in whom the stool culture was negative. It must be remembered that this number of colonies would represent a far greater number of bacteria. In the Cases with Positive Duodenal Cultures Are Typhoid Bacilli Constantly Present in the Bile or Is the Excretion an Intermittent One? A positive solution to this question is important in order: (1) to establish whether there was any best time for culturing the bile; (2) to explain the apparent irregularity in the excretion of the typhoid bacilli in the feces; (3) to ascertain whether 1 negative duodenal cul- ture was sufficient to exclude the existence of a carrier condition. 1. The question of hourly intermittency was studied. Two carriers (for a pecuniary consideration) retained the duodenal tube for 24 hours without taking food, and specimens of pure bile were examined every 2 hours. After each sample was collected, sterile distilled water was forced through the tube in order to prevent the bile from stag- nating and mixing with the next specimen. It was found that typhoid bacilli were constantly present, as is noted in Table VIII. 2 and 3. When duodenal cultures were repeated in carriers at various intervals over a longer period of time, it was found that as a general rule cultures would either consistently continue positive, or if they once became negative they would remain so. This is very different from the feces findings in these carriers. Bacteria are often reported to be absent from the feces for months, only to recur again later on. It seems to be almost definitely established in the literature that the excretion of typhoid bacilli in the feces is intermittent. 20 TYPHOID CARRIERS AND TYPHOID IMMUNITY TABLE VHI. Name. Bile obtained. Ty- phoid present. Time. Quan- tity. Transparency. Color. Haak a.m. 9.00 11.00 p.m. 2.00 4.00 6.30 9.00 a.m. 9.30 11.30 p.m. 2.00 4.00 6.30 9.00 cc. 4 12 13 16 21 6 3 10 9 15 8 22 Turbid. Clear. tt Slightly turbid. Clear. tt it tt Slightly turbid. Clear. it it 1 Yellow-green. « « « It u tt it ti it 11 it tt ti it tt Orange. Yellow-green. « ti Orange-yellow. -f- + ++ ++ ++ ++ + + + + + + Semple and Greig (13), who published a thorough bacteriological investigation of typhoid convalescents, show that carriers may be readily overlooked if the routine feces examination is limited to 3 or 4 cultures during the 6 weeks following defervescence. "K linger (14) published most interesting results in this respect. Date. Stool cul- ture. Date. Stool cul- ture. Date. Stool cul- ture. Date. Stool cul- 'ture. Date, Stool cul- ture. 1904 1904 1904 1904 1905 Jan. 30 — May 3 - Aug. 10 — Oct, 11 + Feb. 13 Feb. 10 + " 10 - " 16 — " 29 " 18 + " 11 + tt 19 + Sept. 12 + Nov. 2 + Mar. 24 + Mar. 26 — June 8 + Oct. 1 — " 24 + Apr. 26 + Apr. 3 + July 2 + " 5 — Dec. 22 + May 12 + + indicates typhoid bacilli present; -.typhoid bacilli absent. The writer is convinced that this intermittency is only apparent; were duodenal cultures made during those periods when the stool is apparently free from bacteria, typhoid bacilli would undoubtedly be A. L. GARB AT 21 present in the bile. There are, however, exceptions to this uniformity of excretion in the bile. Occasionally a negative duodenal culture was interspersed among positive ones. This intermittency was rare and the negative findings would not persist for any length of time; often as early as the next day a positive culture would again be noted. No definite explanation can be offered for this variation. In two patients there was an associated cholelithiasis, so that possibly the stone ob- structed the cystic duct and the bile at the time of culture was ob- tained from a non-infected liver. That such conditions exist was proved by operation and by postmortem examination, and they are reported by the writer further on. In another case, an occasional negative culture amongst numerous positive ones was noted after a cholecystectomy failed to cure the carrier state. In these liver in- fections, intermittency can only arise if the particular negative specimen happens to be drained from a part of the liver which was not infected. Taking into consideration such occasional irregularities, the author requires two consecutive negative bile cultures made at intervals of one or more days instead of a single negative bile culture, before a patient is safely considered free of typhoid bacilli. This precaution applies especially to three groups of cases: (1) those who once showed a positive typhoid culture in the bile; (2) those in whom a surgical operation (cholecystotomy or cholecystectomy) had been performed for the cure of the carrier condition; (3) those who during the acute illness or convalescence manifested symptoms of cholecystitis or cholelithiasis. As will be seen later on, it is these patients who usually become carriers. Should Stool Examinations Be Entirely Discarded in Favor of Bile Cultures or Is There a Type of Pure Intestinal Typhoid Carrier Whose Bacteria Do Not Lodge in the Bile but in the Intestine and Are Therefore Detected Only by Stool Examinations? The existence of a pure intestinal carrier has been a disputed question but a very important one, especially from the standpoint of surgical therapy; for excision of the gall bladder in these cases will not effect a cure. Of our entire series of 164 typhoid cases studied, there was only one patient who from the very beginning of his convales- 22 TYPHOID CARRIERS AND TYPHOID IMMUNITY cence, 5 days after normal temperature, began and continued to show typhoid bacteria in the stools in almost pure culture. In spite of this, 4 duodenal cultures, were made during the first 2 months after normal temperature, and in none of them were typhoid bacilli isolated. The only bacteria found were colon bacilli. There was nothing unusual in this patient's clinical history. He was. 54 years old, and reported sick Aug. 2, 1918, with the usual symptoms of head- ache, lassitude, and fever. He had one typhoid inoculation on July 26, 1918. He ran a comparatively mild typhoid, with fever for approximately 3 weeks. There were no complications and he con- valesced without interruption. One cannot deny that types of pure intestinal carriers do exist, although they comprise only about 2 per cent (1 in 53) of all typhoid carriers, and only about 0.6 of 1 per cent of all typhoid patients. A detailed study of the different types of carriers is taken up in Section III, but it may be said here that an intestinal carrier is very readily differentiated from a bile carrier by stool examinations. In the former, there is no difficulty in detecting the typhoid bacteria. They begin to appear in the stools early during convalescence, and are found practically in every specimen in very great numbers and in almost pure culture. In the bile carriers, if typhoid colonies are found in the stool cultures at all, they appear there with great irregu- larity (in 1 out of 3 to 5 cultures) and in fewer numbers than in in- testinal carriers. For example, on the Endo plate made from the stool of a bile carrier, the ratio between the typhoid and the colon colonies would be approximately 1 to 25 or 75, while in the intestinal carrier this proportion would be reversed. In the latter, it seems as if the flora of the intestine are changed from colon to typhoid. A more detailed discussion of this topic is undertaken in Section III on Typhoid Carriers. What Other Types of Bacteria Were Found in the Duodenal Cultures, either Alone or in Association with the Typhoid Bacillus? Out of 39 duodenal cultures which showed typhoid bacteria, 32 (82 per cent) were pure cultures, 4 (10 per cent) were mixed with colon bacilli, and 3 (8 per cent) with the Staphylococcus albus. The more persistent carriers had pure cultures. However, even in the A. L. GARBAT 23 mixed cultures the typhoid colonies stood out prominently on the Endo plates and were isolated with no difficulty. Table IX shows other bacilli found in 132 cultures which did not contain the typhoid bacillus. TABLE IX. No. of cases. Per cent of cases. Bacteria found. 42 32 Colon bacillus. 13 10 Staphylococcus albus. 5 4 " aureus. 12 10 and colon bacillus. 8 6 An intermediate 2 0.1 Bacillus proteus. SO 38 Sterile. The items of interest in the above figures are, first, the high per- centage of pure typhoid growths, and, secondly, the high percentage of sterile cultures. The last was especially surprising in view of many obstacles which were encountered and which interfered with the usual rules observed in taking sterile cultures. in. TYPHOID FECES CARRIERS, WITH SPECIAL REFERENCE TO A CLASSIFICATION OF THE VARIOUS TYPES AND THEIR SURGICAL TREATMENT. Temporary Versus Permanent Typhoid Carriers. Every typhoid patient who continues to show typhoid bacteria in the excreta (urine or feces) after the acute disease is over, should be classed as a "carrier." This carrier state may ultimately cease, in which case the individual is considered a "temporary carrier," or it may continue the rest of the patient's life and then the individual is classed as a "permanent carrier." Temporary carriers are usually such for only weeks or months and then clear up; occasionally they persist for a year or even longer and then they are known as "chronic" carriers. It is safest, however, to consider every carrier a permanent one until he is proved otherwise. It is difficult to determine how long a time must elapse before a temporary carrier can be classed as a chronic or probably permanent one. It was found that those patients who showed typhoid bacteria in the stool for longer than 3 to 4 months after normal temperature, usually became chronic or permanent carriers. Diagnosis of Feces Carriers. We advise the use of the general term "feces carrier" to include all patients who show typhoid bacteria in the feces, rather than the term "intestinal carrier," because the latter designation should be reserved for a special class of feces carriers. The finding of typhoid bacilli directly in the bile has formed our basis for the designation of feces carriers. This method determines much more readily than stool cultures the probable presence of typhoid bacteria in the intestinal tract, especially during the early part of convalescence from typhoid fever. In our series of 136 cases, 13 feces carriers would have escaped detection if duodenal cultures had not been made. Cultures of the stool alone may suffice for the detection of "chronic" or "permanent" feces carriers because in 24 A. L. GARBAT 25 these the typhoid bacteria have reached to great numbers in the intestines. During the first weeks or months after normal tempera- ture, however, and this is just the important time for the detection of the carrier, it has been demonstrated that, while the bile may be full of typhoid bacteria, few or no typhoid colonies will be seen in the routine stool cultures of some patients. No feces carrier should be dismissed as having been only a tem- porary carrier until repeated duodenal cultures show no typhoid bacilli. It is fully realized that this method is more troublesome for the patient than stool cultures; on the other hand, it is much safer for the community. In the following study of 164 typhoid patients, at least 4 stool examinations by the Endo method were made on each patient. In addition, duodenal cultures were made on 136 of them. It is to be kept in mind that these tests were only started during convalescence, after the patient was free of fever and had remained so; in other words, at a time when the laity and some physicians have a tendency to leave off the usual precautions taken during the acute stage of the infection in order to prevent its spread. Frequency of Various Types. Of these 164 typhoid cases, 39, or 21 per cent, showed typhoid bacilli in the feces during convalescence. In addition, there were 14 cases in whom 3 consecutive stool examinations proved negative, while duodenal cultures showed numerous typhoid bacteria. Thus, adding these together, we find that 53, or 32 per cent, continued as feces carriers after the temperature reached normal. This is a very high percentage in comparison to the figures reported by other writers. per cent Schroder (15) 7.9 von Drigalski (11) 11.0 Lentz (16) 4.0 Klinger (17) 11.6 Semple and Greig (13) 11.6 26 TYPHOID CARRIERS AND TYPHOn) IMMUNITY The typhoid bacilli persisted in the stools of these patients for vary- ing lengths of time. Results in the 53 positive cases are shown in Table X. TABLE X. No. of cases. Per cent of all Per cent oi all Duration of carrier state positive cases. typhoid cases. after normal temperature. weeks 3 s.s 1.8 1 9 17.0 5.4 2 . 10 19.0 6.0 3 7 13.0 4.2 4 4 7.0 2.4 5 3 5.5 1.8 7 6 11.0 3.7 8 2 3.7 1.2 9 1 2.0 0.6 10 1 2.0 0.6 11 1 2.0 0.6 12 2 3.7 1.2 17 1 2.0 0.6 18 1 2.0 0.6 19 2 3.7 1.2 21* Summary in terms of months. months 29 55.0 17.4 i 13 24.0 7.9 2 5 10.0 3.0 3 6 11.0 3.6 4* .- *And longer. At the time of writing, the typhoid bacilli had cleared up in all but 7 carriers who had been observed for 12, 17 (2 cases), 18, 19, and 21 (2 cases) weeks respectively after normal temperature. Judging by the stool cultures, surely 4, if not all of them, will un- doubtedly remain chronic or permanent typhoid feces carriers; this makes 2.4 per cent to 4.2 per cent of all typhoid cases. These figures correspond with the findings of others. A. L. GARBAT 27 per cent Hetsch (18) 4.62 Fornet (19) 0.9 Kayser (20) .5.0 Park (21) 5.0 Kirchner and co-workers (22) 5.0 Frosch (23) 2.47 Mayer (24) 4.0 TYPES OF FECES CARRIERS. The gall bladder, or rather the bile coming from the gall bladder, has been conceded by all as the source of the typhoid bacteria in the feces. By means of duodenal cultures, a finer classification is possible. Feces carriers may primarily originate as one of three types: (a) intestinal carriers, (6) gall bladder (bile) carriers, (c) liver or duct (bile) carriers. Intestinal Carriers. An intestinal carrier is one in whom the typhoid bacteria in the feces come primarily from the intestines and not from an infected gall bladder or liver by means of the bile. The existence of a true in- testinal type of carrier has never been definitely accepted. It is even denied in the recent article of Nichols (41). A negative proof was presumably offered by those cases in the literature in whom chole- cystectomy with excision of the cystic duct, had failed to cure the carrier state. Such evidence in favor of the existence of intestinal carriers is insufficient, because, as will be seen further on, the typhoid bacteria in the feces of patients after cholecystectomy, may originate higher up than the gall bladder, in the liver or gall ducts. Positive evidence for the existence of a pure intestinal carrier type is afforded by means of duodenal cultures and was proved in only one case of our entire series. This was a patient 44 years old who reported sick Aug. 2, 1918, with the usual symptoms of headache, lassitude, and fever. He had had an antityphoid inoculation 7 days previously. He ran a com- paratively mild typhoid fever for approximately 4 weeks. There were no complications and he reached normal temperature and convalesced without interruption. The first routine stool culture made 5 days 28 TYPHOID CARRIERS AND TYPHOID IMMUNITY after normal temperature revealed the great proportion of colonies on the Endo plate to be typhoid colonies. The subsequent specimens of feces continued to show almost a pure culture of typhoid bacteria. This has persisted up to the time of writing (Feb. 20, 1919), 21 weeks after normal temperature and will probably remain so. 2 During the first 2 months of convalescence, however, 4 duodenal cultures revealed no typhoid bacteria in the large quantities of pure bile obtained. Actual dates of examinations are shown in Table XI. TABLE XI. Date of normal temperature. Dates of positive stool cultures. Dates of negative duodenal cultures. Aug. 28 Sept. 2 " 14 " 23 Oct. 10 « 19 " 26 Nov. 16 Sept. 25 Oct. IS " 23 « 29 Thus it can be assumed that types of pure intestinal carriers exist. They comprise only about 0.6 per cent (1 out of 164) of all typhoid patients and 2 per cent (1 out of S3) of all feces carriers, temporary and permanent. Of the permanent feces carriers, 1 out of 5 or 6 may originally have been a pure intestinal carrier. The fact that thereis a type of intestinal carrier is extremely important to know, for, as will be discussed more fully further on, cholecystectomy in this type of case is of no value. It is possible to differentiate with reasonable certainty an intestinal carrier from a bile carrier, by stool and duo- denal cultures, especially in the early period of the carrier state. The intestinal carrier begins to show the typhoid colonies in the feces plate cultures during early convalescence and they persist in almost every stool specimen in very great numbers and in almost pure 2 This prediction came true. The patient was operated upon several months later. Cholecystectomy was performed, but the carrier condition has continued. A. L. GAKBAT 29 TABLE XH. Name. Date of normal temperature. Dates of cultures. Feces cultures. Duodenal cultures. Aug. 27 Sept. 11 " 20 — " 26 + + Oct. 3 — " 11 — " 18 — " 29 + Dec. 2 + Sept. 24 Sept. 14 " 21 — " 28 — Oct. S + " 8 — Nov. 2 + Dec. 2 + " 31 — " 10 Sept. 13 " 24 + " 30 — Oct. 4 — " 11 + Nov. 1 + Dec. 2 + Jan. 2 + " 12 — " 20 — " 12 Sept. . 5 " 11 — " 17 — " 20 + " 21 — " 23 — " 30 — Oct. 7 — - " 20 — " 27 — Jan. 6 + « 9 , + + indicates typhoid bacilli present; — , typhoid bacilli absent. 30 TYPHOID CARRIERS AND TYPHOID IMMUNITY culture. It seems as if the entire intestinal flora are. changed from colon to typhoid. On the other hand, in those cases where the typhoid bacilli in the intestines originate in the bile, the typhoid colonies on plate cultures made from the stool during early convalescence are present either in very small numbers (1 typhoid colony to 50 or 75 colon colonies) or are absent entirely. The diagnosis of the bile carrier state is possible only by duodenal culture and not by feces culture, as is seen in Table XII. The results in this table are in marked contrast to the findings in the intestinal carrier quoted previously (Table XI) . Bile Carriers. The bile may transport the typhoid bacilli into the intestines from either one or both of two sites: (a) the gall bladder, (b) the liver or bile ducts. An artificial experiment could not have been planned to differentiate between these two sources better than the solution of nature offered by two cases of the series. Two patients, Karpinsky and Lehman, manifested symptoms referable to the gall bladder during their convalescence. At this time, direct culture of the bile by means of the duodenal tube showed numerous typhoid bacilli in Lehman and none in Karpinsky. Cholecystectomy with complete excision of the cystic duct was done in both (Major Kammerer). A pure culture of typhoid bacteria was obtained from the contents of each gall bladder. The apparent discrepancy in the case of Karpinsky of the negative bile culture before operation and the positive culture from the gall bladder was explained at the time of operation. A large stone was fixed in the cystic duct which completely occluded this passageway. The gall bladder was contracted, had a very much hypertrophied wall and contained a small amount of thick, greenish mucous fluid. The bile we had cultured from the duodenum before operation had come directly from the liver which was not infected and did not enter the gall bladder which was the only seat of infection! The mucoid bilious fluid in the gall bladder escaped into the duodenum probably only at irregular intervals as was evident by 1 stool examination when few typhoid colonies were present. Far different, however, were the conditions in the other patient (Leh- man) in whom typhoid bacteria were found in the bile from the duodenum before cholecystectomy and also in the bile of the excised gall bladder. At operation there was discovered a large chronically inflamed gall bladder, abnormally thick- ened and with numerous stones but all free so that there was no interference with the lumen of the cystic duct. After complete healing of the abdominal wounds in both patients, duodenal cultures were repeated in each. In Karpinsky (where the bile coming from the liver was sterile) it was found that typhoid bacteria A. L. GARBAT 31 remained absent,. and the patient was cured. In Lehman, however, n spite of complete excision of the gall bladder and the cystic duct, repeated duodenal cultures were positive; probably, even before the operation, the bile which came from the liver and before it entered the gall bladder was already infected and remained infected after cholecystectomy. Detailed study shows that it is this condition of hepatic carrier rather than the condition of intestinal carrier which in most cases accounts for the persistence of the carrier state after removal of the gall bladder and cystic duct. That our patient Lehman was not an intestinal carrier can be readily noted by the results of the stool cultures. Seven routine stool examinations made at intervals of 5 to 7 days from Sept. 5 to Oct. 18 (the date of operation) showed no typhoid colonies, although duodenal examinations on Sept. 20 showed a pure culture of very numerous typhoid bacilli. Again after operation, iso- lated specimens of stool showed no typhoid organisms while duodenal cultures repeatedly demonstrated a heavy growth of typhoid bacilli. A systematic search in the stool was then undertaken and out of 163 Endo plates (embracing a period of 2 weeks) prepared from various dilutions of daily stool specimens, 11 plates were found which showed few scattered typhoid colonies. As was discussed above, such findings in the stool are characteristic of a gall bladder or liver carrier and not an intestinal carrier. Were this patient an intestinal carrier, there would have been no difficulty in detecting typhoid in the stool in great numbers both before and after operation, as was the case in our one example. A more detailed history of these two patients is here given for those who may be interested in any particular data. Case 1. — Alfred Karpinsky, Register No. 571, age 42 years; born in Germany; sailor. Past History. — Nothing of importance is noted in his habits, family history, or venereal history. Has never been sick before. Present Illness.— Admitted to U. S. A. General Hospital No. 12 on Aug. 14, 1918, having been sick with headache and fever for about 8 days before this date at Hot Springs, N. C. Had received one antityphoid inoculation there before admission to the hospital. Examination on Admission to U. S. A. General Hospital No. 12 by Lieutenant Sanders. The patient had no complaints. Roseola in great numbers over thorax, abdomen, back, and thighs. Lungs.— Occasional rales. Heart.— Slightly enlarged to left; blood pressure 100/70. Abdomen.— Not distended, not tender. 32 TYPHOID CARRIERS AND TYPHOID IMMtTNITY Liver.— Edge felt 1 inch below costal margin. Spleen.— Edge palpable \ inch below costal margin. Temperature course.— Fluctuated from 102-103.6° reaching i04°F. only once; temperature continued from Aug. 14, date of admission, to Aug. 27, when it reached normal and continued so (total fever course ap- proximately 21 days). Early Laboratory Examinations. — Blood culture. — Aug. 18, negative. Widal test.— Positive. Blood count.— White blood cells, 10,550: polymorphonuclears, 58 per cent; lymphocytes, 42 per cent. Urine.— Trace of albumin, few erythrocytes. Further Course of Disease.— (Complication.) Patient did not feel sick during fever course. On Sept. 1, 4 days after the beginning of normal temperature, patient complained of pain in abdomen and on examination there was obvious tenderness in right upper quadrant and muscular rigidity. The temperature suddenly rose to 105°F., but came down in 2 days. Blood count. — White blood cells, 11,800: polymorphonuclears, 55 per cent. Diagnosis. — Acute cholecystitis. The abdominal symptoms subsided entirely in 8 days and patient felt well again. On Oct. 8, 1918, about 1 month after the first attack, the patient again com- plained of pain in abdomen and again there was rigidity and tenderness over the gall bladder region. Blood count. — White blood cells, 11,750: polymorphonu- clears, 56 per cent; lymphocytes, 44 per cent. Duodenal culture. — No typhoid bacilli, few colon bacilli, and many leucocytes. This attack was not as severe as the first; the temperature only rose to 101 C F. for several days. The patient consented to operation. Operation Report by Major Kammerer, Oct. 9, 1918. — Cholecystectomy. Many old adhesions about large gall bladder. Riedel's lobe present. Separation of gall bladder from bed of liver somewhat difficult. Gall bladder contained 21 large and small calculi varying from J to 1 inch in diameter, with 1 of the larger stones in the cystic duct. Wall of the gall bladder about \ inch thick; cystic duct ligated at junction with hepatic. Considerable hemorrhage from liver bed controlled by tampon. Cultures from gall bladder contents showed pure growth of Bacillus typhosus. Surgical Course after Operation. — About 1 week after the operation (Oct. 16, 1918), the patient became jaundiced and there was a slight rise in temperature. There seemed to be an obstruction to the flow of bile because the stools became clay colored. The tampons were removed from the wound after which there was a profuse drainage of bile for about 3 weeks. Then the patient's jaundice disappeared; the stools resumed their normal color; the drainage ceased almost entirely; the wound began to close. It was healed entirely on Jan. 10, 1919. Results of stool, urine, and duodenal cultures are shown in Table XIII. Condition at Time of Writing (Feb. 20, 1919).— Patient is feeling perfectly well; is up and about and is to be sent to Fort Oglethorpe. Case 2. — Herman Lehman, Register No. 505, age 34 years; born in Germany; barber. Past History. — Had volvulus at age of 15; otherwise nothing of importance is noted in personal or past history. A. L. GARBAT 33 Present Illness. — Began about July 27, 1918, with weakness, loss of appetite, and slight fever. Examination on Admission to U. S. A. General Hospital No. 12 by Lieutenant Sanders, Aug. 14, 1918. Patient complained of headache, abdominal pains, and deafness in left ear. General condition fair ; scattered roseola over body. Heart. — Action good; pulse slow; blood pressure 100/70. A bdomen. — Median scar from old volvulus operation; pain, tenderness, and rigidity in right upper quadrant. Spleen. — Just palpable. Diagnosis. — Acute cholecystitis complicating typhoid fever. Temperature course. — Patient came in with irregular temperature of 102- 104"F., which gradually subsided in 1 week (Aug. 28, 1918). After several days of normal temperature, the fever slowly rose to 105°F. and then gradually came down again to normal (Aug. 25 to Sept. 14). TABLE XIII. Dates of cultures. Feces cultures. Urine cultures. Duodenal cultures. 1918 Sept. 1 " 10 « 21 " 27 Oct. 3 " 10 + + + « 24 * • 1919 Jan. 1 " 2 • * » " 9 * * * Subsequent to operation. Early Laboratory Examination. — Blood culture. — Aug. 18, negative; Aug. 30, positive. Widal test. — Positive. Blood count. — Aug. 14, white blood cells, 4,850: polymorphonuclears, 68 per cent; lymphocytes, 32 per cent. Urine. — Very faint trace of albumin, occasional hyaline cast. Further Course of Cholecystitis. — This attack lasted for 3 weeks and gradually subsided by. Sept. 5, 1918, when the patient was entirely free of gall bladder symptoms. Summary. — Apparently the patient was first seen towards the end of the course of a typhoid infection, at which time there was also an involvement of the gall bladder. This inflammation subsided, but the patient had a relapse of the typhoid as was evidenced by a positive blood culture on Aug. 30. Further Course of Disease. — With the subsidence of the fever on Sept. 14, the patient began to feel well and had no further complaints either subjectively or objectively until Oct. 12 (almost a month after normal temperature), when he 34 TYPHOID CARRIERS AND TYPHOID IMMUNITY had a typical attack of cholelithiasis requiring hypodermic administration of morphine. The following day he was slightly jaundiced; he had local symptoms of cholecystitis and consented to operation. Blood count. — White blood cells, 8,800: polymorphonuclears, 44 per cent; lymphocytes, 56 per cent. Operation by Major Kammerer, Oct. 18, 1918.— Small gall bladder entirely enclosed within old adhesions which were in part ligated and in part torn away. Cystic duct ligated as closely as possible to entrance into common bile duct. Concentric hypertrophy of gall bladder; very small lumen; contained 1 large and several small irregular stones; gall bladder wall about i inch thickjcystic duct free. Culture from gall bladder contents showed Bacillus typhosus. TABLE XIV. Dates of cultures. Feces cultures. Urine cultures. Duodenal cultures. 191S Sept. S — — " 11 — — « 17 — — " 20 + " 21 — — " 23 — + " 30 — — Oct. 7 — — " 20 * * ' 27 * * 1919 Jan. 1 * • " 6 +* " 9 +* Feb. 10 +* " 27 +' May 11 * " 18 • * Subsequent to operation. Course after Operation— -Patient showed effects of shock after operation but soon recovered. About 1 week later, the wound showed some, drainage of bile which kept up for 2 weeks and then ceased. On Nov. 16, 1918, the wound was entirely healed. Nov. 20, 1918, acute phlebitis of left leg which continued for 2 weeks but then subsided; slight edema of left leg remained. Dec. 22, 1918, slight adenitis of left inguinal glands for 6 days. Results of stool, urine, and duodenal cultures are shown in Table XIV. Condition at Time of Writing (Feb. 20, 1919).— Patient is feeling perfectly well, no trouble of any kind; has gained 17 pounds in weight, but has remained under isolation as a carrier. (He ultimately cleared up.) A. L. GARBAT 35 While, as was seen above, it is possible to differentiate the intestinal carrier from the bile carrier, it is impossible by laboratory means to separate the two types of bile carriers, although it would be extremely important for therapeutic purposes to be able to do so. 8 Then, too, one cannot state absolutely whether in the liver carriers, the bacteria are located in the liver tissue or hepatic ducts. During typhoid fever the damage to the hepatic parenchyma in the form of the small areas of focal necroses might speak for the deposit of the bacteria in the liver; bacteria have been isolated from the walls of the larger divisions of the hepatic duct as well as from the liver substance itself. The most important advice is to try to establish the type of carrier during the early convalescence of the typhoid patient; i.e., at the be- ginning of the carrier state. At this stage, the intestinal carrier is readily differentiated from the bile carrier. If the carrier state con- tinues for a long time, various combinations of these three types may result, thus masking the characteristic features of the original type. . An original bile carrier, for example (either gall bladder or liver), may ultimately resemble (although not really be) an intestinal carrier; that is, the bacteria would become so numerous that the feces would constantly show large numbers of them; possibly, too, the bacilli may lodge in the intestines on their way down and actually multiply there. A case in point occurred in our series. This patient (Ericksen) had the ordinary type of typhoid without any complications and ran a moder- ately high temperature for about 23 days. Of the usual number of stool cultures, only one specimen showed a few typhoid colonies on the Endo plate about 1 month after the beginning of normal tem- perature. During the following weeks, the stooL cultures were again negative, although repeated duodenal tests revealed numerous typhoid bacteria in pure culture. After several .months, the stools too began to show typhoid colonies in greater numbers, and then these gradually 3 By the use of magnesium sulfate instillation into the duodenum, Lyon (25) has lately devised a method by which it is attempted to collect separately bile from different parts of the biliary apparatus. This should be tried in the future, and if proved practical, it may help us greatly in this differentiation of carriers. Instead of magnesium sulfate Stepp (26) injects a peptone solution into the duo- denum; these agents are supposed to cause a contraction of the gall bladder and expulsion of the bile into the duodenum. 