Editorial Comment

Diet and Cardiovascular Disease Prevention

The Need for a Paradigm Shift

Frank B. Hu MD, PhD, a,

aDepartments of Nutrition and Epidemiology, Harvard School of Public
Health, Boston, Massachusetts; and the Channing Laboratory, Department
of Medicine, Brigham and Women's Hospital and Harvard Medical School,
Boston, Massachusetts

J Am Coll Cardiol, Available online 18 June 2007.

The traditional diet-heart hypothesis (that high intake of saturated
fats and cholesterol leads to atheromatous plaque, narrowing of
coronary arteries, and, eventually, myocardial infarction) dates back
to evidence from the beginning of the 20th century (1). Early animal
studies showed a relationship among dietary cholesterol, saturated
fat, and arterial lesions, an effect mediated largely through
elevations in plasma cholesterol. Subsequently, metabolic research
found that dietary intake of specific fatty acids directly influenced
blood cholesterol in humans, and epidemiologic studies identified
elevated serum cholesterol as a strong independent risk factor for
coronary heart disease (CHD). Together with ecologic correlations
between diet and heart disease rates and findings from migration
studies and special populations, these data heavily influenced the
formation of the diet-heart hypothesis.

This hypothesis has since played a major role in shaping national
dietary guidelines. Prevailing recommendations to prevent
cardiovascular disease (CVD) and promote weight loss have called for
diets low in fat (particularly saturated fat) and high in complex
carbohydrates. These recommendations unleashed a proliferation of low-
fat or nonfat products, and, over the past 2 decades, led to a
substantial reduction in the percentage of dietary energy from fat. At
the same time, however, the prevalence of obesity and type 2 diabetes
has increased dramatically. These trends cast doubt on the benefits of
low-fat diets and contributed to a resurgence of interest in low-
carbohydrate diets.

Although the initial intention of the low-fat campaign was to reduce
saturated fat intake, the desire for a simple message designed for the
general public resulted in the incrimination of all fats, despite
clear evidence that this view was not supported scientifically. Human
feeding studies from the 1950s showed that reduction of total fat
intake had no effect on serum cholesterol and that polyunsaturated fat
reduced cholesterol levels (2). In contrast, large prospective cohort
studies and secondary prevention trials indicate that substitution of
unsaturated fats for saturated fats, trans fats, or carbohydrates is
beneficial for CHD prevention, whereas simply reducing total fat has
no effect (1). In addition to fostering misunderstanding among the
general public, the low-fat campaign has spurred a compensatory
increase in consumption of refined carbohydrates and added sugars-an
unintended consequence that likely fueled the twin epidemics of
obesity and diabetes. In metabolic studies, low-fat, high-carbohydrate
diets not only induce high glycemic and insulinemic responses, but
also increase plasma triglycerides and decrease high-density
lipoprotein (HDL) cholesterol.

Carbohydrates are traditionally classified as simple or complex
depending on chemical structure. Simple sugars are typically digested
and absorbed more quickly than complex ones, and thus are thought to
induce more rapid postprandial glucose response. However, numerous
metabolic studies have challenged this belief, and it is now
recognized that many starchy foods (e.g., baked potatoes and white
bread) produce even higher glycemic responses than simple sugars (3).
Different glycemic responses to carbohydrate-containing foods underlie
the development of the glycemic index (GI), a concept introduced by
Jenkins et al. (4) in 1981.

The GI compares blood glucose levels after ingestion of a test food
and a standard weight (50 g) of a reference carbohydrate (glucose or
white bread). It ranks foods based on increase in blood glucose (area
under the curve). The GI largely depends on rate of digestion and
speed of carbohydrate absorption, but the physical form of foods is
also an important determinant. Typically, foods with more compact
granules (low starch gelatinization) and high levels of viscose
soluble fiber (e.g., barley, oats, and rye) are digested at a slower
rate and have lower GI values. Whole-grain products with intact bran
and germ also meet these criteria. Refined carbohydrates (e.g., white
bread), however, are digested more rapidly because grinding or milling
of cereals reduces particle size, removes most of the bran and the
germ, and allows for more rapid attack by digestive enzymes (5).

