Factors Associated With Time in Therapeutic Range Among Thai Patients with Non-valvular Atrial Fibrillation Taking Warfarin: Development of AHfHDP Score

Abstract

Background In Thai patients with non-valvular atrial fibrillation (NVAF), anticoagulation quality with warfarin remains suboptimal. Existing prediction tools for poor international normalized ratio (INR) control, such as the SAMe-TT2R2 score, perform poorly in this population. Identifying population-specific determinants of time in therapeutic range (TTR) may improve risk stratification and inform clinical decision-making. Methods We conducted a single-center, retrospective cohort study of adult patients with NVAF taking warfarin between 2016 and 2019. TTR was calculated using the Rosendaal method, with a target INR range of 2.0–3.0. Logistic regression analysis was performed to identify independent predictors of suboptimal anticoagulation control (TTR <65%). A novel risk score was derived from the regression coefficients and evaluated using receiver operating characteristic (ROC) curve analysis. Results Among 1,322 patients (median age 68 years; 57% female), median TTR was 50.5% (IQR 29.8–69.0), and 70% had TTR <65%. Independent predictors of suboptimal TTR included age ≥65 years, heart failure, hypertension, and drug–drug interactions with warfarin, while prior warfarin use ≥1 year was protective (all P<0.05). These variables were incorporated into the AHfHDP score (range, −2 to 4). The AHfHDP score showed modest discrimination (C-statistic 0.64; 95% CI 0.61–0.67), outperforming SAMe-TT2R2 (C-statistic, 0.53). Conclusions Warfarin anticoagulation control in Thai patients with NVAF remains suboptimal. The AHfHDP score demonstrated modest predictive performance and superior discrimination compared to the SAMe-TT2R2 score. However, its overall predictive ability remained limited, and external validation is required before clinical application.