OBJECTIVE The Cox-maze procedure is an effective established surgical method for elimination of atrial fibrillation. The aim of this study was to evaluate the efficacy and safety of radiofrequency ablation as a surgical adjunct in treating atrial fibrillation and to adapt maze principles to mitral valve surgery using transseptal approaches. METHODS Thirty-six patients underwent radiofrequency modified maze procedure using standard and cooled ablation electrodes in combination with surgery for hemodynamically significant mitral valve disease and chronic, persistent or paroxysmal atrial fibrillation. Fourteen (39%) patients had mitral valve repair and 22 (61%) patients had mitral valve replacement; 33 (91.6%) patients underwent concomitant surgical procedures. RESULTS The cardiopulmonary bypass time was 162+/-38 min, the aortic cross-clamp time - 98+/-22 min. The additional aortic cross-clamp time required for the radiofrequency modified maze procedure was 21+/-3 min. Postoperative mortality was 2.8%. In 7 of 36 pts (19.4% ) significant bradycardia requiring permanent pacing (six - AAI, one - DDD) was seen after the operation. Patients were followed up for 4 months to 2.5 years after surgery (mean 11.2+/-7.4 months). There were no deaths, no thromboembolic complications during follow-up. Freedom from atrial fibrillation or flutter was 100% at the end of operation, but during 0.5-3 months after operation atrial fibrillation or flutter were observed in 41.6% (15/36) patients. In 2 (5%) patients chronic atrial fibrillation developed. Freedom from atrial fibrillation/flutter was 86% (12/14) at 1 year postoperatively. CONCLUSIONS The radiofrequency modified maze as an adjunctive procedure is safe and effective in eliminating atrial fibrillation using standard and cooling-tip ablation electrodes in combination with surgery for mitral valve disease.