To investigate the influence of opiates on insulin action in vivo, we induced mild physiological hyperinsulinaemia (15-20 mU/l) in five trained conscious dogs in the absence or presence of ongoing infusion with the opiate agonist D-met2-pro5-enkephalinamide (DMPE, 0.5 micrograms X kg-1 X min-1), or the opiate antagonist naloxone (1.25 mg followed by 1 microgram X kg-1 X min-1). The effects on glucose production and glucose utilization were measured by isotope dilution using 3-3H-glucose. Glucose fell similarly over 30 min in response to insulin in controls (0.021 +/- 0.003 mmol X l-1 X min-1), and both the DMPE and naloxone studies (0.016 +/- 0.002 mmol X l-1 X min-1 and 0.017 +/- 0.003 mmol X l-1 X min-1, respectively). In control dogs, insulin lowered glucose by transiently suppressing production by 0.028 +/- 0.006 mmol X kg-1 X min-1 at 20-30 min without changing utilization. In contrast, in both the DMPE and naloxone studies insulin lowered glucose by markedly raising utilization at 20 min by 0.094 +/- 0.017 and 0.139 +/- 0.022 mmol X kg-1 X min-1, respectively. Furthermore, insulin failed to suppress production in both DMPE and naloxone studies and, as plasma glucose fell, production rose in both treatment groups at 20 min by 0.045 +/- 0.012 and 0.089 +/- 0.022 mmol X kg-1 X min-1 respectively.(ABSTRACT TRUNCATED AT 250 WORDS)