OBJECTIVE: Ghrelin inhibits sympathetic nervous system (SNS) activity in rodents. We studied the effect of ghrelin on healthy humans, in obesity and in vagotomized subjects. DESIGN: Randomized, double-blinded, placebo-controlled crossover. SUBJECTS: Seven lean [mean body mass index (BMI) 23.6 +/- 0.9 kg/m(2) ], seven morbidly obese (mean BMI 50.9 +/- 4.4 kg/m(2) ) and seven post-gastrectomy subjects (mean BMI 22.0 +/- 1.1 kg/m(2) ). MEASUREMENTS: Subjects were randomized to intravenous ghrelin (5 pmol/kg/min) or saline over 270 min. Subjects had a fixed calorie meal and a free choice buffet during the infusion. Heart rate variability (HRV) was measured. Total power (TP) represents overall autonomic function, low-frequency (LF) power represents sympathetic and parasympathetic activity, and high-frequency (HF) power represents parasympathetic activity. Very low (VLO) frequency represents the frequency band associated with thermogenesis. RESULTS: Preliminary anova analysis, looking at all three subject groups together, showed that ghrelin had an overall highly significant inhibitory effect on TP (P = 0.001), HF power (P = 0.04), VLO power (P = 0.03) and no effect on LF (P = 0.07). Further subset analysis revealed that ghrelin had a significant effect on TP (P = 0.03), borderline effect on LF power (P = 0.06) and no effect on HF power (P = 0.1) in healthy controls. By contrast in obese subjects, ghrelin had no effect on TP (P = 0.3), LF (P = 0.5) and HF (P = 0.06) and also no effect in the vagotomized subjects on TP (P = 0.7), LF (P = 0.7) and HF (P = 0.9). Ghrelin had no effect on the LF/HF ratio. CONCLUSIONS: Ghrelin inhibits SNS activity in healthy controls with a moderate effect on parasympathetic nervous system activity but had no effect on obese subjects. Vagotomized subjects also did not respond to ghrelin, suggesting the vagus nerve is important for the effects of peripheral ghrelin on the SNS.