Author + information
- Received August 15, 2017
- Revision received October 2, 2017
- Accepted October 10, 2017
- Published online January 15, 2018.
- Lennart van Gils, MDa,
- Jochen Wöhrle, MDb,
- David Hildick-Smith, MDc,
- Sabine Bleiziffer, MDd,
- Daniel J. Blackman, MDe,
- Mohamed Abdel-Wahab, MDf,
- Ulrich Gerckens, MDg,
- Stephen Brecker, MDh,
- Vinayak Bapat, MDi,
- Thomas Modine, MD, PhDj,
- Osama I. Soliman, MD, PhDk,
- Andrey Nersesov, BSl,
- Dominic Allocco, MDl,
- Volkmar Falk, MD, PhDm,n and
- Nicolas M. Van Mieghem, MD, PhDa,∗ ()
- aDepartment of Cardiology, Thoraxcenter, Erasmus Medical Center, Rotterdam, the Netherlands
- bDepartment of Internal Medicine II, University of Ulm, Ulm, Germany
- cDepartment of Cardiology, Sussex Cardiac Centre, Brighton and Sussex University Hospitals, Brighton, United Kingdom
- dDepartment of Cardiovascular Surgery, Clinic for Cardiovascular Surgery, Deutsches Herzzentrum München, Technische Universität München, Munich, Germany
- eDepartment of Cardiology, Leeds General Infirmary, Leeds, United Kingdom
- fDepartment of Cardiology, Herzzentrum, Segeberger Kliniken, Bad Segeberg, Germany
- gDepartment of Cardiology, University of Rostock, Rostock, Germany
- hDepartment of Cardiology, St. George's Hospital, London, United Kingdom
- iDepartment of Cardiac Surgery, Guy's and St Thomas' Hospital, London, United Kingdom
- jDepartment of Cardiovascular Surgery, CHRU Lille - Hôpital Cardiologique, Lille, France
- kCardialysis Core Laboratories and Clinical Trial Management, Rotterdam, the Netherlands
- lBoston Scientific Corporation, Marlborough, Massachusetts
- mDepartment of Cardiothoracic and Vascular Surgery, Klinik für Herz-Thorax-Gefässchirurgie, Deutsches Herzzentrum Berlin, Berlin, Germany
- nDepartment of Cardiovascular Surgery, Klinik für Kardiovaskulr Chirurgie, Charite Berlin, Berlin, Germany
- ↵∗Address for correspondence:
Dr. Nicolas M. Van Mieghem, Department of Interventional Cardiology, Thoraxcenter, Erasmus Medical Center, Room Bd 171, ‘s Gravendijkwal 230 3015 CE Rotterdam, the Netherlands.
Objectives The aim of this post hoc analysis from the RESPOND (Repositionable Lotus Valve System–Post-Market Evaluation of Real World Clinical Outcomes) post-market study was to assess the final implantation depth on the contrast aortogram after Lotus valve (Boston Scientific, Marlborough, Massachusetts) transcatheter aortic valve replacement (TAVR) and to correlate with permanent pacemaker implantation (PPI) and paravalvular leak (PVL).
Background Contrast aortography allows for the assessment of implantation depth and PVL during and after TAVR. Previous reports suggested an association between final device position and rates of PPI and PVL.
Methods The RESPOND study was a prospective, open-label, single-arm study in 41 centers evaluating outcomes after Lotus TAVR in routine clinical practice. Aortograms were collected at the Erasmus Medical Center and analyzed by researchers who were blinded to clinical outcomes. The primary analysis correlated implantation depth with PPI and PVL and required aortograms in a coaxial projection. The relation between implantation depth and need for PPI was assessed by multivariate logistic regression, adjusting for pre-defined confounders. A secondary analysis compared PVL analysis by contrast aortography with transthoracic echocardiography (TTE) performed by the independent core laboratory.
Results A total of 724 angiographic studies were included in this analysis. Mean Lotus implantation depth was 6.67 ± 2.19 mm. The overall PPI rate was 35%. PPI rate was lower with shallow implants (<6.5 mm: 21% vs. ≥6.5 mm: 41%; p < 0.001). After adjustment for confounders, implantation depth independently predicted need for PPI (odds ratio per 1-mm increment in depth: 1.200; 95% confidence interval: 1.091 to 1.319; p = 0.002). More than trivial PVL was present in 23% by contrast aortography and in 8% by TTE. Implantation depth was not correlated with PVL by contrast aortography or TTE (p = 0.342 and p = 0.149, respectively). PVL grading by contrast aortography and TTE was concordant in 77%.
