HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Related articles in AJCN Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Gao, S. K Articles by Fitzpatrick, A. L Search for Related Content PubMed PubMed Citation Articles by Gao, S. K Articles by Fitzpatrick, A. L Agricola Articles by Gao, S. K Articles by Fitzpatrick, A. L

Modifications to the Healthy Eating Index and its ability to predict obesity: the Multi-Ethnic Study of Atherosclerosis^1,2,3

¹ From the Department of Epidemiology, University of Washington, Seattle, WA (SKG, SAAB, and ALF); the Department of Psychiatry and Behavioral Sciences, University of Washington, Seattle, WA (LLF); the Division of Epidemiology and Community Health, University of Minnesota. Minneapolis, MN (PJS); and the School of Medicine, Wake Forest University, Winston-Salem, NC (GLB)

See corresponding editorial on page 14.

² Supported by contracts N01-HC-95159 through N01-HC-95165 and N01-HC-95169 from the National Heart, Lung, and Blood Institute and grant 1 R21 AT002152-01 from the National Center for Complementary and Alternative Medicine.

³ Address reprint requests and correspondence to SK Gao, Amgen Inc, 1 Amgen Center Drive, MS28-3A, Thousand Oaks, CA 91320. E-mail: sgao{at}amgen.com.

	ABSTRACT

TOP ABSTRACT INTRODUCTION SUBJECTS AND METHODS RESULTS DISCUSSION REFERENCES

 
Background: The Healthy Eating Index (HEI) is a measure of diet quality developed based on the Dietary Guidelines for Americans (DGA).

Objective: The objective was to assess the diet quality of a multi-ethnic population using and comparing the 2 HEIs, the updated HEI (HEI-05) based on the 2005 DGA and the original 1990 HEI (HEI-90), with the objective of predicting obesity outcomes.

Design: A longitudinal analysis of survey and clinical data from 6236 middle-aged and elderly white, African American, Hispanic, and Chinese participants of the Multi-Ethnic Study of Atherosclerosis (MESA) was conducted. Baseline diet quality was assessed with the use of HEI-90 and HEI-05. Baseline and 18-mo follow-up body mass index (BMI) and waist circumference (WC) data were predicted by using z score multiple regression models, and categorical obesity status was predicted by using multinomial logistic regression.

Results: Overall, the HEI-05 had larger z score β coefficients than did the HEI-90 (eg, in whites, –0.53 compared with –0.48 in baseline BMI, –0.54 compared with –0.47 in follow-up BMI, –1.67 compared with –1.56 in baseline WC, and –1.57 compared with –1.44 in follow-up WC). Among whites only, both HEIs were significant predictors of BMI and WC (all P < 0.001). The odds of being obese rather than normal weight were inversely related to HEI z scores primarily in whites (P < 0.05).

Conclusions: The changes to the 2005 DGA, as reflected by HEI-05, appear to better predict obesity outcomes in this multi-ethnic population, primarily in whites. Additional research on ethnic-specific DGA adherence and its relation to health outcomes is needed.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION SUBJECTS AND METHODS RESULTS DISCUSSION REFERENCES

 
Many diet indexes have been constructed to evaluate people's actual diet against dietary guidelines. The Healthy Eating Index (HEI; 1) is a measure of diet quality that assesses conformance to the Dietary Guidelines for Americans (DGA; 2) by the US Departments of Agriculture and Health and Human Services. The original HEI was based on the 1990 DGA, which has been updated every 5 y; the latest update was publicized in 2005 (1). We are proposing an updated HEI that reflects food and nutrition recommendations by specific energy requirements, which may be a better tool in assessing diet quality and associated health outcomes.

Although HEI is primarily a measure of overall diet quality, it may also be a predictor of obesity. First, HEI considers the recommended caloric intake in the scoring rules, which is also associated with obesity (1, 2). Second, HEI is a marker of energy density that is independent of caloric intake, and an energy-dense diet has a positive association with obesity (3).

