key: cord-0743356-fwfoga4x
authors: Tay, Hui Ping; Wang, Xinyi; Narayan, Sujita W; Penm, Jonathan; Patanwala, Asad E
title: Persistent postoperative opioid use after total hip or knee arthroplasty: A systematic review and meta-analysis
date: 2021-09-19
journal: Am J Health Syst Pharm
DOI: 10.1093/ajhp/zxab367
sha: 6b87bac9749040e0bb11318132b2141690e748ba
doc_id: 743356
cord_uid: fwfoga4x

DISCLAIMER: In an effort to expedite the publication of articles related to the COVID-19 pandemic, AJHP is posting these manuscripts online as soon as possible after acceptance. Accepted manuscripts have been peer-reviewed and copyedited, but are posted online before technical formatting and author proofing. These manuscripts are not the final version of record and will be replaced with the final article (formatted per AJHP style and proofed by the authors) at a later time. PURPOSE: To identify the proportion of patients with continued opioid use after total hip or knee arthroplasty. METHODS: This systematic review and meta-analysis searched Embase, MEDLINE, the Cochrane Central Register of Controlled Trials, and International Pharmaceutical Abstracts for articles published from January 1, 2009, to May 26, 2021. The search terms (opioid, postoperative, hospital discharge, total hip or knee arthroplasty, and treatment duration) were based on 5 key concepts. We included studies of adults who underwent total hip or knee arthroplasty, with at least 3 months postoperative follow-up. RESULTS: There were 30 studies included. Of these, 17 reported on outcomes of total hip arthroplasty and 19 reported on outcomes of total knee arthroplasty, with some reporting on outcomes of both procedures. In patients having total hip arthroplasty, rates of postoperative opioid use at various time points were as follows: at 3 months, 20% (95% CI, 13%-26%); at 6 months, 17% (95% CI, 12%-21%); at 9 months, 19% (95% CI, 13%-24%); and at 12 months, 16% (95% CI, 15%-16%). In patients who underwent total knee arthroplasty, rates of postoperative opioid use were as follows: at 3 months, 26% (95% CI, 19%-33%); at 6 months, 20% (95% CI, 17%-24%); at 9 months, 23% (95% CI, 17%-28%); and at 12 months, 21% (95% CI, 12%-29%). Opioid naïve patients were less likely to have continued postoperative opioid use than those who were opioid tolerant preoperatively. CONCLUSION: Over 1 in 5 patients continued opioid use for longer than 3 months after total hip or knee arthroplasty. Clinicians should be aware of this trajectory of opioid consumption after surgery.

A c c e p t e d M a n u s c r i p t Total hip arthroplasty (THA) and total knee arthroplasty (TKA) are among the most common elective procedures. [1] [2] [3] The demand for these surgeries is projected to increase substantially in the next decade. 4 Postoperative pain after these surgeries is often severe, and opioid use in the immediate postoperative period is indicated. [5] [6] [7] However, there is lack of information regarding how long patients continue to require opioids after hospitalization. Increasing duration of opioid use after surgery has been associated with opioid dependence, abuse, and overdose. 8 Thus, international guidelines for acute postoperative pain recommend that opioids should be used only when necessary, at the lowest effective dose, and for the shortest duration. 9, 10 A better understanding of the trajectory of opioid use following THA or TKA is needed to help guide clinicians and patients for care planning and harm prevention.

Identifying patients who are susceptible to prolonged opioid use allows processes to be implemented to mitigate this risk. In patients undergoing THA or TKA, the surgery itself is meant to alleviate pain by correcting the underlying cause. Thus, prolonged use of opioids postoperatively is not routinely expected. 11 Guidance endorsed by the US Centers for Disease Control and Prevention suggest opioid prescribing for less than 14 days for severe pain after THA or TKA. 11 Also, chronic severe postsurgical pain has been shown to occur in 6% to 15% of patients after THA or TKA. 12 Previous reviews have assessed the prevalence of postoperative opioid use after general surgery and trauma. [13] [14] [15] However, there are no systematic reviews or meta-analyses regarding this topic in patients undergoing THA or TKA. Therefore, the primary aim of this systematic review and meta-analysis was to identify the proportion of adult patients taking opioids at 3 to 12 months after THA or TKA. Our secondary objective was to compare duration of postoperative opioid use between patients who were opioid naïve and those who were opioid tolerant preoperatively.

