NN304 is a long-acting insulin analogue that is acylated with a 14-C-fatty acid chain. Protraction of action of this novel insulin analogue is due not to slow absorption after subcutaneous administration but to reversible binding to albumin. We investigated the pharmacokinetic and pharmacodynamic properties of insulin analogue NN304 (0.3 and 0.6 U/kg) in comparison to NPH insulin (0.3 and 0.6 IU/kg) in 10 healthy volunteers performing a randomised, double-blind, cross-over, placebo-controlled glucose clamp study. During the observation period of 24 hours the areas under the insulin curve for NPH[0.3 IU/kg] vs. NPH[0.6 IU/kg] were 60 vs. 102 nmol min l(-1) (p<0.01) and for insulin analogue NN304[0.3 U/kg] vs. NN304[0.6 U/kg] 490 vs. 932 nmol min l(-1) (p <0.001), suggesting a clear dose-response relationship for both NPH insulin and NN304. The amount of disposed glucose (area under the curve of glucose infusion) differed with statistical significance between the five treatments and was highest with NPH[0.6 IU/kg] (2671 mg/kg) and lowest with placebo (265 mg/kg). However, area under the curve of glucose infusion after treatment with NN304 was only 36% (dose of 0.3 U/kg) and 24% (dose of 0.6 U/kg) of that observed with corresponding doses of NPH insulin. Moreover, increasing dosages of NN304 failed to demonstrate a significant dose-response with regard to the area under the curve of glucose infusion. This study demonstrates that the principle of protracted insulin action of NN304 by reversible binding to albumin is effective in humans albeit at a much lower rate of glucose utilisation when compared to NPH insulin. Thus, in contrast to animal studies NN304 and NPH insulin can not be considered equipotent in humans.