A Prospective Study of Diet Quality and Mortality in Women

abstract & commentary

Synopsis: A dietary pattern characterized by consumption of foods recommended in current dietary guidelines is associated with decreased risk of mortality in women.

Source: Kant AK, et al. JAMA 2000;283:2109-2115.

This study explores the effects of dietary patterns on health outcomes. The study population was 42,254 women between the ages of 35 and 74 who completed a 62-item food questionnaire between 1987-1989. The women were then followed until 1993-1995 as part of the Breast Cancer Detection and Demonstration Project (BCDDP) sponsored by the National Cancer Institute and the American Cancer Society. There were 2065 deaths. Dietary patterns were scored according to the extent to which they conformed to current dietary recommendations to eat a diverse diet containing fruits, vegetables, low-fat dairy products, lean meats, and whole grains. The diet scores were divided into quartiles. Those with the highest intake of recommended foods had a 30% lower risk of all-cause mortality. There was a nonlinear dose response curve, such that women with scores in the top two quartiles exhibited similar health benefits when contrasted with women with scores in the bottom quartile.


In many ways, this study confirms what many would view as common sense. Any decent course in nutrition introduces many of the key concepts tested in this study. These key concepts include the recognition that whole foods contain many beneficial substances other than those currently identified as important nutrients and the notion that diversity in diet is the best bet if one wants to get enough of recognized and unrecognized micronutrients. Although one might be tempted to dismiss this report as confirming the obvious, there are many reasons why this study is important. First, the sample size is large and the study methods as described are exemplary. As Kant and associates note in the discussion, very few studies have examined the effects of dietary patterns upon health in women. Second, although it is often taken as an article of faith that a good diet promotes health, dietary counseling and nutritional science are often viewed as a lesser clinical art and a lesser science, respectively. For instance, nutritional counseling is often done by practitioners other than physicians, it is generally not reimbursed, and it is rarely given more than a cursory endorsement in the physician’s office. Evidence such as that contained in this study is needed if nutritional science and nutritional counseling are to be sanctioned by the medical establishment and payors and introduced into the medical curriculum. Third, patients are prone to gross oversimplifications about what are important nutrients. This leads to the common practice of not eating well and making up for poor food choices with multivitamins and over-the-counter, unregulated "food supplements." Would that things were so easy. I have often wondered if anyone has ever thought about how pompous it is to believe that we can capture in a small pill all that nature has wrought over the eons. As Kant et al point out in the introduction, "complex diets consumed by free-living individuals do not consist of single nutrients or foods, but rather a combination of foods containing multiple nutrients." Based on this insight, current nutritional advice endorses what is commonly termed the "food first paradigm." Supplements are needed only when a nutrient cannot be obtained in sufficient quantities from eating a healthy diet. In general, nutrients are delivered to our bodies in a complex matrix, such as a fruit or vegetable. Therefore, it is difficult and perhaps not scientifically defensible to test the effect of adding a single nutrient to a diet. Nutrients do not come "alone" except when we manufacture them that way and often the reason we manufacture them is to be able to sell them for a profit. Many studies suggest that nutrients interact with one another, leading to synergism among the nutrients, in terms of effects on bodily functions. This is the most compelling rationale for studying dietary patterns rather than the effects of single, isolated nutrients.

People want simple answers. The answer here may seem simple, but in truth it is a complicated recommendation. This study supports the common dictum that one should eat a diverse diet containing fruits, vegetables, whole grains, low-fat meat, and dairy products. This recommendation is simple only if you are not responsible for implementing it or teaching it. It is harder to implement if your nutritional IQ is low, if your budget is meager, or if you don’t have much time and energy for shopping and cooking. To some extent, this describes most Americans. I understand that many or most public schools no longer require home economics, so much of what children will know about nutrition must come from their overloaded, harried parents. When one considers how bombarded we are with advertisements from those selling a quick food fix, it is not surprising that many succumb to poor nutritional habits. Kant et al suggest that having a good diet is, at least in part, also a proxy for other lifestyle variables, including not smoking, exercising, and education. Given these considerations, it is easy to see why a good diet may be a surrogate or marker for a set of interlinked behaviors having to do with health awareness and disciplined lifestyle practices. If the agent in this study were a pharmaceutical product conferring a 30% reduction in all-cause mortality, I have no doubt that many would readily endorse its use. Although the recommendation in this study is "simple," I doubt that it will be for most physicians to "prescribe" such a diet in an era of time constraints and productivity expectations.