Background/Purpose: Total knee arthroplasty (TKA) is an effective treatment for patients with end-stage knee osteoarthritis (OA); however, a growing body of evidence suggests that patients using opioids prior to TKA experience worse pain outcomes immediately following surgery. We evaluated the pain relief achieved from TKA in knee OA patients with and without opioid prescriptions pre-TKA.

Methods: We augmented patient-reported data from a longitudinal study of TKA outcomes with medical record-based information about opioid use. The study included patients >40 years old who underwent primary, unilateral TKA at a tertiary academic medical center. The Pain Catastrophizing Scale (PCS) and patient-reported outcomes, including the Western Ontario and McMaster Universities Arthritis Index (WOMAC), were assessed at baseline and 6 months post-TKA. Demographic and comorbidity information were collected at baseline via study questionnaire. We reviewed the electronic medical record of each subject to identify opioid use from two years pre- to one year post-TKA. We documented the dates of first and last opioid prescription and any additional surgeries done during the study period. We restricted our analyses to those who did not have additional surgery during the study period and who had no more than 2 comorbidities. To address confounding by indication we built a propensity score of opioid use based on age, PCS, comorbidities, and pain prior to TKA. The propensity score ranged from 10% to 75% likelihood of opioid use. The primary analysis included subjects with propensity scores from 20-75%. We compared WOMAC Pain scores (0-100, 100 worst) at 6 months after TKA in persons without opioid use to those with documented pre-TKA opioid use with a general linear model, adjusting for propensity score, PCS, and baseline pain.

Results: Our analytic cohort consisted of 134 patients with mean age 68 years (SD 8); 66% were female and mean pre-TKA WOMAC Pain was 43 (SD 16). Nineteen percent had at least one opioid prescription prior to TKA. Baseline WOMAC Pain scores were higher among those who used opioids pre-TKA (49 vs. 41, p=0.04). PCS was substantially greater among pre-TKA opioid users (18 vs. 12, p=0.004). Adjusted analyses showed that the opioid group had mean 6-month WOMAC pain of 18 points, while the non-opioid group had mean WOMAC Pain of 10 points (p = 0.0017). The adjusted difference in pain from pre-TKA to 6 month post-TKA was estimated at 25 WOMAC points among those who use opioids prior to TKR compared to 33 WOMAC points among those who did not use opioids (p=0.0017).

Conclusion: Patients with pre-TKA opioid use had higher levels of pain catastrophizing and WOMAC Pain at baseline, and, after adjusting for differences in baseline pain, pain catastrophizing, and likelihood of opioid use, had higher pain 6 months after TKA compared to patients without preoperative opioid use. With the substantial societal burden associated with opioid use, clinicians and policymakers may consider limiting the use of these analgesics prior to TKA to optimize the efficacy of the procedure.