36 TYPHOID CARRIERS AND TYPHOID IMMUNITY increased so that soon the larger proportion of the colonies on the Endo plates were typhoid; in other words, the picture resembled the char- acteristics of an intestinal carrier type. A more detailed history of this patient follows. Case 3.— Christen Ericksen, Register No. 518, age 25 years; born in Germany; sailor. Past History. — Nothing of importance is noted in personal history. Present Illness. — Started 4 weeks before admission to hospital with headache, pains in the abdomen, and constipation. Examination on Admission to U. S. A. General Hospital No. 12 on Aug. 13, 1918, by Lieutenant Sanders. The patient had no complaints except constipation; general condition good; tongue coated in center; scattered roseola over body. Heart. — Normal; pulse slow; blood pressure 105/65. Abdomen. — Flaccid. Liver. — Edge palpable just below costal margin. Spleen. — Palpable at costal margin. Temperature course. — Irregular fever from 101-104°F. for 13 days after admis- sion; temperature reached 104° only twice {total fever course approximately 24 days). Early Laboratory Examination. — Blood culture. — Aug. 18, negative. Widal test. — Negative. Blood count. — White blood cells, 5,650: polymorphonuclears, 59 per cent; lymphocytes, 41 per cent. Urine. — No albumin or casts. Further Course of Disease. — Patient did not feel very sick during the disease. After his temperature reached normal (Aug. 27, 1918) convalescence progressed rapidly and uneventfully. Results of stool, urine, and duodenal cultures are shown in Table XV. Condition at Time of Writing {Feb. 20, 1919). — Patient is .eeling perfectly well; has gained 20 pounds in weight; has remained under strict isolation as a carrier. 4 It thus seems plausible to assume that in this case the intestinal carrier condition was secondary to the biliary infection. The reverse condition may occur too; that is, an original intestinal carrier with no bacteria in the bile cultures made early during convalescence may ultimately show typhoid bacilli also in the bile. This transition will be explained in detail in the discussion of ascending infections of the gall bladder. 4 On Mar. 24, 1919, cholecystectomy was performed. Cultures from the gall bladder contents showed Bacillus typhosus. Stool cultures were positive only once, 5 days after the operation, and subsequent stool and duodenal cultures were negative. A. L. GARBAT TABLE XV. 37 Dates of cultures. Feces cultures. Urine cultures. Duodenal cultures. 1918 Sept. 11 — — " 20 — — " 26 + — + Oct. 3 — — " 11 — — " 18 — " 29 + Dec. 2 + " 31 + 1919 Jan. 2 + + " 6 + " 12 — " 21 + • Feb. 4 + " 11 + How Do the Typhoid Bacilli Reach the Gall Bladder? Three possibilities can be given as answers. 1. The generally accepted path for the conveyance of the typhoid bacillus to the gall bladder is by way of the blood to the liver and then into the bile. The damage to the hepatic parenchyma during the typhoid illness may more readily allow the passage of the bacillus into the bile. This injury to the liver tissue is not, however, essential for it has been repeatedly shown (Blachstein (27), Welch (28), and Doerr (29)) that in rabbits which have been inoculated intravenously by typhoid bacilli, the bacteria regularly appear in the gall bladder and often as early as 8 hours after inoculation. Doerr demonstrated that when the hepatic duct was ligatured the bile invariably remained sterile. Also when the cystic duct was ligatured, no bacilli could be demonstrated in the gall bladder. Once the bacteria reach the gall bladder, they grow there as readily as in a culture tube, finding a suitable medium which is periodically renewed and eliminated. If the foci in the liver (or ducts) clear up, then the carrier remains only a gall bladder (bile) carrier; on the other hand, the liver infec- tions may continue and then the patient persists also as a liver (bile) 38 TYPHOID CARRIERS AND TYPHOID IMMUNITY carrier. The factors which govern this are unknown. While the above experiments proved that the bacilli travelled by the bile and not by the blood vessels of the gall bladder itself, there are observers who believe that: 2. The typhoid bacilli may enter the bile by the blood capillaries in the submucosa of the gall bladder wall where they possibly form em- boli. In a patient who died during the first weeks of his typhoid ill- ness, Koch (30) by careful histological examination found nests of typhoid bacilli in the papillae of the inflamed mucosa of the gall bladder wall. These bacteria lay in close proximity to the capillaries and were therefore assumed to be capillary emboli. These findings were also corroborated in animals (Chiarolanza (31)). 3. The ascending route of infection of the gall bladder by way of the common bile duct has not been received with general favor. The author, however, desires to record the following case as evidence that this mode of infection is possible. Case 4. — Gustave Haak, Register No. 524, age 44 years; born in Germany. Past History. — Nothing of importance is noted in his habits, family history, or venereal history. Has never been sick before. Present Illness. — Was at the hospital at Hot Springs since Aug. 2, 1918, with headache and fever. On July 26 had received 1 antityphoid inoculation. ■Examination on Admission to U.S.A. General Hospital No . 12 on Aug 13, 1918, by Lieutenant Wenner. General condition good; tongue dry and coated. Heart. — Sounds fair; dicrotic pulse. Lungs. — Numerous rales. Abdomen. — Soft. Liver. — Edge felt one finger's breadth below costal margin. Spleen. — Just palpab'e and tender. Temperature course. — On admission, temperature was 104.6°F., but the fever came down gradually so that it reached normal on Aug. 28, 1918, and remained so (total fever course approximately 26 days). Early Laboratory Examination. — Blood culture. — Aug. 19, negative. Blood count. — White blood cells, 12,000: polymorphonuclears, 64 per cent; lymphocytes, 36 per cent. Urine. — Faint trace of albumin, occasional hyaline cast. Further Course of Disease. — Patient did not feel sick during fever course. After the temperature reached normal, convalescence progressed rapidly and un- eventfully. Results of stool, urine, and duodenal cultures are shown n Table XVI. Condition at Time of Writing [Feb. 20, 1919).— Patient is feeling perfectly well; has gained 31 pounds in weight. In recapitulation, we had a patient who from the very beginning of his con- valescence, 5 days after normal temperature, began to show an almost pure culture of typhoid bacilli in the feces which continued thus right along for 6 A. L. GARBAT 39 months (up to the time of writing). In spite of these early positive stool cul- tures, 4 duodenal cultures were made during the first 2 months after normal temperature and in none of them were typhoid but orjly colon bacilli found, although large quantities of pure bile were obtained for each culture. TABLE XVI. Dates of cultures. Feces cultures. Urine cultures. Duodenal cultures. 1918 Sept. 2 + — " 11 — + " 14 + — " 23 + — " 25 — " 30 — — Oct. 10 + — " 12 + — " 15 — « 19 + " 23 — " 26 + " 29 — Nov. 9 — " 16 + " 25 — Dec. 2 + + " 16 + 1919 * Jan. 2 + — — " 6 + + " 12 + " 21 + " 23 + Feb. 2 + " 4 + " 11 + " 20 + These findings proved the patient a true intestinal carrier. The fifth duodenal culture, however, which was repeated in 1 month showed few typhoid colonies amongst the colon bacteria. 1 month later, the typhoid colonies in the bile were more numerous and the future duodenal cultures continued to show a profuse pure growth of typhoid bacteria. 40 TYPHOID CARRIERS AND TYPHOID IMMUNITY Those opposed to the general principle of ascending infections of the gall bladder, may voice the criticism that the typhoid bacteria cul- tured from the duodenum in the case described did not come from the gall bladder, but had found their way from the large intestine into the duodenum and were only washed out from the duodenum by the bile which came from a non-infected gall bladder. Naturally, the only positive proof in answer to this criticism would be a culture directly from the gall bladder. 5 However, the writer feels that the complete absence of typhoid bacilli in the bile at first, then the slight invasion, and finally the almost pure culture, speaks for the infection of the bile by the ascending route. In this way, a primary intestinal carrier becomes also a bile carrier. Similarly, the principle of ascending infections may possibly account for a small percentage of liver carriers who originally may not have been Kver carriers, but only gall bladder carriers. Since bile is pro- duced in the liver almost constantly, it is present in the ducts all the time but in a stagnant condition, because it flows into the duodenum only intermittently. It should not be difficult therefore for bacteria to invade the bile ducts of the liver or the liver tissue itself when some infected bile from the gall bladder is dammed back through the cystic duct into the hepatic ducts and liver. Especially is this a possibility the longer the bacteria have been harbored in the gall bladder. This view was brought out forcibly in three patients at the Walter Reed Hospital kindly shown to the writer by Lieutenant- Colonel Nichols. They were all bile carriers; two of them had per- sisted for 12 and 13 years and one for only 12 months. Cholecystec- tomy and excision of the cystic duct cured the patient with the more recent carrier condition but had no effect whatever upon the old carriers. The duodenal contents of the latter showed as many typhoid bacteria after as before the cholecystectomy. The author does not wish to give the impression that the ascending route of infection, 8 Since the time of this original manuscript, Patient Haak was operated upon. On Mar. 18, 1919, a cholecystectomy (without drainage of the common bile duct) was performed by Captain William S. Long. As was predicted, the gall bladder contents showed a pure growth of typhoid bacilli. 1 month later (Feb. 21 and 24), both the stool and duodenal cultures were still positive for Bacillus typhosus. Apparently the infection by this time had extended higher up into the liver or bile ducts. A. L. GARBAT 41 either for the gall bladder or liver carriers, is the frequent or the general occurrence, but merely wishes to point out that this path of infection is undoubtedly a possible one. Surgical Treatment of Carriers. Dehler (32) was the first to suggest operative interference directed towards the cure of the carrier state. In his first 2 patients only drainage of the gall bladder (cholecystostomy) was resorted to. In one of these patients typhoid bacilli were isolated in small numbers from the feces several months after operation. Grimme (33) next reported that he had had the operation of chole- cystectomy performed in a female asylum carrier. IS days after the operation, Bacillus typhosus was found in the feces, but not at a later period. Since then, cholecystectomy has been adopted as the routine surgical procedure for the cure of carrier conditions, but the results have not been uniformly successful (Lorey (34), Loele (35), Huismans (36), Kamm (37), Fromme (38), Dehler (39), Leary (40), Nichols, Simmons, and Stimmel (41)). The cause for the failures is evident if the classification of the various types of carriers is kept in mind. Formerly when cholecystectomy did not cure a typhoid carrier, it was usually attributed to the fact that the cystic duct had not been completely excised. The bacteria were supposed to continue to multiply in the duct as a pocket where bile was retained, simulating a gall bladder but on a very much smaller scale. While this explana- tion may account for the failures in some of the early cholecystec- tomies, in recent times the surgeons are certain to excise the cystic duct completely. The author has personally observed 2 such com- plete excisions where the carrier state remained unaffected. The present study has shown that there are two additional reasons which may account for the failure : (a) Hepatic (or duct) carriers. These patients will not be relieved by cholecystectomy because the infection continues higher up. (b) Intestinal carriers. These too will be influ- enced in no way by complete excision of the gall bladder. The bac- teria will continue to grow somewhere in the intestines. Therefore, if a carrier is diagnosed as an intestinal one, operative treatment should not be resorted to. Only bile carriers should be 42 TYPHOID CARRIERS AND TYPHOID. IMMUNITY treated surgically and the type of bile carrier more amenable for operation is the gall bladder type. Unfortunately, it is impossible to differentiate absolutely between the gall bladder and hepatic types. 8 Taking this into consideration, the author advises instead of cholecystectomy alone, a cholecystectomy plus long continued drainage of the liver through the hepatic duct. The reasons for this ar,e both diagnostic and therapeutic. After the drainage of the bile has set in, it is a simple matter to take cultures of the bile coming from the liver through the rubber drainage tube and ascertain whether typhoid bacteria are present or not. At the same time, the stool should be examined as control. If repeated cultures from the bile prove that the liver or ducts are not infected, drainage may be stopped and the wound may be allowed to close quickly. On the other hand, if the bile is proved to contain typhoid bacteria, drainage should be prolonged with the hope that in this way the infection may clear up. Drainage of the common bile duct outside of the body has the ad- vantage over the natural flow into the intestines in that it keeps the intestines free of typhoid bacteria and in that a greater quantity of bile is drained. Physiology of bile excretion teaches us that although bile is formed more or less continuously, it enters the duodenum only periodically during the time of digestion. The bile during the intervening periods is prevented from entering the intestines because the opening of the common bile duct into the duodenum is closed by a sphincter. The secretion, therefore, backs up into the liver. No bile appears in the duodenum as long as the stomach is empty. When a meal is taken, the entrance of the chyme into the duodenum is followed by an ejection of bile. It would seem therefore that each gush of chyme into the duodenum excites by reflex action an inhibition of the sphincter and the opening of the common bile duct. By means of hepatic drainage, the bile which is formed all the time is also ex- creted all the time, thus constantly flushing the infected area and allowing of no stagnation where the bacteria can grow more readily. The possibility of ridding the liver or ducts of bacteria seems in this way a little more hopeful. Another important consideration is the time of operation. When surgical measures are contemplated, they should be done early when the original type of the carrier state remains uncomplicated. If the A. L. GARBAT 43 typhoid carrier presented no clinical symptoms referable to the gall bladder during the acute stage of the disease or convalescence, 6 months from the onset of normal temperature is a reasonable time to wait for nature to cause the disappearance of typhoid bacteria from the bile before operating. On the other hand, if symptoms of cholecystitis or cholelithiasis 6 were present during any stage of the illness and duodenal cultures persistently showed typhoid bacteria, operative interference may be undertaken even earlier, because these patients usually become chronic carriers. It is to be remembered, however, that not all patients who present gall bladder symptoms during typhoid fever become carriers, so that operation cannot be justified from the carrier standpoint unless duodenal cultures show typhoid bacteria. The author observed 4 carriers in each of whom cholecystectomy (without drainage) was performed. In 2 (Karpinsky and Ericksen) the carrier condition disappeared immediately after operation. In 1 of these (Ericksen) only 1 stool culture made 5 days after operation showed typhoid bacteria, while subsequent cultures both of the feces and duodenum were negative. These were examples of pure gall bladder carriers, the type most amenable for operation. In the third patient (Lehman) the duodenal cultures remained positive for four months after operation, although repeated stool examinations were negative. Except for the duodenal cultures, this case would have passed as a surgical cure and been permitted to leave the hospital be- fore he ceased to be a menace to the community. This is an example of a liver carrier, usually not helped by operation. (The writer is informed that ultimately, 8 months after operation, 2 consecutive negative duodenal cultures were obtained.) In the fourth patient (Haak) both duodenal and stool cultures continued positive after operation. This is an example of an original intestinal carrier with secondary bile infection. It is to be assumed that in many instances consent for operation will not be obtained from the apparently healthy carriers; but when it is, our classification and study of the various types should give us better results than we have had in the past, or abetter understanding of our failures. 