Assessment of the Pediatric Index of Mortality 2 with the PaO2/FIO2 ratio derived from the SpO2/FIO2 ratio: A prospective pilot study in a French pediatric intensive care unit

Objective: The Pediatric Index of Mortality 2 is a mortality prediction tool used in pediatric intensive care units. Arterial blood gas sampling that is required to calculate the PaO2/FIO2 ratio is often unavailable. Several authors have proposed mathematical relations between the PaO2/FIO2 and SpO(2)/FIO2 ratios. The main objective of this study was to assess the validity of the Pediatric Index of Mortality 2 score and three modified Pediatric Index of Mortality 2 scores with the PaO2/FIO2 ratio calculated from the SpO(2)/FIO2 ratio. Design: Prospective observational study of consecutive patients admitted during a 23-month period. Setting: A multidisciplinary French pediatric intensive care unit. Patients: One thousand forty-three patients, 80 of whom died (7.7%). Interventions: None. Measurements and Mains Results: Only 15 of 1043 patients had a PaO2 measurement. SpO(2)/FIO2 ratio was determined for 338 patients (33%) for whom FIO2 was known and SpO(2) was <= 97%. However, for mathematical reasons, PaO2/FIO2 ratio could be calculated from the three SpO(2)/FIO2 equations for 328 (31%), 295 (28%), and 100 (10%) patients, respectively. Discrimination assessed by the area under the receiver operating characteristic curve was 0.86 for Pediatric Index of Mortality 2 and the modified scores. Calibration assessed by the Hosmer-Lemeshow goodness-of-fit test was poor for Pediatric Index of Mortality 2 (p = .04) and one modified score (p = .03) and good for two other modified scores (p = .06, p = .09). Conclusion: This study suggests that the SpO(2)/FIO2 ratio could be used in place of PaO2/FIO2 for calculating Pediatric Index of Mortality 2. This must nonetheless be confirmed by a larger prospective multicenter study. (Pediatr Crit Care Med 2011; 12:e184-e186)