Clinical Nutrition xxx (2015) 1e8
Contents lists available at ScienceDirect
Clinical Nutrition journal homepage: http://www.elsevier.com/locate/clnu
Meta-analyses
Effects of soy protein containing isoflavones in patients with chronic kidney disease: A systematic review and meta-analysis Zhou Jing, Yuan Wei-Jie* Department of Nephrology, Shanghai First People's Hospital, Shanghai Jiao Tong University, Shanghai, People's Republic of China
a r t i c l e i n f o
s u m m a r y
Article history: Received 29 May 2014 Accepted 22 March 2015
Background & aims: Recent studies have demonstrated mixed results on the effects of soy intake in patients with CKD, and this have not been systematically analyzed. We conducted this meta-analysis to identify and evaluate the effects of soy protein intake in patients with CKD. Methods: A comprehensive search of Medline, Embase and the Cochrane Database of Systematic Reviews was performed in December 2013 and updated in April 2014 for any new trials. Randomized trials designed to evaluate the effects of dietary soy in patients with CKD were collected. Weighted mean effect sizes were calculated for net changes using random-effect or fixed-effect model. All statistical analysis were calculated by RevMan software 5.2 available free from the Cochrane Collaboration. Results: 12 studies (280 participants) were included. And we found that dietary soy was associated with significant decrease of serum creatinine, serum phosphorus, CRP (C reactive protein)and proteinuria in the predialysis subgroup. The mean difference was 0.05 mg/dL (95% CI: 0.10, 0.00 mg/dL; P ¼ 0.04) for serum creatinine, 0.13 mg/dL (95% CI: 0.26, 0.01 mg/dL; P ¼ 0.04) for serum phosphorus, 0.98 mg/L (95% CI: 1.25, 0.71 mg/L; P < 0.00001) for CRP, and 0.13 mg/d (95% CI: 0.18, 0.08 mg/d; P < 0.00001) for proteinuria. We did not find any significant change in serum phosphorus, CRP in the dialysis subgroup. Blood urea nitrogen (BUN) was reduced with statistical significance in the soy-treated group compared with control when the predialysis and dialysis subgroup were analyzed as a whole. The pooled estimated effects of change for BUN was 0.37 mg/dL (95% CI: 6.03, 0.11 mg/dL; P ¼ 0.04). No significant change was detected in creatinine clearance, glomerular filtration rate, serum albumin, body weight and body mass index(BMI). Conclusions: Soy protein containing isoflavones intake significantly decreased serum creatinine, serum phosphorus, CRP and proteinura in predialysis patients, while no significant change was found in creatinine clearance and glomerular filtration rate. We also found that soy protein intake could maintain the nutritional status in dialysis patients, though no significant change in CRP, BUN, and serum phosphorus was detected. Future large, long-term RCTs are still needed to clarify the effects of soy protein intake in patients with CKD. © 2015 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
Keywords: Soy protein Chronic kidney disease Meta-analysis Randomized controlled trials
1. Introduction Chronic kidney disease (CKD) has become an important problem affecting human health, and it is associated with increased risk of morbidity, mortality as well as huge healthcare costs [1]. Epidemiological studies indicate that CKD is also an independent
* Corresponding author. Department of Nephrology, Shanghai First People's Hospital, Shanghai Jiao Tong University, 100 HaiNing Road, Shanghai 200080, People's Republic of China. E-mail addresses: [email protected] (Z. Jing), [email protected] (Y. Wei-Jie).
risk factor for cardiovascular events, and the renineangiotensin system (RAS), oxidative stress, inflammation and hemodynamic disorders are reported to be involved in this pathophysiological process [2]. Previous studies have demonstrated the beneficial effects of dietary soy in humans and animal models. Frigolet et al. [3]. concluded that, soy protein is capable of attenuating abnormal expression of renineangiotensin. An in vitro and in vivo study conducted by Jia et al. [4]. found that, genistein, a soy isoflavones, protects against inflammation induced by TNF-a in vascular endothelial cells. And oxidative stress was significantly decreased in postmenopausal women when treated with soy isoflavones [5].
http://dx.doi.org/10.1016/j.clnu.2015.03.012 0261-5614/© 2015 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
Please cite this article in press as: Jing Z, Wei-Jie Y, Effects of soy protein containing isoflavones in patients with chronic kidney disease: A systematic review and meta-analysis, Clinical Nutrition (2015), http://dx.doi.org/10.1016/j.clnu.2015.03.012
2
Z. Jing, Y. Wei-Jie / Clinical Nutrition xxx (2015) 1e8
Moreover, a recent population-based study suggested soy or soy isoflavones intake significantly reduced the risk of postmenopausal breast cancer [6]. These beneficial findings have been applied for the development of preventive strategies for human health and disease. For example, the US Food and Drug Administration suggested that “25 g soy protein per day” may help prevent the risk of coronary heart disease because of reduced serum lipids and lipoproteins [7]. Emerging evidence from animal and human studies demonstrated that dietary soy protein containing isoflavones retard the development and progression of renal dysfunction. Reportedly, dialysis patients in japan where mean soy products consumption is significantly higher than that in the United State tend to live longer than patients in the US [8,9]. Iwasaki et al. [10] found Fisher rats in the soy diet group tend to have longer life than that in the caseintreated group, and the progression of renal dysfunction was obviously delayed when casein protein was replaced with soy protein. As a result, following the publication of several randomized, clinical trials, our attention has been directed to manipulate the dietary protein quality especially by replacing animal protein with soy protein that rich in isoflavones. Ogborn et al. [11] found that soy protein significantly ameliorate aging-related nephropathy in Fisher 344 rat and renal dysfunction in 5/6 nephrectomy mouse models. Furthermore, observational studies in patients with diabetic nephropathy demonstrated substitution of soy protein for animal protein results in decreased glomerular hyperfiltration and less proteinuria, whereas Marion et al. have not achieved the aim of attenuating proteinuria in the soy-treated group compared with control [12,13]. As described above, studies evaluating the effects of dietary soy in the management of CKD have demonstrated mixed findings, and these results were few systematically reviewed. As a result, this meta-analysis was designed to explore the effects of soy protein containing isoflavones intake in patients with CKD. 2. Methods
the effects of soy diets, soy protein or soy isoflavones in patients with CKD (including the dialysis patients) were identified. Only adult patients (over 18 years old) with renal dysfunction or kidney failure were eligible for inclusion. As for the initial screening process, the main exclusion criteria were as follows: participants younger than 18 years old; other outcomes such as bone metabolism, serum lipid levels; pharmacokinetics and basic studies; trials treated with isoflavones-depleted soy protein or isoflavones that is not extracted from soybeans; patients with renal transplantation; reviews and editorials. 2.3. Data extraction and quality assessment Data were collected independently from each included RCT, and any discrepancies were figured out by consensus. For those included trials, data were extracted based on a standardized strategies, including publications, study designs, size of study population, type of diet, characteristics of participants (such as age, gender, and treatment modality, etc.), interventions, the contents of isolated soy protein (ISP) and isoflavones, the main outcomes (such as serum creatinine, 24-h urine protein excretion, blood urea nitrogen, etc.), and duration of follow-up. As for Imani et al. [14]. that did not provide the contents of isoflavones, we calculated it based on the concentration of isoflavones in traditional soy products which is about 3.5 mg isoflavones per 1 g soy protein [15]. Methodological quality and risk of bias of each included RCT were examined carefully using the method described by the Cochrane Collaboration [16]. The items were as follows, 1. Random sequence generation; 2. Allocation concealment; 3. Blinding of participants and personnel; 4. Blinding of outcome assessment; 5. Incomplete outcome data; 6. Selective reporting; 7. Other sources of bias. All seven items were classified as “low risk of bias”, “high risk of bias”, or “unclear risk of bias”. Furthermore, intention-to-treat analysis (ITT) was also used to evaluate integrity of the outcome data.
