OP0174 PARTIAL KNEE REPLACEMENT IS ASSOCIATED WITH A LOWER RISK OF VENOUS THROMBOEMBOLISM AND OPIOID USE THAN TOTAL KNEE REPLACEMENT BUT INCREASED RISK OF LONG-TERM REVISION: A MULTINATIONAL, MULTI-DATABASE, PROPENSITY SCORE-MATCHED, COHORT ANALYSIS INCLUDING OVER 280,000 PATIENTS

Abstract

Background: Knee replacement is one of the few effective treatments for severe knee osteoarthritis. A number of surgical interventions are available, including two main types: partial or total knee replacement. Objectives: We aimed to assess the outcomes of partial compared to total knee replacement. Methods: We conducted a multi-database propensity score-matched cohort study. Data was obtained from 4 US claims databases (IBM MarketScan® Commercial Database (CCAE), IBM MarketScan® Medicare Supplemental Database (MDCR), Optum ® de-identified Clinformatics ® Datamart, Extended - Date of Death (Optum), and Pharmetrics), and 1 UK primary care electronic medical record database (THIN). All people aged 40 years or older at the time of knee replacement surgery were included. Data were mapped to a common data model (OMOP), and processed using analytical tools developed by the OHDSI community. Participants were followed up from the date of their first knee replacement and for up to 5 years. Outcomes included short-term (60-day) post-operative complications (infection, venous thromboembolism, mortality, readmission), opioid use in the 3-12 months post-surgery as a proxy for persistent pain, and 5-year revision risk. Propensity score matching (up to 1:10) was used to control for all available confounders, and negative control outcomes and calibration to minimize the impact of residual confounding. Separate Cox regression models were fitted for each outcome to estimate calibrated HR (cHR) and 95% confidence intervals related to partial, using total knee replacement as a reference group. Results: A total of 32,379 and 250,377 patients receiving partial and total knee replacement respectively were matched. Short-term risk of venous thromboembolism was consistently lower in the former, with cHR ranging from 0.47 [0.32-0.71] in MDCR to 0.76 [0.59-0.99] in CCAE. No differences were observed in the risk of other complications. The use of opioids in months 3 to 12 after surgery was also lower in subjects receiving partial knee replacement, with cHR ranging from 0.70 [0.57-0.90] in THIN to 0.86 [0.78-0.96] in Optum. Conversely, risk of 5-year revision was consistently higher amongst those undergoing partial compared to total knee replacement: cHR ranged from 1.51 [1.24-1.88] in CCAE to 2.16 [1.63-3.15] in MDCR. Conclusion: Partial knee replacement is associated with a 25% to 50% reduction in short-term risk of venous thromboembolism, and a 15%-30% lower risk of persistent pain post-surgery. In contrast, partial knee replacement results in a 50% to >100% excess risk of revision in the longer term. This information should be used to individualize surgical indications for patients with severe knee osteoarthritis. Disclosure of Interests: : Daniel Prieto-Alhambra Grant/research support from: Grants from Amgen, UCB Biopharma and Servier outside the submitted work, Consultant for: UCB Biopharma, Speakers bureau: Amgen, Edward Burn: None declared, James Weaver Shareholder of: Shareholder in Johnson & Johnson, Employee of: Full-time employee of Janssen Research & Development, a pharmaceutical company of Johnson & Johnson., Anthony G Sena Shareholder of: Shareholder in Johnson & Johnson, Employee of: Full-time employee of Janssen Research & Development, a pharmaceutical company of Johnson & Johnson, Henry Morgan Stewart Employee of: Full-time employee of IQVIA, Patrick Ryan Shareholder of: Shareholder in Johnson & Johnson, Employee of: Full-time employee of Janssen Research & Development, a pharmaceutical company of Johnson & Johnson.