The pharmacokinetics of intranasal, an intravenous infusion, and intramuscular glucagon has been studied in 5 healthy subjects and 11 patients with insulin-dependent diabetes mellitus. After infusion the elimination half-life was significantly longer in diabetics (11.9 vs 6.6 min) and the apparent volume of distribution was twice as high in diabetics (0.19 vs 0.37 l.kg-1). The metabolic clearance rates were the same in the two groups (18.9 and 21.3 ml.min-1.kg-1 in controls and in diabetics) and were about twice those previously reported. After 1 mg intranasally the Cmax of immunoreactive glucagon (IRG) was similar in diabetic and in healthy subjects. Administration of a higher dose (2 mg) to diabetic patients produced a higher plasma level, although not proportionately so. The AUC after 1 mg was also similar in controls and in diabetics. The elimination half-life in both groups was similar to the value found after i.v. infusion; it was significantly shorter in controls (5.5 min) than in diabetics (13.8 min). In both groups, mean Cmax was significantly lower than after IM glucagon, the relative bioavailability of 1 mg intranasally vs IM injection being less than 30%. After IM administration, the Cmax and AUC of IRG in controls and in diabetic patients, were identical. The apparent elimination half-life was also similar in the two groups, and was three- to four-times longer (28.6 and 31.4 min) than after infusion or intranasal administration, possibly because estimation of the t1/2 was affected by slow release of the hormone from the site of injection.