2023 Volume 70 Issue 1 Pages 31-42
Fibroblast growth factor-23 (FGF23) is a phosphaturic hormone secreted by the bone in response to dietary phosphate intake. Since the phosphate content in the diet correlates with the protein content, both plant- and animal-based protein intake can increase the serum FGF23 level. However, a higher percentage of energy from plant protein than from animal protein is associated with a lower serum FGF23 level in patients with chronic kidney disease (CKD) in the United States. Since dietary habits differ between Asian and Western populations, we performed a cross-sectional study to determine the association between the percentages of energy from plant and animal proteins and the serum FGF23 level in Japanese CKD patients. In 107 non-dialysis CKD patients (age: 66 ± 9 years; estimated glomerular filtration rate: 56 ± 21 mL/min/1.73 m2), the percentages of energy from plant and animal proteins were assessed using a food frequency questionnaire based on food groups. Venous blood samples were used to measure the serum FGF23, phosphate, 1,25-dihydroxyvitamin D, and intact parathyroid hormone levels. The percentages of energy from plant and animal proteins showed a negative and positive association, respectively, with the serum FGF23 level. Furthermore, isocaloric substitution modeling showed that replacing animal protein with plant protein was associated with a low serum FGF23 level. Our findings suggest that encouraging diets with high plant protein level may prevent an increase in the serum FGF23 level in Japanese CKD patients.
HYPERPHOSPHATEMIA is a major complication of chronic kidney disease (CKD) [1]. It has been reported that an elevated serum phosphate level causes vascular calcification and increases the risk of cardiovascular morbidity [2]. Hence, managing the serum phosphate level is crucial for improving the prognosis of patients with CKD [3]. However, a dietary phosphate restriction or treatment with phosphate binders are not used in the clinical management of patients with early-stage CKD because hyperphosphatemia is a terminal symptom observed in patients with end-stage kidney disease (ESKD).
One of the earliest signs of CKD is an increased circulating fibroblast growth factor-23 (FGF23) level. A decrease in the number of functional nephrons resulting from aging or CKD requires an increase in phosphate excretion per nephron to maintain the phosphate balance. An increased level of phosphate excretion is achieved by an increased level of FGF23 expression, resulting in an elevated serum FGF23 level with age and as CKD progresses [4]. During the progression of CKD, the serum FGF23 level begins to increase much earlier than the serum phosphate level. The FGF23-induced increase in phosphate excretion per nephron can increase the risk of calcium phosphate microcrystal formation in the tubular fluid and induce renal tubular damage and interstitial fibrosis [5]. Once renal damage reduces the number of functional nephrons, the FGF23 level further increases, triggering a deterioration that accelerates the progression of CKD [6]. Therefore, a dietary phosphate restriction to inhibit the elevation of the serum FGF23 level could be a solution for preventing the progression of CKD.
A recent review suggested that plant nutrients and plant-based diets have beneficial effects in patients with CKD [7]. The phosphate from plant proteins exists mainly in the form of phytic acid, which is mildly digestible in humans; consequently, it is less bioavailable than is the phosphate from animal proteins [8]. Therefore, a phosphate restriction can be achieved by replacing animal proteins with plant proteins. A large population-based cohort study of patients with CKD in the United States demonstrated that a higher percentage of energy from plant proteins than that from animal proteins was associated with a lower serum FGF23 level [9]. However, there have been no reports on the association between the percentages of energy from plant and animal proteins and the serum FGF23 level in the Asian population. This is needed, given the differences in the dietary habits between the Asian and Western populations.
The purpose of this study was to examine the cross-sectional associations between the percentages of energy from plant and animal proteins and the serum phosphate and FGF23 levels in Japanese non-dialysis CKD patients. We also estimated the potential impact of replacing animal proteins with plant proteins on serum phosphate and FGF23 levels using isocaloric substitution (IS) modeling. We hypothesized that (i) a higher percentage of energy from plant protein or a lower percentage of energy from animal protein would be associated with lower serum phosphate and FGF23 levels, and (ii) the replacement of animal protein with plant protein would be associated with low serum phosphate and FGF23 levels.
We used data collected during a community-based physical examination conducted between November 2018 and February 2019 at the University of Tsukuba. The protocol was registered with the UMIN Clinical Trials Registry on November 1, 2018, by Kunihiro Yamagata (title: Age-related changes in phosphate metabolism; registry number: UMIN000034741). All participants were residents recruited via local newspaper advertisements, flyers, and webpages or patients presenting to the Department of Nephrology at the University of Tsukuba Hospital. Among the 333 participants, 122 eligible participants met the following inclusion criteria: 1) age ≥45 years and 2) estimated glomerular filtration rate (eGFR) <60 mL/min/1.73 m2, urinary albumin level ≥30 mg/g creatinine, or both [10]. Participants were excluded if they provided insufficient data from the food frequency questionnaire based on food groups (FFQg) (n = 7). In addition, eight participants who violated the precautions before the physical examination, which are described below, were also excluded: ate breakfast (n = 2), took medicine (n = 4), and smoked (n = 2). The final analysis was conducted on 107 patients with CKD. None of the patients underwent phosphate binder therapy. This study was conducted in accordance with the Declaration of Helsinki and was approved by the ethical committee of the University of Tsukuba (approval no. H30-161). Informed consent was obtained from all of the participants.
ProceduresThe participants were instructed to abstain from large meals and refrain from vigorous exercise 1 day before the measurements. They also abstained from caffeine and alcohol consumption during this period. All measurements were conducted in the morning after overnight fasting for approximately 12 hours (only water was allowed). Upon arrival, anthropometric measurements were performed, and venous blood and spot urine samples were collected. Urine samples were not necessarily the morning’s first void urine because they were collected after arriving at the laboratory. Next, hemodynamic parameters were measured after sitting at rest for at least 20 min. Measurements were obtained in a quiet, temperature-controlled room (24–26°C). Subsequently, self-report questionnaire results were evaluated. Medication use was assessed using electronic health records or self-administered questionnaires. Menopausal status was surveyed using self-administered questionnaires.
Dietary assessmentThe FFQg version 5 was used to assess dietary intake (Kenpakusha Co., Ltd., Tokyo, Japan). The FFQg includes questions about 29 food groups, 10 cooking methods, and portion sizes for each food group and cooking method, allowing the estimation of dietary intake by nutrient and food groups for approximately 1 to 2 months. Each question was accompanied by an illustration describing the food group and cooking method to aid in an accurate assessment. The standard portion size was pre-specified for each food item, and participants were asked to report their usual portion size relative to the standard portion size using four options (not eating, <0.5, standard, or >1.5). The intake frequency was determined as eating times within a week. Daily food intake was calculated by multiplying the intake frequency by the standard portion and relative size of each food item in the FFQg using Excel Eiyokun version 8 (Kenpakusha Co., Ltd., Tokyo, Japan) based on the Standard Tables of Food Composition in Japan (7th Edition). For the present analysis, we also estimated the protein intake from plant and animal sources separately. Sources of animal protein included meat and processed meat, fish and shellfish, eggs, milk, and dairy products; sources of plant protein included foods other than animal foods. The percentages of energy from proteins and fats were calculated using the following formula: daily food intake (g/day) × calories per 1 g (4 kcal for protein and 9 kcal for fat)/total energy (kcal/day) × 100. The percentages of energy from carbohydrates were calculated using the following formula: 100 - percentages of energy from proteins and fats. The dietitian provided support for the answers. The reproducibility and validity of this questionnaire have been confirmed [11].
Biochemical determinationsEach blood sample was placed in a serum separator tube and centrifuged at 3,000 rpm for 15 min at 4°C. The serum creatinine, urinary creatinine, serum phosphate, and urinary phosphate levels were assessed using enzymatic analysis. The serum cystatin C level was measured using a latex agglutination turbidimetric immunoassay. The urinary albumin level was measured using a turbidimetric immunoassay. The urinary phosphate and albumin excretion levels were evaluated as ratios of the urinary creatinine level. The serum FGF23 concentration was determined using a commercial enzyme-linked immunosorbent assay kit (Kainos Laboratories Inc., Tokyo, Japan) [12]. The intact parathyroid hormone (iPTH) level was measured using an electrochemiluminescence immunoassay (Roche Diagnostics K.K., Tokyo, Japan). The 1,25-dihydroxyvitamin D (1,25[OH]2D) level was measured by radioimmunoassay (Immunodiagnostic Systems Holdings Ltd., Boldon, United Kingdom). Two eGFR values were calculated for Japanese patients: one was based on the serum creatinine level (eGFRcr), and the other was based on the cystatin C level (eGFRcys). The eGFR values were calculated using the following formulae: eGFRcr = 194 × serum creatinine–1.094 × age–0.287 (×0.739 if female) and eGFRcys = {104 × serum cystatin C–1.019 × 0.996age (×0.929 if female)} – 8 [13, 14]. To improve the accuracy of these estimations, the averages of the eGFRcr and eGFRcys were used for the analysis [15]. The fractional tubular reabsorption of phosphate (TRP) was calculated using the following formula: 1 – (urinary phosphate level × serum creatinine level/urinary creatinine level × serum phosphate level) [16]. The fractional tubular reabsorption rate of phosphate (%TRP) was calculated using the following formula: 100 × TRP [17]. The ratio of the renal tubular maximum reabsorption rate of phosphate to the glomerular filtration rate (TmP/GFR) was calculated using the following formula if the TRP was 0.86 or less: Tmp/GFR = TRP × serum phosphate level [16]. If the TRP was greater than 0.86, the Tmp/GFR was calculated using the following formula: Tmp/GFR = 0.3 × TRP/{1 – (0.8 × TRP)} × serum phosphate level [16].
