BACKGROUND: We recently reported that subcutaneous nerve activity (SCNA) can be used to estimate sympathetic tone. OBJECTIVE: The purpose of this study was to test the hypothesis that left thoracic SCNA is more accurate than heart rate variability (HRV) in estimating cardiac sympathetic tone in ambulatory dogs with myocardial infarction (MI). METHODS: We used an implanted radiotransmitter to study left stellate ganglion nerve activity (SGNA), vagal nerve activity (VNA), and thoracic SCNA in 9 dogs at baseline and up to 8 weeks after MI. HRV was determined based on time-domain, frequency-domain, and nonlinear analyses. RESULTS: The correlation coefficients between integrated SGNA and SCNA averaged 0.74 (95% confidence interval [CI] 0.41-1.06) at baseline and 0.82 (95% CI, 0.63-1.01) after MI (P <.05 for both). The absolute values of the correlation coefficients were significantly larger than that between SGNA and HRV analysis based on time-domain, frequency-domain, and nonlinear analyses, respectively, at baseline (P <.05 for all) and after MI (P <.05 for all). There was a clear increment of SGNA and SCNA at 2, 4, 6, and 8 weeks after MI, whereas HRV parameters showed no significant changes. Significant circadian variations were noted in SCNA, SGNA, and all HRV parameters at baseline and after MI, respectively. Atrial tachycardia (AT) episodes were invariably preceded by SCNA and SGNA, which were progressively increased from 120th, 90th, 60th, to 30th seconds before AT onset. No such changes of HRV parameters were observed before AT onset. CONCLUSION: SCNA is more accurate than HRV in estimating cardiac sympathetic tone in ambulatory dogs with MI.