Infective Endocarditis, Predictors of Mortality and Morbidity


 Objectives:

Infective endocarditis is a disease with high mortality and morbidity. The primary outcome of this study is to assess factors associated with in-hospital mortality in patients undergoing valvular surgery for infective endocarditis. The secondary outcome of this study is to assess the incidence and factors associated with post-operative morbidity; namely post-operative stroke, renal failure and dialysis, complete heart block and return to theatre for bleeding or tamponade.
Methods

Between the years of 2015 to 2019, a total of 89 patients underwent surgery for infective endocarditis at Fiona Stanley Hospital. Data was collected from the Australia and New Zealand Cardiac Surgery Database from 2015 to 2019 as well as patients electronic medical record (EMR). A number of preoperative and perioperative factors were assessed in relation to patient mortality and morbidity. Univariate and multivariate logistical regression analysis was done to assess for the association between factors and in-hospital mortality and morbidity
Results:

A total of 89 patients underwent surgery for infective endocarditis, affecting 101 valves. The mean age of patients was 53.7. A total of 79 patients had a positive blood culture pre-operatively, with Staphylococcus Aureus being the most frequently cultured organism (39%). Fourteen patients (16%) were deemed emergent and underwent surgery within 24 hours of review. A total of five patients died within their hospital stay postoperatively. Variables significantly associated with mortality on univariate analysis were intravenous drug use, emergent surgery, perioperative dialysis, perioperative inotropes, cardiopulmonary bypass (CPB) time and cross clamp time (CCT). Only CBP time was significantly associated with mortality on multivariate analysis. A total of 19 patients (21%) required hemodialysis after surgery, 10 patients sustained a postoperative stroke (11%), 15 patients required to return to theatre (17%) and 11 patients developed a complete heart block post operatively (12%).
Conclusion

There are a number of factors associated with mortality and morbidity in patients undergoing surgery with infective endocarditis. Our study demonstrates a lower mortality rate in these patients than previously quoted in literature. Exposure of prolonged CBP times was the only factor significantly associated with increased mortality on multivariate analysis, although a critical perioperative state was highly significant on univariate analysis.


Introduction
Infective endocarditis is associated with a high mortality and morbidity. The mortality rate in studies has been quoted from 6-25% [1][2][3][4][5] . The risk factors associated with morality in infective endocarditis have been evaluated. , [2][3][4][5][6][7][8][9][10][11][12] . Infective endocarditis is also associated with signi cant morbidity. Patients can sustain complications such as embolization (either systemic or pulmonary), heart failure and cardiogenic shock, disseminated infection, abscess formation and arrythmias including complete heart block. Factors associated with these complications have also been assessed in literature, although to a lesser extent 2,11,13,14 . Emergent surgery is required in 25-50% of these cases 4,7 . The European Society of Cardiology recently published guidelines outlining the indications for surgery 2 . Early or emergent surgery is indicated for patients with infective endocarditis and cardiogenic shock, as well as locally uncontrolled infection. Early surgery is also indicated in patients with high-risk vegetations. Grey areas exist with regards to timing of surgery, especially with respect to surgery in the setting of a pre-operative embolic stroke 2 . The concept of a 'Heart Team' comprising of various specialties to create management plans for individual cases has been shown to decrease mortality in these cases 2 . The primary outcome of this study is to assess factors associated with in-hospital mortality in patients with infective endocarditis undergoing surgery. The secondary outcome of this study is to assess the factors associated with morbidity; namely post-operative stroke, renal failure and dialysis, complete heart block and return to theatre. The identi cation of prognostic factors will be useful for risk assessment and surgical decisionmaking.

Variables
Data was retrospectively collected from the Australian and New Zealand Cardiac Surgery database (ANZSCTS). Between the years of 2015 to 2019, a total of 89 patients underwent surgery for infective endocarditis at Fiona Stanley Hospital. This study focused exclusively on patients with endocarditis treated with open heart surgery. Cases of IE related to nonvalvular cardiovascular devices were excluded. Furthermore, cases of infective endocarditis were retrospectively con rmed using the modi ed Dukes Criteria and patients that did not meet these criteria were excluded 15 . A total of 9 patients were excluded.
Several preoperative and operative factors were identi ed and recorded through a combination of the ANZSCTS database and patients electronic medical record. These factors were selected as they are hypothesized to be associated with an increased rate of mortality and morbidity post surgery. blood cell use (RBC), non red blood cell use (NRBC) and aortic procedure were also recorded. These are summarized in Table 1 below. The de nitions of these factors are in accordance with those set by the ANZSCTS database.

