The development of surgical techniques for the treatment of patients with cardiac arrhythmias was dependent on the availability of intraoperative electrophysiologic mapping systems that provided the ability to determine precisely where to place the lesions of conduction block in each patient. Intraoperative mapping was used to “guide” the surgical procedures for Wolff-Parkinson-White syndrome, elective His bundle ablation, automatic left atrial tachycardias, atrioventricular (AV) node reentry tachycardia, automatic right atrial tachycardias, ischemic ventricular tachcycardia, and nonischemic ventricular tachycardia. Because of the success of map-guided arrhythmia surgery for those arrhythmias, we naturally assumed that any successful surgical procedure for atrial fibrillation (AF) would also be map-guided. Our extensive experimental and clinical mapping studies documented that during AF, multiple macro-reentrant circuits were present in both atria. However, the maps also demonstrated that these macro-reentrant drivers of AF were fleeting in nature and frequently changed locations. It soon became clear that locating these ever-changing circuits by intraoperative mapping and then trying to ablate them with individual lesions (i.e., “chasing” them) was simply not feasible. We actually tried this approach experimentally, but as soon as one reentrant circuit was abolished surgically, another reentrant circuit would appear in a different location, and the AF persisted.
The Maze Procedure Concept
The inability to map-guide AF surgery suggested the need for a single surgical technique that was capable of ablating AF in all patients regardless of the specific electrophysiologic activation pattern present in each patient. Such a procedure would have to take advantage of the single common characteristic of all AF, the simultaneous presence of two or more large macro-reentrant circuits in the atria. We observed that macro-reentrant circuits (drivers) during AF frequently rotated around sites of normal conduction block in the atria (e.g., the orifices of various structures connecting to the atria such as the pulmonary veins, atrial appendages, vena cavae) and the mitral and tricuspid valve annuli.
Despite numerous earlier studies, AF at this time (c. 1986) was still widely considered to be a “chaotic arrhythmia” and as such, untreatable by surgical intervention. In early 1987, I had a conversation completely by chance with the recently retired chairman of the Department of Pure Mathematics and Mathematical Statistics of Cambridge University in the airport at Maastricht in the Netherlands. He asked if I did any research, and when I answered that I was working on a “chaotic” rhythm of the heart, he said, “Well, you know, my people in Cambridge are the world’s experts on the chaos theory.” This chance meeting resulted in my giving a lecture later to his mathematics department in Cambridge on what we knew about AF at the time. I also took multiple examples of our raw data (simple numbers representing the local activation times in the atria during AF) for his faculty to examine, and they confirmed that our data suggested specific patterns of activation that were indeed, not chaotic. This further strengthened our belief that AF was not a chaotic arrhythmia at all but rather one characterized by certain anatomic-electrophysiologic features that were common to all patients with AF. Furthermore, correlation of the normal anatomic structures with the observed interactions between local conduction velocity, local refractory periods, and the area of atrium necessary to sustain these reentrant circuits (i.e., their size) indicated that the number of potential locations in the atria that were capable of sustaining atrial macro-reentry was finite. This meant to us that it should be possible to treat atrial fibrillation surgically. The initial reasoning underlying the design of the Maze procedure that addressed these issues are discussed in other chapters and need not be repeated here. However, continuing experimental investigation and clinical experience further clarified why the Maze procedure was successful in ablating AF and confirmed that afterward, the atria could be activated by a sinus-generated impulse, thus restoring interatrial and AV synchrony.
Potential Left Atrial Drivers During Atrial Fibrillation
The local refractory periods were known to be shorter in the left atrium (LA) than in the right atrium (RA); therefore, the macro-reentrant circuits were potentially smaller in the LA (see Chapter 4 , Figs. 4.14 and 4.15 ). Thus, atrial anatomy limited the potential sites where macro-reentry could develop to drive AF ( Fig. 12.1A – C ). We later learned that after pulmonary vein isolation, a new iatrogenic type of long cycle-length atrial flutter often developed in the left atrium that uses the atrial myocardium and the coronary sinus in the left atrial isthmus between the inferior pulmonary vein orifices and the posterior mitral valve annulus ( Fig. 12.1D ). We called this iatrogenic arrhythmia “atypical left atrial flutter,” but its name was later changed to “peri-mitral flutter,” a troublesome arrhythmia that is addressed in greater detail in Chapter 48 . Thus even though all of these left atrial macro-reentrant drivers of AF had not been specifically documented before our first attempt at the surgical ablation, it was clear where these left atrial drivers could potentially be located ( Fig. 12.2 ).
