Surgical ablation, left atrial appendage occlusion or both? Nationwide registry analysis of cardiac surgery patients with underlying atrial fibrillation

Abstract

Open in new tabDownload slide

OBJECTIVES

The aim of this study was to evaluate in-hospital outcomes and long-term survival of patients undergoing cardiac surgery with preoperative atrial fibrillation (AF). We compared different strategies, including no-AF treatment, left atrial appendage occlusion (LAAO) alone, concomitant surgical ablation (SA) alone or both.

METHODS

A retrospective analysis using the KROK registry included all patients with preoperative diagnosis of AF who underwent cardiac surgery in Poland between between January 2012 and December 2022. Risk adjustment was performed using regression analysis with inverse probability weighting of propensity scores. We assessed 6-year survival with Cox proportional hazards models. Sensitivity analysis was performed based on index cardiac procedure.

RESULTS

Initially, 42 510 patients with preoperative AF were identified, and, after exclusion, 33 949 included in the final analysis. A total of 1107 (3.26%) received both SA and LAAO, 1484 (4.37%) received LAAO alone, 3921 (11.55%) SA alone and the remaining 27 437 (80.82%) had no AF-directed treatment. As compared to no treatment, all strategies were associated with survival benefit over 6-year follow-up. A gradient of treatment was observed with the highest benefit associated with SA + LAAO followed by SA alone and LAAO alone (log-rank P < 0.001). Mortality benefits were reflected when stratified by surgery type with the exception of aortic valve surgery where LAAO alone fare worse than no treatment.

CONCLUSIONS

Among patients with preoperative AF undergoing cardiac surgery, surgical management of AF, particularly SA + LAAO, was associated with lower 6-year mortality. These findings support the benefits of incorporating SA and LAAO in the management of AF during cardiac surgery.

INTRODUCTION

Atrial fibrillation (AF) is a prevalent cardiac arrhythmia associated with significant morbidity and mortality [1, 2]. While pharmacological therapies remain the mainstay of AF management, surgical interventions have gained attention as treatment options. Surgical ablation (SA), performed adjunctively during cardiac surgery, has long shown promise in restoring sinus rhythm and reducing AF recurrence [3]. This approach involves creating linear and point lesions, using radiofrequency and/or cryoablation to disrupt abnormal electrical pathways responsible for the arrhythmia. Previous studies have demonstrated that SA can significantly improve freedom from AF and reduce the need for antiarrhythmic medications [4]. Additionally, it has been associated with improved long-term outcomes [5–8] and a lower incidence of stroke and lower mortality compared to medical therapy alone [9].

The left atrium (LA) appendage is a common site for blood stasis and thrombus formation, which can lead to stroke in patients with AF. Left atrial appendage occlusion (LAAO) aims to minimize the risk of thromboembolic events, particularly stroke, in AF patients [10]. Surgical techniques for LAAO include ligation, stapling, or external device-based occlusion. The recently published randomized trial showed a significant reduction in ischaemic stroke among patients undergoing coronary artery bypass grafting (CABG) surgery with concomitant LAAO [10]. Finally, combining SA with LAAO has emerged as a potentially synergistic therapeutic approach. This combined method aims to address both the underlying arrhythmogenic substrate and the source of thromboembolic risk, providing comprehensive treatment for AF patients.

To date, no prospective study compared those 2 therapeutic surgery-based strategies. While the adoption of concomitant LAAO is growing, the prevalence of concomitant SA remains low. Due to perceived complexity of SA and concerns regarding prolonged cross-clamp time and/or increased risk of permanent pacemaker implantation, together with LAAOS study results supporting adjunctive LAAO, some surgeons might opt for LAAO alone in the AF population. Recent evidence, based on Medicare population, has suggested that SA together with LAAO, when compared to LAAO alone, might had been associated with improved survival [11]. The current study aims to compare the outcomes of different surgical AF treatment strategies, including no-AF treatment, concomitant SA alone, concomitant LAAO alone, or a combination of both with regard to in-hospital outcomes and long-term mortality.

METHODS

Ethical statement

The study was approved by the local bioethics committee of the Medical University of Białystok (approval no. R-1-002/398/2019), due to anonymity of patient’s data written consent was waived.

KROK registry

Data were collected from the KROK (Polish National Registry of Cardiac Surgery Procedures) registry (available at: www.krok.csioz.gov.pl). The registry is an ongoing, nationwide, multi-institutional registry of heart surgery procedures in Poland. Centres enrolling patients in the KROK registry are required to transfer the data concerning every cardiac surgery to the central database in the National Centre for Healthcare Information Systems at the Ministry of Health and are financially liable for data integrity and completeness. Follow-up data regarding mortality were obtained from the National Health Fund—the nationwide, obligatory, public health insurance institution in Poland and incorporated to the registry [5, 7].

Study population

We included all consecutive adult patients (age >18 years), undergoing any cardiac surgery procedure and diagnosed with concomitant AF between January 2012 and December 2022. Patients were excluded if any of the following conditions applied: (i) undergoing transcatheter procedures; (ii) history of prior cardiac surgery; (iii) endocarditis; (iv) cardiogenic shock; (v) acute aortic dissections; (vi) preoperative mechanical circulatory support; and (vii) heart and heart-lung transplantations. Non-sternotomy surgical access was eligible for inclusion and recorded. For patients undergoing surgery, we considered and reported 3 categories of variables as potentially influencing the primary end point: (i) demographics and preoperative conditions: age, gender, EuroSCORE II and its single components; (ii) extent of coronary artery disease and/or valvular and/or aortic disease and (iii) surgical variables: urgency, operative technique (e.g. on-pump versus off-pump for CABG surgery).

Endpoints

The primary end point was survival at follow-up as divided by 4 groups: (i) no-AF treatment, (ii) SA alone, (iii) LAAO alone and (iv) SA + LAAO. Secondary outcomes included early survival (<24-h and 30-day mortality rates), in-hospital complications and lengths of stays in the intensive care unit (ICU) and hospital length of stay (HLoS).

Statistical analyses

Records with >5% of missing data were excluded, the remaining missing data were handled with artificial neural networks. Categorical variables were expressed as frequencies with percentages, while continuous variables were assessed for normality and presented as mean with standard deviation or median with interquartile range as appropriate. The Kruskal–Wallis test with Bonferroni correction was used to determine whether there were any statistically significant differences in continuous variables in multiple subgroups comparison. Chi-squared test for independence or Fisher’s exact test was performed to detect differences in categorical variables. Two-way Kolomogorov–Smirnov test along with multiple sample test of means was further used to confirm the previous. To address selection bias adherence to retrospective analysis, we used inverse probability of treatment weighting (IPTW) in the multiple treatment scenarios using propensity scores. The IPTW model consisted of relevant demographic and operative variables, listed in Supplementary Material, Fig. S1. We used a mps function of the twang package in R, which uses generalized boosted model to calculate propensity scores in the multiple treatment scenario. In principle, with the estimand being the average treatment effect, propensity score of a subject given treatment ‘t’ is calculated as a probability of receiving treatment ‘t’ versus any other treatment [12]. For the weighted statistical tests in the adjusted cohort, we utilized the survey package in R. Weighted odds ratios were calculated for in-hospital outcomes. Sensitivity analyses were conducted to assess the robustness of the results, considering the type of procedure performed and pre-specified patient subgroups. Absolute standardized mean differences were used to assess the balance before and after weighting, reflecting the maximum pairwise difference between any 2 groups in the multiple treatment scenario.

Additionally, a separate analysis was performed to evaluate the impact of adding LAAO to aortic valve (AV) surgery. Propensity scores were estimated based on pretreatment variables, and nearest-neighbour matching was conducted.

Robust Cox proportional hazard ratios (HRs) were calculated, and Kaplan–Meier curves were employed to estimate and visualize the survival function over a 6-year period. For the HR calculations after each year, the follow-up was truncated at specified intervals, with HRs calculated for these periods. Proportional hazards assumption was tested using Schoenfeld residuals and visually assessed.