Blood glucose response is determined by the quality as well as the
quantity of carbohydrates in food. Both factors are reflected in the
concept of glycemic load (GL) (the product of the GI value of a food
and its carbohydrate content). When white bread is used as the
reference, each unit of dietary GL represents the equivalent glycemic
effect of 1 g of carbohydrates from white bread. Data from several
large population-based studies show that dietary GI or GL values have
an inverse association with HDL levels and a positive association with
triglycerides. Epidemiologic studies also show that higher dietary GL,
especially when combined with low intake of cereal fiber,
significantly elevates long-term risk of type 2 diabetes ([6] and
[7]).

Liu et al. (8) first reported a positive association between a higher
dietary GL and risk of CHD in the Nurses' Health Study (NHS). In this
issue of the Journal, Beulens et al. (9) report a similar association
between GL and risk of CVD in the Dutch EPIC (European Prospective
Investigation into Cancer and Nutrition) cohort. The analysis included
15,714 women ages 49 to 70 years without diabetes or CVD at baseline.
During 9 years of follow-up, 556 incident cases of major CVD events
were documented. After adjusting for CVD risk factors and dietary fat
and fiber, the investigators found a significant association between
dietary GL and increased risk of CVD (risk ratio comparing extreme
quartiles = 1.47, 95% confidence interval 1.04 to 2.09, p for trend =
0.03). There was also an association between higher dietary GI and
increased risk of CHD. Similar to the results from the NHS, the
increased risk was more pronounced among overweight and obese women
compared with normal-weight women.

Accurate quantification of dietary intakes in free-living populations
is a major challenge in large nutritional epidemiologic studies. In
the past 2 decades, the semiquantitative food frequency questionnaire
(FFQ) has become the method of choice because of its low cost and
ability to assess usual diet, the main interest in most epidemiologic
studies of diet and chronic diseases. Beulens et al. (9) validated the
FFQ against twelve 24-h recalls; the Spearman correlations between the
FFQ and 24-h recalls ranged from 0.56 to 0.78 for major carbohydrate-
containing foods that contribute most to overall GL (including
potatoes, bread, soft drinks, and sweets). Consistent with metabolic
and other epidemiologic studies, the authors also found an inverse
association between dietary GL/GI and HDL levels.

Measuring dietary GL in mixed meals presents another methodologic
challenge. A major concern has been the relevance of the GI values of
individual foods to glucose and insulin responses to mixed meals.
However, strong correlations have been found between observed GI
values of mixed meals and calculated values based on individual
component foods (10). In epidemiologic studies, GL measures a dietary
pattern characterized by higher intakes of refined carbohydrates and
added sugar rather than absolute glycemic effects of a diet.

In both Beulens et al. (9) and the NHS (8), the greater impact of GL
on CVD in overweight and obese people suggests that adverse effects of
dietary GL may be further aggravated by underlying insulin resistance.
Because two-thirds of Americans are overweight or obese, these
findings have important public health and clinical implications. In
the past several decades, the increase in dietary GL has been
proportional to the decrease of dietary fat as a percentage of energy
intake. This trend, together with the obesity epidemic, creates ideal
conditions for the development of cardiometabolic disorders. For this
reason, reducing dietary GL should be made a top public health
priority.

Several dietary strategies can be used to reduce GL. These include
replacing carbohydrates (especially refined grains and sugar) with
unsaturated fats and/or protein or exchanging whole grains for refined
ones. A combination of these approaches may increase benefits and
adherence. The types or sources of fat and protein used to replace
carbohydrates are as important as the amounts consumed. A recent
comparison of a low-carbohydrate diet high in vegetable fat and
vegetable protein and a low-fat, high-carbohydrate diet showed that
the former was associated with a significantly lower risk of CHD
during 20 years of follow-up in the NHS (11). In the NHS, most of the
vegetable fat came from vegetable oil (e.g., soybean, corn, and canola
oils), olive oil, mayonnaise, peanut butter, and nuts. Most of the
vegetable protein came from whole-grain foods (e.g., dark bread and
cold cereals), legumes (e.g., beans and peas), peanut butter, and
nuts. Benefits of the plant-based low-carbohydrate diet on CHD are
likely to stem from the vegetable fat and protein as well as the
reduced GL in the dietary pattern.

