The Effect of Early versus Late Manipulation Under Anesthesia on Opioid Use, Surgical Complications, and Revision Following Total Knee Arthroplasty.

Abstract

Previous studies of early versus late manipulation under anesthesia (MUA) do not report on postoperative opioid utilization or revisions and focused on small single-institution retrospective cohorts. The PearlDiver Research Program (www.pearldiverinc.com), which uses an all-inclusive insurance database, was used to identify patients undergoing primary total knee arthroplasty (TKA) who received (1) late MUA (>12 weeks), (2) early MUA (≤12 weeks), or (3) TKA only. To develop the control group cohort of TKA-only patients, 3:1 matching was conducted using 11 risk factor variables deemed significant by chi-squared analysis. Complications and opioid utilization were compared through multivariate regression analysis, controlling for age, gender, and Charlson Comorbidity Index. The risk of TKA revision was assessed through Cox-proportional hazards modeling and Kaplan-Meier survival analysis with log-rank test. Between 2011 and 2017, 2,062 TKA patients with early MUA, 1,112 TKA patients with late MUA, and a control cohort of 8,327 TKA-only patients were identified in the database. The percent of patients registering opioid use decreased from 54.6% 1 month pre-MUA to 4.6% (p < 0.0001) 1 month post-MUA following early MUA, whereas only from 32.6 to 10.4% (p < 0.0001) following late MUA. Late MUA was associated with higher risk of repeat MUA at 6 months (adjusted odds ratio [aOR] = 2.74, p < 0.0001), 1 year (aOR = 2.66, p < 0.0001), and 2 years (aOR = 2.63, p < 0.0001) following index MUA. Hazards modeling and survival analysis showed increased risk of TKA revision following late MUA (adjusted hazard ratio [aHR] = 3.50, 95% confidence interval [CI]: 2.77-4.43, p < 0.0001) compared to early MUA (aHR = 2.15, 95% CI: 1.72-2.70, p < 0.0001), with significant differences in survival to revision curves (p < 0.0001). When compared to early MUA at 1 year, late MUA was associated with a significantly increased risk of prosthesis explantation (aOR = 2.89, p = 0.0026 vs. aOR = 0.93, p = 0.8563). MUA within 12 weeks after index TKA had improved pain resolution and significant curtailing of opioid use. Furthermore, late MUA was associated with prolonged opioid use, increased risks of revision, as well as prosthesis explantation, supporting screening and early intervention in cases of slow progression and stiffness. The level of evidence of this study is III.