Relationship between Streptococcus gallolyticus Infective Endocarditis and Pre-Neoplastic Colorectal Lesions

Department of Cardiology, Hospital Universitari Doctor Josep Trueta, CIBER-CV (Girona, Spain), Spain Dirección Territorial de Radiologia i Medicina Nuclear de Girona. IDI. IDIBGI (Girona, Spain), Spain Department of Internal Medicine, Hospital Universitari Doctor Josep Trueta (Girona, Spain), Spain Cardiovascular Genetics Centre, University of Girona-IDIBGI; Department of Medical Science, School of Medicine, University of Girona, Girona, Spain Centro Investigación Biomédica en Red: Enfermedades Cardiovasculares (CIBERCV), Madrid, Spain


Introduction
Infective Endocarditis (IE) due to gut pathogens presents a well-recognized relationship with Colorectal Cancer (CRC) (McCoy and Mason 3rd, 1951;Klein et al., 1977;Gupta et al., 2010;Darjee and Gibb, 1993). Streptococcus gallolyticus (previously Streptococcus bovis) is the most relevant microorganism of gut microbiota involved in patients with underlying occult CRC (Schlegel et al., 2003;Giannitsioti et al., 2007;Corredoira et al., 2008). Emerging studies revealed that 65% of patients diagnosed with Streptococcus gallolyticus IE had a concomitant CRC (Hensler, 2011;Pasquereau-Kotula et al., 2018) and stablished that patients with Streptococcus gallolyticus bacteriemia had 5-fold more CRC (Amado et al., 2015;Takamura et al., 2014). It is usually located in the colonic wall and it could benefit from the presence of a tumor and use these neoplastic sites as a point of entry to cause systemic infections (Boleij and Tjalsma, 2013). The degree of bowel wall involvement that could allow the entry of gut pathogens and, therefore greater IE risk, is unknown. In the setting of Streptococcus gallolyticus IE, it is recommended to rule out occult CRC with colonoscopy during hospitalization (Schreuders et al., 2015;Habib et al., 2015). However, it is still a debate of which colorectal lesions could become the point of entry to cause IE.
The most frequent pre-neoplastic colorectal lesions are colonic adenomas, slow overgrowths of the colonic mucosa that carry a risk of becoming malignant. They are 89 classified as flat, sessile or pedunculated, being the adenoma size the most important carcinogenic potential marker (Lemaire et al., 2017;Lieberman et al., 2008;Ferlitsch et al., 2011). Colonic adenomas are much more prevalent than CRC lesions, being identified in 40% of asymptomatic population over 60 years old (Strum, 2016). Multiple hypotheses suggest that Streptococcus gallolyticus may play an etiological role in cancer development (Boleij and Tjalsma, 2013). However, whether the association of gut pathogens related to IE with colorectal lesions is a consequence of the gastrointestinal lesion or could promote CRC is not well-known. Additionally, it is not established if lower degree dysplastic colorectal lesions could also be associated with systemic entry of gut microorganisms. Since only 44% of IE have an identified portal of entry, pre-cancerous lesions could be the gateway for Streptococcus gallolyticus in some of these cases (Delahaye et al., 2016;Murdoch et al., 2009;Habib et al., 2019).
The aim of this study was to determine the relationship between pre-neoplastic colorectal lesions (high and lowgrade dysplasia) and Streptococcus gallolyticus IE, evaluating those IE cases in which a colonoscopy was performed in the extension study during admission.

Study Design
An observational retrospective single-center study was conducted through data gathering from electronic medical records covering the period from January 2008 to December 2020. Patients with IE diagnosis, baseline Transthoracic Echocardiogram (TTE) and a colonoscopy performed for the IE extension analysis during the event were included. All patients were diagnosed of IE according modified Duke Criteria (Li et al., 2000) and those with only a suspected diagnosis of the illness without confirmation were excluded. Patients were treated during the hospitalization according to European current guidelines (Habib et al., 2015). The Endocarditis Team of our institution decided the multidisciplinary approach in the diagnostic and therapeutic management of all IE patients included. The ethical committee of our institution approved the study. Finally, from 402 patients diagnosed with IE, 71 (18%) underwent a screening colonoscopy during hospitalization and fulfilled all the inclusion criteria ( Fig. 1). Colorectal cancer 36% (n = 8)

