Validation of the IMPROVE Bleeding Risk Score

Recent guidelines recommend assessing medical inpatients for bleeding risk prior to providing chemical prophylaxis for venous thromboembolism (VTE). The International Medical Prevention Registry on Venous Thromboembolism (IMPROVE) bleeding risk score (BRS) was derived from a well defined population of medical inpatients but it has not been externally validated. We sought to externally validate the IMPROVE BRS. We prospectively collected characteristics on admission and VTE prophylaxis data each hospital day on all patients admitted for a medical illness to the Walter Reed Army Medical Center (WRAMC) over an 18 month period. We calculated the IMPROVE BRS for each patient using admission data and reviewed medical records to identify bleeding events. From September 2009 through March 2011 there were 1668 inpatients who met the IMPROVE inclusion criteria. Bleeding events occurred during 45 (2.7%) separate admissions; 31 (1.9%) events were major and 14 (0.8%) were non-major but clinically relevant. 256 (20.7%) patients had an IMPROVE BRS ≥ 7.0. Kaplan-Meier curves showed a higher cumulative incidence of major (p=0.02) and clinically important (major plus clinically relevant non-major) (p=0.06) bleeding within 14 days in patients with an IMPROVE BRS ≥ 7.0. An IMPROVE BRS ≥ 7.0 was associated with major bleeding in Cox-regression analysis adjusted for administration of chemical prophylaxis (OR 2.6, 95% CI: 1.1-5.9; p=0.03); there was a trend toward significant association with clinically important bleeding (OR 1.9, 95% CI: 0.9-3.7; p=0.07). The IMPROVE BRS calculated at admission predicts major bleeding in medical inpatients. This model may help assess relative risks of bleeding and VTE before chemoprophylaxis is administered.