Postoperative opioid utilization associated with revision risk following primary shoulder arthroplasty

Abstract

With a substantial increase in utilization of primary shoulder arthroplasty, it is important to understand risk factors that may signal early failure and need for revision. Recent studies have reported that sustained postoperative opioid use is associated with a higher revision risk after total hip or knee arthroplasty. In this study, we evaluated postoperative opioid utilization as a risk factor for revision after primary shoulder arthroplasty. We conducted a cohort study using data from a United States integrated health care system's Shoulder Arthroplasty Registry. Patients who had a primary elective shoulder arthroplasty were identified (2009-2017); those with cancer or who underwent other arthroplasty procedures (either shoulder, hip, or knee) within the preceding year were excluded. Cumulative daily opioid utilization during the first year postoperative, calculated as oral morphine equivalents (OME), was categorized into 3 exposure groups: high user (≥15 mg OME daily), moderate user (<15 mg OME daily), and no opioid use (reference group). The exposure window was stratified into 2 time periods: postoperative days 1-90 and postoperative days 91-360. Multivariable Cox proportional-hazards regression was used to evaluate the association between postoperative opioid use and aseptic revision risk. The final study sample included 8325 shoulder arthroplasty procedures. Of these individuals, 3707 (45%) received some opioid within the 1 year before the index procedure. We failed to observe a difference in aseptic revision risk between opioid utilization in the first 90 days postoperatively, regardless of dose. After the first 90 days, a higher revision risk was observed for high opioid users compared with nonusers (hazard ratio = 1.62, 95% confidence interval = 1.10-2.41), and no association was observed for moderate users (hazard ratio = 1.25, 95% confidence interval = 0.82-1.91). We found a positive association between opioid consumption and aseptic revision risk after primary shoulder arthroplasty. This study cannot determine if opioids have a direct physiological cause that increases the risk of revision; rather it is likely that opioid consumption is a marker of chronic pain, poor function, and/or poor coping mechanisms. Further study is needed to determine if programs designed to decrease opioid use may impact revision risk after shoulder arthroplasty.