National outcomes of surgical ablation for atrial fibrillation at the time of mitral surgery

Authors: Aminah Sallam, Derrick Y Tam, Qiudong Chen, Allen A Razavi, Michael E Bowdish, Armin Kiankhooy, Joanna Chikwe

Abstract

OBJECTIVES

Current guidelines recommend concomitant surgical ablation at the time of mitral surgery for patients with atrial fibrillation; however, there is a paucity of data on long-term outcomes in this population. We sought to assess long-term clinical outcomes in patients undergoing concomitant surgical ablation at the time of mitral surgery.

METHODS

The United States Centers for Medicare and Medicaid data were used to identify patients undergoing mitral repair or replacement from 2015 to 2019. After excluding prior cardiac surgery, endocarditis and emergencies, we identified 11 410 patients undergoing isolated mitral repairs or replacement with pre-operative atrial fibrillation. Of these, 3268 (29%) received surgical ablation and 8142 (71%) did not. Propensity score matching was performed on 27 baseline characteristics. The primary outcome was freedom from death or stroke at 4 years. The secondary outcome was all-cause mortality at 4 years. Both were assessed using Cox-proportional hazard models.

RESULTS

Propensity matching yielded 3268 well-matched patient pairs (mean 74, 53% female, median CHA2DS2-Vasc score 4). There was no difference in all-cause mortality at 30 days (2.7% with concomitant ablation vs 2.8% without, P = 0.762). Patients undergoing concomitant ablation at the time of surgery had significantly higher freedom from death or stroke at 4 years (81% vs 77%, HR: 0.84, 95% CI 0.74–0.96). However, overall freedom from death between groups was not statistically significant (84% with concomitant ablation vs 82% without, HR: 1.14, 95% CI 0.76–1.01).

CONCLUSIONS

Surgical ablation at the time of isolated mitral surgery is underutilized but associated with improved long-term outcomes.

INTRODUCTION

Pre-operative atrial fibrillation (AF) is present in 30–50% of patients undergoing mitral surgery [1–3]. If left untreated, AF is associated with reduced short-term survival, increased risk of stroke and worse quality of life among these patients [4]. Recent work on concomitant surgical ablation at the time of mitral valve surgery revealed improved one-year mortality, increased rate of freedom from AF at one year and improved quality of life [1, 5–7]. However, concomitant surgical ablation at the time of mitral surgery was also associated with slightly longer cardiopulmonary bypass time, higher risk of implantation of permanent pacemakers (PPMs) and higher rates of in-hospital acute kidney injury [1, 8, 9]. Despite this, concomitant surgical ablation of AF at the time of mitral valve operations has received a class IA recommendation from the Society of Thoracic Surgeons largely on the basis of restoration of sinus rhythm alone [10, 11]. The most recent guidelines from the European Society of Cardiology/European Society of Cardiothoracic Surgeons strongly recommend surgical ablation at the time of left atrial open cardiac surgery based on findings suggesting improved long-term survival and outcomes in multiple single centre analyses [12].

While implementation of concomitant surgical ablation at the time of cardiac surgery is increasing [13], it remains under-treated, with an estimated 40% of patients with pre-operative AF undergoing mitral valve surgery not receiving concomitant ablation [4, 10]. Although concomitant surgery for AF at the time of surgical management for valvular heart disease has been shown to be safe and effective, there remains a paucity of data around long-term outcomes following concomitant surgical ablation at the time of mitral surgery at the national level. Thus, the purpose of this study was to compare long-term survival and event-free survival in patients with pre-operative AF receiving concomitant surgical ablation versus no ablation at the time of mitral surgery in a nationally representative cohort.

PATIENTS AND METHODS

Study overview

We conducted a retrospective cohort study comparing 4-year outcomes in patients with pre-operative AF undergoing first-time isolated mitral valve surgery receiving concomitant surgical ablation vs no ablation using the Centers for Medicaid and Medicare Services Fee-for-Service administrative database between 2015 and 2019 in the USA. This database includes information on beneficiaries’ encounters within the health care system in both the inpatient and outpatient settings, including receipt of therapeutic interventions like surgery. The database also includes information on patient demographics. Clinical comorbidities and outcomes are abstracted using the International Classification of Diseases, 10th Revision, Clinical Modification (ICD-10) diagnosis and procedures codes, and information on out of hospital deaths is acquired via linkage with the Master Beneficiary Summary File, which includes death data on all Medicare beneficiaries. Because the ICD-10 coding schema did not go into effect until 1 October 2015, we excluded patients in our dataset who underwent surgery prior to 1 October 2015, as we would be unable to accurately discern whether or not they received concomitant surgical ablation.

To identify our study cohort of patients undergoing isolated mitral valve repair or replacement, ICD-10 diagnosis and procedure codes (Supplementary Table S1) were used. To identify baseline clinical comorbidities, we looked at the ICD-10 diagnosis codes documented in the index hospitalization record and all inpatient admissions within the prior 2 years (Supplementary Table S2). Patients without a history of a given ICD-10 diagnosis code documented in any of their prior healthcare encounters were considered not to have that clinical comorbidity. After further exclusion of patients with infective endocarditis, any previous coronary artery bypass grafting or valve interventions, the presence of prosthetic heart valves, prior heart transplants or durable ventricular assist devices and patients without Medicare continuous enrolment or at least 1-year of Medicare coverage, we were able to identify our final study cohort of patients with pre-operative AF undergoing isolated mitral valve surgery that received concomitant surgical ablation or not (Fig. 1). The Centers for Medicaid and Medicare Service Data Protection Review Board and the Institutional Review Board at Cedars-Sinai Medical Center approved this study with a waiver of informed consent (STUDY00001188, approved on 2/19/2021).

Study end-points

The primary outcome was freedom from death or stroke at 4 years. The secondary outcome was all-cause mortality at 4 years. Other end-points included post-operative PPM implantation, major bleeding, index hospitalization length of stay and 30-day mortality. Deaths were identified from the Master Beneficiary Summary File, and non-fatal secondary end-points were defined using ICD-10 diagnosis and procedure codes (Supplementary Table S3). Stroke was defined as any cerebrovascular accidents documented during the index hospitalization (that was not present on admission) and any subsequent hospital admission where the principal diagnosis was haemorrhagic or ischaemic stroke. Patients without one of the above events of interest were censored on 31 December 2019. Follow-up time was calculated utilizing the reverse Kaplan–Meier method as the difference in time from surgery to the time where the patient experienced one of the above events of interest or the end of the study period—whichever came first.

Statistical analysis

Trends in utilization of concomitant surgical ablation in all mitral surgical patients with AF was assessed using linear regression. Baseline characteristics were first compared in the overall sample between those receiving concomitant surgical ablation and no ablation. Student’s t-test was used for normally distributed continuous variables, Wilcoxon rank-sum test for non-normally distributed continuous variables, while the Chi-squared test was used for categorical variables. Propensity score matching was performed to account for baseline differences in patient characteristics between patients who received concomitant surgical ablation and no ablation to reduce the effects of confounding. The propensity score for each patient was estimated using a multivariable logistic regression model in which the intervention performed and was regressed on 27 important baseline characteristics that may influence the choice of intervention or that were prognostically important for the outcome. Subjects were matched on the logit of the propensity score using 1 greedy nearest-neighbour matching with a calliper distance of 0.1 times the standard deviation of the logit of the propensity score [14, 15]. Success of matching was assessed by computing the standardized difference of each covariate with a cut-off of 0.1 to denote acceptable balance (Supplementary Fig. S1) [16]. Early events were compared between the two cohorts using the McNemar test for binary outcomes and paired t-test and the Wilcoxon signed rank test for normally and non-normally distributed continuous variables, respectively.

In the matched sample, equality of cumulative incidence function was assessed using a univariate Fine-Gray sub distribution hazard model in which the sub distribution hazard of the outcome was regressed on a single variable denoting treatment status, with a robust variance estimator to account for the matched nature of the sample [17]. For all-cause mortality, Kaplan–Meier survival curves was conducted in the matched sample, using a stratified log-rank test (stratified on the matched pairs) to test the equality of the estimated survival curves [18]. In addition, hazard ratios were estimated using a Cox-proportional hazards model, which incorporated a robust sandwich-type variance estimator to account for the matched nature of the data [19].

Falsification end-point analysis

To assess for the risk of potential unmeasured confounders following propensity score matching, we compared the association between treatment allocation (concomitant surgical ablation vs no surgical ablation) and the incidence of a composite falsification end-point of hospitalization for urinary tract infection or pneumonia [20]. We chose this end-point as it is unlikely to be related to treatment assignment. The cumulative incidence of this end-point was evaluated with death as a competing risk.

Statistical significance was assumed for P < 0.05. The primary outcome was tested initially, while there was no adjustment for multiplicity for secondary or early outcomes. All analyses were conducted with RStudio (version 1.3.959, RStudio: Integrated Development for R. RStudio, PBC, Boston, MA, USA).

RESULTS

Baseline characteristics

Of the 11 410 patients undergoing isolated mitral valve surgery with pre-operative AF, there were 3268 patients in the concomitant surgical ablation group and 8142 in the no surgical ablation group (Fig. 1). Prior to propensity matching, patients who underwent concomitant surgical ablation at the time of mitral surgery had fewer comorbidities than patients who did not (Table 1). Those who received concomitant surgical ablation were younger (74 years vs 75 years, P < 0.001), had lower rates of coronary artery disease (63% vs 69%, P < 0.001), decreased incidence of congestive heart failure (67% vs 76%, P < 0.001), decreased incidence of chronic kidney disease (23% vs 29%, P < 0.001) and were less likely to require urgent surgery (11% vs 14%, P < 0.001). However, the CHADS2-VASC score was similar between the patient groups (median 4.0, IQR 3.0–5.0). Median follow-up time of patients in the study period was 1.97 years (maximum: 4.0 years, interquartile range [IQR]: 1.91–2.05 years).

From 2016 to 2019, the percentage of patients with pre-operative AF undergoing concomitant surgical ablation at the time of their isolated mitral valve surgery increased from 35.8% to 46.8% (Supplementary Fig. S2). However, this trend was noted to not be statistically significant (P = 0.075).

Primary matched cohort

Propensity matching on 27 baseline covariates (including year of procedure) yielded 3268 pairs of patients, with 100% of concomitant surgical ablation patients being matched to a no ablation patient. These pairs were well-matched, with standardized mean differences <0.10 for all covariates.

Among the propensity matched patients, there was no difference in early 30-day mortality between patients who received concomitant surgical ablation and those who did not (2.7% vs 2.8%, P = 0.762). There was no difference in in-hospital stroke (2.0% vs 2.1%, P = 0.732) and in-hospital new PPM implantation (9.4% vs 8.4%, P = 0.101) between the ablation and no ablation groups. There was no difference in index length of hospitalization (median 8.0 days, IQR 6.0–11.0 days in both the groups, P = 0.138). Early outcomes before and after propensity matching are provided in Table 2.

At 4-year follow-up, the primary outcome of freedom from death or stroke was higher in the concomitant surgical ablation group (81% vs 77%, HR: 0.84, 95% CI: 0.74–0.96, P = 0.011, Fig. 2). There was no difference in 4-year survival between the two groups (84% vs 82%, HR 1.14, 95% CI: 0.76–1.01, P = 0.075, Fig. 3).

In our falsification end-point analysis, there was no difference in the 4-year cumulative incidence of urinary tract infection or pneumonia in the matched cohort (Supplementary Fig. S3, HR 1.03, 95% CI: 0.88–1.22, P = 0.71). This suggests little residual confounding in our matched cohort.

DISCUSSION

This national analysis of patients with pre-operative AF undergoing isolated mitral valve surgery with and without concomitant surgical ablation has several key findings. Among older patients with pre-operative AF who undergo mitral surgery, less than one in three receive surgical ablation. Concomitant surgical ablation can be safely performed at the time of mitral surgery without an increase in 30-day mortality or need for new PPM. Finally, concomitant surgical ablation at the time of mitral surgery was associated with a reduction in death or stroke at four years, but no difference in late survival. Together, these findings provide further support for consensus guidelines by demonstrating the longitudinal benefit of concomitant surgical ablation over no ablation for patients with AF presenting for mitral surgery.

