Clinical Spectrum of Infective Endocarditis in a Tertiary Care Centre in Western India: A Prospective Study ()
1. Introduction
Despite the availability of improved diagnostic and treatment amenities, infective endocarditis (IE) continues to have a high morbidity and mortality in the developing world [1] [2] . Several data from the developed world have emphasized regarding the changing spectrum of infective endocarditis [3] -[6] . However, in the face of this changing trend, the incidence and mortality have not declined significantly [7] . In spite of the high incidence of rheumatic heart disease and unrepaired congenital heart defects in India and other developing countries, there are relatively scant data on the profile of IE. In fact, most of these studies have been retrospective in nature [8] -[11] . There are very few prospective studies of IE from the developing world [12] [13] . In the face of globalization and rapid urbanization, we need to answer whether the changing scenario of IE observed in the west also applies to our population. Hence, we undertook a prospective study of IE focusing on the epidemiological trend, clinical, and microbiological profile and in-hospital outcomes of patients admitted to our institute.
2. Materials and Methods
This is a prospective study enrolling all the consecutive definitive cases of IE admitted at U. N. Mehta Institute of Cardiology and Research Centre (UNMICRC), Ahmedabad, India from January 2011 to January 2013. Only the patients who met the modified Duke’s definitive criteria for IE were included in the study. The data regarding the clinical presentation, predisposing factors, precipitating cause, investigative work up and outcome were recorded prospectively. At least three blood culture samples (aerobic and anaerobic) were obtained from three different vein puncture sites at least one hour apart before initiation of antibiotic therapy. Fungal culture was done if the clinical situation deemed necessary. All patients underwent transthoracic echocardiography (TTE) on admission and thereafter, if there was a change in the clinical status. The transesophageal echocardiography (TEE) was done in patients with high clinical suspicion but non-diagnostic TTE, in all patients with prosthetic valve endocarditis (PVE), and in patients with complications of IE (perivalvular abscess, valve perforation/rupture, worsening regurgitation). A serial follow up of all the patients was done at weekly intervals.
All the collected data were entered into IBM SPSS Version 20 for data analyses. Data were expressed as mean ± SD. While comparing culture positive and culture negative cases, paired “t” test was used for continuous variables and test of significance of proportion was used for categorical variables. Stepwise logistic regression analysis was done to determine the independent predictors of mortality and surgery. A p value < 0.05 was considered as significant.
3. Results
A total of 75 definitive cases of IE were admitted to our institute between January 2011 and January 2013. The male to female ratio was 2.26:1. The mean age of the patients was 27.46 ± 17.11 years (range, 7 months to 80 years). About 58 (77.3%) patients were younger than 40 years, with 30 (40%) patients under 20 years of age. Majority of cases (n = 64, 85.3%) fulfilled 1 major (echocardiographic) and 3 minor criteria. The rest 11 (14.7%) patients satisfied 2 major criteria (echocardiographic and blood/valve culture proven). Probable source of infection was present in 7 (9%) patients. The probable sources of infection were permanent pacemaker insertion, wound infection, dental manipulation, pyelonephritis, recent lower segment caesarean section with wound infection, maxillary sinusitis and pneumonia each being present in one patient respectively. Five patients had dental caries. None of our patients were intravenous drug abusers.
3.1. Underlying Heart Disease
Rheumatic heart disease (RHD) was the most frequent underlying heart disease. RHD was present in 31 (41.3%) patients (Table 1).
Table 1. Underlying heart disease in 75 patients of infective endocarditis.
1Intracardiac repair; 2Permanent pacemaker implantation; 3Rupture of sinus of valsalva; 4Blalock taussig shunt.
Among these, the most common lesion was mitral regurgitation in 28 (37.3%) cases followed by aortic regurgitation in 17 (22.7%) cases and combined mitral and aortic valve disease in 14 (18.7%) cases. None of these patients had isolated mitral stenosis or aortic stenosis. Congenital heart disease (CHD) represented the second most common underlying heart disease being present in 26 (34.7%) patients. Of these, 14 (18.7%) cases were uncorrected CHD’S and 12 (16%) cases were operated cases of CHD. Among the patients with uncorrected CHD, the most common disease was ventricular septal defect in 4 (5.3%) cases and bicuspid aortic valve in 4 (5.3%) cases. The most common corrected CHD predisposing to IE was intra-cardiac repair for tetralogy of fallot (TOF) in 8 (10.7%) patients (including one patient with right ventricle to pulmonary artery conduit). Prosthetic valve endocarditis (PVE) accounted for 7 (9.3%) cases. Mitral valve prolapse (MVP) and degenerative heart disease were observed in 5 (6.7%) and 2 (2.7%) patients respectively. In 8 (10.7%) patients, no structural heart disease could be identified.
