- Email: [email protected]
Analysis of atrial fibrillatory rate during spontaneous episodes of atrial fibrillation in humans using implantable loop recorder electrocardiogram
Available online at www.sciencedirect.com
Journal of Electrocardiology 45 (2012) 723 – 726 www.jecgonline.com
Analysis of atrial fibrillatory rate during spontaneous episodes of atrial fibrillation in humans using implantable loop recorder electrocardiogram Pyotr G. Platonov, MD, PhD, a, b,⁎ Martin Stridh, PhD, a, c Mirko de Melis, PhD, d Lubos Urban, MD, e Jonas Carlson, PhD, a, b Giorgio Corbucci, PhD, d Fredrik Holmqvist, MD, PhD a, b a Department of Cardiology, Clinical Sciences, Lund University, Lund, Sweden Center for Integrative Electrocardiology at Lund University (CIEL), Lund, Sweden c Department of Electroscience, Lund Institute of Technology, Lund University, Lund, Sweden d Medtronic BRC, Maastricht, The Netherlands e The National Institute of Cardiovascular Diseases, Bratislava, Slovakia Received 8 March 2012 b
Abstract
Atrial fibrillatory rate (AFR) can predict outcome of interventions for atrial fibrillation (AF); however, AFR behavior at AF onset in humans is poorly described. We studied AFR during spontaneous AF episodes in patients with lone paroxysmal AF who received implantable loop recorders and had AF episodes of 1 hour or more recorded (n = 4). Mean AFR per minute was assessed from continuous implantable loop recorder electrocardiogram using spatiotemporal QRST cancellation and time-frequency analysis. Atrial fibrillatory rate increased from 290 ± 20 to 326 ± 39 fibrillations per minute during the first 3 hours (P b .05) and reached plateau then. Atrial fibrillatory rate beyond the initial 3 hours can, therefore, be considered stable and may be evaluated for prediction of intervention effect. © 2012 Elsevier Inc. All rights reserved.
Introduction
Materials
The shortening of the atrial effective refractory period in association with the initiation and persistency of atrial fibrillation (AF) episodes is a well-described phenomenon, also known as atrial electrical remodeling. Atrial remodeling facilitates the induction of AF and can be observed as acceleration of the atrial fibrillatory rate (AFR) during persistent AF episodes. Most data supporting this theory originate from experimental studies 1 or clinical observations made during AF episodes induced during invasive electrophysiological examinations. 2 Atrial fibrillatory rate behavior at the initiation of spontaneous and often self-terminating AF in humans has mostly been out of reach for clinicians, and with few exceptions, 3,4 experimental findings of AFR acceleration during the first minutes and hours of AF are largely unproven in clinical settings. The aim of our study was to assess AFR behavior at initiation and during the first hours of spontaneous AF episodes in patients with paroxysmal AF using time-frequency analysis applied to electrocardiogram (ECG) retrieved from implantable cardiac monitors (ICMs).
