Factors at Admission That Are Predictive of Increased In-Hospital Bleeding in Acutely Ill Medical Patients: Findings from the International Medical Prevention Registry on Venous Thromboembolism (IMPROVE).

Abstract

Background Although clinical studies have not shown a significant difference between the risk of bleeding in acutely ill medical patients receiving pharmacologic venous thromboembolism (VTE) prophylaxis and those receiving placebo, fear of bleeding may lead physicians to withhold pharmacologic prophylaxis for patients who should receive it. We therefore aimed to determine the incidence of, and risk factors for in-hospital bleeding in hospitalized acutely ill medical patients in IMPROVE, an international, observational registry.Methods Patients aged ≥18 years, hospitalized ≥3 days with an acute medical illness have been enrolled consecutively since July 2002. Exclusion criteria: therapeutic antithrombotics/thrombolytics at admission, major surgery or trauma during 3 months prior to admission, and VTE treatment within 24 hours of admission. Patients bleeding immediately before, or at admission were excluded from this analysis. Factors present at admission and associated with increased risk of in-hospital bleeding (defined as major or clinically significant nonmajor [Büller et al. N Engl J Med 2003; 349:1695–702]) were identified by univariate analysis (p<0.15) and included in a multiple logistic regression model (significant at p<0.05). The model was adjusted for patients' length of stay in hospital.Results Data were from 5960 patients enrolled up to 31 March 2005 in 49 hospitals (12 countries). In-hospital bleeding occurred in 170 (2.9%) patients: 68 (1.1%) major and 102 (1.7%) clinically significant nonmajor bleeding. Independent risk factors for in-hospital bleeding are shown in the Table. In-hospital prophylaxis with low-molecular-weight and unfractionated heparin were not independently associated with an increased risk of bleeding when added to the analysis (p=0.51 and 0.38, respectively). In patients with 0, 1, 2 or ≥3 of these risk factors, the incidences of major in-hospital bleeding were 0.1%, 0.4%, 1.2% and 5.2%, respectively.Conclusions In this unselected patient population, the rate of major in-hospital bleeding was low (1.1%) and similar to that in the MEDENOX study (1.0%), a major clinical study of VTE prophylaxis. Factors that we identified will be valuable for predicting the risk of in-hospital bleeding in acutely ill medical patients.TableFactors predictive of an increased risk of in-hospital bleeding in acutely ill medical patientsFactorOdds ratio95% confidence intervalActive gastroduodenal ulcer5.382.90–10.00Bleeding disorder4.542.02–10.19Hepatic failure3.341.80–6.19Serum creatinine >1.5 mg/dL2.291.63-3.21Current cancer2.081.43-3.03Central venous catheter2.001.31-3.05ICU/CCU stay1.921.23-3.02Immobile ≥ 4 days1.751.24-2.46Ischemic heart disease1.571.02-2.40