Type of Anticoagulant Used After Total Knee Arthroplasty Affects the Rate of Knee Manipulation for Postoperative Stiffness

Background: The aim of this study was to perform a population-level analysis on the effect of different types of anticoagulation on postoperative stiffness after total knee replacement, requiring manipulation under anesthesia. We hypothesized that patients receiving warfarin would have a higher rate of manipulation under anesthesia compared with patients receiving low-molecular-weight heparin. We also hypothesized that aspirin, direct factor Xa inhibitors, and fondaparinux would have no effect on the rate of manipulation under anesthesia. Methods: Using the PearlDiver patient database, we analyzed 32,320 patients who underwent a primary unilateral total knee replacement from 2007 to 2015. Patients were included if they filled a prescription for anticoagulation medication within 2 days of their discharge and were excluded if they were taking a prescription anticoagulation medication except for aspirin) in the 3 months before total knee replacement. The primary outcome was manipulation under anesthesia performed within 6 months after a primary total knee replacement. Results: The most commonly prescribed postoperative anticoagulation was warfarin 38.0%), followed by low-molecular-weight heparin 33.8%). There were 1,178 patients 3.64%) who underwent manipulation under anesthesia within 6 months of total knee replacement. In multivariable analysis using low-molecular-weight heparin as a comparison group and accounting for age, sex, comorbidities, and length of stay, there was a significant increase in the risk of manipulation under anesthesia for patients who received warfarin hazard ratio [HR], 1.17 [95% confidence interval CI), 1.01 to 1.36]; p = 0.032), direct factor Xa inhibitors HR, 1.42 [95% CI, 1.20 to 1.66]; p < 0.001), or fondaparinux HR, 1.33 [95% CI, 1.01 to 1.72]; p = 0.038). Although patients who received aspirin had the same risk estimate as patients who received warfarin, there was not a significantly increased risk of manipulation under anesthesia in patients who received aspirin compared with low-molecular-weight heparin HR, 1.17 [95% CI, 0.72 to 1.80]; p = 0.493). Conclusions: We found an increased rate of manipulation under anesthesia after total knee replacement in patients who received oral anticoagulants including warfarin, direct factor Xa inhibitors, and fondaparinux, in comparison with patients who received aspirin or low-molecular-weight heparin. We recommend that patients receiving oral anticoagulants after total knee replacement should be counseled about associated stiffness. Furthermore, surgeons should take these data into account when selecting thromboprophylaxis for patients after total knee replacement.