Accurate interpretation of cardiopulmonary exercise testing (CPET) relies on age, gender, and exercise modality-specific reference values. To date, clinically applicable CPET reference values derived from a source population of endurance athletes (EAs) have been lacking. The purpose of this study was to generate CPET reference values for use in the clinical assessment of EA.
Prospective data accrued during the clinical care of healthy EA were used to derive CPET reference values and to develop novel equations for V˙O2peak. The performance of these equations was compared to the contemporary standard of care equations and assessed in a discrete EA validation cohort. A total of 272 EA (age = 42 ± 15 years, female = 31%, V˙O2peak = 3.6 ± 0.83 L/min) met inclusion criteria and comprised the derivation cohort. V˙O2peak prediction equations derived from general population cohorts described a modest amount of V˙O2peak variability [R2 = 0.58-0.70, root mean square error (RMSE) = 0.46-0.54 L/min] but were mis-calibrated (calibration-in-the-large = 0.45-1.18 L/min) among EA leading to significant V˙O2peak underestimation. Newly derived, externally validated V˙O2peak prediction equations for EA that included age, sex, and height for both treadmill (R2 = 0.74, RMSE = 0.42 L/min) and cycle ergometer CPET (Cycle: R2 = 0.69, RMSE = 0.42 L/min) demonstrated improved accuracy.
Commonly used V˙O2peak prediction equations derived from general population cohorts perform poorly among competitive EA. Newly derived CPET reference values including novel V˙O2peak prediction equations may improve the clinical utility of CPET in this rapidly growing patient population.