6 The relationship between symptoms of cholecystitis during the typhoid fever and a later carrier state is more fully discussed in the next section. IV PREDISPOSING FACTORS TO THE TYPHOID CARRIER STATE WITH SPECIAL REFERENCE TO CHOLECYSTITIS AND CHOLELITHIASIS. Why Do Some Patients Harbor Typhoid Bacteria in the Gall Bladder Longer Than Others? For this study, the 21 patients who continued to show a positive bile culture were selected. A detailed analysis was made of the clinical course of their disease, in order to discover any possible pre- disposing elements to the carrier state. The following factors were investigated: (a) blood culture, (b) length of illness, (^se- verity of illness, (d) prophylactic inoculation, (e) cholecystitis, (/) cholelithiasis. Did These Patients with Positive Duodenal Cultures Have a Positive Blood Culture during the Active Stage of the Disease? This question is taken up because some readers may ask it. The author, however, feels that all typhoid cases at one time or another during the active stage of the disease have had a bacteremia; and especially those who later developed a positive bile culture. The bac- teria must reach the gall bladder by way of the blood (the only ex- ception to this is possibly the bile carrier due to ascending infection). It is surprising to note, nevertheless, that in only 4 of 19 cases was a positive blood culture reported (2 patients had normal temperature and no blood culture was taken). . Table XVII shows the time during the course of the disease when these negative blood cultures were obtained. table xvn. No. of cases. No. of cultures. Week of disease when culture was taken. 3 3 7 2 2 2 1 1 1st 2nd 3rd 4th 44 A. L. GARBAT 45 While a negative blood culture did not mean that there were no bacteria in the circulation, it did imply that there were no excessive numbers of them and that they did not persist in the blood over an extended period. Of the positive cultures, there were three obtained during the second and one during the third week. No relationship, therefore, can be discovered between a positive or negative blood culture as a predisposing factor towards the typhoid carrier state. Only if a great majority of the cases had shown a positive culture and especially at a stage of the disease when a positive culture is unusual, would it have been permissible to draw any conclusions. Does the Length of the Disease Have Any Bearing upon the Tendency of the Bacilli to Persist in the Bile? The length of the illness was reckoned in terms of the number of days of fever; i.e., from the day when the patient went to bed to the day when the temperature came down to below 100° and stayed there. These fixed dates have always seemed to the author a much more reliable method for determining the actual length of the disease than the way usually estimated; namely, from the time the patient first began to feel ill, to some arbitrary day during convalescence, such as when the patient was able to get out Of bed or leave the hospital, etc. In our 21 cases the duration of fever is shown in Table XVIII. TABLE XVIII. No. of cases. Duration of fever. No. of cases. Duration of fever . weeks weeks 3 2 4 5-6 7 3-4 1 6-7 2 4-5 4 7-8 The average case of typhoid fever runs a temperature for about 4 weeks. It is interesting to note that the majority of the typhoid carriers did not have a shorter acute illness but rather a longer one than the general average. The possibility must be considered that 46 TYPHOID CARREERS AND TYPHOID IMMUNITY this protracted course is not a predisposing factor to a carrier con- dition but one resulting from the ever existing infected bile in the biliary tract and intestines. In this connection it may be mentioned that among these 21 patients there were 4 cases with relapses and 4 cases with recrudescences; i.e., 38 per cent. This percentage is rather high when compared with the 18 relapses and 18 recrudescences (22 per cent) which occurred amongst the total number of cases (164). From the same figures it may be seen that 22 per cent of all the patients who suffer from either a recrudescence or relapse tend to become typhoid carriers, temporary or permanent (8 out of 36). Does the Severity of the Disease Have Any Bearing upon the Predis- position to the Typhoid Carrier State? We differentiated strictly between the length of disease and severity of infection. One may have a long, but at the same time mild in- fection and vice versa. The same holds true for a short illness. In grading the degree of each patient's illness, we merely took into consideration the severity of the infection during the acute stage. A ++++ scale was used. + + + + signifies the severest type of infection where prognosis was very grave from the start and the patient died. + + + signifies a seriously ill patient with a doubtful prognosis during his illness but one who finally recovered. + + is the usual sick typhoid. + means a very mild infection. — means that the patient was clinically not ill at all; were it not for the labora- tory findings, the disease would have remained unrecognized. Of the 21 patients with positive duodenal cultures: 1 was graded — (not sick at all.) 11 were " 4- 5 » « + + 4 « " +++ Of the 4 patients who became permanent typhoid carriers: 3 were graded + 1 was " + + It is evident that 16 of the cases, approximately 80 per cent, were only + and ++ types of infection. This proportion is practically , A. L. GARBAT 47 the same as existed between the entire number of + and + + cases and the total number of typhoid cases at the hospital. The point of importance that the author desires to emphasize is that it is not necessarily the severe types of infection that are prone to become carriers. Are Patients Who Have Received Antityphoid Inoculations and Have Been Infected in Spite of the Inoculations Less Liable to Become Typhoid Carriers? Of the total number of typhoid patients treated at the hospital; 38.0 per cent had received no inoculation. 41.0 " " " " onlyl " 11.8 " " " " 2 inoculations. 2.2 " '• " " 3 " 8.0 " " unknown. Of the 21 patients who persisted with typhoid bacteria in the bile: 45 per cent had received no inoculations; 40 per cent had received only one inoculation; and 15 per cent had received two inoculations. None had a complete course of inoculations. Those who had received only one or two prophylactic injections did not complete their third injection because the illness had already set in, or set in immediately after the last inoculation. While the above figures may speak some- what for the beneficial influence of inoculations, conclusions as to their value cannot be drawn too closely. Cholecystitis. In a study of the cases which manifested symptoms of cholecystitis either as a complication during the disease or as a sequel during con- valescence, a very close relationship to the carrier state was noted. Out of 178 proved typhoid cases, 8 (4.5 per cent) had definite clinical evidence of cholecystitis during the acute illness. A tabulation of the findings in 7 of these patients is given in Table XIX (the eighth died of pneumonia several days after admis- sion before the laboratory examinations could be made). 48 TYPHOID CARRIERS AND TYPHOID IMMUNITY TABLE XIX. i 4) S u Name. I 3 w V u V Ph M u a ■s P 1 .s Si a .3 en el 3 o u + + + 1 S3 § d u Results. 1 2 3 4 5 6 7 Karpinsky . . Kobe Bergen thai.. . Kussner Jannssen. . . . + + + + + + +* +* +t + + + + + Cured (cholecystectomy). Remained carrier even after chole- cystectomy, due to liver infection. Died (cholecystitis and complicating myocarditis and pneumonia). Cured (temporary carrier). <( « « u « « " (never proved a carrier). * Operation, t Post-mortem. Thus it is noted that of the 7 patients who manifested symptoms of cholecystitis 6 ultimately became typhoid carriers. The bacilli were found in the excised gall bladder of 3, in the duodenal contents of 2, and in the feces of 1. Some of the discrepancies of the above table require explanation. Case 1 {Table XIX) showed no typhoid bacilli in the duodenum because as was proved by operation the cystic duct was occluded by a large stone. At the time when the positive stool was obtained, it was probable that the cystic duct was either not yet completely occluded or was only intermittently occluded. Case 3 {Table XIX) never showed typhoid bacilli in either the feces or duo- denum. The patient had recurrent attacks of cholecystitis and died during one of these periods of a myocarditis and a complicating pneumonia. A short review of this man's history will disclose the evidence of the carrier condition. He was admitted to U.S.A. General Hospital No. 12 on Aug. 13, 1918; had been at the hospital in Hot Springs since Aug. 1, 1918; had typical clinical typhoid with marked intestinal hemorrhages and definite myocarditis. Aug. 13 to 26. Temperature gradually came down to normal. Aug. 26 to Sept. 1. Temperature normal. Sept. 1 to Sept. 21. Typical relapse with blood in stools and temperature as high as 104.8°F., gradually coming down. Sept. 21 to Oct. 9. Temperature normal; patient out of bed. A. L. GARBAT 49 Oct. 9 to Oct. 18. Another relapse but milder than the first. Oct. 18 to Nov. IS. Temperature normal; patient out of bed. Nov. 15 to Nov. 21. Temperature up to 103°, condition simulating relapse, with symptoms of cholecystitis for the first time. Nov. 21 to Jan. 2. Temperature nofmal; patient out of bed. Jan. 2 to Jan. 30. Temperature up to 103.5°, with renewed symptoms referable to cholecystitis associated with lobar pneumonia; death. Laboratory Findings. — Widal test. — Positive on admission. (Patient had received no inoculation.) Blood cultures. — 6 negative. Stool. — At no time posi- tive for typhoid bacilli. Urine.— Typhoid bacilli found only once, Nov. 22, 1918. Duodenal culture. — Oct. 10, negative. Autopsy report. — The gall bladder region presented old dense fibrous masses which bound the liver to the diaphragm and formed one large mass of the gall bladder, pylorus, duodenum, and pancreas. After very difiicult dissection, the gall bladder was found deeply buried. It was the size of a pigeon's egg and was contracted down upon four yellow apparently cholesterol stones which entirely filled the cavity. The gall bladder contained a little mucus but no bile. The cystic duct was obliterated. Cultures from the interior of the gall bladder showed Bacillus typhosus; at the same time, cultures from the duodenum, large intestine, liver, spleen, lungs, abdominal fluid, etc., showed no typhoid bacilli. The apparent discrepancy between the positive culture from the gall bladder interior and negative culture from the duodenum before and also after death are readily explained by the com- plete shutting off of the cystic duct both by the stones and adhesions. The bile which was examined during life did not come through the gall bladder but directly from the liver which was not infected. The bacteriological conditions in this patient were identical with those of patient Karpinsky discussed in the chapter on bile carriers. Cases 4 and 5 (Table XIX) showed typhoid bacteria in the bile by duodenal culture. Cases 6 and 7 (Table XIX) had the most severe attacks of cholecystitis. Re- peated attempts at duodenal cultures were unsuccessful. In Case 6 the duodenal tube was passed four times but on no occasion was bile obtained. Similarly, in Case 7 (Table XIX), three attempts were made and on one occasion only bile- stained fluid, but not true bile, was recovered. This did not show typhoid bacilli. It is very unusual to meet with such difficulty. The failures were probably due either to reflex spasms of the pylorus or true pyloric obstruction from adhesions of the gall bladder. The frequent vomiting which these patients suffered from substantiated this view. The fact that the stool examinations in Case 7 did not reveal typhoid bacilli does not, in the light of experiments already discussed, ex- clude the possibility of their existence in the bile. It is possible, too, that here again we were dealing with a contracted and obliterated gall bladder where the typhoid bacilli could be recovered only from the interior of the gall bladder. 