The insulin-antagonistic effect of growth hormone was characterized by infusing the hormone at three different infusion rates (6, 12 or 24 mU.kg-1.min-1) for one h in 11 healthy subjects. The insulin effect was measured with the euglycaemic clamp technique combined with D-(3-3H)-glucose infusion to evaluate glucose production and utilization. A control study with NaCl (154 mmol.l-1) infusion was also performed. The insulin levels during the clamps were similar in all studies (36 +/- 0.2 mU.l-1). Peak growth hormone levels were reached at 60 min (growth hormone 6 mU.kg-1.h-1: 31 +/- 5; growth hormone 12 mU.kg-1.h-1: 52 +/- 4 and growth hormone 24 mU.kg-1.h-1; 102 +/- 8 mU.l-1). The insulin-antagonistic effect of growth hormone started after approximately 2 h, was maximal after 4-5 h (approximately 39% inhibition of glucose infusion rate between control and growth hormone 24 mU.kg-1.h-1) and lasted for 6-7 h after peak levels. The resistance was due to a less pronounced insulin effect both to inhibit glucose production and to stimulate glucose utilization. Growth hormone infusion of 12 mU.kg-1.h-1 induced a similar insulin-antagonistic effect as the higher infusion rate whereas 6 mU.kg-1.h-1 induced a smaller response with a duration of 1 h between 3-4 h after peak levels of growth hormone. The present study demonstrates that growth hormone levels similar to those frequently seen in Type 1 (insulin-dependent) diabetic patients during poor metabolic control or hypoglycaemia, have pronounced insulin-antagonistic effects.(ABSTRACT TRUNCATED AT 250 WORDS)