to the editor: As I ate my lunch of soy products with a side of polyunsaturated fat, I enjoyed Dr. Kelly's article entitled “Diet and Exercise in the Management of Hyperlipidemia.” The article provided an excellent overview of dietary approaches to manage high cholesterol levels. Yet, as I chewed on my food, and on the information in the article, I considered that most people do not eat (or think about) food in ill-defined heterogeneous categories, or split into nutritional constituents.

Take the category of tree nuts. What exactly are tree nuts? Putatively, they are firm, dry, edible morsels that grow on trees. But even if patients have the horticultural savvy to distinguish nuts from other firm, dry, edible morsels (e.g., seeds, legumes, fruit), and to know which nuts grow on trees, is this knowledge helpful in choosing foods to improve cholesterol levels? In fact, most evidence for tree nuts comes from studies of two foods: almonds and walnuts¹ (foods that contain, for example, high levels of monounsaturated and polyunsaturated fats, respectively²). Would the same effects be attained by eating coconut (a tree-borne nut having the highest concentration of saturated fats of any whole food)?² And what are the effects of Brazil nuts (conspicuously excluded from nut studies¹)? Are peanuts tree nuts (sometimes they are included despite not actually being nuts¹)? Such questions underscore the confusion about dietary recommendations based on heterogeneous and ill-defined food categories.

Dr. Kelly also mentions a food constituent, saturated fats, as a prime target for dietary reduction in patients with dyslipidemia. The problem is that people do not eat saturated fat; they eat foods. So what foods should patients avoid to limit their intake of saturated fat? Dr. Kelly describes one option, the Mediterranean Diet, as recommending “low consumption of saturated fats” and “limited consumption of red meat and dairy products.” In fact, with regard to lipids, red meat, dairy, and saturated fats are one and the same; the former two being the principal dietary sources of the latter.³ Unfortunately, few people know this because medical science tends to reduce foods to their constituent parts,⁴ and because the beef and dairy industries work hard to ensure that consumers do not make connections between undesired abstract nutritional constituents and food.⁵

Family physicians must provide clear, evidence-based, actionable advice for patients. Patients eat—and understand—food, not food categories or constituents. Thus, rather than counseling patients in terms of tree nuts and saturated fats, we should give advice about food. For example, “to improve cholesterol levels, eat foods like almonds, walnuts, and peanut butter in place of foods like steaks, cheeseburgers, and ice cream.”

in reply: I completely agree with Dr. Lucan that family physicians “must provide clear, evidence-based, actionable advice for patients.” However, patients vary in their level of knowledge, literacy, and willingness to change their diet. No single approach will suit everyone's needs and preferences.

For example, many foods are processed combinations of ingredients (e.g., breads, cereals, snacks, sauces/dressings, quick meal options). For people who eat these foods, reading a nutrition label is an important skill. We should encourage motivated patients to learn how to make better food decisions by reading and understanding nutrition labels. However, other patients are not willing or able to do this, and will benefit more from specific advice about which foods to eat more or less of. Learning about and adopting a Mediterranean diet is an example of this approach.

Knowing patients as individuals and understanding the benefits of various dietary changes should equip family physicians to help patients choose dietary strategies that will be most effective and most acceptable to them.