An experimental model of phenformininduced lactic acidosis was established in rats. Following a subtotal nephrectomy, renal failure developed (serum creatinine 4.5±0.1mg/100ml and 2.8±0.1 mg/100 ml on the 1st and 8th postoperative days respectively). Immediately after nephrectomy intra-peritoneal phenformin treatment, 16 mg/day, was commenced. Lactic acidosis developed progressively within 8 days, or earlier in the rats with the most severe renal insufficiency. The metabolic pattern was very similar to that observed in diabetic patients with a biguanide-induced lactic acidosis: on the 8th day, 2 h after the last phenformin injection, blood lactate was 10.8±1.0 mmol/1 (controls: 1.50±0.03); pyruvate was 0.56±0.06 mmol/1 (controls: 0.10±0.01) and blood pH: 7.00 ± 0.02 (vs 7.34±0.02); 3-hydroxybutyrate was 1.41±0.37 mmol/1 (vs 0.32 ±0.03); acetoacetate: 0.51±0.15 mmol/1 (vs 0.17 ±0.01), and free glycerol: 0.63 ±0.07 mmol/1 (vs 0.14 ±0.02). Increased concentrations of alanine (1.66±0.26 mmol/1, vs 0.48 ± 0.04 in controls) and low blood glucose levels (23± 8 mg/ 100 ml vs 70 ± 2, after a 12 hours fast) accompanied the lactic acidosis in spite of high glucagon levels (2030±170 pg/ml vs 108±10 in controls) and low insulin/glucagon molar ratio (0.19 vs 6.9 in controls). Normal rats, treated with phenformin at same doses, and nephrectomized rats injected with saline served as controls and remained free of lactic acidosis. Hydroxyphenformin (16 mg/day) injected in nephrectomized rats, was biologically inactive. Glucose production from14C-lactate was 425 ±85 μmol/100 g body wt/h, vs 1050 ± 90 in control animals. Blood lactate specific activity declined more slowly in the lactic acidotic rats than in controls, suggesting that a decrease in lactate utilization contributed to hyperlactataemia more than an increased lactate production.