This cross-sectional study was performed based on data from the Study on the Epidemiology of Psychological, Alimentary Health and Nutrition (SEPAHAN) project, which was a cross-sectional study looking at the prevalence of functional gastrointestinal disorders (FGIDs) and their relationship with lifestyle factors and psychological disorders. Details about SEPAHAN project have been published earlier (21). This study was performed among Iranian general adults working in 50 different healthcare centers affiliated to Isfahan University of Medical Sciences (IUMS) across Isfahan province. To collect information about anthropometric measures, demographic and lifestyle factors, including dietary intakes and physical activity, self-administered questionnaires were distributed among 10,087 subjects, and 8,691 participants returned the completed questionnaires (response rate: 86.16%). In the current analysis, we excluded subjects who reported their total daily energy intake outside the range of 800–4,200 kcal/d. We also excluded those who had missing data on any relevant variable. These exclusions resulted in a dataset of 6,724. All participants provided written inform consent forms. Although the protocol of SEPAHAN study was approved by the Regional Bioethics Committee of Isfahan University of Medical Sciences, the current study was separately approved by the Research Council of School of Nutritional Sciences and Dietetics of Tehran University of Medical Sciences, Tehran, Iran (Ethics code: IR.TUMS.VCR.REC.1398.131).

A self-administered, Willett-format, Dish-based, 106- item Semi-quantitative Food Frequency Questionnaire (DS-FFQ), was used to assess dietary intakes. The questionnaire was designed and validated for use in Iranian adults. Details on design, foods included, and the validity of this questionnaire has been reported elsewhere (22). Briefly, we provided a comprehensive list of foods and dishes commonly consumed by Iranian adults. Then, those foods that were nutrient-rich, often consumed, or contributed to between-person variation were selected. Eventually, this process led to remaining of the 106 food items in 5 various categories in the questionnaire: (1) mixed dishes (cooked or canned, 29 items); (2) grains (different types of bread, biscuits, cakes and potato, 10 items); (3) dairy products (dairy, butter and cream, 9 items); (4) fruits and vegetables (22 items); and (5) miscellaneous food items and beverages (including fast foods, nuts, sweets, desserts and beverages, 36 items). In order to provide precise and accurate estimates, the portion size of foods and mixed dishes as a unit with the same perception were given to all people. Nine multiple choice frequency response categories ranging from “never or <1/month” to “≥12/day” were provided for reporting dietary intakes of participants. The number of response categories for the food list varied from 6 to 9 choices. For foods consumed infrequently, we omitted the high-frequency categories, while the number of multiple choice categories increased for common foods with a high intake. Finally, daily intake of all food items was computed and then converted to grams per day using household measures (23). Daily nutrient intakes of each participant were estimated based on the US Department of Agriculture's (USDA) national nutrient databank (24). The validity of DS-FFQ was examined in a subgroup of 200 participants randomly selected for the SEPAHAN project. All participants in the validation study completed the DS-FFQ at study baseline and 6 months later. During this validation study, participants provided three detailed dietary records that were used as the gold standard. As shown in earlier studies (22), it seems that this questionnaire provides reasonably valid measures of long-term dietary intakes. Also, some recent studies have shown that FFQ was a valid questionnaire for assessing dietary quality scores (25).

Dietary TAC was obtained from previous studies, based on the ferric reducing-antioxidant power (FRAP) values of 100 food items from the 106 selected food items. The food parameters that were not included in the TAC construction included salt, pepper, sugar, sugar loaf, gaz and nabat (traditional Iranian sweets), because the TAC value for these dietary factors was zero. The FRAP assay is a tool measuring the ability of dietary antioxidants to reduce ferric to ferrous ions. The FRAP values express as mmol per 100 grams of foods (mmol/100 g) (26). For similar food items in Iranian culture (e.g., several types of breads), we calculated the overall mean value. Finally, the frequencies of consumption of each food item were multiplied by their related FRAP values and then summed up to obtain dietary TAC for each participant.

