Low-protein diet in chronic kidney disease: from questions of effectiveness to those of feasibility.

A low-protein diet (LPD) as a therapeutic measure in chronic kidney disease (CKD) was suggested by Beale as early as 1869 [1], and the first attempt to evaluate experimentally LPD in humans was fulfilled by Smith in 1926 [2]. In the mid 1960s, Giordano and Giovannetti were the first to show that LPD, supplemented with essential amino acids to achieve neutral nitrogen balance, was able to reduce almost all uraemic signs and symptoms [3, 4]. By lowering blood urea and other nitrogenous waste products, LPD has favourable effects on secondary hyperparathyroidism [5], peripheral resistance to insulin [6], hyperlipidaemia [7], hypertension and acid– base disorders [8]. For instance, Goraya et al. [8] have shown that a vegetarian diet in CKD patients Stage II (60–90 mL/min) significantly decreased the 8-h urine net acid excretion, potential renal acid load, urine albumin-to-creatinine ratio, urine N-acetyl-β-D-glucosaminidase-to-creatinine ratio and the urine transforming growth factor-β-to-creatinine ratio at 30 days, as compared with a control group. They also showed a 30-day greater decrease in systolic blood pressure, plasma and urine excretion of potassium, aldosterone, endothelin and urine excretion of sodium in the vegetarian group. As maintenance dialysis is generally initiated when uraemic symptoms begin, the need to start it may be deferred by LPD [9]. For instance, Walser et al. [10] found that dialysis can be safely deferred by LPD for a median of 1 year after patients reach a glomerular filtration rate (GFR) level of 10 mL/min among non-diabetics and 15 mL/min among diabetics. In the 1980s, the rapid development of kidney replacement therapies led to an enormous increase in expenditure, but mortality and morbidity remained high for patients receiving dialysis. This observation further raised the interest of health providers and researchers in interventions for slowing the deterioration of kidney function in order to delay end-stage renal disease (ESRD). Since that time, many experimental and observational studies have addressed the question of the ability of LPD (protein intake ≤0.8 g/kg/day), or very LPD (protein intake ≤0.3 g/kg/day), to retard the progression of CKD towards ESRD. Fouque et al. [11] identified 46 studies conducted between 1975 and 1991 that addressed this issue in non-diabetic CKD patients, and Pan et al. [12] identified 26 studies published up to 2008 in diabetic CKD patients. Despite the many studies performed over more than 30 years, the effectiveness of LPD in preventing ESRD among diabetic or nondiabetic CKD patients remains uncertain, with largely conflicting results. The Modification of Diet in Renal Disease (MDRD) was the largest randomized clinical trial to test the hypothesis that LPD slows the progression of kidney disease among 1840 patients with various stages of CKD. The primary results published in 1994 were not conclusive with regard to the effectiveness of this intervention [13]. However, following secondary analyses undertaken later, the authors concluded that ‘the balance of evidence is more consistent with the hypothesis of a beneficial effect of protein restriction than with the contrary hypothesis of no beneficial effect’ [14]. Five metaanalyses of studies of the effects of LPD on CKD progression in diabetic and non-diabetic patients have been performed since the early 1990s, four in favour of a beneficial effect [9, 11, 15, 16] and one not [12]. The reasons for the discrepancies between the results of studies conducted to evaluate LPD are of particular interest. Comparison of their designs reveals great heterogeneity: