The Influence of Preoperative Neurological Complications on Outcomes after Surgery for Infective Endocarditis

,


Introduction
Infective endocarditis (IE) is considered a life-threatening cardiac infection with a predilection to involve heart valves [1].In 2019, the global incidence of IE accounted for approximately 1.1 million cases with an annual mortality exceeding 66,300 deaths.Tis burden showed a steep increase in the past three decades.Results from the global burden of disease study showed an annual percentage increase of 1.2% for IE incidence and 0.7% for its associated mortality [2].Interestingly, an annual percentage increase of 4.1% was reported for Europe during the last two decades [3].
Cardiac surgery is indicated in around 40-60% of cases [4,5].In patients where cardiac surgery is indicated but not performed, the thirty-day and one-year mortality rates were as high as 60% and 75%, respectively [4].Surgery is indicated in cases of IE for conditions like heart failure, uncontrolled infection, large vegetation, and as a preventive measure against embolic complications [6,7].Several factors can prohibit or delay indicated surgery in patients with IE including advanced age, female gender [8][9][10], afection of the mitral valve, high surgical risk, and neurological complications (NCs).However, the current guidelines recommend early surgery during initial admission and before antibiotic completion for indicated patients [6,7].Several reports concluded that IE patients presenting with neurological complications (NCs) would have signifcantly higher inhospital and long-term mortality [11][12][13][14][15].In this article, we evaluate a large cohort of IE patients who underwent surgical repair in our tertiary hospital.Te aim of this study is to compare the short-and long-term outcomes in patients who underwent cardiac surgery for IE based on the presence of preoperative NCs.

Patient Population and Study
Design.We retrospectively reviewed patient records which were regularly collected in our institutional database.Between January 2002 and August 2020, 438 consecutive patients who underwent open cardiac surgery in our university hospital due to IE were included in the retrospective study.Patients presented with neurological complications were 89, while 349 patients underwent surgery with no evidence of NCs.NCs were defned as preoperative stroke, transient ischemic attack, or the presence of silent cerebral embolization.
We review and compare baseline characteristics and outcomes of patients who presented with NCs in contrast to those who had no NCs.Te 30-day mortality rate was the primary endpoint outcome.Long-term survival, intraoperative details, and postoperative clinical outcomes such as redo surgery, blood loss, ventilation time, acute renal failure, and NCs were secondary endpoints.Patients with active IE required ongoing antibiotic therapy.

2.2.
In-Hospital Patient Management.All patients were discussed among a multidisciplinary team.Diagnosis of IE was confrmed using the modifed Duke criteria.Blood cultures were obtained from all patients in order to aid diagnosis and identify organisms based on species and sensitivities.All patients had a transthoracic or transesophageal echocardiography to determine the location and size of vegetation, the presence of valve destruction or abscess, and the left ventricular ejection fraction.Antibiotic treatment began as soon as IE was suspected.If the patient's condition was stable, coronary angiography and additional computed tomography (CT) scans were performed, including brain CT, thoracic CT, and pan CT scans, especially in high-risk patients or those who required redo surgery.Ten, patients were referred to our surgical department and scheduled for surgery.If intraoperative evidence of endocarditis was identifed, an intravenous antibiotic regimen was continued for 4-6 weeks postoperatively.All patients with NCs were evaluated by a consultant neurologist, as well as a CT scan of the brain to assess bleeding risks and prognosis.Perioperative risk factors, intraoperative details, and predictors of mortality were all analyzed and reviewed.

Surgical Technique.
Te surgical technique used in our university hospital has been previously described [16,17].
All patients underwent curative surgery performed by senior faculty surgeons and experienced assistants.Direct cannulation of the ascending aorta was used to establish cardiopulmonary bypass (CPB).Venous drainage was performed either by direct cannulation of the right atrium in situations of aortic valve endocarditis or with double cannulation of the superior and inferior vena cava followed by cross-clamping of the ascending aorta in cases of mitral or tricuspid valve infection.Te choice of biological, mechanical prosthesis, or valve repair was made on an individual basis considering patients' age, preferences, and their compliance with long-term anticoagulation, as well as intraoperative details and the degree of macroscopic valve damage.Te diseased areas were extensively debrided and thoroughly irrigated.In situations of signifcant damage or the presence of abscesses, the defective regions were restored via isolated valve replacement or aortic root reconstruction with or without autologous or bovine pericardium.Te standard technique for cardiac de-airing was continuous CO 2 insufation.Transesophageal echo was used for postoperative evaluation and to assess for any residual air during rewarming.