IE Variables
Each patient was analyzed with the same protocol searching data in medical records, starting on the day of hospital admission. Demographic, clinical and microbiological parameters were collected of all cases. Patients underwent a baseline comprehensive TTE with commercially available instruments (Phillips Medical Systems, Bothell, Washington and General Electric Medical Systems, Milwaukee, Wisconsin) as part of a standard clinical diagnostic evaluation. Measurements and recordings were obtained according to current recommendations (Lang et al., 2015). Left ventricular ejection fraction was calculated using the Simpson biplane method. A semiquantitative 4-point scale (with grades of none, mild, moderate and severe) was applied to stratify valvular regurgitation on color 2D Doppler echocardiography clips obtained in multiple standard views. A Transesophageal Echocardiography (TEE) and/or 18F-FDG Positron Emission Tomography/Computed Tomography (18F-FDG PET/CT) were performed according to Endocarditis Team management. Similar data was recorded on follow-up TTE during hospitalization, which was performed as part of clinical follow-up. Data on death and survival were obtained from medical records.

Colorectal Lesions Variables
Data from performed colonoscopies to analyze the origin of IE were collected during admission. Endoscopic findings described in histopathological reports classified the colorectal pre-neoplastic lesions as followed: Tubular adenoma was defined as an adenoma with a diameter of <1 cm and villous adenoma was defined as finger-like epithelial projections or an adenoma diameter of >1 cm. Carcinoma in situ was classified as an adenoma with high-grade dysplasia. High-grade dysplasia is characterized by any one of the following: Cribriform architecture, severe cytological atypia and invasion with a desmoplastic response (invasion confined to the lamina propria) based on the histological classification of World Health Organization (Table 1) (Jass and Sobin, 1989). Those patients without all previous cited criteria were considered adenoma with low-grade dysplasia. The criterion for CRC was the presence of malignant cells beyond the muscularis mucosa. Nonneoplastic colonic diseases were also collected.

Statistical Analysis
The results were presented as mean ± standard deviation for the continuous variables normally distributed, medians for the continuous variables of non-Gaussian distribution and percentages for the categorical variables. The characteristics of patients were compared using the X 2 or Fisher exact test for the categorical variables. The quantitative variables were analyzed using the ANOVA test, in case of normal distribution, or the Kruskal-Wallis test in case of non-normal distribution. The level of significance was set at p<0.05. The statistical analysis was performed using SPSS version 20.0 (SPSS Inc.; Chicago, Illianois, US).

Baseline Characteristics, Clinical Presentation and Echocardiographic Results According to Colonoscopy Findings
The study population was mainly men (71%) with a mean age of 69±15 years. Only 9 patients (13%) had an identified portal of entry: Intravascular catheter (n = 4), haemodialysis (= 2), urinary origin (n = 2) and dental procedure (n = 1). From 62 patients without clear portal of entry, 20 of them (32%) showed colorectal lesions: 13 patients (65%) were pre-cancerous lesions and 7 (35%) were CRC. There were no significant differences in baseline characteristics between all groups (Table 2). A total of 68 patients (96%) debuted with fever and only 15 cases (21%) presented shortness of breath.  (14%) showed abdominal symptoms previous to the colonoscopy: 8 cases (80%) without colorectal lesions, 1 (10%) with a pre-cancerous colorectal lesion and 1 (10%) with CRC. There were no significant differences between all groups in clinical presentation and analytical parameters at the time of diagnosis. As a part of the extension study, 96% of patients (n = 68) underwent to a TEE and 14% (n = 10) to a 18F-FDG PET/TC during hospitalization. Native mitral valve was the most frequent location (35%, n = 25). There were neither statistically significant differences in echocardiographic features between groups (Table 2).

Microbiological Results According to Colonoscopy Findings
Streptococcus gallolyticus was the third most common microorganism (14%; n = 10) and the most frequent in those with colorectal lesions: 50% of IE with pre-neoplastic lesions (n = 7) and 25% of IE with CRC (n = 2). In contrast, only one case (2%) without colorectal lesions presented Streptococcus gallolyticus. From 62 patients (87%) without clear portal of entry, 10 of them (16%) presented Streptococcus gallolyticus IE. The relationship between Streptococcus gallolyticus IE and colorectal lesions is shown in Fig. 2. Both, CRC group (25% Vs 2%; p = 0.007) and pre-neoplastic lesions group (50% Vs 2%; p<0.001) were associated with significant higher incidence of Streptococcus gallolyticus IE. Additionally, the subgroup of cases with low degree pre-cancerous lesions (n = 12) was also related to a higher incidence of Streptococcus gallolyticus IE (n = 6) (50% Vs 2%; p<0.001). The most frequent microorganism in all groups was Staphylococcus aureus (n = 15, 21%).

