The pharmacokinetics of several lung disposition pathways for pulmonary insulin were studied and modeled in the isolated perfused rat lung (IPRL). Insulin solution was administered by forced instillation into the airways of the IPRL as 0.1 or 0.02 ml doses of coarse spray, with or without bacitracin (BAC), N-ethylmaleimide (NEM) and atrial natriuretic peptide (ANP). Each insulin absorption profile was fitted to a kinetic model that incorporated the distribution fraction of the dose reaching the lobar region (DF) and the rate constants for absorption into perfusate ( k a) and non-absorptive loss ( k nal); k nal was shown to be due to the sum of mucociliary clearance and metabolism. Insulin absorption occurred largely by passive diffusion with values for k a = 0.39–0.50 h −1. With DF = 0.91 following 0.1 ml doses, 11.9 ± 3.4% of bioavailabilities were observed in 1 h. In contrast, derived values for k nal = 2.34–3.45 h −1 were significantly larger than the rate constant for mucociliary clearance determined previously in this IPRL (0.96–1.74 h −1) due to lung metabolism. Indeed, BAC, but neither NEM nor ANP, was found to decrease the value of k nal, which suggested that BAC-inhibitable lung ectopeptidases, and not insulin degrading enzyme (IDE), were responsible for this pulmonary metabolism. Shallower lung distribution with DF = 0.73 following 0.02 ml doses resulted in reduced values for k a = 0.27 h −1 and k nal = 2.79 h −1, indicating that these kinetic processes may be lung-region dependent, even within this model and emphasizing the likely importance of reliable lung deposition in vivo.