2.1. Search strategy
2.4. Statistical analyses
A systematic literature search was performed in Medline, Embase and the Cochrane Database of Systematic Reviews for clinical trials evaluating the effects of soy diets, soy protein, or soy isoflavones in patients with CKD in December 2013 and updated in April 2014 for any new trials. The search strategies used were as follows, [(soy OR soybean OR soya OR soy protein OR isoflavones OR isoflavone OR genistein OR daidzein)] AND [(chronic kidney disease OR kidney failure OR chronic kidney failure OR kidney disease OR uremia OR dialysis OR continuous ambulatory peritoneal dialysis OR hemodialysis OR renal replacement therapy OR peritoneal dialysis OR Equilibrium dialysis OR extended daily dialysis)]. Only randomized controlled trials (RCTs) were identified for inclusion without any language limitation in our study. Title and abstract of each article was screened carefully for inclusion at the first step of retrieval process. Then all potentially relevant articles were examined carefully in full text for further identification. Reviews and meta-analyses on effects of soy diets or soy protein in patients with CKD were also checked seriously for any potential trials. Meanwhile, the cited references of each included studies were scanned carefully for additional trials or reviews. Trials were analyzed as a whole once it had more than one publication.
We grouped studies based on the mode of treatment as described above into the non-dialysis group and the dialysis group. Results were also calculated separately for subgroup analysis. Data from each included trials were analyzed using Review Manager (RevMan, Version 5.2, Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2012). Treatment effects were presented as the mean differences between changes and its 95% confidence intervals (CI), and the pooled effects were computed by assigning each trial a weight of the reciprocal of its variance. And if the raw data were unavailable, the variances for the changes of individual trials were calculated according to the methods described by the Cochrane Collaboration. Heterogeneity of treatment effects was evaluated using the x2 test, and the I2 test for inconsistency. The degree of inconsistency among those identified trials was defined by calculating the percentage of total betweenstudy variation because of heterogeneity rather than random variation as the I2 metric using the formula I2 ¼ 100% (Q-df)/Q [17]. We described the results using random-effect model when heterogeneity across those trials was significant, and on the contrary, a fixed-effect model was applied [18]. Though the outcomes evaluated in this meta-analysis have limited number of trials, the funnel plots calculated using the Review Manager 5.2 were used to assess the presence or absence of publication bias for certain outcomes. The P-value threshold for statistical significance was set at 0.05 for effect size, P 0.05 was considered significant.
2.2. Selection criteria Randomized, controlled trials with a crossover or parallel design that were of at least 7 weeks follow-up and designed to evaluate
Please cite this article in press as: Jing Z, Wei-Jie Y, Effects of soy protein containing isoflavones in patients with chronic kidney disease: A systematic review and meta-analysis, Clinical Nutrition (2015), http://dx.doi.org/10.1016/j.clnu.2015.03.012
Z. Jing, Y. Wei-Jie / Clinical Nutrition xxx (2015) 1e8
3. Results 3.1. Trial flow/flow of included studies 655 trials were identified by searching Medline, Embase and the Cochrane Database of Systematic Reviews, 594 of which were excluded at the initial screening. 61 potentially relevant trials were identified for further review, of which 12 trials fulfilled our inclusion criteria. All selected studies were randomized, controlled trials designed to evaluate the effects of soy protein containing isoflavones in patients with CKD. Detailed processes of our study selection are shown in Fig. 1.
