Abstract

Our aim is to investigate the correlation between the mean platelet volume (MPV) levels and platelet counts of the septic children with 28-day mortality risk. MPV at admission (MPV adm ), MPV at 72 < sup > nd < /sup > hour (MPV 72h ) and the difference between these two parameters (ΔMPV 72h-adm ) and platelet counts were recorded retrospectively. The control group consisted of 100 healthy children matched for age, gender, and ethnicity. One hundred eighty six children were enrolled into the study. The study group had significant higher MPV values than those of control group. There were 156 survivors and 32 nonsurvivors in study group with a mortality rate of 17%. Nonsurvivors had significantly higher MPV adm (p: 0.001), MPV 72h (p: 0.001), ΔMPV 72h-adm (p < 0.001) and lower platelet count (p: 0.002) than survivors. MPVadm (OR 2.39), MPV 72h (OR 4.23), ΔMPV 72h-adm (OR 6.4), platelet count (OR 7.3), and need for mechanical ventilation support (OR 9.76) had significant effect on 28-day mortality risk at logistic regression analysis. At the receiver operating characteristic analysis (ROC) the cutoff values for MPV adm , MPV 72h , and ΔMPV 72h-adm were found to be 9 fL, 9.86 fL, and 0.79, respectively. Kaplan-Meier analysis and log-rank test proved that these cutoff values were significantly associated with the time of survival. Septic children who had high MPV levels at admission and whose MPV levels increased during follow up had higher risk of mortality. With the results of further researches targetting large groups of pediatric patients, MPV adm , MPV 72h , and ΔMPV 72h-adm values can be fast and reliable markers for early diagnosis of sepsis and mortality prediction.

Keywords: child, mean platelet volume, mortality, sepsis