The bioavailability of rapid-acting insulin administered as a nasal spray was studied in 6 type 1 (insulin-dependent) diabetic patients. They received long-acting bovine insulin (Ultratardum 40 U/ml, Organon) as basal treatment at 8 a.m. Rapid-acting insulin was also administered at 8 a.m., then at noon and 6 p.m, subcutaneously on day 1 as a 100 U/ml solution and intranasally by aerosol spray as a 100 U/ml and 500 U/ml with 1% (w/v) 9 lauryl ether solution on day 2 and day 3 respectively. On days 2 and 3, the dose of insulin was at least nine times higher than the subcutaneous dose on day 1. Free and total plasma insulin concentrations were assayed after the noon insulin administration. The peaks of the free and total plasma insulin levels were reached earlier and the return to basal levels was obtained earlier after nasal insulin administration than after insulin injected subcutaneously. The bioavailability of nasal spray insulin versus subcutaneous insulin with a 100 U/ml insulin solution was similar to that with a 500 U/ml insulin solution: 5.14 +/- 0.38% (m +/- SEM) and 4.64 +/- 0.46% according to the total plasma insulin level. This study suggests that the bioavailability of nasal spray insulin is not increased by increasing insulin concentration in our experimental conditions.