School of Public Health

Archana Shrestha

Diet, Obesity and Diabetes in Suburban Nepal – A Community-Based Study


Obesity and type-2 diabetes are among the top five risk factors for cardiovascular (CVD) deaths in the world, and their prevalence is rising in Nepal. These diseases result from the interaction of genetic, environmental and nutritional factors. Reducing obesity and diabetes through nutritional modification holds great promise as a strategy to prevent CVD deaths.


This dissertation aimed to assess relative validity and reproducibility of a food frequency questionnaire (FFQ); use the FFQ data to derive major dietary patterns; and to investigate the relationship of the derived dietary patterns with overweight, obesity, and type-2 diabetes in a suburban community of Nepal.


This cross-sectional study was conducted among 1,073 community-based adult participants (18 years of age or older) of the Dhulikhel Heart Study (DHS) in Dhulikhel, central Nepal. We recruited a subset of the DHS (n = 121); administered a 115-item FFQ twice over a three months interval; and collected six 24-hour dietary recalls (HDR). We used Spearman correlation coefficients between weekly servings of 22 food groups estimated by the two FFQs to test reproducibility, and weekly servings of the food groups estimated by the FFQs was compared to an average of six 24 HDR to test the relative validity. Dietary patterns were derived using principal component analysis on the FFQ data from 1,073 DHS participants. International body mass index (BMI) cut-offs were used to define overweight as BMI of 25 kg/m2 or higher, and obesity as BMI of 30kg/m2 or higher. Type 2 diabetes was defined as glycated hemoglobin (HbA1c) of 6.5% or higher. The blood sample was available for 479 participants. We utilized generalized estimating equation (GEE) with multivariate logistic regression (household as cluster) to examine the association of the derived dietary patterns with overweight, obesity, and diabetes adjusting for demographic (age, sex, ethnicity, religion, marital status, income and education), and CVD risk factors (alcohol consumption, smoking, physical activity, and systolic blood pressure).


All crude Spearman coefficients for validity were greater than 0.3 except for dairy products, chips and fries, sausages/burgers, and pizzas/pastas. The de-attenuated Spearman correlation ranged from 0.19 (sausage/burgers) to 0.72 (sweet foods). The reproducibility Spearman correlations for all food groups were greater than 0.5, with the exception of oil (0.41). Four dietary patterns were derived: mixed, fast food, refined grain-meat-alcohol, and solid fats-diary. The refined grain-meat-alcohol pattern was positively associated with overweight (OR 1.19, 95% CI: 1.03 – 1.39; p-value=0.02) after adjusting for demographic and CVD risk factors. The association between fast food and overweight/obesity was significantly modified by age (p=0.01). In adults 40 years of age or older, the fast food pattern was positively associated with obesity after controlling for demographic and traditional risk factors (OR 1.69, 95% CI: 1.19 - 2.39; p-value=0.003). None of the food patterns were significantly associated with type-2 diabetes.


The FFQ that was designed for the DHS was found to be reliable and valid for assessing the intake of food groups in Nepal. Our results suggest that refined grains-meat-alcohol intake is associated with higher prevalence of overweight. Fast food intake is associated with higher prevalence of obesity in older adults (40 years or above). The findings were inconclusive for diabetes. This study adds to the existing literature by identifying dietary pattern and their relationships with overweight and obesity in suburban Nepal, an understudied population in a low-resource setting.