Table 6. Multivariate Analysis of Characteristics Potentially Associated with Survival following Cardiac Arrest during Noncardiac Surgery^*

* Multiple logistic regression with stepwise backward elimination of nonsignificant variables was used to identify a set of independent predictors of survival. Only characteristics found to have some evidence (P < 0.05) of a univariate association with either immediate or hospital survival (tables 3–5) were included in the multivariate analysis. After elimination of nonsignificant variables, probable cause of arrest, time of arrest, and documented hypotension before arrest were found to be independently associated with immediate survival. When the analysis was performed with hospital survival as the dependent variable, probable cause of arrest, time of arrest, documented hypotension before arrest, and diabetes mellitus were found to be independently associated with hospital survival. The results of the final multivariate models are presented using odds ratios (OR), which were calculated on the basis of the parameter coefficients from the logistic regression analysis.

† Standard working hours were 07:00–20:00, Monday through Friday.

CI = confidence interval.