Making Sense of the Maze: Title and subTitle BreakWhich Patients With Atrial Fibrillation Will Benefit?

Atrial fibrillation imposes heavy burdens on both patients and clinical practice. Approximately 2.3 million individuals in the United States have atrial fibrillation and the advancing age of the population,^{1 - 2} along with the increasing prevalence of obesity,³ is likely to further expand this number. Recent clinical trials have provided valuable information regarding management strategies of rate control with anticoagulation as opposed to the maintenance of sinus rhythm, primarily by the use of antiarrhythmic drugs. In general, not only has maintenance of sinus rhythm proven no better than rate control,⁴ but the AFFIRM trial actually demonstrated a benefit of rate control over rhythm control in terms of overall mortality among patients with coronary artery disease and hypertension.⁵ Nonetheless certain patients most likely to benefit from coordinated atrioventricular contraction were underrepresented in these trials, including patients with severe symptoms, hypertrophic cardiomyopathy, severe heart failure, valvular heart disease, and those undergoing cardiac surgical procedures. Some evidence suggests that the impact of one treatment strategy over another may vary among patient categories.⁶ In these patient subsets, there is considerable and growing interest in restoration and maintenance of sinus rhythm using pharmacologic therapies and either surgical or catheter-based nonpharmacologic approaches.

The randomized trial by Doukas and colleagues⁷ reported in this issue of JAMA examines a surgical approach for conversion from atrial fibrillation for the subgroup of patients with severe valvular heart disease already committed to surgical intervention. The so-called “maze procedure” creates an electrical maze in the atrium blocking reentrant circuits responsible for maintaining atrial fibrillation and provides a single electrical path from sino-atrial node to atrioventricular node. The procedure was developed in the laboratory by Cox and Boineau and first applied clinically by Cox in 1987.⁸ The effectiveness of the maze procedure as described by Cox and Boineau is well established. In its developer’s hands, the procedure is reported to restore sinus rhythm in more than 95% of cases.⁹ Cox has modified the technique twice, aiming to improve coordinated contraction of the left atrium resulting in effective transport of blood from atrium to ventricle (ie, improved left atrial transport) and to simplify its performance from a technical standpoint, culminating in the “maze III” procedure.¹⁰ Subsequently, a succession of surgeons have found adding their own modifications to the procedure hard to resist. Accordingly, despite the widespread and imprecise use of the term “maze procedure,” few studies in the literature actually address the efficacy of the true maze procedure as originally developed on the basis of laboratory studies.

The incremental value of the maze procedure, as outlined by its developers, ostensibly lies in its ability to address all 3 adverse consequences of atrial fibrillation: an irregular rhythm, risk of thromboembolism, and loss of atrioventricular synchronization. Restoration of regular rhythm is the primary objective from a symptomatic standpoint as well as from a pathophysiologic perspective in patients for whom atrial fibrillation causes left ventricular dysfunction via tachycardia-induced cardiomyopathy.¹¹ In addition, Cox and Boineau sought to reduce the risk of thromboembolism and to restore the hemodynamic benefits of atrioventricular synchrony. Neither of the latter 2 issues is addressed by atrioventricular nodal ablation and pacemaker insertion.

The surgical maze procedure reliably results in a regular rhythm, which is usually sinus, albeit occasionally at the cost of pacemaker insertion. Thromboembolism risk is reduced in theory via both amputation of the left atrial appendage and reduction of stasis in the atrium through improved atrial contraction.^{9 ,12} There remains some room for debate, however, because thromboembolism risk is likely a function of both the efficacy of atrial contraction and the lack of recurrences of atrial fibrillation whether symptomatic or asymptomatic. Larger clinical trials are necessary to prove this benefit definitively because the stakes are high; for example, stroke in patients with atrial fibrillation is frequently a devastating event.¹³ The effectiveness of the maze procedure in restoring ejection of blood from the atrium into the ventricle during pre-systole, the third of the identified abnormalities, has been equally controversial, with some studies demonstrating effective left atrial transport while in others atrial transport function appears to be impaired.^{14 - 17}

Regardless of the perceived or real benefits, application of the maze procedure has not been widely accepted, particularly for patients with “lone” atrial fibrillation or those in whom surgery would otherwise not be considered. This is due in part to the reluctance to perform open-heart surgery for what is usually perceived to be a relatively benign medical condition. Moreover, the option of transcatheter radiofrequency pulmonary vein isolation in addition to left atrial ablative procedures has introduced another alternative to surgery.¹⁸ From the surgeon’s perspective there has been reluctance to embrace the procedure due to technical complexity. Recent development of devices capable of creating linear lesions as an alternative to surgical incisions, such as radiofrequency and laser energy sources and cryoablative energy sinks, have reinvigorated interest in the field. There have been further modifications of the “maze” lesion set to accommodate the physical limitations of these devices. Whether these modifications will be as effective as the classic maze procedure is unclear. Furthermore, the transmural nature of the lesions is in question. Some have argued that these lesions may actually be proarrhythmic in the early postoperative phase by creating zones of uneven conduction block. Reducing the complexity of the procedure may have come at the price of reduced efficacy.

Therefore, despite almost 2 decades of clinical experience and a recent flurry of interest, fundamental questions about the maze procedure remain; the most important is identifying which patients will benefit from this procedure. The study by Doukas and colleagues⁷ approaches this question from the optimal perspective of a randomized controlled trial. In this study, 97 patients undergoing surgical intervention for mitral valve disease were randomized to also undergo a concomitant left-sided (partial) maze procedure using a unipolar radiofrequency device. All enrolled patients underwent ligation of the left atrial appendage. Three quarters of patients had mitral valve repair, with about one quarter of patients undergoing valve replacement; 11% of patients had an “edge-to-edge” repair (the Alfieri repair),¹⁹ a procedure that has not gained wide acceptance among surgeons because of concerns of subsequent mitral stenosis.

The authors demonstrated that the rate of conversion to sinus rhythm was higher among patients undergoing left-sided maze procedure. At all assessments sinus rhythm was present in 44% of patients undergoing radiofrequency ablation vs 4.5% of control patients at 12 months. Successful restoration of sinus rhythm also was associated with an increase in ejection fraction and a decrease in left ventricular end systolic dimensions, as well as a decrease in plasma B-type natriuretic peptide and an increase in the shuttle walk distance. These outcomes were achieved without increasing the operative mortality rate, leading the authors to conclude that a concomitant maze procedure should be considered during mitral valve surgery among patients with atrial fibrillation.

The strengths of the study by Doukas et al include a well-defined lesion set that closely mimics the classic left-sided maze procedure, a standardized exercise test, and the assessment of biomarkers. However, there are several important limitations. At 1 year, sinus rhythm was present in only 44% of patients undergoing the intervention—a success rate that is disappointing and well below the 90% to 95% success rate Cox reported when the maze procedure is performed in the classic manner with scissors and suture.⁹ Could this be due to nontransmurality of the lesions? Despite some evidence that biatrial lesions provide no advantage in conversion rate over left atrial lesions alone,²⁰ omission of the right-sided lesions remains another possible explanation. Additional factors are the complex mix of patients and the adjunctive procedures performed. For example, a significant number of patients had mitral valve replacement, which may have a profound impact on ventricular performance and functional status postoperatively. Furthermore, operative mortality was 6% in the radiofrequency ablation group and 8% in the control group, making the differential surgical risk of adding a concomitant maze procedure difficult to determine. By comparison, in the Society of Thoracic Surgeons database the operative mortality is less than 2% for mitral repair and approximately 6% for mitral valve replacement.²¹ In addition, the true rate of conversion to sinus rhythm is difficult to assess. Ambulatory rhythm monitors were applied only among symptomatic patients, and intermittent asymptomatic episodes of atrial fibrillation are common.²²

Given these results, what is the place of the maze procedure in the current era? The incremental risk of the procedure must be balanced against the benefit of conversion to normal sinus rhythm in terms of symptom relief, ventricular performance, and risk of thromboembolism. One approach is to add a maze procedure routinely for patients (such as those included in the trial reported by Doukas et al) already undergoing mitral valve repair, with the aim of freeing them from long-term anticoagulation with warfarin. Among patients for whom valve replacement is necessary, a maze procedure is often added if the patient is symptomatic as a consequence of atrial fibrillation, but less often if a mechanical valve prosthesis is to be used and the patient is unaware of the abnormal rhythm. Another intriguing strategy is the application of the maze procedure in a prophylactic manner for patients undergoing mitral valve repair who are thought to be at high risk of converting to atrial fibrillation in the years following successful valvuloplasty. The role of the maze procedure among other patient subsets, such as those with hypertrophic cardiomyopathy, is less clear and will depend on the ability to restore effective atrial transport function in patients with noncompliant ventricles and the benefits of maintaining electrical sinus rhythm.

The surgical maze procedure as an adjunctive form of therapy in patients undergoing cardiac surgery is continuing to evolve. Its precise role can be determined only by additional, well-controlled clinical trials. Above all, long-term follow-up will be particularly important as new technologies become available. The trial by Doukas et al is an important step, but whether patients who do not have indications for valve surgery should undergo minimally invasive surgical maze procedures remains to be determined. Current evidence suggests that the maze procedure is effective in maintaining sinus rhythm in many patients, that atrial function appears to be reasonably well preserved, and that in experienced hands the adjunctive procedure does not add much to surgical morbidity or mortality. Justifiable expectations are that this will result in symptomatic benefit, an improvement in ventricular function, a lower rate of thromboembolism, and perhaps improved survival.^{23 - 24} Proof of concept is more elusive and will require enrolling more patients, addressing difficulties in measuring atrial and ventricular function, and monitoring asymptomatic recurrences of atrial fibrillation and the potential impact on thromboembolic stroke.²²

Although the epidemiology of atrial fibrillation, its morbidity, and in some patient subsets its impact on mortality are well known, whether attempts to maintain sinus rhythm will translate into a meaningful clinical benefit remains to be determined. The logic that sinus rhythm is preferable is obvious but the burden of proof remains. It is hoped that the trial by Doukas et al will be but one of many directed toward such objective verification.