ABC | Volume 114, Nº1, January 2019

Original Article Marques et al. In-hospital mortality in infective endocarditis Arq Bras Cardiol. 2020; 114(1):1-8 Table 2 – Reasons for patients not being candidates for cardiac surgery (31 patients) Cause Patients (n) Significant comorbidities: 21 Dementia and cognitive impairment with dependence in activities of daily living 5 Ischemic stroke with significant post-stroke sequelae in patients with advanced age or multiple comorbidities 5 Advanced age with substantial associated comorbidities 4 Noncompliant HIV patients with poor general clinical conditions 4 Active intravenous drug user with poor general clinical conditions 1 Malignant tumor with poor prognosis 1 Severe alcoholism with significant organ damage 1 Hemorrhagic stroke (1 st month after the event) 3 Active bacteremia in association with an active extracardiac infectious focus 3 Death nearly after the diagnosis (the physicians did not have the chance for surgical referral) 3 Post-autopsy diagnosis 1 stroke with hemorrhagic transformation and one was an HIV patient with three IE episodes that was previously submitted to 2 cardiac surgeries due to IE and with significant associated comorbidities); 4 patients died after the intervention (2 patients due to septic shock, 1 due to cardiac tamponade and in 1 the cause of death was uncertain), resulting in a surgery-related mortality rate of 9%. The other adverse outcomes during hospitalization are described in Table 3. Of the 65 patients that evolved with heart failure, left ventricle systolic dysfunction was observed in 5 patients in the transthoracic echocardiography performed during hospital stay. None of the patients had previously known left ventricular systolic dysfunction. Predictors of in-hospital mortality In the univariate analysis, previous heart failure, apyrexia, Staphylococcus aureus etiology, non-isolation of Streptococcal Species, evidence of paravalvular abscess or valve obstruction in echocardiography, incident heart failure or septic shock and absence of cardiac surgery were significantly and positively associated with in-hospital mortality (Tables 1 and 3). In the multivariate analysis, the significant risk factors of in-hospital mortality identified in the final model were Staphylococcus aureus etiology, blood-culture negative endocarditis, evidence of valve obstruction in echocardiography and clinical evolution with heart failure or septic shock. Cardiac surgery was a protective factor of in-hospital mortality (Table 4). The model 2 that included Streptococcus gallolyticus organism had a numerically lower predictive performance and is described in table 5. Discussion The factors associated to increased risk of in-hospital mortality in our cohort were: development of heart failure or septic shock, valve obstruction in echocardiography, Staphylococcus aureus etiology, blood-culture negative endocarditis and absence of surgical treatment. The in-hospital mortality rate observed was 31.2%, which is slightly higher that the reported in the literature (15-30%). 3-9 It is recognized that one of the main protective factors of mortality is cardiac surgery and it was significant in our cohort. 3,7,11-13 Differently from other studies in which 40–50% of patients undergo cardiac surgery, 4,6,8,11,13,14 in our center only 32.8% underwent cardiac surgery. This can be partially justified by the absence of Cardiac Surgery Department in our center, which can difficult and delay the appropriate discussion with cardiac surgeons, and subsequently it may negatively influence the in-hospital mortality rates. The association of mortality with other factors, such as septic shock and heart failure found in our cohort is well known and expected. 3,5,8,13 The microbiological factors that increased the risk of in-hospital mortality were the expectedly Staphylococcus aureus -related endocarditis 8,15 and blood-culture negative endocarditis 14 (possibly due to the difficulty in the diagnosis and administration of timely and directed therapy in the latter group of patients). Valve obstruction was associated with higher mortality and in half of the patients was related to prosthesis degeneration, followed by the presence of large vegetations. In both etiologies, valve obstruction could contribute to clinical patient worsening, namely with heart failure, with congestive symptoms or low cardiac output, that could lead to multiple organ dysfunction and death. The aortic valve was the most affected (57.5%), differently to other series in which the mitral valve was the most affected. 9 Right-sided IE was observed in 14.2%, a value higher than the 5–10% reported. 11,16 This could be due to the higher incidence of drug users (13.4%), compared to other series, 3,7-9,17 which could be justified by the cultural and social characteristics of our population, and could also contribute to the higher mortality rate observed. 5