Journal of Clinical Endocrinology & Metabolism, doi:10.1210/jc.2004-2350
The Journal of Clinical Endocrinology & Metabolism Vol. 91, No. 3 772-780
Copyright © 2006 by The Endocrine Society
A Critical Evaluation of the Role of Soy Protein and Isoflavone Supplementation in the Control of Plasma Cholesterol Concentrations
Antonella Dewell,
Piper L. W. Hollenbeck and
Clarie B. Hollenbeck
Department of Nutrition and Food Science (A.D., C.B.H.), San Jose State University, San Jose, California 95192-0058; and Department of Vascular Surgery (P.L.W.H.), Veterans Affairs Medical Center, San Francisco, California 94121-1598
Address all correspondence and requests for reprints to: Clarie B. Hollenbeck, Ph.D., Department of Nutrition and Food Science, San Jose State University, One Washington Square, San Jose, California 95192-0058. E-mail: clariebh{at}casa.sjsu.edu.
 |
Abstract
|
|---|
Context: The purpose of this review was to critically evaluate current research on the effect of soy protein and isoflavone supplements on plasma lipoproteins and place the potential role of soy in the prevention of coronary artery disease (CAD) into a clinical perspective.
Evidence Acquisition: An extensive literature search was performed using a variety of medical and scientific databases including Medline, PubMed, Science Direct, Ovid, NIST, and Infotrac to identify relevant articles. Journal articles were cross-referenced for additional sources of information. Articles were evaluated based on level of experimental control as well as statistical, quantitative, and clinical analysis.
Evidence Synthesis: Soy and soy isoflavones have been the object of extensive research investigating their potential hypocholesterolemic effects and possible role in the prevention of CAD. It has been suggested that soy, especially the isoflavones contained in soy, improves lipoprotein levels, thus reducing the risk for CAD. This belief, however, is not uniformly accepted. Moreover, the experimental evidence in support of this notion is not as overwhelming as generally perceived, and the current available data reveal that the discrepancies observed are primarily statistical in nature rather than reflecting actual quantitative differences in the hypocholesterolemic effects detected.
Conclusions: A critical analysis of the investigations to date indicates the data are not quantitatively impressive and raises substantial questions about the clinical importance of the hypocholesterolemic effects observed.
|
Introduction
|
|---|
THE ROLE THAT soy and soy products play in reduction of coronary artery disease (CAD) remains controversial (1, 2, 3, 4). The suggestion that soy may have a hypocholesterolemic effect dates back to 1940 when Meeker and Kesten (5) first observed that rabbits fed raw soybean flour did not develop hypercholesterolemia, compared with rabbits fed casein as a control. These findings are consistent with studies in rhesus monkeys (6). In addition, epidemiological observations in humans suggest that in Asian populations, soy intake is associated with lower serum cholesterol levels (7).
Several mechanisms have been proposed for the potential hypocholesterolemic effect of soy (8). There is evidence that the amino acid composition of soy protein and some of the nonprotein components, such as saponins, isoflavones, and phytic acid, may affect serum cholesterol. What remains to be clarified is whether and to what extent these are responsible for the hypocholesterolemic effects observed (8).
In a review of the health effects of soy in their commentary on the Fourth International Symposium on the Role of Soy in Preventing and Treating Chronic Disease, Messina et al. (2) concluded that "the consumption of even 10 g of isoflavone-rich soy protein per day may be associated with health benefits." In 1999 the Food and Drug Administration approved a claim petition stating that "diets low in saturated fat and cholesterol that include 25 g of soy protein a day may reduce the risk of heart disease" (9). A year later the American Heart Association revised their dietary guidelines to recommend the consumption of "soy protein-containing isoflavones, along with other heart-healthy diet modifications, for those high-risk populations with elevated total and LDL-cholesterol" (10). However, the results of controlled clinical trials in human populations have been inconsistent and raise serious questions regarding the hypothesis that soy (11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29) and/or isoflavones (30, 31, 32, 33, 34, 35, 36, 37, 38) lower serum cholesterol in a clinically relevant way.
The purpose of this review was to critically evaluate the current available data on the effect of both soy protein and isolated isoflavones on cholesterol metabolism as well as the interpretation of these data in the discussion of the role of soy in the prevention of CAD.
|
What Are Isoflavones?
|
|---|
Isoflavones are a class of phytoestrogens, a group of nonsteroidal plant chemicals with estrogen-like activity. The chemical structures of 17ß-estradiol and equol (a phytoestrogen metabolite) are so similar that they are virtually superimposable (Fig. 1
). Specifically, the presence of the phenolic ring and the distance between the hydroxyl groups, which is nearly identical, are considered prerequisites for estrogen binding (39). The estrogenic activity of genistein and daidzein, the predominant isoflavones in soy, is 10–2 to 10–3 that of 17ß-estradiol (40). However, their concentration in the plasma in individuals consuming the amount of soy present in the traditional Japanese diet (50–80 mg/d) can be 100 times higher than the concentration of endogenous estrogens (41). Thus, plasma concentrations of these phytoestrogens would put them in a range consistent with endogenous estrogen effects. It is this chemical and structural similarity to endogenous estrogens that led to the hypothesis that isoflavones may be responsible for the hypocholesterolemic effect of soy (39).
