Association between enterotypes of the gut microbiota and features of ischemic stroke

Abstract

An enterotype is a classification of patients’ gut microbes into three types, and these types differ in their features of cardiovascular disease. We hypothesized that patients have different enterotypes depending on their arteriosclerosis risk factors, stroke subtype, and severity of stroke. Stool samples were prospectively collected from 100 patients (average age: 73.4 years, 38 women) with ischemic stroke. Patients’ data were obtained from their medical records. Phylogenetic analyses of the 16 S rRNA gene (V3–V4 region) extracted from each stool sample were performed. Quantitative analyses of the presence of each bacterial genus in the intestines were performed using a next-generation sequencer. After the number of each genus of gut microbes was extracted, ≥ 30% of the patients with the genus Bacteroides were classified as type I, ≥ 15% with the genus Prevotella were classified as type II, and the rest were classified as type III. We analyzed the association between the patients’ enterotypes and their characteristics (i.e., arteriosclerosis risk factors such as stroke subtype, and severity of stroke). 33 patients had type I, 10 had type II, and 57 had type III, with no overlap. Patients with types I and II had a lower prevalence of dyslipidemia than those with type III (P = 0.028), a lower National Institute of Health and Stroke Scale score at admission (P = 0.025), and the modified Rankin Scale score at discharge tended to be lower (P = 0.094). The enterotype may affect the risk factors and severity of ischemic stroke.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-18378-z.

The gut microbiota has been increasingly attracting attention as the major cause of various diseases because of the recent development of metagenomics. Atherosclerosis is strongly associated with diet. Therefore, the gut microbiota has long been suspected to play a crucial role in the progression of atherosclerosis.

In the gut microbiota, the genus Bacteroides has attracted much research focus because it is frequently found in feces. The association of Bacteroides with a high-fat diet has been reported mainly in Western countries¹. One study reported that people with a high level of Bacteroides in their stool samples had a diet high in protein and fat². Another study showed that the genera Bacteroides, Alistipes, and Bilophila were increased in people who ingested a diet rich in animal fat for several days³. However, people with a high body mass index were reported to have a lower rate of Bacteroides than those with a low body mass index⁴.

In contrast, some researchers from East Asia disagree with the opinion that the genus Bacteroides may detrimentally affect atherosclerosis. A study from Japan showed that the abundance of the phylum Bacteroidetes was decreased and that of the phyla Firmicutes and Actinobacteria were increased in patients with type 2 diabetes compared with healthy subjects⁵. A Chinese study showed that the abundance of Bacteroides and Prevotella was decreased in patients with symptomatic atherosclerotic ischemic stroke and transient ischemic attack⁶.

Arumugam et al. suggested that the human gut microbiota could be stratified into three enterotypes, which could help identify the characteristics of people and their diseases⁷. Yamashita et al. found that, using this classification, patients with coronary heart disease had a low rate of Bacteroides (< 30%)⁸. This study suggests the usefulness of the enterotype classification.

Although several studies have reported on the gut microbiota in patients with ischemic stroke^6,9there have been few reports on the association between enterotypes and atherosclerosis in patients with stroke, especially regarding the association of Bacteroides⁸. Therefore, this study aimed to determine the association of enterotypes with the features of stroke, especially the progression of atherosclerosis.

Methods

We prospectively registered 100 patients (73.4 ± 11.3 years; 62 men, 38 women) who were hospitalized for the treatment of ischemic stroke between October 2021 and December 2022. All of the patients were admitted to our neurological center within 7 days following the occurrence of ischemic stroke. We diagnosed patients with ischemic stroke when they had a rapid onset of focal brain disturbance. We confirmed the findings with diffusion-weighted imaging and conducted carotid ultrasound for the evaluation of carotid stenosis. Informed consent has obtained from all the patients. We examined the patients’ medical records to identify stroke risk factors, such as age, sex, stroke subtype, hypertension, diabetes mellitus, hyperlipidemia, atrial fibrillation, current smoking and drinking habits, and a previous history of stroke and cancer. We also assessed their medical data to collect information on stroke subtype, and outcome at discharge. The stroke subtype was classified using the TOAST criteria¹⁰. Furthermore, SVO_LAA was defined as small vessel occlusion (SVO) or large artery atherosclerosis (LAA) because these are the stroke subtypes related to atherosclerosis. We also took note of National Institute of Health and Stroke Scale (NIHSS) score, standard scale indicating stroke neurological severity, and modified Rankin Scale (mRS) expressing stroke outcome. The outcome at discharge was evaluated by mRS; a score of 0–1 was considered favorable because the patients’ condition was not severe. This study was approved by the ethics committee of Fukuoka University Hospital (IRB no.: H20-08-008). We adhere to the observational cohort guideline of STROBE. All methods were performed in accordance with the relevant guidelines and regulations.

