Surgical Treatment of Atrial Fibrillation at the Time of Mechanical Valve Replacement
October 23, 2012
Editor's Note: This Article of the Month is based on Kim JB, Moon JS, Yun SC et al. Long-Term outcomes of mechanical valve replacement in patients with atrial fibrillation impact of the maze procedure. Circulation 2012;125:2071-80.
Patients with atrial fibrillation undergoing cardiac surgery generally have worse late outcomes compared to patients pre-operatively in sinus rhythm.1-3 Concomitant AF treatment is performed in the minority of patients at the time of cardiac surgery,4 especially in patients undergoing mechanical valve replacement, as the need for anti-coagulation will be unchanged regardless of late rhythm. "Long-term outcomes of mechanical valve replacement in patients with atrial fibrillation: impact of the Maze procedure"5 examined the role of surgical ablation in this population.
Kim et al. retrospectively review outcomes of 569 patients with atrial fibrillation associated with valvular heart disease between 1999 and 2010. All patients underwent mechanical valve replacement. In addition, 317 patients underwent a concomitant maze, but 252 patients had no AF intervention. To reduce the effect of treatment selection bias, propensity score matching was performed for each group and outcomes compared between patients who received maze and those who did not.
The primary end points were all-cause death and thromboembolic events; secondary endpoints were the composite of death, congestive heart failure requiring hospitalization and valve-related complications. Freedom from atrial fibrillation, echocardiographic dimensions, tricuspid regurgitation and New York Heart Association (NYHA) Classification were also compared.
There were no differences in early mortality (1.2%) or major morbidity between the 2 groups. Throughout the follow-up (median duration 63.6 months) freedom from AF off anti-arrhythmic medications ranged from 70% to 91.2% in the maze group and was superior to the control (16.7% to 27.2%, p<0.001). The linearized rates (maze vs. control) were 1.9% vs. 1.5% for death (p=0.73), 0.4% vs. 0.9% for thromboembolic events (p=0.059) and 4.1% vs. 4.6% for the composite endpoints (p=0.27) per patient-year, respectively. Adjusted hazard rations demonstrated that maze patients were at lower risk of thromboembolic events but comparable risk of death and composite outcome.
Subgroup analysis was performed by EuroSCORE: low to medium (0-3) and high (>3). In the low EuroSCORE population, the maze procedure reduced the risk of both thromboembolism (p=0.04) and composite outcome (p=0.049); no differences were observed in the high risk group.
Echocardiographic data >6 months after surgery was available for 459 patients. Propensity matching yielded 103 pairs without differences in baseline characteristics. Compared with preoperative findings, both maze and control patients showed significant reductions in left ventricular (LV) systolic and diastolic dimensions, as well left atrial dimension. Only the maze group demonstrated a significant improvement in LV ejection fraction. Less late tricuspid regurgitation was present after maze, with (p=0.09) or without (p=0.01) repair. Both groups demonstrated significant improvement in NYHA class (0<0.001) but the patients in the maze group had better function (p=0.098).
In AF patients undergoing mechanical valve implantation, the addition of the maze procedure was associated with a reduction in thromboembolic complications and improvements in hemodynamic performance, particularly in those patients at low risk of surgical morbidity and mortality.
Since the results of the results the Atrial Fibrillation Follow-Up Investigation of Rhythm Management (AFFIRM) study6, the benefits of a rhythm control strategy have been elusive in the medical population. Several surgical series have reported benefits of the maze operation, including restoration of sinus rhythm,7 reduction of stroke8 and an improvement of mortality9, even in mechanical valve patients.10 However, most of these early series were small or lacked sophisticated comparison techniques.
Larger series suggest that even with anti-coagulation after mechanical valve replacement, AF patients have a worse survival than those in pre-operative sinus rhythm.11 The impact of surgical correction of AF is beginning to emerge. A prior propensity matched series has suggested that concomitant surgical correction of atrial fibrillation restores survival after treatment population to that of patients without AF, largely due to successful sinus restoration.12 However, this was in a patient population that had predominantly valve repairs or bioprosthetic valves.
In this series, Kim et al. employ sophisticated statistical analysis to assess the impact of the addition of a maze on the mechanical valve population. These data suggest that maze will reduce thromboembolic events in the low-risk mechanical valve patient without effecting survival. In the subsequent sub-analysis of the AFFIRM trial, both sinus rhythm and anticoagulation reduced late mortality.13 If the findings of this study are born out in further investigations, sinus rhythm may not provide additional benefits to survival in the patient on anti-coagulation. However, other benefits, such as improvement in heart function and heart failure symptoms may occur.
The reduction in thromboembolic events in this group is more challenging to explain. It is clouded by the fact that the maze procedure eliminates the Left Atrial Appendage, a frequent origin of embolic material.14 Presumably, thromboembolism is also increased in the presence of the mechanical valve. Although this was potentially reduced by a contracting atrium in sinus rhythm, the international normalized ratio target was only 2.0 to 3.0 in this study. Thus, the benefit of sinus restoration/AF treatment, may have potentially been eliminated by a higher INR target. Recent studies suggest that even in sinus rhythm the annual late stroke rate may be as high as 5%, with patients with poor atrial contraction at higher risk.15 The interplay of all these factors requires more investigation to determine the relative contribution of each to thromboembolic events.
As with every procedure, there is a balance between risk and benefit when treating atrial fibrillation. In the raw data, Kim et al. describe approximately 30 minutes of increased cross-clamp time in patients in whom a Maze was performed. However, the concomitant procedures were different in the groups and one would expect a difference in cross-clamp time. Nonetheless, even prior to propensity matching there was no difference in morbidity or mortality in the Maze group. In a retrospective review of the Society of Thoracic Surgeons database including 12,235 patients with AF undergoing mitral valve surgery, after adjusting for various risk factors, the addition of a Maze only increased cross-clamp time 9 minutes and increased cardiopulmonary bypass time 9 additional minutes; there was no effect on morbidity or mortality.4
Surgical AF treatment clearly offers benefits to subsets of patients undergoing cardiac surgery. Benefits may include a reduction of thromboembolism, improved cardiac function, improved symptoms and even an improvement in survival in select patient populations. In experienced hands, with the proper patient selection, concomitant AF treatment does not appear to add risk to the cardiac procedure. Healthier patients may benefit more than those at higher surgical risk. Future trials must focus on specific homogenous populations to determine the patients who will benefit most.