The 1+1 Trial
A Prospective Trial of a Dual- Versus a Single-Chamber Implantable Defibrillator in Patients With Slow Ventricular Tachycardias
Background— The tachycardia detection interval (TDI) in implantable cardioverter/defibrillators (ICDs) is conventionally programmed according to the slowest documented ventricular tachycardia (VT), with a safety margin of 30 to 60 ms. With this margin, VTs above the TDI may occur. However, longer TDIs are associated with an increased risk of inappropriate therapy. We hypothesized that patients with slow VTs (<200 bpm) may benefit from a long TDI and a dual-chamber detection algorithm compared with a conventionally programmed single-chamber ICD.
Methods and Results— Patients with VTs <200 bpm were implanted with a dual-chamber ICD that was randomly programmed to a dual-chamber algorithm and a TDI of ≥469 ms or to a single-chamber algorithm with a TDI 30 to 60 ms above the slowest documented VT cycle length and the enhancement criteria of cycle length variation and acceleration. The primary combined end point was the number of all inappropriate therapies, VTs above the TDI, and VTs with significant therapy delay (>2 minutes). After 6 months, a crossover analysis was performed. Total follow-up was 1 year. One hundred two patients were included in the study. The programmed TDI was 500±36 ms during the dual-chamber phase and 424±63 ms during the single-chamber phase. For the primary end point (inappropriate therapies, VTs above the TDI, or VTs with detection delay), a moderate superiority of the dual-chamber mode was found: Mann-Whitney estimator=0.6661; 95% CI, 0.5565 to 0.7758; P=0.0040.
Conclusions— Dual-chamber detection with a longer TDI improves VT detection and does not increase the rate of inappropriate therapies despite a considerable increase in tachycardia burden.
Received March 27, 2003; de novo received January 6, 2004; revision received March 30, 2004; accepted April 13, 2004.
The tachycardia detection interval (TDI) in implantable cardioverter/defibrillators (ICDs) is conventionally programmed according to the slowest documented ventricular tachycardia (VT), with a safety margin of 30 to 60 ms. This is associated with a significant risk of VTs above the TDI during follow-up.1 In 1 study, 47 of 659 patients with ICDs suffered 61 VTs above the TDI. Only 7 of these were asymptomatic, 54 (88.5%) caused admission to the hospital, 5 (8.2%) caused syncope, and 6 patients (9.8%) were resuscitated. This suggests a larger safety margin. However, a longer TDI is associated with an increased risk of inappropriate therapies.2–6
The implantation rate of dual-chamber ICDs has increased in recent years,7 partly as a result of dual-chamber pacing indications8 and partly because of the assumption that dual-chamber ICDs provide better detection facilities than single-chamber ICDs.9 However, clear evidence from prospective clinical trials is lacking or the evidence does not favor this view.10–12
The 1+1 Trial addressed the question of whether patients with monomorphic slow VTs and no primary indication for dual-chamber pacing benefit from a long TDI and a dual-chamber detection algorithm compared with a conventionally programmed single-chamber ICD. If the safety margin between spontaneous VTs and the TDI can be increased in a dual-chamber ICD without increasing the risk of inappropriate therapies, the implantation of a dual-chamber ICD would be justified merely for the sake of detection.
The target population included patients with spontaneous or inducible monomorphic VTs with a cycle length ≥300 ms. Patients with a class I or II indication for pacing,8 patients with chronic atrial fibrillation, and patients with slow polymorphic VTs were excluded from the trial.
All patients were implanted with a dual-chamber ICD (Defender III, IV, ELA Medical), which, in a randomized manner, was programmed either (1) to a single-chamber mode with a TDI according to the slowest documented VT with a safety margin of 30 to 60 ms and the discrimination criteria cycle length variation and acceleration or (2) to a dual-chamber mode (Parad, ELA Medical) with a TDI of ≥469 ms, irrespective of spontaneous VTs. If longer TDIs were required, the programming of a safety margin was left to the implanting physician (for the detection algorithms, see Figure 1). After 6 months a crossover was performed, ie, patients with dual-chamber detection were programmed to single-chamber detection and vice versa. Tachycardia storage was always based on a dual chamber to allow for equal evaluation of all tachycardias in both study modes.