3
low-protein diet while the other 6 studies adopted the usual diet, and Anderson et al. [12] was given a standard diabetes diet. Furthermore, 4 trials enrolled participants with diabetes mellitus, and five trials enrolled patients with hypertension. Besides, the exact contents of isolated soy protein (ISP) and isoflavones were presented clear in 8 trials, whereas the Teixeira et al. [20] and the Ahmed et al. [13]. applied soy protein containing isoflavones, and the exact content is unclear. The rest trials, the Anderson et al. [12] and Soroka et al. [19] did not described soy protein as isoflavonedepleted soy was thought to be treated with soy protein containing isoflavones. 3.3. Quality of included studies
3.2. Study characteristics 12 randomized trials (280 participants) designed to evaluate the effects of soy or soy products intake in patients with CKD were identified. Characteristics of the 12 trials were listed in Table 1, and Chen et al. [21] trial was analyzed as two subgroups and calculated separately based on the serum lipid levels of the participants. Based on the treatment modality, we divided these trials into two subgroups: the non-dialysis subgroup and the dialysis subgroup. Half of the included trials compared isolated soy protein containing isoflavones with milk or whey protein, and the rest employed soy diets or soy products such as soy beverage, textured soy protein into the treatment group. The contents of isolated soy protein consumption in our meta-analysis ranged from 14 to 30.7 g/d, and the isoflavones ranged from 35.5 to 61 mg/d. 8 of those included trials adopted a parallel design while the rest with a crossover design. As for the duration of follow-up, it varied from 7 weeks to 4 years across these included studies. And 5 of the studies adopted a
Based on the standard methods recommended by the Cochrane Collaboration for assessing risk of bias, trial quality was various across the included trials. Allocation concealment was adopted in 4 trials, the Anderson et al. [12], the Chen et al. [21], the Tabibi et al. [27], the Teixeira et al. [20]. And it is unclear in the rest 8 trials. Only the Ahmed et al. [13] trial was analyzed based on the intent-to-treat analysis (ITT). The result was shown in Fig. S1, 2. 3.4. Quantitative data synthesis 3.4.1. Serum creatinine A total of 7 trials were identified to evaluate the effects of soy consumption on serum creatinine, only participants in the Siefker et al. [24] trial which was not included in this analysis is on dialysis. As for the non-dialysis participants, serum creatinine was decreased with statistical significance in the soy-treated group compared with control, and the pooled estimated change was
Fig. 1. Flowchart demonstrating the literature screening process and reasons for exclusion are explained.
Please cite this article in press as: Jing Z, Wei-Jie Y, Effects of soy protein containing isoflavones in patients with chronic kidney disease: A systematic review and meta-analysis, Clinical Nutrition (2015), http://dx.doi.org/10.1016/j.clnu.2015.03.012
%
64 ± 18.38 57.5 63 63.6 ± 9.7 59.5 ± 12 53.6 ± 11.6 61.0 ± 2.9 50.3 62.1 ± 12.1 62.5 ± 12.1 52 ± 15 52.3 ± 14.9 43.9 ± 12.2 100 100 100 NAc NAc NAc 36 NAc 100 100 NAc NAc NAc Standard diabetes diet Low-protein dietf Low-protein dietf Usual hemodialysis diet Usual hemodialysis diet Usual hemodialysis diet Usual diet Usual diet Low-protein dietf Low-protein dietf Usual diet Usual diet Low-protein dietf
%
3.4.2. Creatinine clearance and glomerular filtration rate (GFR) 6 trials were identified to evaluate the effects of soy protein intake on creatinine clearance (Fig. 3). We did not find any significant differences in the change of creatinine clearance in the soytreated group compared with control. The pooled effects of soy protein consumption on creatinine clearance was 1.14 ml/min/ 1.73 m2 (95% CI: 3.70, 1.41 ml/min/1.73 m2; P ¼ 0.38), and no heterogeneity was detected (heterogeneity X2 ¼ 6.67, P ¼ 0.25, I2 ¼ 25%) (Fig. 3). Visual inspection of the funnel plot is asymmetric, suggesting the presence of publication bias (Fig. 4). Only 2 trials were included for the analysis of GFR which were determined by [99mTc]diethylenetriamine pentaacetic acid (DTPA) nuclear medicine scanning and 51Cr-EDTA separately. No significant change was found in the soy-treated group compared with control. The pooled effects of soy protein consumption on GFR was 1.30 ml/min/1.73 m2 (95% CI: 4.43, 1.83 ml/min/1.73 m2; P ¼ 0.41). The fixed-effect model was used as no heterogeneity was found (heterogeneity X2 ¼ 0.08, P ¼ 0.78, I2 ¼ 0%) (Table 2).
NAc 8 weeks NAc NAc NAc 6 months NAc NAc 7 months 30 35.5 12 weeks 30 35.5 12 weeks 30.7 36.3 12 weeks 25 54 8 weeks 25 52 13 months 15.5 ± 8.9 45.5 ± 15.7 4 years 17.8 ± 3.3 42.3 ± 18.2 7 weeks 14 49 8 weeks 14 61 8 weeks c c NA NA 8 weeks
RCd RCd RCd RPe RPe RPe RPe RPe RPe RCd RPe RPe RPe
Design Type of diet
3.4.3. Serum albumin A total of 9 trials (including a subgroup) reported data demonstrating the effects of soy intake on serum albumin (Fig. 5). Soy intake did not significantly affect the level of serum albumin, and the pooled estimated change was 0.02 g/dL (95% CI: 0.10, 0.06 g/ dL; P ¼ 0.71). There was no heterogeneity across these trials (heterogeneity X2 ¼ 3.18, P ¼ 0.92, I2 ¼ 0%). And the funnel-plot analysis did not show asymmetry demonstrating that the publication bias cannot substantially affect the result of this meta-analysis (Fig. 6). As for the non-dialysis patients, 4 trials were identified. The fixed-effect model was applied as no heterogeneity was detected (heterogeneity X2 ¼ 0.25, P ¼ 0.97, I2 ¼ 0%). No statistically significant difference was detected in serum albumin in the soy-treated group compared with control, and the pooled effects of the changes was 0.0 g/dL (95% CI: 0.09, 0.09 g/dL; P ¼ 0.97). 5 trials enrolling patients on dialysis were included for analysis of serum albumin. No heterogeneity was found in the dialysis subgroup (heterogeneity X2 ¼ 2.34, P ¼ 0.67, I2 ¼ 0%), and the fixedeffect model was used. We did not find any significant change in serum albumin in dialysis patients in the soy protein treated group compared with control, and the pooled effects of the changes was 0.07 g/dL (95% CI: 0.24, 0.09 g/dL; P ¼ 0.39).
