Data on 470 adults with single in-hospital cardiac arrest resuscitations were analyzed to determine 24-h and discharge survival rates and to identify significant correlates of survival. One hundred fifty-three (33%) patients were alive 24 h after initiation of cardiopulmonary resuscitation; 69 (45% of 24-h survivors, 15% of all patients) were discharged alive. Logit analysis identified the following independently significant correlates of 24-h survival: arrest locations other than emergency room or cardiac care unit, CPR duration less than 15 min, non-cardiac primary diagnosis, non-asystolic dysrhythmia, less than one intravenous and one drip-administered inotrope and absence of pacemaker insertion and defibrillation. Fifty-one (94%) of 54 patients with all of these characteristics were alive 24 h after initiation of CPR. The same variables, as well as age less than 68 years and absence of intubation were statistically associated with discharge survival. Nine (64%) of 14 patients with all of these characteristics were discharged alive. Increased intervention was generally associated with increased mortality. Overall survival rates replicate previous reports and may reflect the effects of diagnosis-related groups policies on the average illness severity of the in-patient population, rather than failure of current CPR methods to improve the probability of survival. Use of the data as baseline for future studies and as a source of hypotheses for research on decision making are discussed.