The combined end point included all VTs above the TDI or with a significant therapy delay (>2 minutes) and all inappropriate therapies. Because both end points are inversely related in each algorithm, the combined end point was assumed to provide a balanced marker of the device performance. Adverse events, device-related complications, and changes in antiarrhythmic drug therapy were documented for the whole study population.
The sample size was based on the assumptions that the risk of VT above the TDI would be 5%,1 the risk of significant therapy would be 5%,13–17 and the risk of inappropriate therapies would be 20% to 30% during the first year.3 A total of 80 patients were supposed to provide a power of 80% to detect a relative risk improvement of the combined end point by 60% in the dual-chamber mode (α error=0.05). The dropout and study violation rate was assumed to be 20%. Because the risk of having 1 event could not simply be transformed into the number of events, the number of 100 patients was stipulated in an ad hoc manner.
All patients were implanted with a Defender III or IV according to current guidelines of ICD implantations.8 An electrophysiological study was performed in all patients. This included a programmed ventricular stimulation with the patient off antiarrhythmic drugs or on amiodarone. When antiarrhythmic drugs were changed, programmed ventricular stimulation was repeated to adjust the TDI.
During implantation, lead selection was left to the implanting physician. The implantation requirements were a sensing of >5 mV for the ventricular electrode and >1.8 mV for the atrial electrode; the pacing threshold had to be <1.5 V/0.5 ms. The device was tested intraoperatively with 2 effective shocks 10 J below the maximum device output. Programming of the ICD according to randomization took place at implantation. Pacing was programmed to backup pacing at a rate of 40 bpm in both modes.
Patients were monitored for primary and secondary outcome events from the implantation date. Follow-up visits were scheduled for every 3 months from the date of implantation, and patients were encouraged to schedule extra visits if frequent shocks, syncope, or unexplained palpitations occurred. At each visit, documentation of all tachycardias stored in the ICD, treated or not treated, especially VTs with detection delay and inappropriate therapies, was obtained. Documentation of VTs that led to hospital admission was obtained from the hospitals. Furthermore, complications due to ICD therapy such as lead dislodgment were documented. Programming adjustments and pharmacological therapy were left to the discretion of the physician performing the follow-up.
All tachycardias were validated in a blinded manner by a central validation committee consisting of 3 experts in ICD electrogram analysis. A decision had to be made for each tachycardia. VT below the detection rate was diagnosed if at least a single-lead ECG could be obtained that revealed the typical characteristics of VT (wide complex, VA dissociation) or if VT was documented in a monitor zone in those patients in whom a slow VT had been suspected on the basis of clinical history (eg, unexplained syncope, palpitations) and a monitor zone had been programmed.
Patients randomized to dual-chamber or single-chamber mode were compared in an intention-to-treat fashion. Analysis was performed with the Kaplan-Meier method, comparing risk of a VT above the TDI or with detection delay and risk of inappropriate therapies during the first and second study phases. This analysis examined the number of patients with an end point. The crossover design also allowed for a comparison of the number of VTs above the TDI, with detection delay or inappropriate therapies in the 2 programmed modes in each patient.18,19 Thus, imbalances between the randomization groups could be neglected. However, the second phase was expected to have fewer events than the first phase. This was tested before the crossover analysis for each end point. If the test revealed an imbalance between the first and second phases, ie, a carryover or period effect, the analysis was performed for each study phase separately. Because there is no tool for a crossover design to evaluate patients who take part in 1 study period but not in the other, eg, as a result of death, only patients who participated in both study phases were evaluated in the intention-to-treat crossover analysis. Receiver operating characteristic (ROC) plots were drawn to depict superiority or inferiority of the dual-chamber compared with the single-chamber mode. The area under the ROC curve is identical to the Mann-Whitney estimator, which is presented as a measure of relevance.
