Comparison of APACHE II, MEES and Glasgow Coma Scale in patients with nontraumatic coma for prediction of mortality

There are numerous prehosital descriptive scoring systems, and it is uncertain whether they are efficient in assessing of the severity of illness and whether they have a prognostic role in the estimation of the illness outcome (in comparison with that of the prognostic scoring system Acute Physiology and Chronic Health Evaluation [APACHE] II). The purpose of the present study was to assess the value of the various scoring systems in predicting outcome in nontraumatic coma patients and to evaluate the importance of mental status measurement in relation to outcome.

In a prehospital setting, postintervention values of the Mainz Emergency Evaluation System (MEES) and Glasgow Coma Scale (GCS) were measured for each patient. The APACHE II score was recorded on the day of admission to the hospital. This study was undertaken over a 2-year period (from January 1996 to October 1998), and included 286 consecutive patients (168 men, 118 women) who were hospitalized for nontraumatic coma. Patients younger than 16 years were not included. Their age varied from 16 to 87 years, with mean ± standard deviation of 51.8 ± 16.9 years. Sensitivity, specificity and correct prediction of outcome were measured using the χ² method, with four severity scores. The best cutoff point in each scoring system was determined using the Youden index. The difference in Youden index was calculated using the Z score. For each score, the receiver operating characteristic (ROC) curve was obtained. The difference in ROC was calculated using the Z score. P < 0.05 was considered statistically significant.

For prediction of mortality, the best cutoff points were 19 for APACHE II, 18 for MEES and 5 for GCS. The best cutoffs for the Youden index were 0.63 for APACHE II, 0.61 for MEES and 0.65 for GCS. The correct prediction of outcome was achieved in 79.9% for APACHE II, 78.3% for MEES and 81.9% for GCS. The area under the ROC curve (mean ± standard error) was 0.86 ± 0.02 for APACHE II, 0.84 ± 0.06 for MEES and 0.88 ± 0.03 for GCS. There were no statistically significant differences among APACHE II, MEES and GCS scores in terms of correct prediction of outcome, Youden index or area under ROC curve.

APACHE II is not much better than prehospital descriptive scoring systems (MEES and GCS). APACHE II and MEES should not replace GCS in assessment of illness severity or in prediction of mortality in nontraumatic coma. For the assessment of mortality, the GCS score provides the best indicator for these patients (simplicity, less time-consuming and effective in an emergency situation.