The present study is part of a cross-sectional study, which investigated the association between dietary patterns with CVD risk factors and OS in patients with T1DM with the approval of Ethics Committee of Tehran University of Medical Sciences (IR.TUMS.REC.1394.1595), and written informed consent was obtained from all the participants. The study was conducted on 273 patients with T1DM from the Iranian diabetes society and Endocrine and Metabolism Research Institute of Tehran University of Medical Sciences in Tehran, Iran. According to the exclusion criteria (individuals who had reported energy intake outside the range of 500–3,500 kcal/day in women and 800–4,200 kcal/day in men) 261 patients were finally analyzed for the non-biochemical assessments and due to financial limitation, 81 patients recruited for biochemical examination using simple random sampling method. Inclusion criteria were patients who were diagnosed with T1DM for at least six months, aged between 18–35 and hemoglobin A1c (HbA1c) ≤ 8%. Participants were not enrolled into the study if they had BMI ≥ 40 kg/m2, drugs intake other than insulin to lower blood glucose (such as metformin), diagnosed CVD, cancer, kidney or liver disease, use of contraceptives, hormones and recombinant drugs, thyroid control drugs, taking weight reducing agents, anti-depressants and anti-anxiety drugs, use of any smoking (cigarette, hookah, tobacco pipe) and pregnancy or lactation.
Demographic and general characteristics
Information about age, sex, duration of DM, educational level, daily insulin dose, type of insulin and dietary supplements intake were collected by questionnaires.
Dietary intake assessment
Individual's typical dietary intake, during the last year, was evaluated using a semi-quantitative food frequency questionnaire (FFQ) with 147 food items , interviewed by a trained dietitian. The validity and reproducibility of the FFQ were determined previously for fruits, vegetables, and energy intake [26,27]. In this FFQ, there is a list of food items with a standard serving size that Iranian people commonly consume . Individuals were asked to report their usual frequency and amount of consumption of food items listed by day, week, and month over the previous year. The values listed for each food item were converted to gram using US Department of Agriculture (USDA) serving sizes whenever possible; if this was not possible, household measures were chosen and were then converted to grams. Energy and nutrient contents of food items were obtained from USDA food composition tables (FCTs) because Iranian FCTs are incomplete. The Iranian FCT was used for traditional food items that are not listed in the USDA FCT. Analyzing the energy and nutrients of each food item was done with the Nutritionist IV software version 3.5.1 , that modified for Iranian food .
The weight was measured by a nutritionist using a digital scale (GAIA 359 PLUS. Jawon Medical Co. Ltd., Gyeongsan, Korea), to the nearest 100 g, with minimal dress and no shoes. Height was measured to the nearest 0.5 cm with a tape while the patient standing without shoes. The BMI was calculated by dividing the weight (kg) by height (square meter). WC was measured using an elastic tape measuring midpoint between the iliac crest and lowest rib . Blood pressure was measured from the right hand of the participants to the nearest 2 mmHg, after at least 10 minutes rest, while sitting on a chair with a mercury sphygmomanometer. Overweight and obesity defined as BMI = 25–29.9 kg/m2 and BMI ≥ 30 kg/m2, respectively , central obesity as WC ≥ 80 cm in women and ≥ 94 cm in men . HTN was defined as blood pressure ≥ 140/90 mmHg .
Physical activity assessment
Physical activity of participants, during the previous week, was measured using the short form of International Physical Activity Questionnaire (IPAQ) . The average time that a person would normally spend on different activities each day was asked. Then, to measure the value of metabolic equivalent task (MET)-hour/week, the sum of frequency and duration of activities multiplied by the MET of activity. The reliability and validity of IPAQ across 12 countries was assessed previously .
After a 12–14 hour fasting at night, a trained nurse collected 5 mL of venous blood from the participants. Blood samples were collected in 2 separate tubes. One of the tubes was for separating serum and another tube containing ethylenediaminetetraacetic acid was used to separate the plasma. In order to separate the plasma samples from cells, the blood was centrifuged in 3,000 rounds for 10 minutes. Then the remaining blood was washed three times with sodium chloride solution 0.9 g/L. Separation of cell membranes was performed by centrifugation for 5 minutes at 4°C. Then hemolytic cells were used to determine the activity of antioxidant enzymes. The serum was separated from the blood by centrifugation for 10 minutes at 4°C. After that, all the blood samples were stored at −79°C. Blood glucose measurements were performed on the day of the test. This study was conducted with observance of the Declaration of Helsinki and the National Ethical Guidelines in Biomedical Research in Iran.
