Experimental data suggest that vascular ATP-sensitive potassium (K(ATP)) channels may be an important determinant of functional hyperaemia, but the contribution of K(ATP) channels to exercise-induced hyperaemia in humans is unknown. Forearm blood flow was assessed in 39 healthy subjects (23 males/16 females; age 22+/-4 years) using the technique of venous occlusion plethysmography. Resting forearm blood flow and functional hyperaemic blood flow (FHBF) were measured before and after brachial artery infusion of the K(ATP) channel inhibitors glibenclamide (at two different doses: 15 and 100 microg/min) and gliclazide (at 300 microg/min). FHBF was induced by 2 min of non-ischaemic wrist flexion-extension exercise at 45 cycles/min. Compared with vehicle (isotonic saline), glibenclamide at either 15 microg/min or 100 microg/min did not significantly alter resting forearm blood flow or peak FHBF. The blood volume repaid at 1 and 5 min after exercise was not diminished by glibenclamide. Serum glucose was unchanged after glibenclamide, but plasma insulin rose by 36% (from 7.2+/-0.8 to 9.8+/-1.3 m-units/l; P =0.02) and 150% (from 9.1+/-1.3 to 22.9+/-3.5 m-units/l; P =0.002) after the 15 and 100 microg/min infusions respectively. Gliclazide also did not affect resting forearm blood flow, peak FHBF, or the blood volume repaid at 1 and 5 min after exercise, compared with vehicle (isotonic glucose). Gliclazide induced a 12% fall in serum glucose (P =0.009) and a 38% increase in plasma insulin (P =0.001). Thus inhibition of vascular K(ATP) channels with glibenclamide or gliclazide does not appear to affect resting forearm blood flow or FHBF in healthy humans. These findings suggest that vascular K(ATP) channels may not play an important role in regulating basal vascular tone or skeletal muscle metabolic vasodilation in the forearm of healthy human subjects.