One hundred two patients were included in the trial, with 52 initially randomized to single-chamber and 50 to dual-chamber detection (Table 1). TDI was programmed to 422±70 ms in the first single-chamber detection phase and 496±27 ms in the first dual-chamber detection phase. During the second study phase, the TDI was increased to 435±59 ms in the single-chamber and 512±38 ms in the dual-chamber detection mode. The safety margin between the spontaneous VT before implantation and the TDI was 41 ms in the first and 81 ms in the second single-chamber phase, whereas the safety margin in both dual-chamber phases varied between 131 and 139 ms.
Mean follow-up was slightly longer in dual-chamber detection mode than in single-chamber detection mode (Table 2), which may favor single-chamber detection in the comparative analysis.
During period 1, all patients received the mode to which they were randomized. During period 2, 7 patients did not receive the intended mode. Three patients died during the first study period, and 2 patients were lost to follow-up. One patient had many slow VTs during the dual-chamber mode. This patient and another withdrew their informed consent. These patients were excluded from the intention-to-treat crossover analysis.
Fifteen patients revealed some time differences in the study phases that exceeded the 2 months allowed in the protocol. Four patients performed an early crossover: 3 to the dual-chamber mode and 1 to the single-chamber mode. The early crossover to the dual-chamber mode was due to many inappropriate therapies in 2 patients and due to slow VTs in 1 patient. One patient performed a crossover to the single-chamber mode because of atrial lead dysfunction, which he did not want to be revised. Six patients performed a late crossover to the dual-chamber mode without a special reason. The second study phase was too short in 5 patients because of death in 3. Two patients were lost to follow-up during the second phase. In summary, 4 patients had a shorter single-chamber study phase because of detection problems with the single-chamber mode: 2 due to many inappropriate therapies and 2 due to slow VTs.
Dual-chamber ICD implantation was successful in all patients. Atrial sensing was 2.49±1.04 mV, pacing threshold was 0.97±0.48 V at 0.44±0.06 ms, and the impedance was 625±149 Ω. Sensing was >5.0 mV in all patients, as requested in the protocol (mean, 6.32±2.09). Mean pacing threshold was 0.66±0.37 V at 0.44±0.06 ms, and mean impedance was 704±187 Ω. Twelve patients received a Defender III, and 90 patients received a Defender IV (ELA Medical). The lowest tested and successful defibrillation energy was 16.6±3.8 J. Mean operation time (from cut to close) was 94±41 minutes. Mean x-ray duration was 11±10 minutes.
Antiarrhythmic Drug Therapy
Antiarrhythmic drug therapy did not differ significantly between the 2 study populations at baseline, and only a few changes were made during both detection modes. Between 33% and 42% of patients received β-blockers in each study phase. Amiodarone was added to the antiarrhythmic drug therapy in 7 patients during the first single-chamber study phase, in contrast to only 1 patient in the dual-chamber mode. Therefore, 27% of patients were treated with amiodarone when they finished the single-chamber phase and started the dual-chamber phase as opposed to 20% of patients when they started the second single-chamber phase.
Tachycardias and Device Performance
Device counters revealed 27 260 detections during the single-chamber and 84 204 detections during the dual-chamber mode. The increase of the TDI by 76 ms increased the tachycardia burden >3-fold.
A total of 1969 tachycardias were counted as sustained and stored with electrograms and therapy data in the ICDs of 95 patients: 692 were VTs, and 1257 were supraventricular tachycardias (SVTs). Of these, 1027 were sinus tachycardias, 99 atrial tachycardias with n:1 conduction, 34 atrial tachycardias with 1:1 conduction, 97 atrial fibrillation, and 20 oversensing episodes (Figure 2). This reveals that sinus tachycardias outnumbered VTs above a cycle length of 460 ms. The opposite was true for tachycardias below a cycle length of 340 ms, in which VTs outnumbered supraventricular tachycardias by >15:1. In the range between 340 and 460 ms, VTs and SVTs reached a 1:1 ratio, and regular SVTs such as atrial tachycardias complicate detection.