Measurements of serum glucose and lipids were performed, using Pars Azmoon kit (Pars Azmoon, Tehran, Iran). The measurement of triglyceride (TG) levels was conducted by colorimetric and photometry. Fasting blood glucose (FBG) and total cholesterol (TC) measurement were conducted by enzymatic colorimetric and single point with photometric method. The measurement of total antioxidant capacity (TAC) and activity of glutathione peroxidase (GPx) and superoxide dismutase (SOD) were measured by commercial kits following the manufacturer's protocol (ZellBio GmbH, Lonsee, Germany). The intra- and inter-assay coefficients of variation for SOD, GPx, and TAC were 5.8% and 7.2%, 3.5% and 4.7%, and 3.4% and 4.2%, respectively. The following formula was used for calculating low-density lipoprotein cholesterol (LDL-C) : LDL-C = TC − high-density lipoprotein cholesterol (HDL-C) −TG/5.0 (mg/dL).
Hypertriglyceridemia was defined as serum TG ≥ 150 mg/dL (1.7 mmol/L) [33,34], low HDL-C as serum HDL-C < 40 mg/dL (1.0 mmol/L) for men and < 50 mg/dL (1.3 mmol/L) for women [33,34], high LDL-C as serum LDL-C ≥ 100 (2.6 mg/dL) , hypercholesterolemia as TC > 200 mg/dL , high HbA1c as HbA1c ≥ 7 , hyperglycemia as FBG ≥ 100 mg/dL  and high LDL-C/HDL-C ratio as LDL-C/HDL-C > 3 in women and > 3.5 in men .
Statistical analysis was performed using SPSS software for Windows (version 23; SPSS Inc., Chicago, IL, USA; 2015) . Individuals who had reported energy intake outside the range of 500–3,500 kcal/day in women and 800–4,200 kcal/day in men were excluded from the study . In this order 3 men and 9 women were excluded from the data analysis and the final analysis was done on 261 participants. At first, energy‑adjusted DPI was determined by residual method , then according to DPI, participants were categorized into tertiles. Normality of data distribution was tested using graph and Kolmogorov-Smirnov test. Comparison of the general and nutritional characteristics of the participants between the tertiles of energy‑adjusted DPI was done using analysis of variance (ANOVA) or χ2 test depending on the type of variables. Means ± SD of CVD risk factors across tertiles of DPI were compared by ANOVA test for crude model and analysis of covariance for adjusted models which adjusted for age, sex, total energy intake (kcal/day), physical activity (MET/hour/week), BMI (kg/m2), diabetes duration (year), total daily insulin dose, education and dietary supplement intake. Further adjustment for intake of saturated fatty acids (SFA), mono-unsaturated fatty acids (MUFA), poly-unsaturated fatty acids (PUFA) and trans fatty acids was done for lipids levels. In addition, dietary intake of sodium and potassium was adjusted for systolic blood pressure (SBP) and diastolic blood pressure (DBP). Logistic regression test was used in crude and adjusted model to determine the odds ratio (OR) of cardiovascular risk factors and their 95% confidence interval (CI) in each tertile of DPI, which adjusted for age, sex, total energy intake (kcal/day), physical activity (MET/hour/week), BMI (kg/m2), diabetes duration (year), total insulin dose (unit/day), education, dietary supplement intake. These potential confounders were identified to be related with DPI according to the present study and previous literatures [18,19,21,22,24]. In all analyzes, the first tertile of DPI was considered as a reference and OR of the cardiovascular risk factor in the other tertiles was calculated towards it. Furthermore, to determine the overall trends of OR across increasing tertiles of DPI, the median of each tertile was used instead of the number of tertiles. P-value lower than 0.05 was considered significant.