Novel surgical ablation through a septal-superior approach for valvular atrial fibrillation: 7-year single-centre experience^†

Abstract

OBJECTIVES

We previously reported favourable short-term results of our ‘trans-septal maze procedure’, a novel technique for creating biatrial lesions through a septal-superior approach during mitral valve surgery. Here, we reviewed the mid-term results of this procedure and determined the impact of restored left atrial (LA) contraction on late outcomes.

METHODS

We examined clinical data of 50 patients with persistent (n = 7) or long-standing persistent atrial fibrillation (AF) (n = 43) (mean period of rhythm disturbance 77 ± 78 months) who underwent a trans-septal maze procedure concomitant with mitral valve surgery and were followed postoperatively for at least 24 months. The mean preoperative LA dimension was 59 ± 9 mm (40–85 mm). The presence of an A wave in Doppler echocardiography was considered to indicate evidence of LA mechanical contraction. Serial echocardiography was performed to evaluate left ventricular and LA dimensions, degree of valvular regurgitation and estimated systolic pulmonary artery (PA) pressure. Follow-up was completed with a mean duration of 59 ± 17 months (27–92 months).

RESULTS

There were no ablation-related complications and 48 patients (96%) were free from AF immediately after the operation. At the latest follow-up, 39 patients (78%) were free from AF, while 28 (56%) presented LA mechanical contraction. Patients who restored LA mechanical contraction were less likely to experience postoperative thromboembolic events (4 vs 23%, P = 0.075), as compared with those who did not restore it. Serial echocardiography showed that patients with restored LA contraction showed improvement in Doppler-derived systolic PA pressure to a greater degree and less incidence of significant tricuspid regurgitation (7 vs 41%, P = 0.006). The Cox proportional hazards models with adjustments for all other covariates revealed LA dimension >60 mm at baseline as an independent risk factor for lack of LA mechanical contraction (adjusted hazards ratio 3.9, 95% confidence interval 1.1–14, P = 0.035).

CONCLUSIONS

Our trans-septal maze procedure may be an effective alternative surgical treatment for eliminating AF during mitral valve surgery. In patients with valvular AF, early surgery is warranted to restore sinus rhythm with LA mechanical contraction, before severe LA dilatation occurs. The impact of LA contraction recovery conferred by AF ablation on postoperative haemodynamic improvements and thromboembolic events remains to be determined.

INTRODUCTION

Atrial fibrillation (AF), the most significant type of arrhythmia seen in patients with mitral valve disease, impairs cardiac function, and is associated with an increased risk of systemic embolization and mortality [1]. The maze procedure, initially described by Cox et al., has been proven to be safe and extremely effective for surgical treatment of AF, and is most often used in patients undergoing mitral valve surgery to enhance the probability of postoperative restoration of sinus rhythm [2].

Optimal exposure of the mitral valve and sub-valvular apparatus plays a key role in successful mitral valve surgery. However, that can be particularly difficult to secure by a conventional right-sided left atriotomy in some patients with a deep chest, left ventricular (LV) hypertrophy, or a relatively small left atrium. This concern can often be resolved by a trans-septal incision or extended trans-septal approach (i.e. septal-superior approach), as described by Guiraudon et al. [3], and has become increasingly popular for both minimally invasive and complex mitral valve operations. We have used a septal-superior approach as the first choice for mitral valve surgery since 2005, and perform concomitant ablation in patients with AF using a combination of bipolar radiofrequency and cryothermy. Although the outcome of ablation surgery via a conventional left atriotomy has been extensively investigated, there are scant data available regarding the efficacy of a biatrial surgical ablation via a septal-superior approach for valvular AF. The purpose of this study was to investigate the short- and long-term outcomes of a modified maze procedure performed in a concomitant manner with mitral valve surgery via a septal-superior approach. Furthermore, we attempted to determine the impact of restored left atrial (LA) mechanical contraction on late outcomes.

MATERIALS AND METHODS

We retrospectively analysed the clinical data for 53 consecutive patients who underwent a trans-septal maze procedure concomitant with mitral valve surgery between March 2005 and December 2009. Of those, 3 were excluded because of non-cardiac-related death within 24 months after surgery. The final study population consisted of 50 patients and the baseline characteristics and surgical data are presented in Table 1. There were 32 males (64%) and the mean age was 65 ± 8 years (47–85 years). All patients presented persistent (n = 7) or long-standing persistent AF (n = 43) at the time of surgery with a mean period of rhythm disturbance of 77 ± 78 months (range, 3–300 months). The mean LA dimension was 59 ± 9 mm (40–85 mm). We did not set exclusion criteria for patients who underwent this procedure during the study period. For mitral valve surgery, valve replacement was performed in 14 (mechanical valve 12, biological 2), mitral valve repair in 34 and open mitral commissurotomy in 2. Nearly all of the patients (98%) underwent a concomitant tricuspid annuloplasty. The final study protocol was approved by an institutional ethical committee and all patients provided informed consent.

Echocardiography

Two-dimensional and Doppler echocardiography procedures were performed by expert echocardiographic examiners using an Aplio 80 (SSA-770A) or ARTIDA (SSH-880CV) ultrasound system (Toshiba Medical Systems Corporation, Tochigi, Japan) equipped with a 3.0-MHz (PST-30BT) or 2.5-MHz (PST-25SX) transducer prior to surgery (baseline), 1 month after surgery and annually thereafter. Anatomic and Doppler measurements were performed according to the recommendations of the American Society of Echocardiography. Systolic pulmonary artery (PA) pressure was calculated by adding the systolic pressure gradient across the tricuspid valve, derived from tricuspid regurgitation, to the estimated right atrial pressure value. The severity of regurgitation was classified as none (0), trivial (1+), mild (2+), moderate (3+) or severe (4+). Transmitral peak velocities of the early (E) and late (A) filling waves were measured in the apical view. The presence of an A wave on the tracings was taken to indicate the presence of an LA mechanical contraction [4–6].

Surgical intervention

Our trans-septal maze procedure has been described in detail [7]. Atrial incisions, ablation lines and sites of cryothermia utilized with our procedure are shown in Fig. 1. In brief, pulmonary vein (PV) isolation was performed under a standard cardiopulmonary bypass using a bipolar device (Atricure^®, Century Medical, Inc., Tokyo, Japan; or Medtronic Cardioblate System, Medtronic, Inc., Minneapolis, MN, USA) (Fig. 1A and B-a,b). After a right atrial incision, bipolar radiofrequency was used to connect the lesion toward the tricuspid annulus from the right atrial incision (Fig. 1A and B-c), with a cryo-lesion applied at the tricuspid annulus anterior portion using a reusable cryoprobe (Frigitronics Cardiac Cryosurgical System; Tonokura Ika, Co., Ltd, Tokyo, Japan) at −60°C for 2 min. Then, from the inferior aspect of the vertical right atriotomy, a bipolar clamp was placed on the superior vena cava (Fig. 1A and B-d), and rotated and extended in a similar fashion onto the inferior vena cava (Fig. 1A and B-e).

The heart was then arrested and the LA appendage amputated. A connecting lesion was created from the amputated appendage into the ablation lines isolating the left PVs (Fig. 1A and B-f). Next, the atrial septum was incised vertically through the fossa ovalis and the right atriotomy extended in a superior manner to join the superior end of the septal incision, then the two joint incisions were extended onto the LA roof in a transverse manner. The orifice of the LA appendage was oversewn using a purse string suture from the inside of the left atrium. Two connecting lesions were created from the LA roof to the left (Fig. 1A and B-g) and right superior PVs (Fig. 1A and B-h and C), respectively. A small incision was then made in the distal site of the right inferior PV and an ablation line was created from the small incision towards the posterior portion of the mitral annulus using the bipolar device (Fig. 1A and B-i and C). A cryo-lesion was added at the posterior portion of the mitral annulus. Subsequently, mitral valve surgery or other combined procedures were performed.

Postoperative management

Postoperative care was consistent with that given following routine cardiac surgical procedures. All patients were monitored continuously for arrhythmias until the time of discharge from the hospital. AF was defined as a rapid irregular rhythm with disorganized atrial activity in which distinct P waves were absent in the standard 12-lead electrocardiogram. Postoperative atrial arrhythmias were treated with either class I or class III anti-arrhythmic drugs. A direct-current cardioversion was performed if deemed necessary for patients who did not convert into sinus rhythm under medical treatment. For patients with episodes of postoperative AF, the same regimen of anti-arrhythmics was continued for 3–6 months after discharge. All patients routinely received oral warfarin administration for at least 3 months regardless of the type of mitral valve surgery and postoperative rhythm. That was discontinued within 3 to 6 months for patients who underwent valve plasty or replacement using bioprosthesis and had no recurrence of AF.

