Background: The Japanese Rhythm Management Trial for Atrial Fibrillation (J-RHYTHM) study showed rhythm control was associated with fewer changes in the assigned treatment strategy compared to rate control in atrial fibrillation (AF). The aim was to describe how antiarrhythmics (AAs) were altered in the rhythm control arm and whether altering AAs would impact long-term outcomes. Methods and Results: Of 390 enrolled patients, 23.5% altered their AAs (drug alteration [DA] group). The hard endpoint (HE) was defined as a composite of death, stroke, embolism, major bleeding or heart failure hospitalization; soft endpoint (SE) was defined physical/psychological disability requiring alteration of treatment strategy. The patients were followed for 1.7 years. No significant difference was noted in the occurrence of HE (4.0% vs 6.5%, P=0.31), but DA-group patients had higher rates of SE (9.3% vs 18.4%, P=0.017) compared to single AA patients. The DA group was also associated with the occurrence of SE after adjustment (HR 1.90, P=0.042). When the DA group was subdivided according to the use of class III drugs or change of drugs between classes, there were no differences in outcomes. Conclusions: The need to change AA was associated with physical/psychological disabilities that seemed not to be relieved simply by changing AAs, and this should be considered as a marker for refractory paroxysmal AF requiring other strategies. (Circ J 2010; 74: 870 - 875)