Use of Platelet Indices for Determining Illness Severity and Predicting Prognosis in Critically Ill Patients

BACKGROUND Decreased platelet (PLT) count is one of the independent risk factors for mortality in intensive care unit (ICU) patients. This study was to investigate the relationship between PLT indices and illness severity and their performances in predicting hospital mortality. METHODS Adult patients who admitted to ICU of Changzheng Hospital from January 2011 to September 2012 and met inclusion criteria were included in this study. Univariate analysis was used to identify potential independent risk factors for mortality. Multiple logistic regression analysis was used to calculate adjusted odds ratio for mortality in patients with normal or abnormal PLT indices. The relationship between PLT indices and illness severity were assessed by the Acute Physiology and Chronic Health Evaluation II (APACHE II) scores or sequential organ failure assessment (SOFA) scores in patients with normal and abnormal PLT indices. The performances of PLT indices in predicting mortality were assessed by receiver operating curves and diagnostic parameters. The survival curves between patients with normal and abnormal PLT indices were compared using Kaplan-Meier method. RESULTS From January 2011 to September 2012, 261 of 361 patients (204 survivors and 57 nonsurvivors) met the inclusion criteria. After adjustment for clinical variables, PLT count <100 × 10 12 /L (P = 0.011), plateletcrit (PCT) <0.108 (P = 0.002), mean platelet volume (MPV) >11.3 fL (P = 0.023) and platelet distribution width (PDW) percentage >17% (P = 0.009) were identified as independent risk factors for mortality. The APACHE II and SOFA scores were 14.0 (9.0-20.0) and 7.0 (5.0-10.5) in the "low PLT" tertile, 13.0 (8.0-16.0) and 7.0 (4.0-11.0) in the "low PCT" tertile, 14.0 (9.3-19.0) and 7.0 (4.0-9.8) in the "high MPV" tertile, 14.0 (10.5-20.0) and 7.0 (5.0-11.0) in the "high PDW" tertile, all of which were higher than those in patients with normal indices. Patients with decreased PLT and PCT values (all P < 0.001), or increased MPV and PDW values (P = 0.007 and 0.003, respectively) had shortened length of survival than those with normal PLT indices. CONCLUSIONS Patients with abnormally low PLT count, high MPV value, and high PDW value were associated with more severe illness and had higher risk of death as compared to patients with normal PLT indices.