3.2. Clinical and Laboratory Profile
The most common presenting feature in the patient was fever 69 (92%) followed by congestive heart failure 32 (42.7%). The most frequent peripheral signs were pallor 69 (92%) followed by clubbing 36 (48%) and splenomegaly 35 (46.7%) (Table 2).
The most common abnormal laboratory parameters were anemia 72 (96%) and raised erythrocyte sedimentation rate (ESR) 72 (96%). The most frequent complications were cardiac 40 (53.3%) followed by neurological 23 (30.7%), renal failure 20 (26.7%) and septic shock 20 (26.7%). Among the cardiac complications, the most common was congestive heart failure (CHF) 32 (42.7%). Five patients had conduction abnormalities, among which 3 (4%) had complete hearts block (CHB) and 2 (2.7%) had first degree atrioventricular block. Five (6.7%) patients with conduction abnormalities had aortic valve endocarditis. The other patient had CHB following intracardiac repair of TOF. 3 (4%) patients developed pericardial effusion among which one patient had tamponade that required tapping. Among neurological complications, the most frequent were cerebrovascular (CV) stroke 15 (20%) followed by Central nervous system (CNS) hemorrhage 4 (5.3%) and encephalopathy 4 (5.3%).
Table 2. Clinical and laboratory findings of 75 episodes in 75 patients of infective endocarditis.
Mycotic aneurysms were detected in 2 (2.7%) patients. Embolic complications occurred in 21 (28%) episodes among which, the most common were CV stroke in 15 (20%) followed by digital infarcts in 3 (4%) episodes respectively. Pulmonary infarct, splenic infarct and renal infarct occurred in 1 patient each. The pulmonary infarct was a consequence of tricuspid valve endocarditis secondary to intracardiac repair for TOF with implanted permanent pacemaker. Renal function was deranged in 20 (26.7%) and hepatic dysfunction occurred in 17 (22.7%) patients.
3.3. Echocardiography
The TTE was performed in all 75 patients while, TEE was done in 17 (22.7%) patients only (Table 3).
TTE was diagnostic in a total of 67 (89.3%) cases and in the rest, TEE was diagnostic. Multiple vegetations (≥2) were more common than single vegetation (65.3% vs. 34.7% episodes respectively). Vegetations of <10 mm and ≥10 mm were nearly equally distributed (49.3% vs. 50.7% episodes respectively). The most common site of vegetation was mitral valve in 37 (49.3%) episodes followed by aortic valve in 29 (38.7%) episodes. Ten (13.3%) patients had vegetations in extravalvular location among which 8 patients were postoperative cases of CHD (right ventricle to pulmonary artery conduit in 3 patients, ventricular septal defect patch (VSD) in 3, right ventricular outflow tract (RVOT) in 2 patients) followed by 2 patients with unrepaired CHD (patent ductus arteriosus (PDA) insertion and pulmonary artery in 1 patient each). Nine (12%) patients had increase in size of vegetations on serial echocardiography. Seven (9.3%) patients had PVE among which one patient had partial dehiscence of prosthesis and four patients had moderate paravalvular leak. Three patients had aortic root abscess and two patients had aortic cuspal perforation causing severe regurgitation.
3.4. Microbiological Profile
Blood cultures were positive in 30 (40%) patients only. Among these, staphylococci were the most frequent isolate in 12 (16%) cases followed by streptococci in 9 (12%) cases (Table 4). However, among all the above culture positive cases, only 11 (14.7%) patients satisfied the duke’s major criteria for culture positivity and in the rest of cases only single culture was positive. In patients with culture negative endocarditis who underwent surgery, tissue culture was negative in all except one patient in which the isolate was pseudomonas.
There were no significant differences in sex ratio, complications, need for early surgery, vegetation characteristics and mortality between culture positive and culture negative group (Table 5). There was significant higher proportion of patients with positive blood culture who had CHD as the underlying heart disease (50.0% vs. 24.4%, p = 0.04) and consequently culture positivity was more in younger age group as compared to older age group (p = 0.027). Also, a higher proportion of patients with negative cultures had prior antibiotic usage in comparison to culture positive patients (68.9% vs. 30.0%, p = 0.002).
Table 3. Echocardiographic profile in 75 episodes of infective endocarditis.
Table 4. Blood cultures in 75 episodes of infective endocarditis.
1Methicillin resistant staphylococcus aureus; 2Methicillin resistant staphylococcus aureus; 3Coagulase negative staphylococci.
3.5. Clinical Course and Outcome
The antibiotics were used according to the underlying heart disease, clinical situation, offending organism and complications. The mean duration of antibiotic therapy was 26.61 ± 12.78 days (range 1 - 45 days). Most commonly, patients received two antibiotics 42 (56%). In successfully treated patients the mean duration of antibiotic treatment was 30.89 ± 10.3 days. The most commonly used antibiotic combination was cephalosporin and aminoglycoside 24 (32%). This was followed by cephalosporin, aminoglycoside and vancomycin 23 (30.7%); and crystalline penicillin with aminoglycoside in 12 (16%) of patients. The other antibiotics used were fluoroquinolones, teicoplanin and linezolid. Patients with either proven/suspected fungal IE were treated with fluconazole or amphotericin.
Only 12 (16%) patients underwent surgery for IE. The most common underlying heart disease was CHD (n = 5) followed by RHD (n = 3), no structural heart disease (n = 2), degenerative valve disease (n = 1) and MVP (n = 1). The reasons for surgery during active IE were uncontrolled CHF in 7 (58.3%) and persistent infection despite adequate antibiotic use in 5 (41.6%). The most common surgery done was mitral valve replacement 3 (25%) followed by double valve replacement (mitral and aortic) in 2 (16.7%) patients. The other surgeries done were VSD closure, VSD closure with aortic valve replacement, PDA ligation, conduit replacement, mitral valve repair, rupture of sinus of valsalva repair and left pulmonary artery plasty in one patient each. The total in hospital mortality of patients who underwent surgery was 8.3% (1 out of 12 patients).
The total in hospital mortality in patients with IE was 22 (29.3%). The RHD and CHD in 9 patients remained the most common underlying heart diseases followed by PVE in 3 patients. The refractory heart failure was the most common cause of death in 7 (31.8%) patients. The other causes of death were sepsis in 6 (27.3%), cerebral embolism in 6 (27.3%) and CNS hemorrhage in 3 (13.6%) patients respectively.
4. Discussion
Though there have been many prospective studies in the developed world [5] , these data are not representative of the developing world. The spectrum of IE is different in India as compared to the west with respect to the predisposing factor’s, underlying heart disease, microbiology and outcome [14] . The rapid urbanization, indiscriminate use of antibiotics, increasing geriatric population, use of prosthetic devices and increasing nosocomial infections has changed the clinical scenario of IE.
Age: The mean age of our patients was 27.46 ± 17.11 years with 77.3% of patients younger than 40 years which is very similar to the previous Indian studies [8] [9] [13] . This is in contrast to the west, were the majority of patients were older than fourth decade with a mean age of 56 ± 17 years [2] [3] [5] . This difference may be attributable to the fact that, this study also included the pediatric population in contrast to the studies from the west which included only the adult patients. Also, RHD which continues to be a rampant problem in the underprivileged population in India predominantly affects younger age group. In addition, the proportion of CHD patients
Table 5. Comparison of blood culture positive and negative episodes of infective endocarditis.
is our study was higher in contrast to the western studies [7] . The higher age group in the west is due to increasing geriatric population, increased incidence of degenerative valvular diseases and reduced frequency of RHD [4] -[7] [15] [16] .
Underlying Heart Disease: RHD was the most common underlying heart disease in this study which accounted for 41.3% of all cases. This is similar to other studies from the developing world in which RHD accounted for 42% to 76% of cases of IE [8] -[11] . This is in contrast to the developed world in view of decreased incidence of rheumatic fever in west [4] [5] [7] . CHD also accounted for significant cohort (34.7%) of IE cases in our study similar to other Indian studies by Choudhary et al. (33%) and Math et al. (39.4%) [8] [13] . This high incidence of CHD in our study is in contrast to a large multi-centric study by Murdoch et al. in which CHD constituted 12% of IE cases [7] . This difference may be due to a large number of uncorrected CHD in Indian population as compared to the west. Out of all patients in CHD group, 12 (46.15%) patients were postoperative patients among which the most common surgery done was intracardiac repair for TOF (n = 8). This may be due to the reason that, our institute is a high volume tertiary referral cardiac surgical centre with active pediatric cardiology and cardiac surgery department. In contrast to the western data in which MVP was reported to be a common underlying heart lesion, this group constituted a very small fraction of patients 5 (6.7%) in our study similar to other Indian studies [8] [9] [13] [17] . In our study, PVE constituted 7 (9.3%) cases of IE which is very similar to other western studies but higher than some previous Indian studies [8] [18] [19] . The incidence of PVE has increased in India from 1% in 1980’s to 10% in 1990’s [9] . This increase may be attributable to the reason that an increasing number of patients are undergoing surgical treatment for valvular heart disease. None of our patients were intravenous drug abusers. Although it is traditionally believed that IE does not affect normal valves, recent large series have reported no underlying structural heart disease in about a third of patients [17] . While in our study 8 (10.7%) of patients had no structurally heart disease.
Blood Culture: In this study about 30 (40%) of blood cultures were positive in 75 episodes of IE. This is similar to most Indian series which have shown blood culture positivity rate ranging from 21% - 47% [8] [13] [20]. However blood culture positivity rates in western series is 90% as noted in the Murdoch et al. series [7] . A major factor for this disparity as compared to west may be related to the higher proportion of prior antibiotic therapy in culture negative patients before admission to hospital in our study (68.9% vs. 30.0%, p = 0.002). Another factor could be related to the underutilization of specialized culture media and various serological tests for fastidious organisms. The low culture positivity rate noted in our study is extremely worrisome, as the optimal therapy of IE depends on the organism isolated on culture and their sensitivity patterns. It is very difficult to draw any conclusion in view of the low culture positivity noted in our study.
Recent data from the western series have shown a shift in the microbiological spectrum with staphylococci being the commonest isolate [3] [7] [21] . This was very similar to our study in which the most common isolate is staphylococci followed by streptococci. This observation is similar to other Indian studies [8] [13] . However, the previous Indian study by Garg et al. has shown streptococci as the most common isolate [9] . The increasing emergence of staphylococcus in west is accounted by an increasing geriatric population, rising drug abuse, increasing nosocomial infections and use of prosthetic devices. However this doesn’t apply to Indian patients as RHD and CHD continue to be the most common underlying heart disease in Indian population. In view of a large culture negative group in our study, it is difficult to draw any inference with respect to the commonest organism. Also, in view of our institute being a tertiary referral centre, our study group represents a skewed population of sick referred patients and may not be representative of the population in general.
Echocardiography: Echocardiography has proved as the most important tool in establishing diagnosis of IE in our study. This was especially due to the low culture positivity rates. Vegetations were detected by TTE of episodes in 67 (89.3%) and in all the remaining patients, TEE was diagnostic. This finding is higher than a study by Taams et al. (86%) and various other studies [9] [22] .
The reason for this disparity may be due to the more young population in our study group and the lean built of the Indian patients thereby facilitating good echo windows. The stepwise approach consisting initial use of TTE followed by TEE in patients with non-diagnostic TTE like PVE and patients with poor echo windows explains the higher incidence of vegetations in our study. This underscores the importance of echocardiography in diagnosis of IE in culture negative cases. The most common site of vegetations in our study was mitral valve followed by aortic valve which is similar to a study by Garg et al. and other Indian studies [8] [9] .
Outcome: In our study, the in-hospital mortality was 22 (29.3%) which is much higher than that reported in the recent western series (12.6% - 17.7%) [6] [7] . This higher figure in comparison to the west is attributable to the differences in study population. This may be due to the relatively sicker, complicated and refractory cases referred to our centre; and relatively late presentation of patients in developing countries. Another important factor is CHF, which is the leading cause of death was present in 32 (42.7%) of episodes of IE. Also the high culture negative rates and lower rates of surgery 12 (16%) may be contributory. A higher proportion 32 (42.7%) of patients with PVE experienced death compared to the NVE that signifies the morbid nature of PVE. This may again be due to lower rates of surgery in this group as a part of reluctance on the cardiologist and cardiac surgeon in operating such morbid patients with PVE. Although the incidence of re-infection of the newly implanted valves is higher in PVE (6% - 15%) [23] , the benefit of redo surgery far outweighs the risk of re-infection. Also, a very high proportion of patients with IE in corrected CHD group died (7 out of 12 cases) as compared to uncorrected CHD group (2 out of 14 cases) due to lesser rates of surgery in the corrected CHD group. Hence, we need to overcome these barriers by employing aggressive surgical strategy in IE affecting prosthetic valves and operated CHD cases.