Patients and signal acquisition
⁎ Corresponding author. Department of Cardiology, Lund University Hospital, SE-22185 Lund, Sweden. E-mail address: [email protected] 0022-0736/$ – see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jelectrocard.2012.05.003
Consecutive patients with paroxysmal AF (n = 14; age, 57 ± 7 years; 12 men) received an ICM (Reveal XT, Model 9529; Medtronic, Inc, Minneapolis, MN) due to symptoms attributable to AF. Patients presenting with persistent or permanent AF were excluded. Following a stabilization phase of 4 to 6 weeks after ICM implantation, patients underwent 46 hours of continuous recording using Holter recorders (DR220; NorthEast Monitoring, Inc, Maynard, MA), which, through a telemetry antenna, were used to store simultaneous and continuous 46-hour recordings of subcutaneous ECG. The study protocol was approved by an institutional medical ethics committee, and all subjects gave their written informed consent. Signal processing The 46-hour long ECG recordings were manually reviewed for presence of spontaneous AF episodes. Episodes were selected for further analysis only if they lasted longer than 1 hour, and AF onset was captured during registration. The ECG containing the entire continuous AF episode was
724
P.G. Platonov et al. / Journal of Electrocardiology 45 (2012) 723–726
then exported for off-line automated computerized processing using AFR tracker software (CardioLund Research AB, Lund, Sweden). The ECG signals were preprocessed, including baseline filtering, beat detection, and cross-correlation–based beat classification. Spatiotemporal QRSTcancellation, in which an amplitude and morphology adjusted average beat is subtracted from each beat in the signal, was then used. 5 Individual beat averages were used for beats belonging to different beat classes. The resulting residual ECG containing mainly the atrial activity was analyzed using sequential atrial signal characterization, 6 which performs time-frequency analysis using overlapping windows of short duration (2.56 seconds long, 1 window per second) to provide second-bysecond trends of the atrial fibrillation rate. With this method, the signal structure of each window is continuously analyzed by means of its harmonic frequency pattern to assure that the corresponding signal contains an oscillatory atrial signal. 7 For intervals containing a fibrillatory signal, the second-bysecond frequency estimates are summarized into 60-second AFR (Fig. 1). Based on minute-to-minute AFR assessments, AFR values were then averaged per 10 minutes and 1 hour for graph constructions and statistical comparisons. Statistical analysis All data are presented as means ± 1 SD unless stated otherwise. Comparisons of AFR values at different time intervals from onset of AF episodes were performed using
Friedman's 2-way analysis of variance by ranks for dependent variables. P b .05 was considered statistically significant. Results Data availability Of 14 patients, 9 had AF episodes exceeding 1 hour recorded during monitoring. Of those, 5 patients had AF onset captured on the ECG onset and underwent AFR assessment; however, in 1 patient, signal quality was insufficient for continuous AFR calculation during the entire registration, mostly due to noise prohibiting AFR measurement. Thus, continuous ECG recordings obtained from 4 remaining patients comprised the data available for analysis. In 1 patient, both the onset and end of AF paroxysm were recorded (total time in AF, 8.5 hours). In remaining 3 patients, AF was still present at the end of the 46-hour recording period. Total duration of recorded AF episodes available for analysis was 10, 33, and 39 hours, respectively. Clinical characteristics Patients included in the analysis (n = 4; mean age, 52 ± 9 years; 3 men) had lone AF without any history of congestive heart failure, ischemic heart disease, stroke, or hypertension and were not receiving antiarrhythmic drugs during ECG recording time. Every patient had a moderately dilated left atrium (left atrial diameter 45 ± 5 mm) and normal left ventricular function.
Fig. 1. Noninvasive assessment of the AFR using spatiotemporal QRST cancellation and time-frequency analysis applied to the bipolar implantable loop recorder ECG lead (X1). rX, residual atrial ECG signal after QRST cancellation.
P.G. Platonov et al. / Journal of Electrocardiology 45 (2012) 723–726 Table 1 Mean AFR during initial 10 minutes of hour 1 though hour 6 of AF episode Patient Age Sex no.
Mean AFR during initial 10 min of hour, fpm. AF onset
Second hour
Third hour
Fourth hour
Fifth hour
Sixth hour
1 2 3 4
307 300 294 260
330 321 326 269
358 327 350 271
365 334 337 270
363 337 310 275
361 338 304 277
40 58 52 67
Male Male Male Female
AFR dynamics at onset of AF episodes In all patients, AFR increased during the first 3 to 4 hours of the AF episode and then reached a plateau. The Table 1 represents the mean AFR values recorded during the first 10 minutes of hour 1 through hour 6. Atrial fibrillatory rate averages per hour showed a significant increase from 290 ± 20 during hour 1 to 311 ± 28 fibrillations per minute (fpm) at hour 2 and further rising to 326 ± 39 fpm at hour 3 (P ± .035; Fig. 2A). The AFR increase phenomenon is even clearer when hourly average AFR values are normalized to the mean value of AFR during plateau observed during hour 4 through hour 6 (Fig. 2B). Fig. 3 represents individual AFR trends in all 4 patients based on the 10-minute averages of AFR. The graphs are based on 82.7%, 99.7%, 50.5%, and 99.6% (patients A, B, C, and D, respectively) of the total number of 1-minute segments that comprise the first 6 hours of the AF episodes. In 1 patient (patient D, Fig. 3), both the onset and end of the AF episode were recorded. The AF episode was characterized by an initial increase of AFR from 250 fpm at AF onset to 270 fpm by the end of the second hour and a further slow increase to a maximum of 280 fpm by the sixth hour. Gradual slowing of AFR was then observed during the final hour preceding spontaneous conversion of the AF episode to sinus rhythm. Discussion The main finding of our study is the demonstration of progressive increase in AFR at initiation of spontaneous AF
A
Friedmans’s P = .007
725
episodes, which continued over a period of 3 to 4 hours until reaching a plateau. To the best of our knowledge, the AFR behavior at AF initiation has been previously addressed in only 2 studies that reported an increase in AFR over the course of 4 to 5 minutes of AF. Bollmann et al 3 were the first to use Holter ECG to characterize AFR behavior during paroxysmal AF; they reported an AFR increase to a mean of 5.8 Hz (348 fpm) by the fifth minute of AF in episodes lasting more than 15 minutes. In a more recent study by Petrutiu et al 4 who looked at spontaneous AF episodes with a median duration of 18 minutes, similar findings were reported, that is, a gradual increase of AFR reaching a plateau at the fourth minute of AF. Although we used similar methodology with noninvasive AFR assessment, our findings to some extent contradict those reported by Bollmann et al 3 and Petrutiu et al 4 because AFR increase in our patients continued far beyond the initial minutes of AF and reached a plateau only by the fourth hour from AF onset. In contrast to the above-cited previous studies, we intended to examine AFR behavior in patients with longer-lasting AF that would be similar to clinical AF and could be grounds for hospital admission. Earlier studies suggested that AFR may have predictive value with regard to sinus rhythm maintenance after electrical cardioversion. 8 Recently, we reported that AFR less than 350 fpm assessed from ECG taken in the emergency department is an independent predictor of spontaneous conversion to sinus rhythm within 18 hours of symptoms onset. 9 To assess properties of AF that might be relevant for emergency department settings, arrhythmia duration in our study exceeded 1 hour in all 4 patients (median, 1290 minutes compared with 18 minutes in the study by Petrutiu et al). The differences in AFR dynamics at AF onset may reflect the difference in whether the atrial substrate is capable or incapable of maintaining long-lasting AF. Notably, we did not observe the abrupt drop of AFR preceding spontaneous AF termination that was reported by Petrutiu et al 10 based on data described earlier. 4 In a single patient who had spontaneous termination of AF recorded by ICM, AFR decrease was gradual over a period of 1 hour, from 280 to 260 fpm immediately before conversion. Similar
B
Friedmans’s P = .025
Fig. 2. Mean AFR per hour during the first 6 hours from AF onset. Gradual AFR acceleration is observed during the first 3 to 4 hours of AF.
726
P.G. Platonov et al. / Journal of Electrocardiology 45 (2012) 723–726
Fig. 3. A to D, Individual AFR trends during AF paroxysms presented as 10-minute averages starting from AF onset in all 4 patients. Case D represents AFR trend during 8.5-hour-long AF episode that had both onset (red line) and end (green line) captured on implantable loop recorder ECG. Note that maximal atrial rate during AF did not exceed 280 fpm and spontaneous conversion to sinus rhythm was preceded by gradual atrial rate decrease over the course of 1 hour.
