- Higher NMRI was associated with lower odds of prevalent MASLD (adjusted OR per unit 0.71), with an approximately linear relationship.
- NMRI demonstrated superior discrimination versus individual components and improved model performance (AUC 0.822; NRI 0.759; IDI 0.164).
- TyG and WHtR contributed most to NMRI; adiposity-predominant and unfavourable profiles had markedly higher MASLD odds (adjusted ORs 6.62, 7.85).
Front Endocrinol (Lausanne). 2026 Sep 30;17:1952423. doi: 10.3389/fendo.2026.1952423. eCollection 2026.
ABSTRACT
BACKGROUND: Metabolic dysfunction-associated steatotic liver disease (MASLD) is closely associated with metabolic, anthropometric, and nutritional characteristics. The Nutritional Metabolic Risk Index (NMRI) integrates these dimensions, but its association with MASLD remains unclear. We evaluated the association of the NMRI with prevalent MASLD and its discriminative performance in a sub-plateau population.
METHODS: This cross-sectional study included 1,551 adults undergoing health examinations at approximately 1,900 m. K-means clustering was used to identify metabolic-nutritional profiles. Associations of NMRI with prevalent MASLD were assessed using multivariable logistic regression and restricted cubic splines. Discriminative and incremental performance were evaluated using receiver operating characteristic curves, net reclassification improvement (NRI), integrated discrimination improvement (IDI), and decision curve analysis, with bootstrap internal validation. Weighted quantile sum (WQS) regression assessed the relative statistical contributions of NMRI components.
RESULTS: Three metabolic-nutritional profiles were identified. Compared with the favorable metabolic profile, the adiposity-predominant and unfavorable metabolic-nutritional profiles had higher odds of MASLD (adjusted ORs, 6.62 and 7.85, respectively; both P<0.001). NMRI was inversely associated with prevalent MASLD (adjusted OR per unit increase, 0.71; 95% CI, 0.68-0.74), with an approximately linear association. NMRI showed higher discrimination than its individual components. Adding NMRI to the basic model improved discrimination (AUC, 0.822), reclassification (NRI, 0.759; IDI, 0.164), and net benefit across most examined threshold probabilities. Bootstrap validation yielded an optimism-corrected AUC of 0.815. WQS analysis assigned larger weights to TyG and WHtR than to GNRI. Associations remained consistent across subgroup and sensitivity analyses.
CONCLUSION: Higher NMRI was consistently associated with lower odds of prevalent MASLD and showed favorable within-sample discrimination in this sub-plateau population. These findings support further evaluation of NMRI as a parsimonious composite indicator integrating nutritional, metabolic, and anthropometric characteristics associated with MASLD.
PMID:42876641 | PMC:PMC13661332 | DOI:10.3389/fendo.2026.1952423
Share Evidence Blueprint
Save to Google Notes
Search Google Scholar
Save as PDF
⭐ My Revision List
