P1.002: Do Pediatric Sepsis Scores Accurately Predict Mortality? Prediction of Outcomes in Critically Ill Pediatric Cancer Patients

Background: Infection is a major cause of death in pediatric cancer patients admitted to the pediatric intensive care unit (PICU). Various sepsis scoring systems have been validated, including: Pediatric Organ Failure Assessment (pSOFA), Pediatric Logistic Organ Dysfunction Score (PELOD-2), and Quick Sequential Organ Failure Assessment (qSOFA). It is unclear if these scoring systems are valid in the pediatric oncology population due to differences in immunosuppression and baseline organ dysfunction.

Methods: A retrospective analysis was conducted on pediatric oncology patients admitted to the Vanderbilt PICU (2007–2024) with sepsis or suspected sepsis. Sepsis scores (pSOFA, PELOD-2, and qSOFA) were calculated using the worst values from PICU admission through 24 hours after admission. 

Scoring systems were calculated using the following values:

pSOFA: PaO2/FiO2, platelets, bilirubin, mean arterial blood pressure (MAP), dose of dopamine/epinephrine/norepinephrine, Glasgow coma score (GCS), serum creatinine

PELOD-2: Age, MAP, lactate, GCS, pupil exam, PaO2/FiO2, PCO2, mechanical ventilation, white blood cell (WBC) count, platelets, serum creatinine 

qSOFA: Respiratory rate, systolic blood pressure, GCS

Additional demographic and clinical data were collected to identify factors associated with PICU mortality. Univariate analysis was performed using Spearman’s correlation to assess associations between continuous variables and Wilcoxon Rank Sum tests to compare differences between groups. Logistic regression was done and the discriminatory ability of each scoring system to predict PICU mortality was evaluated using the area under the receiver operating characteristic (ROC) curve (C-statistic).

Results: Among 105 episodes of ICU admission for sepsis or suspected sepsis, 84 patients remained in the PICU 24 hours after admission. Mortality rate in this population was 16.7%. The pSOFA was associated with mortality (Odds ratio [OR] 1.76, p= < 0.01) and had a C-statistic of 0.88., while the PELOD-2 was also associated with mortality (OR 1.67, p= < 0.01) and had a C-statistic of 0.87. pSOFA and PELOD-2 though correlated with each other (Spearman rho 1.44, p=0.02), when combined into one model and adjusted for being severely neutropenic yielded a C-statistic of 0.91. qSOFA scores were not significantly associated with mortality.

Other factors present at 24 hours that were associated with eventual PICU mortality were acute viral infection (OR 3.77, p = 0.045), elevated lactate (OR 1.25, p = < 0.01), having severe neutropenia (OR 8.88, p=0.041) and needing mechanical ventilation (OR 5.61, p = < 0.01).

Conclusion: PELOD-2 and pSOFA scores at 24 hours post PICU admission were strongly predictive of mortality in this pediatric oncology cohort while qSOFA lacked significance. The mortality predictions were improved after adjusting for the presence of severe neutropenia. This study indicates that using these sepsis scores with adjustment for severe neutropenia could yield superior risk stratification.