Triple valve replacement for rheumatic heart disease: short- and mid-term survival in modern era (2024)

Search for other works by this author on:

Article history

Abstract

INTRODUCTION

Rheumatic heart disease (RHD) constitutes a significant cause of cardiac operations in the Far and Middle East [1, 2]. Triple valve surgery is a complex operation and has an in-hospital mortality rate of about 10–12% [3, 4]. Overall actuarial survival rates at 1 year and 5 years are 84.5 and 75%, respectively [3]. Morbidities such as endocarditis, thromboembolism and anticoagulation-related haemorrhage have been reported [5]. Triple valve replacement (TVR) is a type of triple valve surgery and is challenging for most cardiac surgeons due to its prolonged periods of cardiopulmonary bypass (CPB) and aortic cross-clamping for the patients, who are commonly at late stages of RHD [5, 6].

Advances in myocardial protection and CPB techniques can be credited for the improvement in early survival after triple valve surgery [4]. Most new series include a small number of TVR procedures, as opposed to mitral and aortic valve replacement and tricuspid repair [4, 7, 8]. In old studies, not only were old-generation valves replaced but also perioperative and postoperative care was different from that in the modern era [9–11]. In the modern cardiac surgery era, survival in the early and mid-term periods has yet to be clearly defined. Moreover, the modern cardiac surgery era lacks reliable statistical data owing to the small number of patients included in studies and the different patient characteristics, disease aetiologies and treatment procedures.

The present study describes our experience with 90 patients who underwent TVR for advanced RHD at a single tertiary cardiac centre. We attempted to define early and mid-term clinical outcomes and analyse overall survival, event-free survival and postoperative symptoms in the modern era.

METHODS

In this historical cohort study, TVR was defined as patients who underwent aortic, mitral and tricuspid valve replacement combined during the same surgical procedure. Between September 2003 and October 2013, 90 consecutive patients with rheumatic triple valve disease underwent TVR (with or without coronary artery bypass surgery) at our centre, and their clinical, operative and outcome variables were reviewed and then were followed in the period after their surgery until our study performance. All patients' questionnaire forms were filled on call in March 2014 and if new events had occurred, clinical documents were reviewed. The variables included in our analysis were composed of demographic characteristics, cardiovascular risk factors, intraoperative factors and short- and mid-term postoperative mortality and morbidity rates. The primary outcome was cardiac death and cardiac events, including rehospitalization for cardiac cause, reoperation for cardiac cause, cerebrovascular accidents, anticoagulation-associated major haemorrhage and significant valvular dysfunction (structural valve deterioration, valve thrombosis and paravalvular leakage). Overall survival was defined as patients surviving (without cardiac or non-cardiac death) during the follow-up period after triple valve surgery and event-free survival was defined as survived patients during the follow-up period after the surgical procedure until the first registered event. We tried to report events according to recommended guidelines for reporting mortality and morbidity after cardiac valve replacement [12]. The research proposal was approved by our institutional review board (IRB). At the time of admission, informed consent had been filled by the patients and according to IRB recommendation, questionnaire forms were filled on phone call follow-up after verbal consent.

Patients

There were 71 women and 19 men at a mean age of 47 ± 12 years (range = 23–74 years). Most of the patients were in the New York Heart Association (NYHA) functional class II or III. Eight (9%) patients had undergone previous cardiac intervention: open (1 patient) or closed (5 patients) mitral valve commissurotomy and percutaneous mitral valve commissurotomy (2 patients). Eighty-seven (97%) patients underwent elective surgery. Emergency surgeries were done owing to infectious endocarditis secondary to RHD in 1 patient and refractory decompensated heart failure in 2 patients. The complete clinical profile and operative data of the patients are summarized in Table 1. The main cause of aortic valve, mitral valve and tricuspid valve disorder was rheumatic valve disease. Thirty-six patients (40%) had atrial fibrillation before surgery, which increased to 45 patients after surgery. The complete list of the valve diseases is presented in Table 2.

