Alcohol-related liver disease is one of the most prevalent liver diseases in the United States. Early stages of alcohol-related liver disease are characterized by accumulation of triglycerides in hepatocytes. Alcoholic hepatitis is a severe form of alcohol-related liver disease associated with significant morbidity and mortality. We sought to identify patients who are at greatest risk of death using serum lipids. First, we performed lipidomics analysis on serum samples collected from 118 patients with alcoholic hepatitis to identify lipid markers that are associated with high risk of death. Next, we performed gene set enrichment analysis on liver transcriptomics data to identify dysregulated lipid metabolism in patients who received liver transplantation. Finally, we built a random forest model to predict 30-day mortality using serum lipids. A total of 277 lipids were annotated in the serum of patients with alcoholic hepatitis, among which 25 were significantly different between patients in the deceased and alive groups. Five chemical clusters were significantly altered between the two groups. In particular, acylcarnitine cluster was enriched in the deceased group. Several hepatic lipid metabolism pathways were dysregulated in patients with alcoholic hepatitis who received liver transplantation. The mRNA expression of genes involved in the fatty acid transport into mitochondria and β-oxidation were also dysregulated. When predicting 30-day mortality in alcoholic hepatitis patients using serum lipids, we found that the area under the curve achieved 0.95. Serum lipids such as acylcarnitines may serve as biomarkers to identify alcoholic hepatitis patients at the greatest risk of death.
Keywords: acylcarnitine; carnitine system; lipidomics; random forest.