*Article* **Severity by National Institute of Health Stroke Scale Score and Clinical Features of Stroke Patients with Patent Foramen Ovale Stroke and Atrial Fibrillation**

**Kaito Abe 1,2, Fumiya Hasegawa 1,2, Ryota Nakajima 1, Hidetoshi Fukui 1, Moto Shimada 1,2, Takahiro Miyazaki 1,2, Hiroshi Doi 1,2, Goro Endo 1, Kaori Kanbara 1,2, Yasuyuki Mochida 1,2, Jun Okuda 1,2, Nobuya Maeda 3, Akira Isoshima 4, Koichi Tamura <sup>2</sup> and Tomoaki Ishigami 2,\***


**Abstract:** The comparative severity of patent foramen ovale (PFO)-related stroke in patients without atrial fibrillation (AF) and AF-related stroke in patients without PFO is unknown. Therefore, we compared the severity of PFO-related stroke and AF-related stroke. Twenty-six patients who underwent transesophageal echocardiography (TEE) were diagnosed with cardioembolic stroke from July 2018 to March 2020. Cases with AF detected by electrocardiograms or thrombus in the left atrium or left atrial appendage on TEE were included in the AF-related stroke group. Cases with a positive microbubble test on the Valsalva maneuver during TEE, and with no other factors that could cause stroke, were included in the PFO-related stroke group. This study was designed as a single-center, small population pilot study. The stroke severity of the two groups by the National Institute of Health Stroke Scale (NIHSS) score was compared by statistical analysis. Of the 26 cases, five PFO-related stroke patients and 21 AF-related stroke patients were analyzed. The NIHSS score was 2.2 ± 2.8 and 11.5 ± 9.2 (*p*-value < 0.01), the rate of hypertension was 20.0% and 85.7% (*p*-value = 0.01), and the HbA1c value was 5.5 ± 0.2% and 6.3 ± 1.3% (*p*-value = 0.02) in the PFO-related and AF-related stroke groups, respectively. Compared with AF-related stroke patients, stroke severity was low in PFO-related stroke patients.

**Keywords:** patent foramen ovale and stroke; atrial fibrillation and stroke; cryptogenic stroke; severity of stroke; National Institute of Health Stroke Scale score

#### **1. Introduction**

Stroke results in substantial disability and sometimes causes death [1]. The TOAST classification denotes five subtypes of ischemic stroke: (1) large-artery atherosclerosis, (2) cardioembolism, (3) small-vessel occlusion, (4) stroke of other determined etiology, and (5) stroke of undetermined etiology [2]. Cardioembolic stroke accounts for 15–30% of ischemic strokes [3].

In up to 40% of patients with acute ischemic stroke, there is a stroke of undetermined etiology in TOAST classification (5); this stroke has been labeled as cryptogenic

**Citation:** Abe, K.; Hasegawa, F.; Nakajima, R.; Fukui, H.; Shimada, M.; Miyazaki, T.; Doi, H.; Endo, G.; Kanbara, K.; Mochida, Y.; et al. Severity by National Institute of Health Stroke Scale Score and Clinical Features of Stroke Patients with Patent Foramen Ovale Stroke and Atrial Fibrillation. *J. Clin. Med.* **2021**, *10*, 332. https://doi.org/10.3390/ jcm10020332

Received: 30 December 2020 Accepted: 14 January 2021 Published: 18 January 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

stroke [1,2,4,5]. Major cardioembolic risk sources include atrial fibrillation (AF), recent myocardial infarction, previous myocardial infarction (left ventricular aneurysm), intracardiac thrombus, tumors, rheumatic valve disease, aortic arch atheromatous plaques, endocarditis, and mechanical valve prosthesis, whereas minor or unclear risk sources include patent foramen ovale (PFO), atrial septal aneurysm (ASA), and giant Lambl's excrescences [3]. Evaluation of stroke sources is important for preventing second stroke events.

In the general population, 0.4–1% have AF, and the prevalence increases to 9% in the population aged 80 years or older [6]. The CHADS2 and CHA2DS2-VASc risk scores show the frequent occurrence of stroke and embolism, ranging from 0 (low risk) to 18% event/year (high-risk) among patients with AF [7,8].

For the management of AF, anticoagulant therapy, catheter ablation, and antiarrhythmic drugs are well-established [9]. Recently, transcatheter left atrial appendage closure has been used as a primary therapy for AF patients with contraindications for using chronic oral anticoagulation to prevent stroke [10].

PFO is caused by incomplete fusion of the septum primum and secundum after birth in the cranial portion of the fossa ovalis, and is a common anatomical variant found in about 25% of the general population [11,12]. Stroke with PFO occurs when a systemic venous thrombus travels directly into the systemic arterial circulation [1]. The proportion of stroke patients with PFO is 21–63% [11]. According to a report, cryptogenic stroke patients with PFO were younger and less likely to have conventional vascular risk factors than cryptogenic stroke patients without PFO [11].

