ID 1135 -
Izoflawony sojowe
PL: Izoflawony sojowe
EN: Soy Isoflavones
Pdf: soy isoflavones
1. Charakterystyka żywności / składnika
The food constituent that is the subject of the health claims is soy isoflavones.
Soy isoflavones constitute a wide range of compounds of plant origin, which mainly comprise genistein, daidzein and glycitein, among others (Ma et al., 2008a, 2008b). Soy isoflavones can be consumed as isolated soybean protein (ISP), as whole-soybean foods or extracts, as supplements or as pure compounds (Cassidy et al., 2006).
The Panel considers that the food constituent, soy isoflavones, which is the subject of the health claims, is sufficiently characterised.
2.2. Utrzymanie prawidłowego stężenia cholesterolu LDL we krwi (ID 1135, 1704a, 3093a)
The claimed effects are “cholesterol management / heart health”, “menopause/skin and hair health during menopause/cholesterol management” and “act as phyto-estrogens”. The Panel assumes that the target population is the general population.
In the context of the proposed wordings and the clarifications provided by Member States, the Panel assumes that the claimed effects refer to the maintenance of normal blood LDL-cholesterol concentrations.
Low-density lipoproteins (LDL) carry cholesterol from the liver to peripheral tissues, including the arteries. Elevated LDL-cholesterol, by convention >160 mg/dL (>4.1 mmol/L), may compromise the normal structure and function of the arteries.
The Panel considers that maintenance of normal blood LDL-cholesterol concentrations is a beneficial physiological effect.
3.2. Utrzymanie prawidłowego stężenia cholesterolu LDL we krwi (ID 1135, 1704a, 3093a)
Among the references provided were a number of narrative reviews and textbooks which were either not related to the claimed effect or did not provide any original data which could be used for the scientific substantiation of the claim. A number of human, animal and in vitro studies did not address changes in LDL-cholesterol concentrations; these included references on the effect of soy isoflavones on symptoms associated with menopause, severity of asthma, DNA damage, mood, cognitive function, bone loss, bone fractures, cancer, LDL peroxidation, weight loss, inflammation, cardiovascular disease risk, deposition of cholesterol in the aorta and endothelial function. Some of the references provided did not address the food constituent, which is the subject of the claim, but addressed soy isoflavones in combination with other substances, or red clover isoflavones.
Fifty-five human intervention studies were provided either separately in the consolidated list or in the four meta-analyses which assessed the effect of soy isoflavones as extracts, pure compounds or in soy protein on blood cholesterol concentrations.
Some human intervention studies used soy nuts (Welty et al., 2007b), soy foods in general (Chiechi et al., 2002; Scheiber et al., 2001), soy meal replacement formula (Allison et al., 2003), soy protein containing isoflavones (Maesta et al., 2007; Sagara et al., 2004; Washburn et al., 1999) or a soy extract, fibre and lecithin supplement (Puska et al., 2002), without controlling either for the protein component, the amount of lecithin contained or the macronutrient composition, which could have an effect on blood cholesterol concentrations. In one study (Takatsuka et al., 2000), the nature of the control diet was not specified. Two studies, one of which was uncontrolled, did not report the amount of isoflavones administered (Wong et al., 1998; Yildirir et al., 2001). One study examined the effect of soy isoflavones on post-prandial cholesterol concentrations only (Campbell et al., 2006), one study assessed total cholesterol concentrations without reporting on LDL-cholesterol or other cholesterol fractions (Urban et al., 2001), and one study purported to have evaluated blood cholesterol concentrations but did not report any results (Scambia et al., 2000). In four studies (Basaria et al., 2009; Jenkins et al., 2000; Jenkins et al., 2002; West et al., 2005) some subjects took medication such as hormone replacement therapy (HRT), levothyroxine or cholesterol lowering medication which could have an impact on the claimed effect, and it was unclear whether the sub-group analysis performed in subjects not taking medication in one of these studies (West et al., 2005) was pre-planned or not. In one study, subjects who dropped out were replaced after randomisation (Lichtenstein et al., 2002), introducing possible selection bias. In a further study, subjects in the intervention and placebo groups were significantly different with respect to their total and LDL-cholesterol concentrations at baseline, and it was unclear how this difference was taken into account in the analysis (Petri Nahas et al., 2004). Three cross-sectional studies (de Kleijn et al., 2002; Nagata et al., 1998; Somekawa et al., 2001) were provided which investigated the association between soy consumption and changes in blood LDL-cholesterol concentrations. The Panel notes that the observational study design does not allow controlling with sufficient precision for the macronutrient composition of the diets and medication use, both of which could have an effect on blood cholesterol concentrations, besides isoflavones. The Panel considers that no conclusions can be drawn from these references for the scientific substantiation of the claim.