View larger version (19K):
[in this window]
[in a new window]
|
FIG. 1. Comparison of the chemical structures of estradiol and the isoflavone metabolite equol showing their nearly superimposable characteristics. [With permission of The Journal of Nutrition. Setchell K, Cassidy A 1999 Dietary isoflavones: biological effects and relevance to human health. J Nutr 129(Suppl):758S–767S (41 ).]
|
|
Together with isoflavones, there are two other main classes of phytoestrogens: lignans and coumestans (42). Whereas lignans and coumestans do not contribute significantly to dietary phytoestrogen intake, isoflavones are widespread in leguminous plants and are present in highest amounts in soybeans (43). Isoflavones are associated with the protein fraction (41); hence, they are present only in the whole soybean and other high-protein secondary products (Table 1) (44, 45). Whole soybeans contain the highest concentrations of isoflavones, which becomes progressively lower with increasing degree of processing. The aqueous processing of tofu and the dilution of soy protein with other ingredients in soy milk and second-generation soy products explain the reduced isoflavone content present in these products (46). In a practical sense, it is important to note that whole soybeans are not widely consumed in the Western diet. The most commonly available products in the market place targeting consumers are soy milk, tofu, and second-generation soy foods, which contain as little as 2% of the isoflavone content of whole soybeans. For example, a recent study identified soy milk, soy meat alternatives, and tofu as three of the top five selling soy foods (47). Finally, soy protein concentrates generally contain insignificant amounts of isoflavones because they are obtained through alcohol extraction, a technique that removes most of the isoflavones as well as other important alcohol-extractable hypocholesterolemic phytochemicals from soy (44, 46).
|
The Effect of Soy on Lipoprotein Cholesterol Concentrations
|
|---|
The investigation of the potential cholesterol-lowering properties of soy has followed two approaches. Some researchers have examined soy protein, mostly as soy protein isolate (11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29), whereas others have studied purified isoflavone supplements (30, 31, 32, 33, 34, 35, 36, 37, 38). A pivotal paper in this field, responsible for much of the interpretation of the data on the effect of soy protein on cholesterol metabolism, is an often-referenced meta-analysis of 38 studies, performed from 1967 to 1994 (12). This paper concludes that daily consumption of 31–47 g of soy protein can achieve serum total (TC) and low-density-lipoprotein (LDL) cholesterol reductions of 9.3 and 12.9%, respectively (12). However, when the participants are categorized according to baseline cholesterol concentration, it becomes clear that the changes in cholesterol concentrations were dependent on initial cholesterol concentrations. Figure 2 illustrates the changes in TC and non-high-density lipoprotein (HDL)/LDL-cholesterol according to baseline concentrations from this meta-analysis. It is clear (Fig. 2A) that the cholesterol-lowering effects observed in individuals with baseline TC concentrations less than 260 mg/dl (6.7 mmol/liter) were very modest (3.3–4.4%). Slightly higher reductions in serum TC (7.4%) were reported in those individuals with baseline TC between 260 and 335 mg/dl (6.7–8.7 mmol/liter), whereas reductions in TC concentrations greater than 10% were realized only in individuals with initial TC greater than 335 mg/dl (8.7 mmol/liter). The changes in LDL-cholesterol are more difficult to discuss because some report non-HDL-cholesterol, a derived value, whereas others reported LDL-cholesterol, an estimated value (48). However, as Fig. 2B illustrates, the changes in non-HDL/LDL-cholesterol closely mirror those in TC concentrations.
View larger version (27K):
[in this window]
[in a new window]
|
FIG. 2. Changes in plasma cholesterol (A) and LDL cholesterol (B) as they relate to initial cholesterol levels in a meta-analysis of clinical trials with soy protein (12 ). Data are presented both as mg/dl (top and right axis) and mmol/liter (bottom and left axis).
|
|
It is important to point out that a major limitation of this approach to data analysis lies in summarizing data from studies with very different research designs and study populations. The studies evaluated in this meta-analysis involved mostly men and some premenopausal women and children. More importantly, changes in other dietary components such as saturated fat and cholesterol intake were also present in several of the studies and could have explained the lower serum cholesterol concentrations. Finally, as noted by other reviewers (25, 49) 77% of the studies had 95% confidence intervals that encompassed zero; as a result, these findings should be viewed with a great deal of caution. Nevertheless, the general results of this meta-analysis are consistent with findings that reductions in serum cholesterol are dependent on initial cholesterol concentrations (11, 50). A more recent meta-analysis of 10 studies published between 1995 and 2002 including 959 individuals with baseline cholesterol concentrations between 209 and 255 mg/dl (5.42 and 6.60 mmol/liter) and more homogeneity in experimental designs showed a weak correlation between changes in LDL-cholesterol concentrations and soy associated isoflavone (r = –0.33; P = 0.14). An average of 52 mg soy-associated isoflavones in 36 g daily soy protein resulted in a decrease in LDL-cholesterol by 4% (6.6 mg/dl or 0.17 mmol/liter) and an increase in HDL-cholesterol by 3% (1.2 mg/dl or 0.03 mmol/liter) (51). Thus, it appears from these two metaanalyses that the hypocholesterolemic effects of soy are relatively modest in individuals with plasma cholesterol concentrations less than 335 mg/dl (8.7 mmol/liter).
Finally, the marked increase in the hypocholesterolemic effects of soy in those individuals with TC concentrations greater than 335 mg/dl (Fig. 2
) may be due, in part, to the fact that the number of studies in this category is relatively small and represent studies from a single group of investigators (52), and these investigators believe that there may be differences in the protein components among various cultivars that may explain some of the variability among studies. The potential role of variations in soy proteins in the hypocholesterolemic effects of soy is discussed in more detail below.
|
Unprocessed Soy Protein
|
|---|
Subsequent to the publication of the 1995 meta-analysis (12), investigations have focuses on the consumption of soy mostly as soy protein isolate incorporated into foods or drinks, ranging from 20 to 63 g soy protein per day in men (18, 28); mixed populations (15, 21, 23, 24, 25); and premenopausal (17), perimenopausal (16), and postmenopausal women (13, 14, 19, 20, 22, 26, 27, 29). The results of these studies are summarized in Table 2. Of the 17 studies assessing the effects of soy protein on serum lipoproteins, six (35%) reached statistical significance for TC, whereas eight (47%) reached statistical significance for non-HDL/LDL-cholesterol. Although the majority of these studies showed no significant difference (15, 17, 19, 20, 23, 26, 27, 28, 29), some have reported statistically significant but quantitatively small reductions in non-HDL/LDL and/or TC, ranging from 1.8 to 10% (13, 14, 15, 16, 17, 18, 22, 24, 25), whereas some showed small increases in HDL-cholesterol (13, 20, 23, 25).
