Europace Advance Access published online on June 1, 2007
Europace, doi:10.1093/europace/eum106
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Atrial fibrillation reduces the atrial impedance amplitude during cardiac cycle: a novel detection algorithm to improve recognition of atrial fibrillation in pacemaker patients
1 Universitätsklinikum Freiburg, Medizinische Klinik III, Department of Cardiology and Angiology, Hugstetter Str. 55, D-79106 Freiburg, Germany; 2 Biotronik GmbH & Co. KG, Woermannkehre 1, D-12359 Berlin, Germany
Manuscript submitted 1 February 2007. Accepted after revision 24 March 2007.
* Corresponding author. E-mail address: Thomas.faber{at}uniklinik-freiburg.de
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
Aims In carriers of dual chamber pacemakers and implantable cardioverter-defibrillators (ICD), detection of atrial fibrillation (AF) is crucial for adequate mode switch function and to avoid inappropriate shock delivery. Detection algorithms rely on the atrial rate and on the relationship of atrial to ventricular intracardiac electrograms, but the relative portion of misclassified AF episodes remains high. Although myocardial impedance is a reliable indicator of contraction, little is known about atrial impedance as a marker of atrial arrhythmias.
Methods During an electrophysiological study, we investigated the effect of induced AF on impedance at the right atrial free wall (RAFW) and right atrial appendage (RAA) in 20 patients. Using biphasic square-wave pulses (128 Hz, 200 µA/15 µs), impedance changes were recorded during sinus rhythm (SR-1), atrial pacing at 120 beats/min, AF induced by rapid atrial burst pacing, and after spontaneous AF termination (SR-2).
Results At the RAA, peak-to-peak impedance amplitude during cardiac cycle (
Z) dropped from 51.7 ± 35.3 
(SR-1) or 49.6 ± 30.6 (pacing) to 24.6 ± 22.0 (AF, P 
0.0005), and subsequently increased to 37.7 ± 24.7 (SR-2, P 0.0004 v. AF). At the RAFW, Z changed from 16.2 ± 15.5 (SR-1) or 13.5 ± 9.9 (pacing) to 5.9 ± 4.1 (AF, P 0.003), and to 11.4 ± 10.7 (SR-2, P 0.015). Given a discrimination threshold of 65%, the sensitivity and the specificity of Z to detect AF were 79 ± 18 and 89 ± 14%, respectively (95% confidence interval).
Conclusion AF causes Z drop in pacemaker and ICD recipients. This impedance based algorithm can be used as an alternative method of AF detection.
Key Words: Atrial fibrillation, Myocardial impedance, Mode switch, Cardiac pacing
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
Atrial fibrillation (AF) is the most common arrhythmia in the general population and is even more common in carriers of pacing devices.1
,2 In dual chamber pacemakers and implantable cardioverter-defibrillators (ICD), discrimination of the atrial rhythm is crucial for adequate mode switch function and to avoid inappropriate shock delivery. In addition, the diagnosis of asymptomatic AF episodes is warranted in order to guide antiarrhythmic therapy and appropriate anticoagulation.
Current detection algorithms rely on the atrial rate and on the relationship of atrial to ventricular intracardiac electrograms (IEGM). However, the detection of atrial IEGM can be affected by under- and oversensing. In prospective randomized trials, the relative portion of misclassified AF episodes is approximately 26%.3
Intracardiac electrical impedance can be used to assess volume changes around the electrode in the heart.4 Ventricular myocardial impedance varies with respiration as well as with contraction and is a reliable marker of cardiac inotropy.5,6 Less is known about the variation of atrial myocardial impedance during the cardiac cycle or during AF. We investigated for the first time the effect of acutely induced AF on atrial impedance.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
Patients
Fourteen men and six women with coronary artery disease and a normal to moderately decreased left ventricular function according to transthoracic echocardiography were prospectively studied during an electrophysiological evaluation. Patient characteristics including left atrial size are presented in Table 1. Four patients had interventricular conduction delay (left bundle branch block in one patient and right bundle branch block in three patients). All patients gave written informed consent before inclusion into the study. The study protocol was approved by the Ethics Committee of the Albert-Ludwigs-University of Freiburg (study number: 105/03).
|
Catheter positioning
The study was performed in the conscious, fasting state. The vital parameters were continuously monitored. Two catheters were inserted via the femoral vein and advanced to the heart using fluoroscopic guidance. A steerable multipolar electrophysiological catheter (ViaCath, Biotronik, Germany) was positioned in the right atrial appendage (RAA), and a bipolar electrophysiological catheter with a fixed screw (VascoStim Screw, Vascomed, Germany) was attached onto the right atrial free wall (RAFW). The ECG and IEGM were recorded by a conventional electrophysiological system (EP Lab, Quinton, USA). A stimulation unit was connected (UHS 20, Biotronik).
Impedance measurements
Impedance was measured via a commercially available pacemaker (Inos 2+, Biotronik) linked to a programmer (TMS 1000, Biotronik) and a personal computer. During impedance measurement, the respective electrophysiological catheters were connected to the atrial pacemaker port. The impedance signal was continuously registered at both locations (RAA and RAFW) for 60 s during sinus rhythm (SR-1), atrial stimulation at a rate of 120 beats/min (bpm), after AF induction by high-rate atrial burst pacing, and after spontaneous termination of AF (SR-2).
Impedance measurements were performed in the unipolar configuration, with a reference patch attached to the patient's chest serving as the ‘pacemaker case’. The impedance signal and lead II of the surface ECG were continuously registered and stored in the personal computer.
For impedance measurements, biphasic square-wave current pulses of 200 µA in amplitude and 15 µs in duration were injected by the pacemaker through the tip of the electrode at a rate of 128 Hz. The resolution time was thus 8 ms. The resulting voltage was measured by synchronous demodulation. A band pass filter of 0.1–10 Hz was applied in order to obtain the time varying component of the voltage. Ohm's law was applied to continuously calculate and transmit the time varying component of the impedance to the programmer and the personal computer. In the following, 
will be used as a symbol for the time-varying component of the impedance signal.
The difference between maximum and minimum impedance (peak-to-peak amplitude) is denoted as Z.
Data analysis
The stored curves were analysed offline using commercially available software (MATLAB, The MathWorks, Inc., USA). In a first step, the R-waves on the surface ECG were detected and marked. Premature beats were eliminated using rate and morphology criteria. The curve was divided into heart cycle intervals by triggering on the R-wave of the surface ECG. The curve intervals were analysed individually (beat to beat) as well as after averaging over 20 s. The peak-to-peak atrial impedance amplitude during cardiac cycle (Z) was calculated for the raw data and the averaged signals. Figure 1 illustrates averaging of curves and determination of Z.
|
Calculation of sensitivity and specificity
Sensitivity and specificity of AF detection via atrial impedance was plotted for discrimination thresholds in the range from 5 to 100% in steps of 5%, in a so-called receiver operator characteristic (ROC) curve. For instance, the sensitivity for a discrimination threshold of 70% was determined as the number of patients in whom
Z during AF was <70% of the value for Z during SR-1, divided by the total number of patients. On the basis of the ROC plot, an optimal discrimination threshold was determined.
Statistical analysis
Study results are expressed as mean ± SD unless stated otherwise. Paired variables were compared using the two-sided Student's t-test. P < 0.05 was considered statistically significant. The sensitivity and specificity values for the optimal discrimination threshold according to ROC were calculated as mean and confidence interval of 95%.
