Page 274

mates of losses or waste and no measurements of actual intake. Total carbohydrate availability has declined since 1909; per-capita amounts fell from 493 g/day during 1909-1913 to a low of 378 g/day during 1967-1969 and rose to 413 g/day in 1985 (Chapter 3, Table 3-3). The decline was due to decreased use of flour and cereal products.

From the early part of this century to the 1980s, there was a notable shift in the proportion of total carbohydrate derived from starch and sugars. During 1909-1913, 68% of total carbohydrates came from starch, in comparison to 47% in 1980. Conversely, the contribution of sugars increased from 32% during 1909-1913 to 53% in 1980 (Welsh and Marston, 1982).

Over the past 20 years, the relative contribution of sugars to the food supply has changed. In 1965, sucrose predominated, comprising 85% of total sugars; sugars in corn syrups comprised only 13%. There were no HFCS sweeteners at that time. By 1985, the use of all types of corn syrups had increased to 47% of total sugars but there was a concomitant decline in sucrose use. The marked increase in corn syrup use during the last decade was due chiefly to greater use of HFCS—a popular sweetener of soft drinks and other processed foods. In 1985, HFCS accounted for 30% of the total sugar supply (Glinsmann et al., 1986).

The 1977-1978 Nationwide Food Consumption Survey (NFCS) (USDA, 1984) indicated that carbohydrates furnished an average of 43% of calories, whereas the NFCS Continuing Survey of Food Intakes by Individuals (CSFII) (USDA, 1986, 1987, 1988) suggested that women and children derived closer to 47% of their calories from carbohydrates. Both surveys indicated that children had higher proportionate intakes of carbohydrates than did adults and that women had higher intakes of carbohydrates compared to men of the same age group. Because these surveys did not take into account the percentage of calories from alcohol, the reported percentages of calories from carbohydrates, fats, and proteins are inaccurate. Carbohydrate intake was not affected by region, urbanization, or season; however, it was higher for those below than above the poverty level (USDA 1984, 1987, 1988).

In 1986, the Food and Drug Administration (FDA) estimated that the average daily intake of all sugars by the U.S. population accounted for 21% of total calories—half coming from added sugars and half from naturally occurring sugars. On the average, approximately 4% of calories came from fructose, 9%  from sucrose, and 5%  from sugars in corn syrups (see Chapter 3, Tables 3-6 and 3-7, and Glinsmann et al. 1986).

Evidence Associating Carbohydrate Intake with Chronic Diseases

Noninsulin-Dependent Diabetes Mellitus
Epidemiologic Evidence

Most epidemiologic studies were conducted at the time when distinction was still made between juvenile-onset and adult-onset diabetes rather than the most recently adopted more distinct classifications of Type I, or insulin-dependent diabetes mellitus (IDDM), and Type II, or noninsulin-dependent diabetes mellitus (NIDDM), respectively. Although the studies referenced here generally concern adult-onset diabetes, it seems reasonable to extend the results to all cases of diabetes.

Increased intake of sugars or total carbohydrates is not associated with increased risk of NIDDM. In a prospective study of 9,494 male Israeli government employees who were nondiabetic and 40 years of age or older at baseline, Medalie et al. (1974) found no association between calories from sugars or intake of total carbohydrates and incidence of diabetes mellitus over a 5-year follow-up. In a cross-sectional study of 3,454 employed people in England, Keen et al. (1979) observed that intake of carbohydrates, fats, and protein tended to be inversely correlated with concentration of blood sugar and indices of glucose tolerance; they inferred that the correlations probably were confounded by caloric expenditure. Baird (1972) reported an inverse association between sugar intake and prevalence of previously undetected diabetes among the siblings of diabetic propositi. West et al. (1976) found no association between sugar consumption and occurrence of diabetes in 286 Plains Indians, whose intake of refined sugar ranged from less than 70 g/day to more than 200 g/day.

In a study involving 22 countries, Yudkin (1964) reported a correlation of 0.73 between mean per-capita supply of sugars from 1934 to 1938 and risk of death due to diabetes from 1955 to 1956. West (1978) pointed out that this result was not consistently observed; in a sample of 44 countries, the correlation was only 0.18 for sugar intake in 1951 and diabetes mortality in 1971. Furthermore, sugar consumption is high in several coun-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement