Sample size calculation
In this paper, the Sample size was calculated according to following formula:
N= (([(Z1−α+Z1−β) ×√1−r2]/r)2 +2), When r=0.15 β=0.95 and α=0.05, the N was equal to 634.
Study design and subjects
For this cross-sectional study, 636 T2DM patients aged 35–65 years of both genders (252 men and 384 women), were randomly recruited from diabetes referral centers including Gabric Diabetes Association, Iranian Diabetes Society, and other health centers in Tehran during June 2011 to October 2012. All the subjects had either fasting blood glucose levels of ≥ 126 mg/dl or were under treatment with medication (oral) by a physician. Persons under 35 or over 65 years of age, insulin administrating patients, pregnant or lactating women, were excluded. The current study was approved by the ethics committee of Tehran University of Medical Sciences (TUMS) (Ethnic number: IR.TUMS.VCR.REC.1395.15060) and the informed consent form was obtained from all participants. Demographic and general data including age, gender, job, and education status, smoking, and alcohol abuse, lipid-lowering drug consumption, family history of T2DM, and other diseases were collected through interviews.
Assessment of anthropometric measurement and physical activity
Anthropometric data were measured by standard methods. Weight was measured in the fasting state, with the minimal clothing, by using Seca falcon scales, with an accuracy of 100 g. Height was measured using a Seca height gauge with an accuracy of 0.5 cm according to standard protocols (13). BMI was calculated as body weight (kg) divided by the square of the height (m2). Physical activity was calculated as metabolic equivalent of task (MET h/day) (14) by a validated and reliable physical activity questionnaire (15, 16).
Assessment of dietary intake
The participant’s usual dietary intake during the last year was evaluated through face-to-face interviews conducted by a trained dietitian using a semi quantitative food frequency questionnaire (FFQ) for 148 food items. This questionnaire was validated by Esmaillzadeh et al (17). The subjects were asked to report the frequency of food item consumption in a day, a week, a month or a year. The amounts listed for each food were converted to grams per day using household measures (18). Nutritionist III software (version 7.0, N-Squared Computing) was employed to assess the energy and nutrient intake.
Assessment of dietary indices
Three dietary indices were used for evaluating diet quality: The Healthy Eating Index (HEI), a measure for evaluating alignment of dietary intake according to the 2015-2020 dietary guidelines for Americans (DGA) based on a 1000 kcal/day diet. HEI-2015 score ranged from 0 to 100 and consisted of 13 components (total fruits, whole fruits, total vegetables, greens and beans, seafood or plant protein, and total protein foods can receive a score ranging from 0 to 5 , whole grains, dairy, fatty acids can receive a score ranging from 0 to 5 and refined grains, saturated fats, sodium and added sugars are moderated (higher intakes receive lower scores) (19). We calculated the score based on responses from the FFQs.
The Diet Quality Index-International (DQI-I) was a second dietary measurement which focuses on four main aspects of a healthy diet (variety, adequacy, moderation, and overall balance). The score for each category is calculated as the sum of the scores for each component in that category. Overall food group variety including meat/poultry/fish/eggs/dairy/beans/grain/fruit/vegetable (0-15 points) and Within-group variety for protein source including meat/poultry/fish/dairy/beans/eggs (0–5 points) , protein sources, vegetable, fruit, grain, fiber, protein, iron, calcium, vitamin C group (0-5 points), moderation foods such as total fat, saturated fat, cholesterol, sodium and junk foods (0–30 points), macronutrient and fatty acid (0–10 points) (20). The total DQI-I score (ranging from 0 to 100 points) is the sum of the scores for the four categories.
Moreover, Phytochemical Index (PI), defined as the percentage of dietary calories derived from foods rich in phytochemicals including fruits, vegetables, whole grains, nuts, seeds, vegetable juices, soy products, olive oil. The dietary phytochemical index (DPI) was calculated based on the modified method previously developed by McCarty (21); [PI= (phytochemical-rich foods g/d/ total food intake g/d) ×100].
Biochemical assessment and genotyping
Venous blood samples were collected after 12-h overnight fasting. The total antioxidant capacity (TAC) of serum was measured by spectrophotometry. TAC measurement evaluates the overall power of all antioxidants in the body (21, 22) Serum enzymatic activity of superoxide dismutase (SOD), as an enzymatic antioxidant (22) was estimated by colorimetric method (Cayman Chemical Company, USA). Interleukin-18 (IL-18), Pentrexin-3 (PTX3), and 8-isoprostane F2α (PGF2α) were measured using ELISA method (Shanghai Crystal Day Biotech Co., Ltd). The sensitivity of IL-18 and PTX3 ELISA kit was 28 ng/ l and 0.05 ng/ml, respectively. Genomic DNA was isolated from whole blood using salting-out extraction method (23). Polymerase chain reaction (PCR) was used for genotyping the BDNF Val66met, followed by 8% polyacrylamide gel electrophoresis.
Normal distribution of data was measured using Kolmogorov–Simonov test. Logarithmic transformations were applied to variables with skewed distribution. The subjects were divided in to two genotype groups: Val/Val and Val/Met, Met/Met. The data were presented as frequency (%) for categorical variables and as mean ± SD for continuous variables.
Independent T-test was used to compare the quantitative variables between the two groups and chi-square test was used In order to compare the qualitative variables.
The association between diet quality indices and anthropometric or biochemical parameters were evaluated by one-way analysis of variance (ANOVA).
The interaction between BDNF Val/Met polymorphism and DEI, DQI, and PI on anthropometric indices, serum inflammatory, and oxidative stress markers were tested using analysis of covariance (ANCOVA) test in two multivariate interaction models, before and after adjustment for potential confounders including age, sexuality, smoking, alcohol consumption, and physical activity. The data were analyzed by IBM SPSS (SPSS Inc., Chicago, IL, USA, version 26) and P value < 0.05 was considered as statistically significant.