Statistical Analysis.
We employed descriptive statistics to summarize patient baseline characteristics, operative details, and postoperative outcomes throughout the study.Te normality of continuous variables was assessed by the Kolmogorov-Smirnov test.We used median and interquartile ranges (IQRs) to depict normally and nonnormally distributed continuous data.Categorical variables are shown as frequency distributions (n) and percentages (%).We used Student's t-test to test for group diference when the data were normally distributed.For nonnormally distributed data, we used the Mann-Whitney U test.For categorical variables, the chi-square and Fisher's exact tests were used to provide a univariate comparison between groups.All tests were two-tailed, and P values less than 0.05 were deemed statistically signifcant.IBM SPSS Statistics for Windows (version 29.0) was used to examine the data.Variables associated with 30-day mortality (P ≤ 0.1) were selected according to clinical relevance and included in multivariable logistic regression analysis to assess for their adjusted impact on developing NCs.We used adjusted odds ratio (OR) as the efect measure.In multivariable logistic regression, we controlled for age, female gender, BMI >30, arterial hypertension, atrial fbrillation, acute renal insufciency, combined valve surgery, previous endocarditis, diagnosis on day of surgery, Staphylococcus aureus, mitral valve afection, and vegetation.Survival of the NC and non-NC groups was estimated on right censored data by Kaplan-Meier curves and was compared for diferences by the log-rank test.
In contrast, patients in the non-NC group had a higher rate of previous cardiac surgery (45.0% vs. 27.0%;P � 0.002) and were more likely to have a history of combined valve surgery (22.8% vs. 9.3%; P � 0.005).Tey also had a higher incidence of previous endocarditis episodes (16.6% vs. 5.6%; P � 0.008) and more often arterial hypertension (61.0% vs. 49.4%;P � 0.047).
Patients in the NC group were more likely to be admitted to ICU at presentation (36.0% vs. 18.3%;P < 0.001) and had a higher rate of preoperative embolization (92.1% vs. 11.7%;P < 0.001).Staphylococcus aureus was the causative organism of IE more frequently in the NC group (35.2% vs. 16.1%;P < 0.001).

Follow-Up.
Follow-up data were available for 326 patients in the non-NC group (93.4%) and for 83 patients in the NC group (93.3%).For patients who survived thirty days following the indexed operation, follow-up data were available for 273 patients with no NCs (92.9%) and for 63 patients in the NC group (92.6%).Survival probabilities were similar between both groups at one (90% vs. 93%), fve (73% vs. 78%), and ten (56% vs. 62%) years of follow-up in the 30day survivor cohort.Figures 1 and 2 illustrate the Kaplan-Meier curves for the total cohort and for patients who survived 30 days following the indexed procedure.

