OBJECTIVE: Effective type 2 diabetes management requires prompt intervention if glycemic control is not achieved by nonpharmacological means. This study investigates whether inhaled insulin (INH; Exubera) can achieve target glycemic control in patients failing on diet and exercise. RESEARCH DESIGN AND METHODS: Patients with suboptimal control on diet and exercise (HbA(1c) [A1C] 8-11%) were randomized to 3 months' treatment with either INH before meals (n = 76) or rosiglitazone 4 mg twice a day (n = 69), in conjunction with a diet and exercise regimen. The primary end point was percentage of patients achieving A1C <8.0%. RESULTS: The INH and rosiglitazone groups had comparable baseline A1C values (9.5 vs. 9.4%, respectively). Significantly more patients achieved A1C <8.0% (83 vs. 58%, adjusted odds ratio 7.14 [95% CI 2.48-20.58], P = 0.0003), A1C <7.0% (44 vs. 18%, 4.43 [1.94-10.12]), and A1C < or = 6.5% (28 vs. 7.5% 5.34 [1.83-15.57]) with INH. A1C decrease was greater with INH (-2.3% vs. -1.4%, adjusted treatment group difference: -0.89% [95% CI -1.23 to -0.55]) with final mean A1C values of 7.2 and 8.0% for INH and rosiglitazone, respectively. Hypoglycemia (episodes per subject-month) was higher with INH (0.7 vs. 0.05, risk ratio 14.72 [95% CI 7.51-28.83]), with no severe hypoglycemic episodes. Pulmonary function changes were small and comparable between groups. CONCLUSIONS: INH could be an effective therapy for people with type 2 diabetes early in the course of their disease.