Insulin hyperpolarizes plasma membranes; we tested whether insulin affects ventricular repolarization. In 35 healthy volunteers, we measured the Q-T interval during electrocardiographic monitoring in the resting state and in response to hyperinsulinemia (euglycemic 1-mU · min¹ · kg¹ insulin clamp). A computerized algorithm was used to identify T waves; Bazett's formula was employed to correct Q-T (QTc) by heart rate (HR). In the resting state, QTc was inversely related to indexes of body size (e.g., body surface area, r = 0.53, P = 0.001) but not to indexes of body fatness. During the clamp, HR (67 ± 1 to 71 ± 1 beats/min, P < 0.0001) and plasma norepinephrine levels (161 ± 12 to 184 ± 10 pg/ml, P < 0.001) increased. QTc rose promptly and consistently, averaging 428 ± 6 ms between 30 and 100 min (P = 0.014 vs. the resting value of 420 ± 5 ms). Fasting serum potassium (3.76 ± 0.03 mM) declined to 3.44 ± 0.03 mM during insulin. After adjustment for body size, resting QTc was directly related to fasting plasma insulin (partial r = 0.43, P = 0.01); furthermore, QTc was inversely related to serum potassium levels both in the fasting state (partial r = 0.16, P < 0.04) and during insulin stimulation (partial r = 0.47, P = 0.003). Neither resting nor clamp-induced QTc was related to insulin sensitivity. Physiological hyperinsulinemia acutely prolongs ventricular repolarization independent of insulin sensitivity. Both insulin-induced hypokalemia and adrenergic activation contribute to this effect.