A meta-analysis of randomised controlled trials (RCTs) (Taku et al., 2008), which assessed the effects of consumption of extracted soy isoflavones vs. placebo on blood LDL-cholesterol concentrations, included the individual studies of this type which were provided in the consolidated list (Cheng et al., 2007; Gonzalez et al., 2007) except three (Atteritano et al., 2007; Lissin et al., 2004; Nahas et al., 2007).
In the meta-analysis by Taku et al. (2008), several databases were searched for RCTs published between 1966 and 2007. Reference lists of relevant systematic reviews and meta-analyses were hand searched. Twelve studies met the inclusion criteria of RCTs conducted in adults and published in English, Japanese or Chinese, with duration of 1-3 months. Six trials used a parallel design and the remaining trials used a cross-over design. The study population was post-menopausal women (n=565) in all studies, and doses of total isoflavones were 27-132 mg/day expressed as aglycone equivalents. In most of the studies included, women consumed similar diets with similar amounts of fat, cholesterol and fibre, and no significant body weight changes were reported during the study period. The funnel plots did not indicate publication bias. Using either a random or a fixed (i.e. when studies responsible for significant heterogeneity were removed from analysis) effects model, no effect of isolated soy isoflavone consumption on total (ten studies) or LDL-cholesterol (nine studies) concentrations was observed. The Panel notes that no justification has been provided for excluding studies with a duration longer than three months, and that this meta-analysis does not show an effect of extracted soy isoflavones on blood cholesterol concentrations when consumed for 1-3 months at doses of 27-132 mg/day.
In a randomised double-blind, placebo-controlled parallel study conducted in 389 post-menopausal women (49-67 years), the effect of genistein administration (54 mg/day) for 24 months on total and LDL-cholesterol concentrations was investigated (Atteritano et al., 2007). After a run-in phase of four weeks in which all women followed a fat-restricted diet, women were assigned to consume either genistein (n=198) or placebo (n=191) in the form of capsules. All capsules also contained 500 mg calcium and 400 IU vitamin D. Eighty-five subjects withdrew before completion of the study (48 in the genistein group and 37 in the placebo group). Analyses were performed in the intention-to-treat population. No significant changes in total or LDL-cholesterol concentrations were observed between groups at either 12 or 24 months. The Panel notes that this study does not show an effect of genistein consumption on blood cholesterol concentrations.
In a randomised, double-blind, placebo-controlled parallel study 80 post-menopausal women (mean age approx. 55 years) were allocated to consume daily either 100 mg of isoflavones in 250 mg standardised soy extract (n=40) or placebo (lactose; n=40) for ten months (Nahas et al., 2007). Four subjects dropped out during the study (two in the intervention group (n=38 analysed) and two in the placebo group (n=38 analysed)) and were not taken into account in the analysis. No statistically significant differences were observed between the isoflavone and the placebo group with respect to total or LDL-cholesterol concentrations. The Panel notes that missing data have not been taken into account in the analysis, and that this study does not show an effect of extracted soy isoflavones on blood cholesterol concentrations.
In another randomised, double-blind, placebo-controlled parallel trial, 40 post-menopausal women (mean age 61.6±8.4 years) with moderate hypercholesterolaemia were assigned to consume for six weeks either 2x45 mg/day soy isoflavones in tablet form (n=20) or placebo (n=20), which was identical in composition but isoflavone-free (Lissin et al., 2004). All women completed the study. No significant changes in total or LDL-cholesterol concentrations were observed between groups. The Panel notes that this study does not show an effect of extracted soy isoflavones on blood cholesterol concentrations.
The Panel notes that one meta-analysis including nine RCTs, and three additional RCTs, did not show an effect of extracted soy isoflavones on blood LDL-cholesterol concentrations.
Three meta-analyses and 23 human intervention studies were provided which assessed the effect of soy isoflavones in soy protein against isoflavone-depleted soy protein or animal protein.
The meta-analyses by Zhuo et al. (2004) and Taku et al. (2007) investigated the effects of consumption of soy isoflavones in isolated soy protein (ISP+) compared to isoflavone-depleted ISP (ISP-) on blood cholesterol concentrations in meta-analyses of RCTs published between 1966 and 2003, and between 1990 and 2006, respectively, with duration of 1-3 months, using either a parallel or cross-over design and which provided baseline and endpoint blood cholesterol concentrations. Eight (n=471 subjects) and eleven (n=441 subjects) studies, respectively, met the inclusion criteria. The Panel notes that with the exclusion of studies comparing ISP+ to animal protein these meta-analyses did not include all the studies pertinent to the claim, and considers that no conclusions can be drawn from these meta-analyses for the scientific substantiation of the claim.