Soy protein(þ)/animal protein Soy diet(þ)/animal protein diet ISPb(þ)/casein ISPb(þ)/milk protein ISPb(þ)/milk protein ISPb(þ)/milk protein ISPb(þ)/milk protein ISPb(þ)/whey protein Soy diet(þ)/animal protein Soy protein(þ)/animal protein Soy diet(þ)/control dietg Soy diet(þ)/control dietg Soy protein(þ)/animal protein
g/day
Duration Isoflavones
mg/day
ISPb
f
g
e
c
d
M/F, ratio of male and female. ISP, isolated soy protein. NA, not available. RC, randomized crossover design. RP, randomized parallel design. Low-protein diet, 0.8 g protein/kg/d. Control diet, usual diet without any soy. a
b
100/0 55.6/44.4 100/0 26.3/73.7 27.8/72.2 73/27 40/60 41.2/58.8 43.9/56.1 71.5/28.5 50/50 50/50 15.8/85.2 8 9 14 19 18 26 25 16 41 14 36 36 18 Anderson (1998) [12] Soroka (1998) [19] Teixeira (2004) [20] Chen 1 (2005) [21] Chen 2 (2005) [21] Chen (2006) [22] Fanti (2006) [23] Siefker (2006) [24] Azadbakht (2008) [25] Azadbakht (2009) [26] Imani (2009) [14] Tabibi (2010) [27] Ahmed (2011) [13]
N
%
Predialysis Predialysis Predialysis Hemodialysis Hemodialysis Hemodialysis Hemodialysis Hemodialysis Predialysis Predialysis Peritoneal dialysis Peritoneal dialysis Predialysis
Treatment modality Randomized intervention No. of subjects M/Fa Study
Table 1 Characteristics of 12 randomized controlled trials selected for analysis.
0.05 mg/dL (95% CI: 0.10, 0.00 mg/dL; P ¼ 0.04). The fixedeffect model was used as no heterogeneity was detected (heterogeneity X2 ¼ 4.29, P ¼ 0.37, I2 ¼ 7%) (Fig. 2). Because the Anderson et al. [12] trial that is not estimable was excluded, 5 trials were insufficient to detect publication bias.
100 0 100 0 0 0 44 NAc 100 100 30.6 NAc NAc
y
Z. Jing, Y. Wei-Jie / Clinical Nutrition xxx (2015) 1e8
Diabetes Hypertension Age
4
3.4.4. Serum phosphorus 6 trials were included in this section (Fig. 7). No significant change was demonstrated in the soy intake group, and the pooled estimated mean difference was 0.10 mg/dL (95% CI: 0.22, 0.02, P ¼ 0.10). Heterogeneity in this group of studies was not significant (heterogeneity X2 ¼ 9.83, P ¼ 0.08, I2 ¼ 49.0%). The funnel plot is symmetric demonstrating that publication bias cannot substantially affect the result of this meta-analysis (Fig. 8). As for the single outliers (Siefker et al. trial [24]), the results were not changed significantly after this study is excluded. The pooled effect of changes was 0.14 mg/dL (95% CI: 0.26, 0.01, P ¼ 0.03), and no heterogeneity was detected (heterogeneity X2 ¼ 4.20, P ¼ 0.38, I2 ¼ 5.0%) (Fig. S3). As for the 4 trials in the non-dialysis subgroup, Heterogeneity was not significant (heterogeneity X2 ¼ 4.04, P ¼ 0.26, I2 ¼ 26%). soy consumption significantly decreased serum phosphorus, and the pooled estimated mean difference was 0.13 mg/dL (95%
Please cite this article in press as: Jing Z, Wei-Jie Y, Effects of soy protein containing isoflavones in patients with chronic kidney disease: A systematic review and meta-analysis, Clinical Nutrition (2015), http://dx.doi.org/10.1016/j.clnu.2015.03.012
Z. Jing, Y. Wei-Jie / Clinical Nutrition xxx (2015) 1e8
5
Fig. 2. Net change (and 95% CI) of serum creatinine associated with soy protein containing isoflavones intake in non-dialysis patients.
Fig. 3. Pooled estimated effect of soy protein containing isoflavones on creatinine clearance in non-dialysis patients.
3.4.5. CRP This outcome was reported only in 3 trials. One in the nondialysis subgroup, and the other in the dialysis subgroup. CRP levels was significantly reduced in the soy-treated group compared with control, and the pooled estimated change in CRP level was 0.98 mg/L (95% CI: 1.25, 0.71 mg/L; P < 0.00001). This analysis was dominated by Azadbakht et al. [25] study, contributing 99.8% of the weight. No significant change was detected in the dialysis patients when treated with soy products, and the pooled estimated change was 0.83 mg/L (95% CI: 6.56, 4.91 mg/L; P ¼ 0.78) using the fixed-effects model (Table 2). 3 studies were inadequate to detect publication bias.
Fig. 4. Funnel plot with 95% CI to evaluate publication bias on creatine clearance in the predialysis participants.
CI: 0.26, 0.01 mg/dL; P ¼ 0.04). No significant change was demonstrated in the dialysis subgroup, and the pooled estimated change in serum phosphorus levels was 0.26 mg/dL (95% CI: 0.16, 0.68 mg/dL; P ¼ 0.22).
