This study was designed to determine by multivariate statistical methods the influence of 38 variables on outcome after cardiopulmonary resuscitation (CPR) and to assess neuropsychological status in long-term survivors. The charts of 181 consecutive patients resuscitated in a 1,100-bed University Hospital over a 2-year period were analyzed retrospectively. Of the 181 resuscitated patients, 23 (13%) could be discharged. Outcome was significantly affected by the following variables: presence of shock or renal failure before cardiac arrest (CA) (odds ratio = 10.6; 95% confidence interval = 1.3-85.8 and odds ratio = 13.8; 95% confidence interval = 1.7-109.2, respectively), administration of epinephrine (odds ratio = 11.2; 95% confidence interval = 3.2-39.2) or prolonged CPR (> 15 min) (odds ratio = 4.9; 95% confidence interval = 1.7-13.7). By contrast, when CA occurred in uncomplicated acute myocardial infarction a significantly better prognosis could be demonstrated (odds ratio = 0.2; 95% confidence interval = 0.0-0.6). The 10 long-term survivors investigated lead an independent life and all returned to former occupation. The most common complaint was moderate memory disturbance (five patients). The conclusion is that this study confirms the critical influence of cellular anoxia on prognosis and allows the improved delineation of the situations in which cardiopulmonary resuscitation appears to be hopeless or likely to be successful. The follow up in a small number of survivors has shown a good quality of life and minor neuropsychological sequellae.