Poor exercise capacity and abnormal heart rate (HR) recovery during exercise testing predict poor long-term outcomes. The relationship between these parameters in patients with coronary heart disease (CHD) is unknown. Whether abnormal HR recovery adds to the prognostic value of poor exercise capacity in patients with CHD is unclear. A total of 282 patients (17% women) with stable CHD underwent cardiopulmonary treadmill testing at the end of Phase II cardiac rehabilitation and were followed for a mean of 9.8 +/- 2.9 years. Cox proportional hazards regression with adjustment for low peak oxygen consumption ([peak VO(2)] <19 ml/kg/min for men, <15 ml/kg/min for women), age, and gender was used to test the prognostic significance of HR recovery (HR recovery = peak - 1-minute after peak HR). HR recovery and peak VO(2) correlated mildly (r = 0.35, p <0.001). Both an abnormal HR recovery (<13 beats/min) and a low peak VO(2) were significantly associated with greater mortality in a model including age, gender, low peak VO(2), and abnormal HR recovery (hazard ratio for abnormal HR recovery = 2.16, 95% confidence interval 1.14 to 4.09; hazard ratio for low peak VO(2) = 3.63, 95% confidence interval 2.09 to 6.32). Despite a preserved peak VO(2), the 10-year mortality rate was significantly greater in patients with a HR recovery of <13 beats/min compared with those with a HR recovery of >or=13 beats/min (13.6% vs 5.6%, respectively; p <0.05). In conclusion, in patients with stable CHD undergoing cardiac rehabilitation, the rate of HR recovery provides additional prognostic information beyond the peak VO(2). An abnormal HR recovery identifies a subset of patients at intermediate risk despite a preserved peak VO(2).