Clinical, laboratory, microbiological and echocardiographic characteristics of infective endocarditis in a tertiary care hospital

Características clínicas, de laboratorio, microbiológicas y ecocardiográficas de la endocarditis infecciosa en un hospital de tercer nivel

Infective endocarditis (IE) is a difficult and complex disease. Despite advances in its diagnosis and medical and surgical treatment, IE continues to be associated with high mortality and serious complications ¹.

The epidemiological profile of IE has changed substan tially over the last few years ^2,3. Historically, it affected young patients with underlying predisposing valve disease, predominantly rheumatic, and Streptococcus viridans was the most frequent pathogen ⁴. Over the last few years, the age of patients with IE has increased and degenerative valve disease is the most common predisposing lesion in developed countries where rheumatic fever has ceased to be a public health problem. The number of patients with IE and no prior known history of valve disease has also increased considerably ⁵. The microbiology has also changed, with many recent studies reporting that Staphylococcus aureus is currently the main causative microorganism of IE^6,7.

Numerous studies have evaluated the clinical, micro biological and echocardiographic characteristics of IE in the Western world. However, little is known regarding the epidemiology of IE in our setting, its echocardiographic characteristics, the main microbiological agents involved, the most common clinical manifestations and the complications in our patients. The local behavior of this entity must be observed in order to determine if there is a difference in the clinical profile which could serve as a basis for developing future management strategies.

This study analyzes the clinical, laboratory, microbiologi cal and echocardiographic characteristics of IE in a group of patients at a tertiary care hospital in Medellín, Colombia over a period of three years.

This was designed as a retrospective observational study. All patients over the age of 18 with a diagnosis of IE in their medical chart and seen by the echocardiography service at Hospital General de Medellin between March 2011 and June 2014 were included.

Data were collected from the medical charts regarding de mographic characteristics, age, sex, underlying heart disease, presenting signs and symptoms, comorbidities, laboratory and microbiology results, echocardiographic findings, the need for surgery and complications.

All patients met the following requirements: 1. Over the age of 18 years, 2. Definitive or probable diagnosis of IE according to the modified Duke criteria, 3. Transthoracic and/or transesophageal two-dimensional echocardiography, and 4. Serial blood cultures.

Of the 48 cases included, 46 patients (95.8%) were clas sified as definitive endocarditis and two patients (4.2%) as probable endocarditis according to the modified Duke criteria.

The patients' main characteristics are shown in Table 1; 29 patients (60.4%) were men, and 19 patients (39.6%) were women, with a 1.5:1 male to female ratio. The average age (±SD) was 53.8±19.2 years.

Native valve endocarditis occurred in 36 patients (75%), prosthetic valve endocarditis in three patients (6.3%) and endocarditis associated with an intracardiac device in nine patients (18.8%). With regard to how they acquired the infection, 20 patients (42%) had healthcare-related endocarditis and 28 patients (58%) had community-acquired endocarditis. Most patients had no underlying predisposing heart condition (62.5%).

Fever and fatigue were the patients' most common clini cal manifestations (70.8 and 77.1%, respectively). A heart murmur was reported as a clinical finding in 47.9% of the patients (Table 2).

Thirty cases (62.5%) had positive blood cultures and 18 (37.5%) had negative blood cultures. Of the latter, one patient had a microbiological diagnosis with a valve tissue culture.

Table 3 summarizes the causative microorganisms of the infection during the study. Staphylococcus was the cause in 39.6% of cases and Streptococcus in 14.5%. S. aureus was isolated in 16 patients (33%) and was the most frequent microorganism, followed by Streptococcus sanguis in three patients (6.3%).

The initial diagnosis was made by transesophageal echocardiography in 26 patients (54.2%) and by transthoracic echocardiography in 22 (45.8%). The mean time between the clinical suspicion and the first echocardiogram was 24 hours (15-48 hours). Forty-one patients required only one echocardiogram for diagnosis, seven patients required two echocardiograms for diagnosis. The main echocardiographic findings are summarized in Table 4. Vegetations were found in 45 of the 48 patients (93.7%), most were mobile (40/45, 88.8%), and the mitral valve was the most frequent location (17/45, 37.7%), followed by the aortic valve (16/45, 35.5%). The average size of the vegetations was 17.6±11.3 mm. The most frequent complication was embolism in 21 patients (43.7%), followed by heart failure in 41.7%. The most com mon sites of embolism were the lung and central nervous system (CNS) (nine and eight cases, respectively) (Table 5).

