Ten-year experience with the Cox-maze procedure for atrial fibrillation: How do we define success?

Background. The Cox-maze procedure is the standard to which other surgical treatments of atrial fibrillation (AF) are compared. However, evaluation of new devices and lesion sets is difficult because of variable methods of reporting success in eliminating AF. We analyzed 10-year outcome with the "cut and sew" Cox-maze procedure and present rhythm at last follow-up, interval contact, and actuarial AF freedom. Methods. Between March 1993 and December 2002, 335 patients (211 men) underwent the Cox-maze procedure (age, 22 to 83 years; median, 62 years). Atrial fibrillation was chronic (CAF) in 175 patients and paroxysmal (PAF) in 160. Results. Concomitant mitral valve procedures were performed in 59%, coronary artery bypass grafting in 19%, and tricuspid valve repairs in 7%. Early mortality was 0.9%. During hospitalization, transient AF occurred in 29% of patients and 10% required implantation of a new permanent pacemaker (PPM). Dismissal electrocardiogram was normal sinus rhythm in 64%, junctional rhythm in 18%, AF in 11%, and PPM in 7%. At last follow- up (mean 42 +/- 6 months), 88% of patients were free of AF. However, when analyzed by the Kaplan-Meier method, freedom from AF was lower for patients with preoperative lone PAF (5 years, 90%; 10 years, 64%), preoperative lone CAF (5 years, 80%; 10 years, 62%), and patients undergoing combined maze-mitral valve surgery (5 years, 68%; 10 years, 41%). Conclusions. Ten-year results with the standard Cox-maze procedure confirm high effectiveness, but reporting methods should be standardized to account for patients who have transient atrial arrhythmias during long-term follow- up.