to the initial AFR increase, the differences in the time course of AFR decrease before spontaneous conversion between our findings and earlier reported 10 data may be due to the much longer duration of AF episodes in our study. Moreover, the observation in the setting of AF termination is based on a single case in our study, which suggests that caution should be exercised when interpreting it. Temporal variability in AFR over the course of several hours at AF onset illustrates a potential problem in using AFR for risk stratification or prediction of intervention efficacy. Several earlier studies have shown that lower AFR values are associated with a better outcome of interventions for persistent AF, including electrical, pharmacologic cardioversion or catheter ablation. 8,11,12 In the acute setting of short-duration AF, the value of observed AFR may be dependent on its timing in regard to AF onset as illustrated by our study. However, the practical implication of our findings is that AFR recorded in excess of 3 to 4 hours after AF onset is likely to be of value and, as our group recently showed, can predict spontaneous termination of AF. 9 Despite the uniform pattern of AFR increase during the initial hours of long-lasting AF paroxysms observed in all 4 subjects selected using predefined criteria, the limited number of patients available for analysis pose an important limitation to our study, and caution should be exercised when interpreting these findings. Conclusion Atrial fibrillatory rate acceleration occurs during the initial 3 to 4 hours of AF, after which it reaches a plateau. Our findings should be considered whenever AFR is used as a predictor of clinical outcome early in the course of AF paroxysms because the timing of AFR readings may have data interpretation implications.
References 1. Goette A, Honeycutt C, Langberg JJ. Electrical remodeling in atrial fibrillation. Time course and mechanisms. Circulation 1996;94:2968. 2. Ravelli F, Mase M, Del Greco M, Faes L, Disertori M. Deterioration of organization in the first minutes of atrial fibrillation: a beat-to-beat analysis of cycle length and wave similarity. J Cardiovasc Electrophysiol 2007;18:60. 3. Bollmann A, Sonne K, Esperer HD, Toepffer I, Langberg JJ, Klein HU. Non-invasive assessment of fibrillatory activity in patients with paroxysmal and persistent atrial fibrillation using the Holter ECG. Cardiovasc Res 1999;44:60. 4. Petrutiu S, Sahakian AV, Swiryn S. Short-term dynamics in fibrillatory wave characteristics at the onset of paroxysmal atrial fibrillation in humans. J Electrocardiol 2007;40:155. 5. Stridh M, Sornmo L. Spatiotemporal QRST cancellation techniques for analysis of atrial fibrillation. IEEE Trans Biomed Eng 2001;48:105. 6. Stridh M, Sornmo L, Meurling CJ, Olsson SB. Sequential characterization of atrial tachyarrhythmias based on ECG time-frequency analysis. IEEE Trans Biomed Eng 2004;51:100. 7. Stridh M, Bollmann A, Olsson SB, Sornmo L. Detection and feature extraction of atrial tachyarrhythmias. A three stage method of timefrequency analysis. IEEE Eng Med Biol Mag 2006;25:31. 8. Holmqvist F, Stridh M, Waktare JE, Sornmo L, Olsson SB, Meurling CJ. Atrial fibrillatory rate and sinus rhythm maintenance in patients undergoing cardioversion of persistent atrial fibrillation. Eur Heart J 2006;27:2201. 9. Choudhary M, Holmqvist F, Carlson J, Nilsson H-J, Platonov PG. Low atrial fibrillatory rate is associated with spontaneous conversion of recent onset atrial fibrillation. Europace 2011;13:P524. 10. Petrutiu S, Sahakian AV, Swiryn S. Abrupt changes in fibrillatory wave characteristics at the termination of paroxysmal atrial fibrillation in humans. Europace : European pacing, arrhythmias, and cardiac electrophysiology : journal of the working groups on cardiac pacing, arrhythmias, and cardiac cellular electrophysiology of the European Society of Cardiology 2007;9:466–70. 11. Matsuo S, Lellouche N, Wright M, et al. Clinical predictors of termination and clinical outcome of catheter ablation for persistent atrial fibrillation. J Am Coll Cardiol 2009;54:788. 12. Bollmann A, Husser D, Stridh M, et al. Frequency measures obtained from the surface electrocardiogram in atrial fibrillation research and clinical decision-making. J Cardiovasc Electrophysiol 2003;14:S154.
Journal of Electrocardiology 45 (2012) 723 – 726 www.jecgonline.com
Analysis of atrial fibrillatory rate during spontaneous episodes of atrial fibrillation in humans using implantable loop recorder electrocardiogram Pyotr G. Platonov, MD, PhD, a, b,⁎ Martin Stridh, PhD, a, c Mirko de Melis, PhD, d Lubos Urban, MD, e Jonas Carlson, PhD, a, b Giorgio Corbucci, PhD, d Fredrik Holmqvist, MD, PhD a, b a Department of Cardiology, Clinical Sciences, Lund University, Lund, Sweden Center for Integrative Electrocardiology at Lund University (CIEL), Lund, Sweden c Department of Electroscience, Lund Institute of Technology, Lund University, Lund, Sweden d Medtronic BRC, Maastricht, The Netherlands e The National Institute of Cardiovascular Diseases, Bratislava, Slovakia Received 8 March 2012 b
Abstract
Atrial fibrillatory rate (AFR) can predict outcome of interventions for atrial fibrillation (AF); however, AFR behavior at AF onset in humans is poorly described. We studied AFR during spontaneous AF episodes in patients with lone paroxysmal AF who received implantable loop recorders and had AF episodes of 1 hour or more recorded (n = 4). Mean AFR per minute was assessed from continuous implantable loop recorder electrocardiogram using spatiotemporal QRST cancellation and time-frequency analysis. Atrial fibrillatory rate increased from 290 ± 20 to 326 ± 39 fibrillations per minute during the first 3 hours (P b .05) and reached plateau then. Atrial fibrillatory rate beyond the initial 3 hours can, therefore, be considered stable and may be evaluated for prediction of intervention effect. © 2012 Elsevier Inc. All rights reserved.
Introduction
Materials
The shortening of the atrial effective refractory period in association with the initiation and persistency of atrial fibrillation (AF) episodes is a well-described phenomenon, also known as atrial electrical remodeling. Atrial remodeling facilitates the induction of AF and can be observed as acceleration of the atrial fibrillatory rate (AFR) during persistent AF episodes. Most data supporting this theory originate from experimental studies 1 or clinical observations made during AF episodes induced during invasive electrophysiological examinations. 2 Atrial fibrillatory rate behavior at the initiation of spontaneous and often self-terminating AF in humans has mostly been out of reach for clinicians, and with few exceptions, 3,4 experimental findings of AFR acceleration during the first minutes and hours of AF are largely unproven in clinical settings. The aim of our study was to assess AFR behavior at initiation and during the first hours of spontaneous AF episodes in patients with paroxysmal AF using time-frequency analysis applied to electrocardiogram (ECG) retrieved from implantable cardiac monitors (ICMs).
