To examine the effect of different profiles of insulin administration on glycaemia and thermogenesis, we studied 10 subjects with mild non-insulin-dependent diabetes mellitus on four occasions after a standard mixed meal: (1) with no supplementary insulin (control), (2) with intravenous insulin (1.8U over 15 min = short), (3) as for short but extended over 30 min to simulate the normal initial rise in portal vein insulin levels (medium), (4) as for medium with additional insulin to normalize the profile from 30-60 min (3.6U over 60 min, long). All studies in which supplemental insulin was administered lowered the integrated glucose response above baseline versus the control study (short 76%, medium 71%, and long 56% of control, p = 0.003). The insulin infusions also increased the non-protein respiratory quotient in the first hour following the meal (0.82 +/- 0.01 (control) vs 0.87 +/- 0.01 (short), 0.86 +/- 0.01 (medium) and 0.87 +/- 0.01 (long), p = 0.003) and augmented thermogenesis (7.6 +/- 1.5 (control) vs 10.5 +/- 2.9 (short), 13.0 +/- 1.9 (medium) and 13.2 +/- 2.8% (long), p = 0.02). Total integrated insulin area above baseline was significantly greater in the long study (short 121, medium 111 vs long 179% of control, p = 0.02). Thus the greatest glycaemic benefit in relation to insulinaemia was obtained with the two shorter insulin infusions (short and medium). In conclusion, this study confirms the role of early prandial insulin secretion (or delivery) in limiting prandial glycaemia in NIDDM and increasing thermogenesis and highlights the pivotal role of the timing of elevation of insulin levels in modulating hyperglycaemia and hyperinsulinaemia.