The calculation of the corrected QT interval (QTc) is particularly problematic in patients during atrial fibrillation (AF). The aims of this study were to compare the QTc calculated using Bazett's formula in AF and sinus rhythm (SR) and determine whether alternative methods for QT correction were superior to Bazett's, in an effort to define the optimal method for QT correction in patients with AF. We evaluated consecutive patients with persistent AF admitted for initiation of dofetilide. The QT interval was corrected according to the following formulas: Bazett's, Fridericia, and Framingham. We compared the QTc interval on the last electrocardiogram in AF to the first electrocardiogram in SR. The cohort included 54 patients (age 60 ± 10 years, 80% men) with persistent AF for a median of 36 months. Bazett's overestimated QTc during AF compared with SR (464 ± 34 vs 445 ± 38 ms, p = 0.008); in contrast, Framingham underestimated it (385 ± 48 vs 431 ± 40 ms, p <0.001, respectively). However, there was no significant difference between the QTc interval in AF and SR when assessed by Fridericia (435 ± 33 vs 440 ± 35 ms, p = 0.46). There were 24 dofetilide dose reductions based on Bazett's QTc; this would have been avoided in 33% of patients had Fridericia been used. In conclusion, the commonly used Bazett's formula leads to an overestimation of the QTc during AF. This may result in unnecessary reduction in antiarrhythmic doses and thus drug efficacy. These data suggest that the Fridericia most closely approximates the QTc during AF to QTc during SR.