This study was designed to examine the modification of HEI and its ability in predicting obesity outcomes and to test whether the predictive ability of the 2 indexes differs by ethnicity. Specifically, we wanted to examine 1) whether HEI-05 has a stronger association with cross-sectional obesity outcomes than does HEI-90, 2) whether HEI-05 is a better predictor of follow-up obesity outcomes than is HEI-90, and 3) whether HEI's predictive ability differs by ethnicity.

Moreover, diets are culturally relevant, and different ethnicities have different dietary cultures and distinct rates of chronic diseases (4, 5). More research is needed to study the validity of dietary indexes among non-African minority ethnicities. Thus, a secondary goal of our study was to examine the ethnic-specific diet conformance to the 1995 and 2005 DGAs.

	SUBJECTS AND METHODS

TOP ABSTRACT INTRODUCTION SUBJECTS AND METHODS RESULTS DISCUSSION REFERENCES

 
Data and study population
This study used data from the Multi-Ethnic Study of Atherosclerosis (MESA), a longitudinal population-based study of 4 ethnicities. Details of the MESA study were reported elsewhere (6). Between July 2000 and August 2002, 6814 men and women aged 45–84 y, who were free of any clinical cardiovascular disease and self-identified as white, black, Hispanic, or Chinese, were recruited from 6 US sites (Baltimore, MD; Chicago, IL; Forsyth, NC; Los Angeles, CA; northern Manhattan, NY; and St Paul, MN). Follow-up exams were scheduled every 18 mo. By January 2004, 6230 people had returned for exam 2 with a retention rate of 91.4% and an average follow-up time of 18 mo. Informed consent was obtained from all participants, and the study protocol was approved by the institutional review board at all 6 sites.

Both self-administered questionnaires and structured interviews were used to collect participant information, including a 120-item food-frequency questionnaire (FFQ) at the baseline visit. The FFQ included typical Hispanic and Chinese food and was modified from the validated Insulin Resistance Atherosclerosis Study to accommodate the ethnic diversity of the MESA participants (7).

Outcomes
Obesity outcomes of this study were assessed by using both anthropometric measures [body mass index (BMI; in kg/m²) and waist circumference (WC)] and categorical obesity status. During each visit, height and weight were measured with all participants (excluding pregnant women) wearing light clothing and no shoes. An Accu-Hite Stadiometer (Seca, Hamburg, Germany) was used to measure height, and a Detecto Platform Balance Scale (Titus Home Health Care, Alhambra, CA) was used to measure weight. Height was recorded to the nearest 0.1 cm and weight to the nearest 1 lb and converted to kg, from which BMI was calculated as weight(kg)/(height² in m). WC (abdominal girth) was measured by applying a Gulick II anthropometric tape horizontally at the level of the umbilicus and rounded to the nearest centimeter. Categorical obesity status was calculated by using the criteria of the Centers for Disease Control and Prevention: normal weight, BMI < 25.0; overweight, BMI = 25.0–29.9 as overweight; and obese, BMI 30.0 (8).

Healthy Eating Index
HEI is an index that comprises 10 components, half of which measure how diet conforms to the 5 main pyramid food group servings of grains, vegetables, fruit, milk, and meat/beans, whereas the other half measure intakes of total fat, saturated fat, cholesterol, sodium, and dietary variety (9). The original scoring system totals 100 points (optimal diet) and gives equal weight to all 10 components (0–10 each) while taking minimal account of age and sex differences. The DietSys Nutrient Analysis program was used to convert the FFQ responses in portion sizes and frequencies into average daily intake of nutrients. We used the HEI-90 scoring manual to score all components except for variety (9). For variety, scores were assigned using deciles of intake of the sum of food items consumed in the past 12 mo reported on the FFQ.

