W. C. Willett, MD Department of Nutrition, Harvard School of Public Health, has reviewed the latest studies of low carbohydrate food choices and answers the question of Low-Carbohydrate Diets: A Place In Health Promotion?
Low-Carbohydrate Diets: A Place In Health Promotion?
W. C. Willett, MD Department of Nutrition, Harvard School of Public Health
Lagiou et al.  examined the relationship of low-carbohydrate high-protein diets to mortality in a cohort of 42 000 relatively young Swedish women. They found that high protein intake and a score based on both high-protein and low-carbohydrate predicted higher total and cardiovascular mortality over a 12-year follow-up.
A central question raised by the authors is why in this and a recent study from Greece (not yet published), a positive association between a low-carbohydrate diet and mortality was seen, whereas no association between a low-carbohydrate diet score and incidence of coronary heart disease was observed in the large Nurses’ Health Study cohort of US women . Understanding the differences is important because they relate directly to daily dietary choices and outcomes that are the major causes of morbidity and mortality in western countries. We can set side immediately that the difference in findings were due to fundamental differences in study quality.
All three studies were prospective, meaning that the potentially serious biases in case–control studies were avoided; all used dietary assessment methods with documented validity; and all had quite complete ascertainment of outcomes. The difference in endpoints is also unlikely to account for the apparently contrasting outcomes. Although the Swedish study examined total and cardiovascular mortality, the association with total mortality was primarily due to a stronger association with cardiovascular mortality, which should be approximately comparable to the incidence of coronary heart disease, which was the end-point in the US study. Thus, the apparently contrasting findings are likely to be due to real differences in the study populations, their diets or the methods of analysis. Chance can always play a role, but all studies were large and this is an unlikely explanation.
Differences in the study populations very plausibly contribute to differences in results between the Swedish and US findings (the details of the Greek study are not yet available). Women in the Swedish study were substantially younger (20–49 and followed for 12 years) than in the US study (34–60 years and followed
for 20 years) and leaner (the prevalence of BMI of 30 or more was 11% in the Swedish study and is approximately three times this in US women), and for both of these reasons, the degree of insulin resistance would have been much lower in the Swedish study.
One of the most important findings during the last decade has been that the adverse metabolic effects of high-carbohydrate intake are greatly magnified in the presence of underlying insulin resistance [3, 4]. Consistent with this evidence, we have seen that high dietary glycaemic load (high intake of refined starch and sugar) has little effect on the risk of coronary heart disease in lean women, but nearly doubles the risk in overweight and obese women . Similarly, high glycaemic load was not associated with the risk of stroke in leaner women, but was significantly associated with risk in overweight and obese women . The types of diets consumed by the Swedish and US women may have contributed to the differences in findings.
From many dozens of controlled feeding studies, we know that different types of fat have opposite effects on blood lipids; compared with carbohydrate, trans and saturated fats increase LDL and unsaturated fats reduce LDL . In epidemiological studies, these fatty acids have correspondingly predicted increased or decreased risk of coronary heart disease . Similarly, the type of carbohydrate rather than the total amount of carbohydrate is key in relation to the risk of cardiovascular disease. As noted above, high amounts of refined starch and sugar have adverse metabolic effects and increase risk; in contrast, whole grain/high-fibre forms of carbohydrate have beneficial metabolic effects and are related to reduced risks of cardiovascular disease .
Although the report by Lagiou et al. does not include information on the type of carbohydrate consumed by their participants, from other reports we know that the cereal fibre content (the form of fibre most strongly related to lower risk of cardiovascular disease and diabetes) is approximately three times higher in Swedish compared to US women . Whilst there is little reason to believe that the type of protein influences the risk of cardiovascular disease, the protein package can include large amounts of cholesterol and saturated fat (in the case of animal proteins) or fibre and beneficial micronutrients (with plant proteins). The available data do not allow a comparison of the sources of proteins in the Swedish and US studies. Confounding by other dietary factors could have influenced the findings; for example, it is possible that women with the low-carbohydrate dietary score were generally not health conscious and had a lower consumption of fruit and vegetables, which, from other studies, would be expected to increase the risk of cardiovascular disease.