Surgical Technique
All operations were performed through a median sternotomy and the use of Cardiopulmonary Bypass. Arrest was achieved using hyperkalemic cold blood cardioplegia. Intraoperative transoesophageal echocardiography was also routinely performed in all cases. The valvular procedure was documented from a combination of the ANZSCTS database and the patients electronic medical record. This includes whether the patient had an aortic procedure, aortic valve, mitral valve, tricuspid valve or pulmonary valvular procedure. The number of valves affected was documented, as well as whether the index operation was a valvular replacement or repair. When a prosthesis was used, The prosthesis type was recorded (mechanical or bioprosthetic). In patients who underwent valvular repair, details of the repair were recorded.

Outcomes
Two outcomes were considered. Outcome 1, in hospital mortality, was de ned as death following surgery during the patients initial hospitalization. This also included on-table mortality. Outcome 2 was de ned as the presence of one of four issues post operatively, namely a new cerebrovascular accident (CVA), post-operative hemodialysis, return to theatre (RTT) and new complete heart block (CHB). CVA post op was de ned as the occurrence of a stroke or new central neurologic de cit (persisting > 72 hours) post operatively. RTT was de ned as patients returning to theatre after the initial index operation, within the patients initial hospitalization period. CHB was de ned as the presence of new complete heart block post operatively. The presence/absence of these outcomes was obtained from a combination of the ANZSCTS database and the patients' electronic medical record.

Statistical analysis
Descriptive statistics including number of events, mean and standard deviation was calculated for each preoperative and operative variable. Univariate analysis was rstly conducted to identify variables signi cant associated with either in hospital mortality (outcome 1) or morbidity (outcome 2). Categorical variables were assessed using the chi squared (x 2 ) test to ascertain odds ratios. Fishers exact test was conducted when more than 20% of cells on the contingency table had an expected frequency less than ve. Continuous variables were rst assessed for normality and then with the independent T test to assess for equality of means. Variables that weren't normally distributed were assessed with the Mann-Whitney U test. P values less than 0.05 were deemed as signi cant. Preoperative and operative variables that reached signi cance were then further assessed with multivariate logistic regression analysis using a binary logistic regression model. Statistical analysis was done on IBMM ® SPSS statistics version 25.

