Obesity and risk of bleeding: the SMART study.

Background: Obesity is associated with increased levels of procoagulant factors and decreased fibrinolytic activity. Whether this hemostatic profile protects against clinically relevant bleeding has been scarcely investigated. Objectives: To assess the impact of measures of obesity on the risk of bleeding in a large cohort of patients at increased atherothrombotic risk. Methods: The Second Manifestation of ARTerial disease (SMART) study included 9736 patients aged 18-79 years, followed for a median of 5.9 years. Body mass index (BMI), waist circumference and hip circumference were measured at inclusion. All incident fatal or non-fatal hemorrhagic events were recorded. Results: During follow-up, 359 first bleeding events occurred. In quintile-based analyses, the risk of bleeding was highest in the lowest quintile (Q) of BMI (age and sex-adjusted HR Q2 vs. Q1, 0.68; 95% CI, 0.50-0.94), but there was a threshold effect at low BMI levels (men, < 23.84 kg m(-2); women, < 22.49 kg m(-2)), and the risk estimates for bleeding did not further change across the remaining quintiles (HR Q3 0.81 and Q5 0.75). For waist circumference the relationship appeared to be U-shaped, with the lowest risk of bleeding in quintile 3 (HR Q3 vs. Q1, 0.69; 95% CI, 0.46-1.04). Adjustments for hypertension, hemoglobin level, renal failure, diabetes and use of oral anticoagulants or platelet inhibitors did not affect the results. Conclusion: Obesity was not associated with lower risk of bleeding. Our findings suggest that presumed protection against bleeding due to an apparently efficient hemostatic system may be counterbalanced by other factors in obese subjects.