点击显示 收起
1 From the Jean Mayer USDA Human Nutrition Research Center on Aging, the Department of Nutrition and Genomics, Tufts University, Boston, MA
See corresponding article on page 957.
The first century of atherosclerosis research has been dominated by the lipid hypothesis. Even when new hypotheses were brought to the table (ie, the oxidation hypothesis), they were put in the context of lipids (ie, oxidized LDL cholesterol). The reason for this may be traced to the beginning of the story, when Nikolaj N Anitschkow established the cholesterol-fed rabbit as a model for atherosclerosis research (1). One wonders how different this field might have been had Anitschkow decided to use an animal that is much more resistant to diet-induced atherosclerosis, such as the rat or mouse. During the ensuing decades, atherosclerosis moved from a laboratory curiosity to a major public health concern and to the proposal by some of the diet-heart hypothesis, which proposes a sequence of etiologic relations between the saturated fat content of the diet, serum cholesterol concentrations, and the development of disease. From the beginning, this diet-atherosclerosis connection was far from being unanimously embraced by the scientific community. A notable early example of this enduring controversy was "Keys versus Pickering" (2). The venue was the World Health Organization, and the year was 1955. There and then, Ancel Keys put forward his ideas expecting to be accepted on the spot, but he was challenged by Sir George Pickering, who according to witnesses said, "Yes, and Professor Keys would you be kind enough to cite for us the principle piece of evidence that you think supports this diet-heart theory of yours?" Keys's evidence, at that time, was not convincing and his hypothesis was not accepted, which drove him to build the evidence that would allow him to prove his point. This encounter was one of the driving forces behind the Seven Countries Study, which represents one of the most exciting adventures in early nutritional epidemiology (3). This study, despite its shortcomings, solidified for many the notion that dietary factors and, more specifically, dietary fat and cholesterol were responsible in great part for the rise of cardiovascular disease (CVD) experienced in Western and industrialized societies during the 20th century. Nowadays, however, almost 100 y after Anitschkow's experiments and about 50 y since the launching of the Seven Countries Study, the polemic continues about the role of diet on the development of heart disease (4-6). Whereas medical societies and government bodies have embraced the concept of nutrition as a major player in the epidemic of CVD and potentially in its control (6), some scientists remain skeptical about the diet-cholesterol-heart disease connection (4, 5). Even those who take the diet-heart hypothesis for granted maintain their own differences of opinion by arguing about what constitutes the optimal diet for atherosclerosis prevention and therapy (7).
In this issue of the Journal, Lefevre et al (8) present some results that may not soothe the above indicated controversy. However, the title and the opening statement of the manuscript contain the words that may hold the clue to reconciling the lingering polemic: "Individual variability in response." Lefevre et al's study design fulfills many of the expectations of a well-conducted dietary intervention study, namely, a randomized, double-blind, 3-period crossover controlled feeding design. In addition, these investigators take multiple measurements per dietary phase, which reduces the confounding of intraindividual variability. Moreover, they chemically measured the menus provided, thereby connecting the calculated with the actual composition of the diets. The composition of the diets was conventional: the average American diet (38% of energy as fat and 14% of energy as saturated fatty acids), the Step I diet (30% fat; 9% saturated fatty acids), and the Step II diet (25% fat; 6% saturated fatty acids); the fat content was adjusted by adding or removing milk fat. The diets were fed for 6 wk each to 86 free-living, healthy men aged 22–64 y. Although the biochemical variables presented in the article are basic, these are also the variables used by health professionals to ascertain dietary therapeutic success. Compared with the average American diet, the Step I and Step II diets lowered LDL-cholesterol concentrations by 7% and 12%, respectively. However, the bad news was that plasma HDL-cholesterol concentrations decreased on a similar magnitude, whereas plasma triacylglycerols increased by 14% and 16%, respectively. These results were, on average, similar to those from many previous studies that used similar experimental designs and diets. As in previous studies, the magnitude of individual variability in the response of plasma lipids to dietary intervention was astonishing. The investigators studied the potential influence of certain baseline variables as modulators of dietary response. Their analyses showed significantly smaller reductions in the LDL-cholesterol response to a Step II diet with an increase in percentage body fat, body mass index, or insulin concentrations. However, it is important to emphasize that the correlations were in the range of 0.2–0.3 and that the effect was far from specific. In fact, there were persons in the higher end of the body mass index distribution who had identical responses to those in the middle and lower ranges. Therefore, these data in isolation may not raise too many expectations regarding the influence of anthropometric variables on plasma lipid responses to dietary manipulation. Nevertheless, this study joins other articles that support the notion that the benefits of a hypocholesterolemic diet may not reach their maximum performance in subjects who are overweight or obese (9, 10) or who, as shown here (8), are insulin resistant. This consistency across studies supports the need for further research designed to unravel the bases of these observations. However, given the current data, no compelling evidence exists to translate these findings into clinical practice.