Patients and signal acquisition
⁎ Corresponding author. Department of Cardiology, Lund University Hospital, SE-22185 Lund, Sweden. E-mail address: [email protected] 0022-0736/$ – see front matter © 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jelectrocard.2012.05.003
Consecutive patients with paroxysmal AF (n = 14; age, 57 ± 7 years; 12 men) received an ICM (Reveal XT, Model 9529; Medtronic, Inc, Minneapolis, MN) due to symptoms attributable to AF. Patients presenting with persistent or permanent AF were excluded. Following a stabilization phase of 4 to 6 weeks after ICM implantation, patients underwent 46 hours of continuous recording using Holter recorders (DR220; NorthEast Monitoring, Inc, Maynard, MA), which, through a telemetry antenna, were used to store simultaneous and continuous 46-hour recordings of subcutaneous ECG. The study protocol was approved by an institutional medical ethics committee, and all subjects gave their written informed consent. Signal processing The 46-hour long ECG recordings were manually reviewed for presence of spontaneous AF episodes. Episodes were selected for further analysis only if they lasted longer than 1 hour, and AF onset was captured during registration. The ECG containing the entire continuous AF episode was
724
P.G. Platonov et al. / Journal of Electrocardiology 45 (2012) 723–726
then exported for off-line automated computerized processing using AFR tracker software (CardioLund Research AB, Lund, Sweden). The ECG signals were preprocessed, including baseline filtering, beat detection, and cross-correlation–based beat classification. Spatiotemporal QRSTcancellation, in which an amplitude and morphology adjusted average beat is subtracted from each beat in the signal, was then used. 5 Individual beat averages were used for beats belonging to different beat classes. The resulting residual ECG containing mainly the atrial activity was analyzed using sequential atrial signal characterization, 6 which performs time-frequency analysis using overlapping windows of short duration (2.56 seconds long, 1 window per second) to provide second-bysecond trends of the atrial fibrillation rate. With this method, the signal structure of each window is continuously analyzed by means of its harmonic frequency pattern to assure that the corresponding signal contains an oscillatory atrial signal. 7 For intervals containing a fibrillatory signal, the second-bysecond frequency estimates are summarized into 60-second AFR (Fig. 1). Based on minute-to-minute AFR assessments, AFR values were then averaged per 10 minutes and 1 hour for graph constructions and statistical comparisons. Statistical analysis All data are presented as means ± 1 SD unless stated otherwise. Comparisons of AFR values at different time intervals from onset of AF episodes were performed using
Friedman's 2-way analysis of variance by ranks for dependent variables. P b .05 was considered statistically significant. Results Data availability Of 14 patients, 9 had AF episodes exceeding 1 hour recorded during monitoring. Of those, 5 patients had AF onset captured on the ECG onset and underwent AFR assessment; however, in 1 patient, signal quality was insufficient for continuous AFR calculation during the entire registration, mostly due to noise prohibiting AFR measurement. Thus, continuous ECG recordings obtained from 4 remaining patients comprised the data available for analysis. In 1 patient, both the onset and end of AF paroxysm were recorded (total time in AF, 8.5 hours). In remaining 3 patients, AF was still present at the end of the 46-hour recording period. Total duration of recorded AF episodes available for analysis was 10, 33, and 39 hours, respectively. Clinical characteristics Patients included in the analysis (n = 4; mean age, 52 ± 9 years; 3 men) had lone AF without any history of congestive heart failure, ischemic heart disease, stroke, or hypertension and were not receiving antiarrhythmic drugs during ECG recording time. Every patient had a moderately dilated left atrium (left atrial diameter 45 ± 5 mm) and normal left ventricular function.
Fig. 1. Noninvasive assessment of the AFR using spatiotemporal QRST cancellation and time-frequency analysis applied to the bipolar implantable loop recorder ECG lead (X1). rX, residual atrial ECG signal after QRST cancellation.
P.G. Platonov et al. / Journal of Electrocardiology 45 (2012) 723–726 Table 1 Mean AFR during initial 10 minutes of hour 1 though hour 6 of AF episode Patient Age Sex no.
