Author + information
- Received April 20, 2016
- Revision received June 7, 2016
- Accepted June 8, 2016
- Published online September 13, 2016.
- Antoine H.G. Driessen, MDa,
- Wouter R. Berger, MDa,
- Sébastien P.J. Krul, MDb,
- Nicoline W.E. van den Berg, MDb,
- Jolien Neefs, MDb,
- Femke R. Piersma, RNb,
- Dean R.P.P. Chan Pin Yin, MDb,
- Jonas S.S.G. de Jong, MD, PhDc,
- WimJan P. van Boven, MD, PhDa and
- Joris R. de Groot, MD, PhDb,∗ ()
- aDepartment of Cardiothoracic Surgery, Heart Center, Academic Medical Center/University of Amsterdam, Amsterdam, the Netherlands
- bDepartment of Cardiology, Heart Center, Academic Medical Center/University of Amsterdam, Amsterdam, the Netherlands
- cDepartment of Cardiology, Onze Lieve Vrouwe Hospital, Amsterdam, the Netherlands
- ↵∗Reprint requests and correspondence:
Dr. Joris R. de Groot, Heart Center, Department of Cardiology, Academic Medical Center/University of Amsterdam, Meibergdreef 9, PO Box 22700, 1100 DD Amsterdam, the Netherlands.
Background Patients with long duration of atrial fibrillation (AF), enlarged atria, or failed catheter ablation have advanced AF and may require more extensive treatment than pulmonary vein isolation.
Objectives The aim of this study was to investigate the efficacy and safety of additional ganglion plexus (GP) ablation in patients undergoing thoracoscopic AF surgery.
Methods Patients with paroxysmal AF underwent pulmonary vein isolation. Patients with persistent AF also received additional lines (Dallas lesion set). Patients were randomized 1:1 to additional epicardial ablation of the 4 major GPs and Marshall’s ligament (GP group) or no extra ablation (control) and followed every 3 months for 1 year. After a 3-month blanking period, all antiarrhythmic drugs were discontinued.
Results Two hundred forty patients with a mean AF duration of 5.7 ± 5.1 years (59% persistent) were included. Mean procedure times were 185 ± 54 min and 168 ± 54 min (p = 0.015) in the GP (n = 117) and control groups (n = 123), respectively. GP ablation abated 100% of evoked vagal responses; these responses remained in 87% of control subjects. Major bleeding occurred in 9 patients (all in the GP group; p < 0.001); 8 patients were managed thoracoscopically, and 1 underwent sternotomy. Sinus node dysfunction occurred in 12 patients in the GP group and 4 control subjects (p = 0.038), and 6 pacemakers were implanted (all in the GP group; p = 0.013). After 1 year, 4 patients had died (all in the GP group, not procedure related; p = 0.055), and 9 were lost to follow-up. Freedom from AF recurrence in the GP and control groups was not statistically different whether patients had paroxysmal or persistent AF. At 1 year, 82% of patients were not taking antiarrhythmic drugs.
Conclusions GP ablation during thoracoscopic surgery for advanced AF has no detectable effect on AF recurrence but causes more major adverse events, major bleeding, sinus node dysfunction, and pacemaker implantation. (Atrial Fibrillation Ablation and Autonomic Modulation via Thoracoscopic Surgery [AFACT]; NCT01091389)
The most common arrhythmia, atrial fibrillation (AF) is associated with increased morbidity and mortality. Ablation is indicated for patients remaining symptomatic despite a trial with antiarrhythmic drugs (AADs) (1,2). Therefore, catheter ablation and stand-alone thoracoscopic surgery are increasingly being used. The arrhythmogenic trigger from the pulmonary veins (PVs) is the target for ablation in patients with paroxysmal AF without concomitant atrial or cardiac disease; the mechanism is less well established in patients with advanced AF, defined as persistent AF, enlarged left atria, or previously failed catheter ablation. Various treatment strategies have been advocated, combining more extensive myocardial ablation and ablation of non-PV and nonmyocardial targets, including stepwise catheter ablation approaches (3), in which PV isolation (PVI) is followed by linear left atrial (LA) ablation, ablation of continuous fractionated atrial electrograms (4), or ablation of rotors (5). As it has become clear that the autonomous nervous system plays a central role in initiating AF and in atrial autonomic remodeling (6,7), partial atrial denervation through ablation of the major autonomic ganglion plexus (GP), either alone or in combination with PVI, has been pursued (8,9).
GP stimulation promotes AF by a combined parasympathetic and sympathetic action resulting in action potential duration (APD) shortening and increased sarcoplasmic reticulum calcium release in PV myocardium, allowing early after-depolarizations to emerge and trigger AF (10). Aside from AF induction, GP stimulation affects local and global LA conduction time, consistent with a predominantly parasympathetic effect (11). Thus, the stimulation of the autonomic nerves within the GPs, beyond triggering AF, may also have a proarrhythmic effect on the atrial myocardium that perpetuates the arrhythmia (11).
Studies investigating the role of GP ablation in addition to PVI have demonstrated mixed results (8,12,13), as have nonrandomized studies during concomitant cardiac surgery (14,15). The GPs reside in epicardial fat pads and cannot be ablated endocardially without (much) more atrial myocardial ablation. This may induce post-ablation atrial tachycardias (ATs). However, more rigorous myocardial ablation around the PVs may also lower the chance of reconnection. Second, most studies have focused on patients with paroxysmal AF with few cardiovascular comorbidities. Epicardial ablation during thoracoscopic surgery for AF may allow more selective GP ablation without ablating the underlying atrial myocardium; however, only observational data on thoracoscopic GP ablation are available (16–19).
The aim of the prospective, randomized, controlled AFACT (Atrial Fibrillation Ablation and Autonomic Modulation via Thoracoscopic Surgery) study was to investigate epicardial GP ablation during thoracoscopic surgery for advanced AF. We hypothesized that GP ablation in these patient results in a higher percentage of freedom from AF, without inducing more periprocedural or late complications.
AFACT is a single-center study, performed at the Academic Medical Center in Amsterdam, that enrolled patients between April 2010 and January 2015. The study conformed to the Declaration of Helsinki and was approved by the Institutional Review Board. All patients provided written informed consent.
The inclusion and exclusion criteria are listed in the Online Appendix.
All patients had electrocardiographic documentation of AF and underwent the following pre-operative tests: nontriggered cardiac magnetic resonance imaging angiography for LA anatomy, 24-h Holter recording for AF burden and rate control assessment, transthoracic echocardiography for LA diameter and volume (determined using the Simpson method), spirometry for vital capacity and forced expiratory volume in 1 second to assess the ability to undergo perioperative single-lung ventilation, and treadmill testing to exclude clinically significant coronary artery disease (followed by coronary angiography when appropriate). Hepatic and renal failure, as well as clinically relevant anemia, were excluded. All patients were adequately anticoagulated with vitamin K antagonists or non–vitamin K oral anticoagulant agents (NOAC) ≥4 weeks prior to surgery.
Randomization was computer guided and performed in blocks with varying block sizes at the time of pericardial opening. With 110 patients in each arm (α = 0.8, 2-sided significance level = 0.05), AFACT was powered to detect a 17.5% difference in AF absence after 1 year, on the basis of previous studies (16,20,21). We enrolled 240 patients to allow for about 10% of patients not completing follow-up.
The surgical procedure has been described previously (16). In short, bilateral video-assisted thoracoscopy was performed. Oral anticoagulant agents were discontinued 2 days before surgery, and transesophageal echocardiography excluded LA thrombi before the atria were manipulated. All patients underwent PVI, specifically, ≥6 radiofrequency applications to the PV antrum, until a conductance drop within 10 s was observed (Isolator Synergy clamp, AtriCure, West Chester, Ohio). In patients with persistent AF, additional LA ablation lines were made, specifically, a Dallas lesion set involving a superior line connecting the right and left antral PVI and left fibrous trigone line, connecting the superior line to the left fibrous trigone at the aortic annular level (22) (Isolator Transpolar pen or Coolrail linear pen. AtriCure). All ablation lines were extensively tested for bidirectional block with epicardial electrodes connected to an electrophysiologic system in the operating theater, as reported earlier (16,23,24). The LA appendage was excluded using an endoscopic stapler. Patients stayed in the recovery room for 3 to 6 h and were admitted to the ward afterward. Thorax drains were removed within 24 h, and patients were usually discharged on postoperative day 3.
Before any PV or linear ablation, the anterior right (located in the epicardial fat pad anterior to the right superior and inferior PVs) and inferior right (located in the fat pad inferior to the right inferior PV, extending to the inferior side of the LA posterior wall) GPs were localized using anatomic landmarks and high-frequency stimulation (HFS). The GP between the superior caval vein and aorta was not ablated or tested. Online Figure 1 shows the location of the main GPs (posterior LA view). A positive HFS response was defined as ≥50% increase in the R-R interval. HFS was delivered using a bipolar ablation pen (Isolator Transpolar pen) with cycle length 60 ms, 16 Hz, pulse width 1.0 ms, and output incrementing from 1 to 25 mA.
