Evolving characteristics of infectious endocarditis over 15 years in a primary care regional hospital

Original Article
Cardiovasc Med. 2020;23:w02094

Medical Department, Morges Hospital, Morges, Switzerland

Published on 28.02.2020


Infectious endocarditis (IE) is a rare condition affecting 3 to 10 individuals out of 100,000 [1]. Its presentation can be insidious, delaying its diagnosis for several weeks or even months, and impacting on its course and complications. In recent years, the prognosis of patients with IE has not changed much, except for cases where early surgery has been undertaken [2]. The increase use of echocardiography as a diagnostic tool and the better management of IE are balanced by the increased population of high-risk patients (immunosuppression, diabetes, intravenous drug injection, intracardiac devices). The medical literature on the subject is based on large studies carried out in tertiary reference centres, potentially biased by a selection of more serious cases.
The main objective of this work was to assess the actual presentation and evolution of IE in a primary care hospital and to compare these results with the data in the literature usually derived from tertiary referral centres. Our hypothesis was that our setting could be more representative of the initial presentation of infectious endocarditis in the general population and that IE subtypes may be present specific risks.



We performed a retrospective study of all consecutive cases of IE admitted to the regional hospital of Morges between 2002 and 2017 (15 years). This general hospital (200 beds with 60 beds for internal medicine) is the primary reference centre for a population of 50,000 people in the region of Western Switzerland situated between Lausanne and Geneva. We identified all patients who were diagnosed with IE in the hospital discharge letter during the study period. We crosschecked this patient list with all the echocardiographic examinations performed in the hospital during the same period with a diagnosis of suspected IE. We included in the study all the patients having undergone an echocardiographic examination and hospitalised with the final diagnosis of definite or probable IE, according to the modified Duke criteria [3]. The study was approved by the Ethics Committee of the Lausanne University Hospital and of the Faculty of Biology and Medicine of the University of Lausanne. The clinical data were processed in accordance with the ethical principles defined in the Helsinki Declaration.

Clinical data

Data were extracted from the hospital discharge letters, hospital charts and microbiological reports. The following parameters were analysed: demographics, cardiac and noncardiac predisposing conditions for endocarditis, clinical signs and symptoms at hospital admission, echocardiographic findings, blood cultures results, type of treatment, complications and outcome. According to the modified Duke criteria, patients were classified in three categories: definite IE, probable IE and inconclusive diagnosis. Patients with definite or probable IE were included in the study and analysed. Cases with a final diagnosis of non-infectious marantic endocarditis were also identified and described.
Infectious endocarditis was further classified as right-sided IE (RS-IE), left-sided IE on native valve (LS-NVE), left-sided IE on prosthetic valve (LS-PVE), left-sided IE on biological valve (LS-BVE), left-sided IE on mechanical valve (LS-MVE) or IE associated with a cardiac implantable device (CIED-IE).

Statistical analysis

Continuous variables were analysed as median with interquartile range (IQR). Continuous variables were compared with the Mann-Whitney test for unpaired data. A two-sided p-value <0.05 was considered as statistically significant. Graphics and statistics were performed using Graph Pad Prism 7.02 (Graph Pad Software, La Jolla, CA, USA).

Literature review

The literature search was carried out on PubMed with the following keywords: IE associated with epidemiology, pathogenesis, therapy, as well as with multiple combinations of these terms. Studies published from 2001 to 2017 were screened. For the comparison with our data, three studies were selected on the basis of specific settings: a study from a tertiary care centre in London [4], a study from a primary general hospital in the USA [5] and a national registry in Italy [6] including tertiary reference centres and smaller regional hospitals.


Demographic, clinical, and microbiological characteristics

During the period from 2002 to 2017, 151 cases were screened and 87 patients fulfilling the criteria for a final diagnosis of endocarditis were included (69 definite IE, 15 probable IE according to Duke diagnostic criteria and 3 non-infectious marantic endocarditis); 64 cases were excluded because of an inconclusive diagnosis. Demographic, clinical and microbiological characteristics of patients with definite or probable IE are presented in table 1. The median age of our patients was 68.5 years with a male predominance (78.6%). Diabetes was the most prominent comorbidity (19%), and was particularly frequent in left-sided bioprosthetic valve IE (41.7%), as well as mechanical valve IE (25%) and IE on intracardiac devices (67%). A previous history of IE (13.1%) also emerged as a predisposing factor. A cutaneous entry point (skin lesions, peripheral venous catheters or bladder catheter) (15.5%) was found more often than a dental procedure (4.8%). Of note, 6 patients (7.1%) of our population were intravenous drug users. No patients in our study were on haemodialysis as it was not available in Morges.
Table 1
Demographic, clinical and microbiological characteristics of patients with infective endocarditis.
CharacteristicAll episodesLS-NVELS-BVELS-MVERS-IECIED-IE
Number of IE episodes84 (100%)52 (61.9%)12 (14.3%)12 (14.3%)5 (6%)3 (3.6%)
Age (y)68.5 [55.8;78.3]69.5 [53.8;79.3]70.5 [69;84.3]63 [60;75.3]37 [36;38]64 [58;70]
Male78.6% (66:18)71.2% (37:15)91.7% (11:1)91.7% (11:1)80% (4:1)100% (3:0)
ComorbiditiesImmunosuppression and HIV positive4 (4.8%)2 (3.8%)1 (8.3%)1 (8.3%)00
Diabetes16 (19%)6 (11.5%)5 (41.7%)3 (25%)02 (66.7%)
Cancer4 (4.8%)2 (3.8%)1 (8.3%)001 (33.3%)
Predisposing factorsNative valve disease3 (3.6%)2 (3.8%)001 (20%)0
Intravenous drug abuse6 (7.1%)2 (3.8%)004 (80%)0
Congenital heart disease4 (4.8%)3 (5.8%)01 (8.3%)00
History of endocarditis11 (13.1%)4 (7.7%)3 (25%)3 (25%)1 (20%)0
Portals of entryCutaneous portal of entry (skin and catheters)13 (15.5%)8 (15.4%)1 (8.3%)3 (25%)01 (33.3%)
Dental intervention4 (4.8%)4 (7.7%)0000
Clinical presentationFever (>38°)60 (71.4%)32 (61.5%)11 (91.7%)11 (91.7%)4 (80%)2 (66.7%)
Fever AND cardiac murmur52 (61.9%)31 (59.6%)10 (83.3%)6 (50%)4 (80%)1 (33.3%)
Chills39 (46.4%)21 (40.4%)7 (58.3%)9 (75%)1 (20%)1 (33.3%)
Weight loss14 (16.7%)10 (19.2%)2 (16.7%)1 (8.3%)01 (33.3%)
Night sweats17 (20.2%)7 (13.5%)4 (33.3%)4 (33.3%)2 (40%)0
Dyspnoea, orthopnoea, nycturia41 (48.8%)25 (48.1%)7 (58.3%)5 (41.7%)3 (60%)1 (33.3%)
Oedema, weight gain28 (33.3%)18 (34.6%)4 (33.3%)3 (25%)2 (40%)1 (33.3%)
Signs of left cardiac failure32 (38.1%)18 (34.6%)5 (41.7%)4 (33.3%)3 (60%)2 (66.7%)
Signs of right cardiac failure33 (39.3%)21 (40.4%)4 (33.3%)5 (41.7%)1 (20%)2 (66.7%)
Cardiac murmur68 (81%)47 (90.4%)10 (83.3%)6 (50%)4 (80%)1 (33.3%)
Cutaneous embolisation/complication21 (25%)13 (25%)3 (25%)1 (8.3%)3 (60%)1 (33.3%)
Cerebrovascular embolisation/complication22 (26.2%)14 (26.9%)2 (16.7%)5 (41.7%)1 (20%)0
Echocardiographic featuresVegetations42 (50%)28 (53.8%)6 (50%)2 (16.7%)4 (80%)2 (66.7%)
Abcess/perforation14 (16.7%)9 (17.3%)3 (25%)1 (8.3%)1 (20%)0
New valvular insufficiancy/dehiscence24 (28.6%)18 (34.6%)5 (41.7%)1 (8.3%)00
MicrobiologyStaphylococcus aureus21 (25%)9 (17.3%)3 (25%)4 (33.3%)4 (80%)1 (33.3%)
Viridans group streptococci*19 (22.6%)18 (34.6%)1 (8.3%)000
Enterococcus species8 (9.5%)3 (5.8%)2 (16.7%)2 (16.7%)01 (33.3%)
Coagulase negative staphylococci5 (6%)4 (7.7%)1 (8.3%)000
Streptococcus bovis6 (7.1%)4 (7.7%)1 (8.3%)1 (8.3%)00
Other streptococci7 (8.3%)5 (9.6%)2 (16.7%)000
Other microorganisms6 (7.1%)4 (7.7%)01 (8.3%)01 (33.3%)
Polymicrobial - fungi000000
HACEK organisms5 (6%)2 (3.8%)02 (16.7%)1 (20%)0
Negative blood cultures7 (8.3%)3 (5.8%)2 (16.7%)2 (16.7%)00
ValveMitral35 (41.7%)30 (57.7%)05 (41.7%)00
Aortic39 (46.4%)20 (38.5%)12 (100%)7 (58.3%)00
Tricuspid3 (3.6%)0003 (60%)0
Pulmonary2 (2.4%)0002 (40%)0
Ventricular septal defect2 (2.4%)2 (3.8%)0000
Pacemaker3 (3.6%)00003 (100%)
Surgery33 (39.3%)24 (48%) 2 (16.7%)1 (20%)2 (66.7%)