Computations were conducted using R (twang and survey packages, R Core Team 2021, Vienna, Austria), STATA (StataCorp. 2021. Stata Statistical Software: Release 17. College Station, TX: StataCorp LLC) and SPSS (IBM SPSS Statistics, version 29).

RESULTS

Patients’ characteristics

Overall, 42 510 operated patients with preoperative AF were identified; of these, 33 949 were included in the final analysis. The study flow is illustrated in Fig. 1. Of the included patients 1107 (3.26%) received both SA and LAAO, 1484 (4.37%) received LAAO alone, 3921 (11.55%) SA alone and the remaining 27 437 (80.82%) had no AF-directed treatment. Patients who received concomitant AF treatment were generally younger and had a lower comorbidity burden than the control group (Table 1 lists all baseline characteristics). Also, they more frequently underwent elective procedures, especially mitral valve or multiple valve surgeries (Fig. 2). Supplementary Material, Table S1 lists unadjusted operative characteristics, while Supplementary Material, Tables S2 and S3 show statistical differences between treatment arms. Following IPTW patients were well balanced with respect to baseline characteristics (Supplementary Material, Figs S1 and S2). The prevalence of AF treatment based on the index procedure is depicted in Fig. 2. Patients who underwent SA and/or LAAO were more likely to have minimally invasive procedures. Cardiopulmonary bypass and aortic cross-clamp times were 20 and 22, 23 and 16, 30 and 26 min longer in the LAAO alone, SA alone and LAAO + SA groups, respectively. Among the patients who received SA, 1792 (35.6%) underwent pulmonary vein isolation, 1766 (35.1%) underwent LA box lesion procedures, while in 138 (2.7%) cases, both LA box lesions and pulmonary vein isolation were achieved. Additionally, 1332 (26.5%) patients underwent a concomitant Cox Maze procedure.

Risk-adjusted outcomes

Table 2 outlines in-hospital complications. Patients with SA more often experienced pulmonary complications, extended hospital stays and prolonged ICU stays (>48 h). No significant differences were observed in neurological complications, acute kidney injury (AKI), sternal wound infection, bleeding requiring reoperation, the need for postoperative extracorporeal membrane oxygenation (ECMO) and 24-h mortality. All strategies showed improved 30-day mortality compared to the control group, with statistical significance observed only for SA + LAAO. The incidence of in-hospital complications within ‘pseudopopulations’ after weighting along with the P-values for corresponding statistical test are presented in Supplementary Material, Table S4. Supplementary Material, Tables S5 and S6 report weighted mean- and percentage differences, respectively, for the clinical outcomes. Table 3 shows Cox proportional HRs for 6-year mortality. None of the investigated comparisons violated the proportional hazard assumption besides the one of SA + LAAO versus SA alone (graphical assessment of Schoenfeld residuals, along with the proportional hazard tests are available as Supplementary Material, Fig. S3). All strategies were associated with enhanced survival compared to no treatment. A treatment gradient emerged, with the strongest effect seen in combined treatment and the weakest in LAAO alone. Six-year Cox proportional HRs were: 0.87; 95% confidence interval (CI), 0.76–0.99; P = 0.036, 0.69; (0.64–0.75); P < 0.001 and 0.61; (0.51–0.72); P < 0.001 for LAAO alone, SA alone and LAAO + SA groups, respectively (Fig. 3A), with direction and magnitude of the effect sustained after censoring 1-year mortalities (Fig. 3B). Combined treatment outperformed each strategy individually (0.83 [0.69–1.00]; P = 0.054 compared to SA alone and 0.70 [0.56–0.86]; P = 0.001 compared to LAAO alone), while SA alone yielded better outcomes than LAAO alone (0.83 [0.72–0.97]; P = 0.020) (Supplementary Material, Table S7). Restricted mean survival times for all comparator arms are further available in Supplementary Material, Table S8. Since the distribution of treatment over time was non-homogenous, an additional IPTW was performed including temporal variable and the exact procedure type. The Kaplan–Meier curve along with Cox proportional hazard ratios for years 1–6 are available in Supplementary Material, Fig. S4 and Supplementary Material, Table S7, respectively (the covariate balance after IPTW is shown in Supplementary Material, Fig. S2). The results were consistent with the initial analysis.