In patients randomized to GP ablation, GPs were subsequently ablated. Notably, ablation on the basis of anatomic landmarks was performed when HFS did not evoke a vagal response. On the left side, the superior left and inferior left GPs (located in the fat pads on the LA roof, medial to the left superior PV, and inferior to the left inferior PV, extending toward the LA posterior wall) were identified similarly and ablated in these patients. The ligament of Marshall (between the pulmonary artery and the left superior PV) was subsequently dissected. In both groups, HFS of the 4 major GPs was repeated after all ablation was complete to confirm the absence or presence of a vagal response. Additional GP ablation was applied when necessary.
All patients were treated with colchicine 0.5 mg once daily from the first post-operative day for 30 days to prevent pericarditis. Follow-up at 10 days after discharge was for wound control. Clinical follow-up for endpoints was performed every 3 months with a clinical visit, electrocardiography, and 24-h Holter monitoring. Symptomatic patients were encouraged to obtain additional rhythm recording. The first 3 months formed a blanking period, during which recurrences of AF or other atrial arrhythmias were not considered recurrence (2). All AADs were discontinued after 3 months. When AF was present at the 3-month visit, electric cardioversion was performed. Anticoagulation was discontinued only in patients with CHA2DS2-VASc (congestive heart failure, hypertension, age ≥75 years, diabetes mellitus, prior stroke, transient ischemic attack, or thromboembolism, vascular disease, age 65–74 years, sex category [female]) scores of 0 (or 1 when based solely on sex) at 6 months (1).
The efficacy endpoint was freedom from AF or any atrial tachyarrhythmia lasting ≥30 s, documented on electrocardiography, Holter monitoring, or pacemaker or implantable cardioverter-defibrillator electrogram, without AAD use, as per the guidelines (2).
Safety endpoints included any procedure-related complication, stroke, and bleeding. Major complications were defined as those causing hospital admission within 30 days, inability to complete the procedure, or permanent injury or death. An independent clinical endpoint committee, whose members were unaware of study assignment, adjudicated all efficacy and safety endpoints.
Statistical analysis was performed with SPSS version 23.0 (IBM, Armonk, New York) and R version 3.2.1 for Windows (R Foundation for Statistical Computing, Vienna, Austria). Continuous values are expressed as mean ± SD. Categorical variables are expressed as numbers and percentages. The Mann-Whitney U test, paired Student t test, and Fisher exact test were used for comparisons. For freedom from AF recurrence, event-free survival was plotted and estimated by Kaplan-Meier curves. Clinical parameters associated with AF recurrence were studied using univariate and stepwise multivariate analysis in a Cox regression model. All variables with p values <0.10 in univariate analysis, plus 3 well-established risk factors for AF recurrence (left ventricular ejection fraction, AF duration, and previous PVI) were entered into the multivariate regression. A p value <0.05 indicated statistical significance.
Of 318 patients screened, 240 provided written informed consent. Patients were randomized 1:1 to standard thoracoscopic ablation (n = 123) or additional GP ablation (n = 117) (Figure 1). Patients were 60 ± 8 years of age, and 73% were men. Overall, 68% had enlarged left atria (LA volume index [LAVI] >33 ml/m2), and 43% had severely enlarged left atria (LAVI ≥40 ml/m2). AF was present for 4 years (interquartile range: 2 to 8 years; range: 1 to 35 years). Nearly one-quarter of patients had undergone previous catheter ablation; only 26 patients (11%) had normal left atria and no persistent AF or previous ablation. Baseline characteristics were balanced (Table 1).
Surgical procedure and complications
The procedure durations were 185 ± 54 min in the GP group and 168 ± 54 min in the control group (p = 0.015). The difference was driven by procedures in patients with paroxysmal AF (PVI alone: 144 ± 40 min vs. 127 ± 38 min; p = 0.04). There was no difference in procedure duration when additional LA lines were performed for persistent AF (GP 205 ± 49 min vs. control 202 ± 40 min; p = 0.609).
Isolation of all PVs was confirmed by demonstration of entry and exit block as previously described (23,24). Block across all LA lines was confirmed with differential pacing. After ablation, HFS-evoked vagal response was absent in 100% of patients in the GP group, whereas a residual vagal response could be provoked in at least 1 GP in 87% of control patients (p < 0.001).
