The study, published in Frontiers in Public Health as a cross-sectional analysis, started from a database of 214,582 examination records and applied strict quality-control rules, sequentially excluding records with missing key measurements, missing medical history, participants younger than 18, physiologically implausible outliers, and duplicates. The final cohort had a mean age of 48.15 years, was 58.4 percent male, and showed a hypertension prevalence of 20.4 percent, defined according to the Chinese Guidelines for the Management of Hypertension (2024 Revision) as blood pressure at or above 140/90 mmHg, a self-reported diagnosis, or current antihypertensive medication use. Blood pressure was measured with calibrated electronic devices after a five-minute rest, with two readings averaged and a third taken when the first two differed by more than 5 mmHg.

To make the five indices directly comparable, the researchers converted each to a Z-score and used multivariable logistic and linear regression, adjusting for sex, age, ethnicity, smoking, family history of hypertension, diabetes, coronary heart disease, and serum uric acid. The results were striking. Each one-standard-deviation increase in WC or BMI was associated with a 61 percent higher odds of hypertension (OR 1.61, 95 percent CI 1.58–1.64), WHtR with 60 percent (OR 1.60, 1.57–1.63), and BRI with 56 percent (OR 1.56, 1.54–1.59). ABSI, the mathematically most elaborate index, managed only a 9 percent increase (OR 1.09, 1.07–1.11). In linear terms, a standard-deviation rise in WHtR pushed systolic blood pressure up by 3.95 mmHg and diastolic by 2.28 mmHg, while ABSI moved systolic pressure by just 0.61 mmHg.

Discriminatory power told a similar story. Using the area under the receiver operating characteristic curve (AUC), WHtR and BRI tied for the top spot at an identical 0.676, followed by WC at 0.653, BMI at 0.627, and ABSI at the bottom with 0.612. The DeLong test confirmed that WHtR significantly outperformed WC, BMI, and ABSI (all P < 0.001), while its difference from BRI was exactly zero (P = 1.000). The authors are refreshingly candid about why: BRI is essentially a monotonic transformation of the waist-to-height construct, a function of WC divided by π times height, making the two indices mathematical near-equivalents rather than independent measures. Their identical performance reflects the same underlying central obesity signal, and the choice between them in practice comes down to computational simplicity (WHtR) or theoretical appeal (BRI).

The genuinely novel twist emerged when the team stratified the analysis by age. In young adults aged 18 to 35, discriminatory ability was strongest across the board, and raw waist circumference beat everything, reaching an AUC of 0.763, with WHtR and BRI close behind at 0.754. The odds ratios per standard deviation were also far larger in this group—2.19 for WC, 2.31 for WHtR, and 2.23 for BRI—compared with 1.73, 1.77, and 1.73 in the middle-aged group and 1.29, 1.31, and 1.29 in adults aged 60 and above. Formal interaction tests confirmed the pattern was statistically robust: all five indices showed significant negative interaction with age (all P < 0.001), with WHtR’s interaction coefficient at −0.0147 and BMI’s at −0.0075, demonstrating a broad age-attenuation effect.

The mechanistic explanation the authors offer is technically elegant. In young adults, height varies relatively little between individuals while waist circumference varies widely, driven by lifestyle and dietary habits; normalizing WC by height, as WHtR and BRI do, may therefore introduce noise rather than signal, slightly reducing discriminative ability. With advancing age, blood pressure becomes increasingly determined by multifactorial forces—vascular stiffening, comorbidities, renal changes—which dilute the standalone discriminatory value of any adiposity measure. The team also notes that the attenuation in older adults could partly reflect survivorship bias, since obese individuals with hypertension face higher mortality and are systematically less likely to survive to attend check-ups after 60, which would artificially weaken the observed relationship in that selected, disproportionately healthy sample.

ABSI provides the study’s cautionary tale about mathematical sophistication. Originally designed to characterize body shape more finely by adjusting waist circumference for height and BMI, the index consistently underperformed across every age stratum and every outcome. In the fully adjusted models, ABSI was even weakly inversely associated with diastolic blood pressure in women (β = −0.05, P = 0.320) and in older adults (β = −0.44). The finding aligns with prior meta-analyses and population studies in the United States and China, reinforcing the authors’ conclusion that greater mathematical complexity does not necessarily translate into superior clinical identification capability—a lesson that resonates well beyond anthropometry.

Stratified analyses by sex and ethnicity showed the core pattern was stable. Associations were slightly stronger in men than in women, and all five indices performed similarly in Han and minority ethnic groups. Quartile-based trend analyses revealed consistent positive dose-response relationships across progressively adjusted models, and a sensitivity analysis excluding the 1,463 participants on antihypertensive medication left the ranking unchanged, with WHtR and BRI still showing the highest AUC values. The authors also acknowledge an important limitation: roughly half the original records were excluded for missing data, and the excluded individuals were younger (mean age 35.6 years) and more predominantly male (71.5 percent) than the included cohort, meaning the complete-case analysis may have slightly overestimated true population-level AUCs and association strengths. Data on alcohol, dietary sodium, physical activity, and sleep were unavailable, leaving residual confounding possible.

For clinicians and public health planners, the practical message is concrete. In health check-up and primary care settings, WHtR or BRI can serve as the default preliminary screen across all ages, while raw waist circumference deserves priority in young adults, where its discriminatory power peaks. The indices should be treated as flagging tools, not diagnostic criteria—anyone flagged should undergo confirmatory blood pressure measurement and receive counseling on weight management, sodium reduction, and physical activity. The authors emphasize that these simple tape-measure metrics are not merely passive screening instruments but actionable targets, since interventions targeting adiposity-related risk can translate into measurable blood pressure benefits. As China’s hypertension burden remains formidable despite improving awareness and treatment rates, the idea that the optimal screening tool depends on the patient’s age is a deceptively simple insight that could reshape routine risk identification, though the authors caution that external validation in prospective, diverse cohorts is essential before the ranking can be considered universally applicable.

Subject of Research: Comparison of anthropometric adiposity indices for hypertension risk screening across age groups in Chinese adults

Article Title: A cross-sectional study on anthropometric indices and hypertension in Chinese adults: comparison of discriminatory power and age-stratified analysis

Article References: Lu, S., Chen, N., Li, Z., Zhang, Y., Wang, W., Li, G., Hao, J., Wang, R., Qi, Y., & Qin, Z. (2026). A cross-sectional study on anthropometric indices and hypertension in Chinese adults: comparison of discriminatory power and age-stratified analysis. Frontiers in Public Health, 14, Article 1933929. https://doi.org/10.3389/fpubh.2026.1933929

Image Credits: AI Generated

DOI: 10.3389/fpubh.2026.1933929

Keywords: hypertension, anthropometric indices, waist-to-height ratio, body roundness index, waist circumference, body mass index, ABSI, central obesity, age-stratified analysis, ROC curve, cross-sectional study, Chinese adults