Selective coronary angiography and transthoracic echocardiography were performed in all the patients before surgery. The selection of the prosthesis type was based on the patient's age and the physician's preference. The types of prosthetic valves and their positions are presented in Table 1. In most patients, in the aortic and mitral positions, mechanical valves, and in the tricuspid position, bioprosthetic valves were used (78%), followed by mechanical valves in triple positions (13%) and bioprosthetic valves in triple positions (4%). The ranges of the size of the prosthetic valves in the mitral, aortic and tricuspid positions were 23–29, 17–25 and 27–31 mm, respectively. The majority of the mechanical prostheses were St Jude (St Jude Medical, Inc., Minneapolis, MN, USA), and most of the bioprosthetic valves were Carpentier-Edward pericardial valves (Edwards Life sciences) and Hanco*ck II porcine valves (Medtronic, Inc.).

Operative technique

All the operations were done by the same group of surgeons. All the operations were preformed through a median sternotomy incision. CPB was performed under moderate hypothermia (28°C), utilizing a membrane oxygenator. Myocardial protection was offered through the instillation of a crystalloid cardioplegic solution via the coronary ostia. Topical hypothermia with saline ice slush was employed for better protection of the ventricles. After the excision of the aortic valve, the mitral valve was exposed through the trans-septal approach. The anterior leaflet, with a portion of the posterior leaflet, was excised; and whenever possible, chordal attachment was preserved. The prosthesis was implanted using interrupted horizontal mattress sutures with felt pledgets. The mitral prostheses were oriented in the antianatomical position. After the mitral procedure, the aortic prosthesis was sewn in place also with interrupted horizontal mattress sutures with felt pledgets. The aortic prostheses were implanted perpendicular to the ventricular septum. Finally, the tricuspid valve was replaced. The replacement procedure was performed while the ascending aorta was cross-clamped. The tricuspid prostheses were oriented perpendicular to the septal leaflet.

Anticoagulation

Anticoagulation was initiated with warfarin after the removal of the chest tubes. The International normalized ratio (INR) was the only guide for anticoagulation. Anticoagulation was carried out and controlled by the outpatient clinic of our centre and its cardiologists. The target INR was 2.5–3.5 for the aortic and mitral mechanical prostheses and 2.5–3 for atrial fibrillation patients with triple bioprosthetic valves. When triple bioprosthetic valves were implanted, warfarin was administered in the first 6 months postoperatively for those without atrial fibrillation. Low-dose aspirin was administered to all the patients. Follow-up was done via at least yearly outpatient clinic visits for detailed physical examinations and intermittent transthoracic echocardiography and when clinically needed with transoesophageal echocardiography, correspondence questionnaires or telephone contacts. Complications reported by the patients were documented by reviewing their hospital records.

Statistical analysis

The categorical variables are presented as frequencies and percentages, and the continuous variables are described through mean and standard deviation when normally distributed, otherwise presented as median and interquartile range. Survival curves were calculated using the Kaplan–Meier method, and their 95% confidence intervals (CIs) were computed through the log-transform method. The univariate effects of the variables on overall survival and event-free survival were evaluated using the Cox proportional hazards (PH) model, and the effects were reported through the hazard ratio (HR) with 95% CI. The variables with a P-value <0.2 in the univariate analysis were selected as candidates to enter the multivariable model. A Cox PH model with a backward selection method (with probabilities of entry and removal as 0.05 and 0.1, respectively) was applied to find the multiple predictors of overall survival and event-free survival. The PH assumption was tested using the χ² test of the correlation coefficient between the transformed survival time and the scaled Schoenfeld residuals. The statistical analyses were conducted using SPSS 16.0 and R software (version 3.0.0).

RESULTS

The median follow-up was 41 months (range = 0.13–117.1 months). The total cumulative follow-up was 5015 patient-years. We missed 3 (3%) patients after hospital discharge. Coronary artery bypass grafting in addition to TVR was done in 9 (10%) patients. Five patients had single-vessel disease, 2 had two-vessel disease and another 2 had three-vessel disease. The left internal mammary artery was used in 5 patients and the saphenous vein graft was utilized in the other vessels. The total cardiac death rate was 14% (13 patients: 5 patients died in hospital and the others died during the follow-up period; the cause of mortality was acute myocardial infarction in 2 patients and congestive heart failure in 6). The rate of 30-day mortality after surgery, including in-hospital mortality, was 6% (5 patients). The rate of all-cause mortality was 16% (14 patients). Postoperatively, only 6 patients received the intra-aortic balloon pump during their intensive care unit stay. Four patients died due to multiorgan failure and sepsis, and another one expired owing to mitral valve thrombosis and decompensated congestive heart failure. The median time of hospital stay was 17 days (interquartile: 12–23 days).

In the hospital stay period, the incidence rate of complications was 29% (26 patients); some patients had more than one complication. Seven (8%) patients required the re-exploration of the mediastinum for cardiac tamponade, and 7 (8%) patients required the placement of a permanent pacemaker (epicardial lead) for heart block. Other morbidities included cerebrovascular accident in 3 (3%) patients, renal insufficiency in 3 (3%, Cr >2.0 mg/dl), pulmonary complications (including pneumonia and significant pleural effusion) in 7 (8%), new onset atrial fibrillation in 9 (10%), postoperative blood transfusion in 26 (29%), sternal infection in 2 (2%) and sepsis in 2 (2%), which was treated medically.

In the follow-up period, non-structural valve dysfunction was seen in 4 (4%) patients (moderate paravalvular leakage in one mechanical mitral valve, mechanical aortic valve dysfunction due to pannus formation in 3 patients), and valve thrombosis was seen in 4 (4%) patients (mechanical mitral and tricuspid valve dysfunction due to thrombosis in 1 and 3 patients, respectively), bleeding events in 9 (10%) (intracranial haemorrhage in 2 patients, gastrointestinal bleeding in 6 and muscular haematoma in 1), cardiac rehospitalization in 12 (13%) (1 patient for fever and work-up for endocarditis, 4 for valvular dysfunction and reoperation and 6 for dyspnoea and congestive heart failure), and cardiac reoperation in 4 (4%) (valvular dysfunction in 3 patients: 1 with pannus formation in a mechanical aortic valve and 2 with tricuspid valve thrombosis and 1 for CABG; all of them survived the redo operation), so valve reintervention was done in 3 (3%) patients. No instances of structural valve failure of the mechanical prostheses, thromboembolic event, myocardial infarction or admission because of acute coronary syndrome were observed.

The univariate analysis revealed that age (HR = 1.1; 95% CI: 1.03–1.17, P = 0.004), preoperative diabetes (HR = 4.6; 95% CI: 1.37–14.96, P = 0.013), preoperative hypertension (HR = 4.0; 95% CI: 1.33–12.06, P = 0.013) and pump time (HR = 1.0; 95% CI: 1.00–1.04, P = 0.005) were the predictors of overall survival and that also age (HR = 1.0; 95% CI: 1.00–1.07, P = 0.038), preoperative diabetes (HR = 5.3; 95% CI: 2.27–12.13, P <0.001) and pump time (HR = 1.0; 95% CI: 1.00–1.01, P = 0.040) were the predictors of event-free survival. The findings remained significant in the multivariate analysis regarding overall survival except hypertension and the effect of age was not significant for event-free survival (Tables 3 and 4).

