Diversity of the fecal microbiota in POAF patients
Firstly, we included 45 POAF patients underwent isolated CABG and matched 90 no-POAF patients. The POAF was detected by 7-day Holter monitoring for any atrial tachyarrhythmia episode lasting ≥30 s. The clinical characteristics are shown in table supplementary 1. The baseline clinical characteristics among the POAF and no-POAF patients were similar.
The diversity index indicates the variety and richness of microbial entities in the gut and is known to be associated with AF recurrence and duration of persistent AF.8, 14 To assess gut microbial diversity in POAF patients, Shannon index and chao1 index were used to assess the ɑ-diversity of the microbiota, Principal coordinate analysis (PCoA) was used for the β-diversity. The differences of Shannon indexes between the POAF and the no-POAF patients were analyzed and were not statistically different (Figure 1A, P=0.067). The difference in chao1 index between the POAF and no-POAF patients was statistically significant (Figure 1B, P= 0.043). For the β-diversity, the PCoA analysis showed significant difference between POAF and no-POAF patients (Figure 1C, P<0.01). These results show a distinctive bacterial composition in different groups.
The microbiota at genus level in POAF patients
To further explore discriminative features between POAF and no-POAF groups, Metastats was performed, and 21 distinct OTUs between the POAF and no-POAF groups were found. At the genus level, the composition of bacterial microbiota of the POAF patients and no-POAF was analyzed. Brevundimonas, Caulobacter, Leuconostoc, and Veillonella were the main up-regulated genus of the bacterial microbiota in POAF patients. Klebsiella, Streptococcus and Escherichia-Shigella are significantly down-regulated in POAF patients (Figure supplementary 1 and table supplementary 2).
Berberine prevented POAF: a randomized, double-blind, placebo-controlled trial.
As the shifts in gut microbiome composition in POAF patients, a bacteriostatic agent, berberine were used to altering the gut microbiota and conducted a randomized, double-blind, placebo-controlled trial. Two hundred patients underwent isolated off-pump CABG gave informed consent to participate in this trial and were randomized to the berberine group (n = 100) and the placebo group (n = 100) from January 1, 2020 to September 31 2020 (Figure 2). There is no difference at baseline characteristics between the two groups (Table supplementary 3). All the patients were randomized to receive either berberine or placebo for 5-7 days before surgery. Following cardiac surgery, patients received berberine or placebo via oral or gastric tube until 7 days after the surgery.
Primary Outcomes
All 200 patients were recorded by continuous Holter electrocardiographic and stayed in the hospital for more than 7 days. Thirty-five patients in the placebo group and 20 patients in the berberine group developed POAF during the first 7 days after CABG, respectively (hazard ratio: 0.5; 95% confidence interval: 0.29 to 0.78; p = 0.0143) (Figure 3).
Secondary Outcomes: AF related measurements and clinical outcomes
AF burden, maximum times of AF and maximum ventricular rate were significantly lower in the berberine group (secondary outcomes in supplementary tables). However, length of hospitalization and postoperative ventilator support time did not differ between the 2 groups (secondary outcomes in supplementary tables).
Secondary Outcomes: a significant impact of berberine on the human gut microbiota.
To evaluate how does berberine prevented POAF through gut microbiota, 16S sequencing was performed in the patients allocated to berberine group. Seventy of 100 patients under the treatment of berberine were collected stool samples at both times, one is before berberine administration, the other time point is having berberine for 5-7 days before surgery. Analysis showed that the alterations in the gut microbiome after berberine treatment, regarding to the α-diversity and β-diversity (Figure 4).
Further Metastats analysis indicates that 50 distinct OTUs post berberine treatment. Morganella, Pyramidobacter and Erysipelatoclostridium were increased after berberine treatment. Also, Faecalibacterium, Roseburia, Alistipes and Veillonella were decreased significantly (Figure supplementary 2 and Table supplementary 4). Interestingly, patients with high Veillonella abundances were more likely to developed POAF, however, under the treatment of berberine, Veillonella was decreased more than two folds (Figure 5).
Berberine inhibited intestinal endotoxin levels and systemic inflammation.
As Veillonella was gram-negative diplococci and could produce lipopolysaccharide (LPS) and induce inflammation reactions. We next detected the biomarkers of intestinal endotoxin levels and systemic inflammation. The level of LPS, Interleukin- 6 (IL-6), C-reactive protein (CRP) and neutrophil to lymphocyte ratio (NLR) were similar in both berberine and placebo groups at baseline (Table supplementary 5). After 48 hours of surgery, the LPS, IL-6 and CRP levels were significantly lower in berberine group (Table 1 and Figure 6).
Table 1
The biomarkers of intestinal endotoxin levels and systemic inflammation after 48 hour of surgery
|
berberine placebo
|
P
|
LPS
|
496.1( 460.4-518.3)
|
560.0(528.2-613.5)
|
<0.001
|
IL-6
|
67.32(37.05-137.76)
|
179.18(107.59-283.70)
|
<0.001
|
CRP
|
161.90(157.00-172.30)
|
167.35(161.08-195.50)
|
0.001
|
NLR
|
8.97(6.80-12.46)
|
12.09(9.08-16.52)
|
<0.001
|
NLR: Neutrophil/Lymphocyte Ratio. |
Intestinal endotoxin levels and POAF
Then we investigated the role of LPS in peripheral blood after surgery 48 hours, identifying a significant increase in LPS in patients with POAF (Figure 7) (p<0.0001). Further subgroup analysis found that LPS levels were increased in POAF patients but did not reach statistical significance in berberine groups (Figure 7) (p=0.0591). However, in placebo group, we observed that the LPS levers were significantly higher in POAF patients (Figure 7) (p=0.0034).