AIMS: Poor prognosis associated with blunted post-exercise heart-rate recovery may reflect autonomic dysfunction. This study sought the accuracy of post-exercise heart-rate recovery in the diagnosis of cardiac autonomic neuropathy, which represents a serious, but often unrecognized complication of Type 2 diabetes. METHODS: Clinical assessment of cardiac autonomic neuropathy and maximal treadmill exercise testing for heart-rate recovery were performed in 135 patients with Type 2 diabetes and negative exercise echocardiograms. Cardiac autonomic neuropathy was defined by abnormalities in ≥ 2 of 7 autonomic function markers, including four cardiac reflex tests and three indices of short-term (5-min) heart-rate variability. Heart-rate recovery was defined at 1-, 2- and 3-min post-exercise. RESULTS: Patients with cardiac autonomic neuropathy (n = 27; 20%) had lower heart-rate recovery at 1-, 2- and 3-min post-exercise (P < 0.01). Heart-rate recovery demonstrated univariate associations with autonomic function markers (r-values 0.20-0.46, P < 0.05). Area under the receiver-operating characteristic curve revealed good diagnostic performance of all heart-rate recovery parameters (range 0.80-0.83, P < 0.001). Optimal cut-offs for heart-rate recovery at 1-, 2- and 3-min post-exercise were ≤ 28 beats/min (sensitivity 93%, specificity 69%), ≤ 50 beats/min (sensitivity 96%, specificity 63%) and ≤ 52 beats/min (sensitivity 70%, specificity 84%), respectively. These criteria predicted cardiac autonomic neuropathy independently of relevant clinical and exercise test information (adjusted odds ratios 7-28, P < 0.05). CONCLUSIONS: Post-exercise heart-rate recovery provides an accurate diagnostic test for cardiac autonomic neuropathy in Type 2 diabetes. The high sensitivity and modest specificity suggests heart-rate recovery may be useful to screen for patients requiring clinical autonomic evaluation.