Long-term outcomes of minimally invasive concomitant mitral and tricuspid valve surgery with surgical ablation

Authors: Sungsil Yoon, Kitae Kim, Jae Suk Yoo, Joon Bum Kim, Cheol Hyun Chung, Sung-Ho Jung

Abstract

OBJECTIVES

We compared the outcomes of a right mini-thoracotomy (RMT) versus those of a sternotomy for concomitant mitral and tricuspid valve surgery and surgical ablation.

METHODS

We analysed patients who underwent concomitant mitral and tricuspid valve surgery and surgical ablation at a single institution (mean follow-up: 7 years) after propensity score matching. The primary and secondary outcomes were all-cause death, composite major adverse events (including stroke, reoperation, readmission, permanent pacemaker insertion) and recurrence of atrial fibrillation (A-fib). A subgroup analysis was performed.

RESULTS

A total of 797 procedures (mean 61.6 years; RMT: 45.2%; 66.5%; mitral valve 33.6%) were done; 267 pairs were matched. The 5- and 10-year overall survival in the matched cohort was 92.7% and 86.9% for the RMT group and 92.1% and 83.1% for the sternotomy group (*P = *0.879). Significant differences were not observed in major adverse events (*P = *0.273; hazard 0.76) and A-fib recurrence (*P = *0.080; hazard 0.72). The RMT group had lower rates of postoperative low cardiac output syndrome (*P = *0.019) and acute renal failure (*P = *0.003). Atrial fibrillation high-risk factors (including long-standing A-fib, enlarged left atrium, old age) exhibited significant interactions (P for interaction = 0.002) with the approach regarding A-fib recurrence.

CONCLUSIONS

In this study, an RMT exhibited no significant differences in long-term outcomes compared to a sternotomy, but it could remain a clinically reasonable option. Patients with a high risk of A-fib may have favourable ablation outcomes with a sternotomy.

INTRODUCTION

Since the introduction of minimally invasive cardiac surgery (MICS) using the right mini-thoracotomy (RMT) approach in 1990 [1], it has continued to advance, resulting in an expansion of related surgical techniques. The RMT approach, which was previously limited to mitral valve (MV) procedures, is now commonly used even in cases involving concurrent MV and tricuspid valve (TV) operations. Several studies related to these cases have been published [2, 3]. They demonstrate that the augmented utilization of RMT has yielded favourable outcomes regarding cosmetic aspects, postoperative recovery and short- and mid-term results compared to a conventional median sternotomy.

Compared to the conventional sternotomy, most of these studies have also indicated an association between RMT and prolonged cardiopulmonary bypass (CPB) and operating time [2, 4]. Considering the well-established correlation between prolonged CPB duration and compromised surgical outcomes [5], we decided to analyse the following

First, would outcomes compare to those from a sternotomy, despite anticipated CPB and operating time extension, as we expand the use of RMT in cardiac operations, specifically incorporating ablation in MV and TV operations?

Second, could there be any risk factors impacting the selection of the approach if the surgical outcomes are comparable in both surgical approaches?

To address these questions, we investigated whether the RMT approach is appropriate for a concomitant MV and TV operation, including surgical ablation. In addition, we identified risk factors influencing the approach using subgroup analysis.

MATERIALS AND METHODS

Study cohort

We searched the institutional Cardiac Surgery Database of the Asan Medical Center (Seoul, Korea) to identify adult patients (age ≥ 18 years) who consecutively underwent concomitant MV and TV operations and surgical ablation between January 2003 and September 2022. Among the 1379 patients, those who underwent other combined major cardiac operations were excluded. Other exclusion criteria encompassed patients meeting the surgical ablation exclusion criteria (using microwave or radiofrequency, a history of previous surgical or catheter ablation, only right-side ablation), active infective endocarditis and a lower sternotomy approach. Finally, 797 patients (mean 61.6 years; 66.5%) formed the final study cohort. Approach conversion cases were analysed in conjunction with the RMT group (see Supplementary Material, Fig. S1). The baseline characteristics and surgical records of all patients were complete. The surgical approach was determined at the surgeon’s discretion. Most surgeons attempted MICS; however, a sternotomy was often chosen when a high surgical risk was predicted due to an enlarged heart, severe lung adhesions, combined severe comorbidities or other factors. The study protocol was approved by the institutional review board of the Asan Medical Center (approval 2023-0490; date of 21 April 2023); the requirement for informed patient consent was waived because of the retrospective nature of the study.

Outcomes and definitions

The primary outcome was all-cause death. The secondary outcomes were composite major adverse events (MAEs) requiring hospitalization and atrial fibrillation (A-fib) recurrence. We selected these outcomes from the Valve Academic Research Consortium-3 [6]. The MAEs included stroke, cardiac reoperation, cardiac readmission and insertion of a new permanent pacemaker. We analysed not only composite MAEs but also each one individually. Stroke was defined as the occurrence of neurological deficits, which was validated in imaging studies. Cardiac readmission included all cases requiring hospitalization due to symptoms related to cardiac problems including valve-related events, arrhythmia and congestive heart failure. We restricted cardiac reoperation to cases in which a new cardiac operation was required specifically in connection with concomitant MV and TV operations and surgical ablation. A-fib was classified according to the European Society of Cardiology/European Association of Cardio-thoracic Surgery guidelines into paroxysmal (usually ≤ 48 h), persistent (> 7 days or requiring cardioversion) and long-standing persistent (> 1 year) [7]. Postoperative recurrence of A-fib was defined as any episode of A-fib, atrial flutter or atrial tachycardia lasting longer than 30 s, which occurred after a 3-month blanking period [8]. We diagnosed this condition using consecutive electrocardiograms, a 24-h Holter monitor or echocardiography. Monitoring for A-fib recurrence was performed mainly at the outpatient clinic visit postoperatively. At the end of a 3-month blanking period, the restoration of sinus rhythm was usually assessed by a 24-h Holter monitor or electrocardiography. Thereafter, the rhythm status of the patient was screened every 6 months for 2 years postoperatively and then annually after 2 years from the operation. Recurrence of A-fib was typically managed using electrical cardioversion or class I or III anti-arrhythmic drugs. Patients who underwent mechanical valve replacement were prescribed warfarin to maintain an international normalized ratio of 2.0–3.0. Individuals who received annuloplasty rings or bioprosthetic valves typically received warfarin for 3–6 months. The decision regarding continued anticoagulation thereafter was contingent upon estimated thromboembolic risk and restoration of sinus rhythm, signifying effective atrial contraction.

Subgroup analysis was used to identify the factors affecting the surgical approach. It was conducted by propensity score matching (PSM) covariates that could potentially influence the RMT or were confirmed as risk factors through the Cox proportional hazard (CPH) model. Continuous variables were transformed into categorical variables for analysis, considering cut-off values or previously published studies [7, 9, 10]. Among these covariates, the presence of long-standing A-fib, a significantly enlarged left atrium (LA) (diameter >60 mm [10]) and age ≥65 years were selected to define A-fib high-risk factors responsible for A-fib recurrence. Next, we evaluated whether this composite risk factor impacted surgical approaches. The factors previously identified as risk factors for A-fib recurrence [11] were reaffirmed in our study through the CPH model.

We also assessed early postoperative complications (within 30 days or in the hospital). Low cardiac output syndrome (LCOS) was defined as requiring mechanical support, and bleeding, as requiring bleeding control procedures with the patient under general anaesthesia. Vascular injuries resulting from peripheral cannulation were included in wound problems. Pericardial effusion was defined as necessitating drainage procedures.

Follow-up data were obtained until September 2022, with censoring applied to patients still alive at that time. Patients without any episodes of A-fib recurrence or MAEs were censored at the last outpatient follow-up date. Vital status was obtained from the National Population Registry of the Korea National Statistical Office, and all-cause mortality was followed up in 100% of this cohort.

Surgical method

A sternotomy utilized conventional ascending aorta and bicaval cannulation. An RMT required peripheral cannulation through the femoral artery and vein and the internal jugular vein. In RMT, a 4- or 5-cm horizontal incision was made over the fourth intercostal space between the anterior and median axillary lines. A thoracoscope was then inserted through a 1-cm port at the third or fifth right intercostal space. Myocardial protection strategies utilized antegrade cold crystalloid cardioplegia. Following cardioplegic arrest and aortic cross-clamping (ACC), the MV was exposed through an interatrial groove.

The typical procedure for LA ablation involved conducting the procedure after a left atriotomy in the interatrial groove, facilitating access to the interior of the LA. Lesion sets utilized for LA ablation comprised a box lesion encircling the pulmonary veins; a mitral line extending from the box lesion to the mitral annulus; a line extending from the box lesion to the LA appendage; and an epicardial coronary sinus lesion. For a right-atrial ablation, the lesions included a cavotricuspid isthmus lesion and a line extending from the cavotricuspid isthmus lesion to the superior vena cava. Cryoablation using argon was the method employed. For patients with an LA diameter exceeding 5 cm, LA reduction was typically performed to prevent a macro-reentry circuit. This goal was achieved by resecting an enlarged atrial free wall between the right inferior pulmonary vein and the posterior MV annulus. Treatment of the LA appendage involved external resection, endocardial obliteration using running sutures or the application of an occlusion clip.