A statistically significant association was found be tween Staphylococcus infection and pulmonary embolism (p=0.01), but not CNS embolism (p= 0.390).

There was no statistically significant association between the location of the vegetation and embolism. Out of 17 pa tients with mitral valve endocarditis, eight had an embolism (47%); out of 16 patients with aortic valve endocarditis, eight had an embolism (50%); out of 10 patients with endocarditis related to device cables or catheters, four embolized (40%); and out of six patients with tricuspid valve endocarditis, two embolized (33%).

Patients with mobile vegetations tended to have more complications, although the relationship was not statisti cally significant, except for heart failure (p=0.04) (Table 6). Out of eight patients with CNS emboli, seven (87.5%) had mobile vegetations, and only one patient had non-mobile vegetations. Out of nine patients with pulmonary emboli, eight (88.8%) had mobile vegetations and only one had non-mobile vegetations. One hundred percent of patients with valve perforation as a complication had mobile vegetations.

Most patients had ≥10 mm diameter vegetations (35 patients); of these, 100% had an embolic complication. Of the patients with <10 mm vegetations (10 patients), only three had an embolus, although this relationship did not reach statistical significance.

A statistically significant relationship was found between the size of the vegetations (≥10 mm) and paravalvular leaks (p=0.03) and prosthetic dehiscence (p=0.03).

Among the patients' baseline characteristics, advanced age (over 55 years) was statistically significantly associated with the risk of peripheral embolism (p=0.04); other variables analyzed, such as diabetes mellitus, arterial hypertension or kidney failure were not significantly associated with emboli.

Heart failure was a complication in 20 patients, 19 with native valve endocarditis. Of these, 11 patients had aortic valve vegetations and eight patients had mitral valve veg etations.

There were two cases of aortic root abscess in patients with native valve endocarditis. In one, E. faecalis was iso lated, and in the other, S. lugdunensis.

Perivalvular extension occurred in four patients with na tive valve endocarditis, with aortic valve (2), tricuspid valve (1) and intracavitary (1) vegetations. Staphylococcus was isolated in three patients and E. faecalis in one.

Eleven cases of valve perforation were found, seven involving the mitral valve and four the aortic valve. The most commonly involved germ in these cases was S. aureus (36.4%).

Using multivariate analysis, a significant association was found between mortality as an outcome and septic shock, kidney failure, Staphylococcus infection and the use of im munosuppressants. No association was found with the other variables analyzed (Table 7).

Although the incidence of IE in the general population was not analyzed, the results indicate that despite advances in health care, IE continues to pose a diagnostic and therapeutic challenge for physicians, especially now that the pattern of presentation differs substantially from that of a few years ago. Those at greatest risk of infection are no longer young patients with known rheumatic valve disease; now they are older people with no apparent valve disease ^5,8. In this study, most patients did not have a predisposing underlying heart condition (62.5%), just as reported by Castillo et al. in an epidemiological study in Spain, where in the last seven years, 64% of the IE cases were patients with no predispos ing heart disease ⁵.

The average age of onset was 53.8±19.2 years. The increase in the age of patients with IE coincides with other studies carried out within the last decade ^9,10. This could be explained by greater survival of the population and by the increased number of IE cases related to hospitalization and health care. As in the vast majority of studies, we found a predominance in males, with a

Our study confirmed many clinical characteristics of endocarditis which occur in other parts of the world. For example, most of the IE cases were in native valves (75%) and the mitral valve was most commonly affected, followed by the aortic valve ^10,11.

Infective endocarditis with negative blood cultures may occur in up to 31% of IE cases, and often poses a consid erable diagnostic and treatment challenge. It occurs most frequently due to prior antibiotic use, highlighting the need to suspend the antibiotics and repeat the blood cultures in this situation ¹². In this study, we found a 37.5% prevalence of IE with negative blood cultures.

