PURPOSE: The importance of cardiorespiratory fitness versus adiposity in determining heart rate variability (HRV) is unclear. METHODS: From the Coronary Artery Risk Development in Young Adults study, an observational cohort study, we included 2316 participants (mean age = 45.2 +/- 3.6 yr at year 20, 57% female, 43% Black) with HRV measured in 2005-2006 (year 20) and graded exercise test duration (GXTd) and adiposity measures (body mass index and waist circumference) obtained in 1985-1986 (baseline) and 2005-2006. HRV measures (SD of all normal RR intervals [SDNN] and square root of the mean value of the squares of differences between all successive RR intervals [RMSSD]) were obtained from resting 30-s 12-lead ECG. Cross-sectional associations between GXTd, adiposity, and HRV were assessed at year 20. Longitudinal changes in GXTd and adiposity measures were categorized as >/=10% increase, <10% change (no change), or >/=10% decrease. We used multivariable logistic regression to assess associations of GXTd and adiposity measures with unfavorable versus more favorable HRV (lower 25th percentile vs upper 75th percentile). RESULTS: A 1-SD increment in GXTd was associated with 22% and 32% lower odds of unfavorable SDNN and RMSSD, respectively; associations remained significant after adjustment for adiposity. A 1-SD increment in adiposity measures was associated with 16%-28% higher odds of unfavorable RMSSD; associations were not significant after adjustment for GXTd. Compared with no change/increase in GXTd, longitudinal decrease in GXTd was significantly associated with 55% and 94% higher odds of unfavorable SDNN and RMSSD, respectively, at year 20. These associations remained significant after adjusting for adiposity. CONCLUSION: Cardiorespiratory fitness may be a stronger determinant of HRV than adiposity. Intervention studies are needed to better determine the differential effects of improved cardiorespiratory fitness versus weight loss on HRV.