Background. Patients with acute renal failure (ARF) have high morbidity and mortality rates, particularly if they have serious comorbid conditions. Several studies indicate that in rats with ARF caused by ischemia or certain nephrotoxins, insulin-like growth factor-I (IGF-I) enhances the recovery of renal function and suppresses protein catabolism. Methods. Our objective was to determine whether injections of recombinant human IGF-I (rhIGF-I) would enhance the recovery of renal function and is safe in patients with ARF. The study was designed as a randomized, double-blind, placebo-controlled trial in intensive care units in 20 teaching hospitals. Seventy-two patients with ARF were randomized to receive rhIGF-I (35 patients) or placebo (37 patients). The most common causes of ARF in the rhIGF-I and placebo groups were, respectively, sepsis (37 and 35% of patients) and hypotension or hemodynamic shock (42 and 27% of patients). At baseline, the mean (± SD) APACHE II scores in the rhIGF-I and placebo-treated groups were 24 ± 5 and 25 ± 8, respectively. In the rhIGF-I and placebo groups, the mean (median) urine volume and urinary iothalamate clearances (glomerular filtration rate) were 1116 ± 1037 (887) and 1402 ± 1183 (1430) ml/24 hr and 6.4 ± 5.9 (4.3) and 8.7 ± 7.2 (4.4) ml/min and did not differ between the two groups. Patients were injected subcutaneously every 12 hours with rhIGF-I, 100 μg/kg desirable body weight, or placebo for up to 14 days. Injections were started within six days of the onset of ARF. The primary endpoint was a change in glomerular filtration rate from baseline. Other end points included changes from baseline in urine volume, creatinine clearance and serum urea, creatinine, albumin and transferrin, frequency of hemodialysis or ultrafiltration, and mortality rate. Results. During the treatment period, which averaged 10.7 ± 4.1 and 10.6 ± 4.5 days in the rhIGF-I and placebo groups, there were no differences in the changes from baseline values of the glomerular filtration rate, creatinine clearance, daily urine volume, or serum urea nitrogen, creatinine, albumin or transferrin. In patients who did not receive renal replacement therapy, there was also no significant difference in serum creatinine and urea between the two groups. Twenty patients in the rhIGF-I group and 17 placebo-treated patients underwent dialysis or ultrafiltration. Twelve rhIGF-I-treated patients and 12 placebo- treated patients died during the 28 days after the onset of treatment. Conclusions. rhIGF-I does not accelerate the recovery of renal function in ARF patients with substantial comorbidity.