A c c e p t e d M a n u s c r i p t Methods Protocol and registration. The review was developed and reported following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement. 16 The review protocol was registered in the PROSPERO international prospective register of systematic reviews (registration number CRD42020145241). 17 There was no funding source for this systematic review and meta-analysis. c c e p t e d M a n u s c r i p t postoperative follow-up period was less than 3 months or involved cancer or palliative care or if the data for THA or TKA could not be extracted separately.

After removal of duplicate publications, titles and abstracts were screened for potentially relevant studies. Full-text articles were retrieved and reviewed to identify studies that met the eligibility criteria. Two reviewers independently performed database searches and assessed the eligibility of studies. Any inconsistencies were resolved through discussion with a third reviewer.

Data extraction. Data extracted included author, year of publication, country of origin, study design, data source, sample size, patient age, surgical procedures, duration of follow-up, definition of opioid use, and the proportion of patients taking an opioid at postoperative time points defined for each study. The opioid data was collected for the total sample and also for predefined subgroups. The subgroups of interest were those who had preoperative opioid consumption (ie, opioid-tolerant patients) and those who did not have preoperative opioid consumption (ie, opioid-naïve patients). One reviewer extracted the data, which was verified by a second reviewer. Inconsistencies between reviewers were resolved through discussion with a third reviewer.

Risk of bias assessment. The quality of included studies was assessed independently by 2 reviewers (H.P.T. and X.W.) using the Newcastle-Ottawa Quality Assessment Scale for cohort studies, with a maximum score of 9 stars. 18 The scale has 3 domains: (1) selection of the study groups, (2) comparability of the groups, and (3) ascertainment of outcome. A maximum of 9 stars can be awarded to studies of the highest quality. Discrepancies regarding quality assessment were resolved through discussion or in consultation with a third reviewer.

A c c e p t e d M a n u s c r i p t

The primary outcome measure was the duration of postoperative opioid use, reported here as the proportion of patients taking opioids at 3 months, 6 months, 9 months, or 12 months following THA or TKA. The secondary outcome was the proportion of patients with prolonged postoperative opioid use among opioid-naïve and opioid-tolerant patients.

The definition of prolonged or persistent opioid use has not been established and varies substantially between studies. 19 Based on the definition of chronic postsurgical pain, we used a time period cutoff of 3 months. 12 However, we acknowledge that shorter periods may be more desirable for assessing what is considered optimal use. Quantitative analysis and meta-analysis. All studies that satisfied our selection criteria and research question were included in the systematic review. The results were first categorized by type of surgery (THA or TKA). The overall proportion of patients with opioid use was reported for each time period within these strata. For each type of surgery, the groups were stratified by preoperative opioid use status (opioid naïve or tolerant). The

Metaprop package in STATA, version 16.0 (StataCorp LLC, College Station, TX) was used for the meta-analysis. 20 The results were reported as proportions with 95% confidence intervals (CIs). A random effects model with inverse variance was used. Heterogeneity was determined according to definitions in the Cochrane handbook, with an I 2 value of <40%

indicating "might not be important heterogeneity"; a value of 30% to 60%, "may represent moderate heterogeneity"; a value of 50% to 90%, "may represent substantial heterogeneity"; and a value of >75%, "considerable heterogeneity." 21 Sensitivity analysis. There was considerable heterogeneity present, and we considered that this could be because of the difference between opioid-naïve and opioidtolerant patients. Thus, the results were stratified by preoperative opioid consumption to Some studies specified a minimum threshold of opioid consumption based on dose and duration (eg, a minimum of 20 mg per day of morphine equivalents for 30 continuous days within 6 months after surgery) for a patient to be considered a postoperative opioid user. Participants who did not reach this threshold were considered to be nonusers or a separate group; however, the duration of postoperative opioid use could not be ascertained separately for the latter group. We defined postoperative opioid users as patients who met the specific criteria used in the respective studies and considered the remaining patients as non-opioid users. The same applied to the definitions of preoperative opioid use.