3.4.6. Blood urea nitrogen (BUN) 5 trials were identified. Blood urea nitrogen was significantly decreased in the soy-treated group compared with corresponding control group. Heterogeneity was demonstrated significant across these trials (heterogeneity X2 ¼ 78.08, P < 0.00001, I2 ¼ 95%), and the random-effect model was applied. The overall pooled estimate of the effect of soy protein was 3.07 mg/dl (95% CI: 6.03, 0.11 mg/dL; P ¼ 0.04). 5 trials were insufficient to detect publication bias. 4 trials were included in the non-dialysis group. A randomeffect model was used as a result of significant heterogeneity (heterogeneity X2 ¼ 75.92, P < 0.00001, I2 ¼ 96%). No significant
Table 2 Summary of the effects of Soy protein intake in patients with CKD compared with control. Items
No. of studies
No. of populations (treatment/control)
Test for I2
Heterogeneity P
Analysis model
Test for Z
Overall effect P
Mean difference, 95% CI
CRP Non-dialysis subgroup Dialysis subgroup BUN Non-dialysis subgroup Dialysis subgroup 24 h Urine protein GFR Body weight Body mass index
3 1 2 5 4 1 4 2 3 2
42/40 20/21 22/19 58/61 51/52 7/9 52/53 17/17 43/44 23/23
4% e 52% 95% 96% e 85% 0% 56% 89%
0.35 e 0.15
Contents lists available at ScienceDirect
Clinical Nutrition journal homepage: http://www.elsevier.com/locate/clnu
Meta-analyses
Effects of soy protein containing isoflavones in patients with chronic kidney disease: A systematic review and meta-analysis Zhou Jing, Yuan Wei-Jie* Department of Nephrology, Shanghai First People's Hospital, Shanghai Jiao Tong University, Shanghai, People's Republic of China
a r t i c l e i n f o
s u m m a r y
Article history: Received 29 May 2014 Accepted 22 March 2015
Background & aims: Recent studies have demonstrated mixed results on the effects of soy intake in patients with CKD, and this have not been systematically analyzed. We conducted this meta-analysis to identify and evaluate the effects of soy protein intake in patients with CKD. Methods: A comprehensive search of Medline, Embase and the Cochrane Database of Systematic Reviews was performed in December 2013 and updated in April 2014 for any new trials. Randomized trials designed to evaluate the effects of dietary soy in patients with CKD were collected. Weighted mean effect sizes were calculated for net changes using random-effect or fixed-effect model. All statistical analysis were calculated by RevMan software 5.2 available free from the Cochrane Collaboration. Results: 12 studies (280 participants) were included. And we found that dietary soy was associated with significant decrease of serum creatinine, serum phosphorus, CRP (C reactive protein)and proteinuria in the predialysis subgroup. The mean difference was 0.05 mg/dL (95% CI: 0.10, 0.00 mg/dL; P ¼ 0.04) for serum creatinine, 0.13 mg/dL (95% CI: 0.26, 0.01 mg/dL; P ¼ 0.04) for serum phosphorus, 0.98 mg/L (95% CI: 1.25, 0.71 mg/L; P < 0.00001) for CRP, and 0.13 mg/d (95% CI: 0.18, 0.08 mg/d; P < 0.00001) for proteinuria. We did not find any significant change in serum phosphorus, CRP in the dialysis subgroup. Blood urea nitrogen (BUN) was reduced with statistical significance in the soy-treated group compared with control when the predialysis and dialysis subgroup were analyzed as a whole. The pooled estimated effects of change for BUN was 0.37 mg/dL (95% CI: 6.03, 0.11 mg/dL; P ¼ 0.04). No significant change was detected in creatinine clearance, glomerular filtration rate, serum albumin, body weight and body mass index(BMI). Conclusions: Soy protein containing isoflavones intake significantly decreased serum creatinine, serum phosphorus, CRP and proteinura in predialysis patients, while no significant change was found in creatinine clearance and glomerular filtration rate. We also found that soy protein intake could maintain the nutritional status in dialysis patients, though no significant change in CRP, BUN, and serum phosphorus was detected. Future large, long-term RCTs are still needed to clarify the effects of soy protein intake in patients with CKD. © 2015 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
Keywords: Soy protein Chronic kidney disease Meta-analysis Randomized controlled trials
1. Introduction Chronic kidney disease (CKD) has become an important problem affecting human health, and it is associated with increased risk of morbidity, mortality as well as huge healthcare costs [1]. Epidemiological studies indicate that CKD is also an independent
* Corresponding author. Department of Nephrology, Shanghai First People's Hospital, Shanghai Jiao Tong University, 100 HaiNing Road, Shanghai 200080, People's Republic of China. E-mail addresses: [email protected] (Z. Jing), [email protected] (Y. Wei-Jie).
risk factor for cardiovascular events, and the renineangiotensin system (RAS), oxidative stress, inflammation and hemodynamic disorders are reported to be involved in this pathophysiological process [2]. Previous studies have demonstrated the beneficial effects of dietary soy in humans and animal models. Frigolet et al. [3]. concluded that, soy protein is capable of attenuating abnormal expression of renineangiotensin. An in vitro and in vivo study conducted by Jia et al. [4]. found that, genistein, a soy isoflavones, protects against inflammation induced by TNF-a in vascular endothelial cells. And oxidative stress was significantly decreased in postmenopausal women when treated with soy isoflavones [5].
http://dx.doi.org/10.1016/j.clnu.2015.03.012 0261-5614/© 2015 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
Please cite this article in press as: Jing Z, Wei-Jie Y, Effects of soy protein containing isoflavones in patients with chronic kidney disease: A systematic review and meta-analysis, Clinical Nutrition (2015), http://dx.doi.org/10.1016/j.clnu.2015.03.012
2
Z. Jing, Y. Wei-Jie / Clinical Nutrition xxx (2015) 1e8
Moreover, a recent population-based study suggested soy or soy isoflavones intake significantly reduced the risk of postmenopausal breast cancer [6]. These beneficial findings have been applied for the development of preventive strategies for human health and disease. For example, the US Food and Drug Administration suggested that “25 g soy protein per day” may help prevent the risk of coronary heart disease because of reduced serum lipids and lipoproteins [7]. Emerging evidence from animal and human studies demonstrated that dietary soy protein containing isoflavones retard the development and progression of renal dysfunction. Reportedly, dialysis patients in japan where mean soy products consumption is significantly higher than that in the United State tend to live longer than patients in the US [8,9]. Iwasaki et al. [10] found Fisher rats in the soy diet group tend to have longer life than that in the caseintreated group, and the progression of renal dysfunction was obviously delayed when casein protein was replaced with soy protein. As a result, following the publication of several randomized, clinical trials, our attention has been directed to manipulate the dietary protein quality especially by replacing animal protein with soy protein that rich in isoflavones. Ogborn et al. [11] found that soy protein significantly ameliorate aging-related nephropathy in Fisher 344 rat and renal dysfunction in 5/6 nephrectomy mouse models. Furthermore, observational studies in patients with diabetic nephropathy demonstrated substitution of soy protein for animal protein results in decreased glomerular hyperfiltration and less proteinuria, whereas Marion et al. have not achieved the aim of attenuating proteinuria in the soy-treated group compared with control [12,13]. As described above, studies evaluating the effects of dietary soy in the management of CKD have demonstrated mixed findings, and these results were few systematically reviewed. As a result, this meta-analysis was designed to explore the effects of soy protein containing isoflavones intake in patients with CKD. 2. Methods
the effects of soy diets, soy protein or soy isoflavones in patients with CKD (including the dialysis patients) were identified. Only adult patients (over 18 years old) with renal dysfunction or kidney failure were eligible for inclusion. As for the initial screening process, the main exclusion criteria were as follows: participants younger than 18 years old; other outcomes such as bone metabolism, serum lipid levels; pharmacokinetics and basic studies; trials treated with isoflavones-depleted soy protein or isoflavones that is not extracted from soybeans; patients with renal transplantation; reviews and editorials. 2.3. Data extraction and quality assessment Data were collected independently from each included RCT, and any discrepancies were figured out by consensus. For those included trials, data were extracted based on a standardized strategies, including publications, study designs, size of study population, type of diet, characteristics of participants (such as age, gender, and treatment modality, etc.), interventions, the contents of isolated soy protein (ISP) and isoflavones, the main outcomes (such as serum creatinine, 24-h urine protein excretion, blood urea nitrogen, etc.), and duration of follow-up. As for Imani et al. [14]. that did not provide the contents of isoflavones, we calculated it based on the concentration of isoflavones in traditional soy products which is about 3.5 mg isoflavones per 1 g soy protein [15]. Methodological quality and risk of bias of each included RCT were examined carefully using the method described by the Cochrane Collaboration [16]. The items were as follows, 1. Random sequence generation; 2. Allocation concealment; 3. Blinding of participants and personnel; 4. Blinding of outcome assessment; 5. Incomplete outcome data; 6. Selective reporting; 7. Other sources of bias. All seven items were classified as “low risk of bias”, “high risk of bias”, or “unclear risk of bias”. Furthermore, intention-to-treat analysis (ITT) was also used to evaluate integrity of the outcome data.
