Longitudinal changes in cardiorespiratory fitness: measurement error or true change?

Abstract

This study examined the thesis that the reported Aerobics Center Longitudinal Study (ACLS) mortality reductions associated with improved cardiorespiratory fitness were because of measurement error of serial treadmill tests. We tested the research hypothesis that longitudinal changes in cardiorespiratory fitness of the ACLS cohort were a multivariate function of changes in self-report physical activity (SR-PA), resting heart rate, and body mass index (BMI). We used the results of three serial maximal treadmill tests (T1, T2, and T3) to evaluate the serial changes in cardiorespiratory fitness of 4675 men. The mean duration between the three serial tests examined was: T2 - T1, 1.9 yr; T3 - T2, 6.1 yr; and T3 - T1, 8.0 yr. Maximum and resting heart rate, BMI, SR-PA, and maximum Balke treadmill duration were measured on each occasion. General linear models analysis showed that with change in maximum heart rate statistically controlled change in treadmill time performance was a function of independent changes in SR-PA, BMI, and R-HR. These variables accounted for significant (P < 0.001) proportions (7%, 9%, and 12%) of the change in treadmill time variance. Those men who increased their SR-PA and lowered their BMI and resting heart rate gained the most fitness between serial tests. These results support the research hypothesis tested. Variations in serial ACLS treadmill tests are not just due to measurement error alone, but also to systematic variation linked with changes in lifestyle.