Evaluation of the relationship between heart rate and ventilation for epidemiologic studies

Estimation of pulmonary exposure and dose in air pollution epidemiology has been impaired by the lack of methods for directly measuring ventilation in ambulatory subjects. Heart-rate monitoring offers an approach to estimate ventilation by using ventilation-on-heart-rate (VE-HR) regressions established during exercise testing to estimate ventilation in the field. Conventional methods and protocols for testing were used to evaluate the relationship between VE and HR during three tasks: (1) exercising on a cycle ergometer, (2) lifting, and (3) vacuuming. The relationship between VE and HR was curvilinear and was best fit with linear regression models, using a natural log transformation of VE. Considerable interindividual variability in slopes and intercepts was observed across all types of exercise tests. The variability about the fitted regression lines for individual subjects was minimal; for example, individual R2 values for the maximum exercise test on 15 men ranged from 0.90 to 0.99 (mean = 0.97). The regression slopes established during upper-body exercise were greater by approximately 30%, relative to those derived in lower-body exercise (paired t test, p < .001). However, VE-HR regression slopes derived from tests in which progressively increasing workloads were used were comparable to those obtained during variable and nonprogressive protocols. These findings indicate that predictive accuracy is maximized by deriving VE-HR regressions for individual subjects and for both lower- and upper-body activities.