Clinical follow-up examinations

Every 6 months to 1 year, each patient was assessed in our department and/or by their primary cardiologist, and underwent complete physical, standard 12-lead electrocardiography and 24-h Holter monitoring (indicated for patients with symptomatic or unknown arrhythmia) examinations. Current information was obtained by calling the patient or referring cardiologist. Clinical follow-up examinations were completed for all (100%), with a mean duration of 59 ± 17 months (27–92 months).

Statistical analysis

Continuous variables are presented as the mean ± standard deviation (SD). Categorical variables are shown as frequencies and proportions. Continuous variables were compared using analysis of variance (ANOVA) or a Mann–Whitney U-test, as appropriate. Categorical variables were compared using χ² analysis or Fisher's exact test, as appropriate. Echocardiographic variables over time were analysed by repeated-measures ANOVA, with group, time and group–time interaction effects.

The predictors for lack of LA mechanical contraction were examined with Cox proportional hazards models, with adjustments for baseline demographics, medications, echocardiographic and surgical data. Factors obtaining a P-value <0.1 in the univariate analysis were then entered appropriately in a multivariate fashion. Results are summarized as hazard ratios and 95% confidence intervals (CIs). All P-values are two sided and values of P < 0.05 were considered to indicate statistical significance. Statistical analyses were performed using JMP 7.0 (SAS Institute, Cary, NC, USA), SAS statistical software (version 9.2, SAS Institute) and SPSS (version 17.0, SPSS, Inc.).

RESULTS

Hospital outcomes

There were no intraoperative ablation-related complications and 48 patients (96%) were free from AF immediately after the operation, while 2 (4%) never recovered from AF. Among those 48 patients, 42 (88%) experienced one or more episodes of atrial arrhythmia in the early postoperative phase (<30 days) after surgery, while the other 6 did not have any atrial arrhythmia. The initial onset of early atrial arrhythmia was unexceptionally identified within the first 14 days after surgery, with a peak incidence on postoperative day 1. Some patients spontaneously recovered from AF, or after being treated with an anti-arrhythmic drug or direct cardioversion and finally 34 patients (68%) were free from AF at discharge. The average duration of atrial arrhythmia during hospitalization was 10 ± 8 days.

Ten patients (20%) required temporary ventricular pacing due to profound bradycardia, though most were improved by reducing the dose or discontinuing anti-arrhythmic drugs or beta-blockers. Only 1 (2.0%) patient required an elective implantation of a cardiac resynchronization therapy defibrillation device for clinically relevant LV dyssynchrony and sick sinus syndrome on postoperative day 50. Three patients (6%) had minor transient ischaemic attack (n = 2) and stroke (n = 1) on postoperative days 11, 14 and 25, respectively; however, all recovered without subsequent complications. They unexceptionally received oral anticoagulation at the time of their events and the international normalized ratios (INRs) of prothrombin time were 1.4, 1.8 and 2.1, respectively.

Late outcomes

Seventeen (34%) patients required detailed examination by 24-h Holter monitoring, with a mean number of times of 1.8 (range, 1–4). The rates of freedom from AF at each follow-up examination were 80% (40 of 50), 79% (38 of 48), 70% (16 of 23) and 80% (4 of 5) at 1, 3, 5 and 7 years after surgery, respectively (Fig. 2). The mean value for the PR interval in patients presenting P waves was 158 ± 38 ms (range, 100–250 ms), suggesting no conduction delay in the majority of patients.

During the follow-up period, there were three late deaths; one from pneumonia, one from sudden death and one from unknown cause. Actuarial survival rates were 100 ± 0, 95 ± 4 and 91 ± 5% at 2, 4 and 6 years, respectively. No patients underwent redo mitral valve surgery, while 8 (16%) required permanent pacemaker implantation due to sick sinus syndrome (n = 4) or profound bradycardia with underlying AF (n = 4); for those, mitral valve repair was performed in 3, mitral valve repair plus aortic valve replacement in 3, and mitral valve replacement in 2. A large majority of those patients had documented evidence of sick sinus syndrome preoperatively.

