of protein available for consumption in wholesale and retail markets, however, and not the amounts actually consumed.
According to USDA's Nationwide Food Consumption Survey (NFCS) of 1977-1978, the mean protein intake for all respondents (infancy to over 75 years of age) was 74 g/day or 16.5% of total calories and exceeded the RDA for all 22 age-sex groups (USDA, 1984). In later USDA surveys, conducted in 1985 and 1986, protein intake averaged 15 to 16% of calories for children 1 to 5 years, 16% of calories for women 19 to 50 years, and 16.5% for men 19 to 50 years, regardless of income (USDA, 1986, 1987a,b). There was little variation in intake with race or urbanization. The 1977-1978 NFCS indicated that meats, poultry, and fish contributed approximately 49%, dairy products 18%, eggs 4%, legumes 3%, cereal products 18%, and fruits and vegetables 7 to 8% of the protein in the U.S. diet (USDA, 1983).
Several considerations must be borne in mind in reviewing studies on dietary protein and chronic diseases:
· Because intakes of animal protein and saturated fat tend to be highly correlated, it is not possible in most epidemiologic studies to separate their independent effects.
· Many epidemiologic studies rely on evidence from vegetarians (e.g., complete vegetarians and lacto-ovovegetarians) that should be evaluated carefully for several reasons: The total protein intake of vegetarians is not much lower than that of omnivores; however, the lifestyles of vegetarians are likely to differ from those of omnivores in many ways that may confound the association between vegetable or animal protein intake and health. In addition, there is a lack of consistency among and within some studies regarding the length of time that subjects have followed a vegetarian diet.
· Laboratory animal studies are often conducted with large, nonphysiologic doses of protein. Thus, the applicability of their findings to human populations may be severely limited.
The epidemiologic literature on the etiology of coronary heart disease (CHD) emphasizes the role of dietary fats, particularly saturated fat, rather than dietary protein (see Chapter 7). Because animal protein and saturated fat intake tend to be highly correlated, however, it is not surprising that animal protein intake is positively correlated with CHD mortality as are intakes of total and saturated fats. This is so whether one compares populations among different countries, within countries, or migrant populations, or whether one examines secular trends (Aravanis and Loannidis, 1984; Berkson and Stamler, 1981; Kritchevsky, 1976; Toshima et al., 1984).
Findings from the major cohort studies of heart disease have generally failed to demonstrate an independent effect for total dietary protein. For example, Keys et al. (1986) found no association between 15-year mortality from CHD and dietary protein intake (as a percentage of total calories) in an ecological correlation analysis of 15 male cohorts in seven countries. Similarly, Gordon et al. (1981) found no relationship between age-adjusted mean daily protein intake (based on 24-hour dietary recalls) and the occurrence of CHD over periods as long as 6 years in three prospective cohorts of men (in Framingham, Honolulu, and Puerto Rico). A more recent analysis of the Honolulu cohort confirmed the finding; however, because total caloric intake was lower for CHD cases than for the noncases (reflecting a lower intake of carbohydrates and alcohol), protein as a percentage of total calories was significantly higher for the CHD cases (McGee et al., 1984).
The importance of the source of the protein as a factor in CHD risk is supported indirectly by studies of the effects of different dietary proteins on serum cholesterola well-established risk factor for CHD (see Chapter 7). Soy protein-based diets have been shown to have a substantial serum cholesterol-lowering effect in hypercholesterolemic subjects, and the major decrease is in low-density liproprotein (LDL) cholesterol (Descovich et al., 1980; Gaddi et al., 1987; Goldberg et al., 1982; Sirtori et al., 1979, 1985; Verillo et al., 1985; Widhalm, 1986; Wolfe et al., 1981).
The effect of soy-based protein diets on people with normal serum cholesterol is less consistent. For example, Wolfe et al. (1986) and Carroll et al. (1978b) reported that the substitution of vegetable protein (primarily soy) for meat and dairy protein resulted in a substantial lowering of mean serum cholesterol in healthy adults of both sexes. Van Raaij et al. (1981) reported that substitution of