Because of the low incidence of non-cardiac mortality (one patient), only cardiac survival was calculated, and that patient considered as a patient with incomplete follow up after death date. The 1-year and 4-year rates of overall survival (cardiac survival) were 91.7% (95% CI: 86.0–97.8%) and 89.5% (95% CI: 82.7–97.0%), respectively. The 1-year and 4-year rates of freedom from cardiac events (e.g. cardiac death, cardiac rehospitalization, cardiac reoperation, cerebrovascular events, anticoagulation-related major haemorrhage and significant valvular malfunction) were 83.5% (95% CI: 79.0–91.0%) and 69.5% (95% CI: 59.0–81.7%), respectively. The 1-year and 4-year rates of freedom from cardiac rehospitalization were 94.0% (95% CI: 89.0–99.2%) and 88.0% (95% CI: 80.0–98.6%), respectively. The 1-year and 4-year rates of freedom from cardiac reoperation were 98.8% (95% CI: 96.5–100%) and 93.9% (95% CI: 87.0–98.6%), respectively. The 1-year and 4-year rates of freedom from significant prosthetic valve malfunction (e.g. structural valve deterioration, valve thrombosis and paravalvular leakage) were 96.6% (95% CI: 92.8–100%) and 90.7% (95% CI: 83.7–98.4%), respectively. The 1-year and 4-year rates of freedom from bleeding events were 96.4% (95% CI: 92.4–100%) and 90.7% (95% CI: 83.4–98.6%), respectively (Figs 1 and 2). The 1-year and 4-year rates of freedom from a composite of valvular thrombosis, major bleeding events and thromboemboli were 94.1% (95% CI: 89.2–99.3%) and 88.5% (95% CI: 80.8–96.9%), respectively. The 1-year and 4-year rates of freedom from major adverse valve-related events (a composite of valve-related mortality, all valve-related morbidity and need for new permanent pacemaker or defibrillator within 14 days after valve intervention) were 86.3% (95% CI: 74.9–93.8%) and 78.5% (95% CI: 69.2–89.1%), respectively (Fig. 3). On the last follow-up, from 68 patients, 11 (16%) patients were in the NYHA functional class II and 1 (2%) patient was in Class III. The other (82%) patients were asymptomatic.

In the last transthoracic echocardiographic examination, the mean left ventricular ejection fraction (LVEF) was 47.6% (SE: 6.9%) and the tricuspid annular plane systolic excursion was 15.6 mm (SE: 3.9 mm). The peak and mean gradients in the prosthetic mitral valves were 12.4 mmHg (SE: 5.8 mmHg) and 5.5 mmHg (SE: 3.5 mmHg), the peak and mean gradients in the prosthetic aortic valves were 26.8 mmHg (SE: 15.0 mmHg) and 15.7 mmHg (SE: 9.2 mmHg), and the peak and mean gradients in the prosthetic tricuspid valves were 7.0 mmHg (SE: 2.7 mmHg) and 3.7 mmHg (SE: 1.6 mmHg).

DISCUSSION

RHD is a significant cause of cardiac operations in developing countries [1, 2]. Cardiac valve surgeries for RHD account for a significant portion of valve surgeries. Despite improvements in CPB and myocardial protection techniques, TVR for RHD is a challenging and complex procedure. It is not completely possible to compare our data with those in new TVR series because of the small proportion of TVR procedures in these series compared with double-valve replacement and single-valve repair. Nevertheless, in comparison with the old TVR series, some aspects of the new series chime in with ours insofar as they share at least two valve replacements for most patients, modern CPB, surgical techniques, replacement of new generations of valves and advances in perioperative and postoperative care.

Hospital mortality

In our series, the hospital mortality rate was 6%, which is favourable when compared with that reported in the majority of the old series. This indicates that the range of operative mortality can rise up to 24% [9–11, 13]. In the triple valve surgery studies, the operative mortality rate range is up to 11% [3, 5, 6, 14].

In our series, the aetiology of valve disease was rheumatic fever. Our patients were younger than patients in industrialized countries. Advanced age is associated with decreased cardiopulmonary reserve and increased prevalence of comorbid factors. In case series reported from these countries, RHD is not a dominant cause for valve disease [8]. Younger age and better functional class (7 patients with the NYHA functional class IV) are the possible reasons for the lower hospital mortality rates in our series.

Overall survival

In the present study, the 1-year and 4-year overall survival rates were 91.7 and 89.5%, respectively. In the triple valve surgery studies, the 1-year and 5-year survival rates after triple valve surgery have been reported at 80–84.5% [3, 15] and 70–75% [3, 5, 8, 15, 16], respectively. It seems that the 1-year and 4-year survival rates in our study are more favourable than those in other similar studies.