Other covariatesAnthropometric measurements were performed, and the body mass index (BMI) was calculated as kg/m2. The brachial systolic blood pressure (SBP) and brachial diastolic blood pressure (DBP) were measured using a noninvasive vascular profiling system (Form PWV/ABI; Colin Medical Technology Corp., Aichi, Japan). Venous blood samples were used to measure the fasting blood glucose, hemoglobin A1c (HbA1c), high-density lipoprotein (HDL), low-density lipoprotein (LDL) cholesterol, and triglyceride levels. Hypertension was diagnosed when one of the following criteria was met: SBP ≥140 mmHg, DBP ≥90 mmHg, or a history of hypertensive medication use [18]. Diabetes was diagnosed when one of the following criteria was met: fasting blood glucose level ≥126 mg/dL, HbA1c level ≥6.5%, or a history of diabetic medication use [19]. Dyslipidemia was diagnosed when one of the following criteria was met: HDL cholesterol level <40 mg/dL, LDL cholesterol level ≥140 mg/dL, triglyceride level ≥150 mg/dL, or a history of dyslipidemia medication use [20]. Total physical activity (METs-min/week) was assessed using the International Physical Activity Questionnaire Short Version [21].
Statistical analysisThe data are expressed as the mean ± standard deviation, median (interquartile range), or frequency count. Statistical significance was set at p < 0.05. The statistical analyses were performed using SPSS Statistics for Windows, version 28.0 (IBM Japan Ltd., Tokyo, Japan).
Differences in the indicators, such as the serum phosphate and FGF23 levels, between the two or four groups stratified according to the median percentages of energy from plant and animal proteins, were assessed using analysis of covariance (ANCOVA). The following covariates were selected based on the relevant literature: age, sex, BMI, eGFR, presence of albuminuria (urinary albumin level ≥30 mg/g creatinine or not), total physical activity, serum 1,25(OH)2D level, serum iPTH level, presence of comorbidities (hypertension, diabetes, or dyslipidemia), total energy, percentage of energy from carbohydrates, and percentage of energy from fat [5, 22, 23]. Two-way ANCOVA was performed to examine the interaction between plant protein intake and animal protein intake, and the eGFR. Each participant was allocated to one of four categories based on the median percentage of energy from plant protein or animal protein and the median value of the eGFR. The same covariates used in the ANCOVA, except for the eGFR and albuminuria, were used in the analysis. Paired post hoc comparisons underwent correction using Bonferroni’s method.
IS modeling was used to estimate the potential impact of the partial replacement of energy from animal protein with that from plant protein on the serum phosphate and FGF23 levels [24]. Any one of the exposure variables (percentage of energy from plant or animal proteins) and the same covariates used in the ANCOVA were entered into the IS model. The percentage of energy from each nutrient was scaled to 3% to enhance the interpretability of the results [23]. The results are provided as unstandardized coefficients (β) for the continuous exposure variables.