Mean AFR during initial 10 min of hour, fpm. AF onset
Second hour
Third hour
Fourth hour
Fifth hour
Sixth hour
1 2 3 4
307 300 294 260
330 321 326 269
358 327 350 271
365 334 337 270
363 337 310 275
361 338 304 277
40 58 52 67
Male Male Male Female
AFR dynamics at onset of AF episodes In all patients, AFR increased during the first 3 to 4 hours of the AF episode and then reached a plateau. The Table 1 represents the mean AFR values recorded during the first 10 minutes of hour 1 through hour 6. Atrial fibrillatory rate averages per hour showed a significant increase from 290 ± 20 during hour 1 to 311 ± 28 fibrillations per minute (fpm) at hour 2 and further rising to 326 ± 39 fpm at hour 3 (P ± .035; Fig. 2A). The AFR increase phenomenon is even clearer when hourly average AFR values are normalized to the mean value of AFR during plateau observed during hour 4 through hour 6 (Fig. 2B). Fig. 3 represents individual AFR trends in all 4 patients based on the 10-minute averages of AFR. The graphs are based on 82.7%, 99.7%, 50.5%, and 99.6% (patients A, B, C, and D, respectively) of the total number of 1-minute segments that comprise the first 6 hours of the AF episodes. In 1 patient (patient D, Fig. 3), both the onset and end of the AF episode were recorded. The AF episode was characterized by an initial increase of AFR from 250 fpm at AF onset to 270 fpm by the end of the second hour and a further slow increase to a maximum of 280 fpm by the sixth hour. Gradual slowing of AFR was then observed during the final hour preceding spontaneous conversion of the AF episode to sinus rhythm. Discussion The main finding of our study is the demonstration of progressive increase in AFR at initiation of spontaneous AF
A
Friedmans’s P = .007
725
episodes, which continued over a period of 3 to 4 hours until reaching a plateau. To the best of our knowledge, the AFR behavior at AF initiation has been previously addressed in only 2 studies that reported an increase in AFR over the course of 4 to 5 minutes of AF. Bollmann et al 3 were the first to use Holter ECG to characterize AFR behavior during paroxysmal AF; they reported an AFR increase to a mean of 5.8 Hz (348 fpm) by the fifth minute of AF in episodes lasting more than 15 minutes. In a more recent study by Petrutiu et al 4 who looked at spontaneous AF episodes with a median duration of 18 minutes, similar findings were reported, that is, a gradual increase of AFR reaching a plateau at the fourth minute of AF. Although we used similar methodology with noninvasive AFR assessment, our findings to some extent contradict those reported by Bollmann et al 3 and Petrutiu et al 4 because AFR increase in our patients continued far beyond the initial minutes of AF and reached a plateau only by the fourth hour from AF onset. In contrast to the above-cited previous studies, we intended to examine AFR behavior in patients with longer-lasting AF that would be similar to clinical AF and could be grounds for hospital admission. Earlier studies suggested that AFR may have predictive value with regard to sinus rhythm maintenance after electrical cardioversion. 8 Recently, we reported that AFR less than 350 fpm assessed from ECG taken in the emergency department is an independent predictor of spontaneous conversion to sinus rhythm within 18 hours of symptoms onset. 9 To assess properties of AF that might be relevant for emergency department settings, arrhythmia duration in our study exceeded 1 hour in all 4 patients (median, 1290 minutes compared with 18 minutes in the study by Petrutiu et al). The differences in AFR dynamics at AF onset may reflect the difference in whether the atrial substrate is capable or incapable of maintaining long-lasting AF. Notably, we did not observe the abrupt drop of AFR preceding spontaneous AF termination that was reported by Petrutiu et al 10 based on data described earlier. 4 In a single patient who had spontaneous termination of AF recorded by ICM, AFR decrease was gradual over a period of 1 hour, from 280 to 260 fpm immediately before conversion. Similar
B
Friedmans’s P = .025
Fig. 2. Mean AFR per hour during the first 6 hours from AF onset. Gradual AFR acceleration is observed during the first 3 to 4 hours of AF.
726
P.G. Platonov et al. / Journal of Electrocardiology 45 (2012) 723–726
Fig. 3. A to D, Individual AFR trends during AF paroxysms presented as 10-minute averages starting from AF onset in all 4 patients. Case D represents AFR trend during 8.5-hour-long AF episode that had both onset (red line) and end (green line) captured on implantable loop recorder ECG. Note that maximal atrial rate during AF did not exceed 280 fpm and spontaneous conversion to sinus rhythm was preceded by gradual atrial rate decrease over the course of 1 hour.
