Abstract:

Objective: To compare the differences in gut microbiota and their functional pathways between drug-resistant pulmonary tuberculosis (DR-PTB) patients with and without drug-induced liver injury (DILI), so as to identify gut microbiota and functional pathways associated with the development of DILI. Methods: A nested case-control design was adopted with Hunan Chest Hospital as the research site. Newly admitted inpatients diagnosed with DR-PTB were prospectively enrolled starting from September 8, 2021, to establish a follow-up cohort of DR-PTB patients. All cohort participants were followed up, and liver function tests were performed to monitor the occurrence of DILI during anti-tuberculosis treatment. By November 4, 2022, all patients diagnosed with DILI in the cohort were assigned to the DILI group (68 cases). Using a 1∶1 matching strategy, DR-PTB patients with normal liver function were selected as the control group (68 cases), matched by gender, age (±3 years), duration of anti-drug-resistant TB treatment (±1 month), and sample collection time (±1 month). Metagenomic sequencing was applied to extract DNA information from fecal samples. Software including R was used to analyze gut microbial alpha diversity, beta diversity, differential species, differential functional pathways, and correlations between differential species and differential functions across the two groups. Results: No statistically significant differences were observed in gut microbial alpha diversity and beta diversity between the DILI group and the control group (P>0.05). After adjusting for confounders such as age, eight bacterial species showed significant intergroup differences. Four species were enriched in the DILI group: Dorea longicatena (P=0.007), Prevotella sp. CAG:386 (P=0.012), Prevotella stercorea (P=0.013), and Prevotellamassilia timonensis (P=0.043). The other four species were enriched in the control group: Clostridium sp. CAG:7 (P=0.020), Eubacterium sp. CAG:38 (P=0.002), Faecalibacterium sp. CAG:82 (P=0.047), and Klebsiella michiganensis (P<0.001). At KEGG level 3, pyruvate metabolism (P=0.010), butanoate metabolism (P=0.034), valine, leucine and isoleucine degradation (P=0.009), ethoxylate and dicarboxylate metabolism (P=0.041), and cationic antimicrobial peptide (CAMP) resistance (P=0.031) were enriched in the control group. Starch and sucrose metabolism (P=0.027) and fructose and mannose metabolism (P=0.005) were enriched in the DILI group. The abundances of seven differential functional pathways were positively correlated with the abundance of Klebsiella michiganensis (P<0.05). The pathway of valine, leucine and isoleucine degradation was negatively correlated with the abundances of the other seven differential species including Eubacterium sp. CAG:38 and Faecalibacterium sp. CAG:82 (P<0.05). Ethoxylate and dicarboxylate metabolism was negatively correlated with the abundances of Prevotella stercorea, Prevotellamassilia timonensis and Clostridium sp. CAG:7 (P<0.001). Pyruvate metabolism was negatively correlated with the abundances of Prevotella stercorea and Prevotellamassilia timonensis (P<0.001). Butanoate metabolism was negatively correlated with the abundance of Prevotella stercorea (P<0.001). Conclusion: Differences in gut microbial composition already exist prior to DILI onset between DR-PTB patients who subsequently develop DILI and those who do not. Abnormal functions related to starch, sucrose, fructose, mannose, pyruvate and butanoate metabolism, as well as impaired valine, leucine and isoleucine degradation, may be associated with DILI occurrence in patients with drug-resistant pulmonary tuberculosis.

Key words: Tuberculosis, multidrug-resistant, Drug-induced liver injury, Gut microbiota, Case-control studies, Factor analysis, statistical