Study selection. The initial search yielded 1,363 articles. After removal of duplicates and screening of titles and abstracts, 82 articles were retrieved for full-text review. Thirty A c c e p t e d M a n u s c r i p t articles 22-51 met the eligibility criteria and were included. Figure 1 shows a PRISMA flow diagram that summarizes the selection process. (95% CI, 12%-21%); at 9 months, 19% (95% CI, 13%-24%); and at 12 months, 16% (95% CI, 15%-16%). There was considerable heterogeneity at each time point (I 2 values of 99%-100%). The proportions of patients in the opioid-naïve and opioid-tolerant subsets are reported in Table 3 and also depicted in the forest plots in the supplementary material. We also conducted a sensitivity analysis by excluding studies involving military or veteran populations; the results were similar (Table 3) . Table 4 and also depicted in the forest plots in the supplementary material. In the sensitivity analysis that excluded studies involving military or veteran populations, the opioid consumption appeared to be lower at 12 months (Table 4) reported outcomes at multiple time points. All studies showed a trend of decreasing opioid use as postoperative duration increased. The decline in the proportion of individuals using opioids between 3 and 6 months postoperatively was larger than the decline among those using opioids between 6 and 9 months or between 9 and 12 months postoperatively. For example, in the study by Namba et al, 43 the proportion of opioid users dropped sharply from A c c e p t e d M a n u s c r i p t 42.1% to 32.2% between 3 months and 6 months and then decreased slightly to 29.6% at 9 months after surgery. Data from all 10 studies are reported in Table 2 .

Preoperative opioid use status. A total of 17 studies 22-24,28-36,39,40,45,48,51 stratified the reporting of postoperative opioid use based on whether or not patients had preoperative opioid consumption. Preoperative opioid users showed a higher risk of continuing postoperative opioid than nonusers at all evaluated time points. For instance, at 3 months after THA, the proportion with continued opioid use was 10% (95% CI, 6%-15%) for opioidnaïve patients and 47% (95% CI, 34%-60%) for opioid-tolerant patients. Similarly, 3 months after TKA, the proportions with continued opioid use were 11% (95% CI, 5%-18%) and 60% (95% CI, 53%-68%) for opioid-naïve and opioid-tolerant patients, respectively. These data are reported for each time point for both THA and TKA in Table 3 and Table 4 , respectively.

In the subgroup of preoperative opioid users, studies that were conducted in veterans or in military settings 33-35 reported higher proportions of postoperative opioid users than studies that were conducted in the general population (Table 2 ).

To our knowledge, this is the first systematic review to focus on the duration of opioid use after THA or TKA. Overall, more than 1 in 5 patients had continued opioid use that was sustained for 12 months postoperatively. However, there were substantial differences between those who were opioid naïve versus opioid tolerant preoperatively. For example, while about 10% of patients in the opioid-naïve group continued opioid use postoperatively, up to half of the patients who were opioid tolerant had continued use.

Thus, this is an important factor that should be considered when determining risk of prolonged postoperative opioid consumption. Existing guidelines generally recommend A c c e p t e d M a n u s c r i p t opioid prescriptions of no more than 3 to 7 days for the treatment of acute pain. 10,52,53 Although recovery after TKA or THA may take longer, opioids should not be prescribed for longer than 14 days for most patients. 11 Given the high volumes of THA and TKA procedures, the findings showed that a large proportion of patients might be taking opioids for longer than the recommended duration, which has the potential for harm.

