In a recent article in CHEST (February 2010), Mirsaeidi and colleagues1 found that thrombocytosis (platelet count > 400,000 cells/μL) at the time of hospital admission was a strong predictive factor of mortality in a population of 500 patients with community-acquired pneumonia (CAP). These results contrast with those of numerous previous studies in which thrombocytopenia was the main platelet disorder associated with worse clinical outcome.2-4 We believe this conclusion could be explained by the moderate severity of CAP. Among the studied population, only 58% of patients belong to pneumonia severity index risk class 4 and 5, and admission to an ICU concerned only 17.2% of patients. Mortality was relatively low at 10.8% of patients. The authors recommended new studies focusing on the cause of death to determine if an elevated platelet count in patients with CAP is just a marker of the inflammatory response or if it is in part responsible for an increase in mortality. We have recently published a multicenter retrospective study showing, in 822 patients admitted to an ICU for severe CAP, that severe thrombocytopenia (< 50,000 cells/μL) was an independent predictor of mortality.5 We looked at the impact of thrombocytosis in our patients. The overall ICU mortality rate was 35.4%. Thrombocytosis was present in 70 (5.7%) patients. Among these patients, 24 (34.3%) died. We did not find any difference in outcome compared with patients with thrombocytosis (P < .7). Our patients were more severely ill, with mechanical ventilation required within 12 h following ICU admission in 77.6% of patients and septic shock present in 30.4% of patients. When considering the cause of death according to platelet numbers, we found it was essentially related to sepsis complications in patients with thrombocytopenia (septic refractory shock, n = 18; multiorgan failure, n = 17; ARDS, n = 11; nosocomial pneumonia, n = 8), while in patients with thrombocytosis, the cause of death was mostly related to complications of ICU stay or associated comorbidity (P < .007; COPD, n = 3; cerebral vascular ischemia, n = 2; cancer, n = 2; mesenteric ischemia, n = 1; myocardial infarction, n = 1) (Table 1).