We constructed the HEI-05 using the same components, weighting, and scoring rules as for the HEI-90, but further adjusted the 5 food-group component scores to incorporate 12 levels of caloric need as specified in the 2005 DGA (2). For men aged 45–50 y, the recommended energy levels are 2200 (sedentary), 2500 (moderately active), and 2900 kcal (active). For men aged 51 y, the recommended energy levels are 2000 (sedentary), 2300 (moderately active), and 2600 kcal (active). For women aged 45–50 y, the recommended energy levels are 1800 (sedentary), 2000 (moderately active), and 2200 kcal (active). For women aged 51 y, the energy levels are 1600 (sedentary), 1800 (moderately active), and 2100 kcal (active).

Statistical analysis
We generated descriptive statistics of the main baseline characteristics of the MESA cohort, including age, sex, education, income, BMI, WC, physical activity, smoking, and alcohol use. HEI-90 and HEI-05 scores were computed for the overall population and by ethnicity. The Bland-Altman analysis with Pitman's test of difference in variance was used to test the agreement between HEI-90 and HEI-05.

We described each anthropometric measure by HEI-90 and HEI-05 quintiles of the MESA cohort. A global trend test was used to determine whether there was a pattern between HEI scores and anthropometric outcomes.

In multivariate analysis, we first calculated z scores for both HEI-90 and HEI-05 and then built separate linear regression models using the z scores of either HEI as the independent variable and each continuous anthropometric measure as the outcome variable. We built hierarchical z score regression models for each outcome: one unadjusted, one partially adjusted for sociodemographic variables, and one fully adjusted for sociodemographic variables plus total calorie intake, recreational physical activity (MET hours/wk), smoking status, and alcohol use (10). The use of the z scores enabled us to obtain coefficients from separate regressions that were on the same scale and directly comparable (11). Because log transformation did not change the results, we ultimately used linear regression models with the Huber-White sandwich estimators, which allow for valid inference without the assumption of a normal distribution in the outcome variable (12).

To model categorical obesity status, hierarchical multinomial logistic regression (MLR) was used (13). MLR simultaneously estimated 2 equations, the first comparing the probability of being overweight with being normal weight (reference category) and the second comparing the probability of being obese with being normal weight.

To test HEI's predictive ability by ethnicity, we also looked for possible effect modification by ethnicity in all models described in this section. An ethnicity-stratified result would be presented if the interaction term of ethnicity by HEI was significant (P < 0.05).

Each regression model was limited to participants with a nonmissing value in the outcome variable being modeled. Those with a missing value were not significantly different from the rest of the group in terms of other baseline characteristics (all P > 0.1 for t tests and chi-square tests). Estimates were corrected for multiple comparisons with Bonferroni corrections where applicable and adjusted for clustering within sites with Huber-White estimators. All statistical analyses were done by using STATA 9 software (Stata Corp, College Station, TX).

	RESULTS

TOP ABSTRACT INTRODUCTION SUBJECTS AND METHODS RESULTS DISCUSSION REFERENCES

 
Baseline characteristics of the MESA participants are presented in Table 1. Overall, MESA participants were fairly representative of the US population in the same age brackets (14). However, all characteristics differed significantly by ethnicity.

Table 2

Table 3

Table 4

Table 5

	DISCUSSION

TOP ABSTRACT INTRODUCTION SUBJECTS AND METHODS RESULTS DISCUSSION REFERENCES

In our study, we compared the ability of the HEI-90 with that of the HEI-05 in predicting both cross-sectional and follow-up obesity outcomes. The overall performance of the HEI-05, based on more specific energy levels, turned out to be only slightly better. Both HEIs were able to predict differences in obesity outcomes among the white population. However, there was no consistent evidence supporting such an association in other ethnicities.

In our analysis, the mean score of HEI-90 of the MESA population (61.9) was similar to that of the general US population in the same period, 63.8 (17). In comparison, the mean HEI-05 score of the MESA population was only 55.2. Similar differences were observed across ethnicities, mainly because of the changes in the recommended intake of the 5 main pyramid food groups. For the same energy level, compared with the 1990 DGA, the 2005 DGA recommends increased intakes of milk, vegetables, fruit, and meats but a reduced intake of grains (18). This explains why the MESA participants scored higher on the grain component but lower on all other pyramid food components on the HEI-05 and hence had a lower overall HEI-05 score. These suboptimal diet patterns were consistent with DGAI findings using the Framingham Heart Study Offspring Cohort and with CNPP's HEI-2005 findings using national surveys conducted from 2001 to 2002 (15, 19).

