"Spot" ECGs underestimate atrial fibrillation recurrence after surgical ablation

OBJECTIVE: Historically, success of surgical treatment of atrial fibrillation (AF) has been measured by electrocardiograms (ECGs) at various intervals. However, continuous monitoring of cardiac rhythms by "autocapture" devices has recently become more available and convenient. The concordance of measurements of freedom from AF by these two techniques has not been reported after surgical ablation. METHODS: Between August 2005 and May 2006, 47 patients at a single academic center underwent surgical ablation procedures for AF and had recurrence of AF assessed by both "spot" 12-lead ECG and autocapture event monitoring. Forty-one ablation procedures were concomitant with other cardiac surgery and six were stand alone, nonsternotomy procedures. Agreement between these diagnostic modes was measured using the κ statistic at 3, 6, and 12 months (κ of 1 is perfect agreement, 0 is no agreement). McNemar test was employed to determine whether agreement significantly changed from 3 to 12 months. RESULTS: At 3 months follow-up, spot ECGs suggested that 81% (38 of 47) of surgical patients were free of any AF, whereas 1-week event recordings found only 70% (31 of 44) of patients were free of any AF. At 6 months, spot ECGs estimated that 87% (40 of 46) of surgical patients were free of AF; 1-week event recordings found only 74% (34 of 46) of patients were free of AF. At 12 months, spot ECGs estimated that 84% (26 of 31) of surgical patients were free of AF compared with only 68% (19 of 28) as measured by the 1-week event recorder. The κ measures (with 95% confidence interval) at 3, 6, and 12 months were 0.52 (0.24-0.80), 0.60 (0.32-0.87), and 0.63 (0.32-0.94) respectively, showing only moderate agreement. McNemar test showed no significant shift in agreement from 3 to 6 months (P = 0.7055), 3 to 12 months (P = 1.000), or 6 to 12 months (P = 1.000). There were no deaths or strokes, but one myocardial infarction among these 47 patients during 12 months follow-up. CONCLUSIONS: "Spot" ECGs underestimate the incidence of recurrent AF after surgical ablation for AF and show poor agreement with the more reliable 1-week autocapture event recordings.