Zhan and Ho (2005) conducted a meta-analysis of RCTs published in English between 1995 and June 2002 on the effect of ISP+ or isoflavone extracts on blood lipid concentrations. Twenty-three studies (n=1,381 subjects) met the inclusion criteria (RCTs providing the amount of soy isoflavones and blood lipid concentrations at baseline). The study duration ranged from 3 to 26 weeks. Seventeen studies used ISP+, three used tablets containing extracted isoflavones, and three used textured soy foods. In sub-group analyses, eight comparisons (n=440) were made for ISP+ vs. ISP-, 27 comparisons (n=1,501) for ISP+ vs. casein, and three comparisons (n=135) for extracted soy isoflavones vs. placebo. The Panel notes that in two studies (Puska et al., 2002; Washburn et al., 1999) the control was not appropriate to assess the effects of soy isoflavones in soy protein (i.e. carbohydrates and no
isoflavone-free control, and the intervention and control not matched with respect to the lecithin content). One study included subjects on HRT, levothyroxin and lovastatin (Jenkins et al., 2000), and one study assessed the effects of red clover isoflavones rather than soy isoflavones (Hodgson et al., 1998). It is also unclear which studies were included in the sub-group analyses with respect to ISP+ vs. ISP-, and with respect to ISP+ vs. casein, or the reasons for excluding from the sub-group analyses those studies which used animal proteins other than casein (e.g. whey and total milk protein) as control. The Panel notes the methodological limitations of this meta-analysis and considers that no conclusions can be drawn from this meta-analysis for the scientific substantiation of the claim.
All of the individual studies described below were either provided in the three meta-analyses described above or provided separately in the consolidated list.
In a double-blind, randomised, controlled parallel trial, 202 healthy post-menopausal women (60 to 75 years) with mean baseline LDL-cholesterol concentrations of 4.1 mmol/L were assigned to receive daily 25.6 g soy protein containing 99 mg isoflavones (n=100) or total milk protein (n=102) as a powder for 12 months (Kreijkamp-Kaspers et al., 2004). Power calculations were made for three outcome variables, including total cholesterol. It was estimated that 100 subjects per group had to be recruited to detect a 7.4 % difference in total cholesterol concentrations between groups with a power of 80 %, α=0.05, and a drop-out rate of 25 %. A total of 49 subjects withdrew from the study (24 in the placebo group, 25 in the soy group). A modified intention-to-treat analysis was used including all subjects with at least two measurements, including baseline plus a close-out visit (n=22, 9 in the placebo group (n=87 analysed), 13 in the soy group (n=88 analysed)). No statistically significant differences were found in total or LDL-cholesterol concentrations between the soy and the placebo group. The Panel notes that this adequately powered study does not show an effect of soy isoflavones in soy protein on blood cholesterol concentrations.
In a double-blind, randomised, controlled parallel intervention 213 healthy subjects (105 women, 50-75 years) with mean baseline blood LDL-cholesterol concentrations of approx. 3.85 mmol/L were assigned to consume daily either 40 g of soy protein powder with 188 mg of soy isoflavones (n=105) or casein powder (n=108) to be mixed with beverages for three months (Teede et al., 2001). Thirty-four subjects (19 in the soy group and 15 in the placebo group did not complete the study and were not taken into account in the analysis (n=86 analysed on the soy group and n=93 analysed in the placebo group). No statistically significant differences were observed between groups with respect to total or LDL-cholesterol concentrations. The Panel notes that missing data were not taken into account in the analysis, and that this study does not show an effect of soy isoflavones in soy protein on blood cholesterol concentrations.
The randomised, double-blind, controlled parallel study by Allen et al. (2007) was designed to assess the effects of soy isoflavones in soy protein on LDL-cholesterol concentrations. A total of 216 post-menopausal women (mean age: approx. 57 years) with mean baseline LDL-cholesterol concentrations of approx. 3.6 mmol/L were assigned to consume 20 g/day soy protein containing 160 mg total isoflavones (96 mg aglycones; n=107) or 20 g/day whole milk protein (control; n=109) for 12 weeks after a one-month, single-blind, run-in period to select women with high (>80 %) compliance with the study products. A total of 25 women (14 in the soy and 11 in the control group) dropped out during the study and their baseline measures were used for missing data. The authors reported a statistically significant decrease in total and LDL-cholesterol concentrations after controlling for associated variables (age, race, weight change, change in dietary fat intake, energy expenditure) in the soy compared to the placebo group at week 6 but not at week 12. The Panel notes that carrying forward observations is not an appropriate method of taking into account missing data and that this study does not show a sustained effect of soy isoflavones in soy protein on blood cholesterol concentrations.
A randomised, double-blind, controlled parallel trial which included 182 hyperlipidaemic men and women (44 % male, 30-70 years, mean baseline cholesterol concentrations not reported), assessed the effects on blood cholesterol concentrations of the consumption of a supplement containing 31.5 g/day soy protein and 120 mg/day isoflavones (as aglycones) (n=81 analysed) compared to a milk protein supplement (28.5 g/day; n=78 analysed) for five weeks after a three-week run-in period on the milk protein supplement (Ma et al., 2005). Twenty-two subjects dropped out (six before randomisation and eight in each group during the study), and a further subject was not included in the analysis, which was performed in completers only. No statistically significant differences were found between the two groups with respect to changes in total or LDL-cholesterol concentrations. The Panel notes that missing data were taken into account in the analysis, and that this study does not show an effect of soy isoflavones in soy protein on blood cholesterol concentrations.
In a double-blind, randomised cross-over study 43 male subjects (mean age: approx. 28 years) with mean baseline blood LDL-cholesterol concentrations of 2.74±1.06 mmol/L consumed daily either ISP containing 0.75 mg/kg body weight of isoflavones (around 32 g ISP with 62 mg isoflavones per day) (ISP+), ISP containing low amounts of isoflavones (around 1.6 mg per day) (ISP-) and milk protein powder for periods of 57 days each with a 28-day wash-out period in between (McVeigh et al., 2006). Four subjects dropped out during the study and a further four were excluded from analysis (35 subjects analysed). No statistically significant differences were observed between any of the periods with respect to total or LDL-cholesterol concentrations. The Panel notes that missing data were not taken into account in the analysis, and that this study does not show an effect of soy isoflavones in soy protein on blood cholesterol concentrations.
The double-blind, randomised, controlled parallel intervention by Vigna et al (2000) assessed the effects on cholesterol concentrations of daily consumption of 60 g ISP containing 76 mg isoflavones (n=51) compared to casein (n=53) in 104 post-menopausal women (mean age: approx. 53 years) with mean baseline LDL-cholesterol concentrations of approx. 4.2 mmol/L. Eleven subjects from the ISP group and 14 subjects from the placebo group dropped out during the study, and a further two subjects from the placebo group were excluded from analysis (n=40 analysed in the ISP group and n=37 analysed in the placebo group). No statistically significant differences were found between the two groups with respect to changes in total or LDL-cholesterol concentrations. The Panel notes that missing data were not taken into account in the analysis, and that this study does not show an effect of soy isoflavones in soy protein on blood cholesterol concentrations.
In a double-blind, randomised cross-over study, 42 post-menopausal women (mean age: approx. 55 years) with mean baseline blood LDL-cholesterol concentrations of 2.89±0.1 mmol/L consumed ISP with 107.67 mg of total isoflavones (aglycone units), ethanol-washed ISP with 1.82 mg total isoflavones, and total milk protein, in a daily dose of 25 g protein for periods of six weeks each with a four-week wash-out period in between (Steinberg et al., 2003). A total of 14 women withdrew from the study and were not taken into account in the analysis (28 subjects analysed). There were no statistically significant changes in total or LDL-cholesterol concentrations between any of the periods. The Panel notes that missing data were not taken into account in the analysis, and that this study does not show an effect of soy isoflavones in soy protein on blood cholesterol concentrations.
In a double-blind, randomised, controlled parallel intervention, 69 women (42-62 years) (baseline total or LDL-cholesterol concentrations not reported) received either ISP containing 80.4 mg/day isoflavone aglycone components (ISP+, n=24), 4.4 mg/day isoflavone aglycone components (ISP-, n=24) or control (whey protein, n=20) in the form of muffins and soy or placebo powder as meal replacements for 24 weeks in four different cohorts in which the interventions were equally represented (Dent et al., 2001). One control subject was not included in the analyses. No statistically significant effect of soy isoflavones on total or LDL-cholesterol concentrations between groups was found at any time point (12 and 24 weeks). The Panel notes that this study does not show an effect of soy isoflavones in soy protein on blood cholesterol concentrations.
The paper by Mackey et al. (2000) reported on two human intervention studies, one of which was a sequential study without wash-out period in between and for which it was unclear whether the control diet was matched to the intervention diet with respect to macronutrient composition, which could have an impact on the claimed effect. The Panel considers that no conclusions can be drawn from this study for the scientific substantiation of the claim. The other study reported in the paper was a double-blind, randomised, controlled parallel intervention in 54 post-menopausal women (mean age: approx. 56 years) with total cholesterol concentrations >5.5 mmol/L and mean baseline LDL-cholesterol concentrations of 5.1 mmol/L, who consumed daily 28 mg of soy protein powder containing 65 mg isoflavones (ISP+, n=22 analysed) or <4 mg isoflavones (ISP-, n=24 analysed) after a four-week run-in period in which the dietary guidelines from the National Heart Foundation were followed. Eight subjects did not complete the study (group not reported) and were not taken into account in the analysis. No statistically significant differences were observed between groups with respect to total or LDL-cholesterol concentrations. The Panel notes that missing data were not taken into account in the analysis, and that this study does not show an effect of soy isoflavones in soy protein on blood cholesterol concentrations.
