The Impact of Liver Graft Preservation Method on Longitudinal Gut Microbiome Changes Following Liver Transplant: A Proof-of-concept Study

Abstract

Background and aims: End-stage liver disease is associated with disruptions in gut microbiota composition and function, which may facilitate gut-to-liver bacterial translocation, impacting liver graft integrity and clinical outcomes following liver transplantation. This study aimed to assess the impact of two liver graft preservation methods on fecal microbiota and changes in fecal and breath organic acids following liver transplantation.

Methods: This single-center, non-randomized prospective pilot study enrolled liver transplant patients whose grafts were preserved using either static cold storage or ex situ normothermic machine perfusion (NMP). Fresh stool and breath samples were collected immediately before surgery and at postoperative months 3, 6, and 12. Stool microbiota was profiled via 16S rRNA gene sequencing, stool short-chain fatty acids were measured using gas chromatography/-mass spectrometry, and breath volatile organic compounds (VOCs) were analyzed with selected-ion flow-tube mass spectrometry.

Results: Both cohorts experienced a loss of microbiota diversity and dominance by single taxa. The NMP cohort demonstrated enrichment of several beneficial gut taxa, while the static cold storage cohort showed depletion of such taxa. Various gut bacteria were found to correlate with stool short-chain fatty acids (e.g., lactic acid, butyric acid) and several VOCs.

Conclusions: Fecal microbiota alterations associated with end-stage liver disease do not fully normalize to a healthy control profile following liver transplantation. However, notable differences in microbiota composition and function were observed between liver graft preservation methods. Future research with larger randomized cohorts is needed to explore whether the NMP-associated shift in gut microbiota impacts clinical outcomes and if breath VOCs could serve as biomarkers of the clinical trajectory in liver transplant patients.

Keywords: Breath metabolites; Ex situ normothermic liver perfusion; Microbiota; Short-chain fatty acids; Static cold storage; Volatile organic acids.

Fig. 1. Gut microbiota diversity.

Alpha- and beta- diversity plots visualize the longitudinal changes in the microbial community structure of fecal samples in liver transplant recipients at pre-transplant, three, six, and twelve months post-transplant time points. (A) Box plots demonstrating alpha diversity (Shannon index) values. Box plots show the median, lower, and upper quartiles and are color-coded by SCS, NMP, and healthy control cohorts. (B) PCA depicting beta diversity patterns in SCS vs. NMP cohorts. *p < 0.05; **p < 0.01 (PERMANOVA with Benjamini-Hochberg FDR). PCA, Principal component analysis; SCS, static cold storage; NMP, normothermic machine perfusion; FDR, false discovery rate.

Fig. 2. Longitudinal core microbiome analysis.

Longitudinal core microbiome analysis reveals significant taxonomic alterations in liver transplant recipients at (A) pre-transplant, and (B) three, (C) six, and (D) twelve months post-transplant. Heatmaps represent the longitudinal core microbiome of SCS and NMP cohorts at the genus level as a function of relative abundance. The x-axis represents detection thresholds (relative abundance), ranging from lower (left) to higher (right) values. Color shading indicates the prevalence of each bacterial genus among samples for each abundance threshold. SCS, static cold storage; NMP, normothermic machine perfusion.

Fig. 3. Differential abundance of top taxa.

Differential abundance analysis for (A) SCS and (B) NMP cohorts reveals statistically significant taxa between pre-transplant and 12-month time points. (White’s nonparametric t-test with Benjamini-Hochberg FDR multiple test correction, adjusted p ≤ 0.05). SCS, static cold storage; NMP, normothermic machine perfusion; FDR, false discovery rate.

Fig. 4. Gut microbiota and fecal and breath organic compounds.

All SCS and NMP samples across all time points for fecal and breath samples were pooled. A Spearman correlation heatmap displays the relationships between fecal and breath metabolites and fecal microbiota at the genus level. Red squares represent positive correlations. Blue squares represent negative correlations. White squares represent no correlation. *p < 0.05, **p < 0.01. SCS, static cold storage; NMP, normothermic machine perfusion.

Fig. 5. Correlation of fecal microbiota with factors known to disturb gut microbiota.

All patient samples across all time points were pooled. DAA of ASVs (genus-level) was performed: (A) Comparing patients receiving antibiotics (yes) versus those not receiving antibiotics (no); (B) Comparing patients receiving a PPI or histamine 2-antagonist (yes) versus those not receiving these medications (no); and (C) Assessing patients with varying levels of malnutrition (none, mild, moderate, or severe) at the pre-transplant time point. ASV tables were rarefied to the sample with the lowest number of sequences in each analysis. ASVs were assigned at the genus level, and genera with a relative abundance of more than 0.3% of the total were included in the DAA analysis. FDR estimates were calculated for multiple comparisons, with a significance threshold of p < 0.05. ASVs, amplicon sequence variants; DAA, differential abundance analysis; FDR, false discovery rate; PPI, proton pump inhibitor.