Another important way to reduce GL is to restrict consumption of sugar-
sweetened beverages, which account for nearly 50% of added sugar and
8% of total energy intake in the U.S. diet (12). Between 1977 and
2001, consumption of sugar-sweetened soft drinks increased by 135%.
These beverages, particularly soda, provide little nutritional
benefit. They do, however, cause weight gain and likely increase the
risk of diabetes, fractures, and dental caries. Thus, regular
consumption of sugar-sweetened beverages should be strongly
discouraged.

It is clear that the initial diet-heart hypothesis relating total and
saturated fat to CHD is overly simplistic and that sound nutritional
guidelines must take into account different types of fats and
carbohydrates as well as several other aspects of diet, some more
easily modified than intake of dietary fat. Because multiple lines of
evidence implicate high GL in adverse metabolic effects that increase
risk of diabetes and CVD, it is time to shift the diet-heart paradigm
away from restricted fat intake and toward reduced GL. This change can
help prevent CVD and improve overall health, and, as such, should be
considered a public health priority.

References

1 F.B. Hu and W.C. Willett, Optimal diets for prevention of coronary
heart disease, JAMA 288 (2002), pp. 2569-2578. View Record in Scopus |
Cited By in Scopus

2 R.P. Mensink and M.B. Katan, Effect of dietary fatty acids on serum
lipids and lipoproteins A meta-analysis of 27 trials, Arterioscler
Thromb 12 (1992), pp. 911-919. View Record in Scopus | Cited By in
Scopus

3 D.S. Ludwig, The glycemic index: physiological mechanisms relating
to obesity, diabetes, and cardiovascular disease, JAMA 287 (2002), pp.
2414-2423. View Record in Scopus | Cited By in Scopus

4 D.J. Jenkins, T.M. Wolever and R.H. Taylor et al., Glycemic index of
foods: a physiological basis for carbohydrate exchange, Am J Clin Nutr
34 (1981), pp. 362-366. View Record in Scopus | Cited By in Scopus

5 S. Liu, Intake of refined carbohydrates and whole grain foods in
relation to risk of type 2 diabetes mellitus and coronary heart
disease, J Am Coll Nutr 21 (2002), pp. 298-306. View Record in Scopus

6 J. Salmeron, A. Ascherio and E.B. Rimm et al., Dietary fiber,
glycemic load, and risk of NIDDM in men, Diabetes Care 20 (1997), pp.
545-550. View Record in Scopus | Cited By in Scopus

7 J. Salmeron, J.E. Manson, M.J. Stampfer, G.A. Colditz, A.L. Wing and
W.C. Willett, Dietary fiber, glycemic load, and risk of non-insulin-
dependent diabetes mellitus in women, JAMA 277 (1997), pp. 472-477.
View Record in Scopus | Cited By in Scopus

8 S. Liu, W.C. Willett and M.J. Stampfer et al., A prospective study
of dietary glycemic load, carbohydrate intake, and risk of coronary
heart disease in US women, Am J Clin Nutr 71 (2000), pp. 1455-1461.
View Record in Scopus | Cited By in Scopus

9 J.W.J. Beulens, L.M. de Bruijne and R.P. Stolk et al., High dietary
glycemic load and glycemic index increase risk of cardiovascular
disease among middle-aged women: a population-based follow-up study, J
Am Coll Cardiol 50 (2007) 14-21.

10 T.M. Wolever, M. Yang, X.Y. Zeng, F. Atkinson and J.C. Brand-
Miller, Food glycemic index, as given in glycemic index tables, is a
significant determinant of glycemic responses elicited by composite
breakfast meals, Am J Clin Nutr 83 (2006), pp. 1306-1312. View Record
in Scopus | Cited By in Scopus

11 T.L. Halton, W.C. Willett and S. Liu et al., Low-carbohydrate-diet
score and the risk of coronary heart disease in women, N Engl J Med
355 (2006), pp. 1991-2002. View Record in Scopus | Cited By in Scopus

12 V.S. Malik, M.B. Schulze and F.B. Hu, Intake of sugar-sweetened
beverages and weight gain: a systematic review, Am J Clin Nutr 84
(2006), pp. 274-288. View Record in Scopus | Cited By in Scopus

Reprint requests and correspondence: Dr. Frank B. Hu, Department of
Nutrition, Harvard School of Public Health, 665 Huntington Avenue,
Boston, Massachusetts 02115

* * *

Marilyn

In this recent study high meat intake is associated with higher breast
cancer risk.

http://www.nature.com/bjc/journal/v96/n7/full/6603689a.html