Outcome of patient undergoing redo mitral valve surgery with incidence rate of mitral valve infective endocarditis

Abstract

Background

The incidence of infective endocarditis (IE) in patients undergoing redo mitral valve (MV) surgery was evaluated. The outcomes of all the patients and the patients’ specific characteristics were recorded. The patients were analyzed to further the research of IE in this population.

Method

This was a retrospective review of patients admitted for redo MV surgery with a prospective follow-up of electronic medical records at King Faisal Specialist Hospital and Research Center, Riyadh, Saudi Arabia, from 2009 to 2019. Pre/intra/post-operative factors contributing to mortality, morbidity, and freedom of adverse events were analyzed.

Result

A total of 211 patients underwent redo MV surgery, and 41 patients (19.4%) had IE; and 51% of this subset of patients, 21 individuals, developed IE after the initial MV surgery. MV stenosis was moderate/severe in 50 patients. Furthermore, MV regurgitation was present in 89 patients. Multivariate analysis of the data revealed multiple factors influencing mortality: age, peripheral vascular disease, concomitant procedures, peripheral vascular disease, red blood cell transfusions, preoperative mechanical valves, and active IE. In-hospital Mortality was 10.9%. The one-, five-, and ten-year survival was 88%, 79%, and 69% across all patients.

Conclusion

Although redo MV surgery has acceptable outcomes; the presence of IE or concomitant procedures is a significant health detriment in these patients. Our study highlights the need for careful patient management and more in-depth research in this area to improve patient outcomes.

Introduction

Redo Mitral Valve (MV) surgery is the standard of care for patients with history of MV diseases, such as bioprosthetic dysfunction or failed MV repair, after their initial cardiac surgery [1, 2]. The complications of redo MV replacement include increased risk for major adverse events and 30-day mortality; an independent predictor in the STS risk models is active infective endocarditis (IE), particularly in patients with prosthetic valve endocarditis (PVE) [3, 4]. For patients presenting with IE, cardiac surgery is indicated if presenting with congestive heart failure, prevention of systemic embolization, or uncontrolled infection [5, 6].

Patients with IE undergoing MV surgery experience worse immediate/midterm morbidity and mortality rates compared to those undergoing other MV surgeries [3, 7,8,9]; in particular, PVE patients have a 24–35% mortality rate in cardiac surgery patients [10]. The prevalence of IE is 2–8 patients per 100,000 cases, and it is expected to rise further due to prosthetic valve usage in cardiac procedures [7]. PVE constitutes 10–30% of all IE cases and affects 6% in patients with prosthetic valve [10]. The cardiac valves commonly affected in PVE include the aortic, mitral, and multiple valves with frequencies of 66.5%, 40.7%, and 7.2%, respectively [11]. The current literature for MV redo surgery presents a wide range of outcomes for the 30-day mortality rates. The values range from 1.3 to 17.4% [12, 13]; and if all patients are included without exclusions, especially those undergoing concomitant procedures, reaching 22.4% for the 30-day mortality rate [1].

The authors hypothesized that the main factors predicting mortality are IE (especially active IE) and the presence of any concomitant procedure. This research investigated the outcomes, prevalence of IE, and quality of life of all patients undergoing redo MV replacement at a quaternary hospital.

Materials and methods

Study population

The study retrospectively reviewed and prospectively followed up on the electronic medical records (EMR) of all 211 qualified patients admitted for redo MV surgery at King Faisal Specialist Hospital and Research center (KFSHRC), Riyadh, Saudi Arabia. The duration of antimicrobial treatment was 6 weeks for patients with IE, with the appropriate regimen used based on the blood culture results. The indication for surgeries and antimicrobial therapy followed American Heart Association guidelines. IRB approval number 2,191,219 was obtained on 06/Aug/2019 by KFSHRC and the consent was waived by the research ethics committee. The EMR documented the valvular procedures, the type of procedure on the last intervened valve, the intervention type, EuroSCORE-II, STS score, and all relevant data from patients between 2009 and 2019 with a last follow up in 2021.

Outcomes

The primary endpoint was mortality. The secondary endpoints included IE incidence across MV types, MACE freedom (the absence of cardiac death, myocardial infarction, stroke, and major bleeding), and quality of life assessed using the New York Heart Association (NYHA) score. The NYHA score ranges from I to IV; I indicates no shortness of breath or limitations to ordinary physical activity, while IV indicates shortness of breath at rest.

