How accurately are maximal metabolic equivalents estimated based on the treadmill workload in healthy people and asymptomatic subjects with cardiovascular risk factors?

Maximal exercise capacity expressed as metabolic equivalents (METs) is rarely directly measured (measured METs; mMETs) but estimated from maximal workload (estimated METs; eMETs). We assessed the accuracy of predicting mMETs by eMETs in asymptomatic subjects. Thirty-four healthy volunteers without cardiovascular risk factors (controls) and 90 patients with at least one risk factor underwent cardiopulmonary exercise testing using individualized treadmill ramp protocols. The equation of the American College of Sports Medicine (ACSM) was employed to calculate eMETs. Despite a close correlation between eMETs and mMETs (patients: r = 0.82, controls: r = 0.88; p > 0.001 for both), eMETs were higher than mMETs in both patients [11.7 (8.9 - 13.4) vs. 8.2 (7.0 - 10.6) METs; p > 0.001] and controls [17.0 (16.2 - 18.2) vs. 15.6 (14.2 - 17.0) METs; p > 0.001]. The absolute [2.5 (1.6 - 3.7) vs. 1.3 (0.9 - 2.1) METs; p > 0.001] and the relative [28 (19 - 47) vs. 9 (6 - 14) %; p > 0.001] difference between eMETs and mMETs was higher in patients. In patients, ratio limits of agreement of 1.33 (*/ divided by 1.40) between eMETs and mMETs were obtained, whereas the ratio limits of agreement were 1.09 (*/ divided by 1.13) in controls. The ACSM equation is associated with a significant overestimation of mMETs in young and fit subjects, which is markedly more pronounced in older and less fit subjects with cardiovascular risk factors.