To assess the effect of metyrapone on the early morning plasma glucose (PG) rise, seven NIDDM patients were studied from 2400 to 0900 h on two separate occasions one week apart. During the control study nights, patients received conventional therapy only (diet plus sulphonylurea) whereas on treatment nights, patients received in addition 30 mg/kg metyrapone orally at 2400 h. The plasma glucose (PG) levels from 0530 to 0900 h were significantly higher during the control night than the corresponding values following metyrapone. The control mean PG concentrations increased continuously from a nadir 8.4 +/- 1.1 mmol/l at 0400 h to a maximum of 9.4 +/- 1.1 mmol/l at 0800 h (p less than 0.01). In contrast following metyrapone administration a continuous decline in the PG concentration was noted from 2400 to 0800 h. The plasma glucose levels fell from 9.0 +/- 1.2 at 0400 h to 7.7 +/- 1.0 mmol/l at 0800 h (p less than 0.05). The mean overnight cortisol levels were 167.2 +/- 13.2 and 55.9 +/- 6.4 nmol/l (p less than 0.001) during the control and treatment studies, respectively. The cortisol levels were significantly higher during the control study at all time points from 0400 to 0900 h. No significant changes in insulin, C-peptide, glucagon, GH or catecholamine levels were observed between the two study periods. We conclude that the physiologic early morning rise in plasma cortisol possibly contributes to the pathogenesis of the dawn phenomenon in NIDDM patients.