Analysis of Anticoagulation Therapy and Anticoagulation-Related Outcomes Among Asian Patients After Mechanical Valve Replacement

IMPORTANCE Current international normalized ratio (INR) guidelines are based on trials involving European and US populations. To our knowledge, no adequate study involving Asian patients has been conducted to date. OBJECTIVE To evaluate the association between INR and anticoagulation-related outcomes in an Asian population after mechanical aortic valve replacement (AVR) or mitral VR (MVR). DESIGN, SETTING, AND PARTICIPANTS This retrospective cohort study was conducted between 2001 and 2018, with follow-up until December 31, 2018, among patients who underwent AVR, MVR, or combined AVR-MVR at 3 medical centers and 4 regional hospitals and contributed electronic medical records to the Chang Gung Research Database. Exclusion criteria were missing demographic characteristics, younger than 20 years, fewer than 2 INR records, and having died during the hospitalization of the index surgery. MAIN OUTCOMES AND MEASURES Bleeding and thromboembolic complications were analyzed. The possibility of nonlinearity and cutoff potential for the INR were explored using a logistic regression model, which considered the INR a restricted cubic spline (RCS) variable. RESULTS The study population consisted of 900 patients, with 525 (58.3%) men and 375 (41.7%) women and a mean (SD) age of 52.0 (12.5) years. Overall, 474 (52.7%) received AVR alone, 329 (36.6%) received MVR alone, and 97 (10.8%) received combined AVR-MVR. All patients had at least 2 INR examinations after discharge, providing 16 676 INR records for the AVR group and 18 207 for the MVR and combined AVR-MVR groups. In the AVR group, the RCS model showed that higher risks of composite thromboembolic events were associated with an INR of less than 2.0 or greater than 2.6 vs an INR of 2.0, and a higher risk of bleeding events was associated with an INR of less than 1.8 or greater than 2.4 vs an INR of 2.0. When treating the INR as a categorical variable, the risk of composite thromboembolic and composite bleeding events was significantly higher among patients with INRs less than 1.5 (adjusted odds ratio [aOR], 2.55; 95% CI, 1.37-4.73) and with INRs of 3.0 or greater (aOR, 3.48; 95% CI, 1.95-6.23) vs those with INRs between 2.0 and 2.5.In the MVR and combined AVR-MVR groups, higher risks of composite thromboembolic events were associated with an INR of less than 2.1 or greater than 2.7 vs an INR of 2.5, and a higher risk of bleeding events was associated with an INR of less than 2.1 or greater than 2.8 vs an INR of 2.5. When treating the INR as a categorical variable, the risk of a composite bleeding events was significantly higher among patients with INRs of 3.5 or greater (aOR, 2.25; 95% CI, 1.35-3.76) vs those with INRs between 2.5 and 3.0. CONCLUSIONS AND RELEVANCE Among Asian patients in this study, the incidence of thromboembolic events in the MVR group with INRs in the range of 2.0 to 2.5 was not significantly higher than that among those with INRs in the range of 2.5 to 3.0; in the AVR group, the incidence for those with INRs in 1.5 to 2.0 range was not significantly higher than for those with INRs in the range of 2.0 to 2.5.