Assessing Nutritional Factors for Metabolic Dysfunction-Associated Steatotic Liver Disease via Diverse Statistical Tools.

Abstract

Background: Lifestyle modifications are critical in addressing metabolic dysfunction-associated steatotic liver disease (MASLD); however, the specific macronutrients that most significantly influence the disease's progression are uncertain. In this study, we aimed to explore the role of carbohydrate, fat, and protein intake in MASLD development using decision trees, random forest models, and cluster analysis.

Methods: Participants (n=3,951) from the Korean Genome and Epidemiology Study were included. We used the classification and regression tree analysis to classify participants into subgroups based on variables associated with the incidence of new-onset MASLD. Random forest analyses were used to assess the relative importance of each variable. Participants were grouped into homogeneous clusters based on carbohydrate, protein, fat, and total caloric intake using hierarchical cluster analysis. Subsequently, we used the Cox proportional hazards regression models to estimate hazard ratios (HRs) and 95% confidence intervals (CIs) for MASLD risk across the clusters.

Results: Carbohydrate intake was identified as the most significant predictor of new-onset MASLD, followed by fat, protein, and total caloric intake. Participants in cluster 3, who consumed a lower proportion of carbohydrate but had higher total caloric, protein, and fat intake, had a lower risk of new-onset MASLD than those in cluster 1 after adjusting for confounders (cluster 1 as a reference; cluster 3: HR, 0.90; 95% CI, 0.82 to 0.99).

Conclusion: The study's results highlight the critical role of macronutrient composition, particularly carbohydrate intake, in MASLD development. The findings suggest that dietary strategies focusing on optimizing macronutrients, rather than simply reducing caloric intake, may be more effective in preventing MASLD.