Following an application from the European Natural Soyfood Manufacturers Association (ENSA), the European Vegetable Protein Federation (EUVEPRO) and the Soya Protein Association (SPA), submitted pursuant to Article 14 of Regulation (EC) No 1924/2006 via the Competent Authority of Belgium, the Panel on Dietetic Products, Nutrition and Allergies was asked to deliver an opinion on the scientific substantiation of a health claim related to isolated soy protein (ISP) and reduction of blood LDL-cholesterol concentrations.
The scope of the application was proposed to fall under a health claim referring to disease risk reduction.
The food constituent that is the subject of the health claim is “protein-rich soybean component with limited quantities of macro- and micronutrients”, which upon a request from EFSA was further defined by the applicant as ISP. The Panel considers that ISP (as defined by the applicant) is sufficiently characterised in relation to the claimed effect.
The claimed effect is “reduction of total and LDL-cholesterol in healthy subjects with normal or mildly elevated blood cholesterol; a reduction of total and LDL cholesterol has been shown to reduce the risk of heart disease”. The target population proposed by the applicant is healthy subjects with normal or mildly elevated blood cholesterol levels. The Panel considers that reduction of blood LDL-cholesterol concentrations is a beneficial physiological effect. A reduction in blood LDL-cholesterol concentrations reduces the risk of CHD.
The applicant identified 23 RCTs, which used ISP and one RCT which used the water insoluble fraction of a partially hydrolysed soy protein as interventions for the scientific substantiation of the claim, and presented an unpublished meta-analysis of 23 of these 24 RCTs. In addition, six RCTs, which used whole soy foods, and eight observational studies were provided as well as one animal and two in vitro studies on the mechanism by which ISP could exert the claimed effect. Eleven published meta-analyses and one unpublished systematic review were presented in support of the claim.
The Panel considers that the study which assessed the effect of the water insoluble fraction of a partially hydrolysed soy protein on blood cholesterol concentrations cannot be used to substantiate a claim on ISP owing to its compositional differences, which might have had an impact on the claimed effect. Four of the 23 RCTs on ISP provided did not allow conclusions to be drawn on the effects of ISP on blood cholesterol concentrations owing to inadequate methodology or insufficient reporting.
Fourteen of the remaining 19 RCTs had a parallel design while five had a cross-over design. All subjects (n=1,947) who entered data analysis had normal or moderately elevated blood LDL-cholesterol concentrations at baseline.
The Panel notes that four RCTs with 15-41 subjects per group reported a statistically significant effect of ISP on blood LDL/non-HDL cholesterol concentrations at doses of 20 to 40 g per day consumed for six to 24 weeks. However, the Panel also notes that 14 RCTs with >15 subjects per group/period (four of which had about 90 subjects per group/period), did not report such an effect under similar conditions (20 to 40 g ISP per day given for at least six weeks in all but four studies), and that the effects found in one RCT were not consistent. In addition, most of the RCTs were at high risk of bias, and differences in the results obtained between trials appeared to be unrelated to the dose of ISP used, sample size or study duration.
The applicant also provided a meta-analysis performed on 23 of the 24 RCTs presented as the main body of evidence for the scientific substantiation of the claim. The Panel notes that this meta-analysis did not include any additional studies to those provided individually by the applicant for the scientific substantiation of the claim. The Panel also notes that this meta-analysis includes the four RCTs identified above as having inadequate methodology or insufficient reporting, as well as the one RCT which used the water insoluble fraction of a partially hydrolysed soy protein as intervention, and considers that this meta-analysis does not provide any additional information to the individual studies considered for the scientific substantiation of the claim.
Ten of the 11 published meta-analyses as well as the unpublished systematic review provided were not designed to assess the effects of ISP but rather of isoflavones or of soy protein from different sources, including soy foods, on blood cholesterol concentrations. Although results of (sub-group) analyses for ISP were presented in some of the meta-analyses, all of these contained studies not considered pertinent to the claim by the applicant, or the reporting in the publication was insufficient for a full scientific evaluation. Also, the meta-analysis, which was designed to assess the effect of ISP on blood lipid concentrations, included studies which were considered by the applicant as not pertinent to the claim owing to methodological limitations in the design, to the high doses of ISP used, or to the characteristics of the study population. The Panel considers that no conclusions can be drawn from these meta-analyses and the systematic review for the scientific substantiation of the claim.
The six RCTs and the eight epidemiological studies on whole soy foods, which were provided by the applicant as supportive evidence for the scientific substantiation of the claim, did not allow conclusions to be drawn on the scientific substantiation of the claim on ISP and blood cholesterol concentrations, owing to the differences in macronutrient and fibre composition between whole soy foods and ISP, which might have had an impact on the claimed effect.
With regard to a possible mechanism by which ISP could exert the claimed effect, the applicant suggests that peptides derived from intestinal digestion of ISP can enter the circulation and exert a direct effect on the hepatic metabolism of cholesterol by increasing the expression of the hepatic LDL receptor. As evidence for this mechanism, the applicant provided two in vitro studies on the effect of the soybean 7S globulin α’ subunit and a purified recombinant polypeptide containing the N-terminal extension region of the soybean α’ subunit on LDL-receptor mediated LDL uptake and degradation in a human hepatoma cell line (Hep G2). The applicant also presented one animal study in male Sprague-Dawley rats, which assessed the effect of soybean 7S globulin and soybean 7S globulin α’ subunit on blood cholesterol concentrations and β-VLDL receptor activity. The Panel considers that results from rat studies cannot be extrapolated to humans because of differences in lipid metabolism between these two species, that the evidence provided did not establish that peptides derived from the intestinal digestion of ISP can be absorbed intact and that the evidence provided in the in vitro studies is not sufficient to predict an effect of peptides derived from the soybean 7S globulin α’ subunit on the LDL-receptor mediated LDL uptake and degradation in humans.
In weighing the evidence, the Panel took into account that, under similar conditions four RCTs reported an effect of ISP on blood LDL/non-HDL cholesterol concentrations, whereas 14 RCTs did not report such an effect and another RCT showed no consistent effects. The Panel also took into account that most of these RCTs were at high risk of bias, that differences in the results obtained between trials appear unrelated to the dose of ISP used, sample size or study duration and that the evidence provided in support of a possible mechanism was not convincing.
The Panel concludes that a cause and effect relationship has not been established between the consumption of ISP (as defined by the applicant) and a reduction in blood LDL-cholesterol concentrations.