The tongue coating microbiome is perturbed in atrial fibrillation and partly normalized after catheter ablation

Abstract

Background: There is accumulating evidence linking the microbiome and cardiovascular diseases. Nevertheless, no existing studies have been conducted on atrial fibrillation (AF) and the oral microbiome.

Materials and methods: We collected and sequenced 245 AF tongue-coating samples and 26 AF samples after catheter ablation from Zhengzhou and Guangshan, China. We characterized tongue coating microbiome, constructed microbial classifiers in the discovery cohort, and verified their diagnostic potential in a cross-regional cohort.

Results: Tongue coating microbial richness and diversity were significantly increased in the AF group compared to the control group, indicating increased bacterial colonization. The classifiers based on four optimal tongue coating microbial markers achieved good diagnostic efficiency in AF cohorts, with area under the curve (AUC) of 99.10 and 98.62% in the discovery and validation cohorts, respectively, and 97.97% in the cross-regional cohort. Paroxysmal AF and persistent AF shared similar taxonomic features, but some specific differential bacteria acted in the AF progression. Moreover, the outcomes revealed that catheter ablation contributed to rehabilitating oral bacterial disorders.

Conclusion: This was the first cross-sectional and longitudinal research of oral microbiome in AF patients and the alternations after catheter ablation, which offers promising new perspectives for AF clinical diagnosis and management.

Keywords: atrial fibrillation; biomarkers; catheter ablation; diagnosis; tongue coating microbiome.

Figure 1

Study design and flow diagram. A total of 544 tongue coating samples from Henan Provincial Chest Hospital and Guangshan County People’s Hospital were collected. After rigorous inclusion and exclusion criteria, 529 samples were included for further analysis, including 219 AF, 26 ACA, 29 paired BCA, 229 HC samples from Zhengzhou and 26 AF samples from Guangshan. Tongue coating samples were sequenced using 16S rRNA MiSeq to characterize the microbiome and construct predictive or diagnostic model. AF, atrial fibrillation; ACA, 6 months after catheter ablation; BCA, before catheter ablation; HC, healthy control; PAF, paroxysmal AF; psAF, persistent AF; RFC, random forest classifier; rRNA, ribosomal RNA.

Figure 2

Microbial dysbiosis in the tongue coat of AF patients. (a) A Venn diagram displayed that 2,542 of 4,930 OTUs were common to both AF and HC groups, while 1,069 OTUs were unique to the AF. (b) Rarefaction analysis between the number of samples and the number of OTUs. As the number of samples increased, the number of OTUs approached saturation. Compared with the HC, the number of OTUs in AF was decreased. As estimated by the Chao index, tongue coating microbial richness was significantly increased in AF compared with HC. (c) Shannon-Wiener curves of samples showed that as the number of sequences increased, the Shannon diversity approached saturation. As estimated by the Shannon index, tongue coating microbial diversity was significantly increased in AF compared with HC. The PCoA (d) and NMDS (e) (Bray-Curtis distance) analysis showed the tongue coating microbial taxonomic composition was conspicuously different between the two groups. The ratio of the variance contribution was shown. (f) Average compositions and relative abundance of the bacterial community in both groups at the genus level. (g) At the genus level, 37 microbial features were enriched in AF, while 15 microbial features were depleted. AF, atrial fibrillation; HC, healthy control; OTUs, operational taxonomy units; PCoA, principal coordinate analysis; NMDS, nonmetric multidimensional scaling. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001.

Figure 3

Identification of a tongue coating microbial classifier for AF. (a) Four microbial markers were selected as the best markers set by random forest model. (b) Importance distribution map of the selected microbial markers in the model. The POD value was significantly increased in AF compared with HC, and achieved good diagnostic efficacy in the discovery cohort (c,d), the validation cohort (e,f), and the independent cohort (g,h). AF, atrial fibrillation; HC, healthy control; POD, probability of disease; AUC, area under the curve. Centerline, median; box limits, upper and lower quartiles; circle or square symbol, mean; error bars, 95% CI.

