Study Design and Respondents
This was an analytical cross-sectional study conducted on community-dwelling postmenopausal women in Kuala Lumpur and Selangor, Malaysia. The methodology of the study has been published elsewhere [34]. Briefly, a total of 211 eligible Malaysians, at least 5 years postmenopausal and absence from severe diseases were recruited based on two-stage sampling technique. Ethical approval was obtained from the Ethics Committee for Research Involving Human Subjects (project reference number: FPSK (FR16) P019) with all respondents provided written informed consent prior to study enrollment.
Sociodemographic and physical activity
Sociodemographic information was ascertained using a pre-tested structured questionnaire as discussed elsewhere [34]. The physical activity level of respondents was evaluated using the Global Physical Activity Questionnaire (GPAQ) [35], with the level of physical activity was classified according to Kyu et al. [36].
Physical measurement
Height of respondents was measured using a portable SECA stadiometer while the weight was measured at respondents’ fasting state using a TANITA digital weighing scale. Body Mass Index of respondents was computed as the ratio of weight (kg) to height in meter squared (m2), and the WHO 2000 [37] classification was used to classify BMI. Waist circumferences (WC) of respondents was measured using Lufkin anthropometric measuring tape while blood pressure of respondents was measured using a Digital Automatic BP monitor (OMRON HEM-907, Japan).
Blood collection and biochemical measurement
Fasting blood samples were collected from antecubital veins in EDTA (Becton Dickinson, NJ) and plain tubes by certified phlebotomists. The tubes were immediately placed in the icebox and were transferred to an analytical laboratory at which the blood samples were separated into plasma (glucose, vitamin D and DNA analysis) or serum (lipid profile analysis) and stored at -20°C until subsequent analyses. Fasting plasma glucose was determined by Hexokinase method using the Olympus AU analyzer (Beckman-Coulter, Inc., Fullerton, CA, USA) while fasting serum lipid profiles (total cholesterol, triglyceride, HDL-C and LDL-C) were determined using commercially available kits on a Hitachi 704 Analyzer, which is serviced by Roche Diagnostics. Total cholesterol and triglyceride were analyzed according to Cholesterol Oxidase/ Peroxidase and Glycerol Phosphate Oxidase/ Peroxidase method, respectively. On the other hand, HDL-C was measured by direct HDL method while LDL-C was estimated indirectly using the Friedewald formula. On the other hand, serum levels of 25(OH) vitamin D was determined by using the Siemens ADVIA Centaur Vitamin D Total assay (Siemens, Tarrytown, NY, USA), with the analytical measuring range between 4.2 to 150 ng/mL (10.5 154 to 375 nmol/L).
Estimation of Dietary Acid Load
Dietary intakes of respondents were assessed using a validated semi-quantitative food frequency questionnaire (sFFQ) adapted from the Malaysian Adult Nutrition Survey 2014 [38]. The sFFQ covers 165 food items frequently consumed among Malaysian, along with their standard portion sizes. After receiving detailed instructions from researchers, respondents indicated the typical frequency of consumption of foods and average amount (in household measures, eg cup, bowl, spoons), to allow the estimation of food intake over the past month [39]. Portion sizes were then converted to grams, based on the published household measurement. The validity and reliability of this questionnaire among Malaysian have been assessed previously [38]. Nutrients data (protein, phosphorus, potassium, magnesium and calcium) were then analysed using Nutritionist Pro™ Diet Analysis (Version 3.2, 2007, Axxya Systems, Stafford, TX, USA) software, with Nutrient Composition of Malaysia Foods (Tee, 1997) and Singapore Food Composition Database (Energy and Nutrient Composition of Food, 2011) as the primary databases. Dietary Acid Load of respondents was estimated according to potential renal acid load (PRAL) 33 equation as below:
PRAL (mEq/d) = 0.49 protein (g/d) + 0.037 phosphorus (mg/d) - 0.021 potassium (mg/d)
- 0.026 magnesium (mg/d) - 0.013 calcium (mg/d)
SNP selection, genotyping and quality control analysis
Candidate genes and SNPs were chosen from previously published literature which showed associations with metabolic traits [12, 15, 16, 22, 40-43]. IGF1 (rs35767, rs7136446) and IL6 gene (rs1800796) polymorphisms were selected in this study. SNPs sequences were referred from the https://www.ncbi.nlm.nih.gov/snp/ website. Genomic DNA was extracted from whole blood samples (EDTA tube) using a commercially available DNA extraction kit (QIAamp DNA Blood Mini Kit Qiagen, Hilden, Germany) according to the standard protocol. The extracted DNA concentration was quantified using a Spectrophotometer (Nanodrop, USA) and the qualities of the extracted DNA were assessed using 0.8% agarose gel electrophoresis. After all, samples showed good quantity and quality of DNA, each genotyping was further analyzed using Agena® MassARRAY. After the SNP detection process, Typer Analyzer was used to analyze the output data from Agena® Massaray.
Statistical analyses
Statistical analyses were performed using IBM SPSS 22 (SPPS Inc. Chicago, USA) with the level of significance set at p < 0.05. Hardy-Weinberg equilibrium (HWE) test for genotypic distribution was examined using the Hardy-Weinberg equilibrium exact test. Due to less than 5% of the respondents have TT (rs35767), CC (rs7136446) and CC (rs1800796) genotypes, they were combined with the heterozygous groups (rs35767: TT + CT; rs7136446: CC + CT; rs1800796: CC + CG) for further analysis. Prior to analysis, data quality was performed via SPSS to remove outliers, handling missing values and testing normality.
Descriptive data were presented as mean ± SD and range, or percentage. Metabolic traits cut off points were followed Harmonized criteria [43] considering it is the most updated and recommended by Joint Interim Statement (JIS) committees and suitable for the Asian population [43-45]. Next, comparisons of respondents’ characteristics between two groups genotypes of three SNPs were completed using independent Student’s t-test.
Furthermore, 18 models of three-step hierarchical multiple linear regression analysis were employed to test the contribution of variables as well as to determine the direct and interaction effects of DAL and genetic polymorphisms with each metabolic trait (SBP, DBP, WC, FBG, TG, and HDL and cholesterol ratio). Step one was used to determine the association between adjusted variables and metabolic traits while step two assessed the association between the DAL and gene polymorphism (rs35767, rs7136446 and rs1800796) on the dependent variable (SBP, DBP, WC, FBG, TG, and HDL and cholesterol ratio). These were followed by step three which aimed to add interaction term (DAL*gene polymorphism) and to determine the interaction effect of DAL with gene polymorphism for each metabolic trait.