Three patients (6%) (rheumatic 1, degenerative 2) required hospitalization for stroke at 39, 68 and 74 months after surgery; for those, mitral valve repair was performed in 1 and mitral valve replacement with a mechanical valve in 2, respectively. One patient had a history of cerebral infarction, and an LA thrombus was detected during operation. All received oral anticoagulation at the time of their events and the INRs of prothrombin time were 2.7, 1.8 and 2.7, respectively. Patients who did not restore LA mechanical contraction were more likely to experience postoperative thromboembolic events than those who restored it (23 vs 4%, P = 0.075).

Echocardiographic findings

Thirty-nine patients (78%) were free from AF at the latest follow-up examination (51 ± 20 months). Doppler echocardiography demonstrated evidence of LA mechanical contraction in 28 of these 39 patients; however, 11 showed no A wave on the transmitral filling tracing, despite successful AF ablation. As a consequence, only 56% (28 of 50) had echocardiographic evidence of LA mechanical contraction. Patients were divided into two subgroups and analysed according to the presence or absence of LA contraction (Table 2). At the baseline, the 28 patients who restored LA mechanical contraction (Group 1) had demonstrated a significantly smaller LA dimension and lower ejection fraction as compared with those who failed to restore LA contraction (Group 2).

From baseline to 1 month after surgery, LV dimension was decreased and left ventricular ejection fraction increased at 1 month after surgery, and these improvements were sustained for up to 2 years in both groups. LA dimension was decreased from the baseline to 1 month after surgery in both groups. That improvement was sustained for up to 2 years in Group 1, whereas Group 2 tended to show a gradual re-increase at 2-year follow-up examination.

Group 1 patients showed a decrease in systolic PA pressure to a greater degree at 1 month after surgery than Group 2 and that improvement was sustained for up to 2 years after surgery. In contrast, Group 2 showed a modest decrease in systolic PA pressure from baseline to 1 month after surgery, followed by a gradual re-increase and a high systolic PA pressure value at the 2-year follow-up examination.

The prevalence of significant tricuspid regurgitation (equal to or greater than mild grade) at 2 years after surgery was significantly lower in Group 1 than in Group 2 (7 vs 41%, P = 0.006), suggesting the importance of restoring LA mechanical contraction conferred by AF ablation. Notably, there was no difference in the frequency of concomitant tricuspid repair between the groups (95 vs 100%).

Predictor for lack of left atrial mechanical contraction

Univariate analysis identified advanced age, duration of AF >10 years, F-wave voltage <0.1 mV and LA dimension >60 mm as significant predictors of lack of LA mechanical contraction (Table 3). In addition, multivariate analysis identified LA dimension >60 mm as independently associated with a 4-fold increase in the risk of lack of LA mechanical contraction (adjusted hazards ratio 3.9, 95% CI 1.1–14, P = 0.035). Mechanical LA contraction was less-frequently present in patients with a preoperative LA dimension >60 mm when compared with those with an LA dimension ≤60 mm (19 vs 74%, P < 0.001) (Fig. 3).

Patients who did not recover from AF were more likely to have a large LA dimension at baseline when compared with those who recovered from AF with and without restored LA mechanical contraction (one-way ANOVA P = 0.09). Importantly, in the subgroup analysis among patients who recovered from AF, preoperative LA dimension was significantly larger in patients who did not restore LA mechanical contraction when compared with those who restored it (59 ± 6 vs 54 ± 6 mm, P = 0.037) (Fig. 4).

DISCUSSION

Major findings

This is the first report to describe the long-term results of a modified maze procedure via a septal-superior approach during mitral valve surgery. The major findings of this study were (i) a maze procedure via a septal-superior approach could be safely performed, with a recovery rate from AF of 78% at 59±17 months after surgery, (ii) among 39 patients free from AF at the latest examination, 72% (28 out of 39) had echocardiographic evidence of LA contraction, while the remaining 28% did not, and (iii) a preoperative LA dimension >60 mm was an independent risk factor for lack of LA mechanical contraction. These findings suggest that early surgery is warranted to restore sinus rhythm with LA mechanical contraction before severe LA dilatation occurs. The impact of LA contraction recovery conferred by AF ablation on postoperative haemodynamic improvements and thromboembolic events remains to be determined.