Subgroup analysis

Figure 4 presents subgroup analysis based on the type of surgery. Survival tendencies were consistent across subgroups: CABG, mitral valve replacement or repair (MVR/r), tricuspid valve replacement or repair (TVR/r), CABG + valve and multiple valve (all log-rank P < 0.001), except for the AV surgery subgroup, where LAAO alone was associated with worse long-term survival when compared to the no-AF treatment (log-rank P < 0.001). The above was confirmed in an interaction analysis (Supplementary Material, Table S9). We performed an additional propensity score matching analysis in the subgroup of AV surgery between isolated LAAO and no-treatment groups. The Kaplan–Meier survival curves are presented in Supplementary Material, Fig. S5, whereas the covariate balance is shown in Supplementary Material, Table S10. The result remained significant after matching with HR for LAAO: 1.28; 95% CIs (1.01–1.62); P = 0.041. In the subgroup analyses, the survival benefit was consistent across diverse patient populations. Supplementary Material, Figs S6–S8 display the HRs with CIs for each treatment group compared to no treatment in prespecified subgroups. Additionally, the P-values for interaction with groups are provided.

DISCUSSION

Our study is the first to address comparative long-term survival outcomes between 4 different surgical treatment strategies for AF in patients undergoing general cardiac surgery. Main findings are as follows: (i) there is a low prevalence of SA and LAAO, particularly in non-mitral surgery, suggesting a potential underutilization of this procedure despite guidelines’ endorsement; (ii) there was a survival benefit with either strategy: SA alone, LAAO alone and SA + LAAO over no-AF treatment in terms of 30-day- and long-term survival; (iii) a gradient in the treatment effect was observed with SA + LAAO being associated with highest survival benefit, followed by SA alone and LAAO alone.

Left Atrial Appendage Occlusion Study (LAAOS III) randomized 4770 AF patients to undergo LAAO or not, at the time of their cardiac surgery for another indication [10]. At 3.8 years, LAAO was associated with roughly one-third lower risk of ischaemic stroke or systemic embolism HR 0.67; 95% CI, (0.53–0.85); P = 0.001 but no difference in survival [10]. Earlier, Friedman et al. [13] in the analysis of STS database demonstrated that, among older patients with AF undergoing cardiac surgery, concomitant LAAO, compared with no LAAO was associated with a lower risk of readmission for thromboembolism (0.67 [0.56–0.81]; P < 0.001), all-cause mortality (0.88 [0.79–0.97]; P = 0 0.001) and the composite end point (0.83 [0.76–0.91]; P < 0 0.001) over 3 years. Only recently, another registry was made available, this time focusing on 100 000 Medicare Beneficiaries, and found that, at 3 years, LAAO alone was associated with lower readmissions for stroke (HR 0.73, P < 0.001), mortality (0.75, P < 0.001) and composite outcome (0.79, P < 0.001) [11]. Given the above, LAAO at time of heart surgery, partly because of relative simplicity, has gained wide attention and is performed more often for thromboembolic protection in patients with AF.

SA of AF at the time of cardiac surgery has been associated with clear improvements in short and long-term survival as well as reduction in thromboembolism across the entire populations [9]. Rates of concomitant SA have, therefore, increased along with the emergence of supportive evidence both from randomized trials and registries [6, 9]; however, certain obstacles to wider adoption remain, and in particular in non-mitral surgery: CABG, AVR and aortic surgery [14]. Besides reimbursement strategies, that only recently have been supporting SA at time of CABG in Poland, 1 important barrier to SA at time of heart surgery is incomplete understanding of the added value of SA in some patients as also noted in the Medicare analysis [11].

One of the most important findings of the current analysis is the superiority of LAAO with SA and SA alone over LAAO alone. This finding can be explained by the additional benefit of maintaining sinus rhythm as a protective factor in heart failure progression, whereas LAAO provides antithrombotic protection only. Indeed, it has been shown in heart failure (HF) patients that SA leads to improvement in left ventricle ejection fraction and health-related quality of life [15].

