Glasgow Coma Scale: variation in mortality among permutations of specific total scores

Objective: The objective of this study was to determine whether different s core permutations of the Glasgow Coma Scale (GCS) giving the same GCS total score were associated with significantly different mortality. Design: For each GCS total we compared the mortality associated with each o f the different GCS permutations using a Fisher's exact test. The relations hip between components of the GCS score and mortality was also examined usi ng uni- and multivariate logistic regression. Setting: Data were collected from the intensive care unit at Wellington Hos pital, a multidisciplinary, tertiary referral unit. Patients: We analysed the GCS and mortality data from all admissions over a 4 year period (January 1994-January 1998). Patients with GCS scores of 3 o r 15 were excluded, since these two total scores do not have multiple permu tations, leaving 1390 patients with GCS scores of 4-14 for analysis. Results: The incidence of mortality was significantly different for the dif ferent permutations for total GCS scores of 7, 9, 11 and 14. Conclusions. It is possible for patients to have the same total GCS score, but significantly different risks of mortality due to differences in the GC S profile making up that score. This suggests that GCS scores may be more u seful reported in terms of profiles rather than totals. This could also hav e implications for the use of other scoring systems such as Acute Physiolog y and Chronic Health Evaluation and Simplified Acute Physiology Score.