Aims/hypothesis. The AERx insulin Diabetes Management System (AERx iDMS) is a liquid aerosol device that enables insulin to be administered to the peripheral parts of the lung. This study aimed to compare the pharmacokinetic and pharmacodynamic properties of insulin which is inhaled using AERx iDMS with insulin which is subcutaneously administered. Methods. In total, 18 C-peptide negative patients with Type I (insulin-dependent) diabetes mellitus participated in this randomised, open-label, 5-period cross-over trial. Human regular insulin was administered subcutaneously (0.12 U/kg body weight) or inhaled by means of the AERx iDMS (dosages 0.3, 0.6, 1.2, and 1.8 U/kg body weight). Thereafter plasma glucose was kept constant at 7.2 mmol/l for a 10-h period (glucose clamp technique). Results. Inhaled insulin provided a dose-response relation that was close to linear for both pharmacokinetic (AUC-Ins_{(0–10 h)}; Cmax-Ins) and pharmacodynamic (AUC-GIR_{(0–10 h)}; GIRmax) parameters. Time to maximum insulin concentration (Tmax-Ins) and time to maximum glucose infusion rate (TGIRmax) were shorter with inhaled insulin than with subcutaneous administration. The pharmacodynamic system efficiency of inhaled insulin (AUC-GIR_{(0–6 h)}) was 12.7 % (95 % C. I.: 10.2–15.6). Conclusion/interpretation. The inhalation of soluble human insulin using the AERx iDMS is feasible and provides a clear dose response. Further long-term studies are required to investigate safety aspects, HbA_{1 c} values, incidence of hypoglycaemic events and the quality of life. [Diabetologia (2001) 44: 305–308]