Infective Endocarditis With Oral Bacteria: Association With Findings During Oral Infection Screening


 Background: Oral streptococci represent the causing microorganism for infective endocarditis (IE) in many patients. The impact of oral infections is questioned, and it has been suggested that bacteraemia due to daily routines may play a bigger part in the aetiology of IE. The aim of this study was to examine the association between oral health and infective endocarditis caused by oral viridans streptococci. Methods: A retrospective study was conducted at Haukeland University Hospital (HUH) from 2006- 2015. All consecutive adult patients admitted to hospital for treatment of IE and subjected to an oral focus screening including orthopantomogram, were included. The clinical, radiological and laboratory characteristics of the patients, collected during oral infectious focus screening, were analysed. Patient survival was calculated using Kaplan-Meier and mortality rates were compared using Cox-regression. Results: A total of 208 patients were included, 77% (n=161) male patients and 23% (n=47) female, mean age was 58 years. A total of 67 (32%) had IE caused by viridans streptococci. No statistically significant correlation could be found between signs of oral infection and IE caused by viridans streptococci. The overall mortality at 30 days was 4.3 % (95% CI: 1.6-7.0). There was no statistical difference in mortality between IE caused by viridans streptococci or S. aureus (HRR=1.16, 95% CI: 0.57-2.37, p=0.680). Conclusion: The benefit of oral infection screening and elimination in patients admitted with IE seems uncertain. Based on this study it cannot be concluded that oral conditions and oral infection influence the overall mortality for patients with endocarditis.


Introduction
Infective endocarditis (IE) de ned as an infection of the heart valves or endocardium, displays high morbidity and mortality rates despite improved diagnostics and treatment. IE is commonly caused by Staphylococcus aureus, viridans streptococci followed by enterococci [1]. Less frequent causes are coagulase-negative Staphylococci (CoNS) and bacteria of the HACEK-group, including Haemophilus spp., Aggregatibacter spp., Cardiobacterium hominis, Eikenella corrodens, and Kingella kingae [2,3]. The bacteria manifests as a bio lm that may cause valve dysfunction and insu ciency, aneurysm or more rarely stulas. Since oral viridans streptococci are the second most common cause of IE, the oral infection screening and subsequent elimination of suspected infectious foci have traditionally been considered an important component of the treatment protocol [4].
Predisposing factors for IE are prosthetic heart valves or cardiac devices, valvular or congenital heart diseases, intravenous drug use (IVDU), previous episode of IE, and immunosuppression [3] . Also, recent dental or surgical procedures have been suggested to increase the risk for IE in risk patients [5]. However, up to 40% of patients diagnosed with IE lacks previously known risk factors [6].
The clinical presentation of the disease varies substantially and is often non-speci c [5]. IE caused by S. aureus typically presents as a serious infection with peripheral septic embolization whereas IE caused by oral bacteria most often display a subacute course, where symptoms may last for weeks to months before diagnosed. The mortality rates are higher for IE caused by S. aureus than for viridans streptococci [2]. Nowadays the acute and subacute division are historical and instead description of the valve involved, source of infection, and native or prosthetic valves are considered more relevant since it has implications for the choice of treatment [7].
By leaking through the oral mucosal barrier, oral bacteria can enter the bloodstream and thus reach the endocardium. Invasive dental procedures, such as tooth extraction, oral surgery or subgingival periodontal treatment allow oral bacteria to shed into the bloodstream. Bacteraemia may also occur during normal daily activities such as toothbrushing, ossing, and chewing [8].
Patients with IE are always hospitalized. The treatment consists of long-term intravenous antimicrobials alone or in combination with cardiac surgery. The empirical antimicrobial regimen is individualized according to the identi ed organism and resistance patterns. Heart surgery is performed mainly on the following indications: heart failure due to valve dysfunction, uncontrolled infection, and for prevention of embolism [9]. In addition, during the rst days of hospitalization possible infection foci other than the heart valves, should be identi ed and eliminated. As part of this, patients admitted for IE should routinely be referred to a dentist for investigation and elimination of potential oral infectious foci [10].
In most countries, including Norway, national guidelines recommend that patients with increased risk of IE should be prescribed antibiotic prophylaxis before invasive dental procedures [11]. The rational is that the antibiotic administered just before the procedure, reduces/shortens bacteraemia and thereby suggestively decrease the risk of IE. The e cacy of antibiotic prophylaxis for prevention of IE remains a debated area and solid evidence of its e cacy is still lacking. Another proposed preventive measure is maintaining good oral status to reduce the oral bacterial load and the subsequent bacteraemia [12,13]. Neglected oral hygiene and following periodontal disease are thought to increase the likelihood of bacteraemia in connection to normal daily activities such as chewing and toothbrushing [14].
To date the association between oral status and occurrence of IE caused by oral micro ora is not de nitely con rmed. Furthermore, the e cacy and bene t of the oral screening and infection elimination in IE patients is not known. The aim of the current study was to evaluate if there is an association between oral health status and occurrence of IE caused by viridans streptococci.

