**Background:** Childhood obesity is linked to increased risk of type-2 diabetes and cardiovascular disease in adulthood. India faces a rising burden of non-communicable diseases, partly linked to childhood obesity. While BMI is commonly used to screen for obesity, Indians and South Asians typically have higher abdominal adiposity and lower muscle mass, potentially reducing BMI's discriminatory power. Waist circumference (WC) and waist-to-height ratio (WHtR) are simpler measures of central adiposity. WHtR with a cutoff of 0.5 has been suggested as a robust marker independent of age and sex. This study aimed to examine the accuracy of BMI, WC, and WHtR in predicting hypertension in Indian school children and adolescents from low-middle socioeconomic backgrounds.
**Methods:** School children aged 9–20 years were enrolled from 14 districts across India as part of the SOLID (Study Of Lung function and Its Development) cohort. Study sites were selected to represent geographical, cultural, and ethnic diversity. Children attended residential schools (Jawahar Navodaya Vidyalaya) and mostly belonged to lower-middle socioeconomic strata. Cross-sectional screenings of 2,609 participants (1,210 girls, 1,399 boys) were performed from 2017–2019. Height, weight, and waist circumference were measured using standardized equipment. BMI was calculated as weight (kg) per height (m²). Z-scores for BMI-for-age (zbfa) and height-for-age (zhfa) were calculated using WHO growth charts. Age- and sex-specific normograms for BMI and WC were generated using the Box-Cox-Cole-Green (BCCG) method. Blood pressure was measured using a digital device (Omron HEM 7120) after 5 minutes of rest. Hypertension was classified using National Heart, Lung, and Blood Institute cutoffs. Receiver operator characteristic (ROC) analysis was used to calculate AUCs for each anthropometric index, with optimal cutoffs determined by the closest.topleft method. The Delong test compared predictive ability among indices.
**Key Results:** Among 2,603 participants with blood pressure data, overall prevalence of hypertension (stage 1 and 2) was 10.6% and prehypertension was 16.8%. Stage 2 hypertension was infrequent (~1% in both sexes). Prehypertension was twice as prevalent in boys (21.2%) compared to girls (11.67%) (p<0.001). Stage 1 hypertension was similar between sexes (girls 10.68%, boys 8.6%, p=0.44). Girls had significantly higher odds of central obesity (WHtR > 0.5) (OR=3.1, CI=2.44–3.99) and stunting (OR=1.5, CI=1.21–1.88) compared to boys. Central obesity prevalence was 20.07% in girls vs. 7.45% in boys (p<0.001). For predicting hypertension in girls, AUCs were: zbfa 0.55 (95% CI 0.50–0.60), zwfa 0.64 (95% CI 0.59–0.68), and WHtR 0.66 (95% CI 0.62–0.71). WHtR was significantly better than WC (p=0.02). In boys, AUCs for hypertension were: zbfa 0.67 (95% CI 0.62–0.72), zwfa 0.61 (95% CI 0.56–0.66), and WHtR 0.62 (95% CI 0.57–0.67). BMI was significantly better than WC (p=0.005) and WHtR (p=0.02) in boys. All three indices performed poorly for prehypertension prediction (AUCs 0.54–0.59). Optimal cutoffs for hypertension in girls were: zbfa 0.12 (sensitivity 51.43%, specificity 57.38%), zwfa 0.29 (sensitivity 56.43%, specificity 63.93%), WHtR 0.46 (sensitivity 67.14%, specificity 59.23%). In boys: zbfa 0.29 (sensitivity 63.08%, specificity 66.21%), zwfa 0.41 (sensitivity 50.77%, specificity 69.87%), WHtR 0.44 (sensitivity 60.77%, specificity 55.75%).
**Clinical Implications:** BMI alone is an inadequate screening tool for cardiovascular risk in populations with high central obesity, particularly Indian girls. WHtR, which does not require age- or sex-specific norms and uses a simple cutoff, may be a more practical and equitable screening tool. The study suggests combining BMI with central adiposity markers (WHtR or WC) could improve hypertension risk stratification. The unexpectedly high prevalence of elevated blood pressure (approximately one-third of boys and one-fourth of girls affected when combining prehypertension and hypertension) in a relatively lean, physically active cohort from low-middle socioeconomic backgrounds warrants further investigation. However, these were single blood pressure measurements and not clinical diagnoses. The study's strengths include its pan-India diversity and standardized methodology, though results may not be generalizable to affluent urban Indian populations.