Five hundred eight adequate device interventions occurred during single-chamber detection: 183 VTs were correctly terminated, and therapy was correctly withheld in 325 SVTs. During dual-chamber detection, 1244 correct device reactions occurred, 425 VTs were correctly detected and terminated, and therapy was correctly withheld in 819 SVTs.
Primary End Point
During single-chamber detection, 141 tachycardias in 44 patients were not diagnosed correctly. Seventy-eight of these were inappropriate therapies, 32 were VTs with detection delay in 4 patients, and 31 VTs were above the TDI in 10 patients. Eight patients with slow VTs were hospitalized at least once. One patient was resuscitated during a VT above the TDI. In 2 patients slow VTs were detected during routine follow-up.
During dual-chamber detection, 76 tachycardias in 24 patients were not treated correctly, and 62 were inappropriate therapies in 16 patients. Twelve detection delays occurred during VTs in 4 patients, and 4 VTs were above the TDI in 4 patients. In 2 patients slow VTs were diagnosed because they were borderline. In 2 patients VTs were documented as asymptomatic VTs during routine follow-up.
The risk of a VT above the TDI was 12% during the first 6 months in the single-chamber mode with a mean safety margin of 41 ms, and it remained as high as 9% with a safety margin of 81 ms during the second study phase. The risk of a VT above the TDI was 7% in the first and 0% in the second dual-chamber phase (Figure 3). A VT above the TDI could not be predicted by any baseline characteristics.
For the primary end point of this trial, which was the number of VTs above the TDI or significant therapy delay and the number of inappropriate therapies, a more than moderate superiority of the dual-chamber mode was found (Mann-Whitney estimator=0.6661; 95% CI, 0.5565 to 0.7758; P=0.0040; Table 2, Figure 4).
For the comparison of the number of VTs above the TDI plus detection delay, the crossover analysis revealed a moderate advantage for patients in the dual-chamber mode (Mann-Whitney estimator=0.6647; 95% CI, 0.5347 to 0.7946; P=0.0175; Table 2).
The results for inappropriate therapies varied considerably between the first and second study phases (Figure 3). The dual-chamber ICD was significantly better than the single-chamber ICD in the intention-to-treat crossover analysis (P=0.0079). However, there was a strong indication for a carryover effect (P=0.0472) and a period effect (P=0.0499). During study phase 1, there was a more than moderate superiority of the dual-chamber mode over the single-chamber mode (Mann-Whitney estimator=0.6628; 95% CI, 0.5642 to 0.7614; P=0.0023; Figure 4). The second phase did not reveal any significant differences (Table 2, Figure 3).
Sensitivity and Specificity
The sensitivity was 0.94 in the dual-chamber mode and 0.82 in the single-chamber mode (Mann-Whitney estimator=0.6524; 95% CI, 0.4798 to 0.8249; P=0.1345; Figure 4). The specificity was 0.81 in the single-chamber and 0.93 in the dual-chamber mode (Figure 3).
Secondary End Points: Adverse Events, Shocks, and Device-Related Complications
Six patients died during follow-up, and 1 patient survived an episode of sudden cardiac death due to a VT above the TDI. Five deaths were noncardiac in cause.
The number of shocks for any cause was not significantly different between the 2 studied modes, ie, despite the tremendously increased tachycardia burden, patients in the dual-chamber mode did not experience more shocks (adequate or inadequate) than patients in the single-chamber mode (Table 2).
Twenty device-related complications occurred in 19 patients. There were 5 atrial lead dislodgments, and 1 atrial lead had to be revised because of undersensing. Four ventricular leads dislodged, 2 showed undersensing, and 6 had a technical dysfunction or were recalled by the company because of an impending dysfunction (n=2). Lead complications varied considerably between different centers: 2 centers representing 40 implantations had a complication rate of 21.1% and 23.8%, respectively. The complication rate in the other centers was 12.9%.
