A simple surgical technique to prevent atrial reentrant tachycardia in surgery for congenital heart disease†.

OBJECTIVES
To present and test a simple surgical technique that may prevent atrial reentrant tachycardia following surgery for congenital heart disease. This arrhythmia is one of the commonest long-term complications of such a surgery. It may occur many years (even decades) after the operation. It is usually explained as a late consequence of right atriotomy, which is an inherent component of many operations for congenital heart disease. Right atriotomy results in a long scar on the right atrial myocardium. This scar, as any scar, is a barrier to electrical conduction, and macro-reentrant circuits may form around it, causing reentrant tachycardia. However, this mechanism may be counterchecked and neutralized by our proposed method, which prevents reentrant circuits around right atriotomy scars.


METHODS
The proposed method is implemented after termination of cardiopulmonary bypass and tying the venous purse-strings. It consists of constructing a full-thickness suture line on the intact right atrial wall from the inferior vena cava (IVC) (a natural conduction barrier) to the atriotomy incision. This suture line is made to cross the venous cannulation sites if these are on the atrial myocardium (rather than being directly on the venae cavae). Thus, the IVC, atriotomy and cannulation sites are connected to each other in series by a full-thickness suture line on the atrial wall. If this suture line becomes a conduction barrier, it would prevent reentrant circuits around right atrial scars. This was tested in 13 adults by electroanatomical mapping. All 13 patients had previously undergone right atriotomy for atrial septal defect closure: 8 of them with the addition of the proposed preventive suture line (treatment group) and 5 without (control group).


RESULTS
In all 13 cases, the atriotomy scar was identified as a barrier to electrical conduction with electrophysiological evidence of fibrosis (scarring). In the 8 patients with the proposed suture line, this had also become a scar and a complete conduction barrier. In the 5 patients without this suture line, there was free electrical conduction between the IVC and atriotomy scar.


CONCLUSIONS
The proposed suture line becomes a scar and conduction barrier. Therefore, it would prevent reentrant circuits around atrial scars and their consequent arrhythmias.

This complication may, on occasion, be explained by various different pathophysiological mechanisms, such as possibly persisting residual lesions with pressure or volume overload. However, it is most commonly explained as a late consequence of right atriotomy, which is an inherent component of many operations for congenital heart disease. Right atriotomy results in a long scar on the right atrial myocardium. This scar, as any scar, is a barrier to electrical conduction, and macro-reentrant circuits may form around it ( Fig. 1) causing reentrant tachycardia [13][14][15].
This problem prompted us to devise a simple surgical technique to prevent the formation of reentrant circuits around right atriotomy scars. This technique may be implemented, as an adjunct, in any operation that involves a right atriotomy. With the approval of our ethics committee, this technique is presented and tested in this study.

The proposed surgical technique
The proposed surgical technique is implemented after the termination of cardiopulmonary bypass and removal of the venous cannulae; after the venous purse-strings have been tied. It consists in suturing the intact (devoid of any incision) right atrial free wall in a manner to construct a suture line from the inferior vena cava (IVC) to the right atrial incision (Fig. 2). This suture line is transmural (full-thickness). In addition, it is a double-suture line, i.e. suture line constructed with both ends of the same suture, which are then tied to each other. The suture material is polypropylene-6/0 in neonates, infants and small children, 5/0 in bigger children and 4/0 in adults.
Both the right atriotomy incision and the IVC are barriers to electrical conduction (the IVC being a natural barrier). The proposed added suture line was devised with the idea that it will also become a conduction barrier since it compresses the atrial myocardial tissue within it thus interfering with its function and possibly even causing fibrosis (scarring). Thus, this suture line would prevent the formation of reentrant circuits around right atrial incisions-an effective prophylactic measure against atrial reentrant tachycardia.
This technique is adjusted slightly, in one or two ways, if venous cannulation is implemented via the right atrial wall (which is often the case) rather than by direct cannulation of the venae cavae. These adjustments represent a precaution so that no reentrant circuit may form around the scars of venous cannulation either, even though these scars are probably too small to sustain such circuits [16]. If the IVC cannulation site is on the atrial wall, the direction of the suture line is simply adjusted such that it crosses this site (Figs 3 and 4). If the cannulation site for the superior vena cava (SVC) is also on the atrial wall, a separate suture line is constructed between this site and the atriotomy incision (Fig. 5). These adjustments result in a composite suture line connecting all right atrial incisions (atriotomy and cannulation sites) and the IVC to each other in series such that no reentrant circuit may form around any of them.     This technique is tested in a group of patients that volunteered to participate in this study, to see whether or not the proposed suture line does indeed become a barrier to electrical conduction.