B = biological; CIED = cardiac implantable electronic device; HIV = human immunodeficiency virus; IE = infective endocarditis; LS = left-sided; M = mechanical; NVE-PVE = native-prosthetic valve endocarditis; RS = right-sided
* viridans group streptococci do not include Streptococcus bovis.
The IE clinical presentation was classical, with fever (71.4%) and a cardiac murmur (81%) as the most prevalent manifestations; both were present in 61.9%. One third of the patients (38.1% and 39.3%, respectively) exhibited symptoms of left heart failure or right heart failure. Half of the patients reported chills (46.1%). In a quarter of our population, cutaneous emboli (25%) and neurological emboli (26.2%) were observed. Notably, only half of the patients with a mechanical valve endocarditis had a cardiac murmur. The two main microorganisms identified were Staphylococcus aureus (25%) and viridans group streptococci (22.6%), with the latter predominantly in native valves (34.6%) and the former in the other forms of endocarditis (17.3–100%). Streptococcus bovis, actually classified as a viridans group streptococci, was analysed separately in our study. Of the S. aureus associated cases, a fifth were due to methicillin-resistant S. aureus (MRSA; 4.8 vs 20.2% methicillin-sensitive S. aureus [MSSA]). The other classical microorganisms were less frequent; these included the HACEK organisms (Haemophilus species, Aggregatibacter species, Cardiobacterium hominis, Eikenella corrodens, and Kingella species) which were found five times (three Haemophilus and two Cardiobacterium, all detected in blood cultures) Unusual cases included a native valve endocarditis with abscess of the septum and left main coronary occlusion due to Staphylococcus lugdunensis, a bivalvular endocarditis due to Tropheryma whipplei as the first manifestation of Whipple disease (fig. 1) and a pacemaker lead and device endocarditis due to Campylobacter fetus. Overall, 8.3% of the cases were blood culture negative endocarditis. Microorganism identification tended to be more difficult in prosthetic valve endocarditis (16.7%) than in the other types. No fungal or polymicrobial cases of IE were found in our collective.
Figure 1
Example of non-bacterial endocarditis in a case of pancreatic carcinoma with small nodular and filamentous vegetations on the aortic valve (arrows). LA = left atrium; Ao = Aorta
Left-sided native valves were mostly affected (61.9%), followed by left-sided prostheses (28.6% with an equal distribution of mechanical valve prostheses and bioprostheses); right-sided IE and endocarditis with intracardiac material were much less frequent (6% and 3.6%, respectively). Patients with LS-BVE were significantly older than those with RS-IE. The IE on the left-sided heart valves was almost equally shared between the aortic valve (46.4%) and the mitral valve (41.7%). We observed three cases of IE on extracardiac material (defibrillator or pacemaker electrode leads [3.6%]) and two cases of IE on a ventricular septal defect (2.4%). Trans-thoracic echocardiography was performed in 100% of the cases and transoesophageal when clinically indicated, in 51% of the patients (43/84). Vegetations were identified in 50% of the cases and signs of abscess or valve perforation in 16.7%. Signs of new valvular insufficiency or dehiscence were identified in 28.6%. The mitral valve was the most frequently affected valve in native endocarditis (57.7%), and valves in the aortic position for bioprostheses (100%) and mechanical prostheses (58.3%). In our study, we recorded three cases on non-infectious thrombotic endocarditis or “marantic” endocarditis in the context of widespread cancer. In one case, the endocarditis was the first manifestation of the disease through multiple embolic cerebral infarcts. The echocardiographic appearance of the valve was classical with multiple small “frond-like” vegetations on the valve borders (fig. 2).
Figure 2
Case of endocarditis due to Tropheryma whipplei (Whipple disease). Nodular vegetations on the aortic valve (a) and mitral valve (b). LA = left atrium; LV = left ventricle; Ao = Aorta