2.1. Search strategy
2.4. Statistical analyses
A systematic literature search was performed in Medline, Embase and the Cochrane Database of Systematic Reviews for clinical trials evaluating the effects of soy diets, soy protein, or soy isoflavones in patients with CKD in December 2013 and updated in April 2014 for any new trials. The search strategies used were as follows, [(soy OR soybean OR soya OR soy protein OR isoflavones OR isoflavone OR genistein OR daidzein)] AND [(chronic kidney disease OR kidney failure OR chronic kidney failure OR kidney disease OR uremia OR dialysis OR continuous ambulatory peritoneal dialysis OR hemodialysis OR renal replacement therapy OR peritoneal dialysis OR Equilibrium dialysis OR extended daily dialysis)]. Only randomized controlled trials (RCTs) were identified for inclusion without any language limitation in our study. Title and abstract of each article was screened carefully for inclusion at the first step of retrieval process. Then all potentially relevant articles were examined carefully in full text for further identification. Reviews and meta-analyses on effects of soy diets or soy protein in patients with CKD were also checked seriously for any potential trials. Meanwhile, the cited references of each included studies were scanned carefully for additional trials or reviews. Trials were analyzed as a whole once it had more than one publication.
We grouped studies based on the mode of treatment as described above into the non-dialysis group and the dialysis group. Results were also calculated separately for subgroup analysis. Data from each included trials were analyzed using Review Manager (RevMan, Version 5.2, Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2012). Treatment effects were presented as the mean differences between changes and its 95% confidence intervals (CI), and the pooled effects were computed by assigning each trial a weight of the reciprocal of its variance. And if the raw data were unavailable, the variances for the changes of individual trials were calculated according to the methods described by the Cochrane Collaboration. Heterogeneity of treatment effects was evaluated using the x2 test, and the I2 test for inconsistency. The degree of inconsistency among those identified trials was defined by calculating the percentage of total betweenstudy variation because of heterogeneity rather than random variation as the I2 metric using the formula I2 ¼ 100% (Q-df)/Q [17]. We described the results using random-effect model when heterogeneity across those trials was significant, and on the contrary, a fixed-effect model was applied [18]. Though the outcomes evaluated in this meta-analysis have limited number of trials, the funnel plots calculated using the Review Manager 5.2 were used to assess the presence or absence of publication bias for certain outcomes. The P-value threshold for statistical significance was set at 0.05 for effect size, P 0.05 was considered significant.
2.2. Selection criteria Randomized, controlled trials with a crossover or parallel design that were of at least 7 weeks follow-up and designed to evaluate
Please cite this article in press as: Jing Z, Wei-Jie Y, Effects of soy protein containing isoflavones in patients with chronic kidney disease: A systematic review and meta-analysis, Clinical Nutrition (2015), http://dx.doi.org/10.1016/j.clnu.2015.03.012
Z. Jing, Y. Wei-Jie / Clinical Nutrition xxx (2015) 1e8
3. Results 3.1. Trial flow/flow of included studies 655 trials were identified by searching Medline, Embase and the Cochrane Database of Systematic Reviews, 594 of which were excluded at the initial screening. 61 potentially relevant trials were identified for further review, of which 12 trials fulfilled our inclusion criteria. All selected studies were randomized, controlled trials designed to evaluate the effects of soy protein containing isoflavones in patients with CKD. Detailed processes of our study selection are shown in Fig. 1.