We extracted data from carotid ultrasound for all of the patients. We collected the maximum intima–media thickness of each side of the common carotid artery and calculated their average (average value of the maximum common carotid intima–media thickness), and the presence of plaque and stenosis ≥ 50% on the carotid bifurcation on either side. An atheromatous plaque was defined as a lesion with an intima–media thickness ≥ 1.1 mm.

Gut microbial measurements

A brush-type stool collection kit (TechnoSuruga Laboratory Co., Ltd., Shizuoka, Japan) containing preservation fluid was used to obtain and stably maintain the microbiota in feces, and was preserved in the refrigerator for up to 3 weeks. The kits were then sent to Bioengineering Lab. Co., Ltd. (Sagamihara, Japan) to extract and analyze DNA from stool fluid using the bead-beating method, and to perform 16 S rRNA gene sequencing. DNA extracts were used as templates to amplify the V3–V4 region of each 16 S rRNA gene with the primer pair 341 F/805R using the two-step tailing polymerase chain reaction method. The quality and concentration of the libraries were evaluated using a Fragment Analyzer (Advanced Analytical Technologies). Paired-end sequencing of polymerase chain reaction amplicons was performed on the Illumina MiSeq platform (Illumina Inc., San Diego, CA) at 2 × 300 base pairs.

16 S rRNA sequencing and analysis

Sequencing data were processed using QIIME 2 (ver. 2020.8)¹¹ with the DADA2¹² plugin for denoising, chimera removal, and amplicon sequence variant (ASV) inference. Taxonomic classification was performed using the feature-classifier plugin with the Greengenes (ver. 13_8)¹³ 97% operational taxonomic unit (OUT) reference database, which was applied solely for taxonomy assignment without OTU clustering.

Statistical analysis.

The patients were divided into three groups as follows: patients with a rate of Bacteroides ≥ 30% (enterotype I group); those with a rate of Prevotella ≥ 15% (enterotype II); and those with genera other than enterotype I or II (enterotype III group).

The differences between the patients’ characteristics were statistically analyzed between the three groups. We also analyzed the differences in the patients’ stroke subtype, SVO_LAA, and the TOAST classification between the three groups. If the patients’ characteristics were significantly different among the three groups, age- and sex-adjusted analyses were also conducted. Logistic regression model was used for the adjustment of age and sex. All values of p < 0.05 were considered to denote a significant difference while any variable with p < 0.1 was regarded as a trend toward being significant.

Data availability

All data were analyzed using SPSS v28.0 software (IBM Corp., Armonk, NY).

Results

There were 33 patients in the enterotype I group, 10 in the enterotype II group, and 57 in the enterotype III group. There was no overlap between those who were classified as groups I and II. The median time from stroke onset to the collection of stool samples was 3 days. The median NIHSS score was 2 (Supplemental Table 1). The percentages of hypertension, diabetes mellitus, dyslipidemia, atrial fibrillation, and smoking and drinking habits were 80%, 38%, 47%, 24%, 22%, and 24%, respectively. Regarding the stroke subtype, the percentages of SVO, LAA, and cardioembolic stroke were 17%, 33%, and 22%, respectively, while that of other stroke etiologies and undetermined causes was 28%. Regarding the gut microbiota of the patients, the mean percentages of Bacteroides and Prevotella were 24.7% ± 11.2% and 4.0% ± 8.9%, respectively.

The patients’ background and atherosclerotic risk in the three groups are shown in Table 1. There was a significant difference in the frequency of dyslipidemia (P = 0.028) between the three enterotype groups. The rate of diabetes mellitus appeared to be lower in the enterotype II group than in the other groups, but this was not significant. The NIHSS score at admission was significantly higher in the enterotype III group than in the other groups (P = 0.025). The modified Rankin Scale score at discharge tended to be worse in the enterotype III group than in the other groups (P = 0.094). The degree of atherosclerotic changes evaluated by carotid ultrasound was similar in the three groups.

Table 1.

Patients’ characteristics depending on the enterotype.