to the initial AFR increase, the differences in the time course of AFR decrease before spontaneous conversion between our findings and earlier reported 10 data may be due to the much longer duration of AF episodes in our study. Moreover, the observation in the setting of AF termination is based on a single case in our study, which suggests that caution should be exercised when interpreting it. Temporal variability in AFR over the course of several hours at AF onset illustrates a potential problem in using AFR for risk stratification or prediction of intervention efficacy. Several earlier studies have shown that lower AFR values are associated with a better outcome of interventions for persistent AF, including electrical, pharmacologic cardioversion or catheter ablation. 8,11,12 In the acute setting of short-duration AF, the value of observed AFR may be dependent on its timing in regard to AF onset as illustrated by our study. However, the practical implication of our findings is that AFR recorded in excess of 3 to 4 hours after AF onset is likely to be of value and, as our group recently showed, can predict spontaneous termination of AF. 9 Despite the uniform pattern of AFR increase during the initial hours of long-lasting AF paroxysms observed in all 4 subjects selected using predefined criteria, the limited number of patients available for analysis pose an important limitation to our study, and caution should be exercised when interpreting these findings. Conclusion Atrial fibrillatory rate acceleration occurs during the initial 3 to 4 hours of AF, after which it reaches a plateau. Our findings should be considered whenever AFR is used as a predictor of clinical outcome early in the course of AF paroxysms because the timing of AFR readings may have data interpretation implications.
References 1. Goette A, Honeycutt C, Langberg JJ. Electrical remodeling in atrial fibrillation. Time course and mechanisms. Circulation 1996;94:2968. 2. Ravelli F, Mase M, Del Greco M, Faes L, Disertori M. Deterioration of organization in the first minutes of atrial fibrillation: a beat-to-beat analysis of cycle length and wave similarity. J Cardiovasc Electrophysiol 2007;18:60. 3. Bollmann A, Sonne K, Esperer HD, Toepffer I, Langberg JJ, Klein HU. Non-invasive assessment of fibrillatory activity in patients with paroxysmal and persistent atrial fibrillation using the Holter ECG. Cardiovasc Res 1999;44:60. 4. Petrutiu S, Sahakian AV, Swiryn S. Short-term dynamics in fibrillatory wave characteristics at the onset of paroxysmal atrial fibrillation in humans. J Electrocardiol 2007;40:155. 5. Stridh M, Sornmo L. Spatiotemporal QRST cancellation techniques for analysis of atrial fibrillation. IEEE Trans Biomed Eng 2001;48:105. 6. Stridh M, Sornmo L, Meurling CJ, Olsson SB. Sequential characterization of atrial tachyarrhythmias based on ECG time-frequency analysis. IEEE Trans Biomed Eng 2004;51:100. 7. Stridh M, Bollmann A, Olsson SB, Sornmo L. Detection and feature extraction of atrial tachyarrhythmias. A three stage method of timefrequency analysis. IEEE Eng Med Biol Mag 2006;25:31. 8. Holmqvist F, Stridh M, Waktare JE, Sornmo L, Olsson SB, Meurling CJ. Atrial fibrillatory rate and sinus rhythm maintenance in patients undergoing cardioversion of persistent atrial fibrillation. Eur Heart J 2006;27:2201. 9. Choudhary M, Holmqvist F, Carlson J, Nilsson H-J, Platonov PG. Low atrial fibrillatory rate is associated with spontaneous conversion of recent onset atrial fibrillation. Europace 2011;13:P524. 10. Petrutiu S, Sahakian AV, Swiryn S. Abrupt changes in fibrillatory wave characteristics at the termination of paroxysmal atrial fibrillation in humans. Europace : European pacing, arrhythmias, and cardiac electrophysiology : journal of the working groups on cardiac pacing, arrhythmias, and cardiac cellular electrophysiology of the European Society of Cardiology 2007;9:466–70. 11. Matsuo S, Lellouche N, Wright M, et al. Clinical predictors of termination and clinical outcome of catheter ablation for persistent atrial fibrillation. J Am Coll Cardiol 2009;54:788. 12. Bollmann A, Husser D, Stridh M, et al. Frequency measures obtained from the surface electrocardiogram in atrial fibrillation research and clinical decision-making. J Cardiovasc Electrophysiol 2003;14:S154.