3
low-protein diet while the other 6 studies adopted the usual diet, and Anderson et al. [12] was given a standard diabetes diet. Furthermore, 4 trials enrolled participants with diabetes mellitus, and five trials enrolled patients with hypertension. Besides, the exact contents of isolated soy protein (ISP) and isoflavones were presented clear in 8 trials, whereas the Teixeira et al. [20] and the Ahmed et al. [13]. applied soy protein containing isoflavones, and the exact content is unclear. The rest trials, the Anderson et al. [12] and Soroka et al. [19] did not described soy protein as isoflavonedepleted soy was thought to be treated with soy protein containing isoflavones. 3.3. Quality of included studies
3.2. Study characteristics 12 randomized trials (280 participants) designed to evaluate the effects of soy or soy products intake in patients with CKD were identified. Characteristics of the 12 trials were listed in Table 1, and Chen et al. [21] trial was analyzed as two subgroups and calculated separately based on the serum lipid levels of the participants. Based on the treatment modality, we divided these trials into two subgroups: the non-dialysis subgroup and the dialysis subgroup. Half of the included trials compared isolated soy protein containing isoflavones with milk or whey protein, and the rest employed soy diets or soy products such as soy beverage, textured soy protein into the treatment group. The contents of isolated soy protein consumption in our meta-analysis ranged from 14 to 30.7 g/d, and the isoflavones ranged from 35.5 to 61 mg/d. 8 of those included trials adopted a parallel design while the rest with a crossover design. As for the duration of follow-up, it varied from 7 weeks to 4 years across these included studies. And 5 of the studies adopted a
Based on the standard methods recommended by the Cochrane Collaboration for assessing risk of bias, trial quality was various across the included trials. Allocation concealment was adopted in 4 trials, the Anderson et al. [12], the Chen et al. [21], the Tabibi et al. [27], the Teixeira et al. [20]. And it is unclear in the rest 8 trials. Only the Ahmed et al. [13] trial was analyzed based on the intent-to-treat analysis (ITT). The result was shown in Fig. S1, 2. 3.4. Quantitative data synthesis 3.4.1. Serum creatinine A total of 7 trials were identified to evaluate the effects of soy consumption on serum creatinine, only participants in the Siefker et al. [24] trial which was not included in this analysis is on dialysis. As for the non-dialysis participants, serum creatinine was decreased with statistical significance in the soy-treated group compared with control, and the pooled estimated change was
Fig. 1. Flowchart demonstrating the literature screening process and reasons for exclusion are explained.
Please cite this article in press as: Jing Z, Wei-Jie Y, Effects of soy protein containing isoflavones in patients with chronic kidney disease: A systematic review and meta-analysis, Clinical Nutrition (2015), http://dx.doi.org/10.1016/j.clnu.2015.03.012
%
64 ± 18.38 57.5 63 63.6 ± 9.7 59.5 ± 12 53.6 ± 11.6 61.0 ± 2.9 50.3 62.1 ± 12.1 62.5 ± 12.1 52 ± 15 52.3 ± 14.9 43.9 ± 12.2 100 100 100 NAc NAc NAc 36 NAc 100 100 NAc NAc NAc Standard diabetes diet Low-protein dietf Low-protein dietf Usual hemodialysis diet Usual hemodialysis diet Usual hemodialysis diet Usual diet Usual diet Low-protein dietf Low-protein dietf Usual diet Usual diet Low-protein dietf
%
3.4.2. Creatinine clearance and glomerular filtration rate (GFR) 6 trials were identified to evaluate the effects of soy protein intake on creatinine clearance (Fig. 3). We did not find any significant differences in the change of creatinine clearance in the soytreated group compared with control. The pooled effects of soy protein consumption on creatinine clearance was 1.14 ml/min/ 1.73 m2 (95% CI: 3.70, 1.41 ml/min/1.73 m2; P ¼ 0.38), and no heterogeneity was detected (heterogeneity X2 ¼ 6.67, P ¼ 0.25, I2 ¼ 25%) (Fig. 3). Visual inspection of the funnel plot is asymmetric, suggesting the presence of publication bias (Fig. 4). Only 2 trials were included for the analysis of GFR which were determined by [99mTc]diethylenetriamine pentaacetic acid (DTPA) nuclear medicine scanning and 51Cr-EDTA separately. No significant change was found in the soy-treated group compared with control. The pooled effects of soy protein consumption on GFR was 1.30 ml/min/1.73 m2 (95% CI: 4.43, 1.83 ml/min/1.73 m2; P ¼ 0.41). The fixed-effect model was used as no heterogeneity was found (heterogeneity X2 ¼ 0.08, P ¼ 0.78, I2 ¼ 0%) (Table 2).
NAc 8 weeks NAc NAc NAc 6 months NAc NAc 7 months 30 35.5 12 weeks 30 35.5 12 weeks 30.7 36.3 12 weeks 25 54 8 weeks 25 52 13 months 15.5 ± 8.9 45.5 ± 15.7 4 years 17.8 ± 3.3 42.3 ± 18.2 7 weeks 14 49 8 weeks 14 61 8 weeks c c NA NA 8 weeks
RCd RCd RCd RPe RPe RPe RPe RPe RPe RCd RPe RPe RPe
Design Type of diet
3.4.3. Serum albumin A total of 9 trials (including a subgroup) reported data demonstrating the effects of soy intake on serum albumin (Fig. 5). Soy intake did not significantly affect the level of serum albumin, and the pooled estimated change was 0.02 g/dL (95% CI: 0.10, 0.06 g/ dL; P ¼ 0.71). There was no heterogeneity across these trials (heterogeneity X2 ¼ 3.18, P ¼ 0.92, I2 ¼ 0%). And the funnel-plot analysis did not show asymmetry demonstrating that the publication bias cannot substantially affect the result of this meta-analysis (Fig. 6). As for the non-dialysis patients, 4 trials were identified. The fixed-effect model was applied as no heterogeneity was detected (heterogeneity X2 ¼ 0.25, P ¼ 0.97, I2 ¼ 0%). No statistically significant difference was detected in serum albumin in the soy-treated group compared with control, and the pooled effects of the changes was 0.0 g/dL (95% CI: 0.09, 0.09 g/dL; P ¼ 0.97). 5 trials enrolling patients on dialysis were included for analysis of serum albumin. No heterogeneity was found in the dialysis subgroup (heterogeneity X2 ¼ 2.34, P ¼ 0.67, I2 ¼ 0%), and the fixedeffect model was used. We did not find any significant change in serum albumin in dialysis patients in the soy protein treated group compared with control, and the pooled effects of the changes was 0.07 g/dL (95% CI: 0.24, 0.09 g/dL; P ¼ 0.39).
