Can a clinical decision aid be useful in determining when to discontinue in-hospital cardiac resuscitation?

Journal of Family Practice, June, 2001 by Samantha Sattler, Erik J. Lindbloom

Van Walraven C, Forster AJ, Parish DC, et al. Validation of a clinical decision aid to discontinue in-hospital cardiac arrest resuscitations. JAMA 2001; 285:1602-06.

* BACKGROUND The authors previously derived a simple clinical decision aid (CDA) to be used at the bedside for determining the prognosis of in-patients after resuscitation. The CDA was derived from data obtained from a population of 1077 inpatients who underwent resuscitation at 5 teaching hospitals. The authors observed that all resuscitated patients discharged from the hospital had a witnessed arrest, an initial cardiac rhythm of ventricular tachycardia or ventricular fibrillation, or a pulse within 10 minutes of chest compressions. They proposed that resuscitative efforts may be safely withdrawn if the patient does not meet at least 1 criterion. They noted that 100% of patients not meeting the CDA criteria died in the hospital. However, the 95% confidence interval for this figure ranged as low as 97.1%. The purpose of this study was to apply the CDA using data from a second set of patients to determine its usefulness.

* POPULATION STUDIED This validation study used a registry of data from 3960 attempted resuscitations at a 550-bed community teaching hospital. The investigators included 2181 arrests of inpatients 16 years and older who were pulseless at the start of resuscitation. They excluded patients who arrested in the operating room or neonatal intensive care unit; those who had an initial rhythm other than ventricular tachycardia, ventricular fibrillation, pulseless electrical activity, or asystole; those whose time to initial chest compressions exceeded 15 minutes; and those who received no chest compressions at all. Nearly 300 patients were also excluded because information needed to apply the CDA was missing.

* STUDY DESIGN AND VALIDITY The research team reviewed code sheets and records with hospital charges including cardioversion, defibrillation, or administration of epinephrine. Witnessed arrests were classified as those observed directly or on cardiac monitoring. They compared actual discharge status to that predicted by the CDA. Only 460 (21%) of the 2181 arrests were considered unwitnessed. No further details are given about the 7% who did not have adequate information to include in the study.

* OUTCOMES MEASURED The primary outcome was survival to discharge.

* RESULTS Fifteen percent (N = 327) of the 2181 attempted resuscitations resulted in eventual discharge from the hospital. Of those 327 events, all but 3 were correctly predicted by the CDA (sensitivity = 99.1%). Those 3 patients were among the 269 patients predicted by the CDA to have no chance of discharge (negative predictive value = 98.9%). Because of the generally low success rate of attempted resuscitations, the specificity and positive predictive value of the CDA were low (14.4% and 17.0%, respectively). The negative likelihood ratio of this CDA was 0.064; this would lower the probability of surviving to discharge from 15% to 1.1% if the patient did not satisfy the CDA criteria. Of the 3 surviving patients predicted to have no chance of discharge, 2 were transferred to a nursing home, and 1 died 2 months after discharge.

RECOMMENDATION FOR CLINICAL PRACTICE

A CDA is useful in identifying those patients who have some chance at a successful resuscitation (sensitivity = 99.1%). It assigns a pulseless patient some chance of survival to discharge if any of the following criteria are met: the arrest is witnessed; the initial rhythm is ventricular tachycardia or ventricular fibrillation; or if the pulse is regained within the first 10 minutes of resuscitation. Absence of all these criteria creates a dismal prognosis and supports the decision to discontinue in-hospital cardiac resuscitation. Note that only a minority of inpatients experience unwitnessed cardiac arrests (ie, those patients who are not in the intensive care unit or on cardiac monitoring). This decision aid does not advise when to discontinue resuscitative efforts after a witnessed arrest.

Samantha Sattler, MD Erik J. Lindbloom, MD, MSPH University of Missouri-Columbia E-mail: lindbloome@health.missouri.edu