50 TYPHOID CARRIERS AND TYPHOID IMMUNITY In recapitulation, it is noted that of 7 patients who had attacks of cholecystitis during the course of their typhoid fever, 6 became definite typhoid carriers, either temporary or permanent, with typhoid bac- teria in the bile or feces; the 1 other patient was inadequately ex- amined. These facts assume importance in the light of the surgical therapy in carriers. Cholecystectomy (plus drainage of liver) for typhoid cholecystitis becomes warranted very much earlier when we have the additional knowledge that a typhoid carrier state results in most cases of cholecystitis. Once the symptoms of cholecystitis have manifested themselves, recurrent attacks of pain and fever are almost bound to arise as long as live typhoid bacteria remain in the gall bladder. This is probably the state of affairs that exists in patients operated upon for cholecystitis many years after the typhoid illness, in whom typhoid bacilli are unexpectedly found in the gall bladder. These individuals had undoubtedly been carriers during all that interval. Especially in these patients should duodenal cultures be made as soon after the .operation as feasible, in order that a liver carrier may not be overlooked. If operation for typhoid cholecystitis is delayed and attacks recur, very marked adhesions are developed which may make surgical interference very much more difficult or even dangerous. Such would have been the case in our patient Kobe who died of a compli- cating pneumonia and myocarditis before the operation was under- taken. Even at post-mortem it was impossible to dissect out the gall bladder without great damage to the related organs. Many of the so called prolonged cases of typhoid fever or the cases with frequent relapses are probably typhoid infections of the gall bladder, although the local process may be so mild that definite symptoms of cholecystitis are either overlooked or missing. On the other hand, not all patients of our series who became carriers had manifested symptoms referable to the gall bladder at some- time during the course of the acute disease. Of our 21 patients with persistent positive bile cultures, only 7, or 33 per cent, had presented gall bladder symptoms during the typhoid illness. Amongst the 14 others who did not show such symptoms, 2 were the most persistent carriers. A. L. GARBAT 51 Pathology of the Gall Bladder in Carriers. This division of carriers on the basis of former gall bladder involve- ment explains very logically the two types of gall bladders met with in carriers. In those who do not manifest any gall bladder disturb- ances, the gall bladder may show only very slight pathological changes. The normal glistening appearance of the peritoneal coat may be dulled, the wall may be slightly thickened, the entire organ may not be en- larged, and on being cut open the mucous membrane may present just a thickened or congested appearance. Microscopically, the different layers of the wall may be sharply outlined, but infiltrated by lymphocytes and few leucocytes. In these cases the gall bladder acts purely as a test-tube containing the bile medium in which the typhoid bacteria propagate without affecting the gall bladder itself. In the second type, the gall bladder may present the inflammatory changes of various grades of cholecystitis. The entire organ may be buried in dense adhesions, the walls may be J to f inch in thickness, and on section the entire normal gall bladder appearance may be "obliterated and replaced by a mass of fibrous tissue infiltrated by lymphocytes and polynuclear leucocytes. In between these two extremes,, all degrees of pathological changes may exist. Cholelithiasis. The relationship between cholelithiasis and typhoid fever has already stimulated such a vast amount of literature that only the outstanding features can be reviewed here. The discovery of the typhoid bacillus in cases of cholecystitis and cholelithiasis long after recovery from the primary infection, was made many years previous to the recognition of the typhoid carrier. In 1892 Naunyn (42) had observed that gall stone troubles frequently occurred in persons who had suffered from typhoid fever. Lentz (43) in 1905 was the first to direct attention to the association of gall bladder com- plaints with the carrier state. Since then it has been frequently shown that quite a large number of carriers suffer from gall stones, while in others, though symptoms may have been absent, a condition of cholelithiasis has been very frequently found, whether on examina- tion, post-mortem, or at operations directed towards bacteriological cure of the carrier condition. 52 TYPHOID CARRIERS AND TYPHOID IMMUNITY In our 21 carriers there were 3 (14 per cent) in whom gall stones were discovered. (1) Karpmsky. — Operation about 7 to 8 weeks after onset of illness; 21 stones; 4 very hard and about 1 inch in diameter; 1 about f inch in diameter; 5 about \ inch in diameter; the rest about \ inch in diameter; they were all hard and faceted. (2) Lehman. — Operation 2 months after onset of illness; 1 large, hard, mul- berry-shaped gall stone about \\ inches in diameter and several very small softer stones (all calculi of cholesterol). (3) Kobe. — Post-mortem about 6 months after admission to hospital; 4 stones (cholesterol) ; 2 large stones each about 1 inch in diameter and 2 smaller ones about I inch in diameter. Forster (44) in an analysis of several hundred carriers found the same percentage (14 per cent) of gall stone sufferers. The remaining 85 per cent undoubtedly have some disease of the gall bladder, al- though they give no clinical indication thereof. This accords well with our knowledge thai 90 per cent of all gall stone cases present no symptoms during life. These facts brought to the writer's considera- tion the question whether it is the chronic carrier infection of the gall " bladder which gives rise to the formation of the stones, or vice versa whether it is a preexisting cholelithiasis which predisposes the typhoid patient to the development or continued existence of the carrier state. X-ray examination of the gall bladder region was made in each of our 21 carriers, but in only 2 were the x-ray findings suspicious of stones. These negative results agree fully with the fundamental findings of Blaohstein (27) and Welch (28) who showed that no locus minoris resistenticB in the gall bladder was essential for the development of the carrier condition; mere intravenous inoculation of typhoid bacilli in rabbits is followed by the presence of these organisms in the gall bladder for a very long while— in one instance for 128 days. While these experiments disprove the absolute necessity of a previous gall bladder injury for the formation of a carrier condition, they do not speak against the logical assumption that those patients with a previous cholelithiasis are more prone to become carriers, or to remain such for a longer period of time. As for the relationship between the typhoid bacillus and actual gall stone formation, here also a division of opinion exists. While A. L. GARBAT 53 the typhoid bacillus has been repeatedly isolated from the center of gall stones both in cases of cholelithiasis operated on by the surgeon and in chronic carriers either at operation or at autopsy (Anton and Futterer (45), Droba (46),Blumenthal (47), Levy and Kayser (48)), no definite decision has yet been reached as to whether the typhoid bacillus directly excites the formation of the gall stone and forms a nucleus for it or whether the stone is preformed and later penetrated by the typhoid bacillus. Conforming with these opposing views, some authorities (e.g. Bacmeister (49) ) believed that organisms were only to be found in old stones into which they had penetrated, while other observers (e.g. Cushing (50)) held that only recently formed stones contained bacilli. As is so often the case in medicine, experimental data were furnished by both sides in support of their contentions. On the other hand, gall stones were placed in broth and bile cultures of the Bacillus typhosusandalsooi Bacillus coli, and these bacteria were then recovered from the center. Gilbert and Fournier (51) found that bacilli wandered in when stones were of cholesterol and not in the case of other stones; but then, it is usually the cholesterol stone which develops during typhoid fever. The Aschoff school has recently shown that the cholesterol content of the blood is increased in all long continued acute and subacute septic or pyemic processes and therefore concludes that in typhoid the stones are found in the gall bladder because of stagnation of the bile. The bacteria may later travel into the stone and thus bear no etiological relationship to the stone. On the other hand, embryonic gall stone formation may be observed in test-tubes containing bile cultures of either typhoid or colon bacilli. Bacmeister found that Bacillus coli, Bacillus typhosus, Bacillus proteus, and especially Bacillus pyocyaneus could cause a precipitation of cholesterol from bile. Even true concretions were found after prolonged growth of the bacteria in this medium. Then, too, gall stone formation has been observed after injecting heated bouillon cultures of typhoid bacilli into the gall bladders of rabbits (Gilbert and Fournier (52)). Doerr (53) has met with two concretions of the size of a lentil in a rabbit which had been injected intravenously with typhoid bacilli 40 days before. Gay (54) noted 54 TYPHOID CARRIERS AND TYPHOID IMMUNITY gall bladder concretions in some of his carrier rabbits; and Richardson (55) succeeded in experimentally producing concretions in rabbits by the injection into the gall bladder of agglutinated typhoid bacilli. In the light of these varied experimental results, there can be no doubt that an attack of typhoid fever predisposes a patient to the formation of gall stones, but at present one cannot be certain how far this formation is contributed to by other factors. It is probable that in certain instances the precipitation of cholesterol comes first, to be followed by the entrance of bacteria, while in other instances the stagnation of the bile may cause a precipitation of the cholesterol around a nucleus of epithelial cell debris and typhoid bacteria. The stones formed during typhoid fever may reach a large size in a comparatively short time. In one of our carriers (Kobe) 4 choles- terol stones were removed from the gall bladder 6 months after admission for the acute typhoid. 2 were about 1 inch in diameter and 2 about § inch in diameter; typhoid bacteria were recovered from the center of all. The size of the stones seemed very much larger than would be expected to develop in 6 months. One should not, however, fall into the fallacy of attributing to the typhoid infection all stones found in a patient who gives the history of a recent or old typhoid. In Case 2, the large stone-hard choles- terol calculus about 1J inches in diameter undoubtedly existed long before the typhoid illness which was only 2 months old. The 2 smaller and softer stones, also of cholesterol, were probably of typhoid origin. Unfortunately the stones were dropped into formalin and cultures .made from the center of the stones the next day were found sterile. It may be that in typhoid patients with preexisting choleli- thiasis, the tendency to the formation of stones being already present, the added element of the typhoid infection predisposes the patient to additional and more rapid gall stone formation. V. TYPHOID URINE CARRIERS. A. Irregularity and Intermittency of Bacilluria; New Method for Detecting Urine Carreers. The excretion of typhoid bacilli in the urine from typhoid con- valescents has been studied in 164 patients and entailed approxi- mately 2,000 urine cultures. According to the army rule, every patient was kept under strict typhoid precautions until 3 consecutive urines and stools at intervals of 6 days were free of typhoid bacteria. The first culture of the urine was generally made when the patient's temperature was nearing normal or had already reached normal. It was found that the excretions of typhoid bacilli in the urine during convalescence followed one of three courses: (a) Urines which were at no time positive (3 or more consecutive urines at intervals of 6 days being negative). This comprised 51 per cent of the patients (84 out of 164) . (b) Urines which were at first positive and remained so for a shorter or longer time, then became negative and remained so. This comprised 25 per cent of the patients (41 out of 164). For example, patient Gutte: Sept. 13, 18, 29, positive; Sept. 30, Oct. 7, 26, Nov. 2, negative. (c) Urines which changed, sometimes positive, at other times nega- tive, then positive only to become negative again, etc. This com- prised 23 per cent of the patients (39 out of 164). As typical of such irregularity of findings in the routine urine specimens, results with 4 patients are cited in Table XX. It is these changing urines which particularly interested the writer. The question arose whether it was possible that 3 isolated specimens of urine examined at intervals of 6 days could show no typhoid bacilli, while other specimens taken in between or even later on contain typhoid bacteria. The great danger of overlooking urine carriers would thereby be offered. As the routine cultures were usually made of specimens passed in the morning, the first question to decide was whether the time of voiding had any influence upon the result. 12 55 56 TYPHOID CARRIERS AND TYPHOID IMMUNITY patients were selected and every specimen voided by each patient was carefully collected and cultured separately. All precautions for obtaining a sterile specimen were observed: the head of the penis was washed with bichloride solution; the first part of the urine voided was not kept but used for washing out the urethra; the last part was voided directly into a sterile bottle. At this time, cultures were made only on Endo plates; 1 cc. of each specimen was spread in a thin layer over the entire surface of the media and incubated for 24 hours. TABLE XX. Urine examinations. Urine examinations. Name. Name. Dates. Results. Dates. Results. Adam Sept. 13 " 17 + Sept. 12 " 17 + + " 19 + Oct. 5 — " 25 — " 21 + Oct. 1 + " 28 + " 18 + " 30 — " 20 — Nov. 4 — " 26 " 18 " 18 + Sept. 7 " 16 + Oct. 8 " IS + " 21 + " 21 + " 30 " 30 Nov. 6 " 22 " 29 + + indicates typhoid bacilli present; — , typhoid bacilli absent Every suspicious growth was identified by the Russell double sugar medium and then by serum agglutination. The results in 3 of the 12 cases thus examined are cited in Table XXI. We sought the explanation for the striking results shown in Table XXI. Keeping in mind the general laboratory experience that sterile cultures are often seen on solid media when growths occur in broth, we continued the above plan of separately examining every specimen voided by 5 patients, but this time cultures were made not only on Endo plates, but also in broth (from 1 to 2 cc. of urine). The A. L. GARBAT 57 TABLE XXI. Urine examination. Urine examination. Name. Name. Date. Time. Result. Date. Time. Result. a.m. p.m. Sept. 13 2 p.m. — Trautman . . . Sept. 13 2 5 + 4 + 8 + 7 + " 14 7 + 10 — 10 + a.m. a.m. « 14 11 p.m. 3 + " 15 5 8 Pm. + + 9 + 2 + a.m. 7 + " 15 4 8 + 11 a.m. + i " 16 5 + "*~ 3 p.m. 11 l + #.m. 5 + " 17 9 a.m. — 8 10 + + " 18 3 ». + 11 a.m. " 19 12 — " 17 5 — #.«*. . 8 — 8 — p.m. a.m. 2 — " 20 4 P.m, + 6 9 ■""* " 22 7 a.m. + 10 a.«. + " 23 3 + " 18 6 + " 24 4 + p.m. " 25 1 + ■ 2 — 6 + 6 + £.IK. 9 — 2 + 11 — 8 + a.m. 11 — " 19 5 — a.m. " 26 4 6 + + 58 TYPHOID CARRIERS AND TYPHODD IMMUNITY TABLE XXI— Concluded. Urine examination. Urine examination. Name. Name. Date. Time. Result. Date. Time. Result. p.m. p.m. Sept. 13 9 a.m. + Sept. 19 5 8 + « 14 2 + a.m. 6 + " 20 4 + p.m. p.m. 4 4 a.m. 9 " 15 2 9 p.m. a.m. " 21 8 + " 16 9 a.m. + " 22 10 4 — " 17 4 p.m. + 7 p.m. 5 — 2 — 7 + 5 — a.m. 8 — " 18 8 + a.m. " 19 7 + " 23 2 — 10 + 5 — p.m. 7 — 1 + 2 + | results from 1 case are noted in Table XXII. The other 4 patients gave similar results. It was thus definitely proved that the use of Endo plates alone for the purpose of detecting typhoid bacilli in the urine is insufficient unless the number of bacteria happens to be excessive. The safe procedure is first to obtain a 24 hour growth in broth and then to make an Endo plate therefrom. The irregularity in the excretion of typhoid bacilli in the urine thus seemed to be explained as dependent upon the technique; i.e., the use of solid media only. On further study, however, it was soon found that there was an actual intermittency in the excretion of the typhoid bacilli. That is, certain specimens of urine were found absolutely sterile (also in broth), while others voided within several hours were A. 1. GARBAT 59 full of typhoid bacteria. As examples, we may cite the following cases (Table XXIII). Thus it is surprising to find, for example, a urine (Schubert, Table XXIII) full of typhoid bacilli at 3 p.m., sterile 4 hours later, and full of typhoid bacilli again 3 hours later. It only definitely proves that excretion of typhoid bacteria in the urine follows an intermittent curve. TABLE XXII. § .9 g .9 •8 8 1 •o u 1 ed*» © rt<-> S u Name. Date. Time. 5 & 3S Name. Date. Time. n & ££ S^ 2^ 3 ° •gw "3-S 3W 3& u u u o a.m. a.m. Sept. 27 4 + + Sept. 29 2 — + 9 + + 6 — + p.m. 8 — + 5 + + n. 8 + + 12 — + 11 + + p.m. a.m. 4 __ + " 28 3 6 p.m. + + + + 9 a.m. — + 2 ^_ + " 30 1 + + 6 — + - 8 + + 10 — + p.m. a.m. 4 + + " 29 i — + 11 + + During the entire study, it was noted that although practically the same quantity of urine was used for plating, the intensity of the growth on the Endo plate varied greatly. A quantitative estimation of bacteria in the urine was therefore undertaken in order to determine the approximate curve of excretion in any one patient. A pour plate in agar was made from 0.05 to 0.5 cc. of every specimen voided, incubated for 24 hours, and the number of colonies counted. Control cultures of each urine were made on Endo plates and in broth in order to ascertain any contaminating organisms. Results are shown in Table xxrv. 60 TYPHOID CARRIERS AND TYPHOID IMMUNITY TABLE XXIII. § .3 § ,9 U • a 4) . Name. Date. Time. S-3. u I 11 u Name. Date. Time. "S3 11 o a.m. p.m. Oct. 1 6 10 + + + + Oct. 1 8 a.m. + + P.m. « 2 10 + + 3 + + p.m. 7 — — 1 + + 10 + + 4 — — a.m. 9 + + " 2 i + + a.m. 6 + + " 3 5 + + n + + p.m. p.m. 2 + + s 10 + + + + Wietchel Sept. 27 1 #.m. + + a.m. 3 + + « 3 3 + + 6 + + 7 + + 11 p.m. a.m. 3 — — " 28 6 — + 8 + + ». 