Table 1 presents the participants’ characteristics. The average age was 66 ± 9 years, and 54% of the patients were men. The average serum phosphate level and the median serum FGF23 level were 3.4 ± 0.5 mg/dL and 64 [45–86] pg/dL, respectively. The dietary intake patterns of the participants are presented in Table 2. The average percentages of energy from plant and animal proteins were 6.8 ± 1.1% and 7.8 ± 2.6%, respectively. The total energy and percentages of energy from total protein, plant protein, animal protein, fats, and carbohydrates were significantly associated with the dietary phosphate intake (total energy: r = 0.20, p < 0.05; total protein: r = 0.90, p < 0.05; plant protein: r = 0.20, p < 0.05; animal protein: r = 0.78, p < 0.05; fats: r = 0.49, p < 0.05; carbohydrates: r = –0.72, p < 0.05). The dietary phosphate intake was significantly associated with the serum 1,25(OH)2D level, but it was not significantly associated with the serum FGF23 and iPTH levels (1,25[OH]2D: r = 0.21, p < 0.05; FGF23: r = –0.09, p = 0.35; iPTH: r = –0.12, p = 0.23).
| Age, years | 66 ± 9 |
| Men, n (%) | 58 (54) |
| BMI, kg/m2 | 23 ± 4 |
| eGFR, mL/min/1.73 m2 | 56 ± 21 |
| Urinary albumin, mg/g creatinine | 86 [35–455] |
| Serum phosphate, mg/dL | 3.4 ± 0.5 |
| Serum FGF23, pg/dL | 64 [45–86] |
| Serum 1,25(OH)2D, pg/dL | 54 ± 18 |
| Serum iPTH, pg/dL | 45 [35–65] |
| Urinary phosphate, mg/g creatinine | 486 ± 178 |
| %TRP, % | 83 ± 9 |
| Tmp/GFR, mg/dL | 3.0 ± 0.7 |
| Total physical activity, Mets-min/week | 2,061 ± 2,542 |
| GFR stages | |
| eGFR stage 1 (eGFR: ≥90 mL/min/1.73 m2), n (%) | 8 (8) |
| eGFR stage 2 (eGFR: 60–89 mL/min/1.73 m2), n (%) | 29 (27) |
| eGFR stage 3 (eGFR: 30–59 mL/min/1.73 m2), n (%) | 57 (53) |
| eGFR stage 4 (eGFR: 15–29 mL/min/1.73 m2), n (%) | 13 (12) |
| Albuminuria stages | |
| Normoalbuminuria (urinary albumin level: <30 mg/g creatinine), n (%) | 20 (19) |
| Microalbuminuria (urinary albumin level: 30–299 mg/g creatinine), n (%) | 53 (50) |
| Macroalbuminuria (urinary albumin level: ≥300 mg/g creatinine), n (%) | 34 (32) |
| Hypertension, n (%) | 80 (75) |
| Diabetes, n (%) | 23 (22) |
| Dyslipidemia, n (%) | 67 (63) |
| Post-menopause, n (%) | 45 (92) |
Data are presented as the mean ± standard deviation, median [interquartile range], or frequency counts (%), as appropriate. BMI, body mass index; eGFR, estimated glomerular filtration rate; FGF23, fibroblast growth factor-23; 1,25(OH)2D, 1,25-dihydroxyvitamin D; iPTH, intact parathyroid hormone, %TRP: fractional tubular reabsorption rate of phosphate; Tmp/GFR: ratio of the renal tubular maximum reabsorption rate of phosphate to the glomerular filtration rate.
| Total energy, kcal/day | 1,762 ± 394 |
| Plant protein, % of energy | 6.8 ± 1.1 |
| Animal protein, % of energy | 7.8 ± 2.6 |
| Total fat, % of energy | 29 ± 5 |
| Total carbohydrates, % of energy | 57 ± 6 |
| Phosphate, mg/1,000 kcal | 547 ± 91 |
| Calcium, mg/1,000 kcal | 296 ± 82 |
| Kalium, mg/1,000 kcal | 1,301 ± 311 |
| Vitamin B12, μg/1,000 kcal | 3.6 ± 1.7 |
| Vitamin D, μg/1,000 kcal | 3.5 ± 1.7 |
| Salt equivalents, g/1,000 kcal | 4.7 ± 1.2 |
| Cereals, g/1,000 kcal | 202 ± 61 |
| Potatoes and starches, g/1,000 kcal | 20 ± 17 |
| Green and yellow vegetables, g/1,000 kcal | 47 ± 28 |
| White vegetables, g/1,000 kcal | 71 ± 38 |
| Algae, g/1,000 kcal | 2.6 ± 2.1 |
| Pulses, g/1,000 kcal | 41 ± 30 |
| Fish and shellfish, g/1,000 kcal | 37 ± 25 |
| Meat and processed meat, g/1,000 kcal | 38 ± 21 |
| Eggs, g/1,000 kcal | 17 ± 11 |
| Dairy products, g/1,000 kcal | 76 ± 47 |
| Fruits, g/1,000 kcal | 67 ± 45 |
| Confectionaries, g/1,000 kcal | 34 ± 21 |
| Sugar-sweetened beverages, g/1,000 kcal | 67 ± 105 |
| Sugar, g/1,000 kcal | 4.9 ± 3.8 |
| Nuts and seeds, g/1,000 kcal | 2.2 ± 3.1 |
| Fats and oils, g/1,000 kcal | 6.6 ± 3.5 |
| Seasonings and spices, g/1,000 kcal | 12 ± 7 |
Data are presented as the mean ± standard deviation.