Recently, the DEFENSE, REDUCE, and CLOSE trials demonstrated the superiority of PFO closure over medical management [13–15]. In cryptogenic stroke, detection of PFO is important to select an adequate secondary stroke prevention therapy. Transesophageal echocardiography (TEE) is the gold standard for PFO detection. The microbubble test with Valsalva maneuver is recommended for detecting PFO on TEE to avoid the increasing false negative rate of up to 20% when the Valsalva maneuver is not performed [3,16].

It is known that the severity of ischemic stroke patients with PFO, including patients with AF, is lower than that of patients without PFO, including patients with AF [17]; however, whether the severity of PFO-related stroke in patients without AF is lower than that of AF-related stroke in patients without PFO is unknown. Thus, the purpose of this analysis was to evaluate the severity of PFO-related stroke and AF-related stroke, and to identify the characteristics of both stroke types.

#### **2. Materials and Methods**

#### *2.1. Study Design and Patient Population*

We performed a single-center (Omori Red Cross Hospital) retrospective study on consecutive patients with cardioembolic stroke, including suspected cases on magnetic resonance imaging, who underwent TEE between July 2018 and March 2020.

Patients with AF diagnosed from history, electrocardiogram (ECG) at admission, 24 h-holter ECG monitoring, ECG monitoring in the ward, or patients with thrombi including smoke-like echo with a swirling motion of blood in the left atrium (LA) or left atrial appendage (LAA), which is known to be a marker of a prothrombotic state, were classified as AF-related stroke patients [18]. Patients with PFO without AF were classified as PFOrelated stroke patients.

Patients with mobile aortic plaque were defined as Class V in the Katz Index [19], patients after valve replacement and those with cardiac tumor, infectious endocarditis, Lambl's excrescence on the aortic valve, or those diagnosed with atherosclerotic stroke by neurologists after TEE were excluded from this study. Patients with PFO between the right atrium (RA) and LA diagnosed only by the color Doppler method without passage of microbubbles were excluded from this study.

#### *2.2. Evaluations*

The diagnosis of ischemic stroke was made by neurologists with known experience in cerebrovascular diseases. TEE was performed and evaluated by cardiologists who were well-experienced in echocardiology. TEE was performed with either Vivid E95 (GE Healthcare, Tokyo, Japan) or ALOKA Prosoundα10 (ALOKA, Tokyo, Japan).

#### *2.3. Baseline Study Assessment*

We collected data on patient characteristics (age, sex, height, body weight, and smoking habit); vascular risk factors; the administration ratio of antiplatelet therapy or oral anticoagulant therapy before stroke onset; blood tests (aspartate transaminase (AST), alanine aminotransferase (ALT), lactate dehydrogenase (LDH), alkaline phosphatase (ALP), total bilirubin (T-bil), brain natriuretic peptide (BNP), hemoglobin, HbA1c, D-dimer, low-density lipoprotein (LDL), high-density lipoprotein (HDL), triglycerides (TGs), creatinine (Cre), estimated glomerular filtration rate (eGFR), and creatinine clearance (CCR) (Cockcroft–Gault equation)); HAS-BLED score to assess bleeding risk in AF patients from hypertension, abnormal renal/liver function, stroke, bleeding history or predisposition, labile INR, elderly (>65 years old), and concomitant drugs/alcohol use [20]; National Institute of Health Stroke Scale (NIHSS) [21]; the risk of paradoxical embolism (RoPE) score to assess the likelihood of the cryptogenic stroke being related to PFO based on the scoring items of age, hypertension, diabetes, history of stroke or transient ischemic attack, smoking habit, and cortical infarct on imaging [11]; ECG; 24 h-holter ECG monitoring; ECG monitoring in the hospital ward; thrombus, including smoke-like echo in LA or LAA by TEE; and left ventricular ejection fraction (LVEF) by transthoracic echocardiography.

The TEE was performed under light sedation with propofol. The LA or LAA thrombi were evaluated in all patients by TEE. For all patients who underwent TEE, an intravenous microbubble test during the Valsalva maneuver was performed. In the GORE-REDUCE trial [14], the classification of PFO size was based on the maximum number of microbubbles during the first three cardiac cycles; 0 microbubbles were classified as no shunt, one to five microbubbles as small, six to 25 microbubbles as moderate, and more than 25 microbubbles as large. PFO was diagnosed by the microbubble test using the Valsalva maneuver technique between the RA and LA. Complex PFO was classified as PFO with ASA, or with a long tunnel length of over 8 mm, or with the eustachian valve [22]. Some AF-related stroke patients with LA or LAA thrombi skipped the microbubble test.

#### *2.4. Statistical Analysis*

JMP Pro version 15 software (SAS Institute Japan Inc., Tokyo, Japan) was used for statistical analysis. Data are expressed as mean ± standard deviation for continuous variables and as frequencies and percentages for categorical variables. Baseline characteristics were compared using the Student's *t* test or Welch's *t* test for continuous variables and Fisher's exact test for categorical variables. A *p*-value < 0.05 was considered statistically significant.

#### **3. Results**

#### *3.1. Stroke Classification*

A total of 82 patients were enrolled from July 2018 to March 2020. AF-related stroke was noted in 21 (25.6%) patients, atherosclerotic stroke in 21 (25.6%), PFO-related stroke in five (6.1%), cardioembolic stroke without AF in 13 (15.9%), others (systemic lupus erythematosus, hyperemia, vasculitis, lacunar infarction) in four (4.9%), and cryptogenic stroke in 18 (22.0%) (Figure 1). Cardioembolic stroke without AF included patients with post-valve replacement (four patients), cardiac tumor (two patients), infectious endocarditis (one patient), Lambl's excrescence on the aortic valve (two patients), old myocardial infarction (one patient), and patients with PFO between the RA and LA diagnosed only by the color Doppler method without passage of microbubbles (three patients).

**Figure 1.** Stroke classification of this study.

#### *3.2. Study Population and Patient Features*

Among the 21 AF-related stroke patients, 20 (95.2%) patients had AF and 10 (47.6%) had LA or LAA thrombus, whereas the PFO-related stroke patients had no AF or thrombus in the LA or LAA. The ratio of comorbidity with hypertension in AF-related stroke patients was higher than that in PFO-related stroke patients (85.7%, 20.0%, *p*-value = 0.01). The NIHSS score in AF-related stroke patients was more severe than that in PFO-related stroke patients (11.5 ± 9.2, 2.2 ± 2.8, *p*-value < 0.01). Age, height, body weight, LVEF, and HAS-BLED score were not significantly different between the groups. The comorbidity ratio of dyslipidemia, diabetes, old myocardial infarction, or past stroke history had no significant differences between the two groups. The administration ratio of antiplatelet therapy or oral anticoagulant therapy before stroke onset had no significant differences between the two groups (Table 1).

**Table 1.** Baseline patient characteristics of PFO-related stoke and AF-related stroke.



**Table 1.** *Cont.*

Atrial fibrillation (AF), direct oral anticoagulants (DOAC), left atrium (LA), left atrial appendage (LAA), left ventricular ejection fraction (LVEF), National Institute of Health Stroke Scale (NIHSS), old myocardial infarction (OMI), patent foramen ovale (PFO), years old (y.o); \* Welch's *t* test, † Fisher's exact test.

HbA1c values in AF-related stroke patients were higher than that in PFO-related stroke patients (6.3% ± 1.3%, 5.5% ± 0.2%, *p*-value = 0.02). BNP, D-dimer, hemoglobin, LDL, HDL, TG, Cre, eGFR, CCR, AST, ALT, T-bil, and ALP were not significantly different between the groups (Table 2).

**Table 2.** Baseline patient blood test parameters of PFO-related stoke and AF-related stroke.


Atrial fibrillation (AF), alkaline phosphatase (ALP), alanine aminotransferase (ALT), aspartate transaminase (AST), brain natriuretic peptide (BNP), creatinine clearance (CCR) (Cockcroft–Gault equation), creatinine (Cre), estimated glomerular filtration rate (eGFR), hemoglobin (Hb), high-density lipoprotein (HDL), lactate dehydrogenase (LDH), low-density lipoprotein (LDL), patent foramen ovale (PFO), total bilirubin (T-bil), triglycerides (TG); \* Welch's *t* test.

#### *3.3. PFO Characteristics of This Study*

Among the five PFO-related stroke patients, one patient had a small shunt PFO, two patients had moderate shunt PFO, and two patients had a large shunt PFO. Of the five PFO-related stroke patients, four patients had complex PFO; all complex PFOs had a long PFO tunnel, and had no eustachian valve or ASA. The RoPE score was 4.8 ± 2.4, the average PFO attributable fraction was 36.0 ± 37.3%, and the estimated two years stroke recurrence rate was 12.2 ± 7.2% in this PFO population by RoPE score to assess the likelihood of the cryptogenic stroke being related to PFO (Table 3) [11].


**Table 3.** PFO characteristics, PFO attributable fraction, and estimated two years stroke recurrence rate by RoPE score.

The classification of PFO size: 0 microbubbles were classified as no shunt, one to five microbubbles as small, six to 25 microbubbles as moderate, and more than 25 microbubbles as large.; patent foramen ovale (PFO), risk of paradoxical embolism (RoPE).

#### **4. Discussion**

The novelty of this study is that it is a PFO group diagnosed by TEE with the highly diagnostic Valsalva maneuver technique, and that it includes the severity of stroke between PFO-related stroke in patients without AF and AF-related stroke in patients without PFO.

