EVALUATION OF THE RELATIONSHIP BETWEEN HEART-RATE AND VENTILATION FOREPIDEMIOLOGIC STUDIES

Estimation of pulmonary exposure and dose in air pollution epidemiolog y has been impaired by the lack of methods for directly measuring vent ilation in ambulatory subjects. Heart-rate monitoring offers an approa ch to estimate ventilation by using ventilation-on-heart-rate (V(E)-HR ) regressions established during exercise testing to estimate ventilat ion in the field. Conventional methods and protocols for testing were used to evaluate the relationship between V(E) and HR during three tas ks: (1) exercising on a cycle ergometer, (2) lifting, and (3) vacuumin g. The relationship between V(E) and HR was curvilinear and was best f it with linear regression models, using a natural log transformation o f V(E). Considerable interindividual variability in slopes and interce pts was observed across all types of exercise tests. The variability a bout 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 esta blished during upper-body exercise were greater by approximately 30%, relative to those derived in lower-body exercise (paired t test, p < . 001). However, V(E)-HR regression slopes derived from tests in which p rogressively increasing workloads were used were comparable to those o btained during variable and nonprogressive protocols. These findings i ndicate that predictive accuracy is maximized by deriving V(E)-HR regr essions for individual subjects and for both lower- and upper-body act ivities.