11 + + 12 + + a.m. P.M. " 4 2 — + 2 + 8 + + 7 ^_ + Sept. 29 8 + + a.m. 11 + + « 29 2 + p.m. 7 + 4 + + «. 8 + + 12 _ + a.m. P.m. " 30 6 *""■ + 2 6 — + + 3 + + " 30 6 + + 8 + + 9 + + m. a.m. 12 — — Oct. 1 7 _ __ a.m. p.m. Oct. 1 6 n. 12 + + + + • 1 + 40,000,000 35,000,000 30, 000,000 25,000,000 20,000,000 15,000,000 § 10,000,000 "5 I 2 §000,000 1,000,000 500,000 100,000 60,000 20,000 15,000 10,000 1,000 & 3 5i o 5T o vl ?J — )5l 8 § ty S qy — " o- «\i S . Q 1 V A f I <\l IS / 1 v/ R h* i tv N cA V ♦ ^\ -co. ^ / <°A s \ h r N / s \ , « 1 <■ »— — -t >■■■- ♦■■■■H'" »-— — e _ - E 6- - g e g c E - S~p_- £ £ g p £E- £ S g 6 £ fi 3 P- " i~ CDcM <\! O! CM t~ O CO <- c\i CO «* lO LO r~ 03 o> a — CM CO ** lO CD t~ CO a> o CM S. jj ^J •¥» -j-j ^j ^_; ^j , : _^j ^J ^J ^ «_> 4_i ,j -*-> +-* « 8 o f) u o u o ' i <"> :j y O n o O < > o o o O o to o O CJ o o o i J O O o o o o O O o o o o O Chaet 2. Quantitative estimation of bacteria in urine (Table XXV). A. L. GARBAT 61 TABLE XXIV. Name. Date. Time. Total No. of colonies per cc. Name. Bate. Time. Total No. of colonies per cc. p.m. a.m. Schubert . . Oct. 1 3 60,800 Schubert . . Oct. 8 * 7 " 11 * 10 6,260,000 " 12 * TO. " 14 * 12 80,000 p.m. a.m. " 15 11 232 " 2 6 25,100 a.m. 11 2,640,000 " 16 8 158 P.m, « 17 * 5 2,400 " 18 * 10 6,500 " 19 * 500 a.m. " 3 3 7 p.m. 520,000 11,000 " 20 p.m. 9 a.m. 300 3 " 21 8 400 8 520,000 « 22 * 11 2,240,000 " 23 * a.m. " 24 * " 4 2 500,000 " 25 * p.m. " 26 8 3,400 2 10,370,000 " 27 9 4,200 8 288 " 28 8 2,000 11 3,200 " 29 * a.m. " 30 * " S 4 1,440,000 " 31 * 7 58,000 Nov. 1 * p.m. " 2 * n 20,800 " 3 * cm. " 4 * " 6 s 64 a.m. p.m. " 5 7 200 2 1,200,000 " 6 * 8 40 " 7 * s.»». " 8 * " 7 1 " 9 * 9 4,300 " 10 * 11 15,600 ' All specimens. 11,000,000 10,000,000 9,000,000 8,000,000 7,000,000 6,000,000 5^000,000 4,000,000 3,000,000 3,000,000 I! fe 1,000,000 V 100,000 & I ;§ 50,000 30,000 10,000 1,000 - CO 0> Q Mil 62 TYPHOID CARRIERS AND TYPHOID IMMUNITY It is interesting to note the marked changes; e.g., from 10,000,000 colonies to 200 in the course of 6 hours. Graphically represented, these figures are even more striking. The presence of several sterile specimens among the positive ones should be noted (Chart 1). TABLE XXV. Name. Date. Time. Total No. of colonies per cc. Name. Date. Time. Total No. of colonies per cc. p.m. p.m. Wietchel . . Sept. 30 6 9 24,000,000 6,720,000 Wietchel . . Oct. 5 11 a.m. 10 28,000,000 " 6 8 a.m. p.m. Oct. 1 7 p.m. 2 3 1 1,920 11 6 18,200,000 11 a.m. 8,400,000 " 7 a.m. 7 " 2 7 64 " 9 * p.m. " 10 * 3 2,680,000 " 11 * 7 3,600 " 12 V a.m. m. " 3 7 71. " 12 12 a.m. 800 12 36,400,000 " 13 7 1,200 p.m. " 14 * 3. 122,000 " IS * 6 11,600,000 " 16 • *• a.m. " 17 * " 4 ■ 7 100 " 18 • P.m. « 19 * 9 4,140 " 20 * a.m. " 21 « " 5 7 216,000 •All spe cimens. Another patient (Wietchel) in whom a quantitative determination was undertaken proved of interest because the estimate was begun just before the urine became negative. It is observed that when the excretion of the typhoid bacteria ceases, it does not necessarily do so by a gradual downward curve but by an abrupt intermittent fluctu- ating one (Table XXV) . A. L. GARBAT 63 The fluctuations in this patient's excretion of bacteria were even greater than those in Chart 1. A drop from 36,000,000 to 100,000 occurred within 3 hours. A graphic representation of this patient is seen in Chart 2. An especially instructive tabulation was made on a patient who for a time was considered a permanent urine typhoid carrier (Table XXVI). He continued with a marked excretion of typhoid bacteria in the urine for 6 weeks. Quantitative estimations were begun 19 days after the onset of normal temperature and continued on almost every specimen voided for 5 weeks. As in the foregoing instance (Table XXV), there is no gradual downward curve in the number of bacteria when the urine is nearing the time of complete sterility but an intermittent fluctuating type of excretion. A graphic presentation of this patient's excretion is given in Chart 3. In these patients, all urines continued to be cultured for many days after the specimens were repeatedly sterile, and in all cases there were occasional urines which showed typhoid bacilli. In no instance was there a renewed excretion of typhoid bacteria for any length of time once the urine continued negative in every specimen for several days. To establish more conclusively this intermittent excretion of typhoid bacteria, 3 patients were asked to void every hour during the day and a quantitative estimation of each specimen was made separately. In this way multiplication of the bacteria in the bladder was avoided. Tables XXVII and XXVIII (Charts 4 and 5) tabulate the results of two such experiments. They show definitely that there is an actual fluctuation of the number, of bacteria excreted in the different hourly specimens of urine. At the same time it appears as if the very high figures in the former charts are partly accounted for by a growth of the bacteria in the bladder urine. It is very important to note that there was a sterile specimen passed at 7 p.m. (Table XXVII), while the one before it (6 p.m.) and the one after it (8 p.m.) contained typhoid colonies. In doing one of these experiments, we luckily struck the last hour that the typhoid bacteria appeared in the patient's urine. As will be seen from Table XXVIII, all the urines passed at or before 3 p.m. contained numerous typhoid bacteria, while all the specimens after 64 TYPHOID CARRIERS AND TYPHOID IMMUNITY TABLE XXVI. Name. Date. Time. Total No. of colonies per cc. Name. Date. Time. Total No. of colonies per cc. p.m. a.m. Mueller. . . . Sept. 30 8 1,330,000 Mueller. . . Oct. 7 6 2,980,000 m. " 8 10 6,300,000 12 T*. a.m. 12 2,000,000 Oct. 1 6 #.m. 150,000 p.m. 2 1,280,000 4 5,860,000 3 1,400,000 8 2,160,000 7 141,000 " 2 10 680,000 9 a.m. 1,260,000 p.m. i 1,800,000 « 9 2 1,920,000 4 6 740,000 9 1,050,000 11 456,000 a.m. £.?». " 3 5 p.m. 43,200,000 3 9 1,740,000 528,000 2 10,300,000 10 1,200,000 9 8,640,000 a.m. a.m. " 10 i 60,000 " 4 9 12,300,000 6 11 6,400,000 n 10,400,000 n. p.m. 12 15,800,000 3 8,000,000 £.m. 8 1,400,000 3 2,600,000 10 980,000 5 8,400,000 a.m. a.m. " 11 i 1,120,000 " S 6 11 #.m. 15,800,000 10,000,000 7 p.m. 2 1,320,000 1,400,000 4 9,640,000 5 760,000 5 1,680,000 8 440,000 8 960,000 a.m. m. " 12 2 1,540,000 12 8,600,000 7 600,000 a.m. p.m. " 6 11 p.m. 2 740,000 560,000 2 9 a.m. 1,200,000 1,800,000 3 2,180,000 " 13 6 2,420,000 1 9 1,380,000 Number of bacteria ro ro o> to *> p" o -W o <-n o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o *. o o o "o o o Sept. 30 8p.m. 12m. Oct.l 6am. 4 p.m. 8 " Oct. 2 10a.m. lpm. Oot.3 Oct.4 Oct. 5 Oct.6 Oct.7 Oct-8 10 A. L. GARBAT 65 TABLE XXVI— Continued. Name. Date. Time. Total No. of colonies per cc Name. Date. Time. Total No. of colonies per cc. p.m. a.m. Mueller. . . . Oct. 13 4 1,420,000 Mueller . . . Oct. 20 1 840,000 7 1,960,000 5 520,000 a.m. n. " 14 3 6 1,620,000 2,160,000 12 p.m. 1,880,000 11 1,240,000 3 1,820,000 p.m. 7 1,420,000 7 1,640,000 9 11,200 11 1,280,000 11 1,120 a.m. " 21 6 9,000 " 15 6 1,420,000 a.m. 10 1,260,000 " 22 9 1,880,000 p.m. P.m. i 540,000 i 800,000 4 860,000 2.30 3,800 8 920,000 6 8,200 a.m. 9 5,640 " 16 2 760,000 a.m. 7 1,280,000 " 23 4 5,000 11 940,000 8 2,200 £.!». ». 6 1,450,000 12 152,000 a.m. p.m. " 17 1 750,000 3 52,600 6 900,000 9 40,000 11 1,240,000 " 24 a.m. p.m. 6 3,920 9 840,000 8 2,820 a.m. 10 470,000 " 18 6 p.m. 750,000 p.m. 2 94,000 3 330,000 8 218,000 8 1,060,000 11 300,000 10 510,000 " 25 a.m. a.m. 5 92,000 « 19 4 720,000 n. 6 620,000 12 114,000 £.m. p.m. i 400,000 2 2,920 7 520,000 4 2,880 9 180,000 66 TYPHOID CARREERS AND TYPHOID IMMUNITY TABLE XXVI— Concluded. Name. Date. Time. Total No. of colonies per cc. Name. Date. Time. Total No. of colonies per cc. a.m. p.m. Mueller .... Oct. 26 1 158,000 Mueller . . . Oct. 28 6 300 5 240,000 9 50 10.30 124,000 m. p.m. 12 220 6 1,000 a.m. a.m. " 29 10 240 " 27 i 800,000 " 30 * 7 1,860 " 31 * 11 236,000 Nov. 1 * p.m. " 2 * 2 470,000 " 3 * 8 166,000 " 4 • a.m. " 5 * " 28 7 800 " 6 * 10 650 " 7 * 71. " 8 • 12 1 920 1 * All specimens. TABLE XXVII. Name. Date. Time. Specific grav of urine. ity Total No. of typhoid colonies - per cc. « . m. Nov. 5 9 1.022 27,000 10 1.020 8,200 11 1.022 1.250 12 1.022 340 p.m. 1 1.014 5,130 2 1.016 100 3 1.020 380 4 1.016 10 5 1.022 220 6 1.026 20 7 1.024 8 1.020 170 | 9 1.020 220 A. L. GARBAT 67 that showed no typhoid bacteria. From this particular patient, cultures were made separately of all urine specimens voided for the 30,000 CD en Chart 4. Hourly excretion of urine (Table XXVII). 16 days after that and not one of them showed typhoid bacteria. Such an abrupt and permanent cessation of the typhoid bacilluria is unusual. TABLE XXVIH. Acidity of urine in Name. Date. Time. terms of cc. 0.1 N NaOH per 100 cc. of urine. Total No. colonies per cc. a.m. cc. Nov. 16 8 9 45 26 8,000 12,000 10 21 14,200 11 24 8,400 12 26 16,000 p.m. 1 17 880 2 65 8,400 3 26 9,600 4 45 5 20 6 66 7 54 17,500 (0 (SI I Chart 5. Hourly excretion of urine (Table XXVm). 68 A. L. GARBAT 69 All experiments thus far described prove definitely that culturing a single specimen of urine for the presence of typhoid bacteria and concluding from a negative result that typhoid bacilluria does not exist, is absolutely erroneous. Even two consecutive examinations at intervals of a fixed number of days would surely miss a great per- centage of carriers. The Army rule of not considering a patient free of bacteria until 3 consecutive urine cultures at intervals of 6 days are proved negative, appears very stringent. To our great surprise, even this is not a dependable rule if complete freedom from bacteria is sought. This was determined in one of the patients in whom cultures were made separately of every specimen voided (Schubert). The 3 routine urines which were cultured at intervals of 6 days hap- pened to be sterile, whereas other specimens passed within several hours on the same days were positive; furthermore, isolated urine specimens continued to show typhoid bacilli for 2 weeks after the routine third consecutive negative urine. The possibility of erroneous conclusion with the Army rule as guide, was proven in another way. 25 patients were selected who were apparently free of typhoid bacteria. A sterile 24 hour specimen of urine was collected from each patient. Every time that the patient voided, a sample (the last part) was passed into a large sterile bottle. From this mixed 24 hour specimen, 15 to 20 cc. were cultured in a broth flask and 1 cc. was spread on an Endo plate. We found that 5 out of these 25 cases (20 per cent) showed a positive typhoid growth. All of these patients had shown typhoid bacteria in the urine previously, during early convalescence. As will be seen from Table XXIX there was an interval of 5, 6, 8, 9, and 11 days respec- tively from the time of the last or third consecutive negative urine until the time of the positive 24 hour specimen. In 3 of these 5 patients, the bacteria in the 24 hour specimen were so numerous that a growth was obtained even on the Endo plate. In the other 2, the growth was obtained only in the broth flasks. Undoubtedly these 5 patients were excreting typhoid bacteria in the urine intermittently during the time of the 3 consecutive negative urines and even after that, although the urines had been considered no longer infectious. 70 TYPHOID CARRIERS AND TYPHOID IMMUNITY TABLE XXIX. Name. •g-S Pi o 3 3 ■g 0) > 8 p* ' (A ■si P 3 1 at bo 41 ■ used for fixation. ■3* ll — ■*■■ si a.a ca-a ll •eg I] 33 S.S - a ll 8.1 1* is a. s <*> ++ v-t w (O « "- CI fO N •"* cs as. 0.05 ++ ++ ++ ++ + ++ ++ ++ + ++ ++ 0.02 ++ • + ++ + + + ++ ++ ++ + ++ ++ 0.01 ++ * + ++ ++ + + ++ ++ + ++ ++ 0.005 + * + + + + + + ++ + + ++ ++ 0.002 + + + + + ± ± ± db + + 0.001 + ± ± + ± ± ± + 0.0005 + ± + ± ± + ' Tests were spoiled. 92 TYPHOID CARRIERS AND TYPHOID IMMUNITY Here again it is proved that the greater the number of living bacteria in the circulation, the greater the number of fixation bodies formed. With the same number of bacteria injected, the longer they remain in the blood the more numerous are the complement fixation bodies that are stimulated. Discussion. — While the author realizes that especially in the study of typhoid fever one should not draw conclusions too readily from experimental results in rabbits, still the problems presented here dealt more with the immunologic reactions to the presence of bacteria than with the anatomical or clinical effects caused by them. These experiments should be repeated in a larger series of animals in order to be certain that our results were not dependent upon the exceptional reactions of a few animals. As far as we have gone, how- ever, the experiments substantiate our hypothesis based upon clinical analysis of the cases of typhoid fever; namely, that the persistence of the complement fixation test depended upon the actual number of living typhoid bacteria that invaded the system and the length of time they remained there. Already in 1916 (78) we offered this explanation for the almost constant presence of the reaction during typhoid fever and the almost constant absence after prophylactic immunization. During typhoid fever, one is inoculated with live bacteria instead of organisms killed by heat, and with a com- paratively much greater number of bacteria than in prophylactic immunization. Comparison between Widal Test and Complement Fixation Test during Convalescence. It has already been established (77) that during the acute stage of typhoid fever, the complement fixation test bears no direct relation- ship to the agglutination test. This also holds true during con- valescence. Comparative tests made during the first 2 to 3 months after normal temperature showed: 89 cases with positive complement fixation test and positive Widal test. 11 " " " " " " " negative « " 42 " " negative " " " " positive " " 13 " " " " " " « negative " « A. L. GARBAT 93 It is important to remember that the complement fixation bodies may persist very much longer than the agglutinins. This is evident from Table XXXIX. TABLE XXXIX. Widal test. Complement fixation test. Time of Name. Dilution. examinations after normal 1:50 1:100 Fricke + ± ± ± ± ± + +++ + + + + +++ + + + + + + + + + + ++++ +++ + months Si 5 Heinricksen 5 3* 5 Sh Muza 5 S The Complement Fixation Reaction as a Diagnostic Test during Convalescence. From a diagnostic standpoint, the complement fixation test made for the first time during convalescence, was of aid in several classes of cases. They were: 1. Cases That Were Diagnosed as Typhoid Clinically but Had No Other Laboratory Corroboration for That Diagnosis (Table XL). — This included 15 patients in whom: (a) Widal test was of no help, because of antityphoid inoculations just previous to the onset of the illness. (b) Blood cultures were negative, probably because the patients were received in the third and fourth week of the disease, (c) The urine and feces showed no typhoid bacteria. The complement fixation test in all of these patients was strongly positive (in 13 it was + + + + ; m l ) -j__f-_|_ ;m l ) -|_-f). A + reaction was not considered sufficiently specific when the complement fixation test alone was to be the support for the diagnosis. These cases comprised 9.2 per cent of our typhoid patients; in them the positive complement fixation reaction was the only laboratory test which corroborated the definite clinical diagnosis of typhoid fever. vt I VPHlllll UAKJtLE.J^3 AISIU iirnuiu jj^uu.