Severity comparison by the NIHSS showed that AF-related stroke was more severe than PFO-related stroke. Previous studies [17] also reported that PFO-related stroke was less severe than strokes in the other groups; it was expected that AF-related stroke would be more severe, which was also revealed in our study. Regarding administration of oral medication before stroke onset, antiplatelet therapy was administered to two (40%) patients in the PFO-related stroke group. In the AF-related stroke group, anticoagulation therapy was administered to four (19.0%) patients and antiplatelet therapy to seven (33.3%). It may be necessary to consider the effect of these pre-medications on the severity of stroke.

Generally, it is said that PFO-related stroke patients are younger and have less cardiovascular risk; this study also showed that the ratio of comorbidity of hypertension and high HbA1c in blood tests was significantly lower in the PFO-related stroke patients compared to AF-related stroke patients. Although there was no significant difference in other factors, the age, ratio of comorbidity of dyslipidemia, diabetes, and ratio of smoking habits tended to be low in the PFO-related stroke group, which was similar to the existing report [11]. In particular, the average age of the PFO-related stroke patients was 58.2 years, which is younger than that of the AF-related stroke patients (77.9 years), and it is necessary to consider that the fact that there are few cardiovascular risk factors may also contribute to the low severity of stroke in the PFO-related stroke patients. For younger patients, functional improvement after stroke can be expected, and observation of long prognosis is also important.

It has been reported that the probability of stroke due to PFO is 88%, and the recurrence rate after two years is 2% in the group with the highest RoPE Score [11]. The PFO attributable stroke rate in this study is not high at 36.0%, but the average recurrence probability after two years by the RoPE score is estimated to be high at 12.2%. Therefore, secondary prevention therapies, such as antiplatelet therapy, anticoagulant therapy, or a PFO occluder device, seemed to be important in decreasing the recurrence rate of strokes.

This study has some limitations. This study is a pilot study not registered in clinical trials with an international clinical trials register. As this was a retrospective analysis at a single institution, it is necessary to consider the influence of the small population, five PFO-related stroke patients and 21 AF-related stroke patients. Statistical results may not be sufficient, as power analysis has not been performed in this study. It is also necessary to consider the influence of population bias, such as pre-test probability, because it is a group of cases in which neurologists suspected cardioembolic stroke and who needed TEE. In addition, there is a possibility that latent AF and PFO may be involved in cases of stroke other than cardioembolic stroke diagnosed by neurologists. Since patient prognosis was not observed in this study, it is necessary to observe their prognosis after medical interventions. We need to be careful in interpreting the results of this study.

There were 18 (22.0%) cases in the stroke group with unknown causes in the analysis target, and four of them were suspected of being cardioembolic stroke by the neurologist, but definitive clinical findings were unclear. It seems necessary to evaluate the comorbidity of AF with an implantable electrocardiograph.

In this analysis, TEE with the Valsalva maneuver technique was performed for PFO detection; therefore, PFO detection in this study was highly credible. However, it is said that there are detection limits of this method; hence, it seems that there is a possibility of wrongly classifying other strokes into the cryptogenic stroke group.

In addition, even in the group diagnosed with cryptogenic stroke, the detection rate of thrombus in LA/LAA may have decreased, because at the time of TEE, antiplatelet therapy or anticoagulant therapy had already been performed. In this study, after stroke, anticoagulant therapy was administered to three patients, and antiplatelet therapy was administered to one patient in the PFO-related stroke group. In the AF-related stroke group, anticoagulant therapy was administered to all patients. Of these, anticoagulation therapy alone was administered to 19 (90.5%) patients, and a combination of antiplatelet therapy and anticoagulant therapy was administered to two (9.5%) patients. Therefore, it is necessary to observe the secondary preventive effect on the recurrence rate in the future.

#### **5. Conclusions**

Compared with AF-related stroke patients, stroke severity, the comorbidity of hypertension rate, and the value of HbA1c were low in PFO-related stroke patients.

**Author Contributions:** Conceptualization, K.A.; methodology, K.A.; software, K.A. and T.I.; validation, K.A., F.H., R.N., H.F., M.S., T.M., H.D., G.E., K.K., Y.M., J.O., N.M., A.I., K.T. and T.I.; formal analysis, K.A. and T.I.; investigation, K.A. and T.I.; resources, K.A., F.H., R.N., H.F., M.S., T.M., H.D., G.E., K.K., Y.M., J.O., N.M. and A.I.; data curation, K.A. and T.I.; writing—original draft preparation, K.A.; writing—review and editing, K.T. and T.I.; visualization, K.A. and T.I.; supervision, N.M., A.I., K.T. and T.I.; project administration, K.A. and T.I.; All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the ethical committee of Omori Red Cross Hospital (protocol code: No.20-5, 26 May 2020).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** We are grateful to departments of neurology, neurosurgery, and cardiology staffs for stroke patients' care.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Review* **The Types and Proportions of Commensal Microbiota Have a Predictive Value in Coronary Heart Disease**

