Aims To compare the pharmacodynamic properties of insulin detemir (detemir) and neutral protamine lispro (NPL) insulin using a euglycaemic glucose clamp. Methods In a double-blind, crossover study, 30 patients with C-peptide negative type 1 diabetes were randomly assigned to a single dose (0.4 U/kg) of detemir and NPL. Plasma glucose (PG) was normalized with a variable insulin infusion and then decreased stepwise, followed by a euglycaemic clamp at 5.5 mmol/l over 32 h. Duration of action was defined as time from dosing until PG exceeded 8.3 mmol/l for at least 30 min. Results Duration of action was similar for detemir [23.0 (range 2.2532) h] and NPL [22.0 (9.532) h], p?=?0.55. Using glucose infusion rate (GIR) parameters, detemir showed a flatter pharmacodynamic profile versus NPL: area under the curve, AUCGIR(032)?=?1326 vs. 1841 mg/kg, p?<?0.01 (detemir vs. NPL, respectively); AUCGIR(012)?=?784 vs. 1392 mg/kg, p?<?0.05; AUCGIR(1232)?=?455 vs. 274 mg/kg, p?=?0.051; GIRlate(1232)/GIRearly(012) ratio?=?0.33 vs. 0.04, p?<?0.001. Detemir also showed a lower and later peak of action than NPL [GIRmax 2.0 vs. 3.2 mg/kg/min, p?<?0.01; Tmax 9.1 (95% confidence interval: 3.014.7) vs. 7.0 h (1.815.2)]. Conclusions Detemir and NPL had similar duration of action of approximately 24 h in patients with type 1 diabetes. Compared with NPL, detemir had a flatter profile with a more even distribution of metabolic effect over 24 h.