Reduction of the prolonged terminal elimination phase of gentamicin may be caused by diabetes mellitus, irrespective of the model of diabetes. To test this hypothesis, five normal dogs, three dogs with alloxan-induced diabetes mellitus, and four dogs with naturally occurring diabetes mellitus (all of which were given exogenous insulin to control hyperglycemia) were given 4.4 mg/kg gentamicin intravenously. Serum pharmacokinetics were analyzed using non-compartmental pharmacokinetics assuming a sum of exponential terms. Gentamicin pharmacokinetics during the first 8 h were the same in normal and diabetic dogs. Over 7 days, MRT in normal dogs (5830 +/- 2970 min, mean +/- SD) was longer (P less than 0.01) than in diabetic dogs (136 +/- 164 min). In diabetic dogs, Cls was greater (3.01 +/- 0.86 ml/min/kg) than in normal dogs (1.45 +/- 0.11 ml/min/kg; P less than 0.01), whereas Vd(ss) was smaller in diabetic dogs (0.405 +/- 0.508 l/kg) than in normal dogs (8.56 +/- 4.48 l/kg; P less than 0.01). Serum gentamicin concentrations were less than 0.020 microgram/ml by 2 days in all of the diabetic dogs, but were 0.048 +/- 0.018 microgram/ml at 7 days in normal dogs. Thus, diabetes mellitus, either induced by alloxan administration or naturally occurring, abolished the terminal elimination phase of gentamicin disposition in a non-rodent species.