The predictors for mortality on univariate analysis were age < 20 years; thrombocytopenia, hematuria, hepatic dysfunction, septic shock, neurological abnormalities, CV stroke, CNS hemorrhage, CHF, renal failure and AV block (Table 6).
But on multivariate analysis; the independent predictors for mortality were age (p = 0.030), CHF (p = 0.003), renal failure (p = 0.037) and neurological abnormalities (p = 0.046). Septic shock was of borderline significance (p = 0.062). Garg et al. and Siddiq et al. reported CHF, renal failure and prosthetic valve dysfunction as independent predictors of mortality. Choudhary et al. reported a significantly higher mortality in patients with neurological manifestations [8] [9] [21] . On multivariate analysis, Math et al. reported underlying heart disease other than NVD and septic shock to be independent predictors of mortality [13] .
Surgery: Several studies have shown that surgery is potentially lifesaving and is required in 25% - 50% of cases during active infection [7] [23] . In our series, surgery was performed in only 16% of patients, with major indications being refractory CHF and persistent infection despite adequate antibiotic therapy. This is similar to other Indian studies by Math et al. (15%). This is in contrast to the recent data from developed countries where surgery was performed in 49% - 52% of cases [6] [7] . The major reason for this disparity is the lack of availability and affordability for cardiac surgery and reluctance on the part of cardiologist/cardiac surgeon in undertaking surgery in active endocarditis.
In our study, the most common complication leading to death was CHF 24 (31.8%). On stepwise logistic analysis for predictors of surgery, duration of illness < 45 days and culture positivity were predictors of cardiac surgery on univariate analysis (Table 7).
Table 6. Results of stepwise logistic regression analysis for predictors of lethality of active infective endocarditis.
1Confidence interval; 2Cardiovasular; 3Not significant; 4Central nervous system.
Table 7. Results of stepwise logistic regression analysis for predictors of surgery in active infective endocarditis.
But on multivariate analysis, only duration of illness <45 days was independent predictor of cardiac surgery (p = 0.037). Despite CHF being the commonest indication for surgery, the presence of CHF was not predictive of cardiac surgery on univariate analysis. In a similar study done by Garg et al., CHF and cardiac abscess were independent predictors of cardiac surgery [9] . In our study, the mortality rate in patients who underwent surgery was 6 (8.3%). The chance of re-infection of the newly implanted valve is 2% - 3% [1] . This may be far lower than the mortality rate in patients of IE with heart failure managed only medically (56% - 86%) [23] . Despite cardiac surgery, the most common cause of death in our study was CHF thereby, emphasizing the need for early surgery in these patients.
5. Conclusion
This prospective study emphasizes that the spectrum of infective endocarditis is different in Indian population compared to the west and carries a substantial morbidity and mortality. Infective endocarditis occurs at a relatively younger population than the west. Rheumatic heart disease followed by coronary heart disease still continues to be the most common underlying heart disease in our population. The lower culture positivity rate and lower rates of surgery are worrisome in Indian patients. Strategies need to be designed to improve the culture positivity by employing specialized culture media and other ancillary tests. Congestive heart failure, renal failure, neurological complications, age and sepsis were independent predictors of mortality. Also, prosthetic valve endocarditis and infective endocarditis affecting operated cases of coronary heart disease carry a high mortality risk. Hence, aggressive surgical strategy needs to be adopted to improve the outcome of the high risk patients.
Limitations of the Study
Firstly, the true microbiological spectrum of the patients could not be determined due to prior antibiotic therapy, lack of specialized culture media, serological tests and polymerase chain reaction for fastidious organisms thereby leading to high proportion of culture negative cases. Secondly, our institute being a tertiary referral centre; the data may be skewed to sick, complicated and refractory cases and may not be representative of general population at community level. Thirdly, our study included only duke’s definitive cases thereby excluding the probable cases of IE. Also, the long-term follow-up of patients is not available.