Belly Fat Meets Blood Sugar: New Index Sharpens Long-Term Hypertension Prediction

Hypertension rarely announces itself. It builds silently over years, driven by a tangle of metabolic disturbances that standard checkup numbers only partially reveal. Now, a large community-based study from the NAGALA cohort in Japan suggests that a new generation of composite markers—ones that fuse waist-based measures of abdominal fat with blood glucose and lipid readings—may identify people at risk of developing high blood pressure years before their readings climb. The research, published in BMC Endocrine Disorders, offers one of the most rigorous head-to-head comparisons to date of four competing glycolipid indices, and its findings could reshape how clinicians think about early cardiovascular risk screening.

The study, led by Hui Wu and colleagues at the First Affiliated Hospital of Kunming Medical University, analyzed 15,422 adults drawn from the NAGALA cohort, a longitudinal community database established by Prof. Takuro Okamura at Murakami Memorial Hospital and recruited between 1994 and 2016. Participants were followed for an average of 6.05 years, with some tracked for as long as 13 years. During that window, 780 individuals developed new-onset hypertension. The researchers used this rich dataset to ask a deceptively simple question: which of four composite indices best predicts who will develop high blood pressure, and does adding a measure of central obesity to the classic formulas improve their performance?

To understand the study’s logic, it helps to unpack the indices themselves. The triglyceride-glucose index, known as TyG, is a well-established mathematical combination of fasting triglyceride and glucose levels that serves as a proxy for insulin resistance—the metabolic state in which cells respond poorly to insulin, setting the stage for vascular damage. The monocyte-to-HDL cholesterol ratio combined with TyG, abbreviated MCMI, adds an inflammatory dimension. The lipid accumulation product-based WyG index and the cardiometabolic index, or CMI, each blend lipid and glycemic signals in different proportions. What all four traditionally lack is any direct accounting of abdominal adiposity, the visceral fat that wraps around internal organs and is one of the strongest known bridges between metabolic dysfunction and vascular disease.

The researchers’ central innovation was the construction of a novel waist-adiposity index, or WAI, designed to capture two independent dimensions of central obesity: the absolute waist-to-height ratio and the relative measure of waist circumference adjusted for body mass index. By integrating WAI into three of the four indices—creating TyG-WAI, MCMI-WAI, and WyG-WAI—while keeping CMI as a comparator without the waist component, the team could test directly whether the added adiposity signal earns its place. All indices were z-score standardized to ensure fair comparison under identical statistical adjustment.

The results were nuanced. Three of the four indices—CMI, MCMI-WAI, and TyG-WAI—remained independently associated with incident hypertension in fully adjusted Cox proportional hazards models, each reaching statistical significance. WyG-WAI, by contrast, lost its association once the full set of confounders was accounted for. TyG-WAI emerged as the most consistent performer, showing stable long-term discriminatory performance across the follow-up period and across demographic subgroups. CMI told a more sex-specific story, predicting hypertension only in men. The pattern suggests that different composite markers may capture different facets of cardiometabolic risk, and that no single formula is universally superior.

Perhaps the most technically impressive portion of the study is its exhaustive internal validation of the WAI itself, conducted across eight distinct domains. The index demonstrated known-groups validity, distinguishing participants with metabolic syndrome and those who were overweight from healthier counterparts with p-values below 0.001 in both cases. Convergent validity was confirmed through correlations with waist circumference (r = 0.600) and body mass index (r = 0.357), while discriminant validity was established by showing that the index correlated more strongly with waist-to-height ratio (r = 0.789) than with BMI. In time-dependent receiver operating characteristic analysis, the index achieved an area under the curve of 0.733 at 13 years, indicating reasonably strong long-term discrimination.

The validation also probed whether adding WAI genuinely improves prediction beyond existing models. A likelihood-ratio test confirmed a statistically significant incremental value (p = 0.008), though the authors were candid that the absolute change in the concordance index was negligible (ΔC-index = −0.00021). This honesty matters: a statistically significant but clinically tiny improvement means the index is best viewed as a candidate risk-stratification marker rather than a transformative predictor. Restricted cubic spline analysis revealed a linear dose-response relationship between the index and hypertension risk, with no evidence of nonlinearity (P for nonlinear = 0.280), and sensitivity analyses across four exclusion criteria plus measurement-error propagation—showing a coefficient of variation of just 2.26% when waist circumference was perturbed by one centimeter—confirmed the findings were robust.

One of the most intriguing findings concerns nonlinear risk patterns and subgroup effects. MCMI-WAI exhibited a distinct non-monotonic dose-response relationship with hypertension, meaning the risk curve was not a simple straight line, and showed exploratory evidence of stronger predictive performance in lean men. Separately, the WAI itself showed exploratory associations with incident hypertension in women (hazard ratio 1.216, p = 0.008) and in lean adults (hazard ratio 1.186, p = 0.004). These subgroups are precisely the populations that conventional risk tools often miss, since lean individuals are frequently assumed to be metabolically healthy. The authors are careful to frame these observations as hypothesis-generating, noting that whether the lean-male interaction is genuine requires confirmation through a combined subgroup analysis in independent cohorts.

Translating these findings into clinical practice, the researchers propose a potential two-stage risk-stratification framework: TyG-WAI for general population screening, given its stable long-term performance, with MCMI-WAI reserved for lean subgroups where its nonlinear pattern may add value. Crucially, all four indices rely on routine clinical and laboratory data—fasting glucose, triglycerides, cholesterol fractions, waist circumference, height, and weight—meaning no specialized testing would be required if the framework is validated. The authors emphasize that calibration studies and decision-curve analysis in independent cohorts are prerequisites before any clinical implementation, a caution that reflects the study’s secondary-analysis design using de-identified public cohort data.

The broader significance of the work lies in its methodological rigor and its reframing of how metabolic risk should be measured. Rather than treating glucose dysregulation, lipid abnormalities, and abdominal obesity as separate concerns, the composite indices quantify their joint burden in a single number. The study’s transparent reporting—including the acknowledgment of a negligible absolute C-index change and the explicit call for external validation—models the kind of scientific restraint that viral health claims often lack. For now, the message for readers is measured but meaningful: the combination of belly fat and blood chemistry carries predictive information about future blood pressure that neither alone provides, and researchers are steadily building the tools to read that signal. Whether these indices will earn a place in routine screening depends on the validation studies now needed, but the NAGALA analysis has laid a careful foundation for that next step.

Subject of Research: Comparison of waist-adiposity–glycolipid composite indices for predicting new-onset hypertension in a community cohort

Article Title: Head‑to‑head comparison of four waist-adiposity–glycolipid indices for long‑term prediction of new‑onset hypertension: the NAGALA cohort study

Article References: Head‑to‑head comparison of four waist-adiposity–glycolipid indices for long‑term prediction of new‑onset hypertension: the NAGALA cohort study. (n.d.). https://doi.org/10.1186/s12902-026-02642-y

Image Credits: AI Generated

DOI: 10.1186/s12902-026-02642-y

Keywords: hypertension, glycolipid indices, waist-adiposity index, insulin resistance, NAGALA cohort, metabolic syndrome, central obesity, TyG index, CMI, risk prediction, cardiometabolic risk, cohort study