Elderly versus non-elderly patients undergoing surgery for left-sided native valve infective endocarditis: A 10-year institutional experience

This retrospective study aimed to clarify the short- and mid-term outcomes of elderly patients who underwent surgery to treat left-sided native valve infective endocarditis (LSNIE). Between July 2005 and September 2015, 179 patients underwent surgical treatment for active LSNIE at a single institution. Patients were classified into two groups: ≥65 years (elderly group) and <65 years (non-elderly group). Clinical features, surgical information, postoperative complications, and three-year survival rates were compared. The average ages were 74.2 ± 6.4 and 45.2 ± 12.6 years in the elderly and non-elderly groups, respectively. The elderly group had a higher predicted mortality rate and a lower incidence of preoperative septic emboli-related complications. Echocardiographic assessments of infected valves were generally homogenous between the groups. The elderly patients had a higher in-hospital mortality rate than the non-elderly patients (26.3% vs. 5.7%, P = 0.001). For patients who survived to discharge, the three-year cumulative survival rates were 75.0% ± 8.2% and 81.2% ± 3.4% in the elderly and non-elderly groups, respectively (P = 0.484). In conclusion, elderly patients are at a higher risk of in-hospital mortality after surgery for LSNIE. However, once elderly patients are stabilized by surgical treatment and survive to discharge, the mid-term outcomes are promising.

Each patient fulfilled the modified Duke criteria for IE 12 , and surgical treatment was indicated based on published guidelines prior to 2015 13,14 . General cardiac function, valvular destruction severity, and vegetation size/location were assessed by an experienced cardiovascular physician using transthoracic or transoesophageal echocardiography. Active endocarditis was defined as definitive IE requiring intravenous antibiotic therapy until surgery 15 . After excluding 23 patients with sub-acute IE that resolved after a complete antibiotic treatment course and underwent delayed surgery due to residual valvular insufficiency found by follow-up echocardiography, and 17 with prosthetic valve IE, 179 patients were included. The patients were classified into two age groups: ≥65 years (elderly group, n = 38, 21.2%) and <65 years (non-elderly group, n = 141, 78.8%) according to previous definitions of elderly 10,16 . Preoperative septic emboli-related complications were diagnosed according to clinical symptoms and image studies, including computed tomography and magnetic resonance imaging, which were interpreted by experienced radiologists. If the vegetations were located in the aortic valve, preoperative coronary angiography was avoided to prevent new systemic embolic events. For patients with unstable haemodynamics or cardiopulmonary failure, depending on the inotropic agents and ventilator support, surgery was performed on an emergency basis.
Surgical management. Intraoperative transoesophageal echocardiography was performed by specialized cardiovascular anaesthesiologists to confirm the preoperative diagnosis and extent of the valvular lesion. Using a standard sternotomy approach, cardiopulmonary bypass (CPB) was established via cannulation of the ascending aorta and right atrium or the venae cavae. Cardiac arrest was induced by a single dose of histidine-tryptophanketoglutarate solution or intermittent cold-blood cardioplegic solution through the coronary orifice. The diseased valves were carefully inspected, and the feasibility of valvular repair was considered after complete debridement of the infected tissues. The decision of valve repair or replacement was dependent on the surgeon's discretion regarding the degree of valvular destruction and patient characteristics, including heart function, comorbidities, and preoperative haemodynamics. For infected mitral valves that were repairable, the Carpentier principle was applied 17 , and ring annuloplasty was used to ensure the long-term durability of repair. In Regards to infected aortic valves that were repairable, a pericardium patch was usually used for reconstructing the valvular defects following debridement. If valve replacement was necessary due to limited healthy tissue, the prosthetic choice was based on the patient's individual preference after a detailed discussion before surgery. Before patients were weaned off CPB, haemostasis and good competency of the treated valve were confirmed, and a global assessment of cardiac function was performed. postoperative care and interventions. After undergoing surgery for endocarditis, all patients were transferred to a specialized cardiovascular intensive care unit (ICU) for further treatment and observation. At 8 hours post-surgery, a ventilator-weaning protocol was initiated if there was no active bleeding, unstable haemodynamics, persistent arrhythmia, or signs of organ malperfusion. Renal replacement therapy was applied if acute renal failure developed after surgery, according to the Acute Kidney Injury Network criteria 18 . Further survey imaging, endovascular intervention, and surgical exploration for bleeding or malperfusion were performed, if indicated. Statistical analyses. Statistical analyses were performed using SPSS for Windows (version 22.0, IBM Corp., Armonk, NY). The Shapiro-Wilk test was used to test whether the continuous variables were