Evaluation of an Algorithm to Detect the First Ventilatory Threshold from Heart Rate

Abstract

PURPOSE: Evaluate the accuracy of an algorithm we previously developed to detect VT1 using heart rate and physical characteristics. METHODS: Fifty-three subjects (30M, 23F; 1.73±0.10m; 74.6±13.3kg) of varying age and athletic ability provided consent to complete two days of laboratory testing: an incremental treadmill test on the first day and an incremental over-ground shuttle run on the second day. Ventilatory gases and heart rate were recorded during the treadmill test and only heart rate was recorded during the shuttle run. Heart rate at VT1 (HRVT1) was detected using the ventilatory measurements and compared to HRVT1 estimated by the algorithm. The algorithm detected HRVT1 using a three-step process: 1) Estimate 50-80% of maximum estimated heart rate, within which HRVT1 could occur. 2) Using a 30-second heart rate moving average, detect a heart rate threshold (HRT) at the point of maximum slope within the target heart rate zone. 3) Estimate HRVT1 from the previously developed regression equation with HRT as input. RESULTS: HRVT1 as predicted by the algorithm and ventilatory measures were strongly correlated for the treadmill test (r=0.70) but not for the shuttle run (r=0.42). When grouping subjects by age, sex, and athletic ability, the algorithm most accurately predicted HRVT1 during the treadmill test for younger subjects (20-29y, r=0.77 vs. 30-69y, r=0.35), but performed similarly for both sexes as well as for athletes vs. sedentary individuals. CONCLUSIONS: The algorithm was able to estimate HRVT1 well during an incremental treadmill test, but was limited during the incremental shuttle run; this may be due to the shuttle run being conducted on a different day and during task (i.e. change-of-direction vs. treadmill running).