Demographics and clinical characteristics of probable versus definite IE

The probable IE group, according to the modified Duke criteria, included 15 patients and was significantly associated with prosthetic valves (p = 0.02), although many characteristics were similar to the larger group of definite IE (table 2.). The male preponderance remained high (80%). The clinical presentation was dominated by fever, in the same proportion of patients as in the definite and probable groups (86.7 vs 68.1%, p = not significant). Similarly, the proportion with cardiac murmurs was not significantly different in the two groups (73.3 vs 82.6%, p = not significant). The significant differences were the presence of cardiovascular and neurological emboli or other related complications only in the definite IE group (p <0.02 and p <0.01, respectively).
Table 2
Comparison between definite and probable infective endocarditis (IE).
Number of IE episodes69 (82.1%)15 (7.9%) Native valves51 (73.9%)5 (33.3%) 
Age (y)69 [54;77]67 [60;81]Prosthetic valves16 (23.2%)9 (60%)
Male78.3% (54:15)80% (12:3)Pacemaker2 (2.9%)1 (6.7%)
Comorbidities   Microbiology   
Immunosuppression and HIV positive2 (2.9%)2 (13.3%)0.1446Staphylococcus aureus18 (26.1%)3 (20%)0.7512
Diabetes14 (20.3%)2 (13.3%)0.7247Methicillin-resistant S. aureus3 (4.3%)1 (6.7%)0.5520
Cancer2 (2.9%)2 (13.3%)0.1446Methicillin-sensible S. aureus15 (21.7%)2 (13.3%)0.7247
Predisposing factors   Viridans group streptococci*19 (27.5%)00.0181
Native valve disease2 (2.9%)1 (6.7%)0.4501Enterococcus species5 (7.2%)3 (20%)0.1483
IV drug abuse6 (8.7%)00.5853Coagulase neg staphylococci4 (5.8%)1 (6.7%)>0.9999
Congenital heart disease3 (4.3%)1 (6.7%)0.5520Streptococcus bovis5 (7.2%)1 (6.7%)>0.9999
History of endocarditis5 (7.2%)6 (40%)0.0033Other streptococci4 (5.8%)3 (20%)0.1043
Portals of entry   Other microorganisms5 (7.2%)1 (6.7%)>0.9999
Cutaneous portal of entry (skin and catheters)9 (13%)4 (26.6%)0.2355Polymicrobial – fungi00>0.9999
Dental intervention3 (4.3%)1 (6.7%)0.5520HACEK organisms4 (5.8%)1 (6.7%)>0.9999
Clinical presentation   Negative blood cultures5 (7.2%)2 (13.3%)0.6030
Fever (>38°)47 (68.1%)13 (86.7%)0.2122Valve   
Fever AND Cardiac murmur39 (56.5%)10 (66%)0.5701Mitral27 (39.1%)8 (53.3%)0.3900
Chills30 (43.5%)9 (60%)0.2684Aortic35 (50.7%)4 (26.7%)0.1519
Weight loss14 (20.3%)00.0643Tricuspid3 (4.3%)0>0.9999
Night sweats14 (20.3%)3 (20%)>0.9999Pulmonary1 (1.4%)1 (6.7%)0.3270
Dyspnoea, orthopnoea, nycturia34 (49.3%)7 (46.7%)>0.9999Ventricular septal defect1 (1.4%)1 (6.7%)0.3270
Oedema, weight gain23 (33.3%)5 (33.3%)>0.9999Pacemaker2 (2.9%)1 (6.7%)0.4501
Signs of left cardiac failure27 (39.1%)5 (33.3%)0.7749Echocardiographic featuresETO 63%ETO 40%0.1461
Signs of right cardiac failure27 (39.1%)6 (40%)>0.9999Vegetations39 (56.5%)3 (20%)0.0204
Cardiac murmur57 (82.6%)11 (73.3%)0.4705Abscess / perforation14 (20.3%)00.0643
Cutaneous embolisation/complication21 (30.4%)00.0174New valvular insufficiency22 (31.9%)2 (13.3%)0.2122
Cerebrovascular embolisation/complication22 (31.9%)00.0087Surgery31 (36.9%)2 (13.3%)0.0386
The main differences between the two groups involved the type of infected valve and the infectious agent. Probable IE affected mostly left sided mechanical valves (46.7%) and Enterococcus species were much more frequent than in the definite case of IE (20%); S. aureus (20%) remained another common identified microorganism. We did not encounter any viridans group streptococci in the probable IE collective. In this group, a history of previous IE was significantly more frequently recorded (40%), as well as an identifiable point of infection entry (26.6%). The mitral valve was more often affected. As foreseen, echocardiography was less often suggestive (33.3% of vegetations and new valvular insufficiency in probable IE versus 88.4% in definite IE). Not surprisingly, the majority of patients in the probable IE group were treated with antibacterials only, but there was one case of early and one case of late surgery in whom the diagnosis of IE could not be confirmed. Finally, surgery was significantly less often performed in the probable IE group (p <0.04).

Comparison between left-sided IE on native versus prosthetic valves

Native valve endocarditis represented 61.9% of all left-sided endocarditis cases (n = 76) (table 3.) Diabetes mellitus was significantly more often associated with left-sided prosthetic valve infection (p = 0.03). The clinical presentation was significantly different, with more bacteraemia signs (fever and chills) at diagnosis in the prosthetic group (p <0.01 and p <0.05, respectively). Cardiac murmurs were significantly less often observed in this subgroup (p <0.02). Similarly, as described before in the probable endocarditis group, left-sided prosthetic valve IE was less frequently associated with viridans group streptococci (p <0.005). The main difference between the two subgroups was the type of infected valve: the mitral valve was significantly associated with left-sided native valve endocarditis (p <0.005); on the other hand, the aortic valve was significantly more often affected in the prosthetic subgroup (p = 0.001).
Table 3
Comparison of left-sided infective endocarditis (IE) on native vs prosthetic valves.
Number of IE episodes52 (61.9%)24 (28.6%)     
Median age (y)69.5 [53.8;79.3]69.5 [61.8;79.5]
Male71.2% (37:15)91.7% (22:2)     
Comorbidities   Echocardiographic featuresETO 48%ETO 63%0.3242
Immunosuppression and HIV positive2 (3.8%)2 (8.3%)0.5866Vegetations28 (53.8%)8 (33.3%)0.1382
Diabetes6 (11.5%)8 (33.3%)0.0302Abcess / perforation9 (17.3%)4 (16.7%)>0.9999
Cancer2 (3.8%)1 (4.2%)0.9999New valvular insufficiancy/dehiscence18 (34.6%)6 (25%)0.4406
Predisposing factors   Microbiology   
Native valve disease 2 (3.8%)0>0.9999 Staphylococcus aureus9 (17.3%)7 (29.2%)0.2452
IV drug abuse2 (3.8%)0>0.9999Methicillin-resistant S. aureus2 (3.8%)2 (8.3%)0.5866
Congenital heart disease3 (5.8%)1 (4.2%)>0.9999Methicillin-sensible S. aureus7 (13.5%)5 (20.8%)0.5022
History of endocarditis4 (7.7%)6 (25%)0.0639Viridans group streptococci*18 (34.6%)1 (4.2%)0.0040
Portals of entry   Enterococcus species3 (5.8%)4 (16.7%)0.1978
Cutaneous portal of entry (skin and catheters) 8 (15.4%)4 (16.7%)0.9999 Coagulase neg staphylococci4 (7.7%)1 (4.2%)0.9999
Dental intervention4 (7.7%)00.3013Streptococcus bovis4 (7.7%)2 (8.3%)0.9999
Clinical Presentation   Other streptococci5 (9.6%)2 (8.3%) 0.9999
Fever (>38°)32 (61.5%)22 (91.7%)0.0069Other microorganisms4 (7.7%)1 (4.2%)0.9999
Fever AND cardiac murmur31 (59.6%)16 (66.7%)0.6186Polymicrobial – fungi000.9999
Chills21 (40.4%)16 (66.7%)0.0481HACEK organisms2 (3.8%)2 (8.3%)0.5866
Weight loss10 (19.2%)3 (12.5%)0.7440Negative blood cultures3 (5.8%)4 (16.7%)0.1978
Night sweats7 (13.5%)8 (33.3%)0.0626Valve   
Dyspnoea, orthopnoea, nycturia25 (48.1%)12 (50%)>0.9999Mitral30 (57.7%)5 (20.8%)0.0032
Oedema, weight gain18 (34.6%)7 (29.2%)0.7940Aortic20 (38.5%)19 (79.2%)0.0012
Signs of left cardiac failure18 (34.6%)9 (37.5%)0.8027Tricuspid000.9999
Signs of right cardiac failure21 (40.4%)9 (37.5%)>0.9999Pulmonary000.9999
Cardiac murmur47 (90.4%)16 (66.7%)0.0193Ventricular septal defect2 (3.8%)00.9999
Cutaneous embolisation/complication13 (25%)4 (16.7%)0.5578Pacemaker000.9999
Cerebrovascular embolisation/complication14 (26.9%)7 (29.2%)>0.9999Surgery24 (48%)6 (25%)0.1289
LS = left-sided; NVE-PVE = native-prosthetic valve endocarditis; RS = right-sided

Treatment strategy according to microbiology

The treatment of IE included single (19%) or combined antibacterial treatment (42%), in accordance with the European guidelines for the management of IE [7]. Early surgery was performed in 39% of the cases (fig. 3.). The median interval to surgery was 11 days for the aortic valve (IQR 7,–21) and 12 days for the mitral valve (IQR 9–20). We were not able to demonstrate a higher rate of surgery associated with staphylococcal IE. Single antibacterial therapy was often used for Streptococcus viridans or bovis IE (47% and 33% respectively). In-hospital mortality and 30-day mortality rates were both 6%. Five patients died before the completion of the antibacterial treatment. S. aureus was found in four of them and a coagulase-negative staphylococcus in the last patient.
Figure 3
Treatment modality by microorganism.

Evolving characteristics of IE

To assess an evolving trend in the clinical characteristics of IE, we subdivided the collective into two periods of 7.5 years (table 4.). There 33 cases in the initial period (2002–2009) compared with 51 in the second period (2009–2017). The IE incidence remained constant over the study period (0.31–0.36%). The median age at diagnosis increased significantly (p <0.01) from 55.8 years to 67.1 years in the second period of our study (fig. 4.). The female proportion of affected patients increased non-significantly, from 12% to 27%. In the distribution of microorganisms, we observed a downward trend for viridans group streptococci and HACEK microorganisms. In contrast, there was an increase in the number of cases with other streptococci, S. aureus or negative microbiological examinations. The distribution of affected valves remained constant; however we observed new cases of IE linked to infection of an implantable electronic cardiac device (not observed in the first period). As far as treatment was concerned, antibacterial monotherapy tended to decrease in favour of combined antibacterial treatment, which significantly increased in the second period (p <0.02). The rate of surgical intervention tended to decrease over the study period (48% to 33%), but not significantly.
Figure 4
Evolving trends of age at diagnosis of infective endocarditis.


