Abnormalities in GH release have been found in adults with poorly controlled type I diabetes mellitus. During puberty, circulating GH concentrations transiently increase. To investigate in pubertal diabetic adolescents, the physiological relationship between metabolic control and GH release, we compared spontaneous and GH-releasing hormone (GHRH)-stimulated GH release in six pubertal subjects during poor (study A) and improved (study B) metabolic control. The subjects included two females and four males (mean age +/- SE, 15.5 +/- 1 yr; duration of diabetes, 8.6 +/- 0.9 yr; Tanner stages II-V). Serum samples for glucose and GH determinations were obtained at 20-min intervals over a 24-h period. Significant pulses of GH release were identified using a pulse detection algorithm (Cluster). Fourier expansion time series was used to document the occurrence of significant periodicities in the GH concentration-time data series. All subjects received 1.0 microgram/kg GHRH-44, iv, at 0800 h on the day after the 24-h monitoring for GH. After GHRH administration, samples were taken for glucose and GH determinations over 90 min. The overall mean glucose level (+/- SE) during the 24-h monitoring was 11.5 +/- 0.2 mmol/L during study A and 7.2 +/- 0.2 during study B (P = 0.0001). During the 4 weeks of improved control, glycated hemoglobin fell from 13.9 +/- 1.4% to 11.7 +/- 0.8% (mean +/- SE; P < 0.025). All subjects had significant pulses of GH release during poor or improved metabolic control. Relative to that at night, the daytime pulse frequency was higher in study A (P < 0.025). The overnight pulse frequency increased during study B (P < 0.01). Other pulse parameters, including maximal and incremental pulse amplitudes, pulse width, and interpulse valley mean, did not change during improved control. The mean +/- SE 24-h GH concentration was 4.1 +/- 0.7 micrograms/L during study A and 4.3 +/- 0.8 during study B. The amplitude of the circadian GH rhythm was not different by Fourier analysis. The overall mean glucose +/- SE after GHRH administration was 15.3 +/- 0.2 mmol/L in study A and 6.8 +/- 0.1 in study B. In spite of the marked hyperglycemia during study A, the GH responses were similar during studies A and B. Maximal GH levels were obtained at 15-30 min (mean +/- SE) and were 36.0 +/- 16.9 micrograms/L in study A and 38.7 +/- 18.9 in study B.(ABSTRACT TRUNCATED AT 400 WORDS)