The effects were compared of varying aerosol particle size and aerosolization time within each breath on the metabolic effect elicited by inhalation of a liquid insulin aerosol in comparison with that after subcutaneous injection (s.c.) of regular insulin. In this single-center, open-label euglycemic glucose clamp study, 13 healthy non-smoking subjects received five administrations of insulin in randomized order on separate study days, once by s.c. (0.15 U/kg of regular insulin) and four times by inhalation. Subjects inhaled 1.5 U/kg of liquid insulin aerosol administered by the Aerodose Insulin Inhaler (Aerogen Inc., Mountain View, CA) configured to deliver two aerosol particle sizes--fine [F, 4.4 +/- 0.3 microm (mean +/- SD)] or very fine (VF, 3.5 +/- 0.2 microm)--and two aerosolization times (aerosol released for the first 2 or 4 s after the start of each 5-s inhalation). Glucose infusion rate (GIR) values necessary to keep blood glucose concentrations constant at 5.0 mmol/L were determined over a 6-h period following insulin administration. After inhalation of insulin, the onset of action was substantially more rapid on all four inhalation study days than after s.c. insulin, and the time to maximal action [t(GIRmax) (min)] was reached earlier: F/2 s, 127 +/- 54; F/4 s, 128 +/- 55; VF/2 s, 158 +/- 91; VF/4 s, 132 +/- 72; s.c., 175 +/- 69 (P < 0.0001). The longer aerosolization time (4 vs. 2 s) resulted in higher maximal metabolic action [GIR(max) (mg/kg/min), F/4 s 8.1 +/- 3.6, VF/4 s 8.4 +/- 2.7 vs. F/2 s 6.6 +/- 2.4, VF/2 s 7.2 +/- 2.4 (P = 0.01 for 4 s vs. 2 s, grouped data)], total metabolic activity [area under the curve of GIR 0-6 h (g/kg), F/4 s 1.97 +/- 0.92, VF/4 s 2.14 +/- 0.86 vs. F/2 s 1.56 +/- 0.68, VF/2 s 1.78 +/- 0.60 (P = 0.01)], and relative biopotency [F/4 s 10.6 +/- 4.0%, VF/4 s 11.7% +/- 4.1% vs. F/2 s 8.5 +/- 3.2%, VF/2 s 9.7 +/- 2.4% (P = 0.01)]. None of these summary measures was significantly affected by particle size. No drug- or device-related adverse events were observed. This study shows that aerosolization time, but not particle size, in the ranges studied, had an impact on the metabolic effect elicited by inhaled insulin, allowing rational selection of delivery parameters for further clinical testing. Based on the observed biopotency and the rapid onset of action, inhalation of a liquid insulin aerosol generated by the Aerodose Insulin Inhaler shows promise for covering prandial insulin requirements.