Patient population
Since January 2013 (when this technique was devised), all patients undergoing surgery involving a right atriotomy in our centre have received the proposed suture line, independently of both age and underlying pathology. However, only adults that had undergone surgical closure of ostium secundum atrial septal defects (ASDs) were included in this study in order to keep the patient groups homogeneous. These are patients that had been deemed unsuitable for percutaneous device closure on the grounds of large size and/or inadequacy of edges of the defect. They were recruited, on voluntary basis, by interview giving them detailed information about the rationale and objectives of this study. Patients that had undergone ASD closure prior to 2013 (i.e. without the addition of the proposed suture line) were also recruited in order to form a control group. Patients operated before September 2006 were excluded from this study since they had been operated on by a different surgical team. Thus, this work is a non-randomized retrospective study with historical control.
A total of 33 adults (age 16 years or more) have undergone surgical closure of ASD in our centre since September 2006. All 33 patients were interviewed, of whom 13 agreed to participate in this study and signed a consent form. These consist of 8 patients (5 women, 3 men, age 19-57 years) operated after January 2013 in whom the proposed prophylactic suture line had been implemented, and 5 patients (all women, age 16-53 years) operated before that date in whom this had not been done. All the 13 operations had been carried out, or directly supervised, by the lead surgeon (ARH). All procedures had been carried out uneventfully via midline sternotomy, using standard bicaval normothermic cardiopulmonary bypass and antegrade cold blood cardioplegia. Bypass and cross-clamp times had been less than 30 and 10 min respectively in all cases. The atriotomy incision did not cross the crista terminalis in any patient. All patients had been in sinus rhythm prior to surgery, and their postoperative course had been free of any complications including arrhythmias. At the time of the study, all subjects were asymptomatic and in sinus rhythm; none of them were on any antiarrhythmic medication.

Assessment of the proposed technique
Assessment was done by high-density electroanatomical mapping studies. This was approved by our ethics committee, but with strict conditions: only adults could be studied, and the only aspect that we were allowed to test was whether or not the proposed suture line had become a scar and a barrier to electrical conduction. No permission was given to use isoproterenol or carry out any test of induction of atrial tachycardia, or any ablation therapy, given that these patients were all healthy and asymptomatic volunteers.

High-density electroanatomical mapping studies
These studies were carried out by our electrophysiologists and co-authors (AP, EAR, LGR, ÁAL, AJV) who had no knowledge of which and how many of the recruited patients had received the proposed suture line.
These studies consist in obtaining high-density electrical activation maps of the right atrial wall, using the CARTO 3 system (Biosense Webster, Diamond Bar, CA, USA). This requires two catheters, both of which are introduced via the right femoral vein and advanced to the right atrium. One of the two catheters is used as a stimulation ( pacing) electrode (NaviStar, Biosense Webster). This is positioned onto the right atrial endocardium on the anterior atrial wall very close to the opening of the IVC (very close and anterior to where the proposed suture line would be). The other catheter is for signal detection-the mapping catheter (PentaRay NAV eco, Biosense Webster). Signals are detected on a large number of sites on the atrial wall ( 2000 sites); hence, the term 'high-density' mapping, which is another way of saying 'highresolution'. Mapping was performed under continuous pacing with a 600-ms cycle length.
This system was employed to assess the proposed suture line in two ways: (i) to see whether or not fibrosis (scarring) had occurred along its line; (ii) to see whether or not electrical wave fronts can travel across it, i.e. whether or not it is a barrier to electrical conduction. One may argue that if the first point is fulfilled, the second one is unnecessary, since it is well established that scars are barriers to electrical conduction [17,18]. Nevertheless, both criteria were taken into account in every case.
These tests were carried out according to existing and wellestablished criteria for the detection of scars and conduction barriers [17,19]: Firstly, the system is used to measure voltages at all the reference points. These voltages are colour-coded, thus generating a visual representation of signals and scars-a voltage map (Fig. 6). Contiguous areas with absent electrograms, or electrograms with very low voltages (<0.5 mV), are suggestive of scars.
Then the system is switched to a mode that constructs local activation maps. This mode measures the time taken for each of the 2000 references points on the atrial wall to be activated by the stimulation electrode. The farther a reference point is from the position of the stimulation electrode, the longer it takes to be activated. These activation times are also colour-coded, thus making the wave front, and its velocity, visible (an activation map). This enables us to easily see the trajectory of the wave front, and any barrier to it, by simply following the assigned colour code on the atrial wall (Fig. 7).
Finally, the activation mode is used to measure the activation times between signals detected at adjacent sites on either side of the proposed suture line (or equivalent sites in those that had not received this suture line). These sites are very close to each other in space (separated by only the width of the suture line) such that the activation times between them should not be long, unless the suture line is an electrical barrier (i.e. wave fronts cannot cross it). If the suture line is indeed an electrical barrier, the site immediately on the other side of this is activated later than in those that did not receive this suture line. This is because the wave front, having activated the first site, needs to take an alternative path in order to reach the second site.
This entire process of signal detection is then repeated, this time with the stimulation electrode positioned on the posterior atrial wall close to the IVC opening (very close and posterior to where the proposed suture line would be). Thus, voltage and activation maps are obtained with signals generated from both sides of the proposed suture line (if this is present). The suture line is considered to be a conduction barrier if it blocks conduction in both directions.  : Colour-coded activation maps of the right atrial wall, with the stimulation electrode positioned anteriorly. These were arbitrarily colour-coded in the following sequence: red-orange-yellow-green-blue-purple. Thus, the area that is closest to the stimulation electrode (and therefore activated first) appears red; the farthest area (and therefore activated last) appears purple. By following this colour sequence, we can see the exact trajectory of the wave front. Areas of no activation (scars) were coded dark grey. (A) Atriotomy without the addition of the proposed suture line. The atriotomy scar is easily seen (dark grey). The colour code sequence shows that there is clearly unhindered propagation of the wave front around both ends of the atriotomy scar. (B) Atriotomy with the addition of the proposed suture line. There is a continuous scar (dark grey) from and including the atriotomy down to the IVC. The colour sequence shows that the wave front cannot cross this; it propagates around the cephalad end of the atriotomy scar and reaches the other side of the added suture line last (small purple area). LAT: latent period between signal generation by the stimulation electrode, and its detection by the mapping electrode, measured in milliseconds.

RESULTS
The electroanatomical mapping studies were done 2-13 months (median: 5 months) after surgery for those with the proposed suture line. In those without this suture line, being historical controls, the time intervals between surgery and mapping studies were longer: 26-89 months (median: 37 months).
In all cases, 600-ms cycle signals were generated by the stimulating electrode, and were detected at all reference points on the atrial wall in order to construct high-density (high-resolution) electroanatomical maps. The median number of reference points was 2197 (range: 1947-2448).

Voltage maps
The voltage maps identified the atriotomy scar in all 13 patients (zones of non-detectable signals, or of voltages <0.5 mV). In all 5 patients that had not received the proposed suture line (control patients), there were normal voltages in the zone between the atriotomy scar and the IVC, indicating normal atrial tissue in that zone (Fig. 6A). Conversely, voltage maps in the 8 patients that had received the proposed suture line revealed that this had also fibrosed, i.e. there was a continuous scar line from and including the atriotomy site down to the IVC (Fig. 6B).

