Background <p>Sepsis is frequently combined with acute liver failure (ALF), a critical determinant in the mortality of septic patients. Pyroptosis is a significant form of programmed cell death that plays an important role in the inflammatory response. Research has been conducted to elucidate the relationship between pyroptosis, sepsis, and ALF, but the mechanism of action remains unclear.</p> Methods <p>Datasets relating to sepsis and ALF were obtained from the Gene Expression Omnibus (GEO). The intersection of differentially expressed genes (DEGs) and pyroptosis-related genes for sepsis and ALF was identified. Simultaneously, a gene diagnosis model for sepsis and ALF was developed using machine learning, and the model’s accuracy was assessed through the plotting of the ROC curves and confusion matrix. The Hub genes identified by the model with an area under the curve (AUC) value ≥ 0.7 were used for the investigation of immune cell infiltration to explain the molecular mechanism of sepsis combined with ALF. The precise mechanism of action of these model genes in sepsis combined with ALF was evaluated through animal experiments.</p> Results <p>Machine learning revealed that GABARAP and ITCH may serve as diagnostic biomarkers for pyroptosis in sepsis combined with ALF. The examination of immune cell infiltration indicated that immune dysregulation is present in both sepsis and ALF and preliminarily suggested that GABARAP and ITCH may be pivotal in cellular immunity responses, particularly those mediated by T cells. Animal experiments further validated that in the process of sepsis combined with ALF, the expression level of GABARAP is elevated, while the expression level of ITCH is diminished.</p> Conclusions <p>We found GABARAP and ITCH may serve as diagnostic biomarkers for pyroptosis in sepsis combined with ALF, suggesting their potential involvement in the initiation and advancement of sepsis combined with ALF through cellular immunomodulatory pathways.</p> Clinical trial number <p>Not applicable.</p>