Performance of semiquantitative food frequency questionnaires in international comparisons. Mexico City versus San Antonio, Texas.

International dietary comparisons present a number of methodologic difficulties. We developed two semiquantitative food frequency questionnaires for use in Mexico City and San Antonio, Texas. The Mexico City questionnaire contained 85 food items and the San Antonio questionnaire contained 108. Thirty-six items were common to both questionnaires. The questionnaires were administered to 128 Mexican Americans, 35 to 64 years old, residing in a low-income barrio neighborhood of San Antonio and to 644 similarly aged Mexicans residing in a low-income "colonia" of Mexico City. Major differences in fat and carbohydrate intake as a percent of total kilocalories were observed between the two study groups, with Mexico City residents consuming approximately 18 to 21% of calories from fat and 68 to 72% from carbohydrate compared to 29 to 33% from fat and 48 to 52% from carbohydrate for Mexican Americans from San Antonio. Both of these differences were highly statistically significant (P < 0.001). A number of arguments support the validity of these findings. First, the distribution of high- and low-fat foods was similar on both questionnaires. Also, of the top-ten foods (as contributors to total kilocalorie intake) in San Antonio, seven were high in fat (> 30% of kilocalories), whereas this was true of only three of the top-ten foods in Mexico City.

Genetic and environmental determinants of type II diabetes in Mexican Americans. Is there a "descending limb" to the modernization/diabetes relationship?

Evidence from migrant population studies and secular trend data indicates that environmental factors play a role in the etiology of non-insulin-dependent (type II) diabetes. These environmental factors appear to be concomitants of the process whereby traditional populations become "modernized" or "westernized" and include increased intake of total calories, fat, and sucrose, decreased intake of total and complex carbohydrates, including fiber, and decreased physical exercise. There also appears to be a "postmodernization" process, which we have characterized as the "descending limb of the curve." In Mexican Americans in San Antonio, the prevalence of type II diabetes declines with acculturation to the values, attitudes, and behaviors of "postmodernized" American society. However, examination of the dietary and exercise concomitants of this process revealed a mixed picture. There was some suggestion that Mexican-American women, although not men, had entered onto the descending limb of the curve. However, Native American genetic admixture in Mexican Americans also covaried with affluence and acculturation in such a way that the declining prevalence of diabetes could as easily be due to genetic factors as to environmental factors. The "pancreatic exhaustion" theory holds that resistance to insulin action is a principal lesion leading to hypersecretion of insulin, hyperinsulinemia, and eventual islet cell failure and clinical diabetes. This theory predicts that prediabetic subjects will be hyperinsulinemic. In conformity with this theory, we have shown that subgroups of the Mexican-American population, defined on the basis of family history of diabetes, who would be expected a priori to be enriched with prediabetic subjects, are hyperinsulinemic as predicted.(ABSTRACT TRUNCATED AT 250 WORDS)

Dietary intake and the prevalence of gallbladder disease in Mexican Americans.

Mexican Americans have a high prevalence of gallbladder disease. We examined the contribution of ethnic preferences in food intake to the risk of gallbladder disease in Mexican Americans and non-Hispanic whites. Participants in a population-based health survey were questioned about any history of gallbladder disease, and were interviewed to determine their dietary intake. After adjusting for age, body mass index, and ethnic group, we found that women with the highest intake of total fat and linoleic acid had reduced risks of gallbladder disease, although an opposite trend was observed in men. High levels of sucrose intake and low levels of cholesterol intake were associated with an increased risk for both sexes, but the odds ratios were not statistically significant. Although Mexican Americans and non-Hispanics differed in their intake of several nutrients, the elevated risk of gallbladder disease in Mexican American women was unchanged after ethnic differences in food intake were taken into account. Although the dietary preferences of Mexican Americans as reflected in 24-h diet recall interviews do not appear to explain their high prevalence of gallbladder disease, this finding should be interpreted with caution because of methodologic limitations in measuring habitual food intake.

The role of behavioral variables and fat patterning in explaining ethnic differences in serum lipids and lipoproteins.

Mexican Americans have been previously reported to have greater adiposity, higher triglyceride levels, and lower high density lipoprotein (HDL) cholesterol levels than Anglos. This study investigated the relationship between behavioral variables (caloric balance, cigarette and alcohol consumption, exercise, postmenopausal estrogen or oral contraceptive use) and fat patterning (central vs. peripheral distribution of adiposity) in the San Antonio Heart Study (1979-1982) (n = 2,102) to explain the ethnic difference in lipids and lipoproteins. Adjustment for caloric balance (as reflected by body mass index) narrowed the ethnic difference in triglyceride and HDL levels for both sexes, while adjustment for smoking widened the ethnic difference. For females, the ethnic difference was also decreased by adjustment for alcohol and estrogen use. However, adjustment for these behavioral variables did not completely eliminate the ethnic difference in lipids and lipoproteins in either sex. Increased central adiposity, more characteristic of Mexican Americans than Anglos, was positively associated with triglycerides and negatively associated with HDL levels, especially in females. Fat patterning made a more important contribution to the prediction of triglyceride and HDL levels than did the other behavioral variables (except for caloric balance) and, in general, eliminated ethnic differences in lipids and lipoproteins. Epidemiologists should consider the use of a centrality index to distinguish different types of adiposity since it is easy and inexpensive to measure.