Study design and patient population
This study was based on data from patients diagnosed with IE in the period 2006-2015 at Haukeland University Hospital (HUH), Bergen, Norway. Inclusion criteria were patients admitted to hospital for treatment of IE, diagnosis veri ed according to the modi ed Duke criteria [15] , age above 18, and performed oral focus screening including orthopantomogram (OPG). Exclusion criteria were missing patient records at the oral and maxillofacial surgery [16] department, missing OPG and previous admission to hospital for same episode of IE. In case of repeated admissions during the same year, only the rst episode of IE was registered. Patients were followed until December 31, 2015. The study was approved by the Regional committee for ethics in medical research (REC West, approval no. 2015/1170).

Data management
The data was collected from the following sources: 1) patient records from the department Oral and maxillofacial surgery (OMS); 2) radiographic examination using OPG; 3) a local medical patient register of IE patients [17]. Two standardized case record forms were developed, one for examination of OPG and the other for evaluating the patient records. Data collection included the variables age, gender, last dental visit, subjective complaints regarding oral cavity, temporomandibular status, oral hygiene, caries, periodontal condition, wisdom teeth, maxillary sinus, number of teeth, dental apices status, number of endodontically treated teeth, condition of restorations, suspected infection foci, eventual bone pathology, suggested treatment and conducted treatment. Data regarding hospital admission, IVDU, presence of prosthetic heart valves, acquired or congenital heart valve defects, causing microorganisms, duration of symptoms at time of diagnosis, performed heart surgery during admission period, and mortality, were retrieved from the patient register. Data collection from the patients records from the OMS visits and OPG were performed blinded to all the information in the patient register, including the causing microorganism.
Furthermore, during the evaluation of OPG the examiner did not have access to patient records and vice versa. Repeated calibration exercises were performed before and during onset of data collection both regarding examination of OPG and registration of data from patient records. Also, patient record review and OPG examination were done independently by two different people to prevent bias.

Statistical analyses
Comparison of categorical variables were performed using Chi-square test and Fisher's exact test.
Survival probabilities were calculated using the Kaplan-Meier methods. Cox-regression models were used to compare mortality rates. IBM SPSS PASW Statistics for Windows, Version 25.0 were used for statistical analysis. P-values less than 0.05 were considered statistically signi cant.
A p value ≤ 0.05 was considered statistically signi cant. IBM SPSS PASW Statistics for Windows, Version 25.0 were used for statistical analysis. Statistical calculations were performed using Chi-square test, Fisher's test, and Kaplan-Meier analyses.

Patient characteristics
A total of 208 patients, 77% (n=161) male and 23% (n=47) female patients, admitted for treatment of IE during the study period and ful lling the stipulated criteria, were included. The mean annual incidence rate of IE per 100.000 inhabitants was 7.4 during the period. Clinical and microbiological characteristics of IE cases are shown in Table 1. There was a strong male predominance of the included patients, and the mean age was lower for men compared to women. The most frequent pathogen in declining order was staphylococci 33% (n=70), divided into S. aureus 25% (n=53) and the less common coagulasenegative staphylococci (CoNS) in 8% (n=17). The second most common group were viridans streptococci 32% (n=67). Of the total number of patients IVDU was the most common predisposing risk factor (n= 54), followed by previous episode of IE. The period of symptoms before diagnosed with IE was less than two weeks for 46% (n=96) and more than two weeks for 54% (n=112) of the patients. Information in the patient records regarding last dental treatment was missing for most of the cases. When stated, 12% (n=25) had dental treatment less than three months ago and 5% (n=11) had dental treatment more than three months ago.