The published studies disagree on which causative micro organism is the most frequent; this variation may be related to the study population ^10,11,14,15. The studies which include intravenous drug users report a higher prevalence of S. aureus as the causative agent; other studies predomi nantly include patients with prosthetic valve IE. According to reports from recent studies, the Staphylococcus species have surpassed Streptococcus species as the most frequent causative microorganism of native valve IE ^12,13. In this study, which predominantly included native valve IE cases, with no intravenous drug use cases, the Staphylococcus species were the most frequent (39.6%), especially Staphy lococcus aureus (33%), followed by Streptococcus species (14.5%), indicating that the globally reported shift from S. viridans as the main germ in the past to S. aureus currently, is also true in our study population and is probably related to the same reasons, such as a growing number of invasive procedures (vascular catheters, intracardiac devices, etc.).

The most common presenting symptoms were fatigue and fever; however, the prevalence of fever in our study was lower than that reported by other authors ¹². A total of 29.2% of the patients did not report a fever, which sug gests that the lack of fever should not exclude a diagnosis of IE in a patient with suggestive clinical characteristics.

The lack of a murmur in patients with IE has been de scribed previously; however, the proportion of cases without a murmur (52.1%) was much greater in our study than what was reported in a retrospective study by Fefer P. et al. ¹⁶.

Echocardiography was the cornerstone of IE diagnosis in our study. Vegetations were found in most cases (93.7%). The size of the vegetations (17.6 ± 11.3 mm) was much greater in our study than what was reported in a Spanish registry by Roca B. et al. ¹⁷, in which the average length was 9 mm. Mitral regurgitation was the predominant heart condition in our study (60.4%), similar to what was reported by Winston and Bolger ¹⁸.

The prevalence of embolism was high in our study (43.7%), coinciding with what was published previously by Vilacost et al. ¹⁹. The most frequent sites of embolism were the lung (42.8%) and the central nervous system (38%). The prevalence of spleen embolism was much lower than that reported in the literature, and this may reflect an underdiagnosis due to fewer abdominal imaging studies being routinely performed on these patients ¹².

Several observational studies have evaluated the re lationship between the location of vegetations on the echocardiogram and the risk of complications ^19-21, finding that vegetations on the mitral valve are a risk factor for embolism. In this study, we did not find a statistically significant relationship between the location of the vegeta tions and embolism, with a similar proportion of emboli in patients with mitral and aortic valve vegetations. Mobile vegetations were associated with a greater incidence of complications like embolism, although this association did not reach statistical significance in our study, perhaps due to the size of the sample.

As previously mentioned, the size of the vegetations in this study was much larger than what was reported in other studies; vegetations which were ≥ 10 mm had more frequent

The multivariate analysis showed an increased risk of death from endocarditis in the presence of septic shock, kidney failure, Staphylococcus infection and the use of immunosuppressants. Other studies on this aspect have reached similar conclusions as ours regarding the factors associated with greater mortality ^22-25. However, there are also studies which have found a significant association of other factors with mortality, such as heart failure and advanced age ²⁶. The discrepancies may be explained by the different study designs and the lack of uniformity in the analyzed variables.

Our study has several limitations, especially the sample size, its retrospective design and the low number of serological and molecular tests performed, which limits the detec tion of fastidious organisms such as Brucella, Bartonella, Coxiella and other rare causes of IE which cannot be ruled out with traditional microbiological studies.

This study provides an epidemiological profile of IE in Medellín, Colombia. Although this study included 48 pa tients over a period of three years in a tertiary care hospital, this frequency is comparable to or greater than that reported in other studies, such as that of Senior ²⁷, in which 92 pa tients were included over a longer period (11 years) and from three reference centers, and that of Olaya Sánchez et al. ²⁸ with 34 patients in four years. The sizes of these studies are explained by the low prevalence of infective endocarditis.

This study shows the epidemiological, clinical and echocardiographic characteristics of a population with IE in a tertiary care hospital in Medellín, Colombia. In our population, IE is currently a disease which predominantly affects older people without predisposing underlying heart disease. Most of the cases are native valve endocarditis, with greater involvement of the mitral valve. S. aureus was the most frequent causative microorganism. Echocardiography is the cornerstone of diagnosis. This study helps recognize the prognostic factors which identify patients at higher risk of inpatient mortality in our population.

Special thanks to Dr. Elsa María Vásquez Trespalacios for her help with the statisti cal analysis in this study. To Universidad CES and to Hospital General de Medellín.