These studies have investigated both normocholesterolemic and hypercholesterolemic populations, and, in general, their results confirm previous observations that the greatest reductions are reported in the most hypercholesterolemic populations (Table 2). However, this is not universally true. For example, Cuevas et al. (26) found no significant reductions in either TC or LDL in 18 postmenopausal women with baseline TC of 286 mg/dl (7.4 mmol/liter), the highest of the studies presented in Table 2. Similar findings were reported by Blum et al. (27) in a study with 25 hypercholesterolemic postmenopausal women. Jenkins et al. (24), on the other hand, reported statistically significant, but quantitatively small, reductions of TC and LDL in a population with mean baseline cholesterol concentrations of 261 mg/dl (6.75 mmol/liter). Baum et al. (13) and Potter et al. (14) in a study published in slightly different forms reported a statistically significant decrease in non-HDL-cholesterol (7–9%) and increase in HDL-cholesterol (5%) in moderately hypercholesterolemia subjects (mean TC 255 mg/dl or 6.6 mmol/liter) receiving 40 g soy protein daily as compared with the control group receiving casein. It is important to note, however, that there were no statistically significant differences in plasma cholesterol concentrations between the two groups at the beginning of the study periods (251 vs. 243 mg/dl TC; 197 vs. 189 mg/dl non-HDL), and the cholesterol concentrations were essentially identical at the end of the study period (237 vs. 235 mg/dl TC; 182 vs. 184 mg/dl non-HDL) (13). The apparent statistical differences were achieved only by subtracting the baseline values. The statistical appropriateness of these data manipulation given the concurrent parallel design used in the study is questionable. This study also raises questions of a more practical nature. For example, if one accepts this statistical manipulation as valid, then one is left with the problem of interpreting the clinical relevance of the statistical significance in a group of individuals whose plasma cholesterol concentrations were not significantly different at the beginning of the treatment period and are essentially identical at the completion of the study. Finally, even if one disregarded the statistical handling of the data, the magnitude of the reduction in non-HDL-cholesterol was still very modest (from 197 to 182 mg/dl, i.e. 15 mg/dl) and would be of little clinical significance in terms of reducing the risk of developing CAD. In two recent studies, one in a mixed population (21) and the other in hypercholesterolemic postmenopausal women (19), providing 40–42 g soy protein per day showed quantitatively similar changes (
10% reduction) in plasma lipoproteins as those studies discussed above but failed to reach statistically significance. In the study by Cuevas et al. previously discussed (26), 40 g soy protein incurred even greater apparent changes (16–18%) that, again, did not reach statistical significance. These results reflect the great variability of response among individuals and may partially explain the achievement of statistical significance by some studies and the failure to do so by others. It can be seen by reviewing those studies using unprocessed soy protein that the differences observed are not in the magnitude of the hypocholesterolemic response but rather in whether or not they achieved statistical significance.
|
Alcohol-Extracted Soy Proteins
|
|---|
One confounding variable in studies using soy proteins, often missed in drawing conclusions about the effect of soy on serum cholesterol (7, 17, 22, 23), is the nature of the soy protein used in control populations. In the attempt to determine whether isoflavones are the component in soy responsible for its potential hypocholesterolemic effect, six studies compared soy protein to soy foods with different amounts of isoflavones using alcohol extraction to obtain soy protein with lower or zero isoflavone content (15, 17, 19, 22, 23, 24). This technique, however, is known to deprive soy of not only isoflavones but also other components associated with the soy protein that are known to lower cholesterol concentrations (e.g. saponins, phytic acid, and other alcohol-extractable phytochemicals) (43, 45). For this reason, one cannot exclude the possibility that these components may have played an important role in the comparative hypocholesterolemic effects observed in these studies. A study by Gardner et al. (19) provides insights into the possibility that alcohol extraction of soy protein might have a deleterious effect that may account for the apparent lowering of cholesterol when compared with native soy protein. Specifically, these investigators assessed the effects of soy protein, alcohol-extracted soy protein, and a milk protein in a parallel study design. The results showed no significant difference between the native soy protein and the milk protein on plasma cholesterol concentrations. However, when native soy protein was compared with alcohol-extracted soy protein, the alcohol-extracted soy resulted in significantly greater cholesterol concentrations than the native soy protein. Certainly, this is an important point that needs to be resolved before we can unequivocally conclude that soy protein has a hypocholesterolemic effect as a result of its comparison to alcohol-extracted soy protein.
Thus, the available evidence does not appear to support the conclusion that the differences observed between intact soy protein and alcohol-extracted soy protein are attributable solely to the isoflavone component associated with the intact soy protein (7, 15, 17, 19, 22, 23). Although one investigator (15) clearly acknowledged the use of alcohol extraction as a confounding factor, this has either been missed or not discussed by others (7, 17, 19, 22, 23).
|
Variations in the Composition of Soy Protein as a Mechanism of Variability
|
|---|
It has been suggested that the specific amino acid composition of soy protein may be the responsible agent found in soy (52, 53). Early studies in rabbits (54, 55) provided evidence that amino acids lysine and methionine have moderately hypercholesterolemic effects, whereas arginine tended to lower cholesterol concentrations. Because soy protein has a higher ratio of arginine to lysine and methionine, it has been suggested that this may explain, at least in part, the hypocholesterolemic effects of soy protein (53). Studies in rabbits (54, 55) and gerbils (56) using amino acid patterns of soy protein lend support to this notion but are unable to explain all of the effects observed with soy proteins. Recently a series of studies by Lovati and coworkers (57, 58, 59, 60, 61) suggest that the hypocholesterolemic effects of soy proteins may be mediated through an increased clearance of LDL-cholesterol by up-regulation of LDL receptors on mononuclear cells (61) and hepatocytes (58, 59, 60, 61), which may be down-regulated in individuals with hypercholesterolemia. More specifically, they have identified a polypeptide sequence found in the 7S globulin protein present in soy (the 
'-subunit, corresponding to residues 127–150), which was capable of up-regulating LDL receptors on Hep G2 cells in vitro. Furthermore, they suggest that this mechanism may explain why the effects of soy proteins on LDL-cholesterol concentrations are substantially less in normocholesterolemic individuals and appear greater with increased levels of hypercholesterolemia (assuming, or course, that the down-regulation of LDL receptors is proportional to the degree of hypercholesterolemia). They also suggest, indirectly, that difference in the concentration of this storage protein (7S globulin) among various cultivars of soy may also be responsible for the differential effects reported on LDL-cholesterol.
As provocative as these data are, for this mechanism to function in vivo, the polypeptide sequence would have to be absorbed and presented to hepatocytes and mononuclear cells as the intact polypeptide sequence. Although this would seem to be highly unlikely, it has recently been suggested that proteins and polypeptide fragments may indeed be absorbed intact through enterohepatic circulation (62). At the present time, however, this should be considered very speculative, and much more research is needed before the role of soy and/or soy proteins on cholesterol regulation is fully understood.
|
Soy Protein as Part of an Overall Cholesterol-Lowering Diet
|
|---|
Before leaving the issue of the effects of dietary soy on plasma cholesterol, it would seem important to address the issue of including soy as part of a general cholesterol-lowering diet. Jenkins et al. (63, 64) recently demonstrated that a so-called portfolio approach in the dietary treatment of hypercholesterolemia was nearly as effective as low dose (20 mg) lovostatin (a first-generation statin) in achieving a clinically meaningful reduction in LDL-cholesterol concentration in healthy hyperlipemic men and women, and although there were significant differences in the hypocholesterolemic effects between the portfolio diet (–29.6%) and the statins (–33.3%) at 4 wk, the differences were not quantitatively impressive. The portfolio dietary approach to cholesterol reduction is one that combines a switch from animal-based proteins to plant-based proteins (the adoption of a vegetarian diet), which includes an increase in soy protein, along with cholesterol-lowering dietary changes such as supplementation with plant sterols, and viscous fiber as well as reductions in saturated fat and cholesterol (63, 64). However, it is difficult in the context of the overall dietary changes made in the portfolio diet to assess the importance of soy in achieving the observed results. Additionally, it is also important to realize that the portfolio diet does not represent a solely dietary approach to cholesterol reduction and includes supplementation not achievable by dietary changes alone. Quantitatively similar lowering of LDL-cholesterol (–24%) was reported by Gerhard et al. (65) with diets low in saturated fat and cholesterol and higher in dietary fiber, without the supplementation of soy, plant sterols, or viscous fiber. The question germane to the present review remains, What are the lipid-lowering effects that can be attributed to soy proteins or isolated soy isoflavones? Jenkins et al. (63) concluded that the individual contributions of soy protein, plant sterols, viscous fiber, and almond in the portfolio diet were in the range of 5–7%. This assessment of the contribution of soy protein is consistent with most of the studies that form the basis of this review. They are also consistent with the reported lipid-lowering effects of lean animal proteins (66, 67, 68). In both short-term (66, 67) and long-term (68) studies, the introduction of lean beef, lean fish, and skinless poultry have also been shown to result in reductions of between 5 and 9% in LDL-cholesterol. Thus, it appears from the available information that the hypocholesterolemic effects of soy are comparable with those achieved by dietary reductions in saturated fat and cholesterol or by switching to lean animal proteins.
|
Isolated Soy Isoflavones
|
|---|
In contrast to those studies using soy proteins, investigations using purified isoflavone supplements have consistently reported no significant effects on serum lipoproteins (30, 31, 32, 33, 34, 35, 36, 37, 38). These studies have used a range of isoflavones (40–150 mg/d) and examined their effect on both normocholesterolemic and hypercholesterolemic men and women (30, 31, 32, 33, 34, 35, 36, 37, 38). Of the eight studies using isolated soy isoflavones summarized in Table 2, none of the reported changes in either TC or non-HDL/LDL-cholesterol concentrations reached statistical significance.
In a key crossover study with 19 mildly hypercholesterolemic postmenopausal women, 80 mg isoflavones per day failed to significantly lower serum lipoproteins (31). The authors observed a downward trend in LDL (6%) and an upward trend in HDL (4%), resulting in a nonsignificant reduction (10%) in the LDL to HDL-cholesterol ratio between the placebo and treatment values. However, all these differences, including total cholesterol (3%), were quantitatively small and not statistically significant. In a study of 36 hypercholesterolemic postmenopausal women, a much higher dosage of isoflavones (150 mg/d) for 6 months resulted in quantitatively similar reductions in total and non-HDL cholesterol, which also failed to reach statistical significance (37).
In contrast to all of the major studies using isoflavone extracts, there is a single study (36) that recently reported a significant increase in HDL-cholesterol of 13–22% in postmenopausal women consuming different doses of isoflavones (28–85 mg/d). The relevance of these findings, however, is difficult to interpret, given the absence of a control group in the study design and the fact that these changes did not return to baseline at the end of a 2-month washout period following the treatment phase. It is also important to note that even in this study (36), there were no significant changes to either total cholesterol or LDL-cholesterol at the end of 6 months of treatment. Therefore, the results of controlled experimental studies consistently report the absence of any significant effect of purified isoflavone supplements on plasma lipids and lipoproteins. However, fasting plasma cholesterol concentrations in these studies were generally lower than cholesterol concentrations in the studies using soy protein, and because plasma cholesterol-lowering effects appear to be related to initial cholesterol concentrations, it could be argued that these individuals might not be expected to experience clinically significant reductions. On the other hand, this population (cholesterol between 212 and 230 mg/dl) is at considerable risk for development of CAD, and it is precisely this population that the use of soy as a preventative treatment may have the greatest utility.
Finally, it has been suggested that the potential beneficial effects of soy may be related to the presence of the metabolite equol, produced by intestinal microflora from daidzein, a major isoflavone in soy (69). Equol has been shown to possess higher affinity for the estrogen receptors than the parent isoflavone daidzein. It appears that about 50–70% of the general population can produce equol, possibly due to differences in microflora species in the large intestine (69). Therefore, the estrogenic effects of soy may be stronger in a subpopulation of equol producers. However, to date, there is no published peer-reviewed evidence to support this notion in regard to the effect of soy isoflavones on lipid profiles (70). On the other hand, Nestel et al. (71), demonstrated significant lowering of LDL-cholesterol (9.5%) in men, but not in women, with 40 mg of biochanin (a precursor of genistein)-enriched isoflavones but not formononetin (a precursor of daidzein)-enriched isoflavone isolated from red clover. They concluded that different isoflavones might produce different responses, depending on gender, which may explain the failure to demonstrate significant differences in studies predominantly in women using mixed isoflavones. However, similar evidence does not exist for soy-derived isoflavones.
In summary, it is important to note that the changes reported in the studies using purified isoflavone supplements (30, 31, 32, 33, 34, 35, 36, 37, 38), although not statistically significant, are quantitatively similar to those observed in the soy protein studies (Table 2). Therefore, it appears that the differences between the soy-based studies and the investigations of isoflavone extracts may be more apparent than real, and the discrepancies are merely statistical in nature, rather than reflecting actual quantitative differences in the magnitude of the data (72).
|
Comparative Effects of Isolated Soy Isoflavones as a Therapeutic Hypocholesterolemic
|
|---|
The supplementation with isolated soy isoflavones represents a pharmacologic approach to the management of hypercholesterolemia, and, therefore, comment on the relative effectiveness of this method, compared with more conventional pharmacological approaches, would be appropriate. Isolated soy isoflavones have been introduced in amounts from 28 to 150 mg and have achieved nonstatistical reductions in LDL-cholesterol in range of between 1 and 6.5% (Table 2). Statins, on the other hand, generally have been associated with reductions in LDL-cholesterol of between 30 and 60%, depending on dose and agent used (73, 74). In a large systematic review and meta-analysis, Law et al. (73) reported on data from 164 randomized, placebo-controlled clinical trials using six statins (atrovastatin, lovostatin, rosuvastatin, fluvastatin, pravastatin, and simvastatin). The results of this analysis indicated reductions in LDL-cholesterol ranging from 10 to 38% at low dose (5 mg/d) and between 38 and 58% at high dose (80 mg/d), with a 5–8% lowering in LDL-cholesterol for every 10-mg increase in dose. It is important to note that the percent lowering of LDL-cholesterol reported by Law et al. (73) was also dependent on initial baseline cholesterol concentrations, a response that appears to be common among studies and treatment modalities. Quantitatively similar lowering of total cholesterol and LDL-cholesterol (30–60%) has also been reported by bile binding resins, with lesser effects (5–15%) demonstrated with other therapeutic hypocholesterolemic agents (nicotinic acid, fibrates) (74, 75).
|
Statistical Significance vs. Clinical Significance
|
|---|
In light of the preceding discussion in regard to the statistical nature of the differences observed, it is important to underscore the fact that statistical significance does not provide an evaluation of the clinical importance or significance of the finding, a point that has not been discussed in many of the studies cited above. It is the clinical interpretation of the quantitative differences that have been demonstrated that needs to assume primary importance in the discussion and implementation of any conclusion from this area of research.
Although there is broad general agreement on the establishment of levels of statistical significance, the definition or criteria for establishment of clinical significance are undoubtedly more complex, would include several criteria including baseline cholesterol concentrations, and certainly vary among clinicians. It is not our intention to establish these criteria here; rather we have provided actual quantitative changes in cholesterol concentrations to allow individuals to evaluate the clinical importance of changes observed. In this regard, even in the best-designed studies, the magnitude of the changes reported remains relatively small, ranging from 1.8 to approximately 9%, with the greatest reductions observed in the most hypercholesterolemic populations. For example, the two studies reporting the greatest statistically significant reductions (9%) in non-HDL/LDL-cholesterol also had study populations with the highest baseline cholesterol concentrations (13, 14, 24). In one study, non-HDL-cholesterol was lowered from a mean of 199 to a mean of 184 mg/dl (5.2 to 4.8 mmol/liter) after 6 months of treatment (13, 14); in the other, LDL decreased from 175 to 160 mg/dl (4.5 to 4.1 mmol/liter) (24). Given that the National Cholesterol Education Program guidelines (75) recommend a desirable level of LDL-cholesterol less than 100 mg/dl (2.58 mmol/liter) and the concentrations of non-HDL/LDL-cholesterol after treatment in these studies (13, 14, 24) (184 and 160 mg/dl or 4.8 and 4.1 mmol/liter, respectively), it would be reasonable to conclude that the effects of these declines in non-HDL/LDL-cholesterol concentrations would be of little clinical significance in terms of a reduction in the risk of developing CAD.