Primary End Point Cluster
The combined end point of the 1+1 Trial was the number of inappropriate therapies and VTs above the TDI or with a considerable detection delay >2 minutes. For this primary end point, the 1+1 Trial demonstrated a significant, moderate superiority of the dual-chamber detection with a long TDI compared with conventional single-chamber detection. This finding is in contrast to the results of a trial in which patients were randomly assigned to a dual-chamber or single-chamber ICD, which did not reveal any benefit of the dual-chamber ICD in terms of tachycardia detection.10 However, patients included in this trial were not selected with respect to slow VTs. They had shorter TDIs (mean, 368 ms) than the patients in the 1+1 Trial, and the overlap of SVT and VT cycle length was possibly too small to demonstrate a benefit of the dual-chamber detection in these patients (Figure 2). In contrast, the 1+1 Trial deliberately selected patients who where supposed to have a large overlap between VT and SVT cycle length.
VTs Above the TDI
The risk of VTs above the TDI was 12% in the first and 9% in the second single-chamber study phase, with a mean safety margin of 41 and 81 ms, respectively. The risk was 7% in the first and 0% in the second dual-chamber study phase, with a safety margin of >130 ms (Figure 3). The risk of a VT above the TDI is higher than that reported in a retrospective analysis that revealed a risk of ≈3% per year.1 Patients on antiarrhythmic drugs with conduction-slowing properties had a higher risk in that study (25% for patients on amiodarone in the first year after ICD implantation).1 Significant VT-slowing properties have also been demonstrated for sotalol20,21 and class I antiarrhythmic drugs.22–24 When it is considered that patients with ICDs frequently receive antiarrhythmic drugs to suppress frequent VTs or control SVTs, this would strongly suggest a longer TDI in these patients, at least for some time after ICD implantation or after a change in antiarrhythmic drug therapy.
The increase of the TDI by 76 ms in the dual-chamber study phase increased the tachycardia burden >3-fold. With a specificity of 93%, the dual-chamber algorithm used in this trial prevented an increase of inappropriate therapies despite the increased tachycardia burden. For the number of inappropriate therapies, the dual-chamber mode proved to be superior to the single-chamber mode (P<0.008). However, this superiority was seen only in the first study phase, whereas in the second phase both modes revealed a similar risk of inappropriate therapies, which still implies superiority of the dual-chamber mode because of the larger tachycardia burden. However, this needs further interpretation. Obviously, patients initially randomized to the single-chamber mode posed a larger problem to the algorithm than patients who in the single-chamber mode group in the second study phase. The opposite was true in the dual-chamber mode. This implies that the actual risk of inappropriate therapies depends not only on the device and its specificity but also on the type of SVTs in a specific patient population.12
In summary, the increased specificity of the dual-chamber detection algorithm (0.93) was able to compensate for the tremendously increased tachycardia burden in patients with a very long TDI, and it prevented an increase of inappropriate therapies. Therefore, a lengthening of the TDI is feasible with the dual-chamber algorithm under investigation in patients at risk of slow VTs after ICD implantation or after a change in antiarrhythmic drug therapy.
Tachycardias in Patients With ICDs
In retrospective studies, atrial fibrillation with rapid ventricular response was the most common cause of inappropriate therapies (38%), followed by sinus tachycardias (29%) and oversensing (24%).3 The 1+1 Trial revealed a different distribution of inappropriate therapies. Atrial tachycardias were more frequent than atrial fibrillation. The load of atrial tachycardias with regular n:1 conduction may also suggest a strategy that includes a preventive radiofrequency ablation of atrial tachycardias in patients with a very long TDI.25 The tachycardia burden for monomorphic VTs was also high in this trial. When the favorable results of trials on VT ablation are taken into account,26,27 this may also call for a preventive approach in patients with monomorphic VTs at the time of ICD implantation.
