The association between dietary inflammatory index, dietary antioxidant index, and mental health in adolescent girls: an analytical study

This descriptive-analytical study is a part of a larger study to identify the association of nutrient patterns with mental health in Tabriz, Iran. The study population included adolescent girls aged 14 to 16 years selected from high schools in the five regions of Tabriz, Iran. Sampling and data collection were carried out between November 2017 and July 2018.

The eligibility to participate in the study were: a) being high school students; b) being female; c) being 14–16 years old. Criteria for exclusion from the study included adherence to special diets, the presence of any apparent clinical illness including endocrine and chronic diseases (thyroid disorders, diabetes, heart, and renal failure) based on the patient's self-reported medical history. Also, subjects with caloric intake outside the range of 800–4200 kcal per day were excluded after collecting dietary intake data [60]. Of all eligible female students in each school, those whose parents did not sign the consent form were classified as “refusal to participate in the study” and were not entered the study. Sampling was done in two stages. In the first stage, schools were selected by cluster sampling from different areas of Tabriz (according to the number of high schools in each urban area and the number of students in these schools). Since SES is related to diet and psychological status, Tabriz city was first divided into three areas (good, moderate, and poor) in terms of SES. In the second stage, 3 schools from good, 3 from poor, and 4 from moderate SES status were selected (the total number of schools at this stage has been 10). The sampling was done in selected schools among girls aged 14 to 16 years. In this stage, 352 students from the selected schools were included based on the eligibility criteria by convenient sampling method (approximately 35 students from each school). Two more subjects were excluded after dietary data collection because their calorie intake was outside the range of 800–4200 kcal/day (Fig. 1). Then, a general questionnaire was completed by interviewing with participants.

All anthropometric measurements were performed twice. Then the average of the two measurements was recorded. All participants wore light clothing and no shoes; weight and height were measured using a standardized scale (Seca, Germany) and a portable stadiometer (Seca, Germany). The weight was logged to the nearest 100 g, and height was logged to the closest 0.5 cm. The body-mass index (BMI) was calculated as the ratio of weight in kilograms divided by the square of the height in meters (kg/m²). BMI was reported as BMI z-score standardized for 5–19 years old girls [61]. World health organization cut off points were used to define if participants were severe thin (BMI z-score < -3), thin (BMI z-score < -2), had a normal weight (-2 < BMI z-score < 0 and 0 < BMI z-score <  + 1), were overweight (BMI z-score >  + 1) or obese (BMI z-score >  + 2). Two trained nutritionists participated collecting anthropometric measurements.

Participants were asked to record the type and amount of foods and drinks they consumed in a 3-day food record, which is an open dietary report not yes/no questionnaire. They were asked to record on two specific consecutive weekdays and one weekend day. Participants were trained to fill out the food records, and instructions were provided on how to record the quantity using standard household measures. A trained dietitian finally checked all questionnaires in face-to-face interviews. Dietary intake analysis was performed using a food composition table from the last update of U.S. Department of Agriculture website to extract the food related data [62].

The 3-day food records were used to extract dietary data and calculate DII scores for all participants. In the food record, the portion size of food items is recorded, and then these data are converted to grams per day of macro and micronutrients. The DII was designed based on the literature published through 2010 and updated in 2014, linking diet to inflammation. Individuals’ intakes of food parameters on which the DII is based are then compared to a world standard database. A complete explanation of the DII is available elsewhere [23]. An explanation of validation work, including DII derived from both dietary recalls and a structured questionnaire similar to an food frequency questionnaire and related to high-sensitivity C-reactive protein interval values, is also available [23]. Briefly, for calculating DII, the dietary data were first linked to the regionally representative world database, which provided a robust estimate of each parameter's mean and standard deviation [23]. These then become the multipliers to express an individual’s exposure relative to the “standard global mean” as a z-score. This is achieved by subtracting the “standard global mean” from the amount reported and dividing this value by the standard deviation. This value is then converted to a centered percentile score to minimize the effect of “right skewing” (a common occurrence with dietary data). The centered percentile score for each food parameter for each individual was then multiplied by the respective food parameter effect score, which is derived from the literature review, to obtain a food parameter-specific DII score for an individual. All of the food parameter-specific DII scores are then summed to create the overall DII score for every participant in the study [23]. DII = b1*n1 + b2*n2………..b31*n31, where b refers to the literature-derived inflammatory effects score for each of the evaluable food parameters and n refers to the food parameter-specific centered percentiles, which were derived from this case–control’s dietary data. Of the theoretically possible list of 45 food parameters, a total of 31 were available from the 3-day food records and therefore could be used to calculate DII (energy, carbohydrate, protein, total fat, fiber, cholesterol, saturated fat, monounsaturated fat, polyunsaturated fat, omega-3, omega-6, niacin, thiamin, riboflavin, vitamin B12, vitamin B6, iron, magnesium, selenium, zinc, vitamin A, vitamin C, vitamin D, vitamin E, folic acid, beta carotene, garlic, ginger, onion, turmeric, saffron, pepper).

The DAI for all participants was calculated based on 3-day food records data. For estimating the DAI, each of the same six dietary vitamins and minerals was standardized by subtracting the global mean and dividing the result by the global standard deviation. The calculation of the DAI was done by summing up the standardized intakes of these vitamins and minerals and equal weight, as follows [47]:

The Depression, Anxiety, and Stress scale (DASS) is a validated and reliable questionnaire for assessing psychological disorders [63]. The DASS is a modified version of the Depression Anxiety Stress Scales-42 [63]. The validity and reliability of the questionnaire have been established in a sample of Iranian population with acceptable internal consistency (α = 0.84 to 0.91) and satisfactory convergent validity [64]. The internal consistency of the Persian DASS-21 was also very good in an adolescent sample (α = 0.86) [65].

The DASS consists of 21 items and three subscales for anxiety, depression, and stress, seven items in each category. There is a 4-point scale from 0 to 3 for scoring the items; the 0 is “did not apply to me at all,” and three is “applied to me very much or most of the time.” Item scores are summed for each of the seven-item subscales, ranging from 0 to 21 for each subscale and a total possible score of 63 for the entire scale. Lovibond and Lovibond’s [66] cut-off values were used to rate the severity of each of the outcomes which are as follows:

At the time of designing the study, based on our search, no similar study was found, so the correlation formula was used to determine the sample size. The correlation coefficient between the score of nutrient intake pattern and depression was about 0.15 based on a pilot study. The first type of error was 5%, the study power was 80%, and using the following formula, the sample size was estimated to be 347 people. Increasing to 360 people with anticipation of an overall dropout:

Normally-distributed continuous data are presented as mean and standard deviation (SD); qualitative data are presented as frequency (percent). The Kolmogorov–Smirnov test was used to determine the normality of distribution for continuous variables. One-way analysis of variance (ANOVA) was applied to compare continuous demographic variables. Categorical variables were compared using the chi-square test across tertiles of DII. To examine the mean differences of nutrient intakes across tertiles of DII, One-way ANOVA and Kruskal–Wallis H test were used in normal and non-normal distributed variables, respectively.

Univariate linear regression was conducted to determine the association between DII and DAI with depression, anxiety, and stress. The relationships of the DII and DAI with depression, anxiety, and stress were determined using generalized linear model (GLM) adjusted for confounders in 3 models. The models were defined as follows; model 1: crude, model 2: Adjusted for age and BMI, Model 3: Model 2 + Adjusted for energy intake, family income and mother and father education. The GLM with Gaussian family and identity link were used. Regarding confounding variables, based on prior knowledge and review of articles [23, 40, 59, 67], variables that were the common cause of exposure (DII, DAI) and outcome (Mental health profile) were selected as confounding variables, and their role was adjusted in the analysis. For example, family income is related to both DII, DAI, and mental health. Therefore, it can be considered as a confounding variable in the assessment of this relationship. P < 0.05 was considered statistically significant. Data analyses were performed using SPSS 26 (SPSS Inc., Chicago, IL, USA).