Time course of platelet counts in critically ill patients

Background. Although thrombocytopenia in the intensive care unit (ICU) is associated with a poorer outcome, the precise relationship between the time course of platelet counts and the mortality rate has not been well defined. Objective: To describe the time course of the platelet count in relation to the mortality rate in critically ill patients. Design: Substudy of a prospective, multicenter, observational cohort analysis. Setting., Forty ICUs in 16 countries from Europe, America, and Australia. Patients: Data were collected from all ICU admissions in a 1-month period, excluding patients younger than 12 yrs old and those who stayed in the ICU for <48 hrs after uncomplicated surgery. A total of 1,449 critically ill patients were enrolled, including 257 who stayed in the ICU for >2 wks. Interventions: None. Measurements. Platelet counts were collected daily throughout the ICU stay, together with other measures of organ dysfunction. Thrombocytopenia was defined as a platelet count of <150 X 10(3)/mm(3). A relative increase in platelet count was defined as a 25% increase above the admission value, together with an absolute platelet count of greater than or equal to150 X 10(3)/mm(3). Main Results., For the entire population, the platelet count was lower in the 313 nonsurvivors than in the 1,131 survivors throughout the ICU course. Of the 257 patients who stayed in the ICU for >2 wks, 187 (64%) survived. The platelet count decreased significantly in the first days after admission to reach a nadir on day 4 in both survivors and nonsurvivors. In the survivors, the platelet count returned to the admission value by the end of the first week and continued to rise to become significantly greater than the admission value by day 9. In the nonsurvivors, the platelet count also returned to the admission value after 1 wk, but there was no subsequent increase in platelet count. A total of 138 (54%) patients had thrombocytopenia on day 4, and these patients had a greater mortality rate than the other patients (33% vs. 16%; p < .05). On day 14, 51 (20%) patients had thrombocytopenia, and these patients had a greater mortality rate than the other patients (66% vs. 16%; p < .05). Thrombocytopenia was less common on day 14 than on day 4 (20% vs. 54%; p < .05), but the mortality rate was greater in the thrombocytopenic patients on day 14 than those who were thrombocytopenic on day 4 (66% vs. 33%; p < .05). The ICU mortality rate of nonthrombocytopenic patients on day 14 was also significantly lower in patients with, than without, a relative increase in platelet count on day 14 (11% vs. 30%; p < .05). Conclusion. Platelet count changes in the critically ill have a biphasic pattern that is different in survivors and nonsurvivors. Late thrombocytopenia is more predictive of death than early thrombocytopenia. A relative increase in platelet count after thrombocytopenia was present in survivors but not in nonsurvivors. Although a single measured platelet count is of little value for predicting outcome, changes in platelet count over time are related to patient outcome.