In a randomised, single-blind, controlled parallel trial, 43 obese women (20-65 years) with mean baseline LDL-cholesterol concentrations of approx. 3.0 mmol/L were assigned to consume daily, in the context of an energy-restricted diet, either three soy-based isoflavone-containing (91 g protein, 150 mg isoflavones (aglycones); n=22) or three casein-based (n=21) meal replacement shakes for 16 weeks after a two-week run-in period (Anderson et al., 2007). The shakes were matched with respect to their macronutrient composition, but differed in their calcium, potassium and isoflavone content. Eight women did not complete the study (three in the casein and five in the soy group) and were not taken into account in the analysis (n=17 analysed in the soy group and n=18 analysed in the placebo group). No statistically significant differences were observed between the soy and the casein group with respect to total or LDL-cholesterol concentrations. The Panel notes that missing data were not taken into account in the analysis, and that this study does not show an effect of soy isoflavones in soy protein on blood cholesterol concentrations.
In a double-blind, randomised cross-over trial Merz-Demlow et al. (2000) studied the effects of consuming 75-100 g of soy powder (depending on the subject‟s body weight), which provided an average of 53 g protein and 10 mg (control), 65 mg (low isoflavone) and 129 mg (high isoflavone) soy isoflavones (aglycone units), on cholesterol concentrations in the different phases of the menstrual cycle in 20 women (mean age: 26.3±4.8 years). The intervention covered three menstrual cycles plus nine days, with wash-out periods of 2-3 weeks in between during which an ad libitum diet was consumed. Six subjects did not complete the intervention and a further subject was excluded from analysis (13 subjects analysed). Compared with the control diet the high isoflavone diet lowered LDL-cholesterol concentrations significantly, by 7.6 % and 10 %, in the midfollicular and periovulatory phases, respectively, (p<0.02) but not in the early follicular and midluteal phases. The Panel notes that this study evaluated cholesterol lowering effects during menstrual phases only, and did not show a sustained effect of soy isoflavones in soy protein on blood cholesterol concentrations.
In a double-blind, randomised, controlled parallel trial, 57 post-menopausal women (47-72 years) with mean baseline LDL-cholesterol concentrations of approx. 3.7 mmol/L were allocated to consume 40 g ISP containing either 1.2 mg isoflavones and 0.22 g phytate (LP/LI, n=14 analysed), 1.2 mg isofavones and 0.64 g phytate (NP/LI, n=13 analysed), 85.8 mg isoflavones and 0.22 g phytate (LP/NI, n=14 analysed), or 84.6 mg isoflavones and 0.78 g phytate (NP/NI, n=14 analysed) in the form of smoothies or other foodstuffs for six weeks (Engelman et al., 2005). Two subjects did not complete the study (group not reported) and were not taken into account in the analysis. No statistically significant differences were observed between groups with respect to total or LDL-cholesterol concentrations. The Panel notes that this study does not show an effect of soy isoflavones in soy protein on blood cholesterol concentrations.
In a 15-month, double-blind, randomised, controlled parallel intervention, which involved nine months of dietary intervention and a six-month (intervention-free) follow-up, 65 post-menopausal women (mean age: approx. 55 years) with mean baseline cholesterol concentrations of approx. 3.5 mmol/L were assigned to consume daily either 40 g ISP with 96 mg isoflavones (n=17 analysed), 40 g ISP with 52 mg isoflavones (n=19 analysed) or ISP containing 4 mg isoflavones (n=14 analysed) (Gallagher et al., 2004). Fifteen women (group not reported) dropped out during the study and were not taken into account in the analysis. No statistically significant differences were observed between groups with respect to total or LDL-cholesterol concentrations at the end of the intervention. The Panel notes that missing data were not taken into account in the analysis, and that this study does not show an effect of soy isoflavones in soy protein on blood cholesterol concentrations.
In a double-blind, randomised cross-over trial, Meinertz et al. (2002) studied the effect of consuming total liquid diets composed of casein, soy protein or ethanol-extracted soy protein in 12 subjects (6 women; 22-68 years) with mean baseline LDL-cholesterol concentrations of 2.1±0.5 mmol/L for 32 days each, with wash-out periods of at least 18 days in between in which self-selected solid foods were consumed. The Panel assumes that the diets provided 140 g of ISP with 335 mg isoflavones, 140 g of ISP with 15.4 mg of isoflavones and 139 g of casein. No information on drop-outs was provided. No statistically significant differences were observed between the three diets with respect to total or LDL-cholesterol concentrations. The Panel notes that this study does not show an effect of soy isoflavones in soy protein on blood cholesterol concentrations.