To the best of our knowledge, this is the first study of IE reported from a single primary care hospital in Switzerland. Eighty-four cases were collected over a 15-year period, representing 0.3% of the total number of admissions in our hospital. IE remains a rare, but not an exceptional, pathology in a primary care hospital. Overall, 39% of the cases were ultimately transferred to a tertiary centre, mainly for early cardiac surgery. So, theoretically at least, our population could differ from those of the numerous publications originating mainly from tertiary centres, with referral bias from surrounding primary care structures. Our setting could better reflect the picture of IE in the general population, since more than 60% of the patients never reached a tertiary care hospital. Our typical patient with native valve IE is a 68-year-old man who presents with fever and a cardiac murmur. Blood cultures are positive for S. aureus or for viridans group streptococci. The echocardiography confirms the diagnosis by showing a new valvular insufficiency, vegetations or already a complication of the infectious process in the form of an abscess or a perforation of the valve. Patients with a prosthetic valve IE are typically somewhat older, especially for bioprostheses, more often diabetic, with more cases discharged from hospital with the diagnosis of probable IE. This is explained by the difficulty of confirming the diagnosis in patients carrying a mechanical valve with positive blood cultures in the absence of visible vegetations. In addition to an appropriate antibacterial treatment, early (90% of the time) or late cardiac surgery was necessary in more than a third of the patients.

Comparison with the chosen reference studies

Three reported studies of IE, performed in different environments, were chosen for comparison. First, a study conducted in a primary referral hospital, quite similar to our situation (n = 54 over 5 years) [5], then a publication from a large tertiary referral centre in London (n = 208 over 8 years) [4], and finally, a large multicentric Italian study including secondary and tertiary centres (n = 1456 over 7 years) [6].
The 82.1% of definite IE cases versus 17.9% of probable IE was comparable with equivalent results of the three studies mentioned, with a slight excess of our probable IE, most likely related to the greater proportion of prosthetic valve endocarditis. The hospital mortality observed in our study and in another primary reference centre was 5–6%, which was significantly different from the 18% observed in the tertiary referral centre study. At 6 months, this mortality difference between primary and tertiary referral centres persisted (9.2 vs 27%), reflecting the more severe cases drained by the tertiary centres. Diabetes was the most frequent comorbidity found in our study (19%), as in the Italian reference study. This proportion was almost double in the US study in a primary care centre [4]. Our proportion of intravenous drug users corresponded to the large studies analysed, but was three times lower than in a similar primary care study in the US. These data probably represented the differences in prevalence of diabetes and addiction in the different populations. The rate of recurrent endocarditis was similar to the recurrence rate of studies in secondary and tertiary referral centres. Similarly, a comparable rate of IE linked to the implantable intracardiac devices was observed. The proportion of infected native valves was two third of the population, as in the reference studies. The distribution between aortic and mitral valves matched the study in a tertiary centre, with a predominance of aortic endocarditis.
Our most commonly recovered microorganism was S. aureus (25%), which was similar to the tertiary reference centre. This rate was higher in the US primary care centre, most likely because of a higher proportion of intravenous drug users. Our second most frequent microorganisms were viridans group streptococci (22.6%). In comparison, 30% of viridans group streptococci were isolated in the tertiary centre before 2000, whereas the Italian study found only 15% between 2004 and 2011. This could be partially explained by the evolution of the microbiological epidemiology of endocarditis in recent years, as recently published [8]. The latter publication reported a shift in the microbiology of infective endocarditis, S. aureus being now the most common microorganism. Our negative microbiological samples were significantly fewer than in the reference studies, with a percentage of only 8.4% versus 15–20% in the tertiary centres and 11% in the primary centre in the US. This could possibly be explained by the evolution of direct detection techniques for genetic material in blood or on histopathological examination of operated valves. In our study, we detected 6% of HACEK microorganisms, which was four times more than in a large population based multicentric study comparing HACEK IE characteristics with other types of endocarditis (n = 5591) [9], and similar to our reference studies from secondary and tertiary centres. A plausible explanation for this difference could be the higher proportion of mechanical prosthetic valve carriers in our population; this predisposition has been shown to be an independent risk factor for developing HACEK endocarditis [9]. We observed vegetations in half of our patients. Vegetations were significantly less often detected than in our comparative studies of tertiary centres or population studies, in which vegetations were detected in 62% to 80% of all endocarditis cases. The explanation could be the lower rate of transoesophageal echocardiography and the higher proportion of mechanical prostheses in our study.


Our distribution of the different endocarditis subtypes was comparable to a large recent multicentric study [6]. In prosthetic left-sided IE, more S. aureus were detected, as well as other streptococci. We found fewer viridans group streptococci and coagulase-negative staphylococci when compared with the Italian multicentric study. The HACEK microorganisms remained much more frequent in our study and present in all subtypes. In the large majority of right-sided IE cases, S. aureus was isolated, which was also the case in our comparative Italian study. In cases of probable IE, we described a greater proportion of prosthetic valves, especially mechanical, because echocardiography is often not diagnostic in this group of patients, making it difficult to meet the Duke criteria for definite endocarditis. Transoesophageal echocardiography was performed more often in this group (68 vs 48%) as a result of inconclusive initial transthoracic echocardiography. In our study, we did not use positron emission tomography – computed tomography (PET-C)T techniques, currently recommended in this setting.