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
Figure 2 shows an example of original and averaged
curves during different stages of the electrophysiological study. The calculation of Z after restoration of sinus rhythm (SR-2) was feasible in 16 patients in whom the induced AF terminated spontaneously.
|
In the RAA, the
Z was 55.7 ± 34.9 during SR-1 and 54.2 ± 31.4 during atrial pacing. Z dropped to 37.0 ± 25.6 in all patients following AF induction (–34%, P 0.007; Table 2). Thereafter, Z increased to 43.0 ± 25.3 in all 16 patients with spontaneous AF termination (P 0.003 vs. AF). The observed impedance changes occurred immediately after the change in heart rhythm and remained stable thereafter. The Z results after signal averaging are shown in Figure 3. Because of averaging, the observed impedance variation patterns became more pronounced, with Z having changed from 51.7 ± 35.3 during SR-1 and 49.6 ± 30.6 during atrial pacing to 24.6 ± 22.0 (–53%) after AF induction (P 0.0005). The restoration of SR-2 was associated with Z rise to 37.7 ± 24.7 (P 0.00004 vs. AF).
|
|
The mean
Z measured via the screw electrode at RAFW (20.3 ± 15.7 ) during SR-1 was significantly lower than that for the RAA (P 0.00022; Table 2). Atrial pacing did not influence Z at RAFW (17.4 ± 9.3 ). However, a significant drop to 14.8 ± 10.9 was observed after AF induction (P 0.03). Restoration of SR-2 resulted in Z increase to 16.0 ± 10.4 (non-significant P-value vs. AF). After signal averaging over 20 s, the effects were more pronounced. Z changed from 16.2 ± 15.5 during SR-1 and 13.5 ± 9.9 during atrial pacing to 5.9 ± 4.1 after induction of AF (P 0.003). Restoration of SR-2 resulted in Z rise to 11.4 ± 10.7 (P 0.015 vs. AF). The calculated sensitivity and specificity values for AF detection via atrial impedance are shown in Figure 4 (note that the abscissa is scaled in 100%—specificity). The value of 65% was regarded as an optimal discrimination threshold, yielding the sensitivity of 79 ± 18% (using 95% confidence interval) for the RAA position and 74 ± 20% for the RAFW position. The specificity after spontaneous AF termination was 79 ± 14% (RAA) and 89 ± 11% (RAFW).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
To improve the current detection algorithms for AF in carriers of pacing devices, this prospective trial was designed to investigate impedance as a marker of new onset AF. The following major results were obtained: (i) impedance was measured for the first time in the right atrium in humans during SR and during AF. (ii) AF causes a significant drop in
Z, which returns towards baseline values after restoration of SR-2. (iii) Assuming a discrimination threshold of 65%, sensitivity and specificity values of Z to detect AF were very satisfactory.
Mechanism of impedance
Changes in the unipolar right ventricular impedance during the cardiac cycle are related to the changing content of blood (low impedance) and tissue (high impedance) around the tip of the pacing electrode.6 The curve morphology depends on right ventricular contractility, and thus on the inotropic state of the heart. It is likely that the observed changes in Z during AF follow the same principle and are linked to atrial contraction. A loss of atrial contraction during AF leads to a decrease in atrial wall motion around the electrode, and thus decreasing Z. Different mechanisms such as atrial compliance and pressure which are both increased during AF might also contribute to this phenomenon.7 Further investigation is needed to identify the determinants of impedance.
Detection of AF in pacing devices
A reliable detection of AF in pacemaker and ICD recipients is warranted in order to guide appropriate antiarrhythmic therapy and anticoagulation as well as to avoid inappropriate therapies. In multiple prospective, randomized trials, the rate of inappropriate therapies could not be reduced by atrial detection algorithms compared with single chamber devices.3,8 The rate of misclassified AF episodes was still 26%.
The major drawback of detection algorithms using atrial rate criteria is their vulnerability to undersensing. During AF, the atrial IEGM amplitude may decrease by up to 82%, and become lower than the programmed sensing threshold.9 Even after optimized programming, 14–43% of atrial tachyarrhythmia episodes were not adequately detected in former studies.10,11
Patients' symptoms are not a reliable surrogate for the prevalence of AF episodes, because approximately 50% are asymptomatic.12 The rate of asymptomatic AF episodes is even increased with certain antiarrhythmic drugs or after catheter ablation.13 Therefore, the reliable identification of AF in pacemaker/ICD carriers is a useful tool to guide therapeutic decisions.
