Cardiopulmonary exercise testing for high-risk South African surgical patients.
Aim: To determine the prognostic value of cardiopulmonary exercise testing (CPET) for major vascular surgery in South African patients. Methods: CPET has been used in Durban since October 2004 to predict cardiac risk for high-risk patients undergoing major vascular surgery. A submaximal 'anaerobic threshold' (AT) test was conducted on all high-risk patients. Patients were classified into two groups: 'low AT' where the oxygen consumption at the AT was <1 lml.kg^.min"1 for cycling or < 9ml.kg"1.mkf1 for arm cranking and 'high AT' when the patient surpassed these targets. Analysis of all in-hospital deaths following surgery was conducted by two independent assessors blinded to the CPET test result. Deaths classified as primarily 'cardiac in origin' have been used in this retrospective cohort analysis. Results: The AT measured during CPET was not a statistically significant pre-operative prognostic marker of cardiac mortality. However, the survivors of the patients with a 'low AT' may be identified by their response to increasing metabolic demand between 5 and 7 ml.kg^.min"1. Survivors were more dependent on increasing heart rate, while non-survivors were more dependent on oxygen extraction. When this information is added to the AT, CPET was the only test statistically associated with cardiac mortality, in comparison to Lee's Revised Cardiac Risk Index and the resting left ventricular ejection fraction which were not statistically associated with cardiac death. A hundred percent of patients with a positive test died of cardiac causes, while 11% of the patients with a negative test had cardiac deaths. The risk ratio associated with cardiac death following a positive test was 8.00 [95% CI 3.8-16.9]. The sensitivity was 0.25 [95% CI 0.04-0.64], the specificity was 1.00 [95% CI 0.90-1.00], the positive predictive value was 1.00 [95% CI 0.20-0.95] and the negative predictive value was 0.88 [95% CI 0.74-0.95]. Conclusions: CPET provides valuable prognostic information in our surgical population.