Food-group and nutrient-density intakes by Hispanic and Latino backgrounds in the Hispanic Community Health Study/Study of Latinos

Abstract

Background: Hispanics are a heterogeneous group of individuals with a variation in dietary habits that is reflective of their cultural heritage and country of origin. It is important to identify differences in their dietary habits because it has been well established that nutrition contributes substantially to the burden of preventable diseases and early deaths in the United States.

Objective: We estimated the distribution of usual intakes (of both food groups and nutrients) by Hispanic and Latino backgrounds by using National Cancer Institute methodology.

Design: The Hispanic Community Health Study/Study of Latinos is a population-based cohort study that recruited participants who were 18-74 y of age from 4 US cities in 2008-2011 (Miami, Bronx, Chicago, and San Diego). Participants who provided at least one 24-h dietary recall and completed a food propensity questionnaire (n = 13,285) were included in the analyses. Results were adjusted for age, sex, field center, weekend, sequencing, and typical amount of intake.

Results: Overall, Cubans (n = 2128) had higher intakes of total energy, macronutrients (including all subtypes of fat), and alcohol than those of other groups. Mexicans (n = 5371) had higher intakes of vitamin C, calcium, and fiber. Lowest intakes of total energy, macronutrients, folate, iron, and calcium were reported by Dominicans (n = 1217), whereas Puerto Ricans (n = 2176) had lowest intakes of vitamin C and fiber. Food-group servings reflected nutrient intakes, with Cubans having higher intakes of refined grains, vegetables, red meat, and fats and Dominicans having higher intakes of fruit and poultry, whereas Puerto Ricans had lowest intakes of fruit and vegetables. Central and South Americans (n = 1468 and 925, respectively) were characterized by being second in their reported intakes of fruit and poultry and the highest in fish intake in comparison with other groups.

Conclusion: Variations in diet noted in this study, with additional analysis, may help explain diet-related differences in health outcomes observed in Hispanics and Latinos.

Trial registration: ClinicalTrials.gov NCT02060344.

FIGURE 1.

Age- and sex-adjusted distribution of usual energy intake (kcal/d) by Hispanic and Latino backgrounds (n = 13,285). These distributions (means and percentiles) were estimated by using the National Cancer Institute method with a one-part nonlinear mixed model for repeated 24-h recalls with adjustment for age, sex, field center, weekend (including Friday), self-report intake amount (more, same, or less than usual), and recall sequence (first recall in person and second recall by phone). The distribution of usual energy intake was estimated empirically for the first recall adjusted for the second recall and for the usual self-report amount (as opposed to less or more). The simulated population had the same covariate patterns as the Hispanic Community Health Study/Study of Latinos for sex (women: 53.5%), age (42.4 y), and Hispanic and Latino backgrounds. Usual mean intake was significantly different in Hispanic and Latino backgrounds (overall test P < 0.001).

FIGURE 2.

Age- and sex-adjusted distribution of selected nutrient densities (per 1000 kcal) from food intake only by Hispanic and Latino backgrounds (n = 13,285). Nutrients presented are from food intake; dietary supplements and vitamins were not considered. These distributions (means and percentiles) were estimated by using the National Cancer Institute method with a one-part nonlinear mixed model for repeated 24-h recalls with adjustment for age, sex, field center, weekend (including Friday), self-report intake amount (more, same, or less than usual), and recall sequence (first recall in person and second recall by phone). The distribution of usual energy intake was estimated empirically for the first recall adjusted for the second recall and for the usual self-report amount (as opposed to less or more). The simulated population had the same covariate patterns as the Hispanic Community Health Study/Study of Latinos for sex (women: 53.5%), age (42.4 y), and Hispanic and Latino backgrounds. Usual mean intake was significantly different in Hispanic and Latino backgrounds (overall test P < 0.001). RAE, retinol activity equivalents.

FIGURE 3.

Age- and sex-adjusted mean (±SE) macronutrient composition (percentage of total energy) by Hispanic and Latino backgrounds (n = 13,285). Means were estimated by using the National Cancer Institute method with a one-part nonlinear mixed model for repeated 24-h recalls with adjustment for age, sex, field center, weekend (including Friday), self-report intake amount (more, same, or less than usual), and recall sequence (first recall in person and second recall by phone). The distribution of usual energy intake was estimated empirically for the first recall adjusted for the second recall and for the usual self-report amount (as opposed to less or more). The simulated population had the same covariate patterns as the Hispanic Community Health Study/Study of Latinos for sex (women: 53.5%), age (42.4 y), and Hispanic and Latino backgrounds. Bars with different lowercase letters are significantly different at P < 0.05 (ANCOVA and Tukey-Kramer paired comparisons).