Statistical analysis

Continuous variables are presented as means and standard deviations, whereas non-normally distributed variables are presented as medians with ranges (first and third quartiles). These were compared using the Student t-test for data with normal distribution and the Mann–Whitney U-test for skewed data based on the normality test. Categorical variables are presented as frequencies with percentages. The χ^2^ test or Fisher’s exact test was used for categorical variables, depending on the expected frequency assumption. Several differences in patient characteristics precluded the direct comparison of outcomes (Tables 1 and 2). Therefore, to reduce the potential treatment selection bias, we used PSM to approximate a randomized trial. One-to-one matching without replacement was performed using greedy nearest neighbour matching on the logit of the propensity score with a caliper of width 0.2. A total of 27 covariates were used for PSM and are listed in Supplementary Material, Fig. S2. After PSM, a matched cohort was compared using the paired t-test or the Wilcoxon signed-rank test, depending on normality tests in continuous variables. For categorical variables, the McNemar test was used. The Kaplan–Meier analyses were used to evaluate the conditional probability of all-cause mortality, composite MAEs and recurrence of A-fib. We used the stratified CPH model to compare survival analysis data. In the matched cohort, we employed a multistate model [12] to analyse the cumulative incidence of death, considering the competitive risk of non-cardiac death The multistate model is utilized in research to analyse various disease progressions effectively. The competing risk model, a specific instance of the multistate model, extends the traditional survival analysis framework by accommodating different causes of death. We opted for a multistate model incorporating competing risks instead of the commonly used Fine-Gray model to also account for strata (PSM). Subgroup analysis was conducted in the matched cohort using different risk factors. To assess these factor-dependent effects of the surgical approach on the primary and secondary outcomes, a CPH model was fitted with an interaction term for risk factors and surgical approach. All reported P-values were two-sided, and those < 0.05 were considered significant. The R software, version 4.3.1 (R Core Team; R Foundation for Statistical Computing, Vienna, Austria) and Prism version 9 (GraphPad Software, San Diego, CA, USA) were used for statistical analyses.

RESULTS

Baseline characteristics and operative data

Whole cohort

Table 1 summarizes the baseline characteristics for the whole and the matched cohorts. In the full cohort, the sternotomy group was older than the RMT group, with no significant differences in glomerular filtration rate, hypertension, diabetes, chronic obstructive pulmonary disease (COPD) or cerebrovascular disease. Furthermore, the sternotomy group had a higher prevalence of New York Heart Association functional classes III and IV, along with a greater prevalence of long-standing persistent A-fib. Preoperative echocardiography in this group revealed a higher prevalence of severe tricuspid regurgitation and a larger LA diameter, left ventricular systolic (LVIDs) and diastolic (LVIDd) internal dimensions.

Table 2 depicts the operative data. In the whole cohort, no significant difference was observed in the rate of mitral repair. However, atrium reduction procedures were performed more frequently in the sternotomy group. The bi-atrial maze procedure was more commonly performed in the RMT group. We divided the total research period into 5-year intervals for comparison. The surgical period revealed that the RMT was usually conducted recently.

Matched cohort

All covariates used in the analysis were incorporated into PSM, and most of these exhibited well-balanced standard mean deviation values, displaying no significant differences (Tables 1 and 2). PSM provided a matched cohort consisting of 267 pairs of patients.

Clinical outcomes

Early postoperative outcomes

Whole cohort

Table 3 presents early postoperative outcomes. Of all of the cohorts, the RMT group exhibited longer CPB time (167 vs 144 min, *P < *0.001) and ACC time (105 vs 98 min, *P < *0.001). However, ventilation time (11 vs 13 h, *P < *0.001), intensive care unit (ICU) stay (2 vs 3 days, *P < *0.001) and hospital stay (10 vs 12 days, *P < *0.001) were all shorter in the RMT group.

The entire cohort showed no significant differences in early stroke, postoperative bleeding, pneumonia, wound problems or pericardial effusion. However, LCOS (2.8% vs 6.4%, *P = *0.026) and acute renal failure (ARF) requiring continuous renal replacement therapy (CRRT) (2.5% vs 6.6%, *P = *0.011) were more prevalent in the sternotomy group. Early composite complications, encompassing all complications detailed earlier, displayed higher prevalences in the sternotomy group (18.1% vs 25.2%, *P = *0.020). Early reoperations and deaths did not exhibit any significant differences.

