Outcome from central nervous system injury

In general, coma portends a very poor outcome in ICU patients except when a drug overdose is the cause. Interactions with concurrent diseases and the cause of the comatose state are important variables that modify mortality rates. Recent studies have made it possible to predict irreversibility with a high degree of specificity in many cases. Three approaches show promise for early prediction of poor outcome--multivariate analysis, neurologic findings, and technologic innovations. Multivariate analysis of combinations of clinical conditions common in ICUs appears to have defined some patients whose prognosis is nearly hopeless, empirically. It would seem reasonable to consider withholding or withdrawing life support from these patients so that mere prolongation of death is avoided. Additional studies are needed and are relatively easy to perform with computerized statistical analyses of a readily available database derived from large patient populations in multicenter studies. A second useful approach for predicting irreversibility has arisen from studies using neurologic findings at defined time points after brain injury. The studies of Levy and colleagues exemplify this approach; their data suggest that coma can be declared irreversible with a high degree of confidence in a large subset of patients, delineated by simple sequential neurologic observations. Finally, technologies are evolving that may lead to noninvasive measurements to provide quantitative evidence of irreversible brain damage. Additional clinical studies are needed to validate the predictive value of these approaches to identifying patients who are certain to have a poor ultimate outcome. It then may be possible to predict function and irreversibility of central nervous system injury in individual cases with a high degree of certainty early in the clinical course.