Predicting cardiac complications in patients undergoing non-cardiac surgery

Abstract
The authors prospectively studied 455 consecutive patients referred to the general medical consultation service for cardiac risk assessment prior to non-cardiac surgery, in order to validate a previously derived multifactorial index in their clinical setting. They also tested a version of the index that they had modified to reflect factors they believed to be important. For patients undergoing major surgery, the original index performed less well in the validation data set than in the original derivation set (p<0.05), but still added predictive information to a statistically significant degree (p<0.05). The modified index also added predictive information for patients undergoing both major and minor surgery, demonstrating an area under the Receiver Operating Characteristic curve of 0.75 (95% confidence interval of 0.70 to 0.80). A simple nomogram is presented which will enable conversion of pretest probabilities into posttest probabilities using the likelihood ratios associated with each risk score. It is recommended that clinicians estimate local overall complication rates (pretest probabilities) for the clinically relevant populations in their settings before they apply the predictive properties (likelihood ratios) demonstrated in this study in order to calculate cardiac risks for individual patients (posttest probabilities).