Prediction of survival from resuscitation: a prognostic index derived from multivariate logistic model analysis

Abstract

Despite advances in resuscitation, the ability to predict survival at cardiac arrests remains unsophisticated. We identified the factors determining outcome of all cardiopulmonary resuscitations performed at our institution over a 4-year period, and used a Cox multivariate regression model to design prognostic indices to assess the probability of successful resuscitation and hospital discharge. Cardiac arrests (710) were studied, and 193 (28%) were successfully resuscitated. The most influential variables, judged by the size and significance of their logistic regression coefficients, were rhythm, resuscitation delay, and age (for successful resuscitation), and rhythm, performance of intubation and defibrillation, defibrillation delay, and age (for survival until discharge). The combination of these in a prognostic index reliably predicted both outcome (area under the receiver operating curve of 0.78), and survival until discharge (area under the curve of 0.80).