|
Potential Adverse Effect of Soy Intakes
|
|---|
It is also necessary to point out that not all of the evidence demonstrates beneficial effects from increased soy consumption; the potential adverse effects of soy have been well described (50, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86). Soybeans are rich in antinutrients including protease inhibitors (trypsin and chymotrypsin), lectins (hemaggluttinins), goitrogens, phenolic compounds (tannins and phytoestrogens), phytates, saponins, and antivitamins (to vitamins A, B12, D, and E) (76). These antinutrients have been associated with growth retardation and failure to thrive, pancreatic hypertrophy, hyperplasia and hypersecretion, pancreatic acinar adenomas, and endocrine abnormalities (76, 77, 78, 79, 80, 81, 82). Trypsin inhibitors and lectins are both growth inhibitors. Trypsin is important not only as a protease but also is responsible for the activation of other pancreatic proenzymes (zymogens), whereas lectins are capable of inhibiting growth by binding specific cell-surface receptors on small intestine epithelial cells (as well as lymphocytes) (76, 78, 79, 80, 81). Although high-temperature cooking can reduce some of these growth inhibitors, it does not completely eliminate them (76). In addition, soy has one of the highest concentrations of phytates, an antinutrient that has been shown to block the uptake of several essential elements including calcium, magnesium, copper, iron, and zinc. The phytates present in soy appear to be highly resistant to normal phytate-reducing processes (76, 80, 81).
Concentrations of soy isoflavones in the range of levels found with consumption of soy-based diets have been shown to inhibit thyroxine synthesis inducing goiter and hypothyroidism in infants fed soy-based formulas, in some cases leading to the development of autoimmune thyroid disease (81). In addition, potential harmful effects of soy isoflavones have been recently reported in adults. A recent prospective epidemiological study reported that increases in tofu consumption may lead to increased cognitive dysfunction in Japanese American men (82). Finally, high concentrations of genistein, daidzein, and other isoflavones have been reported to result in genetic abnormalities in a variety of cells including human lymphocytes, oviduct cells, and testis cells and therefore may possess potentially genotoxic effects (50). Although research in many of these areas is not yet well developed, it should raise a note of caution in terms of recommendations to dramatically increase soy consumption or encourage the supplemental use of soy.
|
Conclusion
|
|---|
A critical evaluation of the evidence currently available in the literature on the potential role of soy protein or isolated soy isoflavone supplementation for improving plasma lipoproteins indicates that the data are not quantitatively impressive and raise substantial questions about the clinical importance of the hypocholesterolemic effects. Therefore, it would appear that conclusions in regard to the hypocholesterolemic benefits of soy made by researchers (1, 2, 12) and health agencies (9, 10) are perhaps too premature to make any recommendation for their use as an alternative to established therapies in the management of hypercholesterolemia in populations at risk for CAD (87).
Soybeans are a very healthful food per se, and it is not our intent to discourage their incorporation into the diet. They are a good source of relatively complete plant protein, viscous fiber, unsaturated fat, vitamins, minerals, and phytochemicals. Their substitution for other sources of proteins would certainly increase the variety of nutrient intake in the diet. Soy products can be helpful in displacing animal foods high in saturated fat and cholesterol; however, more comprehensive dietary changes may be needed to induce clinically significant lowering of these risk factors for CAD. In a more practical sense, to reach a dietary intake of soy protein or isoflavones within the range used by the studies reviewed in this manuscript, one would have to consume about 1–2 cups of cooked soybeans or 0.5 to 2 lb of tofu each day (Table 1), a dietary change that most Americans might consider impractical.
|
Footnotes
|
|---|
The authors have no conflict of interest.
First Published Online January 10, 2006
Abbreviations: CAD, Coronary artery disease; HDL, high-density lipoprotein; LDL, low-density-lipoprotein; TC, total cholesterol.
Received December 2, 2004.
Accepted December 20, 2005.