The locations of left atrial macro-reentrant drivers fall into four general categories, which are as follows. (A) Those rotating around the two pairs of pulmonary veins that usually incorporate very little of the posterior left atrial wall as a part of their circuits. (B) Those rotating around the one pair of pulmonary veins that use either the atrial septum or the lateral left atrial free well and incorporate the posterior left atrial wall as a part of their circuits. In addition, one smaller circuit is shown in the left atrial isthmus area because it was recorded in the first patient ever to be mapped with a multipoint computerized mapping system during atrial fibrillation (see Chapter 5, Fig. 5.5). (C) Those within the body of the left atrial appendage (rare) or just below the orifice of the appendage using the so-called “coumadin ridge” (more common). (D) The iatrogenic arrhythmia peri-mitral flutter, which uses the left atrial isthmus, including either the atrial myocardium or the coronary sinus as a part of its circuit (see Chapter 48 ).
Left atrial macro-reentrant drivers of atrial fibrillation (AF). This figure comprising all of the known left atrial macro-reentrant drivers of AF shows that some of them use the posterior wall of the left atrium (LA) as a part of their circuit and some do not. In long-standing persistent AF, these left atrial macro-reentrant circuits represent approximately 70% of the drivers of AF.
Potential Right Atrial Drivers During Atrial Flutter and Atrial Fibrillation
Atrial Flutter
Because the macro-reentrant circuits in the RA require a larger area to develop and sustain themselves than macro-reentrant circuits in the LA. There are only a few RA sites where they can be located. Santucci et al. published an excellent article in 2009 documenting that classic atrial flutter can be caused by one of three potential macro-reentrant circuits in the RA, the so-called atrial flutter waves ( Fig. 12.3 ). All three of these potential RA drivers of atrial flutter use the cavo-tricuspid isthmus (CTI) between the inferior vena cava (IVC) orifice and the tricuspid valve annulus as a part of their macro-reentrant circuit. This explains why the isolated CTI lesion created by catheter ablation usually cures classic atrial flutter ( Fig. 12.4 ).
Right atrial macro-reentrant drivers of atrial flutter. Left panel, Key figure from Santucci and coworkers’ 2009 article showing the locations of each of the three macro-reentrant circuits that can cause classic right atrial flutter. Right panel, Three-dimensional representation of the location of the three macro-reentrant circuits in the left panel that can drive classic right atrial flutter. Only one of these reentrant circuits is present at a time. They are (1) a single large circuit in the right atrium (RA) above the annulus of the tricuspid valve (top panels) that uses the cavotricuspid isthmus (CTI) inferiorly; (2) a single large circuit that uses the right atrial free wall and atrial septum, passing anterior to the superior vena cava (SVC) orifice superiorly and the CTI inferiorly (middle panels) ; or (3) a single large circuit that uses the right atrial free wall and atrial septum, passing posterior to the SVC orifice superiorly and the CTI inferiorly (lower panels). All three of these “atrial flutter waves” use the CTI. Therefore, a single lesion placed across the CTI with either a catheter or surgically is highly effective in the treatment of patients with atrial flutter.
(Reproduced from Santucci PA, Varma N, Cytron J, et al. Electroanatomic mapping of postpacing intervals clarifies the complete active circuit and variants in atrial flutter. Heart Rhythm. 2009;6(11):1586–1595.)
Cavotricuspid isthmus (CTI) lesion for atrial flutter. An endocardial lesion placed across the CTI (dashed blue line ) is extremely effective for the treatment of patients with atrial flutter because all three of the “flutter waves” that can potentially cause atrial flutter use the CTI as a part of their macro-reentrant circuit. (A) CTI lesion across the three “flutter waves” that can drive atrial flutter. (B) No remaining “flutter waves” after CTI lesion for atrial flutter.
Atrial Fibrillation
The contribution of the RA to the maintenance of AF is finally beginning to get its due as a result of multiple mapping studies showing that up to 50% of patients with persistent and long-standing persistent AF have drivers in the RA. In addition to the three “flutter waves,” three additional sites of potential macro-reentrant drivers in the RA are (1) around the superior vena cava (SVC) orifice, (2) around the IVC orifice, and (3) around the base of the RA appendage (RAA; Fig. 12.5 ). This explains why, unlike atrial flutter, an isolated CTI lesion is inadequate to ablate all of the RA drivers of AF because it leaves the possibility of a macro-reentrant driver around the orifice of the SVC and around the base of the RAA ( Fig. 12.6 ). The potential macro-reentrant drivers of AF in the RA are combined with the left atrial drivers to document the basis for the development of a biatrial surgical procedure for AF ( Fig. 12.7 ).
