Adiposity measures offer only modest cardiovascular prediction gains

Waist-to-hip ratio was positively associated with cardiovascular disease, coronary heart disease and stroke, while alternative indices produced the highest observed incremental classification performance for selected outcomes. Overall, gains over body mass index in discrimination were small, and no index was uniformly superior across cardiovascular disease, coronary heart disease, stroke or mortality. The complete-case analysis followed participants for a median of 15.6 years.

During that period, 27,947 incident cardiovascular disease events, defined as coronary heart disease or stroke, and 24,464 deaths from any cause occurred. The researchers compared simple, formula-derived and biomarker-based adiposity indices using multivariable Cox models, restricted cubic splines and time-varying analyses. They examined incident cardiovascular disease, its components and both all-cause and broadly defined cardiovascular mortality.

To assess discrimination, the team used five-fold cross-validated integrated time-dependent areas under the curve over one to 17 years, evaluating each alternative index in place of body mass index within a common framework of conventional risk factors. For each standard-deviation increase in waist-to-hip ratio, the adjusted hazard ratio was 1.11 for cardiovascular disease, 1.12 for coronary heart disease and 1.07 for stroke. The corresponding 95% confidence intervals were 1.09–1.13, 1.10–1.14 and 1.03–1.10.

These estimates describe statistical associations and do not show that changing an adiposity measure would prevent cardiovascular events. The patterns also differed among outcomes, and several associations between adiposity indices and cardiovascular disease or coronary heart disease attenuated during follow-up. The results therefore support neither a single universally best measure nor a uniform association across all outcomes and time periods.

When the researchers tested whether alternative indices improved discrimination, the highest observed gains over body mass index ranged from 0.0001 to 0.0025 across outcomes. Incremental prediction was assessed at 10 years with truncated Harrell’s C-index, integrated discrimination improvement and continuous net reclassification improvement after adding a non-body-mass-index measure to a Framingham risk-factor benchmark that already included body mass index and conventional cardiovascular risk factors. For cardiovascular disease, weight-adjusted waist index yielded a continuous net reclassification improvement of 0.077, with a 95% confidence interval of 0.063–0.093.

That result was driven mainly by the component concerning people who did not experience an event, rather than by a large improvement in distinguishing events from non-events. The corresponding C-index change was 0.00076 and the integrated discrimination improvement was 0.000455, indicating only modest additional discriminatory information in this analysis. Outcome-specific results also differed.

Weight-adjusted waist index had the highest observed total net reclassification improvement for cardiovascular disease and coronary heart disease, waist-to-hip ratio had the highest observed value for broadly defined cardiovascular mortality, and A Body Shape Index had the highest observed value for all-cause mortality. In addition, the analysis used a complete-case sample from UK Biobank, and the reported predictive performance was assessed within the cohort rather than externally validated. Calibration and clinical utility were not established.

The authors concluded that selected non-body-mass-index indices may provide modest, outcome-specific incremental information, but external validation and further assessment of calibration and clinical utility are needed before the findings can support clinical recommendations.

Subject of Research: Comparison of 12 adiposity indices for predicting cardiovascular disease and mortality in a UK Biobank cohort

Article Title: Comparative prognostic value of simple, formula-derived, and biomarker-based adiposity indices for cardiovascular disease and mortality: a UK Biobank cohort study

Article References: Liu, J., Jiang, G., Qu, H., Hao, W., Liu, Y., & Han, D. (2026). Comparative prognostic value of simple, formula-derived, and biomarker-based adiposity indices for cardiovascular disease and mortality: a UK Biobank cohort study. BMC Public Health. https://doi.org/10.1186/s12889-026-29871-z

Image Credits: AI Generated

DOI: 10.1186/s12889-026-29871-z

Keywords: UK Biobank, adiposity, body mass index, waist-to-hip ratio, cardiovascular disease, coronary heart disease, stroke, mortality, risk prediction, cohort study, Comparative, prognostic