|
References
|
|---|
- Clarkson TB 2002 Soy, soy phytoestrogens and cardiovascular disease. J Nutr 132:566S–569S
- Messina M, Gardner C, Barnes S 2002 Gaining insight into the health effects of soy but a long way still to go: commentary on the fourth international symposium on the role of soy in preventing and treating chronic disease. J Nutr 132:547S–551S
- Lichtenstein AH 2001 Got soy? Am J Clin Nutr 73:667–668[Free Full Text]
- Kerckhoffs DAJM, Brouns F, Hornstra G, Mensink RP 2002 Effects on the human serum lipoprotein profile of ß-glucan, soy protein and isoflavones, plant sterols and stanols, garlic and tocotrienols. J Nutr 132:2492–2505[Free Full Text]
- Meeker DR, Kesten HD 1940 Experimental atherosclerosis and high protein diets. Proc Soc Exp Biol Med 45:543–545
- Anthony MS, Clarkson, TB, Bullock BC, Wagner JD 1997 Soy protein versus soy phytoestrogens in the prevention of diet-induced coronary artery atherosclerosis of male cynomolgus monkeys. Arterioscler Thromb Vasc Biol 17:2425–2531
- Nagata C, Takatsuka N, Kurisu Y, Shimizu H 1998 Decreased serum total cholesterol concentration is associated with high intake of soy products in Japanese men and women. J Nutr 128:209–213[Abstract/Free Full Text]
- Potter SM 1995 Overview of proposed mechanisms for the hypocholesterolemic effect of soy. J Nutr 125:606S–611S
- Food and Drug Administration 1999 Food labeling: health claims; soy protein and coronary artery disease. Fed Regist 64:57699–57733
- The Nutrition Committee of the American Heart Association 2000 AHA dietary guidelines. Revision 2000: a statement for healthcare professionals from the Nutrition Committee of the American Heart Association. Circulation 102:2284–2299[Free Full Text]
- Carrol KK 1991 Review of clinical studies on cholesterol-lowering response to soy protein. J Am Diet Assoc 91:820–827[Medline]
- Anderson JW, Johnstone BM, Cook-Newell ME 1995 Meta-analysis of the effects of soy protein intake on serum lipids. N Engl J Med 333:276–282[Abstract/Free Full Text]
- Baum JA, Teng H, Erdman Jr JW, Weigel RM, Klein BP, Persky VW, Freels S, Surya P, Bakhit RM, Ramos E, Shay NF, Potter SM 1998 Long-term intake of soy protein improves blood lipid profiles and increases mononuclear cell low-density-lipoprotein receptor messenger RNA in hypercholesterolemic, postmenopausal women. Am J Clin Nutr 68:545–551[Abstract]
- Potter SM, Baum JA, Teng H, Stillman, RJ, Shay NF, Erdman Jr JW 1998 Soy protein and isoflavones: their effects on blood lipids and bone density in postmenopausal women. Am J Clin Nutr 58:1375S–1379S
- Crouse III JR, Morgan T, Terry JG, Ellis J, Vitolins M, Burke GL 1999 A randomized trial comparing the effect of casein with that of soy protein containing varying amounts of isoflavones on plasma concentrations of lipids and lipoproteins. Arch Intern Med 159:2070–2076[Abstract/Free Full Text]
- Washburn S, Burke GL, Morgan T, Anthony M 1999 Effect of soy protein supplementation on serum lipoproteins, blood pressure, and menopausal symptoms in perimenopausal women. Menopause 6:7–13[Medline]
- Merz-Demlow BE, Duncan AM, Wangen KE, Xu X, Carr TP, Phipps WR, Kurzer MS 2000 Soy isoflavones improve plasma lipids in normocholesterolemic premenopausal women. Am J Clin Nutr 71:1462–1469[Abstract/Free Full Text]
- Teixeira SR, Potter SM, Weigel R, Hannum S, Erdman Jr JW, Hasler CM 2000 Effects of feeding 4 levels of soy protein for 3 and 6 wk on blood lipids and apolipoproteins in moderately hypercholesterolemic men. Am J Clin Nutr 71:1077–1084[Abstract/Free Full Text]
- Gardner CD, Newell KA, Cherin R, Haskell W 2001 The effect of soy protein with or without isoflavones relative to milk protein on plasma lipids in hypercholesterolemic postmenopausal women. Am J Clin Nutr 73:728–735[Abstract/Free Full Text]
- Scheiber MD, Liu JH, Subbiah MT, Subbiah MT, Rebar RW, Setchell KD 2001 Dietary inclusion of whole soy foods results in significant reductions in clinical risk factors for osteoporosis and cardiovascular disease in normal postmenopausal women. Menopause 8:384–392[CrossRef][Medline]
- Teede HJ, Dalais FS, Kotsopoulos D, Liang Y-L, Davis S, McGrath BP 2001 Dietary soy has both beneficial and potentially adverse cardiovascular effects: a placebo-controlled study in men and postmenopausal women. J Clin Endocrinol Metab 86:3053–3060[Abstract/Free Full Text]
- Wangen KE, Duncan AM, Xu X, Kurzer MS 2001 Soy isoflavones improve plasma lipids in normocholesterolemic and mildly hypercholesterolemic postmenopausal women. Am J Clin Nutr 73:225–231[Abstract/Free Full Text]
- Sanders TAB, Dean TS, Grainger D, Miller GJ, Wiseman H 2002 Moderate intakes of intact soy protein rich in isoflavones compared with ethanol-extracted soy protein increase HDL but do not influence transforming growth factor ß1 concentrations and hemostatic risk factors for coronary heart disease in healthy subjects. Am J Clin Nutr 76:373–377[Abstract/Free Full Text]
- Jenkins DJA, Kendall CWC, Jackson CJC, Connelly PW, Parker T, Faulkner D, Vidgen E, Cunnane SC, Leiter LA, Josse RG 2002 Effects of high- and low-isoflavone soyfoods on blood lipids, oxidized LDL, homocysteine, and blood pressure in hyperlipidemic men and women. Am J Clin Nutr 76:365–372[Abstract/Free Full Text]