Matched cohort

In the matched cohorts, the RMT group also exhibited longer CPB time (169 vs 146 min, *P < *0.001) but did not exhibit significant differences in ACC time (*P = *0.082). Ventilation time (11 vs 13 h, *P = *0.015), ICU stay (2 vs 3 days, *P < *0.001) and hospital stay (10 vs 12 days, *P < *0.001) were all shorter in the RMT group, similar to the whole cohort.

The LCOS (1.9% vs 6.4%, *P = *0.019) and ARF requiring CRRT (1.9% vs 7.9%, *P = *0.003) were lower in the RMT group. But other early outcomes did not exhibit any significant disparities (Table 3).

Long-term outcomes

During a mean follow-up of 7 (4.8) years, overall all-cause death was 11.3% in the whole cohort and 10.1% in the matched cohort (see Supplementary Material, Table S1).

Whole cohort

In the whole cohort, the RMT group displayed higher overall survival than the sternotomy group. The overall survival at 5 and 10 years was 93.2% and 88.5% in the RMT group and 90.3% and 79.6% in the sternotomy group hazard ratio (HR): 0.58, *P = *0.013. The RMT group had significantly lower rates of composite MAEs (HR: 0.67, *P = *0.015) occurrence and A-fib recurrence (HR: 0.66, *P < *0.001) (Fig. 2A and C). The overall incidences of composite MAEs occurrence (15.8% vs 24.7%, *P = *0.003) and A-fib recurrence (29.2% vs 41.7%, *P < *0.001) were lower in the RMT group (see Supplementary Material, Table S1).

Matched cohort

After PSM, no significant differences were noted in the overall survival between the 2 groups. The overall survival at 5 and 10 years was 92.7% and 86.9% in the RMT group and 92.1% and 83.1% in the sternotomy group (stratified CPH: HR: 0.96, *P = *0.879) (Fig. 1B). The cumulative incidence of cardiac mortality with non-cardiac mortality as a competitive risk was not statistically significant (*P = *0.100) (see Supplementary Material, Fig. S3). Neither group exhibited significant differences in the rate of composite MAEs occurrence (HR: 0.76, *P = *0.273) and A-fib recurrence (HR: 0.72, *P = *0.080), unlike the whole cohort (Fig. 2B and D). The overall incidences of composite MAEs occurrence were lower in the RMT group (15.4% vs 23.2%, *P = *0.019) but A-fib recurrence did not exhibit significant differences (see Supplementary Material, Table S1). Stroke (*P = *0.100), cardiac reoperation (*P > 0.99), cardiac readmission (P= *0.566) and permanent pacemaker insertion (*P = *0.506) did not exhibit significant differences individually (see Supplementary Table S1).

Subgroup analysis

Subgroup analysis using age ≥65 years, sex, COPD, longstanding persistent A-fib, LA diameter, LVIDd and A-fib high-risk for long-term outcomes yielded consistent trends in the results. In the presence of most of these factors, the RMT group tended to reveal a lower HR (RMT favourable). However, significant P-values for interaction were just observed for age, COPD, LVIDd and A-fib high-risk. These factors exhibit significant interactions with the surgical approach related to secondary outcomes but not to primary outcome (see Supplementary Material, Fig. S4).

LVIDd showed statistically significant interactions with surgical approaches (P for interaction = 0.006) for the MAEs occurrence. Notably, a significant HR was observed in the subgroup with normal LVIDd [9] (HR: 0.42, *P = *0.002), indicating its impact. In the subgroup with abnormal LVIDd, the HR increased to 1.28 but did not reach statistical significance (*P = *0.403) (see Supplementary Material, Fig. S5).

A-fib high-risk factor demonstrated a significant interaction (P for interaction = 0.002) with surgical approaches for A-fib recurrence (Fig. 3). The RMT displayed an HR of 1.86 (*P = *0.048) with A-fib high-risk. Age ≥65 years (*P = *0.013) and COPD (*P = *0.031) had significant P-values for interaction and increased HR but did not show statistical significance in the presence of these risk factors (HR: 1.18, *P = *0.434; HR: 2.59, *P = *0.113) (Fig. 3).

DISCUSSION

We compared outcomes between the RMT and the sternotomy groups with concomitant MV and TV operations and surgical ablation analysing 2 decades of data using PSM. In the long term, none of the outcomes exhibited significant differences. Regarding early outcomes, the RMT group displayed results consistent with those of previous studies. These include prolonged CPB time, but reduced ventilation time, ICU stay and hospital stay. A subgroup analysis was conducted to examine the factors influencing outcomes based on surgical approaches. For patients with A-fib high-risk factors, sternotomy produced more favourable A-fib recurrence outcomes.

