Study population
This study was conducted during the baseline (cross-sectional) phase of a randomized, double-blind trial, aimed at investigating the effect of different doses of one-year vitamin D supplementation on cIMT and other bone and cardiovascular variables in children and adolescents with an age- and sex-specific body mass index (BMI) Z-score ≥ 1 (according to the criteria established by the World Health Organization). Participants, aged 6 to 13 years, were recruited from primary schools located in three Tehran (Iran) districts. The baseline screening took place between June 2016 and March 2017. Individuals were eligible for inclusion if they had no known medical illnesses such as diabetes, liver, or kidney diseases, were not taking any dietary supplements, or using pharmaceutical agents that affect glucose and lipid metabolism. A total of 378 participants met these selection criteria and were enrolled at the baseline. The sample size for the trial was determined by a sample size calculation designed to achieve 80% power to detect a 0.02 mm difference in cIMT with a 0.05 standard deviation (95% confidence interval) between different groups of the original trial.
Measurements
Fat and soft lean mass were assessed by portable bioelectrical impedance analyzer GAIA 359 PLUS 8-contact electrode bioelectrical impedance (BIA) system (Jawon Medical Co. Ltd., Shinsang, Korea). During the measurement, all participants removed their shoes and socks, wore light clothing, and stood with the soles of their feet making contact with the foot electrodes, while holding the hand electrodes in their bare hands. All necessary variables including sex, height, and age, were input into the instrument. The BIA system measured the impedance via a tetra-polar multi-frequency (5, 50, 250 kHz) electrode method. Resistance to the 500 μA current was measured while the participants stood motionless on the analyzer’s platform, their arms raised about 30°, while lightly holding onto the hand electrodes. For the data interpretation, we used the manufacturer’s software in its “standard” setting.
Body weight was measured using the scale function of the GAIA 359 PLUS (100-g accuracy) while the subjects were standing barefoot and wearing light clothing. Height was measured while participants were standing without shoes and shoulders in normal alignment, using a stadiometer with an accuracy of 0.5 cm. Body mass index (BMI) was calculated as weight (in kilograms) divided by height (in meters) squared (kg/m2). Waist circumference (WC) was measured down to the nearest 0.5 cm at the level of the umbilicus, over light clothing, and without any pressure, using a tape meter.
Blood pressure was measured twice, at least 1 min apart, using a mercury sphygmomanometer and the Korotkoff sound technique, with an accuracy of 2 mmHg, after participants had been resting for 15 min on a chair; the average of both measurements served as the participant’s final pressure; systolic/diastolic blood pressures (SBP/DBP) were determined from the first onset/disappearance of sound.
Blood samples were collected from all study participants between 8:00 and 10:00 AM, after 10–12 h of overnight fasting. All the blood analyses were carried out at the RIES research laboratory. Fasting blood glucose (FBG) was measured by enzymatic colorimetric method using glucose oxidase (Pars Azmoon, Tehran, Iran) and the Selectra 2 auto-analyzer (Vital Scientific, Spankeren, Netherlands) with intra- and inter-assay coefficients of variation (CVs) of 1.1 and 1.4%, respectively. Fasting serum insulin was determined by the electrochemiluminescence immunoassay (ECLIA) method, using Roche Diagnostics kits and the Roche/Hitachi Cobas e-411 analyzer (Roche Diagnostics, GmbH, Mannheim, Germany). Intra- and inter-assay CVs were 1.3 and 2.5%, respectively. To monitor the quality of measurements, assayed serum controls in different concentrations were used for the glucose (Pars Azmoon, Tehran, Iran) and insulin assays (Lyphochek Immunoassay Plus Control, Bio-Rad Laboratories).
Physical activity was assessed using the Modifiable Activity Questionnaire (MAQ) to calculate metabolic equivalent task minutes per week; the high reliability (97%) and moderate validity (49%) of the Persian translated MAQ in adolescents have been ascertained previously [
15].
Pubertal status was classified according to the Tanner scale by a well-trained endocrinologist, dividing the participants into 2 groups based on breast and genital stages: pre-pubertal (boys at genital stage I, girls at breast stage I) and pubertal (boys at genital stage ≥II, girls at breast stage ≥II).
Carotid intima-media thickness was measured by a well-trained radiologist (P.D.). Participants were examined in the supine position with the head slightly extended and rotated away from the examiner. The carotid arteries were interrogated using a high-resolution Samsung ultrasound machine (model UGEO WS80A) with a linear-array transducer operating at a frequency of at least 7 MHz. Depth, gain, and focus was adjusted for each participant individually so that the arterial lumen was completely anechoic and in the center of the image. Common cIMT was measured from longitudinal B-mode images of the distal 1 cm of the far wall of each common carotid artery (CCA), between the intimal-luminal and the medial-adventitial interfaces of the carotid artery wall, represented as a double-line density on the ultrasound image. In rare cases where appropriate images of the distal CCA could not be obtained, proximal or mid-CCA images were used for IMT measurement. Measurements were performed using the automated edge-tracking software (automated IMT calculator) which obviated the need to perform manual measurements.
Statistical analysis
Quantitative variables were expressed as the mean ± SD or median (interquartile range) while categorical variables were expressed as percentages. Continuous and categorical variables were compared according to sex using a t-test and a Chi square-test, respectively. Multiple linear regression analysis was used to evaluate the association of markers of insulin metabolism and cIMT. A proportional odds model (POM) was used to test for a possible association between an insulin metabolism marker increase of approximately 1 SD and tertiles of cIMT using the ranges: cIMT< 0.38, 0.38 ≤ cIMT< 0.42, and cIMT≥0.42 mm. The proportional odds assumption was assessed through a chi-squared score test for each covariate to examine whether the assumption was violated. Proportional odds assumptions were generally appropriate. To test whether pubertal stage or gender could affect the contribution of fasting insulin, the interaction terms for [female × fasting insulin] and [pre-pubertal × fasting insulin] were used as separate inputs in our multivariable linear regression models. The likelihood ratio test for interaction terms was not significant. Nevertheless, the interaction term for gender was of biological importance which is why each model was performed in the subset specified by gender. Waist circumference, body fat percentage, physical activity, and Tanner stages were considered as confounding variables. Collinearity was assessed using Pearson’s rank correlation coefficient and variance inflation factor. No collinearity was found between body fat percentage and WC. All analyses were performed using IBM SPSS for Windows, version 20 (SPSS, Chicago, IL, USA) with the significance level set at P < 0·05 (two-tailed).