Prevalence of thrombocytosis in critically ill patients and its association with symptomatic acute pulmonary embolism. A multicentre registry study.

It is uncertain whether thrombocytosis without underlying myeloproliferative diseases is associated with an increased risk of acute pulmonary embolism (PE). We investigated the relationship between thrombocytosis and risk of symptomatic acute PE, and whether Pulmonary Embolism Severity Index (PESI) was reliable in predicting mortality of acute PE. This multicentre registry study involved a total of 609,367 critically ill patients admitted to 160 intensive care units (ICUs) in Australia or New Zealand between 2006 and 2011. Forward stepwise logistic regression was used to assess the relationship between risk of acute PE and platelet counts on intensive care unit (ICU) admission. Acute PE (n=3387) accounted for 0.9% of all emergency ICU admissions. Over 20% of all PE required mechanical ventilation, 4.2% had cardiac arrest, and the mortality was high (14.8%). Thrombocytosis, defined by a platelet count >500x10⁹ per litre, occurred in 2.1% of the patients and was more common in patients with acute PE than other diagnoses (3.4 vs. 2.0%). The platelet counts explained about 4.5% of the variability and had a linear relationship with the risk of acute PE (odds ratio 1.19 per 100x10⁹ per litre increment in platelet count, 95% confidence interval 1.06-1.34), after adjusting for other covariates. The PESI had a reasonable discriminative ability (area under receiver-operating-characteristic curve = 0.78) and calibration to predict mortality across a wide range of severity of acute PE. In summary, thrombocytosis was associated with an increased risk of symptomatic acute PE. PESI was useful in predicting mortality across a wide range of severity of acute PE.