To evaluate the metabolic effects of long-term treatment with recombinant human (rh) GH in short children with chronic renal failure (CRF), annual oral glucose tolerance tests (oGTT) during rhGH therapy for up to 5 y in 53 prepubertal children with CRF on conservative treatment, dialysis, and after renal transplantation were compared with that of 12 age-matched children treated with rhGH for idiopathic short stature. At the start of rhGH treatment, fasting values of glucose, insulin, glycosylated Hb A (HbA1C), triglycerides, cholesterol, glucose, and insulin responses during oGTT were significantly elevated in all patient groups compared with control subjects (p < 0.001). In the total population, fasting and 2-h postprandial glucose concentrations were inversely correlated with GFR and positively with age and methylprednisolone dosage in transplanted patients. Fasting insulin levels were positively correlated with body mass index and inversely with GFR. RhGH treatment was not associated with a change in fasting or stimulated glucose concentrations in any treatment group throughout the observation period. In contrast, serum insulin levels increased during the first treatment year in all groups, resulting in a more marked elevation of integrated insulin levels in transplant (1402 +/- 179 pM) and dialysis (1025 +/- 114 pM) patients compared with conservatively treated patients (829 +/- 94 pM), and control subjects (719 +/- 89 pM) (p < 0.01). Hyperinsulinemia persisted in all treatment groups for up to 5 y of follow-up. In conclusion, age, renal function, and obesity are the major independent predictors of glucose tolerance in children with CRF. Long-term rhGH treatment does not affect glucose tolerance, but aggravates the preexisting hyperinsulinemia in children with end-stage renal disease. In concert with the dyslipidemia of uremia, the rhGH-promoted hyperinsulinemia may contribute to the long-term risk for premature atherosclerosis in patients with childhood onset CRF.