The other relevant outcome of the study relates to the results obtained for the ratio of total to HDL cholesterol, also known as the atherogenic ratio. Higher values have been associated with an increased CVD risk. Therefore, the elevations observed after the Step I and II diets cast some concerns about the efficacy of these low-fat diets to reduce CVD risk. However, experimental design may have been the driving force for the increased ratio of total to HDL cholesterol and triacylglycerol concentrations observed after the low-fat diets. We need to keep in mind that the persons were maintained at constant weights throughout the experiment by adjustments to their dietary intake. However, when low-fat diets are provided ad libitum and result in weight loss, they have the positive effect of reducing LDL-cholesterol concentrations without the potential downsides of decreasing HDL-cholesterol concentrations and increasing triacylglycerol concentrations (11). Moreover, despite the indisputable epidemiologic evidence regarding the protective role of HDL cholesterol, what really matters is the efficacy of reverse cholesterol transport efficiency rather than HDL-cholesterol concentrations (12). Whether the lowered HDL-cholesterol concentrations that result from low-fat diets are associated with functional impairment remains to be elucidated, especially in those situations in which the decrease is not accompanied by increased triacylglycerol concentrations. Therefore, deciding the suitability of a specific diet to reduce CVD risk based exclusively on its effects on the ratio of total to HDL cholesterol may be too vague; more information can be gained from examining in more detail the concentrations and distribution of different lipoprotein subfractions. In addition, several other variables exist that could have made the manuscript more comprehensive. These include measurements of glucose and insulin concentrations at each diet period, as well as measurements of more novel biomarkers, such as adipokines and cytokines. However, in this case, the interpretation of the data may be limited because of the imposed body weight maintenance.
In summary, the study was carried out in a sound manner and the results presented are consistent with previous literature. However, the data reflect the experimental design, which was conceived to minimize confounders rather than to reflect real life. It would be unfortunate if somebody decides to interpret from these findings that the best diet therapy for overweight or obese people to decrease CVD risk would be a high-fat diet consisting of milk products, and the authors judiciously avoid making this an explicit conclusion. Rather, they focus on the dramatic variability in the interindividual plasma lipid responses to diet even under a highly controlled environment and try to gain additional understanding about how factors such as body mass index and insulin sensitivity drive this variability. On the basis of the data, one may think about a hierarchical structure to dietary prevention or therapy for CVD. Thus, if the individual has insulin resistance or is obese, the primary emphasis should be in normalizing those conditions to achieve the maximum benefit of the hypocholesterolemic diets. Whether insulin, body mass index, or waist circumference are better determinants of impaired dietary responses cannot be assessed from the present study because multiple stepwise regression analyses are not advisable when using such highly correlated variables. Moreover, many other factors influence responses, including age, sex, physical activity, alcohol, smoking, and genetics. Their combined and thoughtful use should help in the identification of vulnerable populations or persons that will benefit from a variety of more personalized and mechanistic-based dietary recommendations. This potential for better prevention and therapy can and needs to be developed within the context of nutritional genomics (13) that, as part of systems biology, may provide the tools to achieve the holy grail of dietary prevention and therapy of CVDs. This approach will break with the traditional public health approach of one size fits all. In this regard, the first baby steps can be seen already in the most current version of the US Department of Agriculture pyramid (www.mypyramid.gov). Perhaps at some time in the future, the current controversies will be put to rest. We will be able to identify those persons for whom diet plays no major role in their risk of CVD and this should appease those who defend the diet-heart null hypothesis (4, 5). The same tools will identify those persons who may benefit more from one of the many potentially beneficial diets currently proposed (7).
ACKNOWLEDGMENTS
The author did not have any conflicts of interest.
REFERENCES
Related articles in AJCN: