Validation of Simplified Acute Physiology Score II and Simplified Acute Physiology Score III as mortality and morbidity risk models

Introduction

Simplified Acute Physiology Score (SAPS) II and SAPS III are valuable scoring systems used to predict the risk of mortality in the ICU. The main purposes of this study were to assess the strength of both the scores in predicting the mortality risk, to check which one gained the best performance, and to ascertain whether they also achieved a good level of morbidity risk prediction in our patient population.

Methods

Two hundred and forty-one patients (82 female, 159 male, mean age 53 ± 23), out of 393 admitted to our ICU over 12 months, were included in the study. Exclusion criteria were age <18, and ICU stay <24 hours. SAPS II and SAPS III were prospectively collected. The derived probability of death was calculated according to the original descriptions. To evaluate the capability of morbidity risk prediction of these scoring models, the GIVITI definitions (Italian Group for the Evaluation of Interventions in the ICU) of organ dysfunction and failure were used. Univariate and multivariate analyses were applied. The area under the receiver operating characteristic curve (AUC) was calculated.

Results

Median values for SAPS II and SAPS III were 35 and 58, respectively. Mortality and morbidity rates resulted 16% and 40.5%, respectively. The mean ICU stay was 9 ± 10 days. Although the univariate analysis found similar statistical significances (P < 0.001) for both the scores, at the multivariate analysis only SAPS II maintained a statistical significance (P < 0.001) in predicting the probability of death. Moreover, SAPS II showed a significantly higher AUC (0.91 vs. 0.73, SAPS II vs. SAPS III, respectively, P < 0.05). With respect to morbidity, the AUC of both the scores gained a poor level of predictive power (AUCs = 0.618 and 0.605, SAPS II and SAPS III, respectively). However, in the subgroup of patients admitted to our ICU for major trauma (41% of patients), SAPS III reached a high degree of discriminative strength in predicting the morbidity rate (AUC = 0.81, P < 0.01).

Conclusion

SAPS II seemed a suitable tool in predicting the risk of death, but not morbidity. Conversely, SAPS III, even though it did not gain a sufficient degree of predictive power toward mortality, could be a useful morbidity risk model when applied to major trauma patients. In our ICU, SAPS II and SAPS III are not interchangeable, but they should be carefully used in the right circumstances. This might avoid the pitfalls leading to errors in forecasting mortality and morbidity risk.