Right atrial drivers of atrial fibrillation (AF). The locations of the three potential atrial flutter waves plus the other sites in the right atrium shown to be capable of sustaining macro-reentry and therefore of driving AF as well. The additional sites include reentrant circuits around the base of the right atrial appendage and around the orifices of the superior vena cava and inferior vena cava (IVC). In addition to the three atrial flutter waves that use the cavotricuspid isthmus (CTI) in their circuits, the reentrant circuit around the IVC also uses the CTI as a part of its circuit.
Cavotricuspid isthmus (CTI) lesion for atrial fibrillation (AF). An endocardial lesion placed across the CTI (dashed blue line ) is not as effective for the treatment of patients with AF as it is for those with atrial flutter because potential reentrant sites remain around the right atrial appendage and the superior vena cava (SVC), either of which can still participate (with left atrial reentrant circuits) in driving AF. However, combining a right atrial CTI lesion with the left atrial maze lesions in patients with AF significantly enhances the success rate in comparison with placing no lesions at all in the right atrium (RA). This is because macro-reentry around the right atrial appendage and SVC is quite rare, so most of the right atrial drivers that participate in AF are one of the three atrial flutter waves or the inferior vena cava reentrant circuit. (A) CTI lesion for atrial fibrillation. (B) Remaining potential RA macro-reentrant circuits after CTI lesion for atrial fibrillation.
Two-dimensional representation of the known macro-reentrant drivers of atrial fibrillation (AF) that were originally only theoretical but have since been documented by electrophysiological mapping. This method of illustrating the location of the drivers provides a basis for understanding the results of surgical and catheter ablation for persistent and long-standing persistent AF.
Interruption of the Right and Left Atrial Drivers of Atrial Fibrillation
We knew that one way to ablate AF would be to “breadloaf” the atria into multiple separate compartments, as had been done experimentally some 25 years earlier (see Chapter 5 , Fig. 5.11 ). Although breadloafing the atria would certainly ablate AF ( Fig. 12.8 , left panel ), it was not feasible clinically. The two sequelae of atrial “breadloafing” that negated its clinical feasibility were (1) the total isolation (compartmentalization) of individual segments of the atrium from one another, which precluded any subsequent synchronized activation of the atrial myocardium, and (2) the only part of the atrium that could resume a sinus rhythm postoperatively was the isolated compartment that contained the sinoatrial (SA) node (see Fig. 12.8 ).
“Breadloafing” the atria. (A) The creation of multiple parallel lines of conduction block that separate the atria into multiple electrically isolated compartments ablates atrial fibrillation if the compartments are too small to sustain macro-reentry. (B) However, compartmentalization of the atria means that there is only one compartment that can be activated by the sinus node postoperatively (red shaded compartment). The rest of the atrium cannot be activated by a sinus-generated impulse.
Despite not being clinically applicable, the breadloafing experiments showed that the critical factor that determines the ability to develop atrial macro-reentry is the area of contiguous atrium available to participate in the macro-reentrant circuit. These studies also showed rather inadvertently that the critical area can be limited by placing atrial lesions close enough together to prevent the development of large macro-reentrant circuits. This suggested the possibility of ablating atrial macro-reentry (fibrillation) by avoiding the complete isolation (compartmentalization) of the individual segments and leaving them connected so that they remained a single unit of myocardium from an electrical standpoint. The obvious reasoning was that if the large macro-reentrant drivers could not form, then the atria could not fibrillate ( Fig. 12.9 ). It then occurred to me that if instead of placing parallel lesions as is done with the breadloafing approach, we placed the lesions in a “maze” pattern, atrial macro-reentry (fibrillation) could still be surgically ablated, but in this case, the entire atrial myocardium could be activated by the SA node afterward. Thus the objective of such a “maze procedure” would be twofold: (1) to preclude the development of atrial macro-reentry (fibrillation) and (2) to leave the atria capable of responding to a sinus-generated impulse that could activate both atria.
(A) The dotted circle represents a potential macro-reentrant circuit. The red arrow within that dotted circle indicates that the potential circuit has been stimulated and is now active. (B) As mentioned in the text, these macro-reentrant drivers are relatively large and are therefore vulnerable to being ablated by placing lesions (lines of conduction block) close enough together in the atria to preclude their ability to form. The areas between these lesions do not have to be isolated (as with “breadloafing”) and therefore can remain in continuity with the remainder of the atrium. If such macro-reentrant circuits cannot occur in the atrium, the atria cannot fibrillate.
The next step was to figure out how to create an electrical “maze” in the atria. It is critical to recognize the fact that the myocardium of the RA, atrial septum, and LA are all one piece of atrial muscle. We commonly think of the RA and LA as separate structures because they function on different sides of the cardiovascular system. However, from an electrophysiologic standpoint, they are a single unit of myocardium that is activated in a sequential manner irrespective of their individual hemodynamic functions. And exactly what is an “electrical maze”? It is a pattern of atrial lesions that has one entry site (the SA node), one exit site (the AV node), one route of conduction between the entrance and exit sites, and multiple “blind alleys” along that route. The “blind alleys” provide the means of activating the entire atrial muscle by the SA node. The story of how the Maze procedure came about is described in more detail in Chapter 1 .
The Original Maze-I Procedure
Surgical Technique
The lesions of the original Maze procedure satisfied both goals of a surgical procedure for AF: (1) ablation of the arrhythmia ( Fig. 12.10 , left panel ), and (2) leaving the atria capable of responding to a sinus-generated impulse postoperatively ( Fig. 12.10 , right panel ) (Video 12.1). The lesions were placed close enough together to preclude the development of macro-reentry anywhere in either atrium, thereby ablating AF. The so-called “box lesion” around the four pulmonary vein orifices and the intervening posterior left atrial wall isolated the encompassed portion of the LA from the rest of the atrial myocardium. Otherwise, the entire myocardium in both atria could be activated by a sinus-generated impulse. After the Maze-I procedure, the two atria were completely activated in approximately 140 ms compared with activation of normal atria within approximately 110 ms ( Fig. 12.11 ). This 30 ms difference in the total time required to activate the atria had no effect on the hemodynamics of the heart or on mechanical AV synchrony because electrical activation is well over before mechanical function of the atria is initiated ( Fig. 12.12 ). For example, at a heart rate of 60 beats/min, the heart beats every 1 second or every 1000 ms. This means that in the normal heart with a rate of 60 beats/min, the atria are electrically silent for 890 ms (89%) of each heartbeat. After the Maze procedure, the atria are electrically silent for 860 ms (86%) of each heartbeat at a rate of 60 beats/min. In both normal and post-Maze hearts, the mechanical function of the atria occurs within this electrically silent portion of the heartbeat.
Maze-I procedure. (A) The lesions of the Maze-I procedure precluded the ability of large macro-reentrant circuits to form in the atria and therefore rendered the atria incapable of fibrillating. (B) Placing the lesions in the pattern of a maze provided the ability of a sinus-generated impulse to activate all of the myocardium of both atria except for the atrial appendages, which were removed, and the isolated pulmonary vein cuff that included the pulmonary vein orifices and the posterior left atrial wall between them.
Atrial activation time. (A) The entire myocardium of both atria is activated within approximately 110 ms in normal hearts. (B) After the Maze procedure, the entire myocardium of both atria (except the isolated pulmonary vein cuff) is activated in approximately 140 ms. In both cases, atrial activation is completed long before the atrial muscle contracts in response to the activation.
Atrial electromechanical activity during normal sinus rhythm (NSR) (A) Normal heart. The temporal relationship between atrial electrical activation and atrial mechanical function is illustrated in a patient with a heart rate of 60 beats/min. At this heart rate, the heart beats every 1 second or every 1000 ms. The extremely rapid atrial electrical activation in a normal heart (110 ms, outlined in red ) is over well before atrial mechanical function begins. This results in the atria being electrically silent for 89% of each cardiac cycle in a normal heart at a rate of 60 beats/min (890 ms, outlined in blue ). (B) Post–Maze procedure heart. Atrial electrical activation in a patient with a heart rate of 60 beats/min after the Maze procedure (140 ms, outlined in red ) is also over well before atrial mechanical contraction begins. In these post–Maze procedure patients, the atria are electrically silent for 86% of each cardiac cycle compared with 89% of the cardiac cycle in normal patients. Thus, the Maze procedure does not significantly alter the temporal relationships of the electro-mechanical activity of the atria. E, Electrical activity in the atria; M, mechanical activity in the atria.
The same electro-mechanical relationship is present even when the heart is beating twice as fast in sinus rhythm (i.e., at 120 beats/min). In this case, the atria are electrically silent for 78% of each cardiac cycle in the normal heart and for 72% of each cardiac cycle after the Maze procedure ( Fig. 12.13 ). Again, the atrial mechanical function occurs well after electrical activation has been completed in both normal hearts and in hearts after the Maze procedure. As discussed in previous chapters, great effort was expended in documenting that the atria continued to function after the Maze procedure, and this is largely because the electromechanical characteristics of the atria are not altered significantly by the Maze procedure.
