Diagnostic Performance of Six Insulin Resistance Indices for MAFLD Across BMI Categories in Type 2 Diabetes.

Abstract

Objective: To evaluate the diagnostic utility of six insulin resistance surrogate indices for metabolic dysfunction-associated fatty liver disease (MAFLD) across different body mass index (BMI) categories in patients with type 2 diabetes (T2DM).

Methods: This single-center, cross-sectional study utilized a consecutive sampling method and was conducted at the Department of Endocrinology and Geriatrics, Huaian First People's Hospital (January 2018 to December 2022). Inclusion criteria were age ≥18 years and confirmed T2DM. A total of 2079 hospitalized adult patients meeting these criteria were initially screened. After applying exclusion criteria (including other diabetes types, secondary liver diseases, excessive alcohol consumption, low BMI, and incomplete data), 1205 patients with T2DM were enrolled and stratified by body mass index (BMI) into normal weight (n=429) and overweight/obese groups (n=776). MAFLD was diagnosed based on abdominal ultrasound, yielding 636 cases of MAFLD and 569 non-MAFLD controls. Six insulin resistance surrogate indices-homeostasis model assessment of insulin resistance (HOMA-IR), triglyceride-glucose index (TyG), triglyceride-glucose-body mass index (TyG-BMI), metabolic score for insulin resistance (METS-IR), triglyceride to high density lipoprotein cholesterol ratio (TG/HDL-C), and Chinese visceral adiposity index (CVAI)-were assessed for their diagnostic performance using receiver operating characteristic (ROC) analysis, DeLong's test, and decision curve analysis (DCA).

Results: The overall prevalence of MAFLD was 52.78%, with rates of 34.03% in the normal weight group and 63.14% in the overweight/obese group. In the overweight/obese cohort, TyG-BMI demonstrated superior diagnostic accuracy (AUROC 0.769; optimal cut-off 240.503; sensitivity 76.37%, specificity 66.91%), significantly outperforming all other indices except for METS-IR, which showed comparable efficacy and could serve as an alternative. Among normal-weight individuals, TG/HDL-C exhibited the highest diagnostic performance (AUROC 0.763; optimal cut-off 2.908; sensitivity 70.00%, specificity 73.58%), which was significantly better than HOMA-IR and CVAI, but showed no significant difference compared with TyG, TyG-BMI, or METS-IR. DCA indicated that TyG-BMI consistently provided the greatest net clinical benefit, particularly within the high-risk threshold range (0.4-0.6), regardless of BMI category.

Conclusion: TyG-BMI should be prioritized as the primary screening indicator for MAFLD in overweight and obese T2DM patients, with METS-IR as a valid alternative. In normal weight patients, TG/HDL-C demonstrates excellent clinical utility, with TyG-BMI serving as effective alternatives.