In this cohort study, the subjects were selected from participants of the Tehran Lipid and Glucose Study (TLGS), a population-based prospective study performed to determine the risk factors for non-communicable diseases among the residents of the District 13, Tehran, capital city of Iran (20). The first examination survey was performed from 1999 to 2001 on 15,005 individuals aged ≥3 years, using the multistage stratified cluster random sampling, and follow-up examinations were conducted every 3 years; 2002–2005 (survey 2), 2005–2008 (survey 3), 2008–2011(survey 4), and 2012–2015 (survey 5) to identify recently developed diseases.
Of 14,712 individuals participating at baseline (surveys 3 and 4) , 9057 subjects were randomly selected for dietary assessment based on age- and sex-stratified random sampling, of whom 5,531 subjects aged≥30 years at baseline with completed data were included and followed until 2014. Subjects (n=227) with under- or over-reporting of energy intake (<800 or ≥4200 kcal/day) 15 and also those with a history of CVD (n=414) at baseline were excluded. Finally, after excluding participants who lost follow-up (n=15), 5,102 subjects remained and entered the analysis (Figure 1).
All participants signed a written informed consent form prior to the research. The study was conducted based on the Declaration of Helsinki and the study protocol was accepted by the ethics committee of the Research Institute for Endocrine Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran. All methods were performed in line with their relevant guidelines and regulations. Materials and methods were checked using the strengthening the reporting of observational studies in epidemiology (STROBE) statement.
2.2. Dietary intake measurements
Dietary assessment was performed by a valid and reliable 168-item semi-quantitative food frequency questionnaire (FFQ). Trained interviewers collected information on usual dietary intake through face-to-face private interviews. The frequency of eating each food item on a daily, weekly, or monthly basis was converted to daily intakes, and portion sizes were then converted into grams using household measures 16. Dietary intakes of vitamins E, C, and A, and zinc were calculated and considered in grams per week. Vitamin A consumption was considered as taking retinol and its esterified form.
Physical activity level was assessed using the Persian translation of the modifiable activity questionnaire (MAQ) with high reliability and moderate validity 17. The time and frequency of the light, moderate, high, and very hard intensity activity were obtained according to the list of common activities of daily living over the past year and the results were transformed into metabolic equivalent-hours/week (Met/h/week) 18.
Blood pressure and anthropometric measurements
Systolic and diastolic blood pressures (SBP and DBP) were measured twice (with a 30 s interval) in a sitting position after 15 min of rest.
The body weight was measured to the nearest 100 g, using a digital scale (Seca 707) while the subjects were minimally clothed and without shoes. Height was measured to the nearest 0.5 cm by a tape measure in standing position without shoes and with shoulders in normal alignment. Waist circumference (WC) was measured with a non-flexible tape meter without any pressure to body surface at the level of the umbilicus 19 and was taken at the end of a normal expiration over light clothing. Measurements were recorded to the nearest 0.1 cm.
Blood samples were drawn into vacutainer tubes between 7:00 and 9:00 a.m., after a 12–14-hour overnight fast in sitting position and centrifuged within 30 to 45 min of the collection. All biochemical analyses were performed using a Selectra 2 auto-analyzer at the TLGS research laboratory on the blood collection day. Fasting blood glucose (FBG) concentration was measured on the blood collection day using the colorimetric glucose oxidase procedure. The standard 2-h post-challenge blood glucose test was performed by oral administration of 82.5 g of glucose monohydrate solution (or 75 g of anhydrous glucose) for those who did not take glucose-lowering drugs.
High-density lipoprotein cholesterol (HDL-C) concentration was assessed after precipitation of the apolipoprotein B-containing lipoproteins by phosphotungstic acid. Total cholesterol (TC) and triglyceride (TG) were measured using the enzymatic colorimetric method. For TC assay, cholesteryl ester hydrolase was used to convert cholesteryl ester to cholesterol, which was then oxidized by cholesterol oxidase to cholesterol-4-en-3-one and H2O2. TG was broken down to glycerol and free fatty acids using lipoprotein lipase and glycerol was then phosphorylated to glycerol phosphate by glycerokinase. Glycerol phosphate was then converted to dihydroxyacetone phosphate and H2O2 by glycerol phosphate oxidase. The Friedewald equation (LDL-C = TC − HDL-C − TG/5) was used to calculate LDL-C concentration in samples containing TG 20.
Details of the collection of CVD outcome data have been described elsewhere 21. Coronary heart disease (CHD) events included definite myocardial infarction (diagnostic electrocardiographic [ECG] results and biomarkers), probable myocardial infarction (positive ECG findings plus cardiac symptoms or signs plus missing biomarkers or positive ECG findings plus equivocal biomarkers), proven CHD by angiography, unstable angina pectoris (new cardiac symptoms or changing symptom patterns and positive ECG findings with normal biomarkers), and CHD mortality. CVD was defined as stroke (a new neurological deficit that lasted more than 24 h), CHD events, or CVD death (a fatal stroke or fatal CHD).
Sex-specific multivariable risk functions (“general CVD” algorithms) were derived that incorporated age, TC, HDL-C, SBP, treatment for hypertension, smoking, and type 2 diabetes status 22. Hypertension was defined as SBP≥140, DBP≥90 mmHg, or receiving antihypertensive drug 23.
Statistical analyses were carried out using the Statistical Package for Social Sciences (version 21.0; SPSS). A two-tailed P value <0.05 was used to determine statistical significance. We used a Chi-square test and Student’s t-test for qualitative and quantitative variables to compare the results obtained from the male and female subjects. For non-normal nutritional and biochemical variables (TG concentration), log-transformed values were used for statistical analysis. The hazard ratio (HR) and 95% confidence interval of incident CVD were assessed using multivariable Cox proportional hazard regression models. Person-years of the follow-up were calculated for each individual between the date of study inclusion and the date of diagnosis of CVD, death, or end of the follow-up, whichever came first. Survival time for censored individuals was calculated as the interval between the first and last observation dates. The event date was considered as the middle-time between the date of follow-up visit, at which the events were diagnosed for the first time, and the most recent follow-up visit preceding the diagnosis. The incidence of CVD during the follow-up period was considered as dichotomous variables (yes/no) in the models. Vitamins E, C, and A, and zinc intakes were categorized into quartiles, given the first quartile as the reference. The median of each quartile was used as a continuous variable to assess the overall trends of HRs across quartiles of dietary intakes of vitamins A, E, and C and zinc in the Cox proportional hazard regression models. Trend test was applied for evaluating dose-response effects in association studies. The proportional hazards assumption of the multivariate Cox models were assessed using Schoenfeld’s global test of residuals.
The confounders were selected based on previous studies and included in the uni-variable Cox regression model. A two-tailed P value <0.20 was used for inclusion in the model 24. The Cox regression models were adjusted for age, sex, CVD risk score (continuous), family history of CVD, physical activity (continuous), dietary intakes of total energy, fiber (g/1000 Kcal) and total fat (percentage of energy).