normally distributed. Data were presented as means ± standard deviation for normally distributed variables, whereas medians and interquartile ranges were used to describe non-normally distributed data. The independent t test was performed for comparison of normally distributed data and the Mann-Whitney U-test was used for non-normally distributed data between the two study groups. Dichotomous variables are presented as numbers (n) and percentages (%). Univariate analyses of categorical variables were performed using the chi-square test, or Fisher's exact test to determine inter-group differences, where appropriate. Preoperative and surgical variables found to be significant in the univariate analysis of in-hospital mortality were included in a multivariate logistic regression analysis to identify the independent predictors of in-hospital mortality. Numbers of events (in-hospital mortality) per covariate fell below four or five in the present study; therefore, we used penalization through data augmentation to perform multivariate logistic regression to avoid sparse data bias 19,20 . The Hosmer-Lemeshow test was used to evaluate the goodness-of-fit for the multivariate logistic regression model 21 . The Kaplan-Meier method was used to estimate 3-year cumulative survival, which was compared using the log-rank test. For all analyses, statistical significance was set at P < 0.05.

Results
patient demographics. Table 1 shows the clinical demographics, comorbidities, preoperative conditions, and clinical presentation of the elderly and non-elderly groups. The average ages were 74.2 ± 6.4 and 45.2 ± 12.6 years in the elderly and non-elderly groups, respectively. Overall, 27.4% of patients were female, and diabetes mellitus was the most prevalent comorbidity, accounting for >20% of cases in both age groups, followed by hypertension and end-stage renal disease (ESRD). The average European System for Cardiac Operative Risk Evaluation (EuroSCORE) II estimated that mortality risk was higher in the elderly group (6.6% [4.1-11.7%] vs. 2.4% [1.8-4.1%]; P = 0.001) than in the non-elderly group. Bacteraemia was the most prevalent clinical finding, followed by fever and septic emboli-related complications in both the elderly and non-elderly groups. The elderly group had a lower incidence of septic emboli-related complications (23.7% vs. 42.6%; P = 0.034). There were no differences in the antibiotic courses and length of hospital stays before surgery between the elderly and non-elderly groups. The epidemiology of pathogens was generally similar, and streptococci accounted for >40% in both groups (Fig. 1).
preoperative echocardiographic assessment and surgical information. Table 2 provides detailed information regarding preoperative echocardiographic assessment and surgical variables. In general, there were no differences in the location of the diseased valves, size/number of vegetations, and occurrence of periannular abscess between the groups. Overall, 23.5% of the patients underwent emergency surgery. Regarding the valvular procedures, the rates of repair and replacement were similar in the elderly and non-elderly groups. However, more patients in the non-elderly group underwent mechanical valve replacement for both the aortic and mitral valve. The time span of CPB, aortic cross-clamp, and rate of histidine-tryptophan-ketoglutarate cardioplegic solution usage did not differ between the two groups. In total, 11.7% of patients underwent extracorporeal membrane oxygenation installation in the operating room due to intraoperative myocardial failure. postoperative complications. As shown in Table 3, a significantly higher in-hospital mortality rate was observed in the elderly group compared to the non-elderly group (26.3% vs. 5.7%; P = 0.001). Moreover, the elderly group showed a higher incidence of complications, including new cerebral infarction, prolonged ventilator dependence, and prolonged ICU course. The median hospital stays were 37.5 (16.8-53.0) and 37.0 (28.0-47.5) days in the elderly and non-elderly groups, respectively.

Discussion
An increasing number of elderly patients suffer from cardiovascular diseases, including IE, which can be life-threatening [7][8][9] . However, elderly patients tend to receive a more conservative treatment course due to concerns regarding surgical complications 10,11 . In this single-centre study, a comparative cohort of patients who underwent surgical treatment for active LSNIE is presented, which includes 38 elderly patients aged >65 years. In-hospital mortality rates and postoperative complication rates were higher in this population than in non-elderly patients. However, the mid-term outcomes were still satisfactory for elderly patients who survived to hospital discharge. Due to the increase in average life expectancy and the higher incidence of cardiovascular disease with advancing age, more elderly patients nowadays present for cardiac surgery 22 . However, advanced age is considered a predictor of adverse outcomes following cardiac surgery 23,24 . An increased but acceptable surgical mortality rate has been previously observed among elderly patients with IE 11,25 , which was also observed in our study. Therefore, surgery may be justified as a treatment option for elderly patients with LSNIE and should be considered in selected patients. Thorough consultation between the medical and surgical teams is essential to make this a reliable strategy.
