Association between Helicobacter pylori infection and MASLD: a cross-sectional study with interpretable machine...

Area of Science:

• Hepatology
• Gastroenterology
• Infectious Diseases

Background:

• Metabolic dysfunction-associated steatotic liver disease (MASLD) is increasingly associated with Helicobacter pylori (HP) infection.
• Existing evidence on this association is inconsistent, and the predictive role of HP status in integrated risk models remains unclear.

Purpose of the Study:

• To evaluate the association between HP infection and MASLD.
• To develop and validate interpretable machine learning models for MASLD risk stratification.

Main Methods:

• Analysis of two cohorts: a Chinese hospital-based cohort (n=1,021) and the U.S. NHANES (n=4,870).
• HP infection assessed via urea breath test (China) and serum IgG (NHANES). MASLD diagnosed by ultrasonography (China) and approximated by Fatty Liver Index (NHANES).
• Logistic regression, mediation analysis, and eleven machine learning models were employed, with external validation in NHANES.

Main Results:

• HP infection showed a significant association with MASLD in both the Chinese (OR=4.83) and U.S. (OR=1.22) cohorts.
• Machine learning models achieved strong internal performance (AUC up to 0.8876) and moderate external discrimination (AUC=0.6536).
• SHAP analysis identified HP and LHR as top predictors for MASLD, with TyG indices as secondary contributors.

Conclusions:

• HP infection is associated with MASLD in both Chinese and U.S. populations.
• Interpretable machine learning models show promise for MASLD risk stratification, consistently identifying key predictors.
• Findings are associative due to study design and diagnostic variations; external validation is exploratory.