Employment status and women's protection against coronary heart disease. Findings from the San Antonio Heart Study.

This study examined whether currently employed women are at increased risk of coronary heart disease relative to full-time homemakers. Subjects were 1,041 Mexican-American and non-Hispanic white women aged 25-64 years, residing in households randomly selected from three socioculturally distinct neighborhoods in San Antonio, Texas. No statistically significant differences between employed women and homemakers were found for obesity, total serum cholesterol, low density lipoprotein cholesterol, systolic and diastolic blood pressures, or cigarette smoking. Highly significant differences favoring employed women over homemakers were found for both Mexican Americans and non-Hispanic whites in high density lipoprotein (HDL) cholesterol, ratio of HDL cholesterol to total cholesterol, and triglycerides. These differences were not explained by obesity, exercise, cigarette smoking, alcohol consumption, use of exogenous estrogens, and use of oral contraceptives, or by the healthy worker effect, and were observed at all occupational levels. Employed women ate a less atherogenic diet than full-time homemakers, but it is not clear that this nutritional factor could explain the differences in HDL cholesterol and triglycerides found in this study. The magnitude of the employment status difference in HDL cholesterol for both ethnic groups was in a range (3-4 mg/100 ml) associated with protection against coronary heart disease.

Role of obesity and fat distribution in non-insulin-dependent diabetes mellitus in Mexican Americans and non-Hispanic whites.

Recent data have suggested that central obesity is related positively to the prevalence of non-insulin-dependent diabetes mellitus (NIDDM). We examined whether central obesity (measured by the ratio of subscapular to triceps skinfold) was predictive of NIDDM prevalence independently of overall obesity (measured by body mass index, BMI) in 1231 Mexican Americans and 939 non-Hispanic whites who participated in the San Antonio Heart Study, a population-based survey of diabetes and cardiovascular risk factors. Mexican Americans are characterized by higher rates of NIDDM, greater overall obesity, and more central body fat distribution than age-matched non-Hispanic whites. Using multiple logistic regression with age, ethnicity, BMI, and central obesity as covariates, overall obesity was positively associated with NIDDM prevalence in both sexes (P less than 0.001) but central obesity was related to NIDDM prevalence only in women. Our data suggest that the effect of centrality decreases at higher levels of centrality. While both BMI and centrality narrow the ethnic difference in NIDDM prevalence, Mexican Americans still have an increased risk of NIDDM (odds ratio = 2.33 in men and 1.80 in women), suggesting that other factors, possibly genetic, may also be important determinants of the ethnic differences in NIDDM prevalence.

Dietary intakes of macronutrients among Mexican Americans and Anglo Americans: the San Antonio heart study.

Twenty-four-hour dietary recalls were obtained on 1254 Mexican Americans (MA) and 916 Anglo Americans (AA), aged 25 to 64, as part of the San Antonio heart study, a population-based survey of cardiovascular risk factors from 1979 to 1982. In order to separate the effects of ethnicity from those of socioeconomic status (SES), we sampled subjects in three distinct neighborhoods: a low income MA neighborhood (barrio), a middle income neighborhood, and an upper income, predominantly Anglo, neighborhood. Intakes of protein, fat, and carbohydrate were similar to those found in other dietary surveys (NHANES, LRC). MA females living in the barrio consumed more cholesterol than either Anglos or MAs living in the other two neighborhoods. In MA males, the rise in the Hegsted Score with increasing SES paralleled the rise in LDL cholesterol with rising SES reported previously by our group. Females consumed a less atherogenic diet than males.

Dietary intakes of essential nutrients among Mexican-Americans and Anglo-Americans: the San Antonio Heart Study.

Dietary intakes of essential nutrients were measured as part of a population-based investigation of diabetes and cardiovascular risk factors in Mexican-Americans and Anglo-Americans in San Antonio, Texas. Twenty-four hour dietary recalls were conducted on 2134 individuals residing in three socioeconomically distinct neighborhoods: low-income, middle-income, and upper-income. Mean intakes of calcium, vitamin A and vitamin C were significantly lower among Mexican-Americans than among Anglos. Intake of vitamin C was most affected by socioeconomic status. Females of both ethnic groups consumed less than their RDA for calcium and iron. Intakes of B vitamins, phosphorus, and potassium were adequate, with few ethnic or socioeconomic differences.

Coffee consumption, diet, and lipids.

Recent reports suggest that coffee consumption is associated with increased serum cholesterol and triglyceride concentrations. The authors examined the association between serum lipids and coffee consumption and other caffeinated beverages as part of a population-based study of 1,228 women and 923 men, aged 25-64 years, in San Antonio, Texas, studied between October 1979 and November 1982. The study confirmed a positive relationship between coffee consumption and both total and low density lipoprotein cholesterol in both sexes which persisted after adjustment for age, ethnicity, obesity, cigarette smoking, and alcohol consumption. Neither tea nor cola consumption was associated with changes in serum lipids, suggesting that caffeine alone does not exert a direct effect on lipid levels. The possibility was examined that the coffee-cholesterol relationship might be due to a more atherogenic diet consumed by heavy coffee drinkers. In men, per cent calories from both total and saturated fat and dietary cholesterol intake increased with increased coffee consumption. Similar trends were not observed in women, however. The positive relationship between coffee and cholesterol may therefore be due to confounding effects of other aspects of the diet.