Evolving trends in infective endocarditis in a developing country: a consequence of medical progress?

Background Staphylococcus has replaced streptococcus as the most common cause of infective endocarditis (IE) in developed health care systems. The trend in developing countries is less clear. Aim To examine the epidemiological trends of infective endocarditis in a developing nation. Methods Single-centre, retrospective study of patients admitted with IE to a tertiary hospital in Malaysia over a 12-year period. Results The analysis included 182 patients (n = 153 Duke’s definite IE, n = 29 possible IE). The mean age was 51 years. Rheumatic heart disease was present in 42%, while 7.6% were immunocompromised. IE affected native valves in 171 (94%) cases. Health-care associated IE (HCAIE) was recorded in 68 (37.4%). IE admission rates increased from 25/100,000 admissions (2012) to 59/100,000 admissions (2017). At least one major complication on admission was detected in 59 (32.4%) patients. Left-sided IE was more common than right-sided IE [n = 159 (87.4%) vs. n = 18 (9.9%)]. Pathogens identified by blood culture were staphylococcus group [n = 58 (40.8%)], streptococcus group [n = 51 (35.9%)] and Enterococcus species [n = 13 (9.2%)]. staphylococcus infection was highest in the HCAIE group. In-hospital death occurred in 65 (35.7%) patients. In-hospital surgery was performed for 36 (19.8%) patients. At least one complication was documented in 163 (85.7%). Conclusion Staphylococcus is the new etiologic champion, reflecting the transition of the healthcare system. Streptococcus is still an important culprit organism. The incidence rate of IE appears to be increasing. The rate of patients with underlying rheumatic heart disease is still high.


Introduction
Infective endocarditis (IE) is a universal disease with risk factors, predispositions and outcomes considered similar across continents for decades. An epidemiological transition has been reported over the past decade for developed health care systems where viridans streptococcus has been replaced by staphylococcus as the most common etiological agent. Methicillin-susceptible Staphylococcus aureus (MSSA) is now the most common agent in native valve infective endocarditis (NVIE) and methicillin-resistant S. aureus (MRSA) in healthcare-associated infective endocarditis (HCAIE) [1]. This transition is thought to be a result of ageing populations, the abolition of rheumatic heart disease (RHD) as the traditional major risk factor for IE due to appropriate antibiotic treatment of rheumatic fever and advanced device management

Annals of Clinical Microbiology and Antimicrobials
Page 2 of 9 Sunil et al. Ann Clin Microbiol Antimicrob (2019) 18:43 particularly in cardiac patients [2,3]. Moreover, reductions in IE mortality rates that had been anticipated with improved medical care did not materialize due to the changing epidemiology and the coinciding higher rates of HCAIE [1,4,5]. It is not clear whether developing countries undergo an epidemiological transition of IE to the same extent. Available studies show notable differences in developing countries with much higher rates of culture-negative IE, zoonotic infections and mortality, although with significant regional variations [6,7]. There is a lack of data for developing countries overall and South-East Asia in particular. The aim of this study is to describe the clinical trends, characteristics and outcomes of IE patients in a Malaysian tertiary hospital.

Materials and methods
The study is a single-centre, retrospective study of patients admitted with IE to a tertiary hospital in Kuala Lumpur, Malaysia from 1st January 2005, until 31st December 2017. The hospital caters a population of 1.7 million with an annual adult admission rate of 44,000-46,000/year. The study received approval from the institutional ethics committee (No: 201782-5459).

Patient recruitment and inclusion criteria
Patients were identified from the echocardiography reporting system, the admission logbooks and the hospital electronic medical records (EMR). All patients fulfilling the modified Duke's criteria [8] for definite or possible IE were included in the study. The first admission during the study period was considered the index admission for patients with more than one episode of IE. Patients younger than 18 years were excluded.

Data collection
Data regarding patient demographics, comorbidities and recent hospital admissions were documented. Healthcare-associated infective endocarditis (HCAI) was defined as IE in patients with health care contact before the index hospitalization. Health care contact constituted a history of receiving intravenous therapy (including chemotherapy), transfer from a specialized nursing care facility, haemodialysis, or hospitalization for at least 2 days in the 90 days prior to the index admission. Presenting features and examination findings were recorded. Documented laboratory variables included haemoglobin (Hb), white blood cell count (WBC), platelet count, urine haemoglobin, serum creatinine, erythrocytes sedimentation rate (ESR), C-reactive protein (CRP), serum albumin and troponin I. Results on admission (i.e. baseline) and results after 2 weeks of IE treatment were analysed for all those parameters. Blood results at the time when the IE diagnosis was made were considered baseline results for those patients who developed IE during the admission.
Every patient had at least 4 blood samples taken for blood culture. Automated laboratory systems were utilised for pathogen identification and antimicrobial susceptibility testing. Genus and species of organisms cultured and their respective sensitivities were recorded. The type of echocardiographic modality [transthoracic echocardiography (TTE) vs. trans-oesophageal echocardiography (TOE)] and its results were documented. All available echocardiographic cinematic loops were reviewed by two consultant cardiologists for the presence of rheumatic heart disease (RHD) according to the 2012 WHF criteria for the echocardiographic diagnosis of RHD [9]. Type of antibiotics used for initial empirical treatment and subsequent definite treatment were documented. All patients included in this study completed 4 to 6 weeks of intravenous antibiotics. For all patients undergoing surgery; timing, type of surgery and outcome were recorded.

Outcome measures and statistical analysis
The outcome was recorded as "death" or "discharged alive". Results were analysed using SPSS for Windows version 24. Data are summarized as a percentage or mean ± standard deviation (SD). Categorical variables were compared using Chi-square or Fisher-Exact test as appropriate, while Student t-test was used for continuous variables. A p-value < 0.05 was considered significant.

Recruitment
A search of the electronic medical records database and the admission logbooks yielded 251 patients with the label "IE". The folders of 29 patients treated for IE between 2005 and 2007 were incomplete or could not be retrieved. Twenty-six patients were excluded as they were under 18 years of age. The details of the remaining 196 patients were analysed for the presence of the modified Dukes criteria for definite or possible IE [8]. The modified Duke's criteria for definite IE were fulfilled by 153 patients which were consequently included in the study. Fourteen out of 43 patients who fulfilled the criteria for possible IE did not receive antibiotic treatment as per IE protocol and were excluded from the study. A total of 182 patients remained for analysis as shown in Fig. 1.

Clinical presentation
Most patients 152 (83.5%) reported fever on admission. The mean duration of fever was 20 (± 24) days.

Microbiology data
Positive blood cultures were available for 142 (78%) patients. The mean duration to achieve culture clearance was 6 (± 5) days. All HACEK group organisms (n = 11) were sensitive to ceftriaxone, but only 3 (27.3%) of them were sensitive to ampicillin. Nine out of 13 (69.2%) Enterococcus species were ampicillin-sensitive, while 4 (30.8%) were ampicillin-resistant (sensitive to vancomycin). Microbial agents and respective resistance patterns are shown in Table 1.

Antibiotic prescription
Empirical antibiotics were used while awaiting culture result. The majority of patients (65%) were treated with two empirical agents, typically comprising of a combination of gentamycin and benzylpenicillin or cloxacillin. Definitive therapy consisted of 6 weeks of culture-directed antibiotics for culture positive cases. 72 (39.6%) patients were treated with a single antibiotic while 110 (60.4%) patients received a combination of two antibiotics.

Outcome and complications
The overall mortality rate was high.