Better myocardial protection and CPB techniques, increased experience with triple valve procedure, advances in the treatment of postoperative heart failure, intensive patient follow-up, and extensive education on anticoagulation can be the reasons for the improvement in patient survival. It is worthy of note that the first patient in our study was enrolled 6 years after the date of enrolment in the last published report on triple valve surgery [3]. In addition, younger age and better functional class in our series may be the other possible factors in favour of the more desirable early and mid-term survival rates.

Predictors of overall survival

In our univariate analyses age, diabetes mellitus, hypertension and pump time and in multivariate analysis age, diabetes mellitus and pump time were the predictors of overall survival. Advanced age, NYHA functional class IV, lower LVEF, renal impairment, pulmonary hypertension, TVR, prior mitral valve replacement and emergency operation were the independent predictors of late mortality after triple valve surgery in various studies [5, 7–9, 11, 14–17, 18].

In our study, in most patients, the NYHA functional class was less than IV. In addition, TVR was done for all the patients and TVR was done in one operation. Consistent with some previous studies, the LVEF [5, 14] and renal impairment [14] were not the predictors of overall survival in our study population. Our different results can be due to the lower prevalence of significant LV systolic dysfunction and renal impairment (2%) and the higher prevalence of systemic hypertension (20%) and diabetes mellitus (21%) in our sample. In our study, pump time was an independent predictor of overall survival and valve stenosis or regurgitation severity was not a predictor of overall survival.

Thromboembolism and bleeding events

It has been reported that complications related to prosthetic valves in the triple valve procedure are comparable with those in single- and double-valve replacement [19–21]. Thromboembolism and bleeding are the major problems due to anticoagulation and valve replacement [5, 7, 16]. In our study, thromboembolism did not occur in any patients and the 1-year and 4-year rates of freedom from anticoagulation-related major haemorrhage were 96.4 and 90.7%, respectively. The rates of 10-year freedom from thromboembolism and major bleeding in triple valve surgery have been reported to be in the ranges of 81–88 and 81–83% [5, 6, 16]. The absence of the occurrence of thromboembolism may be because of the lower follow-up duration, higher INR maintenance (2.5–3.5), inherent difference in coagulable states and competence in managing the patients' anticoagulation. Nonetheless, the incidence of major bleeding in our study is comparable with that in similar studies.

Cardiac reoperation

The rates of 1-year and 4-year freedom from cardiac reoperation in our study were 98.8 and 93.9%, respectively. The rates of freedom from reoperation at 5 years have been reported to range from 86.3 to 96% [7, 22, 23]. Our results are, therefore, in agreement with those reported previously. We had no failure of valve repair (because no repair was done) and prosthetic valve endocarditis in our sample.

Paravalvular leakage

In the current study, paravalvular leakage was seen in 1 patient in the follow-up period, which is much lower than the rates reported by the other studies [5]. This may be due to the effectiveness of felt pledgets and interrupted horizontal mattress sutures, the combination of which may have provided valve stability.

TVR may be considered in patients with RHD. For all the complexity of TVR, its short- and mid-term outcomes show considerable improvement on what was reported in older series. We achieved acceptable rates of perioperative mortality for TVR with good short- and mid-term results. The majority of our TVR patients were in the NYHA functional class I or II during the follow-up period, and the mid-term valve-related complications rates were comparable with those of patients undergoing single- or double-valve replacement. These favourable results following TVR in patients with RHD justify the use of this surgical procedure in patients before the NYHA functional class IV.

The most significant limitation in the present study is that it is a single-centre investigation with a small sample size. The outcome, by comparison with similar studies, may have been influenced by concomitant coronary bypass graft surgery in some patients, diverse patient characteristics, different kinds of replaced valves and dissimilar surgical techniques and postoperative care. However, the number of TVR patients in our series is higher than those in recent studies on the survival and prognosis of triple valve surgery.

Conflict of interest: none declared.

REFERENCES

© The Author 2014. Published by Oxford University Press on behalf of the European Association for Cardio-Thoracic Surgery. All rights reserved.

Issue Section:

Adult Cardiac

Email alerts

• Web of Science
• Google Scholar

Citing articles via

More from Oxford Academic