^,.! * 1-4 H « * ■a o "I ■» K| ^ 5* J? * J2 i & §> 2 2 «3 ■§ « *s S I I S B h s -8-5 § Ml .3 .S I v "i"8 ■S J o & Si p 09 be a> o -S 55 ° ^ j a « "3. m 13 5 » s. f 8.2 8? a- S ^ J} o a) q, -m 13 tS a) a 13 I Sj g 2 « si Q<-7S K>.fl -S •i-i ft Q ° 12 Q. TO *i3 slass .23 S Jj Ml' w « S 3 H *s 5 3 O N H H «-l *H *H C^ *-H gJ3 M *-l »H iax -aop3E3JJ0 33l83(j; '3SB3Sip JO X)U3A3g -oqduiA'i jo inso M J ■S °° s + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + 3 •sinspnaoqdjom -Xjod JO )U33 J3J •SI[33 pcO[q 3};qJ4 I«»ox 3 CM ■S B •fl W Sr a »h 2 IS - ■gJI'S J" I 8-1 8 o p 8 8l.fl8 f s p.j Is-e Widal test in dilution 1:50. 4- + + + + + No. of negative blood cultures. No. of typhoid inocu- lations. o o o o o o Dura- tion of fever course. 3 s 00 "1 3 bo < CM CM o CM S o I a. a c s cs u c c 1 1 £ 8 a 1 a: C c ffi a K A. L. GAKBAT 99 8 5 3 % I a a r 11- a 3 I 1 ■3SS3StpjOiC}IMA3S I -oqdauCi jo ;n» a J •siEopnnoqdjoui -iC[Od jo }u» jsj 9 •S n S3 pooiq 9*!t[Ji jBjOi .3 Q O O t» On i •8 3 JJf § I s 3 • B .a a bo •- ill §| 1 K Pa 8 1 o ■aoi^uuituBxa aoijfl •noi^nirawjtD eaoaj •1S»; l«pjAl + + ll 5-s I 60 1 = O J3 I e 0] S5 § I a as a « en a .as si j ; s •a « .2 • a < pa •? w 100 TYPHOID CARRIERS AND TYPHOID IMMUNITY 3. Cases in Which Them Was a Possible Doubt but Which Clinically Did Not Give the Impression of Typhoid Fever, and in Which the Com- plement Fixation Test Was Negative. — The Widal test was of no aid in this group because of previous inoculations; the blood cultures were either negative or were not made because the patients first came under observation when the acute stage of the disease was over and the temperature was normal. There were 5 such instances; the diagnoses were, meningitis, hemiplegia, pleuropneumonia, reactions following inoculation (2 cases) (Table XLIII). All of these had negative fixation tests. Summary. — Adding the number of patients in these 3 groups, 36, one finds that the complement fixation test gave valuable information in 21.5 per cent of the cases diagnosed as typhoid fever during a typhoid epidemic amongst patients only partly, entirely, or not at all immunized by previous inoculation. Especially with the more generalized use of antityphoid inocu- lations, some test is needed to replace the agglutination reaction when the blood culture and other methods are negative. The fixation reaction aims to supply this demand. The Dreyer agglutination method has also been advocated with this object in view (83). Technique of the Complement Fixation Tests. The technique employed in performing the complement fixation tests is the same as that reported in the author's original publication (77). It will be reviewed in part in order to discuss several important additional observations. Antigen. — The preparation of the antigen is the most important element in the reaction. The author's method followed the directions of Wasserman and Citron (84) for preparing artificial aqueous aggressins. Cultures are grown for 24 hours upon agar slants or plates (Kolle's flasks are preferable). The growth is washed off with sterile distilled water, about 1 cc. to an agar slant. The total emulsion is kept in a hot-water bath at a temperature of 60-65 c C. for 24 hours, then transferred to a strong bottle and shaken vigorously with the aid of glass beads for 24 hours. The bacteria are thus thoroughly broken up. The mixture is next centrifugalized for a long time (4 to 8 hours, depending upon the total quan- tity) until all the bacteria have sunk to the bottom and the supernatant fluid is absolutely clear. The latter is carefully pipetted off, and can be preserved for about 6 to 8 months in sealed tubes, not exposed to sunlight or room tempera- A. L. GARBAT 101 ture. Sterile precautions must be taken throughout the preparation. The object of using a total small quantity of distilled water is to eliminate the hemolysis of red blood cells, which may be occasioned by employing large doses of antigen in the test. It is best to titrate the strength of the antigen about 1 week after its prepara- tion as at the end of this period its titer usually remains unchanged. Varying quantities of antigen are mixed with the constant units of complement, hemolysin, and red cells (sheep system), and those amounts are determined which do not of their own accord inhibit hemolysis (Table XLIV). TABLE XLIV. Titration of Antigen. antieen Complement. Hemolysin. 0.2 0.1 0.5 0.02S 0.0125 0.00625 0.5(1 0.5(1 0.5(1 0.5(1 0.5(1 0.5(1 10) 10) 10) 10) 10) 10) 0.5(1:1,000) 0.5(1:1,000) 0.5(1:1,000) 0.5(1:1,000) 0.5(1:1,000) 0.5(1:1,000) Red blood cells. Saline. cc. cc. 0.5(1:20) Up to 2.5. .0.5(1:20) .... 0.5(1:20) 0.5(1:20) 0.5(1:20) 0.5(1:20) .... Result. No hemolysis. u u Complete hemolysis. In accordance with the summation law of Weil and Nakajama (85), the dose of antigen to be employed in complement fixation experiments is represented by one-half of the maximum quantity of antigen that does not of itself bind comple- ment; that is, in the above instance one-half of 0.05 cc, or 0.025 cc, or 0.5 cc. of a dilution 1 to 20. This unit of antigen is then tested with a known typhoid serum to prove that complement fixation is possible. In a recent article (86), studying the comparative antigenic sen- sitiveness of nine typhoid antigens each prepared in a different way, Matsumoto showed that the method of preparation has a decided effect upon the occurrence and degree of complement fixation tests. In his experiments the antigen similar to the author's, prepared from the filtrate of typhoid bacilli autolyzed in distilled water aided by heating at a high temperature, proved least antigenic. Mainly immune sera from rabbits were used for this differentiation and no series of typhoid patients or inoculated persons were studied. At first sight this would speak against the use of such distilled water extract antigens. We do not, however, consider this lessened sensitiveness a disadvan- tage if the test is to differentiate between a serum from a patient with 102 TYPHOID CARRIERS AND TYPHOID IMMUNITY an active typhoid fever and a serum from a person inoculated with typhoid vaccine. In the former individual, the antibodies stimulated by the great excess of typhoid bacilli in the system have been produced in such great numbers that they readily give complement fixation even with a less sensitive antigen. On the other hand, in inoculated persons the number of complement fixation bodies are so few that they do not react with an antigen which is not highly sensitive. In other words, for this purpose it is better to have the antigen not too sensitive. The distilled water extract antigen described by Matsumoto is probably less sensitive than the similar antigen of the author's (Garbat), for Matsumoto places the suspension of bacilli in a water bath at 80°C, while the writer does not allow a temperature over 65°C. Higher temperatures interfere too much with antigenic properties. Matsumoto also noticed that inactivating the human serum at 56°C. for 30 minutes causes a marked reduction in the degree of complement fixation. In the writer's studies, all the sera examined were previously heated thus. Just as with the reduced sensitiveness of the antigen, the writer believes that the reduction in sensitiveness of the serum by its being heated is not a disadvantage when one wishes to differentiate between a serum from an inoculated person and the serum from a typhoid patient. As has been said before, a typhoid fever serum usually has an excess of antibodies and even a slight diminution in their number by heating will still leave sufficient for the reaction. Done in this less sensitive way, as recommended by the author, the complement fixation test when positive is all the more specific for active typhoid fever. If it should be determined by the study of a large number of cases that the more sensitive antigens of Matsumoto react not only with the sera from cases of typhoid fever but also with the sera from inoc- ulated persons, then sensitive antigens combined with unheated serum will have a limited use only in patients who are suffering from a suspicious typhoid illness but who have with certainty not been inoculated, while our less sensitive antigen combined with heated serum will be employed in other suspicious typhoid cases which may have received typhoid inoculations. Then, too, it will have to be A. L. GARBAT 103 shown that the more sensitive antigens do not react with sera from patients with high fever suffering from acute infections other than those of typhoid origin. The less sensitive typhoid antigen has proved specific. As many different typhoid strains should be employed in the pre- paration of the antigen as possible. A highly polyvalent antigen is an element which accounts for a higher percentage of positive com- plement fixation tests (77). Antigens prepared from autogenous strains give the best fixation. The antigen used in the present study contained 15 strains out of a total of 45 strains isolated. It is probable that all the typhoid bacilli were the same, since they were derived from the same epidemic. We were able to employ an auto- genous antigen because the fixation tests were started only during convalescence. Comparative fixation tests done in 17 patients with 2 different antigens, 1 from the autogenous strain and 1 from the Rawling strain, showed 10 positive with the autogenous antigen and only 4 positive with the Rawling antigen. There was no cross fixation with a para- typhoid antigen. Complement. — In the original publication (77), a fixed dose of com- plement, that is 0.05 cc. of the mixed serum from several guinea pigs, was employed for the \ quantity system. This technique has not been changed. Titration of complement was intentionally omitted because it has not as yet been worked out whether titration of com- plement so that no excess is used, may not make the reaction too delicate and thus non-specific. A patient with typhoid fever usually produces so many antibodies that the fixation test is not overshadowed by a slight excess of complement. Water bath incubation for 1 hour was used. Hemolytic System. — The sheep corpuscles system was continued for the same reason. Similarly, the natural antisheep amboceptors in human serum were not taken into account as is done in the Wasserman test. 2 units of hemolysin were used. After the system was added, the tests were read in 20 minutes. CONCLUSIONS. 1. This report is based upon the study of a typhoid epidemic in- cluding 183 patients, and probably arising from polluted drinking water. Laboratory tests were performed in only 164 cases. 2. The typhoid illness was divided into "the acute fever stage" and "convalescence," the first day of the persisting normal tempera- ture being used to separate the two periods. 3. The Endo medium, properly prepared, is very satisfactory for distinguishing the typhoid and the colon bacilli. II. 4. Feces cultures employed for detecting the presence of typhoid bacteria in the intestines, fail in 15 per cent of typhoid carriers. Direct examination of the bile by duodenal cultures is the only ac- curate means for the detection of the carriers. In IS per cent of typhoid convalescents, typhoid bacilli may be discovered in the bile after 3 consecutive stool cultures have been negative and they are not destroyed in the intestines but merely escape detection by stool examination. 5. Feces cultures should not, however, be entirely discarded because of the existence (although very rare) of pure intestinal carriers, in which typhoid bacilli occur in the stool and not in the bile. One such definite intestinal carrier is reported. 6. An absolutely safe indication of the complete absence of typhoid bacteria in the intestinal tract would be offered by 2 consecutive negative bile cultures and 2 consecutive negative feces cultures. No special interval of days between these examinations is required. III. 7. 32 per cent of typhoid patients become feces carriers. Of these, 28 to 29 per cent are temporary carriers while 3 to 4 per cent become permanent carriers. 104 A. L. GAKBAT 105 8. Of the temporary carriers, 17.5 per cent are carriers for 1 month. 8.0 " " " " " 2 months. 3.0 " " " " " 3 " 9. Three distinct types of feces carriers have been found, according to the original source of the infection: (a) liver, (b) gall bladder, (c) intestinal. In the first two only are the bacilli transmitted by the bile. Combinations of these types may result, if the carrier state continues. 10. During early convalescence, one may- attempt by stool cultures to differentiate between bile carriers (liver and gall bladder) and intes- tinal carriers. In the intestinal carriers, the bile cultures show no typhoid organisms, while the stool cultures show almost a pure growth of the bacilli; the bacilli appear in the stool at the beginning of convalescence and continue there persistently. In the bile carriers, the feces cultures may show only few typhoid colonies, or even none at all, while the bile shows typhoid bacilli usually in pure culture. 11. As an explanation for the origin of bile carriers, ascending in- fection of the gall bladder or liver from the intestines is possible, al- though the usual path is by way of the blood and a descending route. 12. Cholecystectomy is curative in pure gall bladder carriers only. Liver and intestinal carriers are uninfluenced by cholecystectomy. 13. Instead of cholecystectomy as the routine operation, in carriers cholecystectomy should be combined with hepatic drainage both for diagnostic and therapeutic purposes. Long continued drainage of the hepatic duct should be carried out in liver carriers. IV. 14. No factors can be cited as definitely predisposing to the carrier condition. 15. Patients who present symptoms of cholecystitis or cholelithiasis during the course of the typhoid illness, usually become carriers, but not all the carriers necessarily present symptoms referable to the gall bladder during the acute illness. 106 TYPHOID CARRIERS AND TYPHODD IMMUNITY 16. Surgical measures to remove the carrier, state should be under- taken very much earlier in those carriers who have had a complicating cholecystitis. 17. Cholelithiasis following typhoid fever may only be associated with and not necessarily induced by the typhoid bacillus. On the other hand, gall stones of unusually large size may be formed during the typhoid illness and originate from the typhoid bacillus. V, Part A. 18. The use of solid media alone (Endo plates) for detecting typhoid bacilli in urine is inadequate. A 24 hour growth in broth is first necessary before plating. 19. The excretion of typhoid bacteria in the urine is of an inter- mittent character. The urine may change from a specimen with no bacteria to one with very many bacteria within several hours. 20. It is misleading to consider the urine of patients free from typhoid bacteria on the basis of 3 consecutive negative exami- nations at 6 day intervals; such a rule fails in 20 to 25 per cent of urine carriers. 21. A negative culture in broth from a 24 hour specimen of urine sterilely collected is the only safe guide. V, Part B. 22. The Urine is as important a factor as the feces in the spread of typhoid fever. 23. 49 per cent of patients present typhoid bacilluria during con- valescence. 24. There is no relationship between typhoid bacilluria and the specific gravity, acidity, or albumin content of the urine. 25. Bacilluria is most frequent during the first week after normal temperature, when as many as 63 per cent of cases show typhoid bacilli in the urine. 26. 