**Lin Chen 1,2, Tomoaki Ishigami 1,\*, Hiroshi Doi 1, Kentaro Arakawa <sup>1</sup> and Kouichi Tamura <sup>1</sup>**


**\*** Correspondence: tommmish@yokohama-cu.ac.jp; Tel.: +81-45-787-2635

**Abstract:** Previous clinical studies have suggested that commensal microbiota play an important role in atherosclerotic cardiovascular disease; however, a synthetic analysis of coronary heart disease (CHD) has yet to be performed. Therefore, we aimed to investigate the specific types of commensal microbiota associated with CHD by performing a systematic review of prospective observational studies that have assessed associations between commensal microbiota and CHD. Of the 544 published articles identified in the initial search, 16 publications with data from 16 cohort studies (2210 patients) were included in the analysis. The combined data showed that Bacteroides and Prevotella were commonly identified among nine articles (*n* = 13) in the fecal samples of patients with CHD, while seven articles commonly identified Firmicutes. Moreover, several types of commensal microbiota were common to atherosclerotic plaque and blood or gut samples in 16 cohort studies. For example, Veillonella, Proteobacteria, and Streptococcus were identified among the plaque and fecal samples, whereas Clostridium was commonly identified among blood and fecal samples of patients with CHD. Collectively, our findings suggest that several types of commensal microbiota are associated with CHD, and their presence may correlate with disease markers of CHD.

**Keywords:** commensal microbiota; coronary heart disease; systematic review

#### **1. Introduction**

In recent years, coronary heart disease (CHD) has remained the leading cause of death worldwide, while statins and other pharmacological agents for coronary secondary prevention have failed to completely protect people against CHD, despite their widespread use [1–3]. Given the unmet need for effective therapies, there has been increasing interest in targeting novel pathways that underlie the pathogenesis of CHD and in establishing a precise system to track its development. Additionally, to observe the progression of the disease, there is an increasingly important clinical value in discovering a predictive biomarker for CHD. Although the molecular mechanisms responsible for the development of CHD are not completely understood, recent studies have highlighted the critical role of commensal microbiota in CHD [4–6], with alterations in the gut microbiota being linked to CHD progression [4]. However, a synthetic analysis of the predictive value of specific types of commensal microbes in CHD patients has not yet been performed. In particular, a cross-site comparison of the types of microbes in the blood, gut, and atherosclerotic plaques of these patient is essential, given the high level of variability observed in the microbiota between different subjects and studies [7–11]. Therefore, in this analysis, we aimed to combine the results from published clinical trials to compare the types, proportions, and sources of commensal microbes in CHD. Our findings revealed several types of commensal microbes common to the atherosclerotic plaques, blood, or gut samples in patients with CHD, and the expression of some specific types of commensal microbes could be used as predictive or disease biomarkers of CHD in the future.

**Citation:** Chen, L.; Ishigami, T.; Doi, H.; Arakawa, K.; Tamura, K. The Types and Proportions of Commensal Microbiota Have a Predictive Value in Coronary Heart Disease. *J. Clin. Med.* **2021**, *10*, 3120. https://doi.org/ 10.3390/jcm10143120

Academic Editors: Maciej Banach, Francesco Giallauria and Massimo Mancone

Received: 10 March 2021 Accepted: 24 June 2021 Published: 15 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

#### **2. Materials and Methods**

#### *2.1. Search Strategy*

We searched several electronic databases (PubMed, Embase, Web of Science, Cochrane Library, and ClinicalTrials.gov) up until 8 March 2020 for prospective, observational clinical studies that have investigated commensal microbes in patients with CHD. We used broad search terms (Additional File 1) describing aspects of 'Gastrointestinal Microbiome' and 'Coronary Disease'. These terms were used in combination with "AND" or "OR". This literature review was performed independently by two investigators, with a third resolving any disputes as needed. The detailed search strategy of PubMed: used: ("Coronary Disease") or (Coronary Diseases) or (Disease, Coronary) or (Diseases, Coronary) or (Coronary Heart Disease) or (Coronary Heart Diseases) or (Disease, Coronary Heart) or (Diseases, Coronary Heart) or (Heart Disease, Coronary) or (Heart Diseases, Coronary) and ("Gastrointestinal Microbiome") or (Gastrointestinal Microbiomes) or (Microbiome, Gastrointestinal) or (Gut Microbiome) or (Gut Microbiomes) or (Microbiome, Gut) or (Gut Microflora) or (Microflora, Gut) or (Gut Microbiota) or (Gut Microbiotas) or (Microbiota, Gut) or (Gastrointestinal Flora) or (Flora, Gastrointestinal) or (Gut Flora) or (Flora, Gut) or (Gastrointestinal Microbiota) or (Gastrointestinal Microbiotas) or (Microbiota, Gastrointestinal) or (Gastrointestinal Microbial Community) or (Gastrointestinal Microbial Communities) or (Microbial Community, Gastrointestinal) or (Gastrointestinal Microflora) or (Microflora, Gastrointestinal) or (Gastric Microbiome) or (Gastric Microbiomes) or (Microbiome, Gastric) or (Intestinal Microbiome) or (Intestinal Microbiomes) or (Microbiome, Intestinal) or (Intestinal Microbiota) or (Intestinal Microbiotas) or (Microbiota, Intestinal) or (Intestinal Microflora) or (Microflora, Intestinal) or (Intestinal Flora) or (Flora, Intestinal) or (Enteric Bacteria) or (Bacteria, Enteric) and (risk\*[Title/Abstract] or risk\*[MeSH:noexp] or risk \*[MeSH:noexp] or cohort studies[MeSH Terms] or group\*[Text Word]). We then used a similar approach with Embase, Web of Science, Cochrane Library, and ClinicalTrials.gov. Following the PICOS (Participants, Interventions, Comparisons, Outcomes and Study design) principle, the key search terms included (P) patients with CHD; (I) detection of the gene of microbiota; (C/O) compare the types of commensal microbiota between the CHD group and the control group; (S) case-control studies or cohort study.