A diagnostic algorithm for AF detection should ideally be fast, robust, and reliable. Changes in Z immediately followed changes in heart rhythm, but remained constant during the specific rhythm. A fast detection of AF onset can be provided. Z shows a pronounced drop of >50% during AF. As Z is determined by mechanical processes, it provides additional independent information from electrical parameters (IEGM). Hence, it does not interfere with reduced IEGM amplitudes during AF and may be considered robust. The decrease in Z after AF onset was observed in most patients (19/20 at RAFW and 18/20 at RAA) and may thus be considered reliable.
Given a discrimination threshold of 65% for AF detection, 79% of all AF episodes were correctly diagnosed with a specificity of 89%. This is comparable with the currently available detection algorithms.14
Study limitations
The impedance measurements were performed at RAFW and in the RAA. Therefore, no information on impedance changes at the right atrial septum was provided. Since most atrial pacing leads are implanted either at the RAFW or in the RAA, the results are applicable to the majority of pacemaker/ICD recipients.
In this acute study, impedance measurements were conducted with electrophysiological catheters. In the chronic state, impedance changes detected with chronically implanted leads might be less stable and less prominent.
The fact that AF episodes were induced by rapid pacing does not reflect the clinical situation. However, paroxysmal AF is usually induced by rapidly firing triggers in the pulmonary veins, which resembles rapid pacing.15 Secondly, pacing at 120 bpm did not affect Z. An influence of short-term rapid burst pacing on changes in impedance is therefore assumed to be low.
All impedance measurements were performed in patients without known history of AF. Atrial fibrosis and remodelling, which can occur in patients with AF of significant duration and even in case of paroxysmal AF, may alter the measured impedance changes.
The duration of the analysed AF episodes was <15 min. In these cases, Z returned to baseline values immediately after restoration of SR. It is unknown whether the termination of long lasting AF episodes followed by atrial stunning may also be accompanied by an immediate rise in Z. Since changes in Z most likely depend on atrial contraction, Z rise may be retarded.
The measurements were performed in a unipolar fashion. Nowadays, most atrial leads are programmed to a bipolar sensing mode. Using advanced software features, the device should be able to detect atrial IEGM in a bipolar mode and measure in a unipolar fashion simultaneously.
It would be of interest to compare different AF detection algorithms (impedance vs. rate) in a randomized trial. Since the atrial IEGM was not continuously registered, this analysis was not performed.
AF was the only arrhythmia induced during the study. No conclusion can thus be made with respect to organized tachyarrhythmias such as atrial flutter or atrial tachycardias.
|
Conclusion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
After induction of AF in humans, the
Z drops by >50% and may serve as an alternative detection tool to diagnose AF in pacemaker and ICD recipients. |
Acknowledgements |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
The study was sponsored by a research grant of Biotronik GmbH, Germany. We thank Dr M. Lippert for his help and technical support during preparation of the manuscript.
Conflict of interest: T.S. F. received a research grant from Biotronik GmbH, Germany (see Acknowledgements). O. S. is an employee (clinical research scientist) of Biotronik GmbH. All other authors did not receive any grants or honoraria.
|
Footnotes |
|---|
B.S. and S.A. contributed equally to the study and the manuscript. 
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionConclusionAcknowledgementsReferences |
|---|
[1] Go AS, Hylek EM, Phillips KA, Chang Y, Henault LE, Selby JV, et al. Prevalence of diagnosed atrial fibrillation in adults: national implications for rhythm management and stroke prevention: the AnTicoagulation and Risk Factors in Atrial Fibrillation (ATRIA) Study. JAMA (2001) 285:2370–5.
[2] Lamas GA, Lee KL, Sweeney MO, Silverman R, Leon A, Yee R, et al. Ventricular pacing or dual-chamber pacing for sinus-node dysfunction. N Engl J Med (2002) 346:1854–62.