Clinical Outcomes According to Colonoscopy Findings During Hospitalization
The mean inpatient stay was 34±71 days. A total of 15 patients (21%) presented in-hospital mortality and 28 cases (39%) required cardiac surgery during the IE event. There were no significant differences in inhospital mortality or treatment by cardiac surgery between the three groups. Fourteen systemic (20%) and eight central nervous system (11%) embolisms were reported, without significant differences between groups (Table 2).

Discussion
This study demonstrates that pre-neoplastic colorectal lesions are associated with IE caused by Streptococcus gallolyticus and even more frequently than CRC. Not only high-degree but also low-degree dysplastic adenomas presented this association. However, colonoscopy findings do not seem to affect the in-hospital prognosis of patients.

Colorectal Lesions and Gut Pathogens Related IE
The association between Streptococcus gallolyticus IE and pre-cancerous colorectal lesions has not been thoroughly explored to date. Our results revealed a strong relation between streptococcal IE and preneoplastic colorectal lesions, even in those cases with low-grade dysplastic adenomas. Although the relationship between Streptococcus gallolyticus IE and CRC is well-established (Ferrari et al., 2008), the mechanisms underlying this association are not well-known.  (Sillanpää et al., 2008;Rusniok et al., 2010). Thus, Streptococcus gallolyticus could have competitive advantage to colonize cancerous sites in the intestine (Fig.  3). Additionally, superior binding to collagen I on heart valves could be responsible for the increased occurrence of Streptococcus gallolyticus IE, compared to other Streptococcus bovis biotypes (Boleij and Tjalsma, 2013). Pre-neoplastic colorectal lesions could also increase collagen IV expression and facilitate systemic infections by Streptococcus gallolyticus with the same mechanism.

Clinical Implications of Association between Pre-Neoplastic Lesions and IE
Pre-neoplastic colorectal lesions such as colonic adenomas are found in up to 40% of 60-year general population and could evolve into CRC by the accumulation of mutations (Strum, 2016;Boleij et al., 2011;Levine and Ahnen, 2006). Although it is necessary to corroborate our hypothesis with larger studies, the fact that these colorectal lesions could act as a gateway for opportunistic gut pathogens, such as Streptococcus gallolyticus, present significant clinical implications. It could be advisable to extend the screening by colonoscopy in all patients with gut pathogens related IE or in those without a clear portal of entry, to rule out occult neoplastic and pre-cancerous colorectal lesions. The colonoscopic screening not only could identify the portal of entry but also detect and eradicate a potential colorectal disease malignancy with clinical implications (Kim et al., 2018;Pericàs et al., 2017). Moreover, in the last years, other advanced imaging techniques, such as 18F-FDG PET/TC, have emerged and may help in the early detection of septic embolisms and occult neoplasms (Pizzi et al., 2015). Furthermore, it would be necessary to analyze the capability of pre-cancerous colorectal lesions for the development of systemic infections and IE risk, in order to carry out preventive strategies and thereby avoid the development of the disease. Finally, patients with highly recurrent or multiple pre-neoplastic lesions may be considered high risk for developing IE due to gut pathogens. Therefore, specific prevention protocols should be considered in these cases.

Limitations
Our study is a retrospective analysis of a tertiary unicentric cohort with a limited number of patients and therefore the hypothesis stated in this study should be corroborated with larger sample size trials. The lack of uniform microbiological classification of Streptococcus bovis and also the confusing colorectal lesion nomenclature system has led to an underestimation of this relationship during the study period. However, patients included in data analysis were classified from all 3 biotypes of Streptococcus bovis with the latest nomenclature and colorectal lesions were graded from the World Health Organization classification to minimise the impact of these variations (Klein et al., 1977;Gupta et al., 2010;Darjee and Gibb, 1993;Schlegel et al., 2003;Giannitsioti et al., 2007;Corredoira et al., 2008;Jass and Sobin, 1989). Finally, the different techniques employed in the diagnosis of IE and colorectal lesions have evolved significantly during the study period, which could lead to differences between the results obtained in the first cases included compared to the latest ones.

Conclusion
Pre-cancerous colorectal lesions are associated with Streptococcus gallolyticus IE. Even low degree dysplastic adenomas present this association. Our results suggest that high and low degree colorectal lesions may act as a gateway for gut pathogens. Hence, it is necessary to rule out occult neoplastic and pre-cancerous colorectal processes in these patients.