#### *2.2. Study Selection*

Prospective, observational, controlled studies that assessed changes in populations of commensal microbes were included if they conducted baseline measurements in a population with CHD, including those with atherosclerosis or acute coronary syndrome (ACS) or chronic CHD (defined as a history of myocardial infarction (MI), percutaneous coronary intervention (PCI), coronary artery bypass grafting (CABG), or a diagnosis confirmed through coronary angiography). Publications without detailed data were excluded from this study. When multiple publications were based on the same source study, we included the publication that had the larger sample size or more relative data.

#### *2.3. Data Extraction*

The following data from each study were extracted: the first author's name, publication year, country of the conducted study, sample size, population, source of commensal microbes (atherosclerotic plaque or blood or fecal samples), region from which the genetic expression was quantified, and the specific types of commensal microbes demonstrating statistically significant changes in expression, whether increased or decreased, in CHD.

#### *2.4. Data Synthesis and Analysis*

The commensal microbes demonstrating statistically significant changes in expression in fecal, blood, or atherosclerotic plaque samples from the 16 included studies were sorted and classified according to the degree to which the populations were increased or decreased. Subsequently, an overall comparison of the gut, blood, and atherosclerotic plaque microbiota was performed, and any microbiota demonstrating a change among

at least two publications for each of the three sample types was recorded. Furthermore, interstudy comparisons of atherosclerotic plaque, blood, and gut microbiota were also performed, and the microbiota that were common to at least two body sites were recorded.

#### **3. Results**

#### *3.1. Study Selection*

The search process and study selection (presented in Figure 1) identified 544 records of interest. Among these, 163 were excluded for being repetitive, and 353 additional articles were excluded from the analysis because they were review articles, published protocols, lab studies, animal studies, or articles deemed not to be of relevance based on their titles and abstracts. The full texts of the 28 remaining articles were obtained. Several studies were subsequently excluded because they did not meet the predefined inclusion criteria, including those with no relevant outcome data (three articles) and those reporting on unrelated topics (two articles). Four studies were excluded due to insufficient information pertaining to the inclusion criteria, and a second article by the same authors was found to be a repetitive report based on a partial dataset. In total, 16 publications, reporting on 16 cohort studies, were selected for inclusion in the analysis [11–26].

**Figure 1.** Flow diagram of the study selection.

#### *3.2. Characteristics of Included Studies*

Ultimately, 16 cohort studies, published from 2011 to 2020, were included in the analysis. These studies reported on the characterization of commensal microbe profiles in CHD patients and analyzed changes in commensal microbe populations in fecal, blood, or atherosclerotic plaque samples. These studies are summarized in Table 1.


**Table 1.** Characteristics of included studies.

ACS, acute coronary syndrome; ACVD, atherosclerotic cardiovascular disease; CAD, coronary artery disease; CHD, coronary heart disease; MI, myocardial infarction.

> In total, this systematic review included data collected from 2210 participants. Three of the studies were conducted in the United States of America, three in Japan, one in Finland, one in France, one in Italy, and seven in China. Three studies reported on patients with ACS, 12 included patients with chronic CHD, and one included patients with atherosclerosis. Thirteen studies analyzed changes in commensal microbes in fecal samples, two of which also analyzed changes in plaque samples, one study reported on changes in the microbiota of plaque alone, while two reported on changes in blood alone (Table 1).

#### *3.3. Overall Comparison of the Gut, Blood, and Atherosclerotic Plaque Microbiota*

We surveyed changes in the atherosclerotic plaque, blood, and gut (feces) bacterial communities to look for commonalities among the studies. The microbiota commonly identified by three publications in patients' plaque samples were classified as Streptococcus, whereas the microbiota commonly identified among nine papers in patient fecal samples were classified as Bacteroides and Prevotela, regardless of the direction of change (increased or decreased) (Tables 2 and 3). After taking the trend of change into consideration, we found that Streptococcus was increased in five studies, whereas Lachnospiraceae was decreased in four studies in the fecal samples of patients (Table 4). Unfortunately, we did not observe any similarities in microbiota between two papers in patient blood samples, which may have been due to only two contributing to the comparisons of bacterial populations in the blood (Table 5). Collectively, these results support the notion that several common types of commensal microbiota that exist in the atherosclerotic plaque, blood, and gut (feces) are associated with CHD.