On the contrary, recent analysis of Medicare patients shown an increased risk for readmission for HF in the SA + LAAO group, compared to no-treatment group. However, as the authors point out, the broad CMS claims diagnosis of heart failure precludes any conclusions. Nevertheless, the potential of SA in mitigating heart failure progression requires further investigation.

Unlike the previous analysis that compared LAAO alone versus SA + LAAO versus no-AF treatment, the current corroborates these findings by adding the SA alone group. It has to be stressed that previous studies included, be design, patients undergoing SA, however detailed analyses are not so far available [11]. On contrary, we were able to demonstrate a gradient arising in terms of long-term survival with all the 4 approaches. Main finding of the current analysis is a superior survival observed in the SA + LAAO group compared to any other group. This supports the potential synergistic benefits of addressing both the arrhythmogenic substrate and reducing the risk of thromboembolic events [4, 5] with potential of weaning of the antiarrhythmic drugs and oral anticoagulation during follow-up. Together with information on early safety as expressed by no difference in in-hospital complications and superior 30-day survival, this current analysis highlights the primary advantage of performing SA + LAAO over SA alone and LAAO alone also for survival benefit. Indeed, an increase in CPB and X-clamp times, elevating the peri-operative risks, in-hospital complications and longer HLoS duration are frequent reasons for no-AF treatment in patients with AF.

A novel finding in our study is the non-favourable prognosis of LAAO alone performed at time of AV surgery and there are several potential reasons for this. Firstly, aortic stenosis, by nature, places increased stress on the heart, leading to development of heart failure; LAAO may interfere with the heart's ability to regulate fluid volume properly due to attenuation of atrial and brain natriuretic peptide release, in turn, leading to volume overload and heart failure’s acceleration [16]. Secondly, the LAA constitutes to about 30% of the LA's volume and has much higher compliance compared to the rest of the LA [17]. It plays a crucial role in protecting the pulmonary circulation from high pressure during LA volume overload. A study by Tabata et al. [18] demonstrated that LAAO led to an increase in mean LA pressure and maximal LA dimension. This proved the importance of the LAA in atrial reservoir function. Consequently, LAAO may contribute to increased pulmonary congestion, which could be particularly problematic in aortic stenosis, where this process might be further amplified [19]. Indeed, a study by Kim et al. [20] found that hospitalization for heart failure was not uncommon after LAAO, and the predictor for this was the left ventricular mass index, which is significantly increased in aortic stenosis patients. Additionally, the LAAO III study demonstrated a higher rate of hospitalization for heart failure among patients with LAAO, although the increase was not statistically significant (7.7% in the occlusion group vs 6.8% in the no-occlusion group). Regarding survival, 1 study indicated a slightly higher mortality following LAAO in isolated AVR, although this increase was not statistically significant [21]. Further research and clinical assessment are necessary to fully understand the effects of LAAO during AV surgery and its potential impact on patient prognosis. Hopefully, the ongoing LAA-CLOSURE randomized controlled trial, which aims to evaluate the results after preventative LAAO during isolated AVR, will offer a definitive answer [22].

Regarding perioperative complications, we found that the rates of AKI were not higher in the LAAO + SA, SA and LAAO group compared to the no-treatment group. It is important to note that our results may differ from previous studies conducted on different patient populations. Several studies focusing on the Medicare population have reported the association between SA and AKI [6, 10]. However, it is worth noting that the Medicare population was older and more likely to have had underlying chronic kidney disease compared to the population in our study. Chronic kidney disease itself is a significant risk factor for the development of AKI, and the presence of comorbidities in the Medicare population may contribute to the observed differences in AKI rates. In our cohort, we observed higher rates of pulmonary complications which encompass pulmonary congestion and pleural effusion in the SA groups. It has been previously observed that SA may increase the risk for pulmonary complications, however, they are mostly temporary and may be effectively managed with appropriate pharmacological therapy [23]. We also observed higher HLoS in the SA groups and prolonged ICU stay. One explanation for this may be a transient rise in cardiac troponins observed after SA, which may prompt clinicians to prolong observation before discharge. However, studies show that the elevated troponins after SA do not impair survival, on contrary, 1 study showed that higher levels of troponins are associated with greater reversal of structural remodelling after SA [24].

Limitations

Our study has several limitations that should be considered when interpreting the results. First, we lacked some baseline echocardiographic data and heart rhythm at follow-up. Second, we did not have access to detailed information about the drug regimens used by the patients or their adherence to drug therapy in particular oral anticoagulation and anti-arrhythmics. Third, KROK registry did not capture data on subsequent catheter ablation or percutaneous LAAO procedures. Fourth, the only available long-term outcome was all-cause mortality; data on stroke rates and heart failure progression is missing to the registry. Finally, the registry, at a time of the analysis did not provide information regarding the AF type and duration nor the ablation energy sources used in SA procedures. Furthermore, no information was available regarding the method employed for LAAO (e.g. ligation, amputation, clipping).

CONCLUSIONS

The current study highlights the underutilization of SA, particularly in non-mitral procedures, and demonstrates the superior survival outcomes associated with concomitant SA combined with LAAO. SA alone appears to offer advantages over LAAO alone, potentially due to the protective effect of maintaining sinus rhythm. Importantly, the rates of perioperative AKI were not elevated in the SA group, providing reassurance regarding its safety. These findings contribute to the existing knowledge on surgical AF treatment strategies and emphasize the potential benefits of incorporating SA and LAAO in the management of AF during cardiac surgery.

A complete list of KROK Investigators is available in the Supplementary Material.

Presented at EACTS Annual Meeting 2023, Vienna, Austria, 5 October.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

FUNDING

The study did not receive any external funding.

Conflict of interest: none declared.

DATA AVAILABILITY

The data underlying this article will be shared on reasonable request to the corresponding author.

Author contributions

Michał Pasierski: Data curation; Formal analysis; Investigation; Methodology; Writing—original draft; Writing—review & editing. Jakub Batko: Data curation; Formal analysis; Methodology; Validation; Writing—review & editing. Łukasz Kuźma: Conceptualization; Methodology; Supervision; Validation; Writing—review & editing. Wojciech Wańha: Conceptualization; Methodology; Supervision; Writing—review & editing. Marek Jasiński: Conceptualization; Methodology; Supervision; Writing—review & editing. Kazimierz Widenka: Conceptualization; Methodology; Supervision; Writing—review & editing. Marek Deja: Conceptualization; Methodology; Supervision; Writing—review & editing. Krzysztof Bartuś: Conceptualization; Methodology; Supervision; Writing—review & editing. Tomasz Hirnle: Conceptualization; Methodology; Supervision; Writing—review & editing. Wojciech Wojakowski: Conceptualization; Methodology; Supervision; Writing—review & editing. Roberto Lorusso: Conceptualization; Methodology; Project administration; Supervision; Writing—review & editing. Zdzisław Tobota: Conceptualization; Project administration; Supervision; Writing—review & editing. Bohdan J. Maruszewski: Conceptualization; Methodology; Project administration; Software; Supervision; Writing—review & editing. Piotr Suwalski: Conceptualization; Investigation; Methodology; Project administration; Supervision; Writing—review & editing. Mariusz Kowalewski: Conceptualization; Data curation; Formal analysis; Investigation; Methodology; Project administration; Software; Supervision; Validation; Visualization; Writing—original draft; Writing—review & editing.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Anders Albage and the other anonymous reviewers for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• AF

  Atrial fibrillation

 
• AKI

  Acute kidney injury

 
• AV

  Aortic valve

 
• CI

  Confidence interval

 
• HLoS

  Hospital length of stay

 
• HRs

  Hazard ratios

 
• ICU

  Intensive care unit

 
• IPTW

  Inverse probability of treatment weighting

 
• LA

  Left atrium

 
• LAAO

  Left atrial appendage occlusion

 
• SA

  Surgical ablation

Author notes