BACKGROUND: Patients with type 1 diabetes require intensive insulin therapy for optimal glycemic control. AIR((R)) inhaled insulin (system from Eli Lilly and Company, Indianapolis, IN) (AIR is a registered trademark of Alkermes, Inc., Cambridge, MA) may be an efficacious and safe alternative to subcutaneously injected (SC) mealtime insulin. METHODS: This was a Phase 3, 2-year, randomized, open-label, active-comparator, parallel-group study in 385 patients with type 1 diabetes who were randomly assigned to receive AIR insulin or SC insulin (regular human insulin or insulin lispro) at mealtimes. Both groups received insulin glargine once daily. Efficacy measures included mean change in hemoglobin A1C (A1C) from baseline to end point, eight-point self-monitored blood glucose profiles, and insulin dosage. Safety assessments included hypoglycemic events, pulmonary function tests, adverse events, and insulin antibody levels. RESULTS: In both treatment groups, only 20% of subjects reached the target of A1C <7.0%. A significant A1C difference of 0.44% was seen favoring SC insulin, with no difference between the groups in insulin doses or hypoglycemic events at end point. Patients in both treatment groups experienced progressive decreases in lung function, but larger (reversible) decrements in diffusing capacity of the lung for carbon monoxide (DL(CO)) were associated with AIR insulin treatment. Greater weight gain was seen with SC insulin treatment. CONCLUSIONS: The AIR inhaled insulin program was terminated by the sponsor prior to availability of any Phase 3 data for reasons unrelated to safety or efficacy. Despite early termination, this trial provides evidence that AIR insulin was less efficacious in lowering A1C and was associated with a greater decrease in DL(CO) and increased incidence of cough than SC insulin in patients with type 1 diabetes.