The improved tachycardia detection in the dual-chamber mode was counterbalanced by an increased risk of lead complications. Twenty device-related complications occurred in 19 patients. The overall complication rate was slightly higher than that reported from other dual-chamber ICD trials (13% to 16%/7.1 to 9 months28,29). The number of atrial leads to be revised was close to the complication rate reported in other dual-chamber trials: 3.4% and 11% in 1.7 and 13.3 months, respectively.28–32 However, the ventricular lead complication rate was approximately twice as high as that reported from other trials.28–32 The lead complications varied considerably between different centers: 2 centers representing 40 implantations had a complication rate of 21.1% and 23.8%, respectively. The complication rate in the other centers was 12.9%. Therefore, the ventricular lead-related complication rate may not be representative. Thus far, the overall complication rate in dual-chamber ICDs demands either an improvement of implantation technique or a ventricular pace, dual chamber sense, dual chamber inhibition system for detection, as provided by some companies.33,34
The single-chamber algorithm under investigation was based on only 2 classic enhancement criteria, acceleration and stability. With these criteria, a specificity of ≈0.81 was reached. Current single-chamber ICDs may have better enhancement algorithms. This, however, needs to be demonstrated for the TDI range addressed in this trial.
VTs above the TDI are frequent in patients with monomorphic VTs and conventionally programmed single-chamber ICDs. Dual-chamber detection with a long TDI improves VT detection and does not increase the rate of inappropriate therapies or shocks despite a considerably increased tachycardia burden.
This study was supported by ELA Medical, Munich, Germany. The statistical model was provided and the final analysis performed by an independent statistician: Dr Volker W. Rahlfs, idv-Datenanalyse und Versuchsplanung, Munich, Germany. The authors thank the investigators and coinvestigators for their participation in the 1+1 Trial. At least 1 investigator from each center that contributed >2 patients is shown as coauthor. All others are named here: Rheinische Friedrich-Wilhelms-Universität Bonn: R. Schimpf, S. Herwig; Stiftsklinikum Augustinum: M. Block, A. Dorcewski; Ludwig-Maximilian-Universität München Klinikum Groβhadern: E. Hoffmann, U. Dorwarth; Herzzentrum Coswig: T. Markert. We especially thank F. Steffgen for his support of the trial and Dr Grönefeld for his contribution to data interpretation and manuscript revision.
Frank Steffgen and Michael Piel were employees of ELA Medical when the trial was started. Frank Steffgen is now working for Medtronic, and Michael Piel is working for Barth EP Systems.
Rosenqvist M, Beyer T, Block M, et al, for the European 7219 Jewel ICD investigators. Adverse events with transvenous implantable cardioverter-defibrillators: a prospective multicenter study. Circulation. 1897; 98: 663–670.
Pinski SL, Fahy GJ. The proarrhythmic potential of implantable cardioverter-defibrillators. Circulation. 1995; 92: 1651–1664.
Gregoratos G, Cheitlin MD, Conill A, et al. ACC/AHA guidelines for implantation of cardiac pacemakers and antiarrhythmia devices: executive summary. Circulation. 1998; 97: 1325–1335.
Deisenhofer I, Kolb C, Ndrepepa G, 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: 134–142.
Wilkoff BL, Kuhlkamp V, Volosin K, et al. Critical analysis of dual-chamber implantable cardioverter-defibrillator arrhythmia detection: results and technical considerations. Circulation. 2001; 103: 381–386.
Garcia Alberola A, Yli Mayry S, Block M, et al. RR interval variability in irregular monomorphic ventricular tachycardia and atrial fibrillation. Circulation. 1996; 93: 295–300.
Zimmermann H, Rahlfs VW. Model building and testing for the change-over design. Biometrics J. 1980; 22: 197–210.
Strickberger SA, Man KC, Daoud EG, et al. A prospective evaluation of catheter ablation of ventricular tachycardia as adjuvant therapy in patients with coronary artery disease and an implantable cardioverter-defibrillator. Circulation. 1997; 96: 1525–1531.
Gonska BD, Cao K, Schaumann A, et al. Catheter ablation of ventricular tachycardia in 136 patients with coronary artery disease: results and long-term follow-up. J Am Coll Cardiol. 1994; 24: 6–14.
Kuehlkamp V, Volosin KJ, Huegl BJ, et al. Worldwide clinical experience with a new dual-chamber implantable cardioverter defibrillator (ICD). Circulation. 1998; 991: 1998.