Use of blood products in patients with anticoagulant-related major bleeding: an analysis of inhospital outcomes.

Abstract

The impact of correcting elevated International Normalized Ratio (INR) values on inhospital mortality in patients with warfarin-associated major bleeding is presented. Using patient information from the database of a large U.S. health system, a retrospective analysis was conducted to (1) evaluate inpatient practice patterns in correcting INR elevations among patients hospitalized with warfarin-related intracranial hemorrhage (ICH) or non-ICH bleeding and (2) test the hypothesis that achieving INR correction, defined as an INR of ≤1.5, at any point during the hospital stay is correlated with lower inhospital mortality. Cox proportional hazards models were constructed to assess predictors of inhospital death. Among the 354 patients who met the study selection criteria, INR correction was achieved in 87.9% overall (92.5% and 85.5% of patients with ICH and non-ICH bleeds, respectively). Patients whose elevated INR values were corrected had significantly lower inhospital death rates than those with uncorrected elevations: 15.3% versus 55.6% (p = 0.010) among patients with ICH and 2.0% versus 11.8% (p = 0.017) among those with non-ICH bleeds. After adjusting for baseline demographics and comorbidities, the correlation between failure to correct INR elevations and increased mortality risk was significant only for patients with ICH (hazard ratio, 8.04; 95% confidence interval, 2.07-31.18; p = 0.003). Results of this study indicated that correction of elevated INR values was associated with a lower likelihood of inhospital death among warfarin-treated patients hospitalized for ICH or non-ICH major bleeding.