To compare the pharmacokinetics of two different concentrations, containing either 40 or 100 IU/ml of short-acting human insulin (Velasulin HM), intermediate-acting human insulin (Insulatard HM), or an insulin mixture (25% shortacting insulin, 75% intermediate-acting insulin; Mixtard HM), three randomized, single-blind, crossover trials were performed using the euglycemic clamp technique. Eighteen healthy volunteers received insulin of either formulation subcutaneously in each of the studies (15 IU Velasulin, 20 IU Insulatard, or Mixtard). The blood glucose levels were maintained constant by glucose infusions. In the trial using Velasulin, the two different insulin concentrations were equivalent regarding the total absorption [area under the curve (AUC) of serum insulin" 126_+28 and 123+35 mU/lx 12 h for U40 and UI00 (2+ SD)], but not in regard to the rate of absorption (t max 1.3 _+ 0.4 and 2.4_+ 1.0 h for U40 and U100). In the case of Insulatard, total absorption was not equivalent (AUC 153 _+ 35 and 128_+ 37 ml/1 x 24 h for U40 and U100), but the rate of absorption was equivalent (t max 4.8 _+-2.9 and 5.3_+4.6h). In the Mixtard series, totM absorption was equivalent (AUC 142_+32 and 128_+22 mU/1 x 24 h), but the rate of absorption was not (t max 2.2+0.9 and 3.2_+4.2 h for U40 and UI00). The glucose requirement was not equivalent in each of the three series. It was higher for the U-IO0 formulation in the Velasulin trial (2363+--578 ml/12 h for U40 and 2601 +820 mI/ 12 h for UIO0) and lower in the Insulatard trial (2203+1271 ml/24h for U40 and 1864+864ml/ 24h for UIO0) and the Mixtard trial (2559+ 914 ml/24 h for U40 and 2067+896 ml/24 h for UIO0). When switching insulin from U40 to UIO0 formulations, more frequent blood glucose moni-Abbreviations. AUC = area under curve; c max = maximal concentration; t max = time when c max is reached; RIA = radioimmunoassay; U40=insulin concentration 40 IU/ml; U100=insulin concentration 100 IU/ml toring in the early phase following the insulin injection is recommended.