Statistical analysis

Descriptive analysis for all patients including demographics, comorbidities, and clinical data from pre-intra-, and post-operative stages were analyzed. Continuous variables were reported as mean ± standard deviation, and categorical variables as frequencies and percentages. Comparisons were made using chi-square tests, Student’s t-test, ANOVA, and McNemar’s with statistical significance set at p < 0.05. Statistically significant variables were included in the multivariate logistic regression. The data analysis was conducted on STATA (version 16) and R package using rBiostatistics.com.

Results

Within the simultaneous 211 redo MV patients, 121 had concomitant procedures and 41 patients had MV IE. Surgery for active IE was indicated for 21 patients at the time of surgery, while 20 patients had previously treated IE before their MV redo. Preoperative variables are listed in Table 1.

Table 1 Preoperative variables (n = 211)

IE occurred after the initial MV surgery in 20 patients (49%) and after the second redo in 8 patients (20%). There were 18 patients with IE who had a mechanical MV vs. another 18 patients with IE who had a bioprosthetic MV, and only 5 patients had IE on the native MV. IE reoccurred in 3 patients (7%); two cases had mechanical MV affected post-surgery (one each after the first and second surgeries), and one case had a bioprosthetic MV affected after the first surgery. Of the 41 IE patients, 21 patients (51%) had positive IE cultures predominantly with staphylococcal species in 12 patients (57%). The organisms identified/suspected were Staphylococcus aureus in four patients, Staphylococcus epidermidis in four patients, other coagulase-negative Staphylococci in four patients, and Brucella in three patients. Additionally, each of the following organisms was identified in separate patients: Streptococci, Candida, Pseudomonas, Enterococci, Klebsiellae, Stenotrophomonas, Coxiella burnetii, Enterobacter cloacae, Mammaliicoccus sciuri, and Propionobacterium acnes.”

Laboratory tests and echocardiography

Renal and liver function tests were performed at three intervals: one week before surgery, immediately before surgery, and immediately after surgery (Table 2). There was a significant difference in INR (p-value < 0.0001) and in ALT (p-value = 0.044) one week prior to surgery vs. immediately post-surgery, while other laboratory results revealed no significant difference.

Table 2 Renal and liver function tests

Echocardiography results, when comparing preoperative data to the first follow-up (averaging 27 ± 116 days), revealed significant improvement in all valvular and cardiac parameters, with the exceptions of pulmonary valve stenosis, Left Ventricular Internal Diameter in diastole (LVIDd), and Left Ventricular Internal Diameter in systole (LVIDs) (Table 3). MV stenosis fully improved from 23.7% preoperatively to 0% postoperatively on first follow-up (p-value < 0.0001). Interestingly, LV function showed a marked decline, with the percentage of patients having moderately to severely reduced LVEF increasing from 9% preoperatively to 15.8% at the first postoperative follow-up. In contrast, Pulmonary hypertension improved significantly from 47.4% preoperatively to 37.3% postoperatively upon first follow-up (p-value = 0.0004). McNemar’s analysis of LVEF and RVEF is shown in Table 4.

Table 3 Echocardiography results pre and post operative

Table 4 Results of McNemar’s analysis

Operative characteristics

The preoperative MV composition was as follows: bioprosthetic in 79 patients (37.4%), mechanical in 63 patients (29.9%), repaired in 60 patients (28.4%), and native in 9 patients (4.3%). Majority of patients underwent a redo MV replacement (98.9%) with a mechanical (59.7%) or bioprosthetic (38.4%) valves, while only 1.9% had a redo valve repair. McNemar test showed there was a significant change from tissue valve preop to mechanical postoperative in 40 patients (P = 0.009). The most common observed concomitant procedure from the 121 patients was valve replacement occurring in 108 of patients (89.3%). From these patients AV was intervened in a total of 49 cases, TV in 84 cases, and AV and TV intervened at the same time in 22 patients. CABG was done in five patients, Aorta repair was done in four patients, and 11 patients had other concomitant procedures done. Post-surgery, 14.2% of the aortic valves and 14.6% of the pulmonary valves were non-native. Most patients had one redo in 139 patients (65.9%), while 16.1% required three redo cardiac surgeries. The average time to last redo was 14.2 years (Range 0.2–53). Intraoperative and postoperative characteristics are summarized in Table 5.

