Diabetes duration modifies the association between AIP and MAFLD in type 2 diabetes.
- 1Department of Endocrinology, Affiliated Dongyang Hospital of Wenzhou Medical University, Dongyang, China.
- 2Vascular Signalling, Molecular Cardiology, Department of Cardiology I-Coronary and Peripheral Vascular Disease, Heart Failure, University Hospital Münster, Münster, Germany.
Abstract
Background
Metabolic dysfunction-associated fatty liver disease (MAFLD) is highly prevalent in patients with type 2 diabetes mellitus (T2DM). Although several insulin resistance (IR)-related indices have been proposed as surrogate markers, their comparative performance and the influence of diabetes duration remain unclear. We compared the associations of eight IR-related indices with MAFLD and evaluated the modifying effect of diabetes duration.Methods
This retrospective cross-sectional study included 5,996 patients with T2DM, of whom 3,087 had MAFLD. Logistic regression was used to evaluate the associations of eight IR-related indices (TyG, TyG-BMI, METS-IR, CHG, CTI, HOMA-IR, SHR, and AIP) with prevalent MAFLD. To facilitate comparison across indices measured on different scales, odds ratios (ORs) were additionally estimated per 1-standard deviation (SD) increase. Receiver operating characteristic (ROC) curve analysis was used to assess discriminative performance. LASSO regression was used for covariate selection, followed by multivariable logistic regression, restricted cubic spline (RCS) analysis, and interaction analysis. Internal validation of the multivariable model was performed using 1,000 bootstrap resamples.Results
All eight IR-related indices were significantly associated with prevalent MAFLD. After standardization, AIP showed the largest association estimate per 1-SD increase (OR = 2.137, 95% CI: 2.007-2.275). AIP also showed the highest, although moderate, discriminative performance among the eight indices (AUC = 0.692). After LASSO-based covariate selection and multivariable adjustment, AIP remained independently associated with prevalent MAFLD (OR = 4.968, 95% CI: 4.100-6.035). Adjusted RCS analysis indicated a nonlinear association between AIP and prevalent MAFLD (P for overall < 0.001; P for nonlinearity = 0.040). The revised multivariable model had an apparent AUC of 0.758 (95% CI: 0.746-0.771), with an optimism-corrected AUC of 0.756 after bootstrap internal validation. Significant interactions were observed between AIP and diabetes duration when duration was modeled continuously (P for interaction = 0.031) and categorically (<5 vs. ≥5 years; P for interaction = 0.001).Conclusions
Among the eight IR-related indices evaluated, AIP showed the largest standardized association with prevalent MAFLD and the highest discriminative performance in patients with T2DM. The association between AIP and MAFLD may vary according to diabetes duration.
Related
- Non-linear association between average daily sleep duration and cardiovascular disease in patients with diabetes or prediabetes
- Association of life course type 2 diabetes duration with adverse cardiac structure independent of current glycaemia in the British MRC National Survey of Health and Development Birth Cohort
- Association Between the Relative Duration of Sliding Scale Insulin Use and the Length of Hospital Stay in Non-Critically Ill Patients with Type 2 Diabetes
- Body mass index modifies the association between occupational noise exposure duration and ultrasound‑detected hepatic steatosis: A cross-sectional study among male factory workers
- Timing matters: a life-course review of midlife type 2 diabetes, exposure duration, and late-life dementia risk
BibTeX
@article{ma2026diabetes,
title = {{Diabetes duration modifies the association between AIP and MAFLD in type 2 diabetes.}},
author = {Jiayu Ma and Xiaokang Hu and Xiaohe Liang and Kangtao Jin},
journal = {Frontiers in endocrinology},
year = {2026},
month = oct,
doi = {10.3389/fendo.2026.1952511},
url = {https://europepmc.org/article/MED/42903117},
note = {CC BY},
}
RIS
TY - JOUR
TI - Diabetes duration modifies the association between AIP and MAFLD in type 2 diabetes.
AU - Jiayu Ma
AU - Xiaokang Hu
AU - Xiaohe Liang
AU - Kangtao Jin
PY - 2026
DA - 2026/10/06
JO - Frontiers in endocrinology
DO - 10.3389/fendo.2026.1952511
UR - https://europepmc.org/article/MED/42903117
AB - Background Metabolic dysfunction-associated fatty liver disease (MAFLD) is highly prevalent in patients with type 2 diabetes mellitus (T2DM). Although several insulin resistance (IR)-related indices have been proposed as surrogate markers, their comparative performance and the influence of diabetes duration remain unclear. We compared the associations of eight IR-related indices with MAFLD and evaluated the modifying effect of diabetes duration.Methods This retrospective cross-sectional study included 5,996 patients with T2DM, of whom 3,087 had MAFLD. Logistic regression was used to evaluate the associations of eight IR-related indices (TyG, TyG-BMI, METS-IR, CHG, CTI, HOMA-IR, SHR, and AIP) with prevalent MAFLD. To facilitate comparison across indices measured on different scales, odds ratios (ORs) were additionally estimated per 1-standard deviation (SD) increase. Receiver operating characteristic (ROC) curve analysis was used to assess discriminative performance. LASSO regression was used for covariate selection, followed by multivariable logistic regression, restricted cubic spline (RCS) analysis, and interaction analysis. Internal validation of the multivariable model was performed using 1,000 bootstrap resamples.Results All eight IR-related indices were significantly associated with prevalent MAFLD. After standardization, AIP showed the largest association estimate per 1-SD increase (OR = 2.137, 95% CI: 2.007-2.275). AIP also showed the highest, although moderate, discriminative performance among the eight indices (AUC = 0.692). After LASSO-based covariate selection and multivariable adjustment, AIP remained independently associated with prevalent MAFLD (OR = 4.968, 95% CI: 4.100-6.035). Adjusted RCS analysis indicated a nonlinear association between AIP and prevalent MAFLD (P for overall < 0.001; P for nonlinearity = 0.040). The revised multivariable model had an apparent AUC of 0.758 (95% CI: 0.746-0.771), with an optimism-corrected AUC of 0.756 after bootstrap internal validation. Significant interactions were observed between AIP and diabetes duration when duration was modeled continuously (P for interaction = 0.031) and categorically (<5 vs. ≥5 years; P for interaction = 0.001).Conclusions Among the eight IR-related indices evaluated, AIP showed the largest standardized association with prevalent MAFLD and the highest discriminative performance in patients with T2DM. The association between AIP and MAFLD may vary according to diabetes duration.
SN - 1664-2392
ER -