The gut hormone, glucagon-like peptide-1 (GLP-1) is a potent insulin secretogogue with potential as a therapy for non-insulin-dependent diabetes mellitus (NIDDM). GLP-1 has been shown to reduce glucose concentrations, both basally, and, independently, in response to a single meal. For it to be an effective treatment, it would need to be administered as a long-acting therapy, but this might not be feasible due to the profound delay in gastric emptying induced by GLP-1. In order to assess the feasibility and efficacy of continuous administration of GLP-1 in NIDDM, we determined the effects of continuous intravenous infusion of GLP-1 (7–36) amide, from 22.00–17.00 hours, on glucose and insulin concentrations overnight and in response to three standard meals, in eight subjects with NIDDM. These were compared with responses to 0.9 % NaCl infusion and responses in six non-diabetic control subjects who were not receiving GLP-1. Effects on beta-cell function were assessed in the basal state using homeostasis model assessment (HOMA) and in the postprandial state by dividing incremental insulin responses to breakfast by incremental glucose responses. To assess possible clinical benefit from priming of beta cells by GLP-1 given overnight only, a third study assessed the effect of GLP-1 given from 22.00–07.30 hours on subsequent glucose responses the next day. Continuous GLP-1 infusion markedly reduced overnight glucose concentrations (mean from 24.00–08.00 hours) from median (range) 7.8 (6.1–13.8) to 5.1 (4.0–9.2) mmol/l (p < 0.02), not significantly different from control subjects, 5.6 (5.0–5.8) mmol/l. Daytime glucose concentrations (mean from 08.00–17.00 hours) were reduced from 11.0 (9.3–16.4) to 7.6 (4.9–11.5) mmol/l (p < 0.02), not significantly different from control subjects, 6.7 (6.5–7.0) mmol/l. GLP-1 improved beta-cell function in the basal state from 62 (13–102) to 116 (46–180) %β (p < 0.02) and following breakfast from 57 (19–185) to 113 (31–494) pmol/mmol (p < 0.02). GLP-1 only given overnight did not improve the glucose responses to meals the next day. In conclusion, continuous infusion of GLP-1 markedly reduced diurnal glucose concentrations, suggesting that continuous GLP-1 administration may be a useful therapy in NIDDM. [Diabetologia (1997) 40: 205–211]