Pediatric preoperative risk factors to predict postoperative ICU admission and death from a multicenter retrospective study.

Abstract

Although some studies have investigated the potential predictors of perioperative mortality, there are few specifically for pediatrics. The aim of the retrospective study was to analyze potential preoperative risk factors and to develop a pediatric preoperative risk prediction score (PRPS), and to predict postoperative ICU admission and the incidence of perioperative death in pediatric patients. Patients who postoperatively admitted to ICU or died (occurred within 30 days after the surgery) from 263 607 pediatric surgical patients with age from 1 day to 14 years old in eight centers in China from October 2010 to September 2013 were retrospectively analyzed. About 5500 non-ICU admission and death patients were randomly selected from those 263 607 patients as controls for analysis comparison. Independent risk factors and a risk model were derived from these analyses, and were further assessed with the likelihood ratio test and the area under the receiver operating characteristic (ROC) curve. There were 1812 ICU admission or death patients but 187 patients' records are incomplete. There were 487 patients with incomplete records among 5500 controls. Collectively, data from 6626 patients were enrolled in final analyses. With multiple logistic regression analysis, age, ASA physical status, SpO2 , prematurity, and unfasted status were found to be independent predictors for critical patients. The AUC value of 0.905 indicated excellent predictive performance between critical and noncritical predictors. Our study revealed that age, ASA physical status, SpO2 , prematurity, and unfasted status are risk factors to predict postoperative ICU admission and death in pediatric patients.