OBJECTIVE Linagliptin (BI 1356) is a highly specific inhibitor of dipeptidyl peptidase (DPP)-4, which is currently in phase III clinical development for the treatment of type 2 diabetes mellitus. Linagliptin exhibits nonlinear pharmacokinetics after oral administration, which are mainly related to concentration-dependent binding of linagliptin to its target, DPP-4. The objectives of the study were to investigate the pharmacokinetics and pharmacodynamics after intravenous administration of linagliptin and to determine its absolute bioavailability (F). METHODS This was a single rising-dose, randomized, four-group, placebo-controlled, single-blind (within dose groups) study. Thirty-six healthy men aged 18-50 years were enrolled and randomized into four sequential treatment groups. Group 1 received linagliptin 0.5 mg intravenously, group 2 received 2.5 mg intravenously and group 4 received 10 mg intravenously. In group 3, subjects underwent a two-way randomized crossover, receiving 5 mg intravenously and a 10 mg oral tablet. Linagliptin concentrations in plasma and urine, as well as plasma DPP-4 activity, were determined by validated assays. Noncompartmental analysis and population pharmacokinetic modelling were performed. RESULTS Linagliptin showed nonlinear pharmacokinetics after intravenous infusion of 0.5-10 mg, with a less than dose-proportional increase in exposure. Noncompartmental parameters were calculated on the basis of total (i.e. bound and unbound) plasma concentrations. The total clearance value was low and increased with dose from 2.51 to 14.3 L/h. The apparent steady-state volume of distribution (V(ss)) increased with dose from 380 to 1540 L. Renal excretion of the unchanged parent compound increased with increasing plasma concentrations from 2.72% in the 0.5 mg dose group to 23.0% in the 10 mg dose group. The terminal elimination half-life was comparable across dose groups (126-139 hours). Because of the nonlinear pharmacokinetics, the standard approach of comparing the area under the plasma concentration-time curve (AUC) after oral administration with the AUC after intravenous administration led to dose-dependent estimates of the absolute bioavailability. Therefore, a population pharmacokinetic model was developed, accounting for the concentration-dependent protein binding of linagliptin to its target enzyme, DPP-4. The model-derived estimates of the V(ss) and clearance of linagliptin not bound to DPP-4 were 402.2 L and 26.9 L/h, respectively. The absolute bioavailability was estimated to be about 30% for the linagliptin 10 mg tablet. CONCLUSIONS The nonlinear pharmacokinetic characteristics and the pharmacokinetic/pharmacodynamic relationship of linagliptin were independent of the mode of administration (intravenous or oral). Because of the nonlinear pharmacokinetics, the standard approach of comparing the AUC after oral administration with the AUC after intravenous administration was inappropriate to determine the absolute bioavailability of linagliptin. By a modelling approach, the absolute bioavailability of the 10 mg linagliptin tablet was estimated to be about 30%.