Risk of major bleeding at different PT-INR ranges in elderly Japanese patients with non-valvular atrial fibrillation receiving warfarin: a nested case-control study

BACKGROUND: Debate continues about the optimal anticoagulation level for elderly Japanese patients with non-valvular atrial fibrillation (NVAF) receiving warfarin. The Japanese Circulation Society guideline has recommended prothrombin time-international normalized ratios (PT-INR) of 1.6 - 2.6 for elderly patients and 2.0 - 3.0 for non-elderly patients, because previous observational studies indicated increased risk of bleeding when the ratio exceeded 2.6. We aimed to reappraise the relationship between PT-INR and the risk of major bleeding in elderly Japanese patients. METHODS: From the electronic medical records, we selected a cohort of elderly (age ≥ 70 years) Japanese patients with NVAF who were prescribed warfarin for the prevention of thromboembolic diseases between November 2010 and March 2014 at Kanto Rosai Hospital. We identified those who developed major bleeding (cases). For each case, we randomly selected two matched controls by adopting a risk-set sampling method defined by calendar date, age, gender, length of warfarin administration, and the prescriber of warfarin. The risk of major bleeding in patients having PT-INR ≤ 1.49, 1.50 - 1.99, 2.00 - 2.49 (the reference), 2.50 - 2.99, and ≥ 3.00 were compared using the conditional logistic regression method. The study protocol was approved by the IRB before the study was begun. RESULTS: Among the cohort of 806 elderly patients, we identified 32 cases and selected 64 matched controls. The overall incidence of major bleeding was 3.5 per 100 patient-years. The odds ratios (95 % confidence intervals) for the risk of developing major bleeding in patients with PT-INR ≤ 1.49 (n = 20), 1.50 - 1.99 (n = 32), 2.00 - 2.49 (n = 18), 2.50 - 2.99 (n = 10), and ≥ 3.00 (n = 16) were 1.0 (0.2, 5.9), 0.3 (0.1, 1.9), 1.0 (reference), 1.2 (0.2, 8.4), and 19.8 (2.0, 198.9), respectively, with a significant difference between ≥ 3.00 and reference. CONCLUSIONS: Among elderly Japanese patients with NVAF, PT-INR 2.0 - 3.0 may be associated with a clinically permissible risk of major bleeding while PT-INR ≥ 3.00 a significant risk. Further studies are warranted to determine whether the risk of major bleeding is significantly lower for PT-INR 2.50 - 2.99 than for PT-INR ≥ 3.00.