Evolving epidemiology
Literature from developed health care systems has been showing an epidemiological evolution of IE over the past decade Table 2. There is a lack of IE epidemiological data from developing countries, thus preventing conclusions regarding local IE trends.
We found an incidence of 31/100,000 IE cases per adult admissions over the entire study period which appears to be higher than in developed countries. In addition, we noted an increasing trend in the IE incidence. There was an almost threefold increase in the incidence rate from 25/100,000 to 59/100,000 over the last 6 years of the study period. Our patient cohort appears to be younger (50.4 ± 17 years) than the IE population of western countries where patients are typically in their 60ies or 70ies. This finding is in keeping with that of other Asian studies conducted in China [10] (mean age 47.8 ± 15.7 years) and Thailand [6] (mean age 47 years, IQR 16-85 years) and likely reflects the overall younger populations of developing nations. Despite this, patients in our study had a high rate of non-communicable diseases like diabetes (29.7%), hypertension (34.6%) and ischaemic heart disease (11.5%). The HCAIE rate was high in our centre (37.4%). A trend towards increasing HCAIE has been observed in many western tertiary referral centres. Toyoda [2] reported a HCAIE rate of 52.9% based on a large American IE database analysis with increased IE mortality in this group of patients. Fowler found that HCAIE is commonly associated with S. aureus infections, particularly methicillin-resistant S. aureus (MRSA) [1]. Our study identified staphylococcus predominance (40.8% staphylococcus vs 35.9% streptococcus). This is in contrast to regional studies done in China and Thailand that demonstrated still streptococcal predominance in their respective cohorts [6,10]. Viridans group streptococci account for only 17-26% of IE in western populations where RHD is now disappearing [11]. Staphylococcus (31-54%) is the most common culprit organism in high-income countries, with methicillin-sensitive S. aureus (MSSA) as the most common agent in native valve infective endocarditis (NVIE) and methicillin-resistant S. aureus (MRSA) for health care associated infective endocarditis (HCAIE) [11]. MRSA has also overtaken coagulase-negative staphylococci as the main cause of PVIE in recent studies [1,12]. Higher staphylococcus IE incidence is typically reported in referral centres and has been shown to correlate with higher mortality [13,14]. The prevalence of RHD in our IE patients (42%) is still much higher than in Western populations where degenerative valvular pathology and valvular prostheses [1,5] constitute the main underlying pathological substrates. The culture-negative IE rate (22%) in our centre is comparable with those reported in developed health care systems (about 10-20%). Relevant serological investigations for Coxiella burnetii, Legionella spp., Brucella spp., Bartonella spp., and Chlamydiae spp. should be performed where epidemiologically applicable [4,11]. Much higher rates of culture-negative IE are commonly reported in regional studies: 30-40% in China and India; 65% in Thailand [6,10,15]. Often, only limited investigations can be carried out due to financial constraints when a more extensive workup would be required to accurately identify the fastidious organism in developing countries.
Echocardiography demonstrated vegetations in 99.5% of cases diagnosed as IE in our study. This high rate is most likely due to improved diagnostic imaging modalities [10]. However, only 39.6% of the patients had a TOE done. This contrasts with reports from developed healthcare systems where the majority of patients undergo TOE to detect complications early. In our study, TTE missed vegetation in 8 (4.4%) patients emphasizing the diagnostic importance of TOE in suspected IE. The mitral valve was the most commonly affected valve.

IE mortality and surgery
The in-hospital mortality rate of 36.7% in our study population was high. Recent studies conducted in developed health care systems reported persistently high mortality rates of 20-22%, 30% and 45% for the in-hospital period, at 1 year and at 5 years respectively [7,12,16,17]. The root cause for the lack of improvement in IE mortality in developed health care systems is believed to be the increasing numbers of older and immunocompromised patients with HCAIE with more resistant organism [1,4,5,11]. The same trend seems to apply to urban tertiary centres of developing countries: increasing rates of HCAIE in a multimorbid, older population resulting in persistently high mortality rates despite optimal medical care. Early surgery has been shown in RCTs to significantly reduce in-hospital death and embolic complications [16,18]. In developed medical systems, about 40-50% of patients are managed surgically whereas the surgical intervention rate in our study was only 32% and might partially explain the rather high mortality. Mortality rates reported in other developing countries, such as China (11%) [10] and Thailand (11.4%) [6] are much lower than local figures and are associated with higher surgery rates (44% and 75% resp.).

Study limitations
As a retrospective study, this study has limitations associated with retrospective data collection. 29 patients identified as IE cases had to be excluded as the medical records could not be retrieved. The data may not be generalizable as the hospital is a tertiary referral centre with inherent selection and referral bias. There are strong urban/rural socio-economic disparities in developing nations with regards to standard of care and health care access. Prospective studies with larger sample sizes are needed to confirm our findings.

Conclusions
Staphylococcus is the new etiologic champion, likely reflecting the transition of the healthcare system to a more developed state entailing more invasive procedures and serving a generally older and more multi-morbid population with increasing HCAIE rates. Despite the epidemiological transition to staphylococcus predominance, RHD remains as an important predisposing factor in our local IE population. Streptococcus viridans is still an important etiological agent of IE. The incidence rate of IE is increasing.