Fig. 1 shows the differences in the serum phosphate levels, urinary phosphate levels, serum FGF23 levels, and Tmp/GFR values between the two groups dichotomized according to the median percentages of energy from plant and animal proteins after adjusting for potential covariates. The serum phosphate level and Tmp/GFR ratio tended to be lower in the higher-plant protein group than in the lower-plant protein group (serum phosphate: F = 3.04, p = 0.09; Tmp/GFR: F = 3.13, p = 0.08). The serum FGF23 level was significantly lower in the higher-plant protein group than in the lower-plant protein group (F = 4.72, p < 0.05). Meanwhile, the urinary phosphate level tended to be higher in the higher-plant protein group than in the lower-plant protein group (F = 3.36, p = 0.07). The serum phosphate level tended to be higher in the higher-animal protein group than in the lower-animal protein group (F = 3.01, p = 0.09). The serum FGF23 level and Tmp/GFR were significantly higher in the higher-animal protein group than in the lower-animal protein group (serum FGF23: F = 7.12, p < 0.05; Tmp/GFR: F = 4.12, p < 0.05). Moreover, the urinary phosphate level was significantly lower in the higher-animal protein group than in the lower-animal protein group (F = 5.14, p < 0.05).
Differences in the serum phosphate level (A, B), urinary phosphate level (C, D), serum FGF23 level (E, F), and Tmp/GFR (G, H) between the two groups were dichotomized according to the median percentages of energy from plant (A, C, E, G) or animal proteins (B, D, F, H). Adjusted for age, sex, BMI, eGFR, presence of albuminuria, total physical activity, serum 1,25(OH)2D level, serum iPTH level, presence of hypertension, presence of dyslipidemia, presence of diabetes, total energy, percentage of energy from carbohydrates, and percentage of energy from fat. FGF23, fibroblast growth factor-23; Tmp/GFR, ratio of the renal tubular maximum reabsorption rate of phosphate to the glomerular filtration rate; BMI, body mass index; eGFR, estimated glomerular filtration rate; 1,25(OH)2D, 1,25-dihydroxyvitamin D; iPTH, intact parathyroid hormone. The data are presented as the mean ± standard error.
The combined effects of the plant and animal protein intake on the serum phosphate level, urinary phosphate level, serum FGF23 level, and Tmp/GFR are presented in Fig. 2. The participants were stratified into four groups based on whether the percentages of energy from plant and animal proteins were higher or lower than the median. The serum phosphate levels, urinary phosphate levels, and Tmp/GFR values did not differ significantly among the four groups (serum phosphate: F = 1.56, p = 0.21; urinary phosphate: F = 2.07, p = 0.11; Tmp/GFR: F = 1.91, p = 0.13). In contrast, the serum FGF23 levels differed significantly among the four groups (F = 2.90, p < 0.05). In the paired post hoc comparison, the serum FGF23 level was significantly lower in the group with a lower animal protein intake and higher plant protein intake than in the group with a higher animal protein intake and lower plant protein intake (p < 0.05).
Combined effects of plant and animal protein intake on the serum phosphate level (A), urinary phosphate level (B), serum FGF23 level (C), and Tmp/GFR (D). The participants were stratified into four groups according to whether the percentage of energy from plant or animal protein was higher or lower than the median. Adjusted for age, sex, BMI, eGFR, presence of albuminuria, total physical activity, serum 1,25(OH)2D level, serum iPTH level, presence of hypertension, presence of dyslipidemia, presence of diabetes, total energy, percentage of energy from carbohydrates, and percentage of energy from fat. FGF23, fibroblast growth factor-23; Tmp/GFR: ratio of the renal tubular maximum reabsorption rate of phosphate to the glomerular filtration rate; BMI, body mass index; eGFR, estimated glomerular filtration rate; 1,25(OH)2D, 1,25-dihydroxyvitamin D; iPTH, intact parathyroid hormone. The data are presented as the mean ± standard error. *p < 0.05 vs. the group with a higher animal protein intake and lower plant protein intake.