Figure 4

Tongue coating microbial characterization among PAF and psAF. (a) A Venn diagram displayed that 1,701 of 2,970 OTUs were shared between PAF and psAF, while only 492 and 777 OTUs were unique to the PAF and psAF, respectively. (b) Rarefaction analysis between the number of samples and the number of OTUs. As the number of samples increased, the number of OTUs approached saturation. As estimated by the Chao index, the tongue coating microbial richness of PAF and psAF was similar. (c) Shannon-Wiener curves of samples showed that as the number of sequences increased, the Shannon diversity approached saturation. As estimated by the Shannon index, tongue coating microbial diversity was similar in PAF and psAF. The PCoA (d) and NMDS (e) (Bray-Curtis distance) showed that there was no significant difference in the tongue coating microbiome distribution between PAF and psAF. The ratio of the variance contribution was shown. (f) Heatmap for the relative abundances of differential OTUs for each sample in the PAF and psAF groups. (g) Average compositions and relative abundance of the bacterial community in the PAF and psAF groups at the genus level. (h) Compared with PAF, two genera were increased, while three genera were depleted in psAF. PAF, paroxysmal AF; psAF, persistent AF; OTUs, operational taxonomy units; PCoA, principal coordinate analysis; NMDS, nonmetric multidimensional scaling. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001.

Figure 5

Alterations in tongue coating microbiome in AF patients 6 months after catheter ablation. (a) Rarefaction analysis between the number of samples and the number of OTUs. As the number of samples increased, the number of OTUs approached saturation. Compared with the HCs, the number of OTUs in ACA and BCA was increased. As estimated by the Shannon index, the tongue coating microbial diversity of ACA and BCA was similar but significantly increased compared with that of the HC. (b) A Venn diagram revealed that 1,407 of 1,951 OTUs in ACA were shared with HC, and 1,580 OTUs were shared with BCA. PCoA (c) and NMDS (d) (Bray-Curtis distance) displayed that the tongue coating microbial distribution in ACA was comparable to that of BCA but distinct from that of HC. The ratio of the variance contribution was shown. (e) The abundance of 15 genera gradually decreased after catheter ablation and 12 genera gradually increased. (f) Average compositions and relative abundance of the bacterial community in the three groups at the genus level. (g) Heatmap for the relative abundances of differential OTUs for each sample in the three groups. HC, healthy control; ACA, 6 months after catheter ablation; BCA, before catheter ablation; OTUs, operational taxonomy units; PCoA, principal coordinate analysis; NMDS, nonmetric multidimensional scaling. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001.

Figure 6

Crucial tongue coating microbial predicted functions correlated with AF progression. (a) Compared with HC, 34 functional modules increased significantly, while 27 functions decreased significantly in AF (LDA > 3, p < 0.05). (b) Compared with PAF, 6 functional modules increased significantly, while 10 functions decreased significantly in AF (LDA > 3, p < 0.05). (c) Based on the LDA selection, 4 functions were significantly increased in BCA, 14 functions were remarkedly raised in ACA, while 16 functions were notably increased in HC (LDA > 3, p < 0.05). AF, atrial fibrillation; HC, healthy control; PAF, paroxysmal AF; psAF, persistent AF; ACA, 6 months after catheter ablation; BCA, before catheter ablation; LDA, linear discriminant analysis.

Figure 7

Associations between the tongue coating microbiome and clinical indicators. Spearman correlations of significantly altered tongue coating microbiome with significantly altered clinical indicators in AF and HC. Red color represents the positive correlation, and green color represents the negative correlation. GLU, glucose; IBIL, indirect bilirubin; BUN, blood urea nitrogen; CRP, C-reactive protein; GLB, globulin; ALB, albumin; TP, total protein; DBIL, direct bilirubin; LDH, lactate dehydrogenase; PLT, blood platelet; CK, Creatine kinase; LDL, low-density lipoprotein; TG, triglyceride; RBC, red blood cells; OTU, operational taxonomy unit. ^*p < 0.05, ^**p < 0.01, ^***p < 0.001.