Postoperative atrial vulnerability after septal-superior approach

Postoperative sinus node dysfunction and increased atrial vulnerability (i.e. conduction disturbance and arrhythmias) remain important considerations with this approach, because the technique involves transecting the sinus node artery and anterior inter-nodal conduction pathway, as well as making incisions in both atria [8]. Berdajs et al. [9] provided a precise anatomical description of the origin and course of the sinus node artery in 50 human hearts obtained from cadavers without previous pathological alterations. They found that the sinus node artery originates from the right coronary artery in 66% of patients (62% from the proximal and 4% from the distal portion), and from the left coronary artery in the remaining 34%. On the basis of their findings, Kovács [9] described in the invited commentary that the sinus node artery would be transected in the large majority of hearts (96%) when a septal-superior approach was applied, meaning that damage to the sinus node blood supply becomes very probable during mitral valve surgery using this approach. The exact role of sinus node ischaemia and its impact on cardiac rhythm are not clearly understood, though once the sinus node artery is totally sectioned, a period of rhythm instability is generally observed during the subsequent 1–2 weeks.

On the other hand, Gaudino et al. showed, in a large prospective randomized investigation, that the use of the septal-superior approach is not associated with an increased incidence of clinical rhythm disturbances either in-hospital or during the mid-term follow-up period [10]. Even electrophysiological studies denied long-term adverse effects on sinus node function and conduction of septal-superior approach [11, 12]. Indeed, we found a relatively high incidence of atrial arrhythmia in the early postoperative period, while two-thirds of patients recovered from AF at discharge, which were almost consistent with those investigations. These findings might be caused by newly developed collateral blood flows to the sinus node or cardiac conduction systems, although further studies are required to better confirm these hypotheses.

Comparison with previous reports of maze procedure via trans-septal/septal-superior approachs or conventional left atriotomy

Some investigators have presented short-term (follow-up period 7–11 months) results of maze procedures performed via a trans-septal or septal-superior approach [13–15] (Table 4). Bunussi et al. [15] reported excellent short-term results, including a recovery rate of sinus rhythm of 80% at 11 ± 6 months after surgery. Despite the limited study period, they found that none of their patients required pacemaker implantation. In contrast, Ishikawa et al. [14] concluded that such exposure should not be considered as a first-line approach, because of the high rate of permanent pacemaker implantation. Aidietis et al. [13] analysed clinical data following a maze procedure in combination with mitral valve surgery via a trans-septal (n = 26) or septal-superior approach (n = 11), and found that the incidence of recurrence of AF or atrial flutter during the early postoperative period (2–15 days) was 42 and 45%, respectively. As compared with their study, a relatively high incidence (88%) of atrial arrhythmia during the early postoperative period (0–30 days) was observed in our study. The possible mechanism of this difference may be explained by the different observation time periods, definition of AF recurrence, or the more advanced age (65 ± 8 vs 55 ± 11 years) of our patients, which has been reported to be an important risk factor for early postoperative AF after open heart surgery [16]. Unfortunately, the lack of their data regarding the incidence of atrial arrhythmia during the acute stage (0–1 days) prevented us from performing a careful comparative investigation of this issue. Notably, our result was partly consistent with results previously reported by Ishii et al. [17], who found that the peak incidence of early postoperative atrial arrhythmia after a conventional maze procedure in 200 patients (Maze I = 30, II = 12, III = 158) occurred on postoperative day 8 and most of their patients (87%) with atrial arrhythmia had their episodes within the first 2 weeks after surgery. Together, these findings suggest that a meticulous assessment of cardiac rhythm is mandatory to avoid postoperative arrhythmia-related complications, particularly within 2 weeks after a trans-septal maze procedure.

Our result in regard to the rate of freedom from AF at the intermediate follow-up point was consistent with that noted in previous studies [13–15, 17–19]. The long-term follow-up period in the present study allowed us to confirm that recovery of cardiac rhythm observed at the intermittent follow-up point remained nearly stabilized during the 5 years after surgery. Application of this approach might also be encouraged by the fact that the incidence of permanent pacemaker implantation during hospitalization in our study was at least not inferior to that seen in patients who underwent a conventional maze procedure [17–19].