Discussion
Infective endocarditis remains a serious and potentially lifethreatening condition.One of the most feared complications of IE is the development of new-onset NCs, with embolic events being the most likely cause.In our study, we aimed to evaluate the impact of NCs on clinical outcomes in patients who underwent surgery for IE.Our results showed that approximately one-ffth of the patients who underwent surgery for IE at our institution had NCs.Patients with NCs were more often female, had a higher incidence of AKI, required more acute dialysis, and were more likely to be admitted to ICU at presentation.Furthermore, NCs were associated with multiple organ embolization as well as Staphylococcus aureus as IE causative organism and more often presented with afection of the mitral valve.Consequently, mitral valve surgery was performed more often in the NC group.Patients with no NCs were more likely to have undergone previous cardiac surgery, received combined valve surgery, experienced previous endocarditis episodes, and presented more frequently with prosthetic valve endocarditis.However, there was no signifcant diference in postoperative length of ICU; moreover, hospital stay was shorter in the non-NC group.In-hospital mortality rate was similar, but 30-day mortality was slightly higher in the NC group.
In a majority of studies, the incidence of NCs in patients with IE who are referred to surgery is reported to range from 20% to 30% [12,[18][19][20][21], with some studies reporting rates as high as 63% [14].Additionally, ischemic events have been identifed as the most common presenting complication [18][19][20][21][22]. Similarly, 20.3% of our patients have experienced preoperative NCs.Multiple studies [12,13,[21][22][23] have concluded that developing preoperative NCs in patients with IE was consistently associated with higher mortality rates.In our study, we focus on NCs in patients who were indicated and underwent surgery for IE.In this subgroup of patients, we did not observe a signifcant diference in postoperative mortality.Interestingly, studies have also reported that mitral valve afection, large vegetations, and infection with Journal of Cardiac Surgery Staphylococcus aureus were independent predictors of NCs.Furthermore, Staphylococcus aureus, large vegetations, and mitral involvement were associated with a markedly increased risk of mortality [21][22][23][24].In an earlier analysis of patients in our cohort, the presence of valvular abscess, older age (>70 years), preoperative neurological defcits, and female gender were independent risk factors for mortality [17].In this updated analysis, although present, the diference in 30-day mortality between patients with NCs and those without did not reach statistical signifcance (23.6% vs. 15.8%;P � 0.08).In this analysis, focusing on NCs in patients who underwent surgery for IE, we noted that infection An earlier study by Diab and colleagues [24] concluded that NCs do not independently infuence outcomes after surgery and that patients presenting with NCs already have high-risk profle.We assume that factors such as mitral afection, larger vegetations, and Staphylococcus aureus infection represent key risks that contribute to both increased incidence of NCs and higher mortality.However, it should be noted that all patients in our study were indicated and underwent surgery.
Previous results showed that earlier antimicrobial therapy can substantially decrease the risk of embolization [12,25].Tis highlights that the early recognition and management of high-risk patients can signifcantly decrease both NCs and mortality rates.
Traditionally, NCs were considered as a major cause of surgery delay in indicated patients.Tis is mainly to avoid the risk of further exacerbation of NCs during heparinization and CPB, which may transform ischemic stroke into hemorrhage or exacerbate baseline hemorrhage [26].However, this is reversed in the contemporary practice, encouraging earlier surgery especially for patients with nonhemorrhagic NCs.Patients who did not receive surgery when indicated had signifcantly worse outcomes [20][21][22][23].Contemporary evidence has established that the most important prognostic factor in those patients is undergoing cardiac surgery when indicated [6,7,[20][21][22][23].Te frequency of female patients was higher in the NC group compared to the non-NC group in our study and, as we had reported previously, there are indications of a delay in the presentation of female patients in our endocarditis cohort [27] and in general [8,9,28].Te higher mortality in female patients [27] has probably contributed to the slightly higher mortality in the NC group in our mixed cohort.However, female gender was not an independent predictor of NCs in our multivariable analysis.
Te aortic valve is most commonly aficted in IE followed by the mitral valve.Our analysis showed a signifcant association between mitral valve involvement and increased risk of NCs.Although patients with prosthetic valve IE present more commonly with intracardiac abscess [16], we observed that patients with native valve IE were at a signifcantly higher risk of NCs.
Tese fndings highlight the importance of considering the location and type of valve afected in IE complicated by NCs.
While a considerable number of patients with IE can have silent brain embolization [26,29], we did not perform routine brain CT scans in all patients, which could have precluded detection of silent NCs and might lead to misclassifying them as patients with no NCs.In our cohort, four patients exhibited radiological evidence of NCs, while lacking any clinical manifestations.Tis small number of patients posed a challenge in predicting outcomes in this subset of patients.Still, Cooper et al. [29] showed that early surgery for those patients was associated with signifcant reduction in three-month mortality (OR: 0.1; 95% CI [0.03-0.6];P � 0.008).
Interestingly, we noted that previous endocarditis was protective from developing NCs.Te protective efect of a previous IE on NCs in subsequent episodes is not fully understood.We hypothesize that initial exposure to infective endocarditis may trigger immune responses or adaptive changes, which may enhance the immune system's ability to defend against or restrict the formation and embolization of vegetations.However, future research might fully elucidate such observation.
It is imperative to consider several limitations when interpreting our results.Firstly, our study is a retrospective observational study, which introduces the possibility of inherent confounding bias and may limit the generalizability of our fndings.Secondly, our cohort only includes IE patients who were indicated for and underwent surgery, limiting the generalizability of our conclusions to a broader population of IE patients.Moreover, the complex and varied nature of patient pathology in IE poses challenges in defnitively assessing the specifc impact of NCs on outcomes.It is crucial to take into account these limitations when interpreting our results.
Our fndings provide insights into the risk factors and outcomes associated with NCs in our subset of patients who underwent surgery for IE.Moreover, we highlight that postoperative outcomes are infuenced by a range of factors and not solely-rather indirectly-dependent on the presence of NCs.
Further research, including prospective studies, is warranted to better understand the infuence of NCs on the outcomes of surgery in IE patients.

Conclusion
In conclusion, our study highlights the risk factors associated with NCs and their impact on the outcomes of surgery for patients with IE.We have identifed multiple risk factors, such as organism type and preoperative AF, which are associated with the development of NCs.Furthermore, these risk factors also serve as predictors of mortality.However, we note that NCs alone do not independently predict shortor long-term survival following surgery for IE.

Table 1 :
(a) Demographic characteristics of the study population.(b) Clinical presentation data.(c) Specifc endocarditis data.(d) Preoperative laboratory parameters.

Table 3 :
Postoperative data and outcomes.