Background and Objectives. The aim of this study was to investigate the factors predicting Pediatric Intensive Care Unit (PICU) mortality and the outcomes in cancer patients admitted to PICU.

Methods. We conducted a retrospective study in 48 consecutive cancer patients admitted to the PICU between January 1, 2015 and January 1, 2018. A total of 48 patients (21 males and 27 females) were enrolled in this study.

Results. The median age was 77 (33,5-149) months. The median duration of PICU stay was 5 (2-9) days. Patients were classified according to their stage of disease. Ten (20.8%) patients were in the remission group, 9 (18.8%) patients were in the induction period and 29 (60.5%) patients were in the progressive diseasegruops. Thirtynine patients (81.2%) had hematological malignancies, 6 (12.5%) had extracranial solid tumors and 3 (6.3%) had intracranial solid tumors. Thirty-seven patients died and the mortality rate was found to be 77.1%. mortality rates were 11%, 88% and 93% for patients in remission,during induction period and in the progressive disease group, respectively (p < 0.01).The most frequent reasons of PICU admission were respiratory failure in 29 (60.4%), sepsis in 12 (25%), circulatory collaps in 2 (4.2%), and other reasons in 5 patients (10.4%). The median PRISM III among survivors was significantly lower than non-survivors (13.1 ± 6.4; vs. 20.7 ± 5.2; p < 0.001). At a cut-off value of 13, the sensitivity of the PRISM III was 94.4% and the specificity was 58.3% (AUC: 0.821). OSDwas present in 41 (85%) patients, 82% of them died (34/41). The presence of MOF, the use of mechanical ventilation and inotrop support were significantly related with mortality. Univariate logistic regression analysis showed that male gender [odds ratio (OR)=5.588, P= 0.041, 95% confidence interval (95%CI) 1.070-29.191], presence of organ system dysfunction[OR=12.143, P= 0.008, 95%CI 1.947- 75.736], need for mechanical ventilation[OR=34.000, P= 0.001, 95%CI 5.272-219.262], IS [OR=8.5, P= 0.001, 95%CI 1.318-54.817]were the predictors ofhigh mortality in pediatric cancer patients. PRISM III score ≥ 13 was a predictive criteria of PICU mortality.

Conclusion. We conclude that the key to improving survival rates is to pick up on this group of patients as soon as possible.We, believe that cancer patients could be saved by earlier evaluation and intervention by the PICU team when they have a less severe disease.

Keywords: PICU, cancer, pediatric