Preoperative Factors
A total of 89 patients underwent surgery for infective endocarditis, affecting 101 valves. The mean age of patients was 53.7, with a minimum age of 16 and a maximum age of 83. Most of the patients were male (n = 67). A large portion of patients (n = 28) had a history of IVDU. In terms of location, 33 patients were from a rural setting. Eight patients were from the South West region, 5 from the Pilbara and 3 from the Kimberley. Fifteen patients identi ed as ATSI. Thirteen patients (15%) were NYHA class 4 preoperatively. The majority of patients (52 patients) were NYHA class 1 preoperatively. Twenty-seven patients had vegetations greater than 20mm on echocardiography, 32 patients had vegetations between 10-20mm and 30 were less than 10mm. The majority of patients who underwent surgery had left-sided disease (82 patients) with 7 patients undergoing surgery for right-sided disease. A total of 79 patients had a positive blood culture pre-operatively, with Staphylococcus Aureus being the most commonly cultured organism (39%). Other common organisms include Enterococcus Faecalis (20%) and Streptococcus Mitis (9%). Ten patients (11%) had culture negative endocarditis. These results are summarized in Table 2. Embolic phenomena were present in 39 patients (44%). The most common site of embolization noted was brain, with 23 patients. Other common sites were skin (6 patients), lungs (5 patients) and spine (4 patients). Two patients had concomitant septic arthritis. Ten patients had multiple sites of embolization. These results are summarized in Table 3.  These results are summarized in Table 4.   Table 5 below. Multivariate analysis revealed that only CBP time was a signi cant predictor of operative mortality, with an odds ratio of 1.05 per minute of additional bypass time (95% CI 1.001-1.101, P = 0.046).   [17][18] . These studies advocate for early surgery in patients presenting in heart failure [17][18][19] . Early surgery for high risk lesions is also supported by literature [20][21][22] . Of these, a randomized control trial by Kang et al demonstrated that early surgery in patients with large left sided lesions (> 10mm) signi cantly reduced morbidity and embolic events 21 . Therefore, the ESC guidelines provide a class 1 indication for early surgery in vegetations greater than 10mm with ongoing embolic phenomena. Uncontrolled infection is a further indication for early surgery. This is supported by evidence, primarily consisting of retrospective cohort studies, demonstrating that locally aggressive infection is associated with a higher mortality rate 10,23 . Of these, a retrospective study by Revilla et al demonstrated that persistent infection is an independent predictor of mortality, where patients who undergo urgent surgery with persistent infection are four-fold as likely to die as patients without persistent infection 10 . At Fiona Stanley hospital, we adopted these guidelines to help with decision making regarding operative timing.
In the current study, the in-hospital mortality rate was 5.6% or 5 out of 89 patients. This is at the lower end of the spectrum of mortality gures quoted by other studies 3,4,6,9,10 . A similar study conducted by Rivas de Oliveira assessed 88 surgical patients between 2005 and 2015, and reported an in-hospital mortality rate of 17% 3 . A study by Dunne et al in a similar Western Australian population with infective endocarditis, treated surgically, reported a mortality rate then of 13% 11 .
One major change reported amongst hospitals during the last decade is the establishment of a dedicated "heart team". This team comprises of Cardiac Surgeons, Cardiologists and Infectious Diseases physicians. A dedicated "heart team" was established at Fiona Stanley Hospital since its initiation in result of a multidisciplinary team (MDT) approach to endocarditis 24,25  Our study identi ed that IVDU, emergent surgery, perioperative dialysis; perioperative Inotropes, prolonged CPB time and prolonged CCT were signi cantly associated with in hospital mortality on univariate analysis. This nding is consistent with previous studies 1,3,6,11 . CPB time was the only factor to be signi cantly associated with death on multivariate analysis, with a mean CBP time of 250.8 vs 130.5 minutes for non-survivors and survivors respectively. Prolonged CPB time is a re ection of operative complexity, predisposes patients to end organ dysfunction, coagulation disorders and is therefore understandably associated with mortality.
In this study, factors associated with a critical perioperative state were strongly associated with postoperative mortality. Of these, perioperative dialysis was one of the major predictors of post-operative mortality, with an in-hospital mortality rate of 57% in this cohort. None of these patients required dialysis prior to their presentation with endocarditis. Studies of patients requiring perioperative dialysis, although chronic, reported a mortality rate of 40% 26 . Elevated peri-operative creatinine clearance is also associated with increased post-operative mortality however did not reach signi cance in this study 3 . Likewise, perioperative inotrope requirement was also associated with mortality post operatively, and re ects the challenges associated with operating on patients in shock.