	EnterotypeI	EnterotypeII	EnterotypeIII	p
	N = 33	N = 10	N = 57
Age	70.6 ± 12.3	76.3 ± 11.5	74.5 ± 10.6	0.21
Gender (female)	12(36.4%)	4 (40%)	22 (38.6%)	0.97
BMI	21.7 ± 3.1	21.9 ± 3.7	22.6 ± 2.7	0.37
Hypertension	25 (75.8%)	9 (90%)	46 (80.7%)	0.6
Diabetes mellitus	13 (39.4%)	1 (10%)	24 (42.1%)	0.15
Dyslipidemia	12 (36.4%)	2 (20%)	33 (57.9%)	0.028
Smoking habit	7 (21.2%)	2 (20%)	7 (12.3%)	0.5
drinking	6 (18.2%)	3 (30%)	13 (22.8%)	0.71
Atrial fibrillation	6 (18.2%)	3 (30%)	15 (26.3%)	0.61
Ischemic heart disease	4 (12.1%)	1 (10%)	7 (12.3%)	0.98
History of stroke	5 (15.2%)	1 (10%)	18 (31.6%)	0.12
History of cancer	3 (9.1%)	0 (0%)	3 (5.3%)	0.206
SVO_LAA	12 (36.4%)	6 (60%)	32 (56.1%)	0.16
NIHSS at admission	1 (0–5)	1 (1–2)	4 (1–6)	0.025
mRS at discharge	1 (1–3)	1 (0–2)	2 (1–3)	0.094
IMT-C10	1.36 ± 0.86	1.34 ± 0.64	1.26 ± 0.53	0.78
maximum of max IMT	2.43 ± 1.13	2.44 ± 1.20	2.41 ± 1.18	0.99
presence of plaque	29 (87.9%)	10 (100%)	55 (96.5%)	0.18
area stenosis ≥ 50%	14 (42.4%)	5 (50%)	18 (31.6%)	0.4

Enterotype I: group of patients with a rate of Bacteroides ≥ 30%; enterotype II: group of patients with a rate of Prevotella ≥ 15%; enterotype III: group of patients with a genus other than that in enterotype I or II.

Age- and sex-adjusted analyses are shown in Table 2. Using the enterotype III group as a reference, the odds ratio of diabetes mellitus was lower in the enterotype II group, but this was not significant. However, the odds ratios of dyslipidemia in the enterotype I and II groups were significantly lower than that in the enterotype III group. The odds ratio of the NIHSS score at admission was significantly lower in the enterotype II group than in the enterotype I group, while that of the enterotype I group was also low. The odds ratio of the modified Rankin Scale score at discharge was also significantly lower in the enterotype II group than in the enterotype I group.

Table 2.

Age- and sex-adjusted analyses.

			univariable			Age and gender adjusted
	number of patients	prevalence	OR	95% CI	p value	OR	95% CI	p value
Diabetes mellitus
EnterotypeIII(reference)	24	42.10%	1			1
EnterotypeI	13	39.40%	0.95	0.40–2.26	0.92	0.92	0.38–2.21	0.85
EnterotypeII	1	10%	0.64	0.34–1.21	0.17	0.62	0.32–1.20	0.15
Dyslipidemia
EnterotypeIII(reference)	33	57.90%	1			1
EnterotypeI	12	36.30%	0.38	0.16–0.90	0.029	0.37	0.15–0.88	0.026
EnterotypeII	2	20%	0.4	0.21-077	0.006	0.39	0.20–0.76	0.006
NIHSS ≥ 3
EnterotypeIII(reference)	34	59.60%	1			1
EnterotypeI	12	36.30%	0.38	0.16–0.89	0.026	0.41	0.17–1.02	0.052
EnterotypeII	2	20%	0.38	0.20–0.73	0.004	0.38	0.20–0.73	0.004
mRS ≥ 3
EnterotypeIII(reference)	25	43.80%	1			1
EnterotypeI	13	39.30%	0.77	0.33–1.80	0.55	0.91	0.37–2.24	0.84
EnterotypeII	1	10%	0.53	0.28–1.01	0.052	0.52	0.27–1.01	0.052

Enterotype I: group of patients with a rate of Bacteroides ≥ 30%; enterotype II: group of patients with a rate of Prevotella ≥ 15%; enterotype III: group of patients with a genus other than that in enterotype I or II; OR: odds ratio; CI: confidnce interval; mRS: modified Rankin Scale.

Among the microbes in the enterotype III group, the rates of the classes Lachnospiraceae and Ruminococcus in the phylum Firmicutes were increased (Table 3).

Table 3.

Main gut microbes in enterotype group III.

Class	Order	Family	Genus	Average(%)	S.D.
Bacilli	Lactobacillales			4.23	4.66
	Bacillales			0.04	0.15
Clostridia	Clostridiales			38.97	10.74
		Ruminococcacaeae		14.02	7.95
			Ruminococcous	3.81	4.71
			Faecalibacterium	4.18	3.61
		Lachnospira		16.62	6.96
			Coprococcus	1.18	1.95
			Roseburia	2.10	2.57
			Blautia	2.04	2.10
		Veillonellaceae		3.36	4.11
Erysipelotrichi	Erysipelotrichale			2.41	2.83

Discussion

In this study, we classified patients with stroke into three enterotypes and examined their association with arteriosclerosis risk factors, stroke subtype, and severity of stroke. This study showed that the enterotype was significantly associated with the frequency of dyslipidemia and stroke severity.

The abundance of Bacteroides is decreased in patients with obesity^14,15 or diabetes¹⁶. However, a study showed that Bacteroides and Prevotella may be increased in people who were more adherent to a Mediterranean diet¹. The Mediterranean diet is characterized by a healthier diet with a higher intake of fiber, vegetables, and fruit, and with a lower intake of sugar and red meat. These findings indicate that characteristics and effects on the human body might vary depending on the region and culture. The classification used in the current study is the same as that from Japan, which indicates the usefulness of identifying atherosclerosis in this study. Therefore, Bacteroides might have beneficial roles in the prevention of arteriosclerosis in Japan.

In this study, the rates of Lachnospiraceae and Ruminococcus were increased in the enterotype III group. An increase in Ruminococcus has been reported to be associated with diabetes¹⁷. In contrast, Lachnospiraceae has been reported to be caused by an increase in non-meat meals and is associated with butyric acid¹⁸. The abundance of Firmicutes is increased in the gut flora of Japanese people with diabetes, which is different from the distribution of the gut microbiota in Western countries. There have been no reports on the role of Lachnospiraceae or Ruminococcus in hypertension, which is considered to be the greatest risk for brain infarction. Further studies are required to determine the role of enterotype III on brain infarction.

The precise mechanisms of Bacteroides and Prevotella on the prevention of atherosclerosis remain unknown. However, the abundance of Prevotella was significantly lower in stool samples of Japanese patients with diabetes, while that of Lactobacillus was higher, than in control subjects¹⁹. In the current study, the rate of Prevotella appeared to be decreased with the presence of diabetes and hyperlipidemia. Bacteroides was found to have a beneficial role in glucose metabolism and was less abundant in patients with type II diabetes mellitus²⁰. Another report showed that a lower abundance of Bacteroides and an increased Firmicutes to Bacteroidetes ratio were associated with the incidence of cardiovascular disease²¹. Bacteroides and Prevotella might contribute to the prevention of atherosclerosis. Therefore, further research is necessary to identify the role of enterotypes in the progression and suppression of inflammation of vessel walls.

The composition of the gut microbiota can be remarkably diverse, and there is no clear definition of a healthy gut microbiota. Inflammation of the vessel walls may lead to the progression of atherosclerosis. Therefore, clarifying whether the progression of arteriosclerosis can be suppressed by regulating the gut microbiota, such as increasing Bacteroides or Prevotella, is necessary.

There are several limitations to this study. First, the number of patients in this study was small. Second, this study analyzed enterotypes and did not address each species, which might be a limitation because the function of the gut microbiota could vary from species to species. Third, we did not collect patients’ dietary data in this study, which could be a limitation. Fourth, it can not be said with certainty that the increases of Bacteroides and Prevotella contribute to the progression of atherosclerosis, because the cross-sectional study does not indicate causal relationship. Furthermore, we referred to the methodology in the paper of Yamashita et al. Although that and ours shared a common in that the patients in both had atherosclerotic disease, the beta-diversity measures are popular method of enterotype classification gut microbiota.

In conclusion, enterotypes may affect the risk factors and severity of ischemic stroke. An increase in abundance of Bacteroides and Prevotella could play a beneficial role in the prevention of atherosclerosis.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

NIHSS

National Institute of Health and Stroke Scale

TOAST

Trials of Org 10172 in Acute Stroke Treatment

SVO

small vessel occlusion

LAA

large artery atherosclerosis

Author contributions

T.O. drafted the manuscript, contributed to the design and conception of the study, acquired data and obtained funding. H.A and M.K performed statistical analysis and obtained funding. Y.B. supervised the study.

Funding

This work was partly supported by the Taiju Life Social Welfare Foundation and the Japan Society for the Promotion of Science (KAKENHI, Grant Nos. 20K10544, 5H04773, 18K1971, 22K19660, and 19K10654).

Data availability

The datasets generated and analyzed within the current study are not publicly available as their publication would be in violation of the Act on the Protection of Personal Information of the Japanese government but are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.