PURPOSE: Available severity of illness scoring systems have not been validated in obstetric patients, and have been shown to overestimate mortality in this population. The purpose of the present study was to estimate the diagnostic utility of APACHE II, SAPS 2, SAPS 3, MPM-II and MPM-III in predicting mortality in obstetric patients admitted to a medical/surgical intensive care unit (ICU).
METHODS: All 2006-2011 obstetric ICU admissions to the Rafael Calvo Maternity Hospital, Cartagena, Colombia (pregnant and up to 6 weeks postpartum) were reviewed. Diagnoses, clinical and laboratory data were collected and severity scores calculated. Approval from the Institutional Review Board was granted. Discrimination was estimated with area under the Receiver Operating Characteristic curve (AUC), and calibration with the Mortality ratio (MR).
RESULTS: 725 women were admitted to the ICU between 2006-2011, for an admission rate of 14.44/1,000 births. The AUC of APACHE II was 0.855 (95%CI, 0.778-0.932), SAPS 2 was 0.865 (95%CI, 0.791-0.940), SAPS 3 was 0.891 (95%CI, 0.833-0.949), MPM-II was 0.866 (95%CI, 0.798-0.934), and MPM-III was 0.915 (95%CI, 0.864-0.965). MR for APACHE II was 0.30 (95%CI, 0.20-0.44), SAPS 2 was 0.51 (95%CI, 0.34-76), SAPS 3 was 0.64 (95%CI, 0.43-94), MPM-II was 0.89 (95%CI, 0.60-1.26), and MPM-III was 1.25 (95%CI, 085-1.78).
CONCLUSIONS: MPM-II and MPM-III appear to be more accurate in predicting mortality in our sample. As previous studies show, APACHE II, SAPS 2 and SAPS 3 overestimate mortality in obstetric critical care patients.
CLINICAL IMPLICATIONS: More studies are needed to validate MPM-II and MPM-III scores in peripartum women to predict mortality and direct care. APACHE II should not be used to predict mortality in this population.