Gastrointestinal haemorrhage in atrial fibrillation patients: impact of quality of anticoagulation control

Aims Gastrointestinal haemorrhage is a common complication of antiplatelet and anticoagulation therapy. The aim of this study is to evaluate the impact of warfarin with different time in therapeutic ranges (TTRs) on the risk of major gastrointestinal haemorrhage in atrial fibrillation (AF) patients compared with aspirin. Methods and results In this observational study, 5426 Chinese AF patients (77.7 +/- 10.7 years, female: 53.1%) with a CHA(2)DS(2)-VASc score of >= 1 were included in the final analysis: of which, 3832 patients (70.6%) were taking aspirin and 1594 patients (29.4%) were taking warfarin, whereas the remaining patients did not receive any anticoagulation. The mean baseline HAS-BLED score was 2.22 +/- 0.93. Among those on warfarin, the median TTR was 39.2%. After a mean follow-up of 3.6 years (19 777 patient-years), 262 patients developed a gastrointestinal haemorrhage requiring transfusion (4.83%), with an annual incidence of 1.32%/year. Annual incidences of gastrointestinal haemorrhage requiring transfusion among patients on aspirin and warfarin were 1.53%/year and 1.00%/year, respectively. For patients on warfarin, the incidence of gastrointestinal haemorrhage increased progressively with higher HAS-BLED scores, from 0.93%/year for those with a HAS-BLED score of <= 1 to 1.68%/year for those with a HAS-BLED score of >= 3, and decreased progressively with an increasing TTR from 1.69%/year for patients in the lowest quartile of TTR to only 0.51%/year for those in the top quartile. Conclusion Aspirin is overall associated with a higher risk of gastrointestinal haemorrhage compared with warfarin, despite the suboptimal TTR in the study population. For patients on warfarin, a HAS-BLED score at baseline prior to commencing anticoagulation positively correlates with gastrointestinal haemorrhage. A poor TTR was associated with a higher risk of gastrointestinal haemorrhage.