Regarding trends of opioid use across the various study periods, it was expected that the proportion of patients taking opioids would decrease with time. However, the decline appeared to be greater in the early postoperative period (from 3 to 6 months after surgery).

This suggests that the likelihood of opioid cessation decreases with the duration of postoperative opioid exposure. A possible explanation for the observed slower decline in the rate of opioid use in the late postoperative period is the increased risk of developing opioid dependence and addiction over time. 54 Also, the longer patients continue to take opioids, the more difficult it may be to achieve cessation of therapy. Clinicians may be less likely to successfully cease therapy in these circumstances. Although approximately half of patients with preexisting opioid use had persistent use of opioids post surgery, we would expect this proportion to be lower because chronic severe postsurgical pain has been shown to occur in 6% to 15% of patients after THA or TKA. 12 There was considerable heterogeneity in duration of opioid use between studies. This could be attributed to a lack of consistency in the definitions of postoperative opioid use. A recent study by Jivraj et al 19 reported a more than 100-fold difference in the rates of persistent opioid use when different definitions were applied to the same patient cohort.

The influence of opioid use definitions on the reported incidence of opioid use could have masked the effect of other factors such as patient characteristics and prescribing pattern.

This makes it difficult to compare results across studies. The issue of inconsistency in A c c e p t e d M a n u s c r i p t definitions has also been discussed by authors of previous studies. 13, 14 In relation to this issue, an expert consensus statement 55 has recommended standard definitions of persistent postoperative opioid use based on preoperative opioid use status. For opioid-naïve patients (ie, those with no history of opioid use in the 90 days before surgery), persistent postoperative opioid use was defined as filling a prescription for at least a 60 days' supply of opioids during days 90 to 365 after surgery. For preoperative opioid users, persistent use was defined as any increase in opioid use relative to baseline during days 90 to 365 after surgery. 55 These definitions could be a good starting point to improve the comparability of future studies. We were unable to apply this definition to the reviewed studies because it was not possible to identify patients fulfilling these criteria due to variable study reporting.

Nonetheless, we found great differences in outcomes between opioid-naïve and opioidtolerant patients, which suggests that the definitions used were able to discriminate between patients adequately. The reason for continued opioid use needs to be elucidated. For example, THA and TKA are meant to alleviate joint pain. If surgery is successful, it is anticipated that chronic opioid use would not be indicated.

Among patients who have THA or TKA, over 20% have persistent opioid use for longer than 3 months postoperatively, and this use may be sustained for over 12 months. 

 A better understanding of the trajectory of opioid use following total hip or knee arthroplasty is needed to help guide clinicians and patients in care planning and harm prevention.

 Thirty studies were identified for full review and ultimate inclusion in this study, and the results showed that 1 in 5 patients may continue opioid use for longer than 3 months after total hip or knee arthroplasty.

 Clinicians should be aware of this trajectory of opioid consumption after surgery and focus on deprescribing.

M a n u s c r i p t A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t 

Inpatient hospitalization, surgery, newborn, alternate level of care and childbirth statistics

Australian Institute of Health and Welfare

Most frequent operating room procedures performed in US hospitals

Projected volume of primary total joint arthroplasty in the US

Pain management associated with total joint arthroplasty: a primer

The prevalence of postoperative pain in a sample of 1490 surgical inpatients

Persistent pain after joint replacement: prevalence, sensory qualities, and postoperative determinants

Postsurgical prescriptions for opioid naive patients and association with overdose and misuse: retrospective cohort study

CDC guideline for prescribing opioids for chronic pain -United States

8) b 9 mo: 226 (6.3) 6 mo: 12,070 (16.3) 9 mo

973 3 mo: 32,638 (30.2) b 6 mo: 22,625 (21.0) b 9 mo: 20,775 (19.2) b 12 mo: 20,752 (19.2) b 3 mo: 22,260 (50.1) 6 mo: 17

A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t A c c e p t e d M a n u s c r i p t