The joint associations of the plant and animal protein intake and the eGFR (four groups stratified according to each median value) with the serum FGF23 level are presented in Fig. 3. There was a significant interaction between the animal protein intake and eGFR in association with the serum FGF23 level (F = 7.56, p < 0.05). In the post hoc comparisons in the patients with a high eGFR, the serum FGF23 level was significantly higher in the group with a higher animal protein intake than in the group with a lower animal protein intake (p < 0.05). Although there was no statistically significant interaction between the plant protein intake and eGFR in association with the serum FGF23 level (F = 1.69, p = 0.20), the serum FGF23 level was significantly lower in the group with a higher plant protein intake than in the group with a lower plant protein intake in the patients with a high eGFR (p < 0.05).
The joint association between the plant (A) or animal (B) protein intake and eGFR (four groups stratified according to each median value) and the serum FGF23 level. p values were calculated from a two-way analysis of covariance after adjusting for age, sex, BMI, total physical activity, serum 1,25(OH)2D level, serum iPTH level, presence of hypertension, presence of dyslipidemia, presence of diabetes, total energy, percentage of energy from carbohydrates, and percentage of energy from fat. FGF23, fibroblast growth factor-23; BMI, body mass index; eGFR, estimated glomerular filtration rate; 1,25(OH)2D, 1,25-dihydroxyvitamin D; iPTH, intact parathyroid hormone. The data are presented as the mean ± standard error. *p < 0.05.
Table 3 presents the results of the IS model. Replacing 3% of the energy from animal protein with energy from plant protein was not significantly associated with the serum phosphate level (β = –0.180; 95% confidence interval: –0.404, 0.045). However, replacing 3% of the energy from animal protein with energy from plant protein was associated with a low serum FGF23 level (β = –0.101; 95% confidence interval: –0.193, –0.010).
| Percentage of energy from animal protein | Percentage of energy from plant protein | |||
|---|---|---|---|---|
| β | 95% CI | β | 95% CI | |
| Serum phosphate level | ||||
| Replacing 3% of energy from animal protein | Dropped | –0.180 | (–0.404, 0.045) | |
| Replacing 3% of energy from plant protein | 0.180 | (–0.045, 0.404) | Dropped | |
| Log of the serum FGF23 level | ||||
| Replacing 3% of energy from animal protein | Dropped | –0.101 | (–0.193, –0.010)* | |
| Replacing 3% of energy from plant protein | 0.101 | (0.010, 0.193)* | Dropped | |
* p < 0.05. The results provided are unstandardized β coefficients for continuous exposure variables. Adjusted for age, sex, BMI, eGFR, presence of albuminuria, total physical activity, serum 1,25(OH)2D level, serum iPTH level, presence of hypertension, presence of dyslipidemia, presence of diabetes, total energy, percentage of energy from carbohydrates, and percentage of energy from fat. BMI, body mass index; eGFR, estimated glomerular filtration rate; FGF23, fibroblast growth factor-23; 1,25(OH)2D, 1,25-dihydroxyvitamin D; iPTH, intact parathyroid hormone; CI, confidence interval.
The present study examined the cross-sectional association between the percentages of energy from plant and animal proteins and the serum phosphate and FGF23 levels in Japanese patients with non-dialysis CKD. The percentages of energy from plant and animal proteins were independently associated with the serum FGF23 level but not with the serum phosphate level. IS modeling showed that replacing the percentage of energy from animal protein with energy from plant protein was associated with low serum FGF23 level. These findings suggest that plant and animal proteins may have different effects on phosphate metabolism in Asian populations, which have different dietary habits than Western populations.
CKD is a non-communicable disease that is defined as any abnormality in kidney structure or function, regardless of its cause, lasting for 3 months or longer [25]. The prevalence of CKD is increasing in our society, which is aging. Elderly individuals with a low eGFR and patients with CKD with an eGFR of less than 60 mL/min/1.73 m2 account for more than 10% of the total population [26]. In Japan, the number of CKD patients has reached approximately 13.3 million (one in eight adults) [27]. In the present study, a higher percentage of energy from plant protein or a lower percentage of energy from animal protein was associated with a lower serum FGF23 level. These results suggest that, even when not strictly vegetarian, patients with CKD following a diet with a higher percentage of energy derived from plant protein than from animal protein may benefit from a low serum FGF23 level, a potent risk factor for cardiovascular events and CKD progression in patients with non-dialysis CKD.
