点击显示 收起
1 From the Division of Nephrology, Moffitt/Mount Zion General Clinical Research Center, University of California, San Francisco, San Francisco, CA 94131.
2 Address reprint requests to A Sebastian, Division of Nephrology, Moffitt/MZ General Clinical Research Center, University of California, San Francisco, 40 Crags Court, San Francisco, CA 94131. E-mail: anthony_sebastian{at}msn.com.
See corresponding article on page 916.
Isoflavones belong to the class of molecules in foods called isoflavonoids, members of a heterogeneous group of molecules (referred to as phenolics) that have the common feature of at least one hydroxyl-substituted aromatic ring system. Isoflavonoids are found almost exclusively in plants in the Leguminosae family (legume family). Of all common plant foods, soybeans appear to have by far the highest concentrations of isoflavonoids, specifically the isoflavones daidzein, genistein, and glycitein (1).
Although not steroids, isoflavones have a molecular structure that resembles that of estradiol closely enough that they have partial agonist or antagonist effects on estrogen receptors in humans (2), and thus they fall into a class of plant molecules referred to as phytoestrogens. Considerable interest has focused on the use of soy-based products, isolated soy isolates, and synthetic isoflavones as estrogen substitutes in postmenopausal women, in particular to inhibit hypoestrogenically induced loss of bone mass. In a recent review of clinical trials and epidemiologic studies of the effect of soy isoflavones on bone, Messina et al (3) concluded: "Fifteen clinical trials were identified that examined the effects of isoflavones or isoflavone-rich soy protein on bone mineral density. Most trials were conducted for 1 year or less and involved relatively few (<30) participants per group. The findings from these studies are inconsistent but generally suggest that isoflavones reduce bone loss in younger postmenopausal women. Similarly, the limited epidemiologic data generally show that[,] among Asian populations[,] isoflavone intake is associated with higher bone mineral density."
In this issue of the Journal, Spence et al (4) report on their laboratory's different approach to the question of isoflavone's effects on bone. They ask whether isoflavones influence the body's calcium economy, as shown by their effects on net calcium balance and on bone calcium deposition and resorption modeled by measurements of calcium kinetics with calcium radioisotopes.
Because studies of isoflavones' effects on bone admittedly yield conflicting results, from a priori considerations one might wonder how the results of a study of isoflavones on calcium physiology might help to resolve the issue, whatever the findings. Positive effects on calcium balance would accord with the positive bone studies, and negative effects would accord with the negative bone studies. Thus, one might ask, what compels the conduct of a calcium kinetic study?
Spence et al answer that question convincingly. They argue that diet confounders and between-subject variation could account for the inconsistencies in the isoflavone studies that have bone as an outcome measure. They point out that their study design of calcium balance and calcium kinetics provides comparisons in the same subjects with strict dietary control and allows for mechanistic interpretation.
Indeed, the investigators have performed an impressively rigorous, diet-controlled, blinded crossover study in a group of 15 postmenopausal women who randomly ate a fixed baseline diet supplemented with a soy protein isolate devoid of isoflavones (soy-minus diet), a soy protein isolate containing isoflavones (soy-plus diet), or a casein-whey protein isolate (control diet) for 28 d. The investigators could not detect a greater effect of soy with isoflavones (soy-plus diet) than of soy without isoflavones (soy-minus diet) on net calcium balance, on bone deposition or resorption (from the kinetic data), or on bone formation or resorption (from the biochemical marker data). They report a negative study, then, which on superficial interpretation would support previously reported published negative studies of a beneficial effect of isoflavones on bone.
Does this rigorously controlled, randomized, crossover study falsify the hypothesis that isoflavones have beneficial effects on calcium balance and bone health in postmenopausal women? Probably not, for several reasons. As Spence et al asserted, "Our study cannot exclude the possibility that soy isoflavones increase calcium retention at high doses or that soy isoflavones are ... more effective immediately after menopause when bone turnover is higher." Alternatively, isoflavones may act most effectively in those postmenopausal women with the lowest endogenous estrogen concentrations; the investigators might consider examining that possibility among their subjects.
Finally, in negative studies, the question always arises as to whether the study had sufficient power to detect a clinically important effect. On that issue, with respect to the comparative results on calcium metabolism in the soy-plus and soy-minus diet arms of their study, the investigators remained silent.
In the studies comparing the soy-minus and control (casein-whey) diets, Spence et al could detect no significant difference in net calcium balance. One wonders whether they might have detected a statistically significant, and potentially clinically important, difference had they studied a larger number of women. They found a net calcium balance of –70 mg/d with the soy-minus diet and of –159 mg/d with the control diet, a >100% difference amounting to 89 mg/d—but not a statistically significant difference, possibly because of a large between-subject and within-subject variability. Because the mean values had very high CVs, and yet differed by a factor of 2, a larger sample size might have shown a significant difference. I find it intriguing that the 89-mg/d nonsignificantly lower net calcium balance compares favorably with the corresponding 61-mg/d significantly lower urinary calcium excretion.
