Prognostic risk-stratified score for predicting mortality in hospitalized patients with thrombotic thrombocytopenic purpura: nationally representative data from 2007 to 2012.

Abstract

Despite proven efficacy and increased availability of therapeutic plasma exchange (TPE), mortality for patients with thrombotic thrombocytopenic purpura (TTP) remains high with a limited understanding of those at highest risk of death. This study utilized the Nationwide Inpatient Sample (2007-2012) to derive a prognostic score for mortality in hospitalized TTP patients. Odds ratios of death with various putative risk factors adjusted for age, sex, and race were calculated (adjOR). Weighted mean of adjOR estimates were incorporated in a risk-stratified score. Among 8203 hospitalizations with TTP as primary admission diagnosis who underwent TPE, 613 deaths were identified (all-cause mortality, 7.5%; median time-to-death, 9 days; interquartile range, 4-14 days). In multivariable logistic regression, arterial thrombosis (adjOR 6.7, 95% confidence interval [CI], 1.1-40.9), intracranial hemorrhage (adjOR, 6.1; 95% CI, 1.6-23.2), age at least 60 years (adjOR, 3.5; 95% CI, 2.1-5.6), renal failure (adjOR, 2.6; 95% CI, 1.5-4.5), ischemic stroke (adjOR, 2.4; 95% CI, 1.2-5.0), platelet (PLT) transfusions (adjOR, 2.2; 95% CI, 1.2-4.1), and myocardial infarction (adjOR, 2.3; 95% CI, 1.2-4.6) were significant independent predictors of mortality in TTP patients who underwent TPE. A prognostic weighted mortality prediction scoring system incorporating arterial thrombosis, intracranial hemorrhage, age, renal failure, ischemic stroke, PLT transfusion, and myocardial infarction showed very good discrimination and was predictive of 78.6% deaths. Early and targeted therapy for high-risk individuals should be used to guide management of TTP patients for improved survival outcomes.