Table 5 Intraoperative and postoperative characteristics

The average transfusion products required during surgery were 4.2 units of red blood cells (RBC), 6.3 units of platelets, and 4 units of fresh frozen plasma (FFP). Preoperative mechanical valves were associated with more bleeding (mean 4.73 units) compared to other MV types (mean 3.54 units) with a p-value of 0.0017. Patients with preoperative native valve repair had significantly fewer bleeding events (mean 2.11 units, p-value 0.006) than all other valves (mean 3.85 units). The average cardiopulmonary bypass and aortic cross clamp time during surgery were 150.2 min and 105.1 min, respectively. The distribution of patients on mechanical ventilation was as follows: 40.38% patients for less than a day, 25% patients for more than four days, and 34.62% patients between 2 and 4 days.

The Kaplan-Meier Curve for Survival estimates is shown in Fig. 1. One-, five-, and ten-year survival for all patients was 88%, 79%, and 69%. There was a significant survival difference in patients with concomitant procedures vs. patients who did not have a concomitant procedure (P = 0.003), with a hazard ratio of 3.246 (95% CI: 1.417 to 7.434, p = 0.00535). In contrast, the hazard ratio for all IE patients showed an increased risk (HR = 1.84, 95% CI 0.88–3.84), however, it was not statistically significant (P = 0.1). There was a significant improvement of patients from New York Heart Association (NYHA) classes III and IV to classes I and II upon last follow up (p-value < 0.001).

Fig. 1

Kaplan-Meier survival probability curve over time for patients who underwent redo MV surgery. The time is measured in years along the x-axis, while the y-axis represents the cumulative survival probability, ranging from 1 (100% survival) to 0 (0% survival). (A): Solid Black Line: 170 Patients without IE 1-5-10 year survival was 90%-80%-66%. Red Line: 41 Patients with IE 1-5-10 year survival was 75%-75%-75%. (B): Solid Black Line: 90 Patients without concomitant procedures 1-5-10 year survival was 98%-80%-80%. Red Line: 121 Patients with concomitant procedures 1-5-10 year survival was 80%-77%-61%. (C): Solid Black Line: All 211 redo MV patients 1-5-10 year survival was 88%-79%-69%. Dashed line: 95% confidence interval 1-5-10 year survival was 83–91%-68-86%-50-81%

Univariate regression analysis

The primary objective of this study was to identify factors influencing mortality utilizing Cox regression model. For every one-year increase in age, there was a 4.1% rise in mortality risk (P = 0.002, CI 1.02–1.07). Specifically, conditions like diabetes mellitus and pre/post-operative stroke presented hazard ratios (HR) of 2.5 (P = 0.0185, CI 1.23–5.12) and 3.1 (P = 0.009, CI 1.33–7.12), respectively; furthermore, peripheral vascular disease (PVD) had the highest HR at 6.1 (P = 0.0065, CI 2.14–17.45). Significant determinants encompassed a history of dialysis (HR 3.4, P = 0.004, CI 1.47–7.78), the presence of a preoperative mechanical valve (HR 2.3, P = 0.014, CI 1.19–4.48), undergoing concomitant procedures (HR 3.2, P = 0.0021, CI 1.41–7.42), and having active IE (HR 3.8, P = 0.001, CI 1.78–8.15).

Preoperative left and right ventricular ejection fractions were linked to heightened mortality risks, with HRs of 2.6 (P = 0.034, CI 1.07–6.23) and 3.8 (P < 0.001, CI 1.95–7.45), respectively. Elevated levels of creatinine (P = 0.007), urea (P < 0.001), total bilirubin (P < 0.001), ALT (P < 0.001), and AST (P < 0.001) were also associated with a rise in mortality. Redo cardiac surgery was associated with a significant 44.3% increase in mortality risk (P = 0.0063, CI 1.13–1.84). During surgery, PCC and FFP administrations were associated with increased mortality risks of 11.9% (P = 0.004, CI 1.04–1.21) and 5.2% (P < = 0.001, CI 1.04–1.21), respectively.

Most patients (89.1%) were discharged from the hospital. There was a significant association (P < 0.001) between patients who received Extracorporeal membrane oxygenation (ECMO) and intra-aortic balloon pump (IABP).

Multivariate regression analysis

The Cox proportional-hazards regression analysis for the 211 cardiac surgery patients, with a mean follow-up of 2.9 years, showed a significant correlation between the evaluated variables and postoperative mortality (Fig. 2). In contrast, regression analysis on 24 patients, all of whom passed away within 25 days post-surgery revealed no significant associations.

