Accuracy of VO2max prediction equations in older adults

Purpose: We explored the accuracy and bias of prediction equations (ACSM and Foster) in older, deconditioned men and women. We also examined the predictors of VO2max to further understand which variables affect respiratory fitness in the elderly. Methods: One hundred seventy-one community dwelling, men (72.6 +/- 4.8 yr) and women (71.0 +/- 5.1 yr) screened in a clinical trial were retrospectively examined. VO2max was measured using a standardized protocol with gas exchange measured. Measured VO2max values were compared with prediction equations via mean difference analyses, and bias was explored using Bland-Altman analyses. Regression analysis determined significant predictors of measured VO2max. Alpha was P less than or equal to 0.05. Results: In men and women, Foster, 21.6 +/- 5.7 and 17.9 +/- 5.1 mL(.)kg(-1.)min(-1), respectively, was not significantly different from measured VO2max, 21.7 +/- 4.8 and 17.3 +/- 4.0, respectively. ACSM overestimated VO2max in men and women, 26.3 +/- 8.2 and 20.9 +/- 7.3, respectively. By using Bland-Altman plots, ACSM showed significant overestimation bias in more fit women (r = 0.29), whereas Foster showed no estimation bias in either gender. Significant predictors of VO2max were gender, BMI, age, treadmill grade, and speed, with an equation R-2 of 0.70. A measure of current activity levels did not make it into the final model (P = 0.0505) but is worthy of future consideration using more sensitive measures than ours. Conclusion: ACSM is not appropriate for use when treadmill testing older adults. We believe the Foster equation's VO2max prediction accuracy is acceptable, showing no bias along a continuum of aerobic capacity.