Outcome of cardiopulmonary resuscitation and predictors of resuscitation status in an urban community teaching hospital.

Medical records of all expired patients as well as all patients designated on billing logs as having received cardiopulmonary resuscitation (CPR) during a 6-month period were reviewed. Patients were considered to have been 'coded' if they were found unresponsive and if the advanced cardiac life support (ACLS) protocol of the American Heart Association (AHA) was subsequently initiated. Of 105 patients who received CPR, 98 died during their hospital stay. Of the seven remaining patients, four had undergone coronary by-pass graft surgery, one was discharged in a persistent vegetative state, one died during an admission 2 months later, and one patient was transferred to another institution where he died. Various factors were studied in an effort to determine how patients on whom resuscitation was attempted differed from those who died without ever having received CPR. Patients who underwent CPR at least once during their hospitalization were more likely to have had cardiac diagnoses on admission (P < 0.001), to have been postoperative (P = 0.02), to have been admitted to a monitored bed on admission (P < 0.001) to have received more days of intensive care (P < 0.001) and to have received more specialist consultations (P = 0.004). Patients not receiving CPR were more likely to have had a primary diagnosis of neoplastic disease (P < 0.001), stroke or intracranial hemorrhage (P = 0.02) or dementia (P < 0.001). Age, race, or gender did not differ significantly between the two groups.