The association between age of hospitalized patients and the delivery of advanced cardiac life support

OBJECTIVE: To determine the extent of variability in the administration of advanced cardiac life support (ACLS) and to determine if age is associated with variability. DESIGN: Retrospective cohort. SETTING: Urban teaching hospital. PATIENTS: One hundred twenty-two adult inpatients without a ''do-not-resuscitate'' order who suffered cardiopulmonary arrest during 1993. MEASUREMENTS AND MAIN RESULTS: Of the total, 35 (29%) survived the arrest and 87 (71%) died. Among the nonsurvivors, two patients received no ACLS and six were not intubated, despite the inclusion of intubation in all ACLS protocols. Of the 87 nonsurvivors, 31 had a single electrocardiographic rhythm during their arrest and should have had similar ACLS trials. However, the 9 nonsurvivors with ventricular fibrillation received a range of 0 to 17 interventions, the 11 with electromechanical dissociation received 1 to 22, and the 11 with asystole received 9 to 14. Based on a protocol-derived definition of a minimal trial of ACLS (a ''short ACLS trial'') for all 87 nonsurvivors, age greater than 75 was associated with receiving a short trial. Dependent functional status and being on a medical service were also associated with a short ACLS trial. In a logistic regression model including these variables as covariates, age remained significantly associated with a short ACLS trial; odds ratio, 9.71 (95% confidence interval 1.68, 56.1). CONCLUSIONS: Wide variability exists in the administration of ACLS at the studied site. The finding that some patients receive no ACLS suggests that physicians at this site may be making bedside determinations of the likelihood of its benefit based on individual patient characteristics. The association between older age and short ACLS trials among all nonsurvivors suggests that age is most important of these characteristics.