Dental health status
The number of teeth was recorded as an index of previous and current oral health. One fourth of the patients had a full dentition of 28 or more teeth, while the remaining had a reduced dentition to different extent. Twelve patients were edentulous. From the dental records the treating surgeons' concluded that there was a possible oral infection foci in 27% (n=56) patients, while in the retrospective analyses of the OPG, the corresponding number was found to be 48% (n=99). In ve patients there were no comments in the patient records regarding the results of the oral infection screening (Table 1). Furthermore, the suggested dental treatment after screening was in several cases not ful lled and, in some cases, the eventual need for treatment not commented at all, Table 2.
According to the patient records 24% (n= 47) of the patients had no need for dental treatment. Extractions of teeth was the dominant suggestion of treatment 24 % (n= 49), endodontic treatment was advocated for 2% (n= 4), while 22% (n= 46) needed a combination of different treatments. For the remaining 28 % of the patients (n= 62) the recommended treatment was not speci ed or di cult to interpret.
From the dental records the surgeon had commented signs of periodontal disease in 19 % (n=39) of the patients, no periodontal disease for 12% (n=25) and for 69% (n=144) information about periodontal disease was missing. From the retrospectively investigated OPG 79 % (n= 164) had signs of periodontal disease, while 21% (n=44) had normal marginal bone levels. The difference between records from the clinical examination and the retrospective analyses regarding periodontal status was signi cant (p<0.001).

Association between oral health and causing microorganisms
Chi-square test showed that there were no statistically signi cant association between IE caused by viridans streptococci and suspected oral infection foci according to patient records or OPG. Furthermore, no association between IE caused by viridans streptococci and poor oral status regarding periodontal condition, apical infection, caries, or lesion in jawbone could be seen (Table 3).

Association between oral health and survival
The risk of mortality was higher for patients with reduction of marginal bone height registered on OPG compared to patients with normal marginal status, independent of causing bacteria species. For patients with IE caused by viridans streptococci the risk of mortality was twice as high as compared to patients with IE caused by other bacteria when the marginal bone height was reduced according to OPG (Fig. 1). There was no risk difference noted for patients with radiological signs of apical infection, where the highest mortality rate could be seen in the group of patients with IE caused by viridans streptococci, but without apical signs of infection (Fig. 2).

Discussion
To date it is a widely accepted approach to subject IE patients admitted to hospital to an oral infection focus screening and elimination as part of the treatment protocol [4,10]. In fact, the rate of these screenings is sometimes used as a quality parameter [18]. Although these routines seize resources and are mutilating and costly for the patient, there is no solid scienti c evidence to support these precautions. Until proven otherwise, screening is a reasonable approach because of the seriousness of IE and the indirect indication of link to oral status through causing microorganism [4]. However, considering the costs for the health care system, society and patients, the value and effectiveness of infection focus screening of IE patients should be investigated.
The current study could not con rm a signi cant association between oral infection foci and IE caused by viridans streptococci. In recent years, the association between invasive dental procedures causing bacteremia and subsequently giving rise to IE is vividly debated [8,19]. An extension of this debate could be to question whether ongoing local oral infections are a risk factor, and if so, what type of conditions that are of relevance. In case of oral infection causing systemic symptoms, the uncertainty is less since this type of infection, by de nition, causes shedding of oral bacteria into the bloodstream.
The study con rms high mortality rates for IE and are in range of other studies [1,20,21]. The high mortality rates might be explained by high prevalence of IVDU and comorbidities, but other explanatory factors cannot be excluded. The mortality rate for patients with IE caused by oral viridans group streptococci were similar as for the other bacteria. The results show that in patients with marginal bone loss the mortality was twice as high when IE was caused by viridans streptococci, compared to other bacteria. In contrast, there was no difference in mortality in patients with IE caused by viridans streptococci and presence of apical radiolucency. Earlier research shows that periodontal disease enhances bacteremia caused by viridans streptococci and has therefore been suggested to be a potential risk factor for IE [22]. Despite this, according to the dental records the impression was that the surgeons' focus was periapical infection, rather than periodontitis. The higher mortality in the group with periodontal marginal bone loss suggests attention to, former or ongoing, periodontal disease when screening patients suffering from IE. An eventual association between marginal bone loss and mortality could either be due to a direct cause of IE or constitute a marker for reduced oral and/or general health [23]. However, these results must be interpreted with caution due to low number of patients and retrospective study design.
Timing of periodontal treatment is another issue worth discussing [24]. Acute periodontal treatment has the potential to worsen the bacteremia during hospital admission [25]. Other studies fail to show that postoperative course changes due to periodontal treatment [26]. Patients with periodontitis show larger tendency to poor oral hygiene, due to increased pocket depths and higher incidence of bleeding. Poor oral hygiene may also cause gingivitis which weaken the mucosal integrity, increasing the risk of bleeding and bacteremia due to normal activities. For this reason, poor oral hygiene can be regarded as a risk factor for IE since the normal infection defense barrier is weakened because of in ammation [14].
In 2007 ESC and NICE started revision of the antibiotic prophylaxis [27]. The arguments for changing the routines for antibiotic prophylaxis was lack of convincing scienti c evidence for association between transient bacteremia and IE caused by invasive dental procedures in relation to the role of everyday routines [12]. Some studies show an association between invasive dental procedures and subsequent development of IE [28], while others fail to convincingly show such relationship [12]. Since our study did not reveal any association, we can to some extent support the studies suggesting that the daily routines, as chewing and toothbrushing, are a greater threat for individuals at risk for IE [8]. Today all antibiotic usage must be weighed against possible costs of resistance development. Consequently, because of weak association between invasive dental procedures and development of IE, some countries have changed their guidelines and refrain from, or decreased, the use of antibiotic prophylaxis for these patients [29,30]. In the current study, the association to previous dental treatment preceding IE could not be evaluated due to frequent missing data. The fact that the surgeons refrain from including this in the patient record is noteworthy, considering the common notion of its importance. In Norway, the national guidelines stipulate that antibiotic prophylaxis should be given to all patients with high risk of IE [31]. Thus, it can be assumed that this precaution has been applied by the patients' dentist in most cases. This could be one of the explanations for the lack of information regarding previous dental treatment in the patient records.
Suggested treatment by the dental surgeon was not executed for more than one third of the patients. Possible explanations for this can be due to either discharge from hospital or that the health status was too compromised to allow dental treatment. According to the patient records the focus was to eliminate apical infections. Due to lack of time before heart surgery and lack of equipment to perform endodontic procedures, the suggested treatment in several cases were extraction of teeth. Some suggest that dental treatment just before heart surgery increases the risk of mortality, while others claim that it makes no difference [24,26]. Whether the postoperative outcome will differ between patients receiving dental treatment before cardiac valve surgery (CVS) or not are unconclusive according to a meta -analysis from 2019 [32].
In summary, the results of the study indicate that there is no association between potential local oral infection foci and IE caused by oral bacteria. Considering the retrospective design of the study, and the seriousness of IE, these results should be interpreted with caution. However, considering the cost for both patients and the health care system, the results call for further studies, preferentially prospective, to verify these ndings.

Conclusions
The bene t of oral infection screening and elimination in patients with IE seems uncertain and there does not seem to be a straightforward association between ongoing dental infection and the development of IE caused by oral microorganisms. Furthermore, a risk hierarchy for different type of dental infections and the subsequent risk for IE could not be determined.

Declarations
Ethics approval and consent to participate Informed, passive consent was obtained from all patients still alive at time of registry.
The study was approved by the Regional committee for ethics in medical research (REC West, approval no. 2015/1170).

Consent for publication
Not applicable.

Availability of data and materials
The data that support the ndings of this study were taken from a local register, but restrictions apply to the availability of these data, which were used under license for the current study, and so are not available publicly. Data are available from the corresponding author on reasonable request and with permission from SJ.

Competing interests
The authors declare that they have no competing interests.