Dietary Plant and Animal Protein Intake and Decline in Estimated Glomerular Filtration Rate Among Elderly Women

A 10-Year Longitudinal Cohort Study

Amélie Bernier-Jean; Richard L. Prince; Joshua R. Lewis; Jonathan C. Craig; Jonathan M. Hodgson; Wai H. Lim; Armando Teixeira-Pinto; Germaine Wong

Disclosures

Nephrol Dial Transplant. 2021;36(9):1640-1647. 

In This Article

Results

Of the 1460 participating women randomized in the original study, 1445 completed the baseline FFQ, and 1381 had at least one measurement of their kidney function. We excluded three participants for implausible energy intake and a further four participants for having a protein intake further than 3.5 SDs from the mean on all FFQs. In total, 1374 women contributed to the analysis (Figure 1).

Figure 1.

Flow diagram of the study population. 1A total of 977 were excluded for taking medications that could affect bone mass, 199 for being unlikely to survive the duration of the study, 44 for participating in another clinical trial and 2 for refusing to be assigned to the placebo.

Table 1 summarizes the characteristics of the study population at baseline. The average eGFR was 65.6 (SD = 13.1) mL/min/1.73 m2, and 367 (33%) participants had an eGFR <60 mL/min/1.73 m2. At 10 years, the average eGFR was 60.2 (SD = 15.8) mL/min/1.73 m2 and 343 (46%) participants had an eGFR <60 mL/min/1.73 m2. The average decline in eGFR was 0.64 mL/min/1.73 m2/year (95% CI 0.56–0.72).

The mean (SD) total protein intake throughout the study was 75.2 (23.6) g/day, or 1.15 (0.41) g/kg of weight/day. At baseline, 248 (18%) participants consumed less than the57 g of protein/day recommended for women >70 years old with a normal kidney function in Australia,[22] while 507 (34%) had a high protein intake (>1.3 g/kg/day). The mean (SD) animal and plant protein intake for the three assessments were, respectively, 51.4 (22.9) g/day [0.75 (0.31) g/kg/day] and 28.9 (9.5) g/day [0.40 (0.15) g/kg/day]. On average, 63% (SD = 11) of the plant protein intake came from grains, 27% (SD = 9) from fruits and vegetables, 3% (SD = 7) from legumes and 3% (SD = 4) from nuts. As expected, higher plant-sourced protein intake was associated with higher overall plant consumption (Supplementary data, Figure S1). Supplementary data, Table S3a summarizes the characteristics of the study population at baseline stratified by quartile of plant protein intake. On average 39% (SD = 16) of the animal protein intake came from dairy products, 31% (SD = 15) from meat, 15% (SD = 11) from fish, 9% (SD = 7) from poultry and 5% (SD = 4) from eggs. Supplementary data, Table S3b summarizes the characteristics of the study population at baseline stratified by quartile of animal protein intake. Despite grouped evidence of participants with higher consumption of animal protein to also have high consumption of protein from plant sources, after adjustment for total energy intake, those consuming more plant-derived protein had a reduced intake of animal protein (Pearson correlation coefficient = −0.41, Supplementary data, Figure S2).

Association Between Plant Protein Intake and Age-related Decline in eGFR

In the unadjusted analysis, increased plant protein intake was strongly associated with a slower decline in eGFR (P = 0.01 for interaction with time). For each 10 g increase in plant protein intake, the yearly decline in eGFR was 0.13 mL/min/1.73 m2/year (95% CI 0.03–0.23) less. Adjusting for the adverse effects of age at baseline did not materially affect the relationship between plant protein intake and the slope of decline in eGFR (P = 0.03 for interaction with time). For each increase of 10 g of protein from plants, the yearly decline in eGFR was 0.12 mL/min/1.73 m2/year (95% CI 0.01–0.23) less. Multi-variable analysis adjustment for age, BMI, diabetes status, physical activity, usage of antihypertensive drugs, history of CHD or ischemic cerebrovascular disease and total energy intake did not substantially alter the relation between a higher intake of plant protein and a slower decline in eGFR (P = 0.03 for interaction with time) (Table 2). For each increase of 10 g in plant protein intake, the eGFR decline was slowed down by 0.12 mL/min/1.73 m2/year (95% CI 0.01–0.23) (Figure 2). Furthermore, the association between plant protein and the rate of eGFR decline was not modified by the presence of CKD, diabetes or HTN at baseline (P for three-way interaction = 0.81, 0.50 and 0.24, respectively) (Figure 3). Likewise, the estimate for plant protein on the slope of eGFR was similar in participants who survived until the end of the 10-year follow-up versus those who did not (P for three-way interaction = 0.69). As expected, older age, increasing BMI, usage of anti-hypertensive drugs, diabetes, CHD and ischemic cerebrovascular disease were associated with a lower eGFR. Interestingly, physical activity was associated with a higher eGFR (Table 2).

Figure 2.

Predicted decline in eGFR for a 10 g difference in plant and animal protein intake. (a) Predicted decline in eGFR for 30 g versus 40 g/day of plant protein intake; (b) predicted decline in eGFR for 50 g versus 60 g/day of animal protein intake. Covariables were set at: age at baseline = 75 years old, BMI = 26.9 kg/m2, total energy intake = 6763 kJ/day, no diabetes, CHD or cerebrovascular accident and not taking anti-hypertensive medication; and for (a) animal protein intake = 48.9 g/day and (b) plant protein intake = 26.2 g/day.

Figure 3.

Associated change in the yearly rate of decline in eGFR/year for subgroups.

An analysis of the food sources of plant protein found a slower average decline in eGFR with higher protein intakes from fruits, vegetables and nuts (P for interaction with time <0.001, 0.02 and 0.03, respectively). However, protein from grain foods, legumes and beans were not related to eGFR change (Supplementary data, Table S4).

Association Between Animal Protein Intake and Age-related Decline in eGFR

In neither the unadjusted analysis nor the age-adjusted analysis was animal protein intake significantly associated with the rate of decline in eGFR (P = 0.30 for interaction with time). Furthermore, we found no association between animal protein intake and the rate of decline in eGFR after adjusting for multiple potential confounders (change in the rate of decline in eGFR for each 10 g of animal protein intake increase was 0.01 mL/min/1.73 m2/year (95% CI −0.04 to 0.05; P = 0.84). Restricting the analysis to nondairy animal protein did not alter the results (−0.01 mL/min/1.73 m2/year, 95% CI −0.07 to 0.04; P = 0.68).

In a subgroup analysis, animal protein intake appeared to be associated with a more rapid decline in eGFR among participants with a baseline eGFR <60 mL/min/1.73 m2 (Figure 3). However, the three-way interaction term for this subgroup did not reach statistical significance (P = 0.051). The association between animal protein and the rate of eGFR decline was not modified by the presence of diabetes or HTN at baseline or by whether the participant survived or died before the completion of the follow-up (P for three-way interaction = 0.89, 0.46 and 0.84, respectively) (Figure 3).

An analysis of food sources of protein from each food group using tertile of intake did not identify any association of protein from meat, processed meat, poultry, fish and dairy with eGFR decline. However, higher intake of protein from eggs was associated with a slower decline in eGFR (Supplementary data, Table S5).

processing....