The prevalence data for atrial fibrillation (AF) are dated. The present retrospective study estimated the current and projected prevalence of AF and atrial flutter (AFL) in the United States using a large national database. Claims data drawn from July 2004 to December 2005 from the MarketScan research databases from Thomson Reuters were used to identify patients aged >or=20 years with nontransient AF and/or AFL and age- and gender-matched controls without these conditions. Of the 21,648,681 patients in the databases, 242,903 (1.12%) had nontransient AF and/or AFL (222,605 AF only, 5,376 AFL only, and 14,922 AF and AFL). Patients with AF only, AFL only, and AF and AFL had a greater (p <0.001) prevalence of co-morbidities, including hypertension (62.0%, 61.3%, and 57.0%, respectively) and coronary artery disease (43.0%, 44.7%, and 44.5%, respectively), than matched controls (45.1% hypertension and 19.4% coronary artery disease). Applying the US Census Bureau population estimates to the prevalence rates for AF and/or AFL in the databases, it was estimated that 3.03 million persons in the United States had AF only, 0.07 million had AFL only, and 0.19 million had AF and AFL in 2005. The projected prevalence for 2050 was 7.56 million for AF only, 0.15 million for AFL only, and 0.44 million for AF and AFL. In conclusion, the current prevalence of AF and AFL is high and is projected to increase considerably by 2050. The current and projected increases in the prevalence of AF are greater than predicted by a previous sentinel study and might reflect more than the aging of the population.