Conclusions In this post hoc analysis of the RESPOND study PPI was highly correlated with implantation depth, whereas PVL was not. Higher Lotus implantation may reduce need for PPI.
Transcatheter aortic valve replacement (TAVR) is recommended for symptomatic severe aortic stenosis in patients at elevated surgical risk (1–7). Multiple transcatheter heart valve (THV) designs are commercially available (8). The Lotus valve (Boston Scientific, Marlborough, Massachusetts) is a mechanically expanding system and includes an adaptive seal. These features make it completely repositionable and retrievable for precise placement as well as minimizing paravalvular leak (PVL) (9,10). The RESPOND (Repositionable Lotus Valve System–Post-Market Evaluation of Real World Clinical Outcomes) study was a prospective post-market study including 1,014 patients from 41 centers and confirmed Lotus valve safety and efficacy with an 2.6% all-cause mortality and 2.2% disabling stroke rate at 30 days with more than mild PVL in 0.3% and permanent pacemaker implantation (PPI) undertaken in 30% of patients (11). TAVR in its current form no longer requires general anesthesia and relies mostly on fluoroscopic guidance. Operators use contrast aortography to determine the implantation depth, device position relative to the coronary ostia, and final PVL assessment. The aim of this post hoc analysis from the RESPOND study was to assess the final implantation depth on the contrast aortogram after Lotus TAVR and to correlate with need for PPI and PVL.
Study population and design
The design and outcomes of the RESPOND post-market study (NCT02031302) have been reported elsewhere (11). In brief, 1,014 patients with elevated operative risk were treated with Lotus TAVR and prospectively enrolled. An independent core laboratory (Cardialysis, Rotterdam, the Netherlands) analyzed the transthoracic echocardiography (TTE).
Clinical events were reported through electronic clinical research forms using the latest Valve Academic Research Consortium-2 criteria (12); all events were monitored by a contract research organization, and an external independent medical reviewer adjudicated death and stroke. The final contrast aortograms after implantation of the Lotus valve were collected and transferred to the Erasmus Medical Center for centralized uniform and blinded analysis. All patients provided written informed consent for participation in the RESPOND study. The primary objective of this study was to correlate depth of Lotus implantation with need for PPI and occurrence of more-than-trivial PVL by contrast aortography. A secondary analysis looked at the concordance of PVL grading between contrast aortography and pre-discharge TTE as assessed by the independent core laboratory.
The current analysis used the as-treated population from the RESPOND study (n = 996) and excluded 132 patients with a pacemaker before TAVR. Of the 864 patients without a pacemaker a final contrast aortogram was not acquired in 140 patients, thus 724 cases were available for the final analysis.
For the PVL analysis contrast aortograms were submitted to the following quality check: 1) presence of sufficient contrast volume; 2) pigtail located >2 cm above the aortic annulus; and 3) no wire across the Lotus valve. Of the 724 cases, 36 (5%) did not meet these criteria and were thus removed from the PVL analysis (Figure 1).
Previous reports suggested an error in measuring the implantation depth if the aortogram had been obtained in a noncoaxial projection (13). To address this matter, all participating centers were requested to perform a baseline and final aortogram in the same coaxial C-arm projection with the 3 coronary cusps aligned. Ultimately, 506 of 724 (70%) aortograms were acquired using a coaxial projection. Only aortograms in coaxial projection were used for the primary implantation depth analysis.
Dedicated trained clinical researchers, who were blinded to clinical and echocardiographic results, analyzed all aortograms. Measurements were performed with Cardiovascular Angiographic Analysis System version 5.11.2 (Pie Medical Imaging, Maastricht, the Netherlands). Core measurements consisted of the final implantation depth at the noncoronary and left coronary cusps. The final implantation depth was considered as the average of the depth at the noncoronary cusp and left coronary cusp. Interobserver variability was evaluated using intraclass correlation coefficients in 10 randomly chosen subjects. Depth differential of the Lotus valve was measured as the difference in implantation depth at the noncoronary cusp and left coronary cusp. The degree of PVL was evaluated using the Sellers criteria (14) as follows: grade 1 (mild PVL), a limited amount of contrast enters the left ventricle during diastole resolving with each beat without reaching the apex of the left ventricle; grade 2 (moderate PVL), the contrast enters and fills the entire left ventricle up to the apex but the opacification remains less than in the ascending aorta; grade 3 (moderately severe PVL), complete left ventricular opacification with similar contrast density compared with the ascending aorta; and grade 4 (severe PVL), opacification of the entire left ventricle at a higher density than the ascending aorta. All TTE data were obtained from the independent core laboratory. PVL grading by TTE was according to Valve Academic Research Consortium-2 criteria (12).
Continuous data are presented as mean ± SD and categorical variables as counts and percentages. Categorical data were compared with the chi-square test for trend and Z-test for proportions. Due to missing site-reported left ventricular outflow tract (LVOT) dimensions by computed tomography and the derived oversizing relative to the LVOT (38% missing), multiple imputation was performed for completion of this single variable, assuming this variable was missing at random (15). Five imputation steps were performed, using the following predictor variables: age, sex, weight, height, history of heart failure, history of coronary artery disease, atrial fibrillation, EuroSCORE (European System for Cardiac Operative Risk Evaluation), pre-existent conduction disturbances, annulus diameter (area derived), pre-dilatation, repositioning, valve prosthesis size, implantation depth, and need for PPI. Logistic regression was applied to evaluate the association between implantation depth and PPI using a predefined multivariate model with inclusion of the following clinically suspected confounders: age, sex, body mass index, coronary artery disease, pre-existent conduction disturbances (i.e., first-degree atrioventricular block, right bundle branch block [RBBB], left bundle branch block, left anterior fascicular block), repositioning, pre-dilatation, high enrolling centers (>50 patients enrolled), oversizing relative to the LVOT, and depth differential of the prosthesis. The p values from multivariate analysis were adjusted for multiple comparisons using the Bonferroni correction. When p values exceeded 1.0 after Bonferroni correction the p value was truncated at 1.0. Data analysis was performed in SPSS version 21 (SPSS Inc., Chicago, Illinois). A 2-sided p value of <0.05 was considered statistically significant.
A total of 724 patients within the RESPOND study were eligible for this post hoc analysis (Figure 1). Baseline demographics are listed in Table 1. Mean age was 81 ± 7 years and 46% were men. Mean Society of Thoracic Surgeons Predicted Risk of Mortality Score was 6.1 ± 7.0%. Coronary artery disease was present in 54% of patients. A total of 36% had a history of congestive heart failure. Pre-existent conduction disturbances were present as follows: first-degree atrioventricular block in 13%, left bundle branch block in 9%, RBBB in 6%, and left anterior fascicular block in 4%. Procedural characteristics are listed in Table 2. The 3 available Lotus valve sizes were equally distributed among patients: 23 mm in 28% of patients, 25 mm in 40%, and 27 mm in 32%. Balloon pre-dilatation was performed in more than one-half of the patients (54%). Repositioning was attempted in one-third (32%).
All-cause death at 30 days occurred in 12 (1.7%) patients, which were all due to cardiovascular causes. At 30 days, clinical stroke had occurred in 20 (2.8%) patients, of which 14 (1.9%) were disabling strokes.
PPI and PVL
PPI was required in 254 (35%) patients. The primary causes for PPI were high-degree atrioventricular block (67%) and bradycardia (9%). In 23% of patients PPI was performed for other reasons (e.g., bifascicular block, second-degree atrioventricular block Mobitz I type).
By contrast aortography, more than trivial PVL was present in 164 (23%) patients: no PVL in 554 (77.2%) patients, mild PVL in 149 (20.8%) patients, moderate PVL in 14 (1.9%) patients, and moderate-to-severe PVL in 1 (0.1%) patient. By TTE, more than trivial PVL was present in 54 (7.9%) patients: none or trace in 628 (92.1%) patients, mild in 53 (7.8%) patients, and moderate in 1 (0.1%) patient.
The interobserver variability for measuring the implantation depth was small (intraclass correlation coefficient 0.964; 95% confidence interval [CI]: 0.858 to 0.991). Final mean implantation depth below the aortic annulus (i.e., average of depth at the noncoronary and left coronary cusp) was 6.67 ± 2.19 mm. Implantation depth had a normal distribution (Figure 2).
Relation between implantation depth and PPI
PPI rates per implantation depth quartiles are listed in Table 3. There was a clear trend toward higher PPI rates with deeper implants (p < 0.001).
