Author + information
- Received January 23, 2011
- Revision received March 21, 2011
- Accepted April 27, 2011
- Published online August 9, 2011.
- Claudia Herrera Siklódy, MD⁎,⁎ (, )
- Thomas Deneke, MD‡,
- Mélèze Hocini, MD§,
- Heiko Lehrmann, MD⁎,
- Dong-In Shin, MD‡,
- Shinsuke Miyazaki, MD§,
- Susanne Henschke, MD†,
- Peter Fluegel, MD†,
- Jochen Schiebeling-Römer, MD⁎,
- Paul M. Bansmann, MD‡,
- Thomas Bourdias, MD§,
- Vincent Dousset, MD§,
- Michel Haïssaguerre, MD§ and
- Thomas Arentz, MD⁎
- ↵⁎Reprint requests and correspondence:
Dr. Claudia Herrera Siklódy, Electrophysiology Department, Herz-Zentrum, Südring 15, D-79189 Bad Krozingen, Germany
Objectives We compared the safety of different devices by screening for subclinical intracranial embolic events after pulmonary vein isolation with either conventional irrigated radiofrequency (RF) or cryoballoon or multielectrode phased RF pulmonary vein ablation catheter (PVAC).
Background New devices specifically designed to facilitate pulmonary vein isolation procedures have recently been introduced.
Methods This prospective, observational, multicenter study included patients with symptomatic atrial fibrillation referred for pulmonary vein isolation. Ablation was performed using 1 of the 3 catheters. Strict periprocedural anticoagulation, with intravenous heparin during ablation to achieve an activated clotting time >300 s, was ensured in all patients. Cerebral magnetic resonance imaging was performed before and after ablation.
Results Seventy-four patients were included in the study: 27 in the irrigated RF group, 23 in the cryoballoon group, and 24 in the PVAC group. Total procedure times were 198 ± 50 min, 174 ± 35 min, and 124 ± 32 min, respectively (p < 0.001 for PVAC vs. irrigated RF and cryoballoon). Findings on neurological examination were normal in all patients before and after ablation. Post-procedure magnetic resonance imaging detected a single new embolic lesion in 2 of 27 patients in the irrigated RF group (7.4%) and in 1 of 23 in the cryoballoon group (4.3%). However, in the PVAC group 9 of 24 patients (37.5%) demonstrated 2.7 ± 1.3 new lesions each (p = 0.003 for the presence of new embolic events among the 3 groups).
Conclusions The PVAC is associated with a significantly higher incidence of subclinical intracranial embolic events. Further study of the causes and significance of these emboli is required to determine the safety of the PVAC.
Left atrial catheter ablation is an increasingly recommended treatment for atrial fibrillation (AF). Although many different techniques have been described, pulmonary vein isolation (PVI) remains a basic and central endpoint (1,2). However, PVI performed with conventional radiofrequency (RF) (externally irrigated RF) ablation catheters remains a technically demanding technique, with a relatively long learning curve.
To facilitate these procedures, new devices have been developed to allow faster and easier PVI. Two of these devices have been approved for routine clinical practice in Europe: the cryoballoon (3) and the multielectrode duty-cycled RF ablation catheter (4).
However, published data on these new devices mainly focus on its efficacy. Safety data are scarce. In particular, although procedure-related strokes or transient ischemic attacks occur in approximately 1% of patients undergoing conventional RF ablation (5), recent evidence suggests a higher rate of subclinical cerebral emboli detected by imaging (6). We aimed to compare the safety profile of these devices by assessing the incidence of new subclinical embolic lesions by cerebral magnetic resonance imaging (MRI) before and after PVI.
We prospectively included 74 patients (68% men, age 61 ± 9 years) with highly symptomatic, drug-resistant paroxysmal (62%) or persistent AF referred for catheter ablation to 3 high-volume centers participating in the study. Exclusion criteria were long-standing persistent AF, the need for further left atrial ablation other than PVI, and any contraindication to undergoing MRI.
The study protocol was approved by local ethics committee/boards, and all patients signed an informed consent.
Catheter ablation procedures
All patients underwent PVI alone with: 1) a conventional externally irrigated RF catheter; 2) an Arctic Front cryoballoon (Medtronic, Minneapolis, Minnesota); or 3) a multielectrode duty-cycled RF pulmonary vein ablation catheter (PVAC) (Medtronic Ablation Frontiers, Carlsbad, California). All ablations were performed by operators with sufficient expertise in the respective techniques. Specifically, RF ablation procedures were performed in all 3 participating centers, cryoballoon ablations in Bad Krozingen, Germany, and PVAC procedures in Bordeaux, France, and Köln, Germany. The procedures were performed with the patients under either general anesthesia or conscious sedation, according to the usual practice at each center.
Conventional RF PVI
Externally irrigated RF PVI consisted of a large circumferential encircling of both ipsilateral pulmonary veins (PVs). The technique is described in detail elsewhere (7). In short, vascular access was obtained through the right femoral vein, and a quadripolar catheter (Xtreme, Sorin SPA, Milan, Italy) was positioned in the coronary sinus. Double transseptal access was obtained and a 3-dimensional geometry of the LA was reconstructed with a 3-dimensional electroanatomic mapping system (EnSite Velocity, St. Jude Medical, St. Paul, Minnesota, or Carto 3, Biosense Webster, Diamond Bar, California). The PV ostia were carefully localized by combining the electrical signal information with selective PV angiography. RF energy was applied using an externally irrigated-tip catheter (Cool Path Duo, St. Jude Medical or Thermo-Cool, Biosense-Webster) with a power between 25 and 35 W and an irrigation rate of 17 to 40 ml/min according to the manufacturer's recommendations to target a temperature of ≤45°C. Circumferential ablation was performed on the posterior wall more than 1 cm and on the anterior wall more than 5 mm away from the pre-defined PV ostia. When residual PV conduction was present after circumferential ablation around ipsilateral veins, PVs were mapped sequentially using a circular mapping catheter (Optima, St. Jude Medical or Lasso 2515 Variable, Biosense-Webster) to localize the earliest PV potential. Further RF ablations were placed on the circumferential line against the site with earliest PV potential recorded on the circular mapping catheter. To place the catheter as close as possible to the ablation line, the diameter of the circumferential mapping catheter was adjusted to the diameter of the PV antrum.
The ablation procedure was performed with the patients under general anesthesia and continuous transesophageal echocardiographic monitoring. The technique for proper echocardiographic visualization of the PVs and their flow was previously described (8). One single transseptal puncture was made under transesophageal echocardiographic guidance and a steerable sheath (Flexcath, Cryocath Medtronic) was placed in the left atrium (LA). Ablation was performed using a double-coated over-the-wire cryoballoon (Arctic Front, Medtronic). The inner lumen of the cryoballoon was connected to a continuous pressure monitoring system. Balloon size was selected in accordance with the diameter of the PVs, as measured by transesophageal echocardiography (TEE). If all PVs had a diameter <20 mm, the procedure was performed using a 23-mm cryoballoon; if any PV was >20 mm, we used a 28-mm balloon. The deflated cryoballoon was advanced over an extra-rigid 0.032-mm guidewire placed inside each PV (GuideRight Superstiff, St. Jude Medical or Amplatz Extra-Stiff, Cook Medical, Bloomington, Indiana), and inflated in front of the venous ostium. After inflation, the balloon was advanced to achieve occlusion of the PV, as assessed by TEE (9) and/or by the pressure curve (10). Each PV was frozen twice over 5 min in the best occlusive position. During ablation of the right PVs, we performed continuous phrenic nerve pacing from the superior vena cava to promptly detect phrenic nerve injury. If PV dimensions were too heterogeneous, the size of the balloon was switched to ensure proper occlusion. In case of persistent conduction after cryoballoon ablation, electrical isolation was segmentally completed using an 8-mm cryoablation catheter (Freezor Max, Medtronic).
Phased RF PVI Using PVAC
After gaining access through a single transseptal puncture, a 10-F sheath (Frontier Advance, Ablation Frontiers Medtronic or Bard Channel, Bard Electrophysiology Division, Lowell, Massachusetts) was placed in the LA. Hand-held left atrial contrast injection during high-rate ventricular pacing was performed for simultaneous unselected PV angiograms. The PVAC is a decapolar, steerable, 9-F, over-the-wire circular mapping and ablation catheter used in conjunction with the multichannel RF generator (GENius, Medtronic). The system is capable of synchronous application of duty-cycled phased unipolar and bipolar RF energy over all 10 PVAC electrodes, where bipolar refers to adjacent electrodes. The PVAC was directed over a guidewire placed in each of the PVs toward the PV antrum, and 5 bipolar electrograms of PV potentials were recorded over adjacent electrodes. Energy delivery was applied in either a 4:1 or 2:1 bipolar-to-unipolar ratio. Energy was limited to a maximum of 10 W with an operator-defined temperature target of 60°C (temperature controlled, power limited). Several overlapping RF applications, 60 s each, were made while repositioning the PVAC, until all PV potentials were eliminated (11). The predefined diameter of the PVAC (25 mm) is designed to deliver RF applications in the antral part of the PVs. The PVAC was also used for validation of PV bidirectional block.
After ablation in each group, PVs were mapped using a circular mapping or the PVAC. The common endpoint of all ablation procedures was complete electrical PV isolation, documented at least 30 min after the last application. No further left atrial ablation was permitted in any group.
All patients received oral anticoagulation 4 weeks before the procedure with a target international normalized ratio (INR) of 2.0 to 3.0. Anticoagulation was reduced or stopped 2 to 3 days before ablation, and patients presenting an INR <2.0 received subcutaneous low molecular weight heparin as a bridge to the procedure. The INR was checked at least 1 day before the procedure in all patients. On the day before ablation, they underwent TEE to rule out thrombi. An adjusted dose of intravenous heparin was administered immediately after the first transseptal access to target an activated clotting time (ACT) >300 s before starting ablation to obtain homogeneous levels of anticoagulation during ablation in all our patients (12,13). Transseptal sheaths were continuously perfused with heparinized saline solution to avoid clot formation or air embolism. No protamine was administered at the end of the procedure. Immediately after the removal of the catheters, oral anticoagulation was resumed, and all patients received intravenous or low molecular weight heparin until the INR was >2.5.
Neurological examination and imaging
A systematic clinical neurological examination (including cranial nerve, motor and sensory function, and gait assessment) was performed at admission, the day after the procedure, and 4 to 5 days after the procedure before discharge by the ward physician. Furthermore, all patients were monitored in the intensive care unit overnight after the procedure, including evaluation of consciousness and basic cranial nerve, motor, and sensory function every 15 min during the first 6 h.
Cerebral MRI was performed 1 day before and 1 or 2 days after ablation, using a 1.5-T unit (Siemens, Erlangen, Germany). The pre-ablation imaging protocol consisted of a T2-weighted axial fluid-attenuated inversion recovery sequence (TI = 2,500 ms, TR = 9,000 ms, TE = 116 ms, 6.0-mm slice thickness, 400 × 512 matrix) and a diffusion-weighted echo planar imaging sequence (TR = 3,500 ms, TE = 87.0 ms, TA = 00.13 ms, 6.0-mm slice thickness, 256 × 256 matrix). For each diffusion-weighted imaging sequence, the apparent diffusion coefficient map was obtained. After ablation, a T1-weighted spin echo sequence and T2-weighted imaging were also performed. All sequences were centered on the axis defined by a line passing between the anterior and posterior cerebral commissures.
On the post-ablation MRI, an acute embolic lesion was defined as a focal hyperintense area detected by the fluid-attenuated inversion recovery sequence, corresponding to a restricted diffusion signal in the diffusion-weighted imaging sequence, confirmed by apparent diffusion coefficient mapping to rule out a shine-through artifact. The number, localization, and size of the focal lesions were analyzed.
Two certified radiologists blinded to the ablation technique independently analyzed all MRI scans. In case of disagreement, the case was reviewed by a third radiologist.
The study was designed to include 240 patients (80 in each ablation group), based on a statistical power of 0.8 to detect an incidence of embolic events 25% higher in any group compared with the others, assuming an α error of 0.05. Because it was a safety study, 2 interim analyses were planned after the inclusion of 75 and 150 patients. The study was terminated in agreement with the ethics committee after the first interim analysis because the incidence of embolic events in the PVAC group was already significantly higher.
Continuous variables are expressed as the mean ± SD and are compared using 1-way analysis of variance. A post hoc analysis using Bonferroni correction was performed if significant differences could be detected. For analysis involving only the 2 RF groups, we used the Student t test. Categorical variables were compared using the chi-square test. In that case, the p value threshold for statistical significance was adjusted for multiple comparisons (0.05/3 = 0.017). The odds ratio was calculated using binary logistic regression, including parameters showing significant levels of correlation with the presence of new embolic lesions on univariate analysis. Statistical significance was set at p < 0.05.
Baseline characteristics of the 74 patients included in the study are detailed in Table 1. There were slightly more patients with persistent AF in the externally irrigated RF group, whereas LA size was similar among the groups. The thromboembolic risk profile was well balanced among the groups (CHA2DS2VASc score: 1.7 ± 1.5 in the externally irrigated RF group, 1.7 ± 1.3 in the cryoballoon group, and 1.3 ± 1.0 in the PVAC group; p = 0.29).
Seventeen of the 74 patients presented in AF, and AF occurred during the procedure in another 10 patients. Twenty-three patients underwent cardioversion during the procedure. The need for cardioversion was similar in all 3 groups (30% in the externally irrigated RF group, 22% in the cryoballoon group, and 42% in the PVAC group; p = 0.33).
All PVs could be successfully isolated after 41 ± 14 min of RF in the externally irrigated RF group and 25 ± 9 min of RF in the PVAC group (p < 0.001). Despite the significantly shorter ablation time, the total delivered energy was similar in both RF groups (68,240 ± 23,084 J for the externally irrigated RF group and 80,540 ± 43,447 J for the PVAC group; p = 0.28). Energy was delivered in a 4:1 mode in 54% of patients and in a 2:1 mode in the rest. In the cryoballoon group, 5 of 23 patients (22%) were ablated with the 23-mm balloon, 16 of 23 (70%) with the 28-mm balloon, and 2 of 23 (9%) with both. In 14 of the 92 targeted PVs (15%), supplementary focal cryoablation was necessary to complete PVI.
Total procedure durations were 198 ± 50 min in the externally irrigated RF group, 174 ± 35 min in the cryoballoon group, and 124 ± 32 min in the PVAC group (p < 0.001 for PVAC versus externally irrigated RF and cryoballoon after Bonferroni correction). Total fluoroscopy times were 34 ± 17 min in the externally irrigated RF group, 27 ± 13 min in the cryoballoon group, and 24 ± 10 min in the PVAC group (p = 0.051 for PVAC versus externally irrigated RF after Bonferroni correction). Procedure characteristics for all 3 groups are summarized in Table 2.
Transient phrenic nerve paralysis occurred in 2 patients in the cryoballoon group and resolved completely before discharge. One cryoballoon procedure had to be aborted because of pericardial effusion just after transseptal access through an aneurysmal interatrial septum. The procedure was rescheduled 1 month later and was successfully performed through a puncture done with an NRG RF transseptal needle (Baylis, Toronto, Ontario, Canada). Finally, 1 patient in the externally irrigated RF group presented with a groin hematoma requiring surgical revision, and another presented with generalized edema that resolved with diuretic therapy.
Incidence of new embolic events after ablation
No overt neurological events were observed, and findings on clinical neurological examination were normal in all patients before and after ablation. Post-procedure MRI of 2 of 27 patients (7.4%) in the externally irrigated RF group and 1 of 23 (4.3%) in the cryoballoon group revealed a new embolic lesion (Fig. 1). In the PVAC group, 9 of 24 patients (37.5%) presented a median of 3 acute lesions (range, 1 to 5 acute lesions) on post-procedure MRI (Fig. 2). The presence of 1 or more new embolic lesions differed significantly among the 3 groups (p = 0.003), with the highest rate in the PVAC group. The incidence of new embolic events was similar in the 2 energy delivery modes (6 of 13 in 4:1 and 3 of 11 in 2:1, p = 0.42). In 2 patients, macroscopic charring formation could be observed on the PVAC electrodes at the end of the procedure: 1 of them showed 2 new lesions on post-intervention MRI and the other patient none. The new lesions were broadly distributed over both hemispheres (15 right-sided and 12 left-sided lesions), both supratentorial and infratentorial, but were located preferably in the vertebrobasilar territory (16 of 27, 59%), supporting its cardiac origin. The median lesion size was 5.5 mm (interquartile range: 3.7 to 6.8 mm). Table 3 provides a summary of these data. Intraprocedure anticoagulation was comparable in those with or without new embolic lesions. Mean intraprocedure ACT was 316 ± 33 s in patients with versus 322 ± 54 s in patients without emboli (p = 0.649). The need for cardioversion was also similar: 18% (2 of 11) in patients with versus 33% (21 of 63) in patients without embolic events (p = 0.49). Univariate analysis searching for predictors of new embolic events included catheter type, age, sex, hypertension, diabetes, structural heart disease, LA size, type of AF, rhythm, cardioversion, and mean ACT during ablation. The use of a PVAC remained the only significant predictor of new ischemic events (odds ratio: 9.4, 95% confidence interval: 2.2 to 39.3; p = 0.002).
Recent studies have raised safety concerns about subclinical embolic events during PVIs (6,14). The current study aimed to assess whether the ablation technique (conventional open-irrigated RF, multielectrode phased RF, or cryoballoon ablation) has an impact on the incidence of new embolic events detected by cerebral MRI. All 3 groups presented new subclinical embolic lesions after PVI. However, the incidence and number of new lesions were significantly higher in patients treated with a PVAC, with a predilection for the vertebrobasilar territory, favoring a cardiac origin.
Thromboembolic events after left atrial catheter ablation procedures
Although the incidence of clinical thromboembolic events after AF catheter ablation is low (0.94%) (5), recent studies have shown a substantial incidence of subclinical microembolic events, ranging from 8% to 14% after cryoenergy or externally irrigated RF procedures (6,14–16).
Embolic events related to a PVI have several potential causes. They may occur as a result of dislodgment of a preexisting left atrial thrombus by catheter manipulation. TEE was performed in all our patients before the procedure to exclude this possibility. Fresh thrombus can also form on the sheath, the catheter, and over the newly created myocardial lesions. As a matter of fact, systematic use of intracardiac echocardiography during left atrial ablation procedures detected as many as 10% of fresh thrombi attached to a sheath or a mapping catheter (17,18). To minimize thrombus formation, all patients received adequate anticoagulation both during and after ablation. Air and tissue debris can also be responsible for embolism. During the procedure, all sheaths (including the 15-French Flexcath sheath, Cryocath Medtronic) were carefully flushed and constantly perfused with heparinized saline solution, making an air embolism unlikely. Finally, charring (carbonization) is a real source of concern, especially with nonirrigated catheters (19). Only 1 of our patients with a new embolic lesion presented gross charring on the catheter electrodes after the procedure. However, larger charring formations might have dislodged during catheter withdrawal inside the sheath.
The incidence of new embolic lesions found in externally irrigated RF and cryoballoon groups was among the lowest reported, similar to the one reported by Neumann et al. (14) in the MEDAFI (Microembolization During Ablation of Atrial Fibrillation) trial (7.9%). Our study population was also similar to that of the MEDAFI trial, with almost two-thirds of cases being paroxysmal AF, and none being long-standing AF. The type of AF, however, was not associated with embolic events in any of the previous trials. Another possible explanation for the relatively low embolic event rate in our cohort is the ablation strategy. Indeed, 8 of 33 patients who presented with acute intracranial lesions after the procedure in the study by Gaita et al. (6) received wide-area ablation including linear lesions and atrial defragmentation. Expanding the set of lesions created also increases the surface of damaged myocardium and therefore the substrate for thrombus formation.
Clinical significance of subclinical intracranial embolic events after AF catheter ablation
All intracranial embolic events observed in our study were asymptomatic. Therefore, they would have remained unrecognized in everyday clinical practice.
Even without previous catheter ablation, patients with AF have been reported to have an increased prevalence of silent cerebral infarctions on routine MRI monitoring compared with matched patients in sinus rhythm (20). AF patients have also been reported to have a higher incidence of dementia than those in sinus rhythm (21). A recent study by Schwarz et al. (15) found a significant neuropsychological decline in patients after AF catheter ablation. Although not statistically related to the incidence of embolic lesions in this small study (with only 2 new silent lesions), these events could be clinically meaningful from a neurological point of view. Other cardiovascular interventions associated with a notable rate of microembolic events, such as cardiac surgery (22–24) and carotid stenting (25), have been correlated with post-intervention neuropsychological deficits, although the embolic load could not always be correlated with the degree of cognitive deficits.
Silent intracranial embolism and PVAC technology
Although not randomized, our 3 study groups had a similar clinical thromboembolic risk profile and received the same degree of anticoagulation during the procedure. Therefore, the difference observed in the rate of new embolic events seems to be catheter specific.
Duty cycling and phasing during energy delivery are supposed to allow periodic cooling of the tissue and prevent heating-related generation of thrombus or charring. However, this is highly dependent on tissue contact and local blood flow, which are parameters that are extremely difficult to assess on individual electrodes. This might lead to overheating of some poles in some catheter positions and therefore to carbonization. Software modifications of the generator may help decrease temperature overshoots during intermittent contact. Another possible burdensome point is the lack of irrigation, especially considering that the PVAC delivered the same amount of energy as an externally irrigated RF catheter in almost half of the time. Irrigated-tip catheters are routinely used in left atrial catheter ablations because of a lower rate of thrombus formation or carbonization compared with nonirrigated 4-mm catheters (19). An intensified anticoagulation regimen or additional platelet inhibition may reduce the embolic rate when using PVAC technology, but may also increase the risk of bleeding (especially secondary intracerebral hemorrhage). The use of dabigatran instead of vitamin K inhibitors might also have had an impact on the rate of embolic events. Further studies might help to clarify this issue. Finally, routine use of intracardiac echocardiography might have enabled identification and prevention of at least some of these embolic events.
Patients were not randomized to the different techniques. Nevertheless, there were no differences in the clinical characteristics of the patients. Each ablation technique was intentionally performed only by operators at high-volume centers, ensuring significant experience with the technique.
Neither an exhaustive neurological examination nor standardized neuropsychological testing was performed in our patients. Post-procedure minor neurological deficits or neurocognitive alterations may have been missed in this study.
Used in everyday clinical management of intraprocedure anticoagulation, the ACT remains a very dynamic parameter. A mean ACT >300 s cannot be taken as a guarantee of a consistent anticoagulation throughout the procedure. Systematic ACT testing every 20 min could have improved anticoagulation levels and prevented some embolic events.
Finally, the sample size and event count were very small. This did not allow us to search for additional predictors of new embolic events with multivariate analysis.
PVI performed with the PVAC presents a notably higher incidence of subclinical intracranial embolic events than if performed with an externally irrigated RF catheter or a cryoballoon. Improvement in PVAC technology and further studies to clarify the origin of these embolic lesions are mandatory to reduce the rate of silent embolisms during ablation procedures in the LA.
Dr. Herrera Siklódy has received consulting and speaker honoraria from Medtronic and speaker honoraria from Biosense Webster. Dr. Hocini has received speaker honoraria from Biosense Webster. Dr. Miyazaki has received fellowship support from St. Jude Medical. Dr. Schiebeling-Römer has received fellowship support from St. Jude Medical. Dr. Haïssaguerre has received speaker honoraria from Biosense Webster. Dr. Arentz has received consulting honoraria from Medtronic. All other authors have reported that they have no relationships to disclose.
- Abbreviations and Acronyms
- activated clotting time
- atrial fibrillation
- international normalized ratio
- left atrium
- magnetic resonance imaging
- pulmonary vein
- multielectrode duty-cycled radiofrequency pulmonary vein ablation catheter
- pulmonary vein isolation
- transesophageal echocardiography
- Received January 23, 2011.
- Revision received March 21, 2011.
- Accepted April 27, 2011.
- American College of Cardiology Foundation
- Calkins H.,
- Brugada J.,
- Packer D.L.,
- et al.
- Neumann T.,
- Vogt J.,
- Schumacher B.,
- et al.
- Beukema R.P.,
- Beukema W.P.,
- Smit J.J.,
- et al.
- Cappato R.,
- Calkins H.,
- Chen S.A.,
- et al.
- Gaita F.,
- Caponi D.,
- Pianelli M.,
- et al.
- Arentz T.,
- Weber R.,
- Burkle G.,
- et al.
- Neumann T.,
- Kuniss M.,
- Conradi G.,
- et al.
- Schrickel J.W.,
- Lickfett L.,
- Lewalter T.,
- et al.
- Ren J.F.,
- Marchlinski F.E.,
- Callans D.J.
- Yokoyama K.,
- Nakagawa H.,
- Wittkampf F.H.,
- Pitha J.V.,
- Lazzara R.,
- Jackman W.M.
- Das R.R.,
- Seshadri S.,
- Beiser A.S.,
- et al.
- Pugsley W.,
- Klinger L.,
- Paschalis C.,
- Treasure T.,
- Harrison M.,
- Newman S.