In plant proteins, phosphate exists predominantly in the form of phytate, which most mammals cannot absorb. Therefore, the amount of phosphate absorbed from plant-based diets should be lower than that from animal-based diets with the same phosphate content. A previous study of nine patients with CKD reported that a 1-week vegetarian diet intervention resulted in reductions in the serum phosphate and FGF23 levels [7]. Since a nearly 100% plant-based diet may not be sustainable, Moorthi et al. conducted an experiment in which 13 patients with CKD followed a 4-week diet in which 70% of protein intake was plant-derived [28]. They concluded that a 70% plant protein diet was safe, well-tolerated, and effective in lowering the urinary phosphate level [28]. In a previous study examining the differences in the metabolic responses to phosphate among different foods, the phosphate in red meat had a higher absorption rate than the phosphate in dairy products and whole grains [29]. Collectively, these results show that limiting the percentage of energy from animal protein, especially that from red meat, and increasing the percentage of energy from plant protein may be effective in controlling the serum phosphate and FGF23 levels.
In the present study, the percentages of energy from plant and animal proteins were not significantly associated with the serum phosphate level. This may be due to tight physiologic regulation of the serum phosphate level by multiple factors, such as the FGF23 level, in our study patients (mean age: 66 ± 9 years; mean eGFR: 56 ± 21 mL/min/1.73 m2) [30]. Hyperphosphatemia is a terminal symptom that is observed only in patients with ESKD [31]. A recent review recommended considering a dietary phosphate restriction or treatment with phosphate binders in patients with CKD with an elevated FGF23 level [5]. However, a dietary phosphate restriction can cause an insufficient intake of essential macronutrients, especially protein [32]. The Hemodialysis Study reported that a dietary phosphate restriction was associated with a poor nutritional status and high mortality in patients undergoing hemodialysis [33]. Another study also suggested that rigorous dietary protein restriction (i.e., phosphate restriction) can lead to a state of malnutrition, termed protein-energy wasting, which is associated with an increased risk of mortality [34]. Although there is no association between the percentages of energy from plant and animal proteins and the serum phosphate level, replacing animal protein with plant protein may be a viable alternative to restrict the intake of dietary phosphate without resulting in a worsening nutritional status (i.e., phosphate restriction without protein restriction).
Increased phosphate excretion is achieved by increasing the FGF23 expression; as the serum FGF23 level increases, the Tmp/GFR value decreases, and the urinary phosphate level increases [35]. However, both the serum FGF23 level and the Tmp/GFR ratio were significantly higher in the higher-animal protein group than in the lower-animal protein group. Moreover, the urinary phosphate level was significantly lower in the higher-animal protein group than in the lower-animal protein group. Because spot urine sample was used in this study, the urinary phosphate level was likely affected by dietary phosphate intake and the TmP/GFR value may not have correctly reflected the actions of FGF23.
The association between the percentages of energy from plant and animal proteins and the serum FGF23 level remained significant after adjusting for potential covariates, including the serum iPTH level. In addition to FGF23, iPTH plays an important role in phosphate metabolism [5]. However, in the present study, the serum iPTH level was not significantly associated with the percentages of energy from plant and animal proteins (data not shown). Nishida et al. reported that iPTH, but not FGF23, can rapidly respond to dietary phosphate loading [17]. The serum level of iPTH is likely to show different responses immediately following the consumption of plant or animal protein.
A prospective cohort study of healthcare professionals in the United States suggested that the median percentages of energy from plant and animal proteins were 4% and 14%, respectively [36]. The Japan Public Health Center-based prospective cohort study suggested that the median percentages of energy from plant and animal proteins were 6.7% and 7.7%, respectively [23]. Moreover, it has been reported that the sources of plant and animal proteins in these two diets may differ because fish and soy product consumption is higher in Japanese populations than in Western populations [23]. Our results suggest that obtaining the required amount of protein from plant-based foods may have health benefits in terms of phosphate metabolism, not only for Westerners but also for Asians, whose diet is different from that of Westerners.