**Table 2.** Microbiota commonly identified by at least two publications in the atherosclerotic plaque samples of those with coronary heart disease.


The column on the right indicates the number of publications for which the microbiota were identified in plaque samples.



The column on the right indicates the number of publications for which the microbiota were identified in fecal samples.



The columns in the middle and on the right indicate the direction of change in the expression of microbiota populations in the feces of patients with coronary heart disease (CHD), and the number of publications reporting the change.

**Table 5.** Microbiota commonly identified in blood samples.


The columns at the right indicate the microbiota that were found in blood samples from two papers.

#### *3.4. Comparisons of Atherosclerotic Plaque, Blood, and Gut Microbiotas between Studies*

One of the main purposes of this study was to search for microbial communities that were commonly observed between gut and atherosclerotic plaque samples, or between gut and blood samples of those with CHD. Although all three sites typically express distinct microbial communities, the comparison of the overall bacterial community compositions revealed commonalities in the microbiota between the atherosclerotic plaque and fecal samples, and between the blood and fecal samples. However, no specific microbial communities were present in all three sample types. Table 6 summarizes the microbial communities that were identified in fecal samples and at least one other sample type in at least two studies. Clostridiales populations were observed in both the blood and fecal

samples of patients, whereas Veillonella, Proteobacteria, and Streptococcus communities were identified in atherosclerotic plaque and gut samples in at least two studies. Interestingly, only Streptococcus was reported to have increased in all eight studies. Collectively, these results indicate that several specific types of commensal microbiota coexist in the atherosclerotic plaque and gut or in the blood and gut of patients with CHD.

**Table 6.** Microbiota populations common to at least two body sites.


The columns in the middle and on the right indicate the types of microbiota that were found in both the plaque and fecal samples or in the blood and fecal samples of those with coronary heart disease.

#### **4. Discussion**

In this study, we compared the bacterial compositions of the microbiota of the blood, gut, and atherosclerotic plaque from 16 relevant studies of patients with CHD. This approach allowed us to generate a relatively comprehensive description of the microbial communities associated with CHD. This comparison of blood, gut, and atherosclerotic plaque samples was necessary to identify members of the normal microbiome that may translocate from one body habitat to another where they may contribute to disease. We specifically identified several common types of commensal microbiota that existed or coexisted in atherosclerotic plaques, blood, and gut that were associated with CHD. However, we did not observe any similarities between the microbiota populations in the blood and plaque, or between the blood, plaque, and gut of the patients assessed by the 16 included studies, which may have been due to the paucity of relevant studies meeting the inclusion criteria. Our findings suggest that specific types of commensal microbiota, such as Streptococcus, Lachnospiraceae, and Clostridiales, may have a stronger predictive value in CHD. Moreover, the atherosclerotic plaque and blood microbiota may, at least in part, be derived from those present in the gut.

In this study, we found that the expression of Streptococcus was increased in the gut but was also present in atherosclerotic plaque samples; this may represent a previously unappreciated core member of the atherosclerotic plaque communities. This organism is known to be implicated in endocarditis [27]; however, the role of Streptococcus in CHD has not yet been reported. Its increased expression in the gut and its presence in atherosclerotic plaques suggests that it may directly affect the pathogenesis of atherosclerosis. Recently, some studies have revealed that the gut microbiome directly affects immune responses that regulate chronic inflammatory diseases, such as atherosclerosis [4,5], and it is becoming clear that microbiota-derived bioactive compounds can signal to distant organs, contributing to the development of cardiovascular disease states [28]. In addition, the molecular mechanism involving the "molecular mimicry" of microbial antigens has also been found to be associated with atherosclerosis [29]. For example, Binder et al. showed that pneumococcal vaccination decreases the formation of atherosclerotic lesions through a molecular mimicry mechanism between Streptococcus pneumoniae and oxidized lowdensity lipoprotein (LDL) [30], while our previous clinical research revealed that many autoantibodies that differ from those found in chronic autoimmune diseases are associated with atherosclerosis [31]. In parallel with these findings, a recent study reported that autoantibodies produced by B lymphocytes are present in plaques and may cross-react with the outer membrane proteins of bacteria, as well as with a cytoskeletal protein involved in atherogenesis [32]. Moreover, Saita et al. also demonstrated that B cells present in both the coronary and carotid plaques of patients with cardiovascular diseases locally produce antibodies that are capable of reacting in response to antigens of the gut microbiota and that they may cross-react with self-antigens. Furthermore, immunoglobulin G1 (IgG1) is

secreted in human coronary atherosclerotic lesions and recognizes the outer membrane proteins of Enterobacteriaceae [33]. These findings demonstrate that in human atherosclerotic plaques, a local cross-reactive immune response may occur, wherein antibodies cross-react with a bacterial antigen and a self-protein. In addition, antibodies and B lymphocytes could play an important role in these disease processes [31,32].