Traditional surgery with CPB remains the standard treatment option for IE. However, the aging process is associated with structural and functional changes in various organ systems, possibly reducing the tolerance for haemodynamic fluctuation, and increasing coagulopathy and the systemic inflammatory response induced by CPB. Therefore, higher in-hospital mortality and complication rates are expected in this population. In the present study, the elderly group demonstrated a higher EuroSCORE II-estimated risk, which also addresses the inferior short-term outcome. The aging process is associated with numerous ionic, molecular and biochemical www.nature.com/scientificreports www.nature.com/scientificreports/ changes in the heart 22,26 . These age-related changes can affect cardiac morphology and reduce the physiologic function. Therefore, cardiac aging results in decreased mechanical and contractile efficiency, and may increase the rate of cardiomyocyte apoptosis after cardiac surgery. In our study, approximately 20% of elderly patients required mechanical support following surgery, and >60% of the mortality in the elderly group was associated with postoperative myocardial failure. Although the rates of postoperative mechanical support were similar in the two groups, the aging heart may exhibit inferior recovery compared with younger hearts.
Septic emboli-related complications occur in 20-50% of patients with IE, especially when the diseased valve is on the left side 27,28 . The central nervous system is the most common destination of embolism, followed by the spleen, kidneys, lungs, and liver. In our study, the elderly group had a relatively lower incidence of preoperative septic emboli-related complications than the non-elderly group. We suggest several reasons for this finding. First, contrast-enhanced imaging may have been arranged more conservatively for patients in the elderly group due to the lower reserved renal function. Second, the symptoms and presentations of emboli-related complications in the elderly may not be apparent and, therefore, could have been overlooked. When the emboli-related complications have progressed to a serious stage, the timing of surgical treatment may be delayed, and the outcome is compromised. Finally, no patients in the elderly group experienced preoperative intracranial haemorrhage or cerebral mycotic aneurysm. These two cerebral complications could increase the surgical risk and induce permanent neurologic deficits. According to previous studies 15,29 , 13% of hospital mortality and 24% of postoperative cardiac and cerebrovascular complications were reported in this high-risk subgroup. As these complications are extremely serious and life-threatening, elderly patients with these conditions may be reluctant to undergo surgery.
The elderly population is associated with a high incidence of comorbidities, including chronic kidney disease (CKD), and relative risk factors, including hypertension and diabetes, which predispose them to CKD 30 . According to a previous multicentre study 31 , patients with CKD had an increased perioperative mortality rate and reduced long-term survival after cardiac surgery. In our study, the elderly group showed worse renal function compared to the non-elderly group before surgery (49.2 (24.6-74.6) vs. 86.6 (55.9-126.7) ml/min/1.73 m 2 ; P = 0.001). Furthermore, ESRD was one of the independent predictors of in-hospital mortality for the elderly  www.nature.com/scientificreports www.nature.com/scientificreports/ these influences can be magnified. Therefore, we suggest that elderly patients with CKD who present with clinical signs of IE should undergo early assessments and receive careful surgical management to optimize the outcomes.
Elderly patients who underwent surgery for IE were associated with higher in-hospital mortality and morbidity rates owing to their comorbid conditions and delayed timing of surgery. However, the late surgical outcomes are rarely reported. Similar short-term results were presented in our study. Furthermore, elderly patients who survived to discharge could have a comparable 3-year survival to that of the non-elderly group. Therefore, we suggest that a guideline-directed surgical strategy according to the presence of complications, which include embolism events, large vegetation, heart failure, or uncontrolled infection, would be beneficial to improve the mid-term outcomes of elderly patients with LSNIE.
Limitations of this study. Our study had several important limitations. First, because the study used a retrospective and nonrandomized control design with a small sample size, bias might exist that influenced the homogeneity of the study groups and the stability of the multivariate logistic regression model. Second, as this cohort spanned a period of >10 years, the technology of CPB and myocardial protection, as well as strategies for