Activation maps
In the 5 control patients, activation maps revealed that wave fronts propagated simultaneously around both the superior and inferior extremes of the atriotomy scar (in particular, in the zone between this scar and the IVC), having previously identified and marked the atriotomy scar using voltage maps (Fig. 7A). Conversely, in all 8 patients that had received the suture line, this had also become a bidirectional barrier to propagation of electrical wave fronts. In these 8 cases, wave front propagation was exclusively around the superior extreme of the atriotomy scar. Consequently, the areas of the atrial myocardium on the other side of suture line were activated last (Fig. 7B).
The activation times between adjacent points on either side of the suture line, or equivalent sites in those that had not received this suture line, were also measured in all 13 patients. In the 8 patients that had received the suture line, these times were relatively long since propagation of the wave front from one reference point to the other could not be direct but deviated around the atriotomy scar. The activation times were 88-130 (median 104) ms when stimulated anteriorly (counterclockwise propagation on the map) and 90-110 (median 90) ms when stimulated posteriorly (clockwise propagation on the map). Activation times between signals detected at similar sites in the 5 patients that had not received the suture line (control patients) were much shorter: 29-51 (median: 29) ms when stimulated anteriorly, and 10-51 (median: 31) ms when stimulated posteriorly. These activation times are illustrated in Fig. 8.

DISCUSSION
The proposed suture line on intact atrial myocardium becomes a scar and a barrier to electrical conduction. This prevents the formation of reentrant circuits around right atriotomy scars, thus preventing reentrant tachycardia caused by such circuits which is the commonest long-term complication following surgery for congenital heart disease [2,13,14].
This arrhythmia may be prevented with the use of at least two other surgical methods which deserve a brief mention. One alternative method is to extend the right atriotomy incision onto the IVC. Such an extended incision would naturally prevent the formation of reentrant circuits. However, this seemingly simple method is often difficult to achieve without first removing the IVC cannula, especially in small children. This would complicate the operation significantly as it requires circulatory arrest. In addition, this method may be further complicated by the fact that the IVC is occasionally quite thin-walled such that the suture line may tear. Such a tear would require further wider suturing which may induce stenosis. The other alternative is radiofrequency ablation, or cryo-ablation, of the same area [20]. This is a simple technique. However, it is very expensive, especially for a preventive measure against something that may or may not occur decades later. Conversely, the technique proposed in this study is always feasible, quick and easy, does not remotely complicate the operation and costs practically nothing. Therefore, this is our preferred method since it is supported by our findings, despite the limitations of this study.
The limitations of this study must be recognized: (i) It is a small study; only a total of 13 patients agreed to participate. (ii) It is nonrandomized and retrospective. (iii) The control group is historical. (iv) It is a short-term study; it tells us nothing about what may happen to these patients in the long run. Although the proposed suture line may act as an effective measure against atrial reentrant tachycardia by preventing the formation of reentrant circuits around atriotomy scars, it may cause other problems not yet come to light, including possibly other types of arrhythmias. These may only be revealed in long-term studies. In fact, it may be argued that the proposed technique should not be viewed as truly prophylactic until such long-term data are available. (v) Electrophysiological assessment did not include any test of induction of atrial reentrant tachycardia, since such tests had not been approved by our ethics committee. Our ethics committee had confined our study to gathering only one piece of information:

ORIGINAL ARTICLE
A.-R. Hosseinpour et al. / Interactive CardioVascular and Thoracic Surgery whether or not the proposed suture line becomes a scar and a barrier to electrical conduction. (vi) It does not include any histological examination of the suture line to confirm directly that this is indeed fibrosed (converted into scar tissue). Assessment of fibrosis was electrophysiological. Nevertheless, this is still considered as quite strong and reliable evidence of scarring [17][18][19]. (vii) Only patients that had undergone ASD closure were studied despite the fact that we implement the proposed suture line in all cases involving a right atriotomy. This arbitrary choice was an attempt to make our patient population as homogeneous as possible while still having a reasonable number of patients to study. This precautionary measure may well have been unnecessary since the objective of this study was simply to assess whether or not suturing the intact atrial myocardium results in an electrical barrier-a pathophysiological process likely to be independent of intracardiac morphology.
Despite the above-mentioned limitations of this study, the results are very encouraging and in favour of this technique. We do this routinely, on adults and children alike, whenever we perform a right atriotomy.