To assess the effect of a low protein diet (LPD) on renal function and metabolic control in three sub-groups of patients with type 2 diabetes those with or without nephropathy. A randomized clinical trial was conducted on 60 patients with type 2 diabetes in primary care -19 with normoalbuminuria, 22 with microalbuminuria, and 19 with macroalbuminuria-. All patients experienced a screening phase during the 3 months, and were designated according to percentages of daily caloric intake (e.g., carbohydrates 50%, fat 30%, and 20% of protein). After this period, they were randomly assigned to receive either LPD (0.6-0.8 g/kg per day) or normal protein diet (NPD) (1.0-1.2 g/kg per day) for a period of 4 months. Twenty nine patients received LPD and 31 received NPD. Primary endpoints included measures of renal function (UAER, serum creatinine and GFR) and glycemic control (fasting glucose and glycosylated hemoglobin A1c). Renal function improved among patients with macroalbuminuria who received LPD: UAER decreased (1,280.7 +/- 1,139.7 to 444.4 +/- 329.8 mg/24 h; p < 0.05) and GFR increased (56.3 +/- 29.0-74.2 +/- 40.4 ml/min; p < 0.05). In normoalbuminuric and microalbuminuric patients, there were no significant changes in UAER or GFR after either diet. HbA1c decreased significantly among microalbuminuric patients on both diets (LPD, 8.2 +/- 1.6-7.2 +/- 1.8%; p < 0.05; NPD, 8.8 +/- 1.9-7.1 +/- 0.8%; p < 0.05) and among macroalbuminuric patients who received NPD (8.1 +/- 1.8-6.9 +/- 1.6%; p < 0.05). A moderated protein restriction diet improved the renal function in patients with type diabetes 2 and macroalbuminuria.