Heterogeneity in risk and potential pathogenic associations of NAFLD among distinct prediabetic phenotypes in young and middle-aged adults.

OBJECTIVE

Prediabetes exhibits a high comorbidity rate with non-alcoholic fatty liver disease (NAFLD), significantly increasing the risks of diabetes progression and other complications. This study aims to clarify the differences in NAFLD risk among different prediabetes phenotypes and explore their underlying associations.

METHODS

A cohort of 2127 prediabetic individuals aged 18-59 years was enrolled. Latent Profile Analysis (LPA) was employed to classify phenotypes based on age, insulin resistance score, and body mass index. Machine learning algorithms and mediation analysis were used to investigate the risk features of NAFLD and the phenotype-specific pathogenic pathways.

RESULTS

Four distinct prediabetes phenotypes were identified, each demonstrating significant differences in the probability of NAFLD. The prevalence of NAFLD was highest (87.93%) in Class 1, characterized by young individuals with severe insulin resistance and obesity. Conversely, the lowest prevalence (18.39%) was observed in middle-aged individuals with mild insulin resistance and normal weight. In Class 1 individuals, hepatocellular damage showed the strongest association with NAFLD. Young individuals with moderate insulin resistance and overweight exhibited a higher NAFLD risk associated with the activation of the metabolism-immunity-inflammation axis. Among middle-aged individuals with moderate insulin resistance and overweight, neutrophilic inflammation was the most relevant factor. Additionally, some individuals displayed exacerbated metabolic abnormalities associated with elevated triglycerides, indirectly increasing the risk of NAFLD.

CONCLUSIONS

The study characterized novel prediabetes phenotypes and their pathogenic associations with NAFLD, providing a new perspective for targeted prevention and management. This research approach addresses the limitations of the "one-size-fits-all" management model and contributes to the improvement of endocrine disease management.