Major bleeding occurred in 9 patients (LAVI 40.3 ± 11.6 ml/m2, which was not different from patients without bleeding; p = 0.833), all in the GP group (p < 0.001). Bleeding was managed thoracoscopically in 8 patients, resulting in termination of the procedure and thoracoscopic reoperation after several weeks in 4 patients, and in 3 patients, at least 1 of the LA lines could not be completed. Furthermore, 1 major bleed consisted of thoracoscopic access-port bleeding, necessitating a second thoracoscopic procedure. In 1 patient, a sternotomy was needed after trocar perforation of the right ventricle and the left anterior descending coronary artery, resulting in tamponade followed by sternotomy with suturing of the perforation and arterially grafting the left anterior descending coronary artery. Minor bleeds, all managed thoracoscopically without affecting the procedure, occurred in 12 patients, 6 in each group (p = 1.00).
Average hospital admission lasted 5.3 ± 2.1 days (range: 2 to 15 days) and 5.0 ± 1.8 days (range: 3 to 14 days) in the GP and control groups, respectively (p = 0.242). Symptomatic sinus node dysfunction, necessitating admission to the cardiac medium care unit, treatment with isoprenaline, and/or (temporary) pacing occurred in 12 GP patients and 4 control subjects (p = 0.038). Pacemakers were implanted in 3 patients for sinus node dysfunction with asystolic pauses during admission. Shortly after discharge, another 3 pacemakers were implanted because of sinus node dysfunction: syncope in 2 patients and postoperative total atrioventricular (AV) block in 1 patient. All 6 pacemakers were implanted in GP patients (p = 0.013). These patients had no pre-existing conduction disorders, apart from 1 with first-degree AV block (P-Q interval 210 ms). Major procedure-related complications occurred in 22 GP patients and 10 control subjects (p = 0.022) (Table 2).
Four patients died during follow-up, none procedure related and all in the GP group (p = 0.055) (Table 3). One-year follow-up was incomplete for 9 patients (4%), who were considered lost to follow-up. Complete information on the primary endpoint was ascertained in 227 patients (95%). AF recurrences during the blanking period were noted in 40 of 117 GP patients and 36 of 123 control patients (p = 0.407), with cardioversions performed in 25 and 28 patients, respectively.
At 1-year follow-up, no AF recurrences were observed in 70.9% and 68.4% of patients in the GP and control groups (p = 0.696) (Central Illustration) and 82% of patients were off AADs. Cardioversions were performed in 22 and 21 patients, respectively. There were no differences according to clinical AF type; AF was absent in 80.0% and 74.5% (p = 0.512) of patients with paroxysmal AF and 65.7% and 62.9% (p = 0.767) of those with persistent AF in the GP and control groups, respectively (Figure 2). In subgroup analysis, there were no differences between GP ablation and control (Figure 3).
AT was the most common recurrent arrhythmia and occurred significantly more often in patients in the GP group (78.1% AT and 21.9% AF) than in the control group (51.4% AT and 48.6% AF) (p = 0.026).
There were no differences in mean heart rate on Holter monitoring at baseline or 3-month follow-up. After discontinuation of AADs, mean heart rate on Holter monitoring increased by 5.5 ± 19.2 beats/min versus 2.6 ± 17.7 beats/min at 6 months (p = 0.013) and by 6.4 ± 19.2 beats/min versus 2.3 ± 16.1 beats/min at 9 months (p = 0.004) in the GP and control groups, respectively. At 1 year, mean heart rate did not differ from baseline in either groups, although office heart rate during follow-up was higher in the GP group than in control subjects during follow-up (data not shown).
Univariate and multivariate regression analysis was performed on AF recurrence (Figure 4). LAVI was the only determinant of AF recurrence (hazard ratio: 1.38 per 10-ml increase; 95% confidence interval: 1.03 to 1.83; p = 0.029) that remained significant in multivariate analysis.
The first randomized study of GP ablation in thoracoscopic surgery, AFACT is the largest study to date of minimally invasive AF surgery. Our study demonstrated that in patients with advanced AF, of whom 68% had enlarged LAs, GP ablation did not reduce AF recurrence (Central Illustration). GP ablation was associated with more major procedural complications (19% vs. 8%), in particular bleeds leading to termination of the procedure or sternotomy, and significantly more pacemaker implantations because of sinus node dysfunction and AV block. Additionally, recurrences were more often ATs in the GP group than in control subjects. There was no difference in AF recurrence rates during the so-called blanking period between groups.
