Thrombocytopenia and thrombocytosis at time of hospitalization predict mortality in patients with community-acquired pneumonia

Abstract

Background: Platelets are inflammatory cells with an important role in antimicrobial host defenses. We speculate that an abnormal platelet count may be a marker of severity in patients with community-acquired pneumonia (CAP). The objectives of this study were to evaluate if abnormal platelet count in hospitalized patients with CAP was associated with 30-day mortality and to compare platelet count and leukocyte count as predictors of 30-day mortality.

Methods: We performed a retrospective cohort study of 500 consecutive patients hospitalized with CAP at the Veterans Hospital of Louisville, Kentucky, between June 2001 and March 2006 to investigate the association of platelet count and leukocyte count with 30-day mortality. Predictor variables were platelet count and leukocyte count. Abnormal platelet count was < 100,000/L (thrombocytopenia) and > 400,000/L (thrombocytosis). The outcome variable was 30-day mortality. To control for potential confounding, a propensity score that incorporated 33 variables was used.

Results: Platelet count was strongly associated (P = .0009) with 30-day mortality, whereas no association was observed for leukocyte count (P = .5114). High platelet counts resulted in a significantly increased risk of mortality.

Conclusions: Thrombocytopenia and thrombocytosis are associated with mortality in patients hospitalized with CAP. When evaluating an initial CBC test in patients with CAP, an abnormal platelet count is a better predictor of outcome than an abnormal leukocyte count.