The double-blind, controlled parallel intervention by Gardner et al. (2001) assessed the effect of daily consumption during 12 weeks of 42 mg ISP containing 80 mg of isoflavones (ISP 80; n=31 analysed) or trace amounts of isoflavones (ISP-, n=33 analysed), or milk protein (n=30 analysed) to be mixed with beverages in 115 post-menopausal women (mean age: approx. 60 years) with mean baseline LDL-cholesterol concentrations of approx. 4.2 mmol/L. During the four-week run-in phase in which the milk protein supplement was consumed, and before randomisation, 15 women dropped out, and a further six (three in the ISP 80, one in the ISP- and two in the milk protein group) did not complete the study after randomisation and were not taken into account in the analysis. It was estimated that 30 subjects per group would provide 80 % power to detect a difference in change in LDL-cholesterol concentrations of 10 %, using an α=0.05. A statistically significant difference between the ISP 80 and the ISP- group was observed with respect to total and LDL-cholesterol concentrations at the end of the study (-0.27 vs. -0.02 mmol/L, p=0.03 and -0.38 vs. -0.09 mmol/L, p=0.005, respectively). No significant differences were observed between any of the ISP groups compared to the milk protein group. The Panel notes that missing data were not taken into account in the analysis and that this study shows an effect on blood cholesterol concentrations of ISP+ when compared to ISP-, but not if ISP+ was compared to animal protein. The Panel also notes that the results of this study are inconsistent with respect to the effects of soy isoflavones in soy protein on blood cholesterol concentrations.
In a randomised, cross-over study (Ashton and Ball, 2000), 45 healthy males (34-62 years) with mean baseline LDL-cholesterol concentrations of 3.68±0.86 mmol/L consumed a tofu-based diet (290 g tofu/day) providing 80 mg isoflavones per day, and a meat-based diet (150 g cooked lean red meat), for four weeks each with a two-week wash-out period in between. All subjects consumed similar vegetarian breakfasts, lunches and snacks on both diets, and diets were comparable with respect to macronutrient composition (except for the amount of cholesterol they provided). Three subjects did not complete the study and were not taken into account in the analysis (42 subjects analysed). Compared to the meat diet, total cholesterol concentrations were significantly lower in the tofu diet period (mean difference: 0.23, 95 % CI 0.02, 0.43), whereas no statistically significant differences were observed for LDL-cholesterol concentrations. The Panel notes that missing data were not taken into account in the analysis, and that this study shows an effect of soy isoflavones in soy foods on total blood cholesterol concentrations.
In a randomised, blinded, controlled parallel intervention, 82 post-menopausal women (aged 45-55 years) with mean baseline LDL-cholesterol concentrations of 3.4 mmol/L were assigned to consume daily either 150 mg soy protein and 100 mg isoflavones in capsules (n=41) or placebo capsules containing the same amount of isoflavone-free soy protein plus glucose (n=41) for four months (Han et al., 2002). One subject from each group dropped out during the study. Total and LDL-cholesterol concentrations significantly decreased in the isoflavone group compared to placebo (mean change: -26.6±1.2 mg/dL vs. 0.2±0.4 mg/dL, and -13.3±1 mg/dL vs. 5.5±1.2 mg/dL, for total and LDL–cholesterol concentrations, respectively, p<0.001). The Panel notes that only basic statistical analyses were performed, and that this study shows an effect of soy isoflavones in soy protein on blood cholesterol concentrations.
In a double-blind, placebo controlled, parallel trial, 159 subjects (25 women; mean age: 52 years) with LDL-cholesterol concentrations ≥4 mmol/L, after having followed an American Heart Association Step I diet for 4-24 weeks and having mean baseline LDL-cholesterol concentrations of approx. 4.8 mmol/L, were randomised to one of four groups to consume in the form of beverages, while continuing the Step I diet: 30 g ISP with 111 mg isoflavones and 10 g cotyledon fibre (n=39): 50 g ISP with 185 mg isoflavones and 16.6 g cotyledon fibre (n=40): 30 g casein and 10 g cellulose (n=40) or 50 g casein and 16.6 g cellulose (n=40) for 16 weeks (Tonstad et al., 2002). The study was designed to have 90 % power to detect a 10 % relative change in total cholesterol concentrations (assuming an SD of 10 %) between groups for pair-wise comparisons with a two-sided p<0.05. Nineteen subjects did not complete the study and a further 10 were not included in the per protocol analysis (five in the 30 g ISP group (n=34 analysed), four in the 30 g casein group (n=36 analysed), nine in the 50 g ISP group (n=31 analysed) and 11 in the 50 g casein group (n=29 analysed)). In addition, all subjects who received the intervention for four weeks or longer were included in a modified intention-to-treat analysis. In the per-protocol analysis there was a statistically significant difference in the decrease in total and LDL-cholesterol concentrations between the two ISP and the two casein groups, as pooled together (difference in change: -0.24, 95 % CI -0.43, -0.04 and -0.26, 95 % CI -0.43, -0.09, p=0.01 for the interaction between time and treatment). The modified intention-to- treat analysis yielded similar results. The interaction between time and dose was not significant. The Panel notes that separate analyses for the different groups as compared to the respective controls were not reported. The Panel notes that this study shows an effect of soy isoflavones in soy protein on blood cholesterol concentrations.
