Head-to-head comparison of 35 metabolic adiposity indices for predicting incident diabetes in middle-aged and older adults with clinical obesity: evidence from the CHARLS and ELSA cohorts

Comprehensive head-to-head analyses of metabolic adiposity indices for diabetes mellitus (DM) prediction are lacking for middle-aged and older adults with clinical obesity, while quantitative evidence documenting performance differences between Chinese and British populations remains insufficient. We included participants from the China Health and Retirement Longitudinal Study (CHARLS) and the English Longitudinal Study of Ageing (ELSA). Thirty-five metabolic obesity-related indices were systematically examined, comprising the triglyceride-glucose index (TyG), C-reactive protein-triglycerides-glucose index (CTI), cholesterol-high-density lipoprotein-glucose index (CHG), eight core adiposity indices, and their combinations. For each, we assessed baseline level, cumulative exposure, and longitudinal changes. Cox models and restricted cubic splines assessed associations and non-linearity for incident DM. Predictive performance was compared using C-index, net reclassification improvement, and integrated discrimination improvement. Subgroup and sensitivity analyses (including inverse probability weighting, propensity matching, and Fine‑Gray regression for competing mortality) tested robustness. Mediation analyses explored the mediating roles of estimated glucose disposal rate (eGDR) and glycated hemoglobin (HbA1c). Among the composite metabolic adiposity indices, CHG-waist-to-height ratio (CHG-WhtR) was among the leading performers in CHARLS, whereas CTI-WhtR was among the leading performers in ELSA; however, pairwise differences among the top-ranked indices were not statistically significant. For baseline level, per-SD increases yielded hazard ratios (HRs) of 1.69 (95% confidence interval [CI]: 1.58–1.82) for CHG-WhtR and 1.79 (1.60–2.00) for CTI-WhtR, with corresponding tertile 3 versus 1 h of 3.85 (2.99–4.94) and 5.08 (3.42–7.55). For cumulative exposure, the per-SD HRs for cumCHG-WhtR and cumCTI-WhtR were 1.48 (1.19–1.83) and 1.77 (1.35–2.30), and the tertile 3 versus 1 h were 2.42 and 5.48 (all P < 0.05). Regarding longitudinal changes, compared with the stable low-level group, the sustained high-level groups showed significantly elevated risks, with HRs of 3.20 for CHG-WhtR and 5.83 for CTI-WhtR (all P < 0.05). When added to the basic model, CHG-WhtR improved the C-index from 0.743 to 0.751, with an NRI of 21.68% and an IDI of 0.56%; CTI-WhtR improved the C-index from 0.769 to 0.797, with an NRI of 42.83% and an IDI of 1.78% (all P < 0.05). In contrast, adding cumCHG-WhtR or cumCTI-WhtR to the basic model did not improve the C-index. Composite metabolic adiposity indices predicted incident DM in middle-aged and older adults with clinical obesity, although pairwise differences among top-ranked indices were not statistically significant and cross-cohort differences should be considered hypothesis-generating. Baseline measurements supported risk stratification; however, lack of incremental C-index gain from cumulative exposure metrics in landmark-based analyses does not indicate that repeated measurements lack clinical value.

Authors

Institutions

Publication Details

Journal
BMC Endocrine Disorders
Published
2026-11-19
DOI
https://doi.org/10.1186/s12902-026-02733-w
Primary Topic
Diabetes, Cardiovascular Risks, and Lipoproteins
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Head-to-head comparison of 35 metabolic adiposity indices for predicting incident diabetes in middle-aged and older adults with clinical obesity: evidence from the CHARLS and ELSA cohorts

Qianhua Feng, Jie Li, Qiang She, Jie An et al.
BMC Endocrine Disorders
Diabetes, Cardiovascular Risks, and Lipoproteins
article

Head-to-head comparison of 35 metabolic adiposity indices for predicting incident diabetes in middle-aged and older adults with clinical obesity: evidence from the CHARLS and ELSA cohorts

Qianhua Feng, Jie Li, Qiang She, Jie An, Ya Luo, Min Xu, Meng Yu, Dandan Yang
article en

Abstract

Comprehensive head-to-head analyses of metabolic adiposity indices for diabetes mellitus (DM) prediction are lacking for middle-aged and older adults with clinical obesity, while quantitative evidence documenting performance differences between Chinese and British populations remains insufficient. We included participants from the China Health and Retirement Longitudinal Study (CHARLS) and the English Longitudinal Study of Ageing (ELSA). Thirty-five metabolic obesity-related indices were systematically examined, comprising the triglyceride-glucose index (TyG), C-reactive protein-triglycerides-glucose index (CTI), cholesterol-high-density lipoprotein-glucose index (CHG), eight core adiposity indices, and their combinations. For each, we assessed baseline level, cumulative exposure, and longitudinal changes. Cox models and restricted cubic splines assessed associations and non-linearity for incident DM. Predictive performance was compared using C-index, net reclassification improvement, and integrated discrimination improvement. Subgroup and sensitivity analyses (including inverse probability weighting, propensity matching, and Fine‑Gray regression for competing mortality) tested robustness. Mediation analyses explored the mediating roles of estimated glucose disposal rate (eGDR) and glycated hemoglobin (HbA1c). Among the composite metabolic adiposity indices, CHG-waist-to-height ratio (CHG-WhtR) was among the leading performers in CHARLS, whereas CTI-WhtR was among the leading performers in ELSA; however, pairwise differences among the top-ranked indices were not statistically significant. For baseline level, per-SD increases yielded hazard ratios (HRs) of 1.69 (95% confidence interval [CI]: 1.58–1.82) for CHG-WhtR and 1.79 (1.60–2.00) for CTI-WhtR, with corresponding tertile 3 versus 1 h of 3.85 (2.99–4.94) and 5.08 (3.42–7.55). For cumulative exposure, the per-SD HRs for cumCHG-WhtR and cumCTI-WhtR were 1.48 (1.19–1.83) and 1.77 (1.35–2.30), and the tertile 3 versus 1 h were 2.42 and 5.48 (all P < 0.05). Regarding longitudinal changes, compared with the stable low-level group, the sustained high-level groups showed significantly elevated risks, with HRs of 3.20 for CHG-WhtR and 5.83 for CTI-WhtR (all P < 0.05). When added to the basic model, CHG-WhtR improved the C-index from 0.743 to 0.751, with an NRI of 21.68% and an IDI of 0.56%; CTI-WhtR improved the C-index from 0.769 to 0.797, with an NRI of 42.83% and an IDI of 1.78% (all P < 0.05). In contrast, adding cumCHG-WhtR or cumCTI-WhtR to the basic model did not improve the C-index. Composite metabolic adiposity indices predicted incident DM in middle-aged and older adults with clinical obesity, although pairwise differences among top-ranked indices were not statistically significant and cross-cohort differences should be considered hypothesis-generating. Baseline measurements supported risk stratification; however, lack of incremental C-index gain from cumulative exposure metrics in landmark-based analyses does not indicate that repeated measurements lack clinical value.

BMC Endocrine Disorders
Zunyi Medical University (CN), Second Affiliated Hospital of Chongqing Medical University (CN), First People’s Hospital of Zunyi (CN)
Diabetes, Cardiovascular Risks, and Lipoproteins
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon