Korean J Transplant.  2022 Nov;36(Supple 1):S268. 10.4285/ATW2022.F-4075.

A potential role of gut microbiome in predicting of early acute rejection after kidney transplantation

• Kim J
• Park S
• Cho H
• Park JI
• Han A
• Min S
• Ha J
• Kim YS
• Lee H

• KMID: 2535495
• DOI: http://doi.org/10.4285/ATW2022.F-4075

Abstract

Background
Gut microbiota may affect host immunity and therefore it may be associated with the immunologic response of kidney transplantation (KT) recipients, given their risk for allograft rejection. The aim of this study is to explore the association between gut dysbiosis and early acute rejection (EAR) in KTs.
Methods
Stool samples from a tertiary hospital were collected from KT recipients before transplantation and metagenomic shotgun sequencing was performed for taxonomic profiling and detection of microbiota-derived genes. Their clinical data were gathered, including demographic factors, immunologic risks, immunosuppressive treatment, and the outcome defined as biopsy-proven EAR within 2 weeks of KT. Using a trainset, EAR prediction models were built in combination of clinical data, microbio-ta taxonomy, and microbiota-derived gene families, and tested the change of statistical power for EAR prediction and possibility for validation in the independent cohort.
Results
A total of 78 and 71 stool samples were collected for a train-test set and a validation set, respectively. EAR was found in 26 (33.3%) and 20 (28.2%) in each set, respectively. There was no specific difference in clinical characteristics except a higher proportion of hypertension in validation set. Recipients experiencing EAR showed a higher body mass index, number of HLA mismatch, and a higher rate of delayed graft function compared to nonrejection recipients. In taxonomic profiling, the abundance of Bacteroides eggerthii, Phascolarctobacterium faecium, and Desulfovibrio piger decreased in rejection group. In gene families analysis, a total of 532 genes including 85 metabolism pathway related genes were differentially expressed. Predictability of EAR was enhanced (C-statistics, 0.77; 95% confidence interval, 0.59–0.86) when we add on the differently expressed microbiota-de-rived genes and taxonomic profiles to the clinical model (C-statistics, 0.57; 95% confidence interval, 0.35–0.67). In addition, the optimized model showed a modest performance with 64.8% of accuracy in the validation cohort.
Conclusions
The gut microbiome-driven metagenomic signatures may have an additive role in predicting EAR after KT when combined with clinical and immunologic features.