Among our patients, 39% underwent cardiac surgery in addition to antibacterial treatment with a median interval to surgery of 11 days. Three main indications for surgery were applied: valve dysfunction leading to heart failure, uncontrolled infection and prevention of the risk of embolism [1]. This surgical rate was quite similar to the 46% found by a study of the French registry of endocarditis [10], but lower than our comparative study of a tertiary centre, where it exceeded 50% [4]. Our results could be explained by the presentation of less severe cases and a short delay between the first symptoms and the diagnosis of IE, linked to immediate access to a cardiologist and an infectious diseases specialist in our setting in primary care [11].
Half of the Streptococcus viridans IE cases were treated with single antibacterial course (ceftriaxone for 4 weeks in 90% of cases), without the need for combination antibacterial therapy. However, viridans group streptococci were related to the same surgical rate as the other microorganisms. When needed, antibacterial combination therapy most often included penicillin, cephalosporins, vancomycin, aminoglycosides and rifampicin for a median duration of 6 weeks. For rare microorganisms, HACEK cases and in cases with negative microbiology, cardiac surgery was required in more than 50% of our patients. This could be possibly explained by a longer delay before diagnosis and by the difficulty in microorganism detection on culture or polymerase chain-reaction (PCR) examinations. A diagnosis bias towards surgical management is also evident because a large proportion of these cases (nine patients) were identified retrospectively by PCR examination of the excised material after surgery.

Temporal trends

We observed a significant increase in age at diagnosis over the 15 years of the study. This observation corresponds to the evolution observed in the literature [11]. The London study made in the 1990s also had much younger patients [5] as compared with the recent studies. The median age of our patient was comparable to a recently performed large study in Olmsted County [12]. As previously described by Correa de Sa et al. [13], we were able to show a tendency to an increase in the age (75 vs 71 years) of women with IE over the past 15 years in Morges. In the large comparative studies, one third of the patients were women. We do not have plausible explanations for the low proportion of women in the early period of our study. As in a major French study [14], which showed an increase of S. aureus to more than 25%, we also observed this trend, reaching this percentage at the end of our study.


The main limitation of our study is the small size of our population, inherent to a single primary reference centre setting. In addition, this was a retrospective, observational study with inherent biases linked to heterogeneous data collection in files over time. We sometimes encountered difficulties in obtaining a definite diagnosis based on the modified Duke criteria, because a few cases were rapidly considered as endocarditis and treated as such. In some cases, a larger number of positive blood cultures would have been required to diagnose an endocarditis as certain, but we do not feel that these inaccuracies impacted on the final results of our study. In the presence of bacteraemia in patients carrying prosthetic valve or with a previous history of IE, clinicians tended to be more suspicious in their diagnostic approach.


