Bleeding events and associated factors in a cohort of adult patients taking warfarin in Sarawak, Malaysia.

Evidence is emerging that rates of adverse events in patients taking warfarin may vary with ethnicity. This study investigated the rates of bleeds and thromboembolic events, the international normalised ratio (INR) status and the relationship between INR and bleeding events in Malaysia. Patients attending INR clinic at the Heart Centre, Sarawak General Hospital were enrolled on an ad hoc basis from May 2010 and followed up for 1 year. At each routine visit, INR was recorded and screening for bleeding or thromboembolism occurred. Variables relating to INR control were used as predictors of bleeds in logistic regression models. 125 patients contributed to 140 person-years of follow-up. The rates of major bleed, thromboembolic event and minor bleed per 100 person-years of follow-up were 1.4, 0.75 and 34.3. The median time at target range calculated using the Rosendaal method was 61.6% (IQR 44.6–74.1%). Of the out-of-range readings, 30.0% were below range and 15.4% were above. INR variability, (standard deviation of individuals’ mean INR), was the best predictor of bleeding events, with an odds ratio of 3.21 (95% CI 1.10–9.38). Low rates of both major bleeds and thromboembolic events were recorded, in addition to a substantial number of INR readings under the recommended target range. This may suggest that the recommended INR ranges may not represent the optimal warfarin intensity for this population and that a lower intensity of therapy, as observed in this cohort, could be beneficial in preventing adverse events.