In a double-blind, randomised, controlled parallel intervention, 156 subjects (62 females; 20-70 years), with baseline LDL-cholesterol concentrations between 3.6 and 5.2 mmol/L after a one-month run-in period on a National Cholesterol Education Program (NCEP) Step I diet, and a further month on casein beverages (in addition to the NCEP Step I diet), were allocated to consume daily one of the following: 25 g of soy protein with either 3 mg isoflavones (ISP 3; n=28 analysed), 27 mg isoflavones (ISP 27; n=27 analysed), 37 mg isoflavones (ISP 37; n=30 analysed) and 62 mg isoflavones (ISP 62; n=30 analysed) or 25 g casein (n=31 analysed) for nine weeks (Crouse et al., 1999). The study was designed to enrol 30 participants per group to evaluate 25 (assuming drop-outs) and have 95 % power to detect a 6 % relative change in LDL-cholesterol concentrations between groups for pair-wise group comparisons at the 5 % two-sided level of significance. Twelve subjects (three in the casein, four in the ISP 3 and five in the ISP 27 group) did not complete the study. Plasma lipids were measured at weeks 8 and 9, and an average of these two measurements was taken for analysis. The analysis was reported to have been performed on an intention-to-treat basis, although from the information provided it appeared that data from ten subjects were not taken into account in the analysis. Only the ISP 62 group significantly reduced total cholesterol concentrations, i.e. by 4 % (-0.25 mmol/L; 95 % CI -0.49, -0.0003), and LDL-cholesterol concentrations, i.e. by 6 % (-0.27 mmol/L; 95 % CI -0.47, -0.06) as compared to casein. There was a statistically significant dose-response relationship between the intake of isoflavones and a decrease in total and LDL-cholesterol concentrations (p=0.01 and p=0.02 for the trends for total and for LDL-cholesterol concentrations, respectively). No effect on either total or LDL-cholesterol concentrations was
observed for ISP 3 or ISP 27. The Panel notes that this study shows an effect of soy isoflavones in soy protein on blood cholesterol concentrations.
In a double-blind, randomised controlled parallel trial, 81 post-menopausal women (mean age: approx. 60 years) with baseline plasma cholesterol concentrations between 6.2 and 7.8 mmol/L and mean baseline blood non-HDL-cholesterol concentrations of approx. 1.4 mmol/L, after having followed a 2-week run-in period on an NCEP Step I diet, were studied during three separate 24-week cohorts in which subjects were assigned to consume daily, in the context of an NCEP Step I diet, 40 g ISP containing 56 mg aglycone isoflavones (ISP 56; n=23 analysed), 40 g ISP containing 90 mg aglycone isoflavones (ISP 90; n=21 analysed) or 40 g casein and non-fat dry milk in the form of baked goods or ready-to-mix beverages and soups (Baum et al., 1998). Seven subjects dropped out during the intervention, and a further eight were not taken into account in the analysis (group not reported). No statistically significant differences were observed with respect to total cholesterol concentrations for either of the two ISP groups compared to placebo at any time point. Non-HDL cholesterol concentrations were significantly different between the ISP 56, the ISP 90 and the control groups (adjusted (for baseline means) mean difference: -0.28 and -0.25 mmol/L; p=0.03 and p=0.04, respectively) at the end of the study. The Panel notes that missing data were not taken into account in the analysis, and that this study shows an effect of soy isoflavones in soy protein on blood non-HDL cholesterol concentrations.