Of the 10 HEI components, the highest mean HEI subscore was obtained for sodium with both scales. It was probably due to the imprecise measurement of sodium on the FFQ, because sodium intake is known to be difficult to assess from dietary reports. Low subscores in food variety attributed to the lower overall HEIs of the Chinese Americans.

Our finding of inverse associations between HEI-90 and obesity outcomes was consistent with the majority of previous studies. For example, the inverse relation between HEI quintiles and BMI were supported by several studies that used either a population-level survey or specific groups of nurses and health professionals (15, 20, 21). The third National Health and Nutrition Examination Survey (NHANES III) found a graded increase in the odds ratio of obesity across the 3 HEI categories (22). However, these studies were not designed to assess BMI or obesity as outcome variables and thus were unadjusted for many of the confounders.

Using different measurements of obesity outcomes (BMI, WC, and obesity status) and after extensive control for potential confounders, we found that the HEI-05 scores were slightly more discerning than were the HEI-90 across all obesity outcomes, which support the validity of current dietary guidelines from which they were derived. Fogli-Cawley et al also found that individuals with higher DGAI scores had lower BMI (15). However, their study was designed mainly as a validity study and used cross-sectional data only, and their index was not directly comparable with the original HEI. The unique contribution of this study to the literature included not only a comparison of both HEIs by ethnicity but also the assessment of longitudinal associations.

Regarding the HEIs' performance in predicting obesity outcomes, we found them useful in whites, only fair in the Chinese and Hispanics, and very poor in the African Americans. It is possible that the HEI component scores apply differently to the latter 2 ethnic groups. These ethnic differences were most pronounced when BMI cutoffs (ie, obesity status) were used as the outcome, which may provide data pertinent to the long-standing debate on whether BMI cutoffs for obesity should differ by ethnicity (23, 24). A recent meta-analysis showed that, for the same percentage body fat, African Americans have a 1.3-unit higher BMI than whites, whereas the Chinese have a 1.9-unit lower BMI than whites (23). This is probably due to differences in bone mineral density between ethnic groups. This finding was also consistent with the trend we found in our analysis. Ethnic-specific cutoffs for BMI may prove to be of greater clinical value in this context. Because of limitations related to the study design, however, it was not possible to calibrate the cutoffs because it could also be an artifact of diet measurement errors.

The main limitation of this study was the use of FFQs to capture nutrient and calorie intakes. Despite being a better reflection of long-term diet, FFQs tend to have a lower correlation with true diet and are prone to greater measurement errors (25). For example, calorie intakes are underreported by an average of 30% on FFQs compared with objective measures of energy expenditure, and the amount varies depending on the characteristics of the responder, especially BMI. In our study, however, this should not have been a huge issue because the underreporting of calorie intakes that occurs more often in persons with a higher BMI would make our estimates on the association between HEI and obesity more conservative. Another limitation of this study was the snapshot measurement of diet, which failed to take account of any dietary changes over the 18-mo study period. The study would also benefit from more waves of data and a longer follow-up time as the MESA study continues into the future.

According to the American Dietetics Association, lack of understanding of nutrition guidelines and the misconceptions about "good" versus "bad" food are 2 of the major obstacles to a healthy diet (26, 27). These factors may have been present in the MESA cohort, but may not have been equally important across ethnicities. Whites and African Americans may be more likely to have misconceptions about "good" versus "bad" food as a result of earlier versions of dietary recommendations promoted in the United States (27). In contrast, Chinese and Hispanics, because of their more recent migration and lesser adaptation to the US culture, are less able to match the US-centric nutritional guidelines to their ethnic foods. The difference in HEI scores observed in this multi-ethnic group indicated that future dietary guidelines should be developed and promoted to address the dietary needs of different ethnic groups (26).