[3] Theuns DA, Klootwijk AP, Goedhart DM, Jordaens LJ. Prevention of inappropriate therapy in implantable cardioverter-defibrillators: results of a prospective, randomized study of tachyarrhythmia detection algorithms. J Am Coll Cardiol (2004) 44:2362–67.
[4] Rushmer RF, Crystal DK, Wagner C, Ellis RM. Intracardiac impedance plethysmography. Am J Physiol (1953) 174:171–4.
[5] Irnich W. Impedance conductance resistance: definitions: measurement principles. In: Thoracic Impedance Measurements in Clinical Cardiology—Winter UJ, Klocke RK, Kubicek WG, Niederlag W, eds. (1994) Stuttgart: Georg Thieme. 4–10.
[6] Schaldach M, Hutten H. Intracardiac impedance to determine sympathetic activity in rate responsive pacing. Pacing Clin Electrophysiol (1992) 15:1778–86.[CrossRef][Medline]
[7] Leistad E, Christensen G, Ilebekk A. Effects of atrial fibrillation on left and right atrial dimensions, pressures, and compliances. Am J Physiol (1993) 264:H1093–H1097.[Web of Science][Medline]
[8] Deisenhofer I, Kolb C, Ndrepepa G, Schreieck J, Karch M, Schmieder S, et al. Do current dual chamber cardioverter defibrillators have advantages over conventional single chamber cardioverter defibrillators in reducing inappropriate therapies? A randomized, prospective study. J Cardiovasc Electrophysiol (2001) 12:134–42.[CrossRef][Web of Science][Medline]
[9] Neuzner J, Sperzel J, Pitschner HF, Schwarz T, Ehrlich W, Schulte B, et al. Bipolar atrial sensing thresholds in sinus rhythm and atrial tachyarrhythmias. A comparative analysis in patients with DDDR pacemakers. Europace (1999) 1:135–9.
[10] Palma EC, Kedarnath V, Vankawalla V, Andrews CA, Hanson S, Furman S, et al. Effect of varying atrial sensitivity, AV interval, and detection algorithm on automatic mode switching. Pacing Clin Electrophysiol (1996) 19:1735–9.[Medline]
[11] Leung SK, Lau CP, Lam CT, Tse HF, Tang MO, Chung F, et al. Programmed atrial sensitivity: a critical determinant in atrial fibrillation detection and optimal automatic mode switching. Pacing Clin Electrophysiol (1998) 21:2214–9.[CrossRef][Medline]
[12] Patten M, Maas R, Karim A, Muller HW, Simonovsky R, Meinertz T. Event-recorder monitoring in the diagnosis of atrial fibrillation in symptomatic patients: subanalysis of the SOPAT trial. J Cardiovasc Electrophysiol (2006) 17:1216–20.[CrossRef][Web of Science][Medline]
[13] Hindricks G, Piorkowski C, Tanner H, Kobza R, Gerds-Li JH, Carbucicchio C, et al. Perception of atrial fibrillation before and after radiofrequency catheter ablation: relevance of asymptomatic arrhythmia recurrence. Circulation (2005) 112:307–13.
[14] Wood MA, Ellenbogen KA, Dinsmoor D, Hess M, Markowitz T. Influence of autothreshold sensing and sinus rate on mode switching algorithm behavior. Pacing Clin Electrophysiol (2000) 23:1473–8.[CrossRef][Medline]
[15] Haissaguerre M, Jais P, Shah DC, Takahashi A, Hocini M, Quiniou G, et al. Spontaneous initiation of atrial fibrillation by ectopic beats originating in the pulmonary veins. N Engl J Med (1998) 339:659–66.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  S. Tzeis, G. Andrikopoulos, C. Kolb, and P. E. Vardas Tools and strategies for the reduction of inappropriate implantable cardioverter defibrillator shocks Europace, August 15, 2008; (2008) eun205v1. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||