| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Editorial |
Channing Laboratory, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, and Department of Epidemiology, Harvard School of Public Health, Boston, Massachusetts
Gertrude H. Sergievsky Center, Taub Center for Alzheimer’s Disease Research, and Department of Neurology, Columbia University, and Division of Epidemiology, Columbia School of Public Health, New York, New York
Channing Laboratory, Department of Medicine, Brigham and Women’s Hospital, Harvard Medical School, and Department of Epidemiology, and Department of Nutrition, Harvard School of Public Health, Boston, Massachusetts
Much of the growing interest in diet and the development of dementia has focused on antioxidant vitamins. Clearly, there is a plausible basis for this attention, and convincing data support a free radical hypothesis for brain aging. Yet, the report by White et al. in this issue [1] that high tofu intake may be associated with cognitive impairment and brain atrophy is an important reminder that it is time to move beyond antioxidants and investigate additional dietary factors which could be associated with cognitive impairment, among the most common problems of aging.
Soy-based foods, with high levels of phytoestrogens, are certainly a good place for initiating such research. It is widely theorized that estrogen-related factors may play an important role in cognitive function. In animal models, estrogen appears to enhance the cholinergic [2,3] and glutamatergic [4] neurotransmitter systems, which regulate learning and memory. Estrogen has also been shown specifically to promote neuron survival [5] and increase synaptic density [6] and efficacy [7] in areas of the brain critical for memory. Estrogen receptors are dense in the basal forebrain [8], the major source of cholinergic innervation to the cerebral cortex and hippocampus. Recent studies in women demonstrate differences in cerebral blood flow and activation patterns among postmenopausal estrogen users compared to nonusers during performance of cognitive tasks [9], and several epidemiologic studies have reported that postmenopausal hormone therapy may reduce the risk of Alzheimer’s disease [10].
As a "natural" source of estrogen, soy foods and supplements have been touted as the all-purpose answer to healthy aging; women in Japan, where soy intake is particularly high, have lower rates of several estrogen-related diseases (e.g., breast cancer and osteoporosis) than their American counterparts. Yet, the lower rates of breast cancer (a "high-estrogen" disease) and osteoporosis (associated with low estrogen levels) in Japan might suggest both agonist and antagonist properties of phytoestrogens, rendering it difficult to predict how soy will interact with estrogen receptors in the brain. The results observed in the present issue of the Journal of the American College of Nutrition are therefore particularly important; it is the first epidemiologic study based on individual subjects to examine specifically the effects of soy consumption on brain aging. White et al. report, contrary to their own and other’s expectations, that men who consumed greater amounts of tofu during midlife appeared to score worse on cognitive tests, to have lower brain weight and to demonstrate ventricular enlargement on MRI compared to men with lower tofu intake. For example, cognitive impairment was identified in four percent of men with the lowest tofu intake, compared to 19% of those with the highest intake; low brain weight was observed in 12% of subjects with the lowest tofu intake and 40% of men in the highest category. Supporting evidence is provided by apparent dose-response effects (i.e., increasing risk of a poor outcome with increasing levels of tofu consumption) and by the subjects’ wives (logically assumed to have comparable tofu consumption to their husbands), in whom a similar association between tofu and lower cognitive scores was noted. Clearly, these results are interesting, although must be considered preliminary; relatively few subjects consumed tofu at high levels, and the confidence intervals around the estimates of effect presented are wide, indicating the limited precision of these results.
In general, diet is difficult to study; investigations of diet and cognitive impairment in the elderly introduce particular concerns. First, information is often collected retrospectively; recall may be a problem, and bias can easily be introduced when the healthy subjects are more likely to remember diet accurately than subjects who are beginning to experience cognitive problems. In addition, in a retrospective assessment, and particularly in the more common cross-sectional assessment (when information on both disease status and diet are collected simultaneously), it may be hard to discern whether specific dietary patterns led to the disease or whether the onset of disease symptoms led to specific dietary patterns.
Second, diet information is often collected only at a single point in time. Given the long latency period for cognitive impairment, information on long-term diet or diet in the distant past may be critical; a study of nuns found that linguistic ability at age 22 predicted risk of low cognitive test scores and Alzheimer’s disease 58 years later [11], strongly suggesting the importance of earlier life events to development of cognitive impairment. In practical terms, it seems likely that prevention could be most effective in the earliest stages of disease development, further supporting the potential need for information from many years prior to the manifestation of symptoms of cognitive dysfunction. Two strengths of the report by White et al. are their ability to combine dietary data from two different assessments up to nine years apart and the prospective data collection from subjects during midlife, approximately 20 years prior to the cognitive evaluations and MRI scans.
Nonetheless, we stress again that these provocative data must be regarded as preliminary; extensive additional investigation of the association between tofu and cognitive function, as well as dementia, in further studies and other populations will be necessary before we can use this information for the public health. Importantly, we do not know if tofu itself was the cause of these numerous indications of accelerated brain aging or if tofu is merely a marker for some other unfavorable exposure. For example, in this population, men with higher tofu intake (and more traditional diets) likely came from poorer immigrant families and perhaps experienced more childhood privation, which may be related to their brain development and subsequent cognitive function. In addition, since tofu is part of a traditional diet, it may simply mark a dietary pattern which might be harmful to the brain; accurately disentangling the individual components from the pattern could be a difficult task. Finally, the men with high tofu intake also experienced more strokes than men with lower intake (stroke is known to be more common in Japan than the US); thus it is possible that vascular causes of cognitive impairment contributed to the observed results. Although the authors attempt to address these issues, they cannot eliminate the possibility that their results may be due to some other, unidentified factor or factors related to both tofu intake and brain aging.
In addition, a plausible biologic hypothesis is generally an important part of judging epidemiologic relations. While high tofu intake may lead to lower plasma estrogen levels [12], we do not know how tofu influences estrogen levels in the brain; we also know very little about estrogen effects in men. Furthermore, data have not even consistently indicated that low endogenous estrogen levels are directly related to cognitive function in non-demented subjects [13]. The authors also posit a non-estrogen mediated hypothesis for the effects of tofu, namely that soy inhibits hippocampal tyrosine kinase and may block long-term potentiation (the likely mechanism by which humans learn and remember). Still, considerably more work must be done to substantiate this hypothesis.
Finally, the single measures of outcome used in this study may have limited value, as cognitive function and brain structure change over time. Factors which predict these measures at one point may or may not be the same as those which predict decline over time; fundamentally, the public health interest is in preventing cognitive decline, as the steepness of the decline trajectory likely provides an early marker for risk of the more clinically relevant result—dementia.
Cognitive impairment is one of the most common conditions affecting the elderly. In our rapidly aging society, the problem will become only more widespread. Yet, there is currently no available cure; thus, understanding ways to prevent early cognitive dysfunction is essential. Dietary factors may prove to be important for prevention, but data are sparse,and few large population-based studies of diet and cognitive function have been established. This should be a priority for future research.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