The various hierarchical levels of extrinsic (nuclei and axonal fields in the brain), intrathoracic (spinal cord ganglia), and intrinsic (major GP and the atrial neural network, consisting of axons and smaller ganglia extending from these) autonomous nervous systems are interdependently connected but can function independently once disconnected. Injection of cholinergic agents into the GPs, electric stimulation of nerves, and pacing-induced AF produce proarrhythmic autonomic hyperactivity (10,25), leading to shortening of the atrial and PV APD (parasympathetic effect) and increases intracellular [Ca2+] (sympathetic effect), resulting in triggered firing and induction of AF (10). GP stimulation also directly affects atrial myocardial electrophysiology in a proarrhythmic manner, consistent with a predominantly parasympathetic action (11). Disconnecting extrinsic from intrinsic cardiac innervation resulted in shortened regional refractory periods in dogs and an increased burden of AF or AT starting after 4 to 5 weeks (26).
Thus, removing the inhibitory effect of the extrinsic on the intrinsic autonomous cardiac innervation causes proarrhythmic GP hyperactivity, which provides the rationale for GP ablation. Indeed, Katritsis et al. (8) demonstrated fewer recurrences in patients with paroxysmal AF randomized to catheter ablation of 4 major GP areas and Marshall’s ligament in addition to PVI compared with either treatment in isolation. An anatomic approach to GP ablation conferred less AF recurrence than an evoked vagal response–based localization (27). These were catheter ablation studies, and it cannot be discerned whether these effects were caused by GP ablation or by more atrial myocardial ablation, resulting in more rigorous PVI (8). Observational thoracoscopic studies showed that GP ablation combined with PVI conferred a high rate of freedom from AF (16). However, in advanced AF, when electric, structural, and autonomic remodeling has occurred, GP ablation would be expected to be ineffective (28). Many minimally invasive studies reported observational evidence on procedures where PVI (with or without LA lines) was standardly combined with ablation of the GPs (16–19), but a systematic review found no benefit of GP ablation in this setting, which may relate to the advanced nature of AF in these patients and autonomic remodeling (29). Indeed, Mao et al. (30) demonstrated that 8 weeks after GP ablation in dogs, APD was shorter than in sham-operated animals, and AF inducibility increased. In contrast, the acute effect of GP ablation results in APD prolongation and decreased AF inducibility but coincides with abundant reinnervation of the atrium, which may promote AF inducibility. Such reinnervation might have contributed to our findings, supported by the observation that increased heart rate (on Holter monitoring) in the GP group was no longer present after 1 year. Similarly, there is abundant clinical evidence for structural and autonomic remodeling and increased non-PV triggers in advanced and long-standing AF (6,7,31).
AFACT randomized patients with advanced AF, undergoing thoracoscopic surgery for AF. There was no difference in AF recurrence in the entire population or in subgroups with paroxysmal or persistent AF. It has been suggested that autonomic hyperactivity, occurring over the hierarchical gradient from the GP via the axonal field to the atrial neural network, is responsible for this (32). Instead, we observed more major bleeding in the GP ablation group. Whereas these bleeds may not be caused directly by GP ablation, more rigorous surgical dissection in the patients randomized to GP ablation and the longer procedure time might have promoted bleeding. Indeed, several bleeds occurred during positioning of the ablation clamp around the right PVs, just after dissection and ablation of the anterior right and inferior right GP. Patients with major bleeding did not have larger LAs. Eight of 9 major bleeds could be managed thoracoscopically but resulted in termination of the procedure and resumption after weeks or in an inability to subsequently complete the ablation lines.
Furthermore, GP ablation more often resulted in conduction disorders requiring prolonged monitoring or isoprenaline therapy as well as pacemaker implantation in 6 patients (vs. 0 control subjects). As the procedural endpoint of GP ablation is not defined, and the absence of HFS-evoked vagal response merely shows the interruption of the trajectory between GP and end organs (i.e., sinus node or AV node), it can be questioned whether denervation was complete in the GP ablation group or, conversely, whether there was no inadvertent GP ablation during PVI or ablation of additional lines in the control group. We found that a residual HFS-evoked vagal response of at least 1 GP was present in 87% of control patients compared with 0% in the GP ablation patients. Moreover, there was an increase in heart rate in the GP group, reaching statistical significance after 6 and 9 months, but not in the control group. Finally, sinus node dysfunction occurred significantly more frequently in the GP group, as did pacemaker implantation, indicating that the autonomic modulation was indeed different among groups. The observation that recurrences constituted AT (as opposed to AF) more frequently in the GP groups might point to more inadvertent myocardial ablation during GP targeting, despite an epicardial approach.