Deeper implantation depth was associated with PPI by univariate analysis (odds ratio [OR] per 1-mm increment: 1.206; 95% CI: 1.102 to 1.319; p < 0.001) in the primary analysis (Figure 3). When patients with noncoaxial projections were also included the implantation depth was still associated with an increased risk of PPI (OR: 1.172; 95% CI: 1.091 to 1.258; p < 0.001) (Online Figure 1).
After adjustment for predefined potential confounders, deeper implantation remained an independent predictor for PPI requirement (OR per 1-mm increment: 1.200; 95% CI: 1.091 to 1.319; adjusted p = 0.002) (Figure 4). Additionally, repositioning also independently predicted PPI (OR: 1.927; 95% CI: 1.261 to 2.945; adjusted p = 0.028). When patients with noncoaxial projections were also included both implantation depth (OR: 1.157; 95% CI: 1.074 to 1.247; p = 0.004) and pre-existent RBBB (OR: 3.283; 95% CI: 1.654 to 6.518; p = 0.014) independently predicted PPI (Online Figure 2).
Relation between implantation depth and PVL
There was no correlation between implantation depth and presence of more than trivial PVL by TTE (p for trend = 0.149) or contrast aortography (p for trend = 0.342) (Figure 5).
Concordance of PVL grading between imaging strategies
PVL grading by contrast aortography and TTE were concordant in 77% of patients (Figure 6). A total of 19% of patients had more-than-trivial PVL by contrast aortography but not by TTE at discharge. Conversely, 4% of patients had more than trivial PVL by TTE at discharge and not by contrast aortography. In all patients with moderate PVL by contrast aortography (n = 14) the PVL was either mild (n = 7) or none/trace (n = 7) by TTE. One patient had moderate-to-severe PVL by contrast aortography and died during the index procedure without echocardiographic analysis. Moderate PVL by TTE was present in 1 patient. Interestingly, this patient had no visible PVL by aortography.
This post hoc analysis from the RESPOND study reports advanced insights from contrast aortography after Lotus TAVR. First, meticulous contrast aortography technique following Lotus TAVR is a prerequisite for adequate evaluation of PVL and Lotus depth of implantation. Second, deeper Lotus implants lead to more frequent PPI. Third, contrast aortography allows for a proper PVL assessment and seems more sensitive than TTE. Interestingly, all patients with more than mild PVL by contrast aortography (Sellers grade ≥2) had either no or trivial or mild PVL by TTE. Contrast aortography thus offers important insights for Lotus TAVR guidance, including prediction of PPI and acceptable PVL evaluation in conscious patients.
TAVR is increasingly executed under local anesthesia with or without conscious sedation with favorable results (16,17). The obvious advantages of local anesthesia include patient comfort, hemodynamic stability, rapid detection of complications (stroke, vascular complications), faster recovery, and shorter intensive care and hospital stay (18). Transesophageal echocardiographic guidance is challenging with TAVR under local anesthesia. Therefore, valve positioning and evaluation of PVL predominantly relies on fluoroscopic guidance and contrast aortography.
Pre-procedural planning by 3-dimensional multislice computed tomography scanning and new TAVR designs improved transcatheter valve sizing and positioning (8) with excellent clinical outcome and low PVL rates (11,19,20). Two randomized trials have established TAVR feasibility in patients at intermediate operative risk (5,6), and challenging anatomies (21). However, high PPI rates remain a matter of concern with new generation devices: ∼13% with the SAPIEN 3 (Edwards Lifesciences, Irvine, California) (20), ∼17% with the Evolut R (Medtronic, Minneapolis, Minnesota) (19), and ∼30% with the Lotus (10,11).
Conduction disorders following TAVR may occur because of the close proximity of the atrioventricular node, bundle of His, and left bundle branch to the native aortic annulus (22,23). It seems plausible that deeper implants of a THV result in more interaction with the conduction system that could lead to more PPI.
Several studies reported a correlation between CoreValve (Medtronic) implantation depth and PPI (24–26). Also with balloon-expandable THVs there seems to be a relation between implantation depth and persistent conduction disturbances or PPI (27–29). Currently, aiming for higher implants is recommended in order to reduce the need for PPI.
A subanalysis from the randomized REPRISE-II (Repositionable Percutaneous Replacement of Stenotic Aortic Valve Through