Dietary Recommendations: Why Are They So Confusing?

by Karen Switkowski, MPH, MS

---

I often hear people express frustration with nutritional guidelines and recommendations. It can be difficult to find a reliable source of information given the influence of politics, the media, and the food and beverage industry on nutrition research. Even when studies are conducted according to high scientific standards and reported appropriately in the media, they often contradict one another or are difficult to interpret. One example is the much-hyped resveratrol, a compound found in red wine, dark chocolate, and berries. When initial studies showed that resveratrol might have beneficial effects for reducing risk of cardiovascular disease (CVD) and associated morbidities, the media was quick to disseminate the message that everyone should be indulging regularly in wine and chocolate bars. However, recent studies in humans have shown that resveratrol (in dietary or supplement form) has no effect on CVD risk and may actually be harmful in certain contexts such as physical activity and pregnancy. Conflicting nutrition research findings like these are very common. Why can’t we get a straight answer about nutrition?

Despite its importance to public health, nutrition is a complex science and there is considerable disagreement within the field about the best way to study it and translate research findings into recommendations. One issue is that it is difficult (and often extremely expensive) to conduct well-designed research studies, and nutrition science has its own unique challenges. Randomized controlled trials (RCTs), considered the gold standard for research studies, are often impossible to conduct for nutrient exposures. When RCTs are successfully completed, results are frequently inconsistent with epidemiologic studies. Many RCTs are conducted in animals and/or use extremely high doses of a nutrient or compound that are not practical for real-life consumption. In the resveratrol example, most of the initial studies were conducted on animals and we soon learned that a human would have to consume upwards of 1,000 liters of red wine daily in order to get the dose found to be beneficial in mice.

Assessing diet accurately in human populations is not easy. Common research methods, including food frequency questionnaires (FFQs) and food records, rely on research participants to remember what they are eating, accurately gauge portion sizes, and provide honest reports of consumption. Additionally, dietary recommendations are developed with the goal of improving health outcomes, which generally result from long-term, cumulative exposures to certain foods or dietary patterns. However, dietary assessments are often conducted at only one time point (or a series of repeated, isolated time points) and are assumed to represent long-term intake.

There is also the question of whether we should consider diet in terms of nutrients, foods, food groups, or dietary patterns (or some combination of all of these). This depends on the disease or health outcome that we are interested in studying. For example, hypertension can be linked directly to consumption of a particular nutrient (sodium), but CVD involves a more complex interplay between dietary fats, antioxidants, fiber, and other nutrients, in as well as lifestyle factors besides diet. Even if a clear exposure à outcome relationship can be identified, many other factors including exercise, stress, and genetic predisposition often contribute to individual susceptibility to diet-related diseases. When single nutrients are studied in an effort to isolate their effect on a particular outcome, results often doesn’t translate to real-life situations (nutrients are consumed as part of foods, where they interact with other food components).

Although these issues make it difficult to provide clear dietary recommendations from the results of even well-designed research studies, consumers of nutrition science can be assured that researchers, clinicians and policymakers are constantly trying to improve through ever-evolving methods, such as evidence-based research.