GUEST EDITORIAL Can Implantable Devices Detect and Pace-Terminate Atrial Fibrillation? CARSTEN W. ISRAEL* and S. SERGE BAROLD† *From J. W. Goethe University Hospital, Department of Medicine, Division of Cardiology, Frankfurt, Germany, and †Tampa General Hospital, Tampa, Florida Modern pacemakers are equipped with algo- rithms to detect atrial fibrillation (AF) and perform automatic mode switching for pacing rate control. Counters reporting the number of mode switch- ing episodes allow estimation of the incidence and clinical impact of AF. Refined counter functions provide histograms depicting the duration and maximum atrial rate of detected AF episodes. 1 De- vices can store individual AF episodes and even- tually display them with marker and cycle length annotations, and more recently, with atrial elec- trograms. 2 The advent of stored electrograms has added an important dimension in the evaluation of appropriate and inappropriate AF detection: the latter is most frequently due to farfield oversensing of ventricular signals. 3,4 Fitted with these mem- ory functions, implantable devices act like power- ful monitors capable of documenting the impact of AF therapy (drugs, pacing, and ablation). Contem- porary implanted devices can continuously mon- itor the cardiac rhythm and record episodes in- terpreted as tachyarrhythmias by storing electro- grams that can be subsequently analyzed to es- tablish whether device diagnosis was appropriate. This storage process is substantially more sensitive in the detection of asymptomatic or short-lived ar- rhythmias than any other diagnostic tool includ- ing event recorders and daily transtelephonic ECG transmission, the latter providing only snapshots of several seconds. Therefore, device derived pa- rameters such as the “AF burden” (percentage of time during the preceding follow-up period spent in AF) have already replaced traditional study end- points, such as the time to recurrence of the first symptomatic AF episode. 5-7 Characterization of onset patterns may increase our knowledge about AF mechanisms and may promote more powerful strategies to maintain sinus rhythm. Devices have also attempted to automatically classify detected atrial tachyarrhythmias as AF and non-AF. Some devices rely on the maxi- mum detected atrial rate. For instance, a max- imum atrial rate of >300 versus 250 versus Address for reprints: Carsten W. Israel, M.D., J. W. Goethe University, Department of Medicine, Division of Cardiol- ogy, Theodor-Stern-Kai 7, 60590 Frankfurt, Germany. Fax: +49/69/6301-6341; e-mail: C.W.Israel@em.uni-frankfurt.de Received June 17, 2003; accepted June 17, 2003. 180 beats/min may imply that AF, atrial flutter, or atrial/sinus tachycardia has occurred. However, the detected maximum rate may be misleading with double sensing of broad but discrete atrial signals. 8 Some pacemakers and implantable car- dioverter/defibrillators (ICD) now offer advanced algorithms that automatically classify detected atrial tachyarrhythmias as AF and non-AF. These dedicated algorithms perform non-AF/AF classifi- cations on the basis of detected rate (median atrial cycle length of 12 intervals) and regularity (sec- ond shortest and second longest of 12 atrial cycle lengths above or below 25% different from median cycle length) 9,10 The cut-off values for the range of rates, are programmable and allow some over- lap. Therefore, an “AT detection zone” may range from 200 to 340 ms together with an “AF detec- tion zone” from 100 to 270 ms. In the overlap zone (in this example median atrial cycle lengths be- tween 200 and 270 ms), the classification relies on the detected regularity. This device classification of non-AF and AF has also been used in several studies with pacemakers 7,11 and ICDs 5,12-18 capa- ble of terminating atrial tachyarrhythmias by atrial antitachycardia pacing (ATP; burst+, ramp, and manual 50-Hz burst in pacemakers, additionally automatic 50-Hz burst and atrial shocks in ICD sys- tems). A success rate of 39%–71% of atrial ATP has been reported for tachycardia episodes in the “AT zone.” 7,11,12,14,16-18 However, in some studies ATP and 50-Hz burst therapies have also been applied in the zone classified as AF by the devices, with a success rate of 14%–30% for aggressive atrial ATP and 50-Hz burst (Table I). These results are at vari- ance to our understanding of AF as an arrhythmia characterized by multiple reentry circuits, lack of an excitable gap, or by an excitable gap allowing, at best, local capture. Experimental and clinical stud- ies have consistently shown that atrial ATP can only change some of the reentrant circuit exten- sions and directions, but cannot terminate AF. 19 The contradictory results on AF termination by pacing, based on electrophysiological studies versus device memory information, have created a rather unproductive discussion as to whether AF can actually be terminated by pacing. There are several possible explanations for the discor- dant results: (1) a considerable proportion of AF episodes may terminate fortuitously regardless of PACE, Vol. 26 October 2003 1923