Ferritin-derived composite indices are associated with MAFLD and support interpretable risk prediction: evidence from NHANES 2017–2023 and NHANES III
Clinical and Experimental Hepatology · 2026
DOI: 10.5114/ceh.2026.166064 ↗Kurz gefasst
Fachartikel in der Zeitschrift Clinical and Experimental Hepatology, 2026. Verfasst von Lu Wang und 5 weiteren Autor:innen.
Abstract (Originalsprache)
Ferritin-derived composite indices are associated with MAFLD and support interpretable risk prediction: evidence from NHANES 2017–2023 and NHANES III Lu Wang 1 , Weigen Wu 2 , Yuqing Wang 3 , Yuanzi Huangsha 3 , Chunxian Zeng 4 , Ruizhi Wang 3 Department of Laboratory Medicine, Xinyu Iron and Steel Group Co. Ltd., Central Hospital, China Center of Hepato-Pancreato-Biliary Surgery, Department of Pancreato-Biliary Surgery, The First Affiliated Hospital of Sun Yat-sen University, China Department of Laboratory Medicine, The First Affiliated Hospital of Sun Yat-sen University, China Shenzhen Hospital, Southern Medical University (Shenzhen School of Clinical Medicine, Southern Medical University), China Clin Exp HEPATOL 2026; 12, 4: DOI: https://doi.org/10.5114/ceh.2026.166064 Online publish date: 2026/11/19 View full text Get citation Get citation ENW EndNote BIB JabRef, Mendeley RIS Papers, Reference Manager, RefWorks, Zotero AMA Wang L, Wu W, Wang Y, Huangsha Y, Zeng C, Wang R. Ferritin-derived composite indices are associated with MAFLD and support interpretable risk prediction: evidence from NHANES 2017–2023 and NHANES III. Clinical and Experimental Hepatology - Manuscripts Accepted. 2026. doi:10.5114/ceh.2026.166064. APA Wang, L., Wu, W., Wang, Y., Huangsha, Y., Zeng, C., & Wang, R. (2026). Ferritin-derived composite indices are associated with MAFLD and support interpretable risk prediction: evidence from NHANES 2017–2023 and NHANES III. Clinical and Experimental Hepatology - Manuscripts Accepted. https://doi.org/10.5114/ceh.2026.166064 Chicago Wang, Lu, Weigen Wu, Yuqing Wang, Yuanzi Huangsha, Chunxian Zeng, and Ruizhi Wang. 2026. "Ferritin-derived composite indices are associated with MAFLD and support interpretable risk prediction: evidence from NHANES 2017–2023 and NHANES III". Clinical and Experimental Hepatology - Manuscripts Accepted. doi:10.5114/ceh.2026.166064. Harvard Wang, L., Wu, W., Wang, Y., Huangsha, Y., Zeng, C., and Wang, R. (2026). Ferritin-derived composite indices are associated with MAFLD and support interpretable risk prediction: evidence from NHANES 2017–2023 and NHANES III. Clinical and Experimental Hepatology - Manuscripts Accepted. https://doi.org/10.5114/ceh.2026.166064 MLA Wang, Lu et al. "Ferritin-derived composite indices are associated with MAFLD and support interpretable risk prediction: evidence from NHANES 2017–2023 and NHANES III." Clinical and Experimental Hepatology - Manuscripts Accepted, 2026. doi:10.5114/ceh.2026.166064. Vancouver Wang L, Wu W, Wang Y, Huangsha Y, Zeng C, Wang R. Ferritin-derived composite indices are associated with MAFLD and support interpretable risk prediction: evidence from NHANES 2017–2023 and NHANES III. Clinical and Experimental Hepatology - Manuscripts Accepted. 2026. doi:10.5114/ceh.2026.166064. Aim of the study This study investigated associations between four ferritin-derived composite indices – ferritin-to-albumin-to-lymphocyte ratio (FALR), ferritin-to-albumin ratio (FAR), ferritin-to-lymphocyte ratio (FLR), and ferritin-to-hemoglobin ratio (FHR) – and metabolic dysfunction-associated fatty liver disease (MAFLD), characterized their dose–response patterns, and developed an externally validated, interpretable machine-learning model for risk stratification.Material and methods We performed a secondary analysis of de-identified public-use data from NHANES 2017–2023 as the development cohort (n = 8,444, 3,512 cases) and NHANES III as the external validation set (n = 12,918, 2,608 cases). We calculated four indices: FALR, FAR, FLR, and FHR. Weighted logistic regression, restricted cubic splines, and threshold models were used to test the associations. We built a gradient boosting machine (GBM) model with 12 predictors selected by the intersection of six feature-selection algorithms and validated it externally. SHAP and LIME provided interpretability.
Results
In fully adjusted models, FAR and FHR were most strongly associated with MAFLD (highest vs. lowest quartile [Q4 vs. Q1]: odds ratio [OR] = 1.69, 95% CI: 1.34-2.14 and OR = 1.65, 95% CI: 1.28-2.13, respectively; both p for trend < 0.001). Applying the same cutpoints to NHANES III gave ORs of 2.92 (2.35-3.62) for FAR and 3.00 (2.37-3.82) for FHR. All four indices showed nonlinear threshold effects, with inflection points at 60.42 (FAR) and 15.10 (FHR). The GBM achieved an internal AUC of 0.731 (0.711-0.750) and an external AUC of 0.699 (0.688-0.710), with accuracy 0.649, sensitivity 0.624, and specificity 0.655.
Conclusions
FAR and FHR show consistent, nonlinear associations with MAFLD across two large U.S. cohorts. The externally validated GBM provides moderate discrimination and a web-based tool that may support exploratory risk identification; prospective studies are required to confirm its clinical utility and temporal validity. Keywords MAFLD, ferritin-derived composite indices, dose-response relationship, machine learning, NHANES
Zitieren
APA
Wang, L., Wu, W., Wang, Y., Huangsha, Y., Zeng, C., & Wang, R. (2026). Ferritin-derived composite indices are associated with MAFLD and support interpretable risk prediction: evidence from NHANES 2017–2023 and NHANES III. Clinical and Experimental Hepatology. https://doi.org/10.5114/ceh.2026.166064
Vancouver
Wang L, Wu W, Wang Y, Huangsha Y, Zeng C, Wang R. Ferritin-derived composite indices are associated with MAFLD and support interpretable risk prediction: evidence from NHANES 2017–2023 and NHANES III. Clinical and Experimental Hepatology. 2026. doi:10.5114/ceh.2026.166064
BibTeX
@article{Wang2026,
author = {Wang, Lu and Wu, Weigen and Wang, Yuqing and Huangsha, Yuanzi and Zeng, Chunxian and Wang, Ruizhi},
title = {Ferritin-derived composite indices are associated with MAFLD and support interpretable risk prediction: evidence from NHANES 2017–2023 and NHANES III},
journal = {Clinical and Experimental Hepatology},
year = {2026},
doi = {10.5114/ceh.2026.166064}
}
EUVIMED