Our study showed similar evolutional trends as tertiary referral centres with some significant differences, possibly related to the primary care setting of our hospital. We showed an increase in age at diagnosis over the 15 years of the study. We confirmed the predominance of S. aureus observed in the global epidemiology of IE. The inverse evolution for viridans group streptococci, which have been proportionally less frequently observed in recent years, was also suggested in our study, although not significantly. A definite diagnosis of IE was significantly associated with native valve involvement as compared with mechanical prosthetic valve involvement, which was more often associated with a probable IE. In the latter group, we observed a higher rate of HACEK microorganisms. However, the meaning of this finding is limited by the small number of patients in this study. We also noted a lower rate of cardiac surgery referral than in the literature. This could be explained by the recruitment of less severe cases in a primary care setting compared with tertiary referral centres.
Table 4: Temporal trends of infective endocarditis characteristics.
 2002–20092009–2017     p-value
   n%   n%
Incidence33 on 14,2210.36%51 on 1,4500.31%
Staphylococcus aureus721%1427%0.6144
Methicillin-resistant S. aureus26%24%0.6438
Methicillin-sensible S .aureus515%1224%0.4145
Viridans group streptococci1030%918%0.1928
Enterococcus species515%36%0.2533
Coagulase negative staphylococci13%48%0.6438
Streptococcus bovis13%510%0.3959
Other streptococci26%510%0.6990
Other microorganisms13%510%0.3959
Polymicrobial – fungi00%00%0.9999
Negative blood cultures26%510%0.6990
Left sided endocarditis    
Right sided endocarditis    
Type of valve    -
Ventricular septal defect00%24%0.5172
Age (median)55.8 67-1 0.0027
Antibacterial monotherapy927%714%0.1578
Antibacterial double therapy824%2753%0.0126
Antibacterial + surgery1648%1733%0.1767
We would like to thank Dr De Rahm, vice-director of Morges Hospital, for his help in providing access to the patients’ records and Dr Eric Dayer for his editorial assistance.
Dr Nicolas Dayer, Hopital Riviera Chablais, Route du Vieux Séquoia 20, CH-1847 Rennaz, nico.dayer[at]hotmail.com
1 Cahill TJ, Prendergast BD. Infective endocarditis. Lancet. 2016;387(10021):882–93. doi:. http://dx.doi.org/10.1016/S0140-6736(15)00067-7 PubMed
2 Slipczuk L, Codolosa JN, Davila CD, Romero-Corral A, Yun J, Pressman GS, et al.Infective endocarditis epidemiology over five decades: a systematic review. PLoS One. 2013;8(12):e82665. doi:. http://dx.doi.org/10.1371/journal.pone.0082665 PubMed
3 Li JS, Sexton DJ, Mick N, Nettles R, Fowler VG, Ryan T, et al.Proposed modifications to the Duke criteria for the diagnosis of infective endocarditis. Clin Infect Dis. 2000;30(4):633–8. doi:. http://dx.doi.org/10.1086/313753 PubMed
4 Wallace SM, Walton BI, Kharbanda RK, Hardy R, Wilson AP, Swanton RH. Mortality from infective endocarditis: clinical predictors of outcome. Heart. 2002;88(1):53–60. doi:. http://dx.doi.org/10.1136/heart.88.1.53 PubMed
5 Fatima S, Dao B, Jameel A, Sharma K, Strogatz D, Scribani M, et al.Epidemiology of Infective Endocarditis in Rural Upstate New York, 2011 - 2016. J Clin Med Res. 2017;9(9):754–8. doi:. http://dx.doi.org/10.14740/jocmr3131w PubMed
6 Rizzi M, Ravasio V, Carobbio A, Mattucci I, Crapis M, Stellini R, et al.; Investigators of the Italian Study on Endocarditis. Predicting the occurrence of embolic events: an analysis of 1456 episodes of infective endocarditis from the Italian Study on Endocarditis (SEI). BMC Infect Dis. 2014;14(1):230. doi:. http://dx.doi.org/10.1186/1471-2334-14-230 PubMed
7 Habib G, Lancellotti P, Antunes MJ, Bongiorni MG, Casalta J-P, Del Zotti F, et al.Wytyczne ESC dotyczące leczenia infekcyjnego zapalenia wsierdzia w 2015 roku [2015 ESC Guidelines for the management of infective endocarditis]. Kardiol Pol. 2015;73(11):963–1027. Polkish. doi:. http://dx.doi.org/10.5603/KP.2015.0227 PubMed
8 Murdoch DR, Corey GR, Hoen B, Miró JM, Fowler VG, Bayer AS, et al.; International Collaboration on Endocarditis-Prospective Cohort Study (ICE-PCS) Investigators. Clinical presentation, etiology, and outcome of infective endocarditis in the 21st century: the International Collaboration on Endocarditis-Prospective Cohort Study. Arch Intern Med. 2009;169(5):463–73. doi:. http://dx.doi.org/10.1001/archinternmed.2008.603 PubMed
9 Chambers ST, Murdoch D, Morris A, Holland D, Pappas P, Almela M, et al.; International Collaboration on Endocarditis Prospective Cohort Study Investigators. HACEK infective endocarditis: characteristics and outcomes from a large, multi-national cohort. PLoS One. 2013;8(5):e63181. doi:. http://dx.doi.org/10.1371/journal.pone.0063181 PubMed
10 Iung B, Doco-Lecompte T, Chocron S, Strady C, Delahaye F, Le Moing V, et al.; AEPEI Study Group. Cardiac surgery during the acute phase of infective endocarditis: discrepancies between European Society of Cardiology guidelines and practices. Eur Heart J. 2016;37(10):840–8. doi:. http://dx.doi.org/10.1093/eurheartj/ehv650 PubMed
11 N’Guyen Y, Duval X, Revest M, Saada M, Erpelding M-L, Selton-Suty C, et al.; AEPEI study group. Time interval between infective endocarditis first symptoms and diagnosis: relationship to infective endocarditis characteristics, microorganisms and prognosis. Ann Med. 2017;49(2):117–25. doi:. http://dx.doi.org/10.1080/07853890.2016.1235282 PubMed
12 DeSimone DC, Tleyjeh IM, Correa de Sa DD, Anavekar NS, Lahr BD, Sohail MR, et al.Temporal trends in infective endocarditis epidemiology from 2007 to 2013 in Olmsted County, MN. Am Heart J. 2015;170(4):830–6. doi:. http://dx.doi.org/10.1016/j.ahj.2015.07.007 PubMed
13 Correa de Sa DD, Tleyjeh IM, Anavekar NS, Schultz JC, Thomas JM, Lahr BD, et al.Epidemiological trends of infective endocarditis: a population-based study in Olmsted County, Minnesota. Mayo Clin Proc. 2010;85(5):422–6. doi:. http://dx.doi.org/10.4065/mcp.2009.0585 PubMed
14 Selton-Suty C, Célard M, Le Moing V, Doco-Lecompte T, Chirouze C, Iung B, et al.; AEPEI Study Group. Preeminence of Staphylococcus aureus in infective endocarditis: a 1-year population-based survey. Clin Infect Dis. 2012;54(9):1230–9. doi:. http://dx.doi.org/10.1093/cid/cis199 PubMed