**Background:** Gestational diabetes mellitus (GDM) affects approximately 20% of pregnancies in India and is associated with adverse short- and long-term outcomes for mothers and children, including increased risk of type 2 diabetes and cardiovascular disease. Diet before and during pregnancy is a modifiable risk factor, but evidence from low- and middle-income countries, particularly India, is sparse. The BANGLES study aimed to identify peri-conceptional diet patterns in South Indian women and examine their association with GDM risk.
**Methods:** BANGLES was a prospective observational study conducted in two hospitals in Bangalore, India, between June 2016 and October 2017. A total of 785 women with singleton pregnancies, <17 weeks' gestation, no prior diabetes, and intending to stay in Bangalore until ≥28 weeks were recruited. Peri-conceptional diet was assessed at recruitment (5–16 weeks) using a validated 224-item food frequency questionnaire (FFQ) administered by trained nutritionists. Women recalled their diet during the month before they knew they were pregnant. Principal component analysis (PCA) of 68 food groups (derived from a combination of a priori and a posteriori methods) identified three diet patterns. GDM was assessed at 24–28 weeks using a 75-g oral glucose tolerance test (OGTT) with WHO 2013 criteria (fasting ≥5.1 mmol/L, 1-h ≥10.0 mmol/L, or 2-h ≥8.5 mmol/L). A subset of women (n=35) had a 50-g glucose challenge test per obstetrician instruction. Multivariate logistic regression examined associations between diet pattern z-scores and GDM, adjusting for a priori confounders (age, gravidity, PCOS, hypothyroidism, family history of T2DM, urban/rural residence, family type, physical activity, SLI score, education, supplement use) with and without BMI.
**Key Results:** GDM prevalence was 21.9% (157/714). Three diet patterns were identified: (1) High-diversity, urban (HDU) pattern — characterized by diverse home-cooked and processed foods, whole grains (chapati), fruits, vegetables, dairy, nuts, seeds, egg, poultry, meat, fast food, and sweets; associated with older, more affluent, better-educated, urban women. (2) Rice-fried snacks-chicken-sweets (RFCS) pattern — low diet diversity with white rice, chicken, fried snacks, sweets, and banana; associated with younger, less-educated, lower-income, rural, joint-family women. (3) Healthy traditional vegetarian (HTV) pattern — home-cooked vegetarian foods including millets (ragi ball), fermented dairy, tender coconut water, pulses, greens, and nuts; associated with less-educated, more affluent, rural, joint-family women. The HDU pattern was associated with significantly lower GDM risk after adjusting for confounders (aOR 0.80 per SD, 95% CI 0.64–0.99, P=0.04). This association weakened after further adjustment for BMI (aOR 0.84, 95% CI 0.67–1.05, P=0.12), suggesting partial mediation by BMI. Neither the RFCS nor HTV pattern was significantly associated with GDM risk. Higher BMI was the strongest predictor of GDM (P<0.001). Higher HDU scores were associated with lower BMI (β: −0.42, 95% CI −0.806 to −0.038, P=0.03). The first three PCA components explained 18.6% of variance.
**Clinical Implications:** The HDU pattern's association with lower GDM risk supports global recommendations to increase diet diversity for optimal health. The finding that BMI was the strongest risk factor and potentially mediated the diet–GDM association underscores the importance of achieving a healthy pre-pregnancy BMI. The presence of both healthy and unhealthy foods within the HDU and RFCS patterns suggests low awareness about healthy vs. unhealthy foods among Indian women, highlighting the need for public education. Higher SES was associated with greater diet diversity, indicating that national policies should aim to make micronutrient-rich foods (wholegrains, fruits, vegetables, dairy, nuts, poultry) more affordable. The authors note that a single 'universal healthy diet pattern' may not be relevant across India's diverse food cultures, and future pooled data studies are needed to design culturally appropriate interventions.