Introduction:
Mortality risk predictions are improved with routine assessment of cardiorespiratory fitness (CRF). Accordingly, an American Heart Association Scientific Statement suggests routine clinical assessment of CRF in apparently healthy adults minimally using non-exercise prediction equations, which can be calculated from common health metrics. However, no study has assessed the ability of non-exercise CRF prediction equations to accurately detect longitudinal changes.
Hypothesis:
Changes in estimated CRF (eCRF) would be related to directly-measured changes, yet appreciable misclassification would occur at the individual level.
Methods:
The sample included 987 apparently healthy adults (324 females; mean±SD age 43.1±10.4 years) who completed 2 cardiopulmonary exercise tests (CPX) at least 3 months apart (3.2±5.4 years follow-up). The change in eCRF from 27 distinct non-exercise prediction equations was compared to the change in directly-measured CRF determined from CPX. A change of ≥5% was used to classify participants as having a directional increase or decrease in CRF. Analysis included Pearson product moment correlations, standard error of estimate (SEE) values, the Benjamini-Hochberg procedure to compare eCRF with directly-measured CRF, and chi-squared tests to examine the impact of follow-up time on the percentage of participants correctly identified as having a directional increase or decrease in CRF.
Results:
The change in eCRF from each equation was correlated to the change in directly-measured CRF (
P
<0.001) with R
2
values ranging from 0.06-0.43 and SEE values ranging from 0.9-5.9 ml·kg
-1
·min
-1
. For 16 of the 27 equations, the change in eCRF was significantly different from the change in directly-measured CRF. When classifying directional changes, the prediction equations correctly categorized an average of 54% of individuals as having increased, decreased, or no change in CRF. When examining the influence of follow-up time, the average percentage of individuals correctly classified as having a directional increase in CRF was greater when the time between tests was ≤8months (54%) compared to ≥2years (28%). In contrast, the average percentage correctly classified as having a directional decrease in CRF was lower with tests ≤8months apart (8%) compared to ≥2years (73%).
Conclusions:
As hypothesized, discernible variability was found in the accuracy between non-exercise prediction equations and the ability of equations to accurately assess changes in directly-measured CRF over time. Considering the appreciable error that prediction equations had with detecting even directional changes in CRF, these results suggest eCRF may have limited clinical utility.
Fasting glucose levels, incident diabetes, subclinical atherosclerosis and cardiovascular events in apparently healthy adults: A 12-year longitudinal study