Finally, the methods of analysis used in the two studies to describe a low-carbohydrate diet were not identical. In the US study, a score was created by expressing fat, protein and carbohydrate as percentage of energy, and then ranking participants by a score based on low-carbohydrate and high protein and fat intakes. In the Swedish study, scores were calculated for protein and carbohydrate and a combination score was calculated based on low-carbohydrate and high protein intakes; these scores were then included in a statistical model with total energy and saturated fat. Although the scores create complexity because they are positively related to protein and inversely related to carbohydrate intake, with total calories and saturated fat in the model they reflect a substitution for unsaturated fat, which in turn is known to reduce the risk of heart disease. For this reason, the findings of the Swedish study may partly reflect the known benefits of unsaturated fat on the risk of cardiovascular disease.
Although the complexities of human diets make comparisons of patterns across populations difficult, the most likely interpretation of the superficially inconsistent Swedish and US findings is that both are correct within the context of their respective populations and their diets. Within the relatively lean and younger Swedish women, replacement of largely whole grains with the protein packages consumed in that population probably have an adverse effect on cardiovascular disease.
In contrast, within a heavier and older population of US women, replacement of mainly refined starch and sugar with almost any source of calories is not likely to increase the risk of cardiovascular disease. This conclusion emphasizes the most fundamental conclusion of research into diet and cancer during the last decade: it is the type of fat and type of carbohydrate rather than the total amounts of either that influence risk of cardiovascular disease.
Thus, in the Nurses’ Health Study, a low-carbohydrate pattern that emphasized vegetable sources of protein and fat was associated with a significantly lower risk of coronary heart disease . This finding is reinforced by the results from a recent large controlled feeding study in which replacement of complex carbohydrates with either protein or monounsaturated fat reduced cardiovascular risk factors . Thus, there is every reason to believe that a moderately low carbohydrate intake (approximately 30–40% of energy) can result in low risk of cardiovascular disease if the proteins and fats are mainly from plant sources (but do not include hydrogenated oils) and the remaining carbohydrates are from whole grains, fruit and vegetables.
However, the findings of Lagiou et al. suggest that most Swedish women who reduce their carbohydrate intake may not be adopting this version of a low-carbohydrate diet and are adversely affecting their risk of cardiovascular disease.
1 Lagiou P, Sandin S, Weiderpass E et al. Low carbohydrate-high protein diet and mortality in a cohort of Swedish women.
J Intern Med 2007
2 Halton TL, Willett WC, Liu S et al. Low-carbohydrate-diet score and the risk of coronary heart disease in women. N Engl J Med 2006; 355: 1991–2002.
3 Jeppesen J, Hollenbeck CB, Zhou MY et al. Relation between insulin resistance, hyperinsulemia, postheparin plasma lipoprotein lipase activity, and postprandial lipemia. Arterioscler Thromb Vasc Biol 1995; 15: 320–4.
4 Willett WC, Stampfer M, Chu N, Spiegelman D, Holmes M, Rimm E. Assessment of questionnaire validity for measuring total fat intake using plasma lipid levels as criteria. Am J Epidemiol 2001; 154: 1107–12.
5 Liu S, Willett WC, Stampfer MJ et al. A prospective study of dietary glycemic load, carbohydrate intake, and risk of coronary heart disease in US women. Am J Clin Nutr 2000; 71: 1455–61.
6 Oh K, Hu FB, Cho E et al. Carbohydrate intake, glycemic index, glycemic load, and dietary fiber in relation to risk of stroke in women. Am J Epidemiol 2005; 161: 161–9.
7 Mensink RP, Zock PL, Kester AD, Katan MB. Effects of dietary fatty acids and carbohydrates on the ratio of serum total to HDL cholesterol and on serum lipids and apolipoproteins: a meta-analysis of 60 controlled trials. Am J Clin Nutr 2003; 77: 1146–55.
8 Hu FB, Willett WC. Optimal diets for prevention of coronary heart disease. JAMA 2002; 288: 2569–78.
9 Park Y, Hunter DJ, Spiegelman D et al. Dietary fiber intake and risk of colorectal cancer: a pooled analysis of prospective cohort studies. JAMA 2005; 294: 2849–57.
10 Appel LJ, Sacks FM, Carey VJ et al. Effects of protein, monounsaturated fat, and carbohydrate intake on blood pressure and serum lipids: results of the OmniHeart randomized trial. JAMA 2005; 294: 2455–64.
Reprinted from Journal of Internal Medicine 261; 363–365