Observations on the transition from intermittent to permanent atrial fibrillation

Background Quantitative data on the frequency with which transition from in termittent to permanent atrial fibrillation occurs are lacking. We conducte d this study to determine the proportion of patients with intermittent atri al fibrillation who progress to permanent atrial fibrillation and to invest igate baseline clinical characteristics that might predict such a progressi on. Methods This retrospective cohort study included 231 patients who were seen with intermittent atrial fibrillation at a university hospital-based clini c from January 1978 through December 1997. Patients' medical records and el ectrocardiograms were reviewed and data were collected for all clinic visit s through May 1998. The proportion of patients who remained free of transit ion from intermittent to permanent atrial fibrillation was calculated by th e Kaplan-Meier method. A Cox proportional hazards model was used to determi ne the effect of some baseline characteristics on this transition. Results The number of patients who remained Free of transition from intermi ttent to permanent atrial fibrillation was 92% (95% confidence interval 88% -96%) at 1 year and 82% (95% confidence interval 75%-88%) at 4 years. Among 5 baseline characteristics (age, sex, structural heart disease, atrial fib rillation at presentation, and use of an antiarrhythmic medicine before pre sentation), the 2 significant predictors of progression from intermittent t o permanent atrial fibrillation were age (P = .0003) and being in atrial fi brillation at presentation (P = .0006). The hazard ratio associated with 10 years of advancing age was 1.82 (95% confidence interval 1.31-2.51), and t he hazard ratio associated with atrial fibrillation at presentation was 3.5 6 (95% confidence interval 1.73-7.34). Conclusions Approximately 18% of patients who had intermittent atrial fibri llation were permanently in atrial fibrillation after 4 years of follow-up. Age and being in atrial fibrillation at presentation were the only 2 impor tant clinical variables identified in predicting such a progression.