Prevalence of obesity in a rural Asian Indian (Bangladeshi) population and its determinants

This study was part of a large longitudinal epidemiological study on diabetes and related cardiometabolic diseases in rural Bangladesh which has been described previously [17]. This population-based cross-sectional study was conducted in a rural community called Chandra, 40 km. north of Bangladesh’s capital, Dhaka during March to December 2009. A total of 10 villages were randomly selected from 25 villages with a population of approximately 20,000 aged ≥20 years. In order to determine the required sample size, the formula: n = Z²PQ/d² was used, where, Z = 1.96, P for prevalence (DM & impaired glucose regulation) rural area from the previous study [18] i.e. 0.15; Q = 1-P i.e. 0.85 and d = allowable error of known prevalence i.e. 0.09 x 0.15. Ideally, it should be 0.05 x 0.15; but, to be on safe estimation with minimum sample size we allowed only 9 % (or .09) error of prevalence. Thus the calculated sample size was, n = 2687. Around 3000 individuals (both male and female) aged ≥20 years of those 10 villages were invited to participate in this study by following simple random procedure from the record of the voters lists number; among them 2376 (79.2 %) agreed individuals were investigated. Regarding non-response, survey timetable in early morning with 2-3 h study time and secondly, the improvement of healthcare facilities in Government Community Clinics may reduce the interest in obtaining a health assessment through the survey. This epidemiological survey was conducted in 20 selected spots of 10 villages. The inclusion criteria's were: both gender, aged ≥20 years and willing to participate. Exclusion criteria included pregnant women, and subjects with self-reported or medical history of myocardial infarction, renal disease, liver disease, tuberculosis, malignant disease and any severe infection at the time of screening.

Once the selection procedure was completed, participants were requested to visit a nearby field center. The socio-demographic data was collected by interviewing the participants using a predesigned pretested questionnaire. Survey procedures included completing a questionnaire on socio-demographic, physical activity, dietary and smoking information, anthropometric measurements, blood pressure measurement and laboratory investigations. Anthropometric measurements, such as height, weight and waist and hip circumferences, were taken with the participants wearing light clothes and without shoes. The weight was taken to the nearest 0.1 kg by modern electronic digital LCD weighing machines (Best Deluxe Model; Bathroom, Dhaka, Bangladesh) placed on a flat surface. The scales were calibrated everyday against a standard (20 kg). Height was taken while the participants stood in erect posture, touching the occiput, back, hip and heels on a straight measuring wall, while the participants looked straight ahead. BMI was calculated as the weight (kg) divided by square of the height (m²). Waist circumference was measured by placing a tape horizontally midway between the lower margin of the last palpable rib and iliac crest on the mid-axillary line. Hip circumference was measured at a level parallel to the floor, at the largest circumference of the buttocks. WHR was then calculated from waist and hip circumference (cm). Food consumption and portion sizes were calculated with the 24-h recall method. Participants were asked to report on all foods and drinks consumed in the previous 24 h (the previous day), in direct chronological order from the first foods in the morning to the last foods before going to bed. A range of local household utensils: glasses, spoons, cups and plates were used for estimating the amount of foods and beverages actually consumed by the respondents. The use of these local utensils acted as visual aids to increase the accuracy of portion size estimations. In some instances the respondent was asked to supply her own utensils for the recall. To obtain the weight/g equivalents of foods, the Bangladesh Food Photo Manual was used, which converts food items of different sizes and composition to gram equivalents. In addition, the respondents were asked to provide information on the following: type of fat/oil used for cooking, preferred cooking method, meal consumption patterns, number of family members use the same kitchen. National Cholesterol Education Program guideline was followed for daily intake of carbohydrate (>55 %), protein (15 %) and fat (>30 %) [19].

Cut-off values for overweight and general obesity for both sexes were BMI 23-24.9 kg/m² and BMI of ≥25 kg/m² respectively; cut-off values for central obesity including waist circumference for male and female were ≥90 and ≥80 cm, WHR for male and female were ≥0.90 and ≥0.80 [14, 15]. Based on the monthly expenditure socio-economic condition was classified as low (<6000 Bangladeshi Taka [BDT, 1 USD = 84 BDT]), medium (6000-11000 BDT) and high (>11000 BDT). Education level graded as illiterate: unable to write and read; having primary and secondary education and college and above. Physical activity was graded on the ordinal scale of 1-3, corresponding to light, moderate and heavy, according to the activity level based on their occupation. For the purpose of data analysis, these results were transformed into a binary variable - inactive (grade1) and active (grade 2 and 3). Smoking habit was classified as either current or non/ex-smoker.

The study was carried out in accordance with the declaration of Helsinki as revised in 2000 and all procedures involving for this study were approved by Regional Ethics Committee (REK) of Norway and the Ethics Review Committee (ERC) of Diabetic Association of Bangladesh for Medical Research. Since approximately 52.1 % of the adult population is illiterate [13], an informed verbal consent was received from all subjects or guardians of the subjects prior to inclusion in the study to avoid selection bias. The participants were also verbally informed of their right to withdraw from the study at any stage, or to restrict their data from the analysis. After the verbal information, a printed copy of their rights was given. Both the Ethics Review Committees were satisfied with the voluntary participation, maintenance of the rights of the participating subjects and confidential handling of personal information by the study team and has approved the consent procedure.

The present analysis is based on 2,293 participants (842 male and 1,451 female) for whom all the variables were available. Both STATA 11 for Windows (STATA Co., College Station, TX, USA) and PASW statistics version 18 for Windows (SPSS Inc., Chicago, IL, USA) were used as needed. Means and Percentages with 95 % confidence intervals adjusted for age were given for normally distributed continuous variables and categorical variables as needed. Age specific and age standardized prevalence by direct standardization method were estimated on the basis of 2001 census data before performing statistical tests [20]. Differences between the groups of means and proportions adjusted for age were tested by analysis of covariance (ANCOVA) and logistic regression. Chi-squared test was used to test the trend. In the analysis, general (defined by BMI) and central obesity (defined by WC) were considered as the dependent variable. Socio- demographic and dietary factors were included in the regression analysis to identify variables independently associated with overweight and obesity. In the analysis, only those variables which were identified to be independently associated in univariate analyses were included in multinomial models. Statistical inference is based on 95 % confidence intervals (CIs) and the significance level was set at 0.05.