Post-maze left atrial mechanical contraction

Several previous studies have addressed the issue of post-maze LA mechanical function and filling capacities [4–6]. Feinberg et al. [4] found that 61% of patients who underwent a maze procedure had evidence of LA contraction after a mean follow-up period of 8 ± 7 months. Interestingly, Yuda et al. [5] investigated the success rate of maze procedures for restoring atrial contraction in patients with and without LA structural remodelling. They reported that effective LA mechanical contraction was identified in 66% of patients with less LA remodelling (mean LA dimension 55 ± 8 mm), whereas it was seen in only 21% of those with advanced LA remodelling (mean LA dimension 73 ± 14 mm), indicating a significant difference between the two groups. Our univariate analysis showed that an increasing LA dimension was associated with an increased risk of lack of LA mechanical contraction (hazard ratio 1.1 per 1-mm increase in LA dimension, 95% CI 1.0–1.1, P < 0.001), supporting their results. The proportion of patients with effective LA contraction seen in our study was also acceptable, given that the mean value of LA dimension at baseline was 59 ± 9 mm. Notably, we found a significant difference in LA dimension between patients with and without LA mechanical contraction in the subgroup analysis (Fig. 4), leading us to speculate that those with a large LA dimension at baseline are less likely to restore LA mechanical contraction, even when they recover from AF.

Impact of restored left atrial mechanical contraction on clinical results

Continued AF after mitral valve surgery is a well-known risk factor for progression of tricuspid regurgitation, while several previous studies have reported that such progression was prevented by a successful maze procedure [20–22]. Kim et al. found that the incidence of significant tricuspid regurgitation at the latest examination was significantly higher in patients with AF who did not undergo a maze procedure when compared with those who did as a concomitant procedure (39.7 vs 14.6%). Interestingly, in their subgroup analysis with patients with AF who had undergone a maze procedure, absence of right atrial mechanical activity was identified as an independent parameter for the progression of tricuspid regurgitation [20]. In the present study, the prevalence of significant tricuspid regurgitation was also significantly greater in patients without LA mechanical contraction than in those with it. Given that most patients received concomitant tricuspid annuloplasty, it is likely that a tricuspid annuloplasty procedure does not completely prevent postoperative progression of tricuspid regurgitation, particularly in patients who did not restore LA mechanical contraction after surgery. We speculate that recovery and maintenance of LA mechanical activity are of great value for preventing the progression of tricuspid regurgitation after mitral valve surgery, possibly yielding favourable haemodynamic function (improvement in systolic PA pressure) in the late follow-up period. In contrast, the lack of LA mechanical contraction might result in an abnormality of the post-capillary component (elevation in LA pressure), which plays, at least partly, a role in deteriorating systolic PA pressure, thereby progressing tricuspid regurgitation in the late follow-up.

The incidence of postoperative thromboembolic events was slightly high as compared with that in the previous reports, ranging 2.1–11% [23, 24]. The difference may be explained by different observation time periods (i.e. a longer follow-up in our study), definition of thromboembolic events, degree of LA remodelling (LA mechanical function) or patient demographics (i.e. slightly older in our study cohort) at baseline. We found a relatively high rate of postoperative thromboembolic events in patients without mechanical LA contraction as compared with those with it, which was consistent with a previous report by Buber et al. [6]. They investigated the association of the lack of LA mechanical contraction after a maze procedure with the increased risk of occurrence of thromboembolic stroke and found that such a lack was independently associated with a 5-fold increase in stroke risk. These results suggest that patients with a dilated LA (>60 mm) are less likely to benefit from a concomitant maze procedure, thus early surgical intervention is warranted to restore sinus rhythm with LA mechanical contraction before severe LA dilatation occurs. Importantly, the absence of LA mechanical contraction after a maze procedure should alert physicians to the possible risk of future thromboembolic events. Further studies are required to examine the need for chronic anticoagulation treatment for patients with LA mechanical standstill.

LIMITATIONS

The primary limitation of the present study is its retrospective nature, thus the data are susceptible to various possible sources of bias. Inclusion of patients with different etiologies for mitral valve disease and those who had undergone concomitant surgical intervention might have influenced the results. However, we did not observe significant effects of those factors on recovery from AF or restoration of LA mechanical contraction after surgery. Another limitation is that atrial contraction and its contribution to ventricular filling were studied using pulsed Doppler echocardiography, because the presence of A-wave velocity can provide only indirect assessment of LA mechanical contraction [25]. This approach can induce false results specifically in patients with prosthetic valves. The lack of data regarding right atrial mechanical contraction could not allow us to determine its impact on the physiology of the right-sided heart component. Finally, episodes of postoperative asymptomatic AF may have been missed, because electrocardiographic measurements and Holter recordings were routinely made during visits to the outpatient clinic or when symptoms arose.

