Both the original (Goldman and colleagues') and modified (Detsky and colleagues') cardiac risk indexes have been shown to be useful for detecting patients at high risk for postoperative cardiac events. Dr. Schiff's point that they are useful teaching tools is correct: They provide a guide to the clinical variables that place the patient at high risk. Because these indexes were validated more than 10 years ago, the specific numerical risk attached to a high score may be less applicable in today's practice environment. Both the original and modified cardiac risk indexes would be suitable as a first screening assessment. Our reasons for recommending the modified cardiac risk index instead of the original are provided in the guidelines.

Neither of the cardiac risk indexes have shown discrimination between low- and intermediate-risk patients [1, 2]. Because most events occur in intermediate-risk patients (the largest group), the cardiac risk indexes will miss these. These indexes need to be supplemented with a more sensitive index, such as that by Eagle and colleagues or Poldermans and colleagues to detect patients at very low risk for postoperative events.

We also agree with Dr. Cohn that patients undergoing nonvascular surgery are not a homogeneous group. Unfortunately, the literature has tried and failed to demonstrate that any noninvasive tests can further stratify risk in this group. Accordingly, despite widespread use of functional capacity in preoperative assessment, the only available perioperative literature shows only weak predictive ability of this characteristic.

A brief description of the College's guideline development process may help alleviate some of Dr. Cohn's concerns about the validity and potential bias of this guideline. In our case, the College document was drafted by two authors and was reviewed as many times as necessary by the Clinical Efficacy Assessment Subcommittee. It was then sent for two rounds of external review (with incorporation of concerns and comments) and final Subcommittee review before approval by the College's Education Committee and Board of Regents. In our situation, this meant nine iterations. Because of this exhaustive process, the guidelines themselves are published as College policy, not as individual authors' opinions. In cases in which the guideline approval process could not be completed (as in recommendation of ß-blocker use), the authors specifically stated so in an authors' addendum.

Valerie A. Palda, MD

St. Michael's Hospital

Toronto, Ontario M5B 1W8, Canada

Allan S. Detsky, MD

Mount Sinai Hospital

Toronto, Ontario M5G 1X5, Canada