Several studies have reported that a minimally invasive approach is comparable to a sternotomy for concomitant MV and TV operations [2–4]. Given this background, we evaluated whether adding more surgical procedures, especially surgical ablation, to the minimally invasive approach would still yield comparable outcomes. We hypothesized that the RMT could yield comparable long-term outcomes in this operation. All long-term outcomes were compared before and after PSM. However, not all outcomes exhibited statistically significant differences after PSM. So in our study, it is challenging to draw conclusions regarding the long-term outcomes when comparing 2 approaches.

Nevertheless, we have also confirmed the following results in this study. Although the statistical power is limited, the HRs of the RMT group consistently were low. Subgroup analysis revealed significantly better outcomes for the RMT group compared to the sternotomy group in MAEs and A-fib recurrence, particularly in patients without COPD, under 65 years old, without A-fib high-risk and with normal LVIDd. Additionally, in early outcomes, there are significant advantages in terms of lower LCOS and ARF rates. These complications can potentially have adverse effects on long-term outcomes [13]. Considering all these results, we can draw 2 conclusions. First, in this study, the long-term outcomes of an RMT did not exhibit significant differences compared to those of a sternotomy. However, based on other results, the RMT could be a clinically reasonable option in this kind of procedure. Second, for patients with A-fib high-risk, sternotomy could be a better choice for A-fib recurrence outcomes.

In the whole cohort, the sternotomy group had unfavorable baseline characteristics. These results could lead us to argue that sternotomy was more frequently selected for high-risk patients due to its shorter CPB and ACC times [2, 14]. Therefore, although an RMT could be a reasonable approach in this study, there was a need to analyse risk factors affecting outcomes based on surgical approaches. Thus, we conducted subgroup analyses based on the approach.

Subgroup analysis revealed that patients with A-fib high-risk could potentially benefit from sternotomy regarding long-term A-fib recurrence. Some previous studies have reported that a more simplified lesion set for ablation is frequently accomplished when utilizing MICS for surgical ablation. Moreover, the chances of incomplete ablation are high in cases with a very large atrium or poor visibility [15, 16]. Nowadays, although RMT is considered a reasonable alternative to sternotomy in this kind of operation, there are specific circumstances or patient characteristics where choosing a sternotomy over an RMT might be valid. Therefore, the selection of the surgical approach should be carefully considered in specific groups, such as patients with A-fib high-risk.

Surgical ablation performed concomitantly during cardiac surgery has been well-established for its effectiveness [7]. As MICS has advanced, several studies have been conducted to investigate the efficacy of surgical ablation through MICS [16]. However, comparing MICS to a sternotomy is relatively rare. Even compared to a sternotomy, these outcomes cannot solely be attributed to the approach due to the inconsistency of the combined operations. [15]. Our study was meticulously designed to ensure uniformity in operations accompanied by ablation, and different factors were adjusted through PSM. This approach allowed us to comprehensively understand the effects of surgical ablation compared to RMT with a sternotomy.

Limitations

This study had certain limitations. It was a retrospective and observational study. The results could have been influenced by selection bias. The study encompassed patients who underwent concomitant MV and TV operations and surgical ablation performed by 8 surgeons over 20 years. Thus, unmeasured variables, such as variations in surgical expertise, learning curves, advancements in endoscopy for MICS and in ablation device, could not have been considered comprehensively. We conducted PSM to overcome these issues but there are still some standardized mean differences above 0.1 (LA diameter, LVIDs/d). Because all of these were under 0.15, they could potentially reduce statistical significance. Finally, we were unable to find statistical significance in the long-term outcome, thus failing to identify clear long-term differences between the 2 approaches. The sample size of our study was insufficient to achieve statistical power. Enrolling nearly 800 patients posed challenges, but conducting a study with an even larger number of patients (each group requires more than 1200 patients to achieve statistical power) is necessary for a thorough evaluation of the results. Further research involving large numbers of patients is necessary moving forward. Due to these limitations, we could not draw clear long-term conclusions. Instead, we made conclusions within clinically acceptable bounds.

It is clinically crucial to mention that this research was conducted in a high-volume centre where MICS is extensively performed. Randomized trials are warranted to confirm the results of this study.

CONCLUSIONS

The long-term outcomes of an RMT showed no significant differences compared to a sternotomy in concomitant MV surgery and TV surgery and surgical ablation. However, based on short-term outcomes and subgroup analysis, an RMT could be a clinically reasonable option for these concomitant procedures.

Patients with A-fib high-risk factors could have different A-fib recurrence outcomes depending on the surgical approach. If A-fib control is clinically important and the patient has A-fib high-risk factors, selecting a sternotomy might be preferable.

Supplementary Material