Fig. 2

Results of the multivariable cox regression for mortality risk factors in redo mitral valve patients. PreOP = Preoperation, PostOP = Postoperation

Comment

Our study’s primary focus was to identify factors affecting mortality in high-risk patients undergoing redo MV surgery. The results showed each yearly increase in age had a 4.1% and each redo surgery had a 44.3% increase in mortality risk. While medical conditions like diabetes mellitus and pre/post-operative stroke were significant determinants, with hazard ratios (HRs) of 2.5 and 3.1, respectively. Peripheral vascular disease (PVD) stood out as a significant risk factor with the highest HR at 6.1. However, it is crucial to highlight that only 2.8% of our patient cohort had PVD, which make the result underpowered and may introduce a bias in our finding. Other significant preoperative factors included a history of dialysis, presence of a preoperative mechanical valve, and active IE, with HRs of 3.4, 2.3, and 3.8, respectively.

Our results showed that patients undergoing redo MV surgery have an acceptable in-hospital mortality rate of 10.9% and significant long-term survival rates. The Kaplan-Meier curve revealed a higher postoperative mortality trend for IE patients compared to non-IE patients (Fig. 1A). However, long-term mortality was lower in IE patients than in non-IE patients. This difference can be attributed to the significant increase in mortality among redo MV patients with concomitant procedures beyond 8 years (Fig. 1B). This may suggest that MV IE mortality risk is higher in redo MV surgery in the short-term only while patients with concomitant procedures have worse short and long-term prognosis. Future research should investigate whether this trend persists in long-term follow-up.

The adoption of biologic valve prostheses for MV replacement rose greatly, increasing from 16.8% in 1993 to 53.7% in 2013 [14]. However, our study showed significant change (P = 0.009) in patients who had bioprosthetic (37.4%) or mechanical (29.7%) valves preoperatively compared to the postoperative composition which was mainly mechanical (59.7%) and bioprosthetic (38.4%). This suggested more careful selection for bioprosthetic MV in patients, which may lessen the need for redo cardiac surgeries.

High-risk patients faced greater risks when undergoing redo MV surgery, particularly those with IE or needing concomitant procedures. Our results showed a heightened IE reoccurrence risk in patients with mechanical MVs compared to bioprosthetic. However, Rutledge et al. noted equal IE susceptibility across valve types in the first post-replacement year [15]. The MitraClip procedure, a type of percutaneous edge-to-edge MV repair, presents a low IE risk [16]. In our cohort, IE occurred in 18 patients with bioprosthetic and 18 with mechanical valves, followed by four with native and one with repaired MV. IE on MV reoccurrence was in two mechanical and one bioprosthetic valves. This may suggest patients undergoing redo mechanical MV surgery require careful follow-up and prophylaxis for IE.

Our cohort showed an overestimation with EUROSCORE-II (22.5 ± 15), and an underestimation with STS score (3.6 ± 3.2). This is contrary to Keenan et al. where EUROSCORE-II underestimated mortality [17]. Onorati et al. reported that EuroSCORE-II and STS-score effectively predict hospital mortality when EuroSCORE-II is > 13 or STS-score is > 7.4. However, lower score values lacked accuracy, especially in redo-patients. Interestingly, STS-score was more predictive than EuroSCORE-II for mitral patients, emphasizing the importance of using both scoring systems [13]. We need to have both scoring system, and this mandates the need to have a specific scoring system that is more accurate and involves the elements of patient demographics and rheumatic heart disease.

Fortunately, the majority (89.1%) of our high-risk cohort were successfully discharged from the hospital. A significant association was identified between the intraoperative use of ECMO and IABP, emphasizing the importance of these interventions in patients with low left ventricle ejection fraction. Utilizing the Cox proportional hazards model, we found age to be a crucial determinant in the risk of MACE post-cardiac surgery, with other significant contributors being diabetes mellitus and tracheostomy.

Further analysis involving 211 all redo MV patients, with an average follow-up of 2.9 years, revealed associations between several parameters and postoperative mortality. These included age, peripheral vascular disease, concomitant surgical procedures, and preoperative left ventricular function, among others. However, a focused study on a subset of 24 patients that passed away within 25 days post-operatively, did not show significant correlations with postoperative MACE events.

In conclusion, several factors were significantly associated with increased mortality risk in redo MV surgery patients. Identifying and addressing these factors can help improve patient outcomes. These results highlight the importance of close follow-up and vigorous complication prevention after redo MV surgery in patients with MV IE and concomitant redo MV surgery.

Limitations

The study was conducted at a single quaternary hospital, predominantly involving patients with rheumatic heart disease. This specific patient demographic may introduce potential bias in the observed outcomes.