- Lichtenstein AH, Jalbert SM, Adlercreutz H, Goldin BR, Rasmussen H, Schaefer EJ, Ausma LM 2002 Lipoprotein response to diets high in soy or animal protein with and without isoflavones in moderately hypercholesterolemic subjects. Arterioscl Thromb Vasc Biol 22:1852–1858[Abstract/Free Full Text]
- Cuevas AM, Irribarra VL, Castillo OA, Yañez MD, Germain AM 2003 Isolated soy protein improves endothelial function in postmenopausal hypercholesterolemic women. Eur J Clin Nutr 57:889–894[CrossRef][Medline]
- Blum A, Lang N, Vidger F, Israeli P, Gumanovsky M, Lupovitz S, Elgazi A, Peleg A, Ben-Ami M 2003 Effects of soy protein on endothelium-dependent vasodilation and lipid profile in postmenopausal women with mild hypercholesterolemia. Clin Invest Med 26:20–26[Medline]
- Sagara M, Kanda T, NJelekera M, Teramoto T, Armitage L, Birt N, Birt C, Yamori Y 2003 Effects of dietary intake of soy protein and isoflavones on cardiovascular disease risk factors in high risk, middle-aged men in Scotland. J Am Coll Nutr 23:85–91
- Kreijkamp-Kaspers S, Kok L, Grobbee DE, de Haan EHF, Aleman A, Lampe JW, van der Schouw YT 2004 Effect of soy protein containing isoflavones on cognitive function, bone mineral density, and plasma lipids in postmenopausal women: a randomized controlled trial. JAMA 292:65–74[Abstract/Free Full Text]
- Nestel PJ, Yamashita T, Sasahara T, Pomeroy S, Dart A, Komesaroff P, Owen A, Abbey, M 1997 Soy isoflavones improve systemic arterial compliance but not plasma lipids in menopausal and perimenopausal women. Arterioscl Thromb Vasc Biol 17:3392–3398[Abstract/Free Full Text]
- Nestel PJ, Pomeroy S, Kay S, Komesaroff P, Behrsing J, Cameron JD, West L 1999 Isoflavones from red clover improve systemic arterial compliance but not plasma lipids in menopausal women. J Clin Endocrinol Metab 84:895–898[Abstract/Free Full Text]
- Hodgson JM, Puddey IB, Beilin LJ, Mori TA, Croft KD 1998 Supplementation with isoflavonoid phytoestrogens does not alter serum lipid concentrations: a randomized controlled trial in humans. J Nutr 128:728–732[Abstract/Free Full Text]
- Samman S, Lyons Wall PM, Chan GS, Smith SJ, Petocz P 1999 The effect of supplementation with isoflavones on plasma lipids and oxidisability of low density lipoprotein in premenopausal women. Atherosclerosis 147:277–283[CrossRef][Medline]
- Howes JB, Sullivan D, Lai N, Nestel P, Pomeroy S, West L, Eden JA, Howes LG 2000 The effects of dietary supplementation with isoflavones from red clover on the lipoprotein profiles of post menopausal women with mild to moderate hypercholesterolaemia. Atherosclerosis 152:143–147[CrossRef][Medline]
- Simons LA, von Konigsmark M, Simons J, Celermajer DS 2000 Phytoestrogens do not influence lipoprotein levels or endothelial function in healthy, postmenopausal women. Am J Cardiol 85:1297–1301[CrossRef][Medline]
- Clifton-Bligh PB, Baber RJ, Fulcher GR, Nery ML, Moreton T 2001 The effect of isoflavones extracted from red clover (Rimostil) on lipid and bone metabolism. Menopause 8:259–265[CrossRef][Medline]
- Dewell A, Hollenbeck CB, Bruce B 2002 The effects of soy-derived phytoestrogens on serum lipids and lipoproteins in moderately hypercholesterolemic postmenopausal women. J Clin Endocrinol Metab 87:118–121[Abstract/Free Full Text]
- Uesugi T, Fukui Y, Yamori Y 2002 Beneficial effects of soybean isoflavone supplementation on bone metabolism and serum lipids in postmenopausal Japanese women: a four-week study. J Am Coll Nutr 21:97–102[Abstract/Free Full Text]
- Setchell KDR 1998 Phytoestrogens: the biochemistry, physiology, and implications for human health of soy isoflavones. Am J Clin Nutr 68:1333S–1346S
- Miksiceck RJ 1994 Interaction of naturally occurring nonsteroidal estrogens with expressed recombinant human estrogen receptor. J Steroid Biochem Mol Biol 49:153–160[CrossRef][Medline]
- Setchell KDR, Cassidy A 1999 Dietary isoflavones: biological effects and relevance to human health. J Nutr 129:758S–767S
- Davis SR, Murkies AL, Wilcox G 1998 Phytoestrogens in clinical practice. Integr Med 1:27–34
- Tham DM, Gardner CD, Haskell WL 1998 Potential health benefits of dietary phytoestrogens: a review of the clinical, epidemiological, and mechanistic evidence. J Clin Endocrinol Metab 83:2223–2235[Abstract/Free Full Text]
- Wang H, Murphy PA 1994 Isoflavone content in commercial soybean foods. J Agric Food Chem 42:1666–1673
- Coward L, Barnes NC, Setchell KDR, Barnes S 1993 Genistein, daidzein, and their ß-glycoside conjugates: antitumor isoflavones in soybean foods from American and Asian diets. J Agric Food Chem 41:1961–1967
- Anderson RL, Wolf WJ 1995 Compositional changes in trypsin inhibitors, phytic acid, saponins and isoflavones related to soybean processing. J Nutr 125:581S–588S
- 2001 SPINS/Soyatech Study in Association with Arthur D. Little. Soyfoods: The U.S. Market 2001. Bar Harbor, ME: Soyatech
- Friedewald WT, Levy RI, Fredrickson DS 1972 Estimation of the concentration of low-density lipoprotein cholesterol in plasma, without use of the preparative ultracentrifuge. Clin Chem 18:499–502[Abstract]
- Nestel P 2002 Role of soy protein in cholesterol-lowering. How good is it? Arterioscl Thromb Vasc Biol 22:1743–1744[Free Full Text]
- Sirtori CR 2001 Risks and benefits of soy phytoestrogens in cardiovascular diseases, cancer, climacteric symptoms and ost
Top
Abstract
Introduction