Besides being present in atherosclerotic plaques, Streptococcus expression was also observed to be simultaneously increased in the gut. Recent studies have found that low levels of microbiota can also enter the bloodstream to systemically induce chronic, lowgrade inflammation [3,34]. Generally, the intestinal mucosal barrier plays a critical role in preventing the translocation of bacterial components. This barrier is efficient when the microbiome is complex and stable; however, under certain conditions, such as those induced by diets high in fat and cholesterol or in certain diseases, major alterations to the composition of the host microbiota can occur, which have in turn been associated with increased intestinal permeability [35–38]. When the intestinal mucosal barrier becomes compromised, commensal microbes or commensal microbe-derived molecules can readily enter the bloodstream and exert systemic effects, which include the induction of infection or chronic low-grade inflammation and immunoreactivity, affecting multiple immune cell populations; this phenomenon has been found to be prevalent in atherosclerosis [39]. However, the presence of Streptococcus in the blood was not identified in at least two of the relevant studies included in our analysis. Recently, another mechanism has been identified through which bacteria could reach the atherosclerotic plaque, which involves phagocytosis by macrophages at epithelial linings (e.g., of the gut and lung). Upon phagocytosis, macrophages become activated; once they reach the activated endothelium of the atheroma, they leave the bloodstream to enter the atheroma and transform into cholesterol-laden foam cells [40]. In support of this mechanism, patients with cardiovascular disease exhibit a two-fold increase in the number of C. pneumonia-infected peripheral blood mononuclear cells compared with that of controls [11]. Furthermore, the bacteria have been shown to only be present in atheromas and not in healthy aortic tissues in mice [41], and they have been identified in human atherosclerotic plaques [42]. Thus, infected macrophages may specifically target bacteria in atheromas. It remains a possibility that Streptococcus can reach atherosclerotic plaques via the systemic circulation and directly promote local inflammatory cascades or elicit a specific immune response, such as molecular mimicry, thereby indirectly influencing host metabolism and systemic inflammation; however, the specific mechanisms by which Streptococcus regulates the development of atherosclerosis remain unknown and require further investigation.

Study Strengths and Limitations

Although some studies investigating the relationship between commensal microbiota and CHD have been conducted previously [6], our synthetic analysis is the first to investigate the types of commensal microbiota that are commonly associated with CHD. Moreover, our findings revealed several specific types of commensal microbiota that commonly exist or coexist in the atherosclerotic plaque, blood, or feces of patients with CHD. This could be valuable knowledge for future studies investigating this association, as it reveals that the microbiota of the atherosclerotic plaque may, at least in part, be derived from the gut and that these specific types of commensal microbiota may have predictive value for CHD.

A limitation of this study was that only 16 relevant publications met the criteria for inclusion, with only three contributing to a comparison of bacterial populations in the atherosclerotic plaque and with only two studies contributing to that in the blood. This limited our ability to generate a precise descriptive summary of the 'real-world' changes in relative commensal microbiota populations. Furthermore, there were several differences in the methodology of these studies, such as the region of genetic quantification, which could have influenced the results or may even have been significant sources of potential inaccuracy. In addition, the types and proportions of commensal microbiota could also have been influenced by several patient characteristics, such as diet (including regional differences) [43]. The addition of new evidence to the field will significantly reduce the effects of these limitations in future analyses.

Further studies should investigate specific types of commensal microbiota, as well as factors that modulate or inhibit their activity. In addition, more information is required to verify the predictive value and mechanisms of commensal microbiota in CHD patients.

#### **5. Conclusions**

In summary, this study revealed key types of bacteria associated with CHD, as well as several types that were simultaneously present in the atherosclerotic plaque, blood, or gut. In addition, the atherosclerotic plaques and blood samples of patients contained numerous bacteria from different phyla. Our findings strongly support the hypothesis that the gut can be a source of atherosclerotic plaque- and blood-associated bacteria. Although our findings are based on the data from a limited number of studies, they clearly suggest that several specific types of commensal microbiota have a predictive value for CHD. More prospective studies are needed to further evaluate this relationship and to identify the mechanisms that drive it.

**Author Contributions:** Conceptualization, L.C. and T.I.; methodology, H.D. and K.A.; data curation, L.C.; writing—original draft preparation, L.C.; writing—review and editing, L.C.; supervision, K.T. and T.I. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the National Natural Science Foundation of China (grant number 81900388). The funding source had no involvement in the study design, in the collection, analysis, and interpretation of data, in the writing of the report, or in the decision to submit the article for publication.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

