Accuracy of prediction equations to estimate submaximal VO2 during cycle ergometry: the HERITAGE Family Study

It was hypothesized that more accurate equations for estimating submaximal (V) over dot O-2 during cycle ergometry could be developed if more independ ent variables were used in the equation. Purpose: The purposes of this stud y were: (1) to develop new equations for estimating submaximal (V) over dot O-2 during cycle ergometry; and (2) to examine the accuracy of the newly d eveloped equations and those of the American College of Sports Medicine (19 95), Berry et al. (1993), Lang et al. (1992), Latin and Berg (1994), and Lo nderee et al. (1997). Methods: Subjects (715 men and women, ages 16-65 yr, from the HERITAGE Family Study) completed a maximal cycle ergometry test, t wo submaximal trials at 50 W and 60% of (V) over dot O-2max, hydrostatic we ighing, and stature and body mass measures before and after 20 wk of cycle ergometry training. Regression analysis generated prediction equations usin g pretraining data from the 60% trials. Results: No equation with more inde pendent variables was better than an equation that used only power output. This equation, HERITAGE-1, with only power output was cross-validated using the "jackknife" technique. Paired t-tests, mean differences, SEEs, and Es were used to compare the (V) over dot O-2, estimated by HERITAGE-I and thos e of previously published equations with the measured (V) over dot O-2 at 6 0% of (V) over dot O-2max. Conclusions: HERITAGE-I was slightly better than the equations of ACSM, Lang et al., and Latin and Berg using pretraining d ata but was not better when using post-training data. All four of these equ ations were superior to the equations of Berry et al. and Londeree et al.