6.7 per cent of all typhoid cases, or 13.6 per cent of all cases of bacilluria, remain positive for 1 to 2 months after the absence of fever. In only 1.2 per cent of all typhoid cases, or 2.5 per cent of all positive cases, does the bacilluria continue for 2 to 3 months. A. L. GARBAT 107 27. No definite explanation can be given for the intermittency with which the typhoid bacilli are excreted. It was suggested that an infection in the pelvis of the kidney or its calyces might account for the condition. Under these circumstances, the pelvis of the kidney bears the same relation to the kidney as does the gall bladder to the liver. VI. 28. In 30 per cent of typhoid patients, the complement fixation bodies disappear from the blood by the 2nd month of convalescence, while 40 per cent of patients still show a positive fixation at the end of 6 to 8 months after the temperature has become normal. 29. A persisting complement fixation test is dependent directly upon the number of bacteria which have invaded the body and the length of time they remained there. 30. In 80 per cent of patients with recurrences or relapses, and 85 per cent of carriers, a strong complement fixation test persists. How- ever, not all carriers, not even all permanent ones, necessarily continue to give the fixation test. 31. In 9.2 per cent of our cases of typhoid fever the complement ^fixation examination was the only laboratory test which confirmed the clinical diagnosis of typhoid fever; the reaction was performed for the first time as late as 1 to 3 months after the onset of normal tempera- ture. The test was of valuable assistance in 21.5 per cent of our cases. The author again desires to express his appreciation to Colonel F. F. Russell for his constant cooperation and interest in thisproblem. Thanks are also due to the commanding officers at U. S. A. General Hospital No. 12, Lieutenant-Colonel William H. Smith and Lieutenant- Colonel J. B. McCreary, who greatly facilitated the work by placing the clinical material at the author's disposal and granting him all the required laboratory equipment and assistance. Without this assistance this study could not have been completed and therefore sincerest thanks are due to Sergeant Paul Piel for his ever ready aid and his patient and untiring efforts with the duodenal cultures and urine work, as well as to Lieutenant Schleussner, Lieutenant Whiting, Corporal Case, Corporal Rudolf Piel, Miss Nichols, Miss Warner, Sergeant Owens, and Private Wintermeyer. BIBLIOGRAPHY. 1. Vaughan, V. C, Jr., /. Am. Med. Assn., 1920, lxxiv, 1074. 2. Gay, F. P., Typhoid fever considered as a problem of scientific medicine, New York, 1918, 2. 3. Typhoid in the large cities of the United States in 1919, Eighth Annual Re- port, /. Am. Med. Assn., 1920, lxxiv, 674. 4. Schumacher, Klin. Jahrb., 1909, a, xxi, 209; 1909-10, b, xxii, 263. 5. Koch, R., Sitzung des wissenschaftlichen Senats bei der Kaiser-Wilhelms- Akademie, Nov. 28, i902. 6. Frosch, P., Festschrift zum sechzigsten Geburtstage von Robert Koch, Jena, 1903, 691. 7. von Drigalski, quoted by Endo, S., Centr. Bakteriol., lte Alt., Orig., 1904, xxxv, 109. 8. Donitz, W., Festschrift zum sechzigsten Geburtstage von Robert Koch, Jena, 1903, 297. 9. Garbat, A. L., /. Am. Med. Assn., 1916, lxvii, 1493. 10. Medical News, New York City, /. Am. Med. Assn., 1920, lxxiv, 1263. 11. von Drigalski, Centr. Bakteriol., lte Abt., Orig., 1904, xxxv, 776. 12. Jurgens, Z. Mm. Med., 1904, lii, 44. 13. Semple, D., and Greig, E. D. W., Scientific memoirs by officers of the medical and sanitary departments of the government of India, No. 32, Calcutta, 1908. 14. Klinger, Arb. k. Gsndhtsamte, 1909, xxx, 584. 15. Schroder, quoted by Park, W. H., and Williams, A. W., Pathogenic micro- organisms, Philadelphia and New York, 7th edition, 1920, 389. 16. Lentz, Klin. Jahrb., 1905, xiv, 475. 17. Klinger, Arb. k. Gsndhtsamte, 1907, xxv, 214. 18. Hetsch, H., Centr. Bakteriol., lte Abt., Ref., 1909, xliii, 161. 19. Fornet, W., Munch, med. Woch., 1910, lvii, 218. 20. Kayser, H., Arb. k. Gsndhtsamte, 1906, xxiv, 176; 1907, xxv, 223. 21. Park, W. H., J. Am. Med. Assn., 1908, li, 981. 22. Frosch, P., Gafiky, G., Kirchner, M., Kruse, W., von Lingelsheim, W., Loffler, F., and Pfeiffer, R., KUn. Jahrb., 1908, xix, 471. 23. Frosch, P., Klin. Jahrb., 1908, xix, 537. 24. Mayer, G., Centr. Bakteriol., lte Abt., Orig., 1910, liii, 234. 25. Lyon, B. B. V., /. Am. Med. Assn., 1919, lxxiii, 980. 26. Stepp, W., Munch, med. Woch., 1918, lxv, 587. 27. Blachstein, A. G., Bull. Johns Hopkins Hosp., 1891, ii, 96. 28. Welch, W. H., Bull. Johns Hopkins Hosp., 1891, ii, 121. 29. Doerr, R., Centr. Bak'eriol., lte Abt., Orig., 1905, xxxix, 624. 30. Koch, J., Z. Hyg. u. Infectionskrankh., 1909, lxii, 1. 108 A. L. GARBAT 109 31. Chiarolanza, R., Z. Hyg. u. Infectionskrankk., 1909, lxii, 11. 32. Dehler, MUnck. med. Wock., 1907, liv, 779, 2134. 33. Grimme, Milnch. med. Wock., 1908, lv, 16. 34. Lorey, A., MUnck. med. Woch., 1908, lv, 15. 35. Loele, W., Deutsck. med. Woch., 1909, xxxv, 1429. 36. Huismans, L., MUnch. med. Woch., 1909, Ivi, 1956. 37. Kamm, W., Munch, med. Woch., 1909, lvi, 1011. 38. Fromme, A., Deutsch. Z. Chirurg., 1910, cvii, 578. 39. Dehler, MUnch. med. Woch., 1912, lix, 857. 40. Leary, T. J., /. Am. Med. Assn., 1913, lx, 1293. 41. Nichols, H. J., Simmons, J. S., and Stimmel, C. O., /. Am. Med. Assn., 1919, lxxiii, 680. 42. Naunyn, JB., Klinik der Cholelithiasis, Leipsic, 1892. 43. Lentz, Klin. Jahrb., 1905, xiv, 475. 44. Forster, J., Munch, med. Woch., 1908, lv, 1. 45. Anton, B., and Fiitterer, G., MUnch. med. Wock., 1888, xxxv, 315. 46. Droba, S., Wien. klin. Wock., 1899, xii, 1141. 47. Blumenthal, F., Munck. med. Wock., 1904, li, 1641. 48. Levy, E., and Kayser, H., Mttnck. med. Wock., 1906, liii, 2434. 49. Bacmeister, Munck. med. Woch., 1907, liv, 1866; 1908, lv, 211, 283, 339. 50. Cushing, H. W., Bull. Jokns Hopkins Hosp., 1898, ix, 91. 51. Gilbert, A., and Founder, L., Compt. rend. Soc. biol., 1896, xlviii, 155. 52. Gilbert, A., and Fournier, L., Compt. rend. Soc. biol., 1897, xlix, 936. 53. Doerr, R., Wien. Mm. Wock., 1905, xviii, 884. 54. Gay, F. P., Typhoid fever considered as a problem of scientific medicine, New York, 1918, 118. 55. Richardson, M. W., /. Boston Soc. Med. Sc, 1898-99, iii, 79. 56. Vincent, H., and Muratet, L., in Keogh, A., Medical and surgical therapy, i, Infectious diseases, New York and London, 1918, 187. 57. Simon, C. E., Human infection carriers, New York, 1919. 58. Beck, M., and Ohlmuller, W., Arb. k. Gsndhtsamte, 1906, xxiv, 138. 59. Rovsing, T., Klinische und experimentelle Untersuchungen liber die infek- tiosen Krankheiten der Harnorgane, Berlin, 1898. 60. Houston, T., Brit. Med. J., 1899, i, 78. 61. Young, H. H., Jokns Hopkins Hosp. Rep., 1899-1900, viii, 401. 62. Brown, T. R., Med. Rec, 1900, lvii, 405. 63. Petruschky, J., Centr. Bakteriol., lte AM., 1898, xxiii, 577. 64. Gwyn, N. B., Pkiladelpkia Med. J., 1901, vii, 81. 65. Besson, A., Practical bacteriology, microbiology and serum therapy (medical and veterinary), translated from the 5th French edition by Hutchens, H. J., London, 1913, 367. 66. Wood, F. C., Chemical and microscopical diagnosis, New York and London, 3rd edition, 1917, 612. 67. Horton-Smith, P., Lancet, 1900, i, 915. 110 TYPHOID CARREERS AND TYPHOn) IMMUNITY 68. Schiider, V., Deutsch. med. Woch., 1901, xxvii, 762. 69. Cole, R. I., Bull. Johns Hopkins Hasp., 1901, xii, 203. 70. Richardson, M. W., Boston Med. and Surg. J., 1903, cxlviii, 152. 71. Osier, W., and McCrae, T., Modern medicine, Philadelphia and New York, 2nd edition, 1913, 142. 72. Greaves, F. L. A., Brit. Med. J., 1907, ii, 75. 73. Meyer, K., and Ahreiner, G., Mitt. a. d. Grenzgeb. d. Med. u. Chirwg., 1909, xix, 486. 74. Cagnetto, G., and Zancan, A., Beitr. path. Anat. u. allg. Path., 1907, xli, 434. 75. Horton-Smith, P., Goulstonian Lectures, Brit. Med. J., 1900, i, 827. 76. Pick, W., Mttnch. med. Woch., 1910, lvii, 1523. 77. Garbat, A. L., Am. J, Med. Sc, 1914, cxlviii, 84. 78. Garbat, A. L., /. Immunol., 1916, i, 387, 391. 79. Felke, Mttnch. med. Woch., 1915, lxii, 578. 80. Hage and KorS-Petersen, Deutsch. med. Woch., 1915, xli, 1328. 81. Schone, C, Mttnch. med. Woch., 1908, Iv, 1063. 82. Henderson-Smith, quoted by Ledingham, J. C. G., and Arkwright, J. A., The carrier problem in infectious diseases, London and New York, 1912, 125. 83. Dreyer, quoted by Krumbhaar, E. B., and Smith, W. B., /. Infect. Dis., 1918, xxiii, 126. 84. Citron, J. B., Immunity, methods of diagnosis and therapy and their practical application, translated by Garbat, A. L., Philadelphia, 2nd edition, 1914. 85. Weil, E., and Nakajama, H., Mttnch. med. Woch., 1906, liii, 1001. 86. Matsumoto, M., /. Immunol., 1920, v, 111. INSTITUTE FOR MEDICAL RESEARCH Under the head of Monographs of The Rockefeller Instilulefor Medical Research are published from time to time scientific papers which are so extensive, or require such elaborate illustrations, as to render them unsuitable for current periodical issues. The Monographs are published at irregular periods, determined by the available material on hand. A number of free copies, which may be 50 or less according to the size of the Monograph, are given to the author. Monographs will be sent post paid on application, payable in advance. Remittances should be made by draft or check on New York, or by postal money order, payable to The Rockefeller Institute for Medical Research, Avenue A and 66th Street, New York, N.Y. No. 1. Simon Flexner and J. W. Jobling. Studies upon a transplantable rat tumor. Plates 1—16. J. W. Jobling. The biology of the mixed tumor of the rat. Maud L. Menten. Experi- ments on the influence of radium bromide on a carcinomatous tumor of the rat. J. W. JbBLiNG. Spontaneous tumors of the mouse. Plates 17-28. J. W. Jobling. Transplan- tation experiments in Macacus rhesus with a carcinomatous teratoma from man. (Issued June 30, 1910.) Price, $1.00. No. 2. William Travis Howard and Oscar T. Schultz. Studies in the biology of tumor cells. Plates 1-6. (Issued February 15, 1911.) Price, $1.00. No. 3. Benjamin T. Terry. Chemo-therapeutic trypanosome studies with, special reference to the' immunity following cure. (Issued March 15, 1911.) Price,$1.00. No. 4, Francis W. Peabody, George Draper, and A. R. Dochez. A clinical study of acute polio- myelitis. Plates 1-13. (Issued June 24, 1912.) Price, $1.00. No. 5- J. P. Slmonds. Studies in Bacillus vielchii, with special reference to classification and to its relation to diarrhea. (Issued September 27, 1915.) Price, $1.00. No. 6. James L. Stoddard and Elliott C. Cutler. Torula infection in man. Plates 1-9. (Issued January 31, 1916.) Price, $1.00. No. 7. Oswald T. Avery, H. T. Chickertng, Rufus Cole, and A. R. Dochez. Acute lobar pneu- monia. Prevention and serum treatment. Plates 1-3. (Issued October 16, 1917.) Price, $1.00. No. 8. Ernest C. Dickson. Botulism. A clinical and experimental study. Plates 1-10. (Issued July 31, 1918.) Price, $1.00. No. 9. James Howard Brown. The use of blood agar for the study of streptococci Plates 1-34. (Issued January 21, 1919.) Price, $2.00. No. 10. William G. MacCallum. The pathology of the pneumonia in the United States Army campsduring the winter of 1917-18. Plates 1-53. (Issued April 16, 1919.) Price, $1.50. No. 11. Frederick M. Allen, Edgar Sttllman, and Reginald Fitz. Total dietary regulation in the treatment of diabetes. 62 charts. (Issued October 15, 1919.) Price, $4.50. No. 12. Lewis H. Weed, Paul Wegeforth, James B. Ayer, and Lloyd D. Felton. A study of experimental meningitis. A series of papers from the Army Neuro-Surgical Laboratory. Plates 1-17. (Issued March 25, 1920.) Price, $L00. No. 13. Werner Marchand. The early stages of Tabanida: (horse-flies). Plates 1-15. (Issued November 15, 1920.) Price, $2.00. No. 14. S. T. Darling and W. G. Smtllie. Studies on hookworm infection in Brazil, First paper. (Issued February 1, 1921.) Published for the International Health Board. No. 15. I. J. Kligler. Investigation on soil pollution and the relation of the various. types of privies to the spread of intestinal infections. (Issued October 10, 1921.) Published for the Inter- national Health Board. No. 16. Abraham L. Garbat. Typhoid carriers and typhoid immunity/ (Issued May 10, 1922.) Price, $1.25. No. 17. Wilson G. Smtllie. Studies on hookworm infection in Brazil. Second paper. (In press.) Published for the International Health Board. No. 18. PA-Levene. Hexosamines, their derivatives, and mucins and mucoids. (In press.) All inquiries, correspondence, and subscriptions should be sent to the DIVISION OF PUBLICATION THE ROCKEFELLER INSTITUTE FOR MEDICAL RESEARCH AVENUE A AND «6th STREET, NEW YORK, N. Y. PUBLICATIONS OF THE ROCKEFELLER INSTITUTE tf£ FOR MEDICAL RESEARCH THE JOURNAL OF EXPERIMENTAL MEDICINE Edited by SIMON FLEXNER, M.D. The Journal of Experimental Medicine is designed to cover the field of experimental medicine. It is issued monthly, two volumes of over 600 pages each appearing in a year. The Subscription price per year (two volumes) is $5.00, payable in advance. Single copies cost 75 cents. Remit- tances should be made by draft or check on New York, or by postal money order, payable to The Journal of Experimental Medicine, Mount Royal and Guilford Avenues, Baltimore, Md., or Avenue A and 66th Street, New York, N. Y. THE JOURNAL OF BIOLOGICAL CHEMISTRY Edited by STANLEY R. BENEDICT, New York, N. Y. LAFAYETTE B. MENDEL, New Haven, Conn. HENRY D. D AKIN, Scarborough, N. Y. DONALD D. VAN SLYKE, New York, N. Y. The Journal of Biological Chemistry is designed foT the prompt publication of original investigations of a chemical nature in the biological sciences. It is issued monthly. Each volume consists of as many numbers as are required to make a total of between 500 and 600 pages. The subscription - price is $3.00 per Volume to subscribers in the United States and Canada, payable in advance; $3.25 to those in other countriesi The number of volumes issued yearly averages four. The price of single copies varies; it is given on the back cover of each number. Remittances should be made by draft or check on New York, or by postal money order, payable to The Joiirnal of Biological Chemistry, Mount Royal and Guilford Avenues, Baltimore, Md., or Avenue A and 66th Street, New York, N. Y. THE JOURNAL OF GENERAL PHYSIOLOGY Edited by JACQUES LOEB, New York, N. Y. W. J. V. OSTERHOUT, Cambridge, Mass. The Journal of General Physiology is devoted to the explanation of life phenomena on the basis of the physical and chemical constitution of living matter. It is issued bimonthly, one volume of about 600 pages appearing in a year. The subscription price per year (one volume) is $5.00, payable in advance. Single copies cost $1.00. Remittances should be made by draft or check on New York, or by postal money order, payable to The Journal of General Physiology, Mount Royal and Guilford Avenues, Baltimore, Md., or Avenue A and 66th Street, New- York, N. Y. STUDIES FROM THE ROCKEFELLER INSTITUTE FOR MEDICAL RESEARCH The published results of investigations conducted in the Institute, or under its grants, are repub- lished at irregular intervals in volumes designated Studies from The, Rockefeller Institute for Medical Research. The subscription price is $2.00 per volume, payable in advance. A special rate of $1.00 per Volume is made to those who subscribe to The Journal of Experimental Medicine, The Journal of Biological Chemistry, or The Journal of General Physiology. Remittances should be made by draft or check on New York, or by postal money order, payable to The Rockefeller Institute for Medical Research, Avenue A and 66th Street, New York, N. Y. MONOGRAPHS OF THE ROCKEFELLER INSTITUTE FOR MEDICAL RESEARCH For information regarding the Monographs see the inside of this cover. SEMIANNUAL LIST The Institute publishes a Semiannual list of all papers by members of the staff and those work- ing under grants from the Institute, stating the title and plate of publication of these reports. The Semiannual List will be sent upon application. All inquiries, correspondence, and subscriptions should be sent to the DIVISION OF PUBLICATION THE ROCKEFELLER INSTITUTE FOR MEDICAL RESEARCH ATBNUB A AND 6*TH STREET) NEW YORK, N. T.