Aim: Pulmonary embolism is a life threatening disease and one of the main causes of in-hospital mortality. There are multiple risk stratification scoring systems for the forecast of the outcomes in patients with pulmonary embolism.
Material and methods: We conducted a prospective observational cohort study to evaluate the comparative validity of four prognostic models: pulmonary embolism severity index (PESI), simplified PESI score, PESI risk classes and shock index, for predicting short-term mortality in acute pulmonary embolism. The PESI and PESI--simplified scores and shock-index were calculated. PESI scores were segregated into risk class (I-V) obtaining PESI classes. Shock index was dichotomized into 0 (for value < 1) versus 1 (for value > 1) risk groups. We determined the area under Receiver Operating Characteristic curve, the sensitivity, specificity, likelihood ratio (LR+, LR-) for PESI score, simplified PESI score, PESI risk classes and shock index, to compare the ability of these scoring tools.
Results: The cohort consisted of 362 subjects. The mortality was 21.54%. The sensitivity, specificity and likelihood ratios were different between scores but the area under Receiver Operating Characteristic curve showed values in range 0.6-0.7.
Conclusions: The assessed prognostic scores performed comparably for identifying pulmonary embolism patients at short-term risk of mortality, even for PESI score segregated in risk classes and dichotomized version of shock index.