Studies published between 1980 through the end of 1996. Medline and bibliographies found in published reviews were used to locate studies.

1. Studies evaluating the effect of protein intake on the rate of decline in renal function over at least 6 months.

2. Controlled studies with either RCT, prospective study (not RCT), or evaluation period preceding the study period.

3. Data on rate of decline in renal function reported for both control and study groups.

4. Random effects and fixed effects models were used for data analysis.

Recruitment

Studies published from 1980 through the end of 1996 using Medline and bibliographies found in published reviews.

Design

Meta-analysis.

Blinding used (if applicable)

Not applicable.

Intervention (if applicable)

Studies were reviewed independently by two of the authors. Study quality indicators were examined: (1) publication as a peer-reviewed report, (2) random allocation to treatment and control group (3) use of parallel control group design, (4) assessment of dietary protein intake measuring urea nitrogen accrual, and (5) analysis of results by intention to treat.

24 controlled (randomized and nonrandomized) trials:

13 RCT: unweighted mean dietary protein content was .68 +/- .11 g/kg/d in treatment groups (.66 +/- .10 g/kg/d in nine studies that included mostly nondiabetic patients and .73 +/- .15 g/kg/d in the four studies that included mostly diabetic patients) and 1.01 +/- .32 g/kg/d in the control groups.

11 nonrandomized trials: mean dietary protein content was .61 +/- .14 g/kg/d in the treatment groups and 1.02 +/- .28 g/kg/d in the control groups.

Statistical Analysis

Weighted regression analysis was used to determine the reasons for the differences in the results of the 13 randomized trials along with the 11 other nonrandomized controlled trials.

Timing of Measurements

Studies published from 1980 through the end of 1996 using Medline and bibliographies found in published reviews.

Dependent Variables

• The rate of decline in renal function using the slope of serial GFR
• Creatinine clearance
• Urea nitrogen.

Independent Variables

• Randomized Trial:  Dietary protein intake .68 +/- .11 g/kg/d in treatment groups (.66 +/- .10 g/kg/d in nine studies that included mostly nondiabetic patients and .73 +/- .15 g/kg/d in the four studies that included mostly diabetic patients) and 1.01 +/- .32 g/kg/d in the control groups.
• Nonrandomized Trial:  Dietary protein intake .61 +/- .14 g/kg/d in the treatment groups and 1.02 +/- .28 g/kg/d in the control groups.

Control Variables

Initial N: 13 randomized controlled trials included 1,919 patients (range, 14 to 840 patients/study).  The other 11 nonrandomized controlled trials included 2,248 patients (range 12 to 840 patients/study).

Attrition (final N): See above.

Age: Not mentioned

Ethnicity: Not mentioned

Other relevant demographics: Not mentioned

Anthropometrics:

Location:  Worldwide studies

The pooled results demonstrated that protein restriction reduced the rate of decline in renal function by 0.53 mL/min/yr (95% CI, 0.08 to 0.98 mL/min/yr).

It appeared that the magnitude and variability of the treatment effects were inversely proportional to the size of the studies. In a funnel plot of treatment effects vs. study sample size, the smaller studies tended to more often report positive results, indicating possible publication bias in favor of low-protein diets.

With removal of 1 large study with a small difference in protein intake between treatment and control groups, the results of the remaining 12 studies were: Protein restriction reduced the rate of decline in renal function by 0.66 mL/min/yr (95% CI, 0.18 to 1.14 mL/min/yr). The mean protein intake was 0.61+0.14 g/kg/d in the treatment group and 1.02+0.28 g/kg/d in the control group, a difference of 0.41+0.26 g/kg/d.

The effect of dietary protein restriction (GFR decline in treatment – control) was substantially less in RCT vs. nonrandomized (regression coefficient, -5.2 mL/min/yr; (95% CI, -7.8 to –2.5 mL/min/yr; P<0.05) and relatively greater in those with diabetes vs. nondiabetics: 5.4 mL/min/yr; 95% CI, 0.3to 10.5 mL/min/yr; P<0.05), while there was a trend toward greater effects with each additional year of follow-up and the duration of follow-up was short in most trials.

Multiple linear regression coefficients and 95% CIs for independent effects of variables on the difference in the rate of GFR decline between low-protein and control diets. A positive coefficient indicates that the variable favored a low-protein compared with a control diet.

The results of this meta-analysis suggest that reduced dietary protein intake retards the rate of decline in GFR among patients with renal disease. However, the magnitude of the effect of dietary protein restriction was relatively small.

The present analysis indicated that the effects of a low-protein diet may be greater in patients with diabetes. However the number of diabetic patients studied was small.

Analysis suggests that better therapies are needed to retard the rate of GFR decline in patients with renal disease.