AF recurrence was absent in 76.8% and 64.4% of patients with paroxysmal and persistent AF, respectively, without the use of AADs after a single thoracoscopic procedure. These were slightly higher recurrence rates than in our earlier study in less affected patients and higher than those reported for full maze surgery with cardiopulmonary bypass and biatrial ablation (16,33). An earlier systematic review reported 69% freedom from AF (without the use of AADs) after minimally invasive surgery (29). AFACT’s single-procedure surgical ablation results compared favorably with catheter ablation in this group of patients (3).
The optimal lesion set in this setting remains a matter of debate and might have affected outcomes in AFACT. Similarly, progression of the underlying disease might have contributed to AF recurrence (34). These hypotheses were not tested systematically, because AFACT tested GP ablation on top of a systematic AF ablation protocol.
GP ablation on top of PVI and additional LA lines in patients with advanced AF did not reduce AF recurrence but resulted in more major procedural complications, in particular major bleeds and pacemaker implantations. Therefore, GP ablation should not be performed in these patients.
COMPETENCY IN PATIENT CARE AND PROCEDURAL SKILLS: In patients with advanced AF that is long-standing and associated with LA enlargement or unsuccessful catheter ablation, augmenting PVI with thoracoscopic GP ablation increased the risk for procedural complications, including bleeding and the need for pacemaker implantation, without reducing the incidence of recurrent of AF.
TRANSLATIONAL OUTLOOK: Further research is needed to determine if particular subgroups of patients with AF can be identified who would benefit from thoracoscopic GP ablation in conjunction with other therapeutic interventions.
The authors acknowledge Wim ter Smitte and Carel Kools for excellent technical assistance.
For inclusion and exclusion criteria and a supplemental figure, please see the online version of this article.
This study was funded in part by personal grants to Dr. de Groot from the Dutch Heart Foundation (2009T021) and NWO/ZonMW (106.146.310). Drs. Driessen and de Groot are consultants for AtriCure. Dr. de Groot received an unrestricted research grant from AtriCure. Dr. de Groot is a consultant for Daiichi Sankyo; and has received research funding from AtriCure and St. Jude Medical. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Driessen and Berger contributed equally to this work.
- Abbreviations and Acronyms
- antiarrhythmic drug
- atrial fibrillation
- action potential duration
- atrial tachycardia
- ganglion plexus
- high-frequency stimulation
- left atrial
- left atrial volume index
- pulmonary vein
- pulmonary vein isolation
- Received April 20, 2016.
- Revision received June 7, 2016.
- Accepted June 8, 2016.
- 2016 American College of Cardiology Foundation
- Camm A.J.,
- Kirchhof P.,
- Lip G.Y.H.,
- et al.
- Calkins H.,
- Kuck K.H.,
- Cappato R.,
- et al.
- Scherr D.,
- Khairy P.,
- Miyazaki S.,
- et al.
- Nademanee K.,
- McKenzie J.,
- Kosar E.,
- et al.
- Narayan S.M.,
- Krummen D.E.,
- Shivkumar K.,
- Clopton P.,
- Rappel W.-J.,
- Miller J.M.
- Katritsis D.G.,
- Pokushalov E.,
- Romanov A.,
- et al.
- Krul S.P.J.,
- Berger W.R.,
- Veldkamp M.W.,
- et al.
- Krul S.P.J.,
- Driessen A.H.G.,
- Van Boven W.J.,
- et al.
- Yilmaz A.,
- Geuzebroek G.S.C.,
- Van Putte B.P.,
- et al.
- Boersma L.V.A.,
- Castella M.,
- Van Boven W.J.,
- et al.
- Han F.T.,
- Kasirajan V.
- Beyer E.,
- Lee R.,
- Lam B.-K.
- De Groot J.R.,
- Driessen A.H.G.,
- Van Boven W.J.,
- et al.
- De Groot J.R.,
- Berger W.R.,
- Krul S.P.J.,
- et al.
- Mao J.,
- Yin X.,
- Zhang Y.,
- et al.
- Di Biase L.,
- Burkhardt J.D.,
- Mohanty P.,
- et al.
- Shen X.,
- Scherlag B.J.,
- He B.,
- et al.
- Weimar T.,
- Schena S.,
- Bailey M.S.,
- et al.
- Kottkamp H.