The aims of the study are to evaluate the effect of low-phosphate diet on FGF23 level and to determine the optimal amount of dietary phosphate restriction in hemodialysis patients. In particular, the investigators will assess the comparing effect of pre-specified low-phosphate diets, very low-phosphate diet, phosphate-to-protein ratio (PPR) value of 8 mg/g, versus low-phosphate diet, PPR value of 10 mg/g, on the change of FGF23 and phosphate level.

In patients with hemodialysis, the prevalence of cardiovascular disease (CVD) is high, and is the leading cause of death. Among several cardiovascular risk factors of hemodialysis patients, elevated fibroblast growth factor-23 (FGF23) level is common, and plays major role in the development of CVD with independent pathophysiologic mechanisms. Evidence from animal studies demonstrated that low-phosphate diet reduced FGF23 level. Clinical trials assessing the effect of dietary phosphate restriction on FGF23 focused on non-dialysis populution. However, little is known about the effect of low-phosphate diet on FGF23 in hemodialysis patients who have higher prevalence of hyperphosphatemia and severely elevated FGF23 level. In addition, current clinical guideline, based on evidence from observational studies of non-dialysis population, has recommended that dietary phosphate intake should be restricted to 800-1000 mg/day (adjusted for dietary protein needs) when serum phosphate levels are greater than 5.5 mg/dL in those with kidney failure. For hemodialysis population, the optimal amount of dietary phosphate restriction has not been determined. The aims of the study are to evaluate the effect of low-phosphate diet on FGF23 level and to determine the optimal amount of dietary phosphate restriction in hemodialysis patients. In particular, the investigators will compare the effect of pre-specified low-phosphate diets, very low-phosphate diet, phosphate-to-protein ratio (PPR) value of 8 mg/g, versus low-phosphate diet, PPR value of 10 mg/g, on the change of FGF23 and phosphate level.

It is to conduct a randomized, active-controlled trial with cross-over design at a hemodialysis unit of tertiary teaching hospital in Northern Taiwan. Subjects with aged older than 20 years, end-stage renal disease undergoing thrice-weekly hemodialysis for more than three months, having adequate dialysis (urea reduction ratio equal to or greater than 65%) and the most recent serum phosphate level greater than 5.5 mg/dL or between 3.5 and 5.5 mg/dL with regular phosphate binder use will be randomly assigned into two groups: those in group A will receive 2-day diet with known PPR of 8 mg/g, followed by 5-day washout period and then receive another 2-day diet with PPR of 10 mg/g. The opposite order of diets will be prescribed in group B. The study diets will be prepared and cooked at hospital cafeteria. Dietary compositions of the study diets were analyzed before the start of the study. Primary outcome measures are difference in change-from-baseline intact FGF-23 level between the two dietary interventions. Secondary outcomes include changes in serum phosphate, intact parathyroid hormone and C-terminal FGF-23 level.

Since food additives include readily absorbable inorganic phosphorus, only natural food sources were chosen for study diets. All study food items had the following unique characteristics including: 1. Using locally produced raw materials. 2. Meeting healthy and safety requirements. 3. Complying with national quality standards. Prior to enrollment of the eligible patients, the study diets were prepared and cooked with the food hygiene practice using Hazard Analysis and Critical Control Points (HACCP) system at hospital cafeteria and dietary composition of study diets were analyzed for chemical analysis. With reference to Association of Official Analytical Communities (AOAC) Official Method 984.27, phosphorus, and calcium were determined by inductively coupled plasma-optical emission spectrometer (ICP-OES) analysis with the detection limit of 0.1 mg/L. In brief, the sample weight were obtained, the edible portions of samples were ashed at high temperature, digested in nitric acid, and used inductively coupled plasma to determine their actual contents of phosphorus and calcium. With reference to Taiwanese official methods, study diets were measured for analyses of protein, fat, saturated fat, sugar, moisture, and ash. Carbohydrates were calculated by the formula: 100 - (Protein + Fat + Moisture + Ash) (g/100 g). Calories were calculated by the formula: Protein (g) x 4 kcal + Fat (g) x 9 kcal + Carbohydrate (g) x 4 kcal.

Based on the measured values of food items, dietitian had crafted low-phosphate diets. Less than 800 mg per day of phosphate amount is designed to fulfill the current clinical recommendation. Two different contents of phosphate diets were prepared to find out the optimal amount of dietary phosphate. Each of the diets was designed to have similar calcium, protein and total caloric contents but only differ in phosphate contents. To enhance nutrition, and to reduce phosphate amount and bioavailability, study diets were designed to fulfill the following criteria including high protein diet (≧1.2 g/kg/day), adequate calories (≧ 30 kcal/kg/day), low phosphate-to-protein ratio (< 10 mg/g), and higher percentage of plant source of phosphate than that of animal source. In addition, meats were sliced and boiled for 30 minutes before cooking to reduce the amount of phosphate while preserving protein content.

Inclusion Criteria:

• Subjects with end-stage kidney disease undergoing thrice-weekly hemodialysis for more than three months, the most recent serum phosphate level greater than 5.5 mg/dL or between 3.5 and 5.5 mg/dL with regular phosphate binder use, and having adequate dialysis (urea reduction ratio equal to or greater than 65%)

Exclusion Criteria:

• Subjects with serum albumin less than 2.5 g/dL, having psychiatric disorders, mental retardation or poor adherence