This study is among the first to have examined the relation between HEI and obesity outcomes using the 2005 DGA. By highlighting the specific nutrient gaps in the diet of each ethnic group, our study will help shape more effective policy interventions for promoting healthy diets and combatting the obesity epidemic. Because DGA is the most authoritative dietary recommendations given to the general US population, more studies are needed to assess the scientific value and clinical relevance of the 2005 DGA and to compare the efficacy of different DGA scoring algorithms for different health outcomes, especially in multiple ethnicities.

	ACKNOWLEDGMENTS

 
We thank the other investigators, the staff, and the participants of the MESA Study for their valuable contributions. A full list of participating MESA investigators and institutions can be found at http://www.mesa-nhlbi.org.

The authors' responsibilities were as follows—SKG: conducted the study design, statistical analysis, and drafting; ALF: supervised the work and secured data and funding; SAAB, LLF, PJS, and GLB: significantly contributed to the analysis and interpretation of the data, provided content-area expertise, and critically revised the manuscript. None of the authors had a conflict of interest.

	REFERENCES

TOP ABSTRACT INTRODUCTION SUBJECTS AND METHODS RESULTS DISCUSSION REFERENCES

 

1. Kennedy ET, Ohls J, Carlson S, Fleming K. The Healthy Eating Index: design and applications. J Am Diet Assoc 1995;95:1103–8.[Medline]
2. US Department of Health and Human Services, US Department of Agriculture. Dietary guidelines for Americans, 2005. 6th ed. Washington, DC: US Government Printing Office, 2005.
3. Ledikwe JH, Blanck HM, Kettel Khan L, et al. Dietary energy density is associated with energy intake and weight status in US adults. Am J Clin Nutr 2006;83:1362–8.[Abstract/Free Full Text]
4. Brown L, Mussell K. Ethnic and regional foodways in the United States: the performance of group identity. Knoxville, TN: University of Tennessee Press, 1984.
5. Lethbridge-Cejku M, Schiller JS, Bernadel L. Summary health statistics for US adults: National Health Interview Survey 2002. Vital Health Stat 10 2004;Jul;(222):1–151.
6. Bild DE, Bluemke DA, Burke GL, et al. Multi-ethnic study of atherosclerosis: objectives and design. Am J Epidemiol 2002;156:871–81.[Abstract/Free Full Text]
7. Liese AD, Schulz M, Moore CG, Mayer-Davis EJ. Dietary patterns, insulin sensitivity and adiposity in the multi-ethnic Insulin Resistance Atherosclerosis Study population. Br J Nutr 2004;92:973–84.[Medline]
8. Center for Disease Control and Prevention. Overweight and obesity: defining overweight and obesity. Version current 18 October 2007. Internet: http://www.cdc.gov/nccdphp/dnpa/obesity/defining.htm (accessed 11 Dec 2005).
9. Kennedy E, Ohls J, Carlson S, Fleming K. The healthy eating index final report. Alexandria, VA: US Department of Agriculture, Food and Nutrition Service, 1994.
10. Ainsworth BE, Jacobs DR Jr, Leon AS, Richardson MT, Montoye HJ. Assessment of the accuracy of physical activity questionnaire occupational data. J Occup Med 1993;35:1017–27.[Medline]
11. Fung TT, McCullough ML, Newby PK, et al. Diet-quality scores and plasma concentrations of markers of inflammation and endothelial dysfunction. Am J Clin Nutr 2005;82:163–73.[Abstract/Free Full Text]
12. Rogers WH. Regression standard errors in clustered samples. Stata Tech Bull Reprints. 1993:3:88–94.
13. Long JS, Freese J. Regression models for categorical dependent variables using Stata. 2nd ed. College Station, TX: STATA Press, 2005.