In a double-blind, randomised cross-over study, 23 post-menopausal women (mean age: approx. 57 years) with mean baseline blood LDL-cholesterol concentrations of approx. 3.5 mmol/L consumed daily 63 g ISP with 7.1±1.1 mg isoflavones (control diet), 65±11 mg isoflavones (low isoflavone diet), and 13±22 mg isoflavones (high isoflavone diet) for periods of 93 days each with a 26-day wash-out period in between (Wangen et al., 2001). Four subjects dropped out during the first diet period, and a further one was excluded from the analysis (18 subjects were analysed for the control and high isoflavone diet and 17 for the low isoflavone diet). Repeated measures analysis of variance showed a statistically significant difference between the three diets with respect to LDL-cholesterol concentrations (p=0.01), but not with respect to total cholesterol concentrations. The Panel notes that missing data were not taken into account in the analysis, and that this study shows an effect of soy isoflavones in soy protein on blood LDL-cholesterol concentrations.
In a double-blind, randomised cross-over trial, Hermansen et al. (2001) studied the effect on blood cholesterol concentrations of consuming 50 g ISP with >165 mg soy isoflavones and 20 g soy cotyledon fibre as compared to 50 g casein and 20 g cellulose for six weeks with a three-week wash-out period in between in 25 type 2 diabetic subjects of whom 20 completed the intervention (6 women, mean age approx. 64 years, mean baseline LDL-cholesterol concentrations: approx. 3.6 mmol/L). Analysis was performed in the population of completers. The percentage mean treatment difference between the two periods was statistically significant for LDL-cholesterol concentrations (10±15 %, p<0.05), but not for total cholesterol concentrations. The Panel notes that missing data were not taken into account in the analysis, and that this study shows an effect of soy isoflavones in soy protein on blood LDL-cholesterol concentrations.
In a double-blind, randomised, controlled parallel intervention, 92 men (23-74 years) with baseline total cholesterol concentrations between 5.69 and 7.76 mmol/L were allocated, after a three-week run-in period on an NCEP Step I diet, to consume daily either 50 g ISP with 95 mg isoflavones (n=15 analysed), 40 g ISP, 10 g casein and 76 mg isoflavones (n=17 analysed), 30 g ISP, 20 g casein and 57 mg isoflavones (n=18 analysed), 20 g ISP, 30 g casein and 38 mg isoflavones (n=15 analysed) or 50 g casein (n=16 analysed) in the form of baked goods or ready-to-mix beverages for six weeks in three separate cohorts (Teixeira et al., 2000). Eight subjects dropped out during the study, and a further three were not taken into account in the analysis (groups not reported). Non-HDL cholesterol concentrations were significantly different in all ISP groups compared to control at six weeks (adjusted (for baseline) mean change: -0.182 (50 g ISP), -0.095 (40 g ISP), -0.163 (30 g ISP),
-0.139 (20 g ISP) vs. +0.224 (control) mmol/L, p<0.05). Total cholesterol changes were significant for the 50 g, 30 g and 20 g ISP groups (but not for the 40 g ISP group) as compared to control at six weeks (adjusted mean change: -0.167 (50 g ISP), -0.115 (30 g ISP), -0.126 (20 g ISP) vs. 0.222 (control) mmol/L, p<0.05). The Panel notes that missing data were not taken into account in the analysis and that this study shows an effect of soy isoflavones in soy protein on blood cholesterol concentrations.
The Panel notes that out of the 23 human intervention studies in which soy isoflavones were consumed in soy protein from which conclusions could be drawn for the scientific substantiation of the claim, 14 studies (n=1,286 subjects, 12-93 subjects per group/period) with durations between one and 12 months using isoflavone doses of around 60-330 mg per day did not show an effect of soy isoflavones on blood cholesterol concentrations in subjects with slightly elevated or high blood LDL-cholesterol concentrations, whereas eight studies (n=712 subjects, 15-42 subjects per group/period) with durations between six weeks and six months using isoflavone doses of around 30-185 mg per day showed a statistically significant effect in subjects with normal to high blood LDL-cholesterol concentrations, and one study (n=94 subjects, around 30 per group) led to inconsistent results with respect to the effect of soy isoflavones on blood LDL-cholesterol concentrations. The Panel also notes that most of the studies had some methodological limitations, and that the inconsistent results reported appear unrelated to the dose of isoflavones used, the study duration, the sample size or the baseline characteristics of subjects with respect to blood cholesterol concentrations.
In weighing the evidence, the Panel took into account that one meta-analysis of nine RCTs and three additional RCTs did not show an effect of extracted soy isoflavones on blood cholesterol concentrations and that the evidence provided by 23 RCTs in which soy isoflavones were consumed in soy protein is inconsistent.
The Panel concludes that a cause and effect relationship has not been established between the consumption of soy isoflavones and maintenance of normal blood LDL-cholesterol concentrations.
Warunki i możliwe ograniczenia stosowania oświadczenia
Consumption of at least 25 g of soya protein (or 40-90 mg soy isoflavones) a day, as part of an overall diet low in saturated fat