OBJECTIVE: To assess heart-rate variability (HRV) measures of interictal electrocardiography (ECG) for drug-resistant epilepsy and to relate the findings to the outcome of vagus nerve stimulation (VNS) treatment. METHODS: Time-domain, frequency-domain, and nonlinear analyses were used to analyze preoperative HRV measures in 32 patients with drug-resistant epilepsy who had received VNS implants at the same hospital and 32 healthy age- and sex-matched control subjects. HRV measurements based on ambulatory 24 h ECG recordings were analyzed to identify seizure reduction 1 year after VNS treatment. Responders were defined as having at least 50% seizure reduction 1 year after treatment. RESULTS: Patients with drug-resistant epilepsy had significantly lower time domain (SDNN, RMSSD, pNN50), frequency domain (VLF, LF, HF, TP), and nonlinear (SD1, SD2) HRV measurements than matched healthy controls. None of the analyzed HRV measures of the responders differed significantly from their controls, whereas those of the nonresponders had significantly lower RMSSD, pNN50, HF, and SD1 than the responders. SIGNIFICANCE: The preoperative HRV indices demonstrate that nonresponders have more pronounced impairment of their cardiac autonomic function than the responders. Presurgical HRV measurements representing parasympathetic cardiac control or vagal tone were significantly associated with the responsiveness to VNS. Thus the measurements show promise for predicting the reduction of seizure frequency after VNS treatment.