In the IS model, the association between the percentages of energy from plant and animal proteins and the serum FGF23 level remained significant after adjusting for the eGFR. In the two-way ANCOVA, there was a significant interaction between the animal protein intake and eGFR and their association with the serum FGF23 level. In the patients with a high eGFR, the serum FGF23 level was significantly higher in the group with a higher animal protein intake than in the group with a lower animal protein intake. In contrast, when we examined the joint associations of the plant protein intake and eGFR with the serum FGF23 level, no statistically significant interaction was observed. However, in the patients with a high eGFR, the serum FGF23 level was significantly lower in the group with a higher plant protein intake than in the group with a lower plant protein intake. The serum FGF23 level begins increasing at CKD stages 2–3, which is thought to result from an increase in the phosphate excretion per nephron as physiological compensation for the decrease in the functional nephron number [4, 6]. Once the functional nephron number is too low to maintain phosphate balance by further increasing FGF23, phosphate retention and hyperphosphatemia ensue [4, 6]. Based on these reports, as CKD progresses, the dietary modification of replacing animal protein with plant protein may no longer have a sufficient effect on lowering the serum FGF23 level, and additional treatment, such as phosphate binders, may be necessary. As this was a cross-sectional study, further interventional studies are needed to confirm this hypothesis.
Concerns have been raised regarding the nutritional tolerance of plant-based diets in patients with CKD [37]. In several previous studies, patients with CKD on a vegetarian diet, including patients with ESKD, as well as vegetarians with normal renal function, maintained a good nutritional status [38, 39]. In contrast, most studies reporting beneficial effects of plant nutrients and plant-based diets in patients with CKD are mechanistic, basic science studies or observational clinical studies. Thus, the recommendation of plant-based food in routine clinical practice should be carefully applied to patients with CKD by monitoring adverse effects. Vegetarian individuals have been reported to be at risk for deficiencies in vitamin B12 and vitamin D, which are important nutrients for patients with CKD [40, 41]. Although there are many potential benefits of plant-based diets in patients with CKD, a recent review did not recommend or suggest that patients with CKD should be routinely prescribed or educated on plant-based diets [7]. Further research is needed to clarify the advantages and disadvantages of plant-based diets in patients with CKD and to make recommendations for clinical practice.
This study has several limitations. First, it is a cross-sectional study based on a small sample size, and we could not explain a causal relationship between the percentages of energy from plant and animal proteins and the serum FGF23 level. Future studies with long-term plant-based diet interventions in patients with CKD are needed to confirm whether a plant-based diet lowers the FGF23 level. Second, we used a self-administered questionnaire to assess dietary intake. This could potentially reduce the strength of the association between the exposure and outcome variables, owing to regression dilution bias. However, the self-administered questionnaire used to evaluate the dietary intake in the present study (i.e., the FFQg) included questions about the consumption frequency and portion size and the use of food illustrations allowed for an accurate assessment. Furthermore, the percentage of energy from each nutrient was automatically calculated from the participants’ answers. Collectively, the FFQg can be useful for evaluating the percentages of energy from plant and animal proteins. Third, differences in the net phosphorus absorption between plant and animal proteins were not considered in this study. Fecal excretion and 24-hour urine excretion should also be evaluated in the future to resolve this issue.
The present study demonstrated that a lower percentage of energy from animal protein and a higher percentage of energy from plant protein were associated with a lower serum FGF23 level in Japanese patients with CKD. We also demonstrated that replacing animal protein with plant protein is associated with a low serum FGF23 level. These findings suggest that replacing dietary animal protein with plant protein may provide a method to attain phosphate restriction without protein restriction and prevent an increase in the serum FGF23 level.
The authors would like to thank all participants of this study. We also thank our (S. M.’s) laboratory members and Ms. Michiru Hotta at the University of Tsukuba for providing technical assistance. This work was supported in part by a grant-in-aid for scientific research (KAKENHI) from the Ministry of Education, Culture, Sports, Science, and Technology, Japan (19H03995), and the MEXT Leading Initiative for Excellent Young Researchers (grant number JPMXS0320200234). Masaki Yoshioka, Masahiro Matsui, and Shoya Mori were recipients of a grant-in-aid for research fellowships from the Japan Society for the Promotion of Science for Young Scientists (21J10316, 20J20892, and 21J10952). Natsumi Nishitani was a recipient of a grant-in-aid for research fellowships from Japan Science and Technology (JPMJSP2124).
None of the authors have any potential conflicts of interest associated with this research.