What accounted for that significantly lower finding for urinary calcium excretion in the soy-minus diet than in the control diet? By design, only the dietary amino acid composition differed between the 2 groups, so it seems logical to look to amino acid differences for the accounting factor. Differences in amino acid composition without differences in protein intake might affect urinary calcium excretion through amino acid–specific differences in the renal glomerular filtration rate or through differences in renal tubular calcium transport. The soy protein supplied 20% less sulfur-containing amino acids (SAAs) than did the casein-whey protein, which would translate to 20% lower rates of sulfate excretion and net endogenous acid production (NEAP), both of which would predictably reduce urinary calcium excretion. Sulfate complexes calcium and thereby reduces its renal tubular absorbability: the less urinary sulfate, the more calcium absorbability. Less NEAP promotes calcium retention in bone, thereby delivering less calcium for glomerular filtration, and it also directly enhances calcium absorption by the renal tubule. If a significantly lower SAA intake constituted the only calcium-affecting factor that differed between the soy-minus and the control arms of the study, one would be compelled to suspect that it accounted for the significantly lower urinary calcium excretion with the soy-minus diet than with the control diet. The investigators clearly acknowledge that likelihood.
Because the investigators found no significant difference in renal net acid excretion (an index of NEAP) between the soy-minus and the control arms, one might postulate that bone release of calcium-accompanying base titrated NEAP in both arms proportionately to the NEAP—ie, the soy-minus diet titrated NEAP less than did the control diet. That would mean less bone release of calcium with its accompanying base in the soy-minus arm and consequently less calcium delivery to the kidney and less urinary calcium excretion, the latter of which accords with the investigators' observations. Although no significant differences between the soy-minus and the control arms in bone calcium deposition or resorption emerged from the calcium kinetic studies, one cannot know, as noted earlier, whether the sample size had the power to detect significant small differences, on the order of 60 mg/d.
In summary, it would seem we cannot interpret the data of Spence et al as conclusively excluding the possibility that a soy protein–supplemented diet (without isoflavones) promotes a lower negative calcium balance in postmenopausal women than does a casein-whey protein–supplemented diet of equal protein content. Although only a small daily difference might result between the 2 regimens, the cumulative effects over time could have substantial effects on bone calcium content.
A critical analysis of the known relations between dietary protein and bone health would extend this editorial's scope, although the following consideration seems pertinent. Increasing protein intakes likely can have anabolic effects on bone, as a result of the amino acid substrate's provision for building bone matrix and of an amino acid–stimulatory effect that increases the bone growth–promoting factor insulin-like growth hormone I (5, 6). However, high protein intakes increase endogenous acid production, which can have offsetting catabolic effects on bone, and which, therefore, might contribute to the negative findings in the study of Spence et al. Because base-loading metabolic alkalosis, on the other hand, has anabolic effects on bone (7) and because metabolic acidosis reduces serum insulin-like growth hormone I concentrations (8), the combination of a net-base-producing and alkalosis-producing diet and a high-protein diet might eliminate age-related declines in bone mass and optimize peak bone mass achievement during growth. Higher protein intakes show favorable longitudinal effects on bone mineral density with diets supplemented with the alkalinizing calcium salt citrate malate plus vitamin D, the latter of which may have improved base absorption by stimulating calcium absorption (9).
Those considerations might have relevance to the implications of the study of Spence et al. Because the diets they used yielded net acid on metabolism, the negative effects on calcium metabolism caused by a positive NEAP might have restricted any potential positive effects of the supplemented proteins (or isoflavones) on calcium metabolism. Perhaps the full calcium and bone "anabolic" effects of protein, isoflavones, or both require the facilitating effect of a concurrent net-base-producing diet.
If isoflavones prove useful for treatment or prevention of postmenopausal bone loss, this author recommends caution with respect to increasing the intake of soybean products to increase isoflavone intake. Paleonutritionists have pointed out that legumes entered the human food supply late in human evolution, and that, therefore, natural selection may have had insufficient time to eliminate maladaptations or to generate adaptations to human ingestion of legumes (10). Cordain (10) pointed out that legumes contain a wide variety of antinutrient compounds that influence multiple tissues and systems, and that normal cooking procedures do not always eliminate them [see also Liener (11) and Messina (12)]. In contrast, other groups caution against the use of isoflavone supplements because overuse may elicit the adverse effects of estrogenic activity (13). It appears we have much yet to learn about the health effects of ingesting soy products and soy isoflavones (13).
ACKNOWLEDGMENTS
The author had no conflicts of interest with the article by Spence et al or with their study.
REFERENCES
Related articles in AJCN: