The guidelines for the assessment and management of perioperative risk from coronary artery disease in patients undergoing major noncardiac surgery include a comprehensive review and management recommendations [1, 2]. In their management strategy (Appendix Figure 1), the authors rely heavily on the modified cardiac risk index developed by Detsky and colleagues [3]. I have several concerns about the bedside use of this modified index. One issue is whether the patients in Detsky and coworkers' 1986 study [3] are representative of all patients undergoing noncardiac surgery. I do not think they are representative because the authors "enrolled only those patients seen by our consultation service for whom a question of cardiac risk arose during the work-up of the referring surgical service or during our consultation" [3]. Thus, the study patients were referred patients, not consecutive unselected patients.

Detsky and colleagues' study recognized differences in the performance of clinical prediction rules in different settings and institutions [3]. The authors of that study recommend that "institutions record their experience with the indexes in order to calculate local likelihood ratios ... ." It would be interesting but cumbersome to calculate and periodically update these ratios.

A clinical prediction rule should be validated by prospective testing in another group of patients, preferably with different clinicians [4]. I do not believe that the modified cardiac risk index has been prospectively validated. The modified index was most useful for predicting risk in high-risk patients (class III; likelihood ratio, 10.6) [2, 3]. However, many of the postoperative cardiac complications occurred in patients who were in class I or II [3].

The modified cardiac risk index and Goldman and colleagues' original cardiac risk index [5] are helpful as teaching tools for attending physicians and residents and provide a framework for the preoperative evaluation of cardiac risk. I have cited some of the limitations of the modified cardiac risk index as a bedside tool for evaluating perioperative cardiac risk.