Editor,The preoperative assessment of the likelihood of a postoperative cardiac event is complex. 1 The Revised Cardiac Risk Index (RCRI) is a commonly used scoring system for the stratification of cardiac risk of patients undergoing major non-cardiac surgery. 2 The RCRI scores patients according to six clinical categories: high-risk surgery (thoracic, abdominal and supra-inguinal vascular surgery); history of ischaemic heart disease (IHD); history of congestive heart failure; cerebrovascular disease; insulin-dependent diabetes; and renal failure. Since the publication of the original article in 1999, the RCRI has become a widely used stratification tool for cardiac risk. A recent meta-analysis has examined the predictive value of the RCRI in these studies. 3 It concluded that the RCRI performed moderately well at discriminating between low and high perioperative risk. However, the authors felt that it performed poorly in vascular cohorts, and that the studies included were of variable quality. This report highlights the need for further studies evaluating the RCRI.We report the results of a single centre observational study of consecutive patients undergoing major vascular surgery (aortic surgery, infra-inguinal bypass surgery, amputation). Ethical approval for this study [REC reference number 04/193(1)] was provided by the West of Scotland Research Ethics Committee (Chairperson Dr A Binning) on 2 March 2004. All patients gave written consent to the study. The patients were preoperatively scored according to the RCRI criteria. 2 The presence of other risk factors was obtained from the patient's history and medical records. Operations were recorded as elective or expedited (within 3 weeks of initial presentation). Patients operated as an emergency, within 24 h of an unplanned admission, were excluded. Postoperative screening for cardiac events was performed by daily clinical assessment, serial ECGs and troponin I measurement. The primary outcomes were major adverse cardiac event (MACE) (non-fatal myocardial infarction and cardiac mortality) and all-cause mortality. Patients were followed-up for 6 weeks following surgery. Statistical analysis was performed using SPSS (Version 15) statistical software package (SPSS, Chicago, IL). Categorical variables are presented with totals and percentages, and analysed with Chi-square, Fisher's exact test or Mann-Whitney test as appropriate. Receiver operating characteristic (ROC) curves were plotted to model the efficacy of the RCRI. The area under the curve (AUC) was calculated.A total of 252 patients were included in the study: aortic surgery 25.8%; bypass surgery 39.7%; and lower limb amputation 34.5%. Thirty-nine patients (15.5%) had a postoperative MACE, and 20 patients (7.9%) died within the postoperative 6-week period. Applying the RCRI to the cohort as a whole, the rate of MACE for RCRI class 1, 2, 3 and 4 was 13.4, 14.9, 18.6 and 16.7%, respectively (P ¼ 0.858) (Table 1). Similarly, the rate of perioperative mortality for these groups was 7.3, 9.2, 6.8 and 8.3%, r...