Embolic phenomena occurred 39 patients (43.8%). The most common site of emboli was the brain (22 patients) followed by skin and lungs. Other studies have also quoted equally high rates of embolic events 10,27 . Likewise, in these studies, the brain was the most common site of embolism 10,27 . Preoperative stroke is a highly relevant complication of infective endocarditis due to the risk of hemorrhagic transformation and postoperative neurological deterioration. Guidelines provide class 2A evidence to delay surgery by a month in the presence of intracranial haemorrhage 2 . As a result, we adopted a low threshold to conduct a CT brain, explaining the higher rate of cerebral emboli compared to other sites in this study. Embolic phenomena and cerebral emboli were linked to the incidence of preoperative stroke on univariate analysis, however was not associated with in-hospital mortality.
In terms of organism, Staphylococcus Aureus was most commonly cultured and present in 39% of patients. This was followed by Enterococcus Faecalis and Streptococcus Mitis in 20% and 9% of patients respectively. Eleven percent of patients had culture negative infective endocarditis. The prevalence of Staphylococcus Aureus is a feature in other studies also 6,10, . There has been a reported shift in the epidemiology of infective endocarditis, away from Streptococcus species and HACEK (Haemophilus species, Aggregatibacter species, Cardiobacterium hominis, Eikenella corrodens and Kingella) organisms towards Staphylococcus Aureus 28, 29 . This was also evident in our study, with only 15 patients culturing Viridians Streptococci. There was one case of HACEK endocarditis. Staphylococcus Aureus has been linked to a higher mortality rate in surgically treated endocarditis 30,31 . It is also linked to locally aggressive infection, higher rates of embolization and septic shock 30,31 .Other studies do not demonstrate a relationship between Staphylococcus Aureus and mortality or morbidity 1,3,6,10, . Our study also did not demonstrate a relationship between Staphylococcus Aureus and in-hospital mortality as well as postoperative complications. Our institution favors early surgery for patients with Staphylococcus Aureus endocarditis.
Complications after surgery for infective endocarditis were not uncommon. Ten patients (11%) had a postoperative stroke. Identi able risk factors were cerebral emboli, pre-operative creatinine, perioperative cardiogenic shock, perioperative respiratory failure, perioperative ionotropic requirement and emergent procedure. Other studies have demonstrated a similar incidence of post-operative stroke 10,11,27 . Only one other study investigated risk factors associated with post-operative stroke 11 . Post-operative stroke is a debilitating issue, and some centers advocate for delaying surgery to minimize the risk of hemorrhagic transformation 22,31 . Others demonstrate that the overall mortality bene t from early surgery outweighs this risk 33  postoperatively. On multivariate analysis, cardiogenic shock and pre-operative dialysis were independently associated with the incidence of post-operative dialysis. Post-operative renal failure is linked to a critical perioperative state and is associated with an increased risk of mortality 6,10,34,35 .
Conduction abnormalities are an early indication of an infectious process expanding to involve the membranous interventricular septum, often in cases with aortic valve endocarditis. A total of 11 patients (12%) had complete heart block, all of whom received a pacemaker. The incidence of which is comparable to that published in other studies 6,36 .
This is a retrospective observational study with inherent biases in data collection. A larger prospective study may enable us to explore more factors associated with mortality and morbidity. Our small patient numbers and the small number of in-hospital deaths have limited the use of multivariate analysis to evaluate risk factors for in-hospital mortality. Fiona Stanley Hospital is a new institution, and data is available over a period of 4 years. As a result, long term morbidity and survival data was not explored by this study and therefore Kaplan-Meier survival data was not conducted. A long term follow up of our patients would be bene cial to assess whether the low in-hospital mortality rate is also translated into long term survival.

Conclusion
This study reports the morbidity and in-hospital mortality of 89 patients undergoing valvular surgery for infective endocarditis at a single institution. It demonstrates that our in-hospital mortality rate fares well compared to other published data. We adopt a MDT approach to the management of infective endocarditis, which is attributable to the lower mortality rate. A number of factors were associated with mortality: IVDU, emergent surgery, perioperative dialysis; perioperative Inotropes, prolonged CPB time and prolonged CCT, with prolonged CPB times reaching signi cance on multivariate analysis. Of these factors, a perioperative hemodialysis requirement was strongly associated with post-operative mortality.
A prolonged CPB time reached signi cance on multivariate analysis. Post-operative complications were not uncommon, with 11% of patients sustaining a post-operative CVA, 21% requiring post-operativedialysis, 11% requiring return to theatre and 12% demonstrating a complete heart block. A larger prospective study may enable us to explore more factors associated with mortality and morbidity. Ethics approval was granted from the Hospitals review board (approval number 33939)

Consent for Publication
Not applicable. There was no personalised patient information.

Availability of data
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests Funding Not Applicable

Author Contributions
The corresponding author AE was involved with data collection, analysis and writing of the paper, the authors AS, UA and KS were involved in the subsequent editorial process.