Soy protein(þ)/animal protein Soy diet(þ)/animal protein diet ISPb(þ)/casein ISPb(þ)/milk protein ISPb(þ)/milk protein ISPb(þ)/milk protein ISPb(þ)/milk protein ISPb(þ)/whey protein Soy diet(þ)/animal protein Soy protein(þ)/animal protein Soy diet(þ)/control dietg Soy diet(þ)/control dietg Soy protein(þ)/animal protein
g/day
Duration Isoflavones
mg/day
ISPb
f
g
e
c
d
M/F, ratio of male and female. ISP, isolated soy protein. NA, not available. RC, randomized crossover design. RP, randomized parallel design. Low-protein diet, 0.8 g protein/kg/d. Control diet, usual diet without any soy. a
b
100/0 55.6/44.4 100/0 26.3/73.7 27.8/72.2 73/27 40/60 41.2/58.8 43.9/56.1 71.5/28.5 50/50 50/50 15.8/85.2 8 9 14 19 18 26 25 16 41 14 36 36 18 Anderson (1998) [12] Soroka (1998) [19] Teixeira (2004) [20] Chen 1 (2005) [21] Chen 2 (2005) [21] Chen (2006) [22] Fanti (2006) [23] Siefker (2006) [24] Azadbakht (2008) [25] Azadbakht (2009) [26] Imani (2009) [14] Tabibi (2010) [27] Ahmed (2011) [13]
N
%
Predialysis Predialysis Predialysis Hemodialysis Hemodialysis Hemodialysis Hemodialysis Hemodialysis Predialysis Predialysis Peritoneal dialysis Peritoneal dialysis Predialysis
Treatment modality Randomized intervention No. of subjects M/Fa Study
Table 1 Characteristics of 12 randomized controlled trials selected for analysis.
0.05 mg/dL (95% CI: 0.10, 0.00 mg/dL; P ¼ 0.04). The fixedeffect model was used as no heterogeneity was detected (heterogeneity X2 ¼ 4.29, P ¼ 0.37, I2 ¼ 7%) (Fig. 2). Because the Anderson et al. [12] trial that is not estimable was excluded, 5 trials were insufficient to detect publication bias.
100 0 100 0 0 0 44 NAc 100 100 30.6 NAc NAc
y
Z. Jing, Y. Wei-Jie / Clinical Nutrition xxx (2015) 1e8
Diabetes Hypertension Age
4
3.4.4. Serum phosphorus 6 trials were included in this section (Fig. 7). No significant change was demonstrated in the soy intake group, and the pooled estimated mean difference was 0.10 mg/dL (95% CI: 0.22, 0.02, P ¼ 0.10). Heterogeneity in this group of studies was not significant (heterogeneity X2 ¼ 9.83, P ¼ 0.08, I2 ¼ 49.0%). The funnel plot is symmetric demonstrating that publication bias cannot substantially affect the result of this meta-analysis (Fig. 8). As for the single outliers (Siefker et al. trial [24]), the results were not changed significantly after this study is excluded. The pooled effect of changes was 0.14 mg/dL (95% CI: 0.26, 0.01, P ¼ 0.03), and no heterogeneity was detected (heterogeneity X2 ¼ 4.20, P ¼ 0.38, I2 ¼ 5.0%) (Fig. S3). As for the 4 trials in the non-dialysis subgroup, Heterogeneity was not significant (heterogeneity X2 ¼ 4.04, P ¼ 0.26, I2 ¼ 26%). soy consumption significantly decreased serum phosphorus, and the pooled estimated mean difference was 0.13 mg/dL (95%
Please cite this article in press as: Jing Z, Wei-Jie Y, Effects of soy protein containing isoflavones in patients with chronic kidney disease: A systematic review and meta-analysis, Clinical Nutrition (2015), http://dx.doi.org/10.1016/j.clnu.2015.03.012
Z. Jing, Y. Wei-Jie / Clinical Nutrition xxx (2015) 1e8
5
Fig. 2. Net change (and 95% CI) of serum creatinine associated with soy protein containing isoflavones intake in non-dialysis patients.
Fig. 3. Pooled estimated effect of soy protein containing isoflavones on creatinine clearance in non-dialysis patients.
3.4.5. CRP This outcome was reported only in 3 trials. One in the nondialysis subgroup, and the other in the dialysis subgroup. CRP levels was significantly reduced in the soy-treated group compared with control, and the pooled estimated change in CRP level was 0.98 mg/L (95% CI: 1.25, 0.71 mg/L; P < 0.00001). This analysis was dominated by Azadbakht et al. [25] study, contributing 99.8% of the weight. No significant change was detected in the dialysis patients when treated with soy products, and the pooled estimated change was 0.83 mg/L (95% CI: 6.56, 4.91 mg/L; P ¼ 0.78) using the fixed-effects model (Table 2). 3 studies were inadequate to detect publication bias.
Fig. 4. Funnel plot with 95% CI to evaluate publication bias on creatine clearance in the predialysis participants.
CI: 0.26, 0.01 mg/dL; P ¼ 0.04). No significant change was demonstrated in the dialysis subgroup, and the pooled estimated change in serum phosphorus levels was 0.26 mg/dL (95% CI: 0.16, 0.68 mg/dL; P ¼ 0.22).
3.4.6. Blood urea nitrogen (BUN) 5 trials were identified. Blood urea nitrogen was significantly decreased in the soy-treated group compared with corresponding control group. Heterogeneity was demonstrated significant across these trials (heterogeneity X2 ¼ 78.08, P < 0.00001, I2 ¼ 95%), and the random-effect model was applied. The overall pooled estimate of the effect of soy protein was 3.07 mg/dl (95% CI: 6.03, 0.11 mg/dL; P ¼ 0.04). 5 trials were insufficient to detect publication bias. 4 trials were included in the non-dialysis group. A randomeffect model was used as a result of significant heterogeneity (heterogeneity X2 ¼ 75.92, P < 0.00001, I2 ¼ 96%). No significant
Table 2 Summary of the effects of Soy protein intake in patients with CKD compared with control. Items
No. of studies
No. of populations (treatment/control)
Test for I2
Heterogeneity P
Analysis model
Test for Z
Overall effect P
Mean difference, 95% CI
CRP Non-dialysis subgroup Dialysis subgroup BUN Non-dialysis subgroup Dialysis subgroup 24 h Urine protein GFR Body weight Body mass index
3 1 2 5 4 1 4 2 3 2
42/40 20/21 22/19 58/61 51/52 7/9 52/53 17/17 43/44 23/23
4% e 52% 95% 96% e 85% 0% 56% 89%
0.35 e 0.15
