Serum Biomarker Identification of Damp-Heat Pattern in Patients with Chronic Liver Diseases.

OBJECTIVE

To identify serum diagnostic biomarker for damp-heat (DH) pattern in chronic liver diseases using transcriptomics and metabolomics.

METHODS

Patients with chronic hepatitis B (CHB) or metabolic dysfunction-associated fatty liver disease (MAFLD) were categorized into DH and non-DH pattern groups. Serum RNA profiles were analyzed via RNA-seq/microarray, and metabolites were quantified by ultraperformance liquid chromatography coupled with tandem mass spectrometry (UPLC-MS/MS). Biomarker screening and validation employed discovery (88 cases) and validation (85 cases) cohorts, utilizing Rank-in analysis and Least Absolute Shrinkage and Selection Operator (LASSO) regression for gene selection; reverse transcription-polymerase chain reaction (RT-PCR) and targeted UPLC-MS/MS for expression validation; random forest and receiver operating characteristic (ROC) curve analysis (with area under the curve, AUC) for diagnostic assessment.

RESULTS

Compared with the non-DH pattern group, patients with DH pattern showed significantly elevated liver injury markers and reduced apolipoprotein A1 (P<0.05). Integrated transcriptome analysis (Rank-in) identified 315 dysregulated gene sets, primarily enriched in chemokine signaling. LASSO selected 27 genes for RT-PCR validation, confirming 5 differential genes. Metabolomics revealed 25 differential metabolites (discovery cohort), with 7 showing ⩾ 2-fold change; 6 maintained consistent trends in validation. A random forest model combining 4 genes (PTPN22, CTSD, TBX21, STAT4) and 2 metabolites (pyroglutamic acid and glutamic acid) achieved a validation AUC of 0.828 for DH diagnosis.

CONCLUSION

A multi-omics diagnostic model incorporating 4 genes and 2 metabolites demonstrates promising diagnostic potential for DH pattern in patients with chronic liver diseases. (registration No. ChiCTR2000037248).