CONCLUSION

Our trans-septal maze procedure may be an effective alternative surgical treatment for eliminating AF during mitral valve surgery. The present findings suggest that, in patients with valvular AF, early surgery is warranted to restore sinus rhythm with LA mechanical contraction, before severe LA dilatation occurs. The impact of LA contraction recovery conferred by AF ablation on postoperative haemodynamic improvements and thromboembolic events remains to be determined.

Conflict of interest: none declared.

ACKNOWLEDGEMENTS

The authors thank Kiyoshi Yoshida (Section of Clinical Engineering, Japan Labor Health and Welfare Organization, Osaka Rosai Hospital) for excellent management of the cardiopulmonary bypass procedures, and Yuji Masaki and Minako Furukawa (Section of Clinical Echocardiography, Japan Labor Health and Welfare Organization, Osaka Rosai Hospital) for their assistance with echocardiographic data collection. We also thank Mariko Yamashita for her great help of clinical data collection. This research was partially supported by research funds to promote the hospital function of the Japan Labor Health and Welfare Organization.

REFERENCES

APPENDIX. CONFERENCE DISCUSSION

Dr G. Bruinsma(Zwolle, Netherlands): In your paper you made reference to the work of Dr Stulak which was published in the Annals of Thoracic Surgery in 2007 and dealt with the issue of how we define success in the surgical treatment of atrial fibrillation. Success was defined as restoring sinus rhythm, and you managed to achieve sinus rhythm in 78% at the latest examination.

The second goal, restoring atrial contraction, shows, at least in your paper, that in 44% there was no evidence of LA contraction, and you observed 23% cerebral infarction, which appeared to be not significant compared to the 4% cerebral infarction rate in patients with LA contraction.

This brings me to the third goal of success, and it means is this Guiraudon approach the safest approach, as Dr Stulak reported only 1% stroke and my associate, Dr Sie, reported 2.1% stroke in their direct LA approach through the Waterston's groove.

Dr Kainuma: You made reference to the frequency of cerebral infarction; 12% of our patients in this study experienced cerebral infarction, which is relatively high as compared to previous reports. And of the six patients who experienced cerebral infarction, in three of them this might have been caused by blood coagulopathy because the cerebral infarction occurred in the early operative period, within the 30 days after surgery. All six patients with cerebral infarction were prescribed warfarin at the time of occurrence of the event.

Dr Bruinisma: Okay. My next question deals with the timing of the first echo. You report 96% freedom from atrial fibrillation immediately after the operation. There are no data on the echo findings with regard to the LA contraction.

Dr Kainuma: We analysed the presence of LA mechanical contraction, only at the latest examination and unfortunately do not have the data relating to the presence of LA mechanical contraction immediately after surgery in this series. So 96% of patients were free from AF, confirmed by electrocardiogram, but echocardiographic assessment of the presence of LA mechanical contraction was not performed in all patients immediately after surgery.

Dr Bruinsma: Okay.

Dr C. Muneretto(Brescia, Italy): As the previous paper, your manuscript addressed the issue about how we affect atrial function by means of the lesion set. In fact, theoretically we can cure any type of atrial fibrillation if we divide the left and the right atrium into small pieces, but the problem is that we will achieve sinus rhythm without any contractile function.

So because there is a lot of evidence that demonstrates that if you do not destroy the left atrial anatomy when you restore sinus rhythm, you can recover a nice function and contractility of the left atrium. Based on your results in which you compromise the atrial function in a significant percentage of patients, would you like to change your lesion set? This because the aim of the treatment of atrial fibrillation is not to achieve sinus rhythm per se but to restore contractility of the left atrium. Otherwise, if you should do the procedure only for rhythm control, you could ablate and pace. It would be much easier and uneventful. Therefore according to that, I mean restoring atrial contractility, are you planning to change your lesion set or not?

Dr Kainuma: We prefer the approach for mitral valve surgery and surgical ablation described here in the presentation. And as you commented, we should recognize that more effective surgical ablation restoring atrial contractility should be explored. But I am not sure at this time if there are more effective or safer methods of surgical ablation for patients with this condition. We think that right-sided surgical ablation will be required for preventing atrial flutter, and the left atrial surgical ablation, including box lesion set, would be necessary for improving the success rate of atrial fibrillation.

Author notes