14. US Census Bureau. Statistical abstract of the United States, 2002: the national data book. Washington, DC: US Government Printing Office, 2002.
15. Fogli-Cawley JJ, Dwyer JT, Saltzman E, McCullough ML, Troy LM, Jacques PF. The 2005 Dietary Guidelines for Americans Adherence Index: development and application. J Nutr 2006;136(11):2908–15.
16. Guenther PM, Reedy J, Krebs-Smith SM, Reeve BB, Basiotis PP. Development and evaluation of the Healthy Eating Index-2005: technical report. Center for Nutrition Policy and Promotion, 2007. Internet: http://www.cnpp.usda.gov/HealthyEatingIndex.htm (accessed 15 January 2008).
17. Basiotis PP, Carlson A, Gerrior SA, Juan WY, Lino M. The Healthy Eating Index: 1999-2000. Washington, DC: US Department of Agriculture, Center for Nutrition Policy and Promotion, 2002. (CNPP-12.)
18. US Department of Agriculture, Center for Nutrition Policy and Promotion. The food guide pyramid. Washington, DC: US Government Printing Office, 1992. (USDA Home and Garden Bulletin no. 252.)
19. Guenther PM, Juan W, Reedy J, et al. Diet quality of Americans in 1994–96 and 2001–02 as measured by the Healthy Eating Index-2005. Nutrition Insight December 2007;37. Internet: http://www.cnpp.usda.gov/Publications/NutritionInsights/Insight37.pdf (accessed 15 January 2008).
20. McCullough ML, Feskanich D, Rimm EB, et al. Adherence to the Dietary Guidelines for Americans and risk of major chronic disease in men. Am J Clin Nutr 2000;72:1223–31.[Abstract/Free Full Text]
21. McCullough ML, Feskanich D, Stampfer MJ, et al. Adherence to the Dietary Guidelines for Americans and risk of major chronic disease in women. Am J Clin Nutr 2000;72:1214–22.[Abstract/Free Full Text]
22. Guo X, Warden BA, Paeratakul S, Bray GA. Healthy Eating Index and obesity. Eur J Clin Nutr 2004;58:1580–6.[Medline]
23. Deurenberg P, Yap M, van Staveren WA. Body mass index and percent body fat: a meta analysis among different ethnic groups. Int J Obes Relat Metab Disord 1998;22:1164–71.[Medline]
24. Stevens J. Ethnic-specific cutpoints for obesity vs country-specific guidelines for action. Int J Obes Relat Metab Disord 2003;27:287–8.[Medline]
25. Schatzkin A, Kipnis V, Carroll RJ, et al. A comparison of a food frequency questionnaire with a 24-hour recall for use in an epidemiological cohort study: results from the biomarker-based Observing Protein and Energy Nutrition (OPEN) study. Int J Epidemiol 2003;32:1054–62.[Abstract/Free Full Text]
26. Westenhoefer J, Pudel V. Pleasure from food: importance for food choice and consequences of deliberate restriction. Appetite. Jun 1993;20(3):246–9.

Related articles in AJCN:

Defining dietary consumption: is the sum greater than its parts?

Martha L Slattery
AJCN 2008 88: 14-15. [Full Text]  

This article has been cited by other articles:

				S. K. Gao, A. L. Fitzpatrick, B. Psaty, R. Jiang, W. Post, J. Cutler, and M. L. Maciejewski Suboptimal Nutritional Intake for Hypertension Control in 4 Ethnic Groups Arch Intern Med, April 13, 2009; 169(7): 702 - 707. [Abstract] [Full Text] [PDF]

				G. C. Frank Breakfast: What Does It Mean? American Journal of Lifestyle Medicine, March 1, 2009; 3(2): 160 - 163. [PDF]

				M. L Slattery Defining dietary consumption: is the sum greater than its parts? Am. J. Clinical Nutrition, July 1, 2008; 88(1): 14 - 15. [Full Text] [PDF]

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Related articles in AJCN Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Gao, S. K Articles by Fitzpatrick, A. L Search for Related Content PubMed PubMed Citation Articles by Gao, S. K Articles by Fitzpatrick, A. L Agricola Articles by Gao, S. K Articles by Fitzpatrick, A. L