A method described by Kant et al. (27, 28) was used for scoring dietary diversity. This method was based on five groups including grains, vegetables, fruits, meats, and dairy, all food groups in the USDA food guide pyramid. The grains group was composed of seven components: refined bread, macaroni, whole grain bread, corn flakes, rice, biscuit, and refined flour. As we had no data about intake of corn flakes, we decided to consider six components. Fruit was defined by summing up fruit and fruit juice, berries, and citrus fruits. In terms of vegetables, we summed up mixed vegetables, potato, tomato, other starchy vegetables, legumes, yellow vegetables, and green vegetables. The group of meat was composed of red meat, poultry, fish, and eggs and the group of dairy was composed of milk, yogurt, and cheese.

Participants were considered as a “consumer,” and scored as 1, for each component of food groups if they had intakes higher than median levels; otherwise they were given the score of 0. Then the scores for components in each food group were summed up to have total score of that food group. Then, we divided total scores obtained in each group to the number of components in that group. This value was then multiplied by 2. Total DDS for each participant was then computed by summing up the figures for food groups. For example, in the grains group, if a person had dietary intakes of whole grain bread, macaroni, and rice higher than the median values, her or his score was calculated as (3/6) × 2 = 1. Therefore, the diversity score for the grains group would be 1 for that person. After computing the diversity score for the other four groups in that person, total DDS was computed. Therefore, minimum and maximum scores of total dietary diversity for each participant were between 0 and 10.

To calculate AHEI-2010, a method designed by Kennedy et al. was used (29–31). AHEI-2010 consisted of eleven components: fruit, vegetables, whole grains, nuts and legumes, long-chain n-3 fats (DHA and EPA), PUFA, alcohol consumption, sugar-sweetened drinks and fruit juice, red and processed meats, trans-fat and sodium. In the current study, alcohol consumption was not included into the score, because of the lack of information in the original dataset. To construct the index, first we obtained energy-adjusted intakes of the above-mentioned components by using the residual method (32). Next, participants were classified based on decile categories of energy-adjusted intakes of these components. As scoring by deciles would be least prone to misclassification, we used decile categories of components instead of other classifications. Individuals in the highest deciles of fruits, vegetables, whole grains, nuts and legumes, long-chain n-3 fats and PUFA were given a score of 10, and those in the lowest decile of these items were given a score of 1. Individuals in the other deciles of these components were assigned the corresponding scores. Regarding sugar-sweetened drinks and fruit juice, red and processed meat, trans-fatty acids, and sodium intake, the lowest decile was given a score of 10 and the highest decile was given a score of 1. Those in deciles 9, 8, 7, 6, 5, 4, 3, and 2 of these components were given scores of 2, 3, 4, 5, 6, 7, 8, and 9, respectively. The whole AHEI-2010 was computed through summing up the scores of its components ranging from 10 to 100.

Required information on other variables including age, sex, marital status, smoking status, and education was obtained from demographic and medical history questionnaires. Physical activity was assessed using the General Practice Physical Activity Questionnaire (GPPAQ) (18). Based on participants' responses, they were classified into 4 categories; (1) inactive, (2) moderately inactive, (3) moderately active, (4) active. However, in the current study, due to low number of subjects in some of the above-mentioned categories, individuals in the “inactive” and “moderately inactive” groups were combined and were defined as those with “sedentary physical activity”. Similarly, individuals in the “moderately active” and “active” categories were combined and then defined as “physically active”. Anthropometric measures including weight and height were assessed using a self-administered questionnaire. Body Mass Index (BMI) was calculated by dividing weight (kg) to height (m²). The correlation coefficient for computed BMI from self-reported values, and the one from measured values was 0.70 (P < 0.001).

We classified participants based on tertiles cut-off points of dietary TAC. General characteristics of study participants across tertiles of dietary TAC were presented as means ± SDs for continuous variables and percentages for categorical variables. To examine the differences across tertiles, we used ANOVA for continuous variables and chi-square test for categorical variables. The multivariable-adjusted means for AHEI and DDS across tertiles of dietary TAC were computed and compared using ANCOVA. In these analyses, energy intake was controlled for in the first model. Further adjustments were made for age (continuous) and sex (male/female) in the second model. BMI (continuous) was controlled for in the third model. Cross-classification of participants across tertiles of TAC, DDS and AHEI was examined. In this analysis, exact agreement was defined when individuals were classified in the same tertiles based on TAC and DDS or AHEI. When individuals were classified in the opposing tertiles, this was considered as gross misclassification. All statistical analyses were done using the Statistical Package for Social Sciences (version 20; SPSS Inc.). P < 0.05 was considered as statistically significant.