Despite multiple consensus guidelines recommending concomitant surgical ablation at the time of mitral surgery, our findings show that less than one-third of patients with pre-operative AF undergoing mitral surgery received concomitant ablation. In our analysis, treated patients were typically younger with fewer comorbidities although their risk of stroke from AF was similar based on their CHADS-Vasc score. A national analysis done by Mehaffey and colleagues using the Society of Thoracic Surgeons National Database found that only 64.6% of patients with pre-operative AF undergoing mitral surgery received concomitant ablation, and the utilization of concomitant ablation decreased by 2.82% per year [21]. Like our study, this analysis found that surgical ablation was underutilized among patients with AF undergoing mitral surgery; they found that older patients were less likely to receive surgical ablation, which is consistent with our analysis.

There are many hypothesized reasons for the underutilization of surgical ablation during cardiac surgery, including concerns over the additional time on cardiopulmonary bypass required to perform the concomitant procedure, patient risk and unclear benefit in older patients [22]. Given these concerns and limited clarity on the long-term benefits of surgical ablation, some surgeons may elect to under-treat AF in older patients. Our findings confirm that concomitant surgical ablation is associated with long-term event-free survival among older patients with AF requiring mitral surgery.

Prior work has demonstrated no difference in 30-day mortality among cardiac surgery patients with AF undergoing concomitant surgical ablation for pre-existing AF compared to those not receiving treatment [23]. However, in a retrospective propensity matched analysis using the Polish National Registry found that patients who received surgical ablation at the time of mitral surgery were younger (63.8 ± 8.7 years vs 66.1 ± 9.0 years; P < 0.001), and had 20% improved survival at 12 years (HR 0.82, 95% CI: 0.70–0.96, P = 0.011) [24]. These prior studies were limited by a smaller sample size [23], or included primarily younger patients in the surgical ablation treatment group [23, 24]. Our study confirms these findings, but is unique in that it highlights the safety and long-term benefit of isolated surgical ablation for AF in a nationally representative sample of older patients. A recent analysis by Mehaffey and colleagues showed that surgical ablation with left atrial appendage closure was associated with improved survival in a heterogeneous population of coronary and valvular heart disease population. Our study exclusively focuses on a more homogenous patient population undergoing mitral valve surgery—this group of patients typically have large left atria and are at higher risk for AF related sequelae. The nature of mitral surgery, which involves opening of the atria tends to allow for more complete ablation procedures and closure of the left atrial appendage [25].

The 2024 European Heart Rhythm Association expert consensus guidelines strongly recommend the utilization of a biatrial Cox MAZE procedure at the time of left atrial open surgery on the basis of improved late outcomes from non-randomized single centre studies. In the updated 2023 Society of Thoracic Surgeons guidelines for patients with AF, there is Class I for surgical ablation in non-emergent cases given its long-term benefits. These recommendations are supported in our national analysis of patients >65 years for improved stroke-free survival [12, 26].

This study must be interpreted in the context of some important limitations. First, given our use of Medicare Data, the generalizability of these findings to a younger cohort may be more limited as our cohort was not only older but had multiple comorbidities as reflected by a median CHADS-VASC score of four. Additionally, while we used propensity matching to account for known confounders, unmeasured or unknown confounders may still exist. However, we were able to match all patients in our treatment group to a control patient, and the results of our falsification end-point analysis demonstrate that the association between the treatment exposure and study outcomes are likely unbiased by these unknown or unmeasured confounders. The administrative nature of this dataset precludes our ability to ascertain surgical ablation lesion sets, rhythm status during follow-up, and use of anti-coagulation during follow-up. We recognize that there is heterogeneity in how surgical ablation procedures are performed and defined by surgeons. Finally, due to inconsistencies in coding of left atrial appendage occlusion (LAAO) in the Medicare dataset, as well as a lack of information surrounding LAAO approach, we chose not to include LAAO as a covariate in our analysis.

In summary, concomitant surgical ablation at the time of mitral valve surgery was associated with improved event-free survival from death or stroke at 4 years in a nationally representative contemporary cohort of elderly patients. This study provides more evidence in support of current consensus guidelines recommending concomitant surgical ablation at the time of mitral surgery. More work is needed to prospectively determine the mechanistic and physiologic underpinnings of this observed phenomenon to better inform clinical practice.

Supplementary Material