Usefulness of RBC distribution width and C-reactive protein to predict mortality in pediatric non-cardiac critical illness.

INTRODUCTION: We aimed to assess the performance of red blood cell distribution width (RDW), C-reactive protein (CRP) or the combination of both to predict clinical outcomes in pediatric non-cardiovascular critical illness. MATERIALS AND METHODS: We analyzed 404 pediatric non-cardiovascular critically ill patients admitted to pediatric intensive care unit (PICU). Potential predictors were identified using multivariable logistic regression. We also calculated the power of RDW and CRP additive to pediatric critical illness score (PCIS) to predict mortality with calculation of C-index value, integrated discrimination improvement (IDI) and net reclassification improvement (NRI) indices. RESULTS: RDW and CRP independently predicted PICU mortality. The C-index value of PCIS with respect to prediction of PICU mortality was greater than that of RDW and CRP. The combination of RDW or CRP or both with PCIS did significantly increase C-index value for predicting mortality (all p < 0.01). Addition of RDW or CRP or their combination to PCIS provided IDI of 7%, 1.1% and 9.4% (p = 0.009, 0.01 and 0.003) and NRI of 15.9%, 13.1% and 19.6% (p = 0.002, 0.043 and 0.002), respectively. CONCLUSIONS: In pediatric non-cardiovascular critically ill patients, RDW and CRP could serve as independent predictors of PICU mortality and addition of RDW or CRP or both to PCIS significantly improves the ability to predict PICU mortality.