One hundred twenty-eight patients with different renal diseases and chronic renal failure, stratified according to the underlying disease, were enrolled in a randomized controlled trial to investigate the effects on the rate of decline of renal function of two diets, a controlled protein diet (CPD) of 1 g protein/kg ideal body-weight (i.b.w.)/day, and a low-protein diet (LPD) of 0.6 g protein/kg i.b.w./day, given for 27.1 +/- 21.8 months. Dietary compliance was assessed by a dietary questionnaire, dietary interviews and measurement of 24-h urinary urea excretion. At the end of 6 months, actual mean protein intake was higher than expected (1.06 +/- 0.25 g/kg i.b.w./day) in CPD patients, and (0.80 +/- 0.21 g/kg i.b.w./day) in LPD patients: values were similar at 12 and 18 months after the time of enrollment. The end-point, defined as halving of creatinine clearance, was reached in 40% of patients on CPD, and in 28.6% of those on LPD (P = 0.038 by comparative life-table analysis). Multivariate regression analysis confirmed that CPD was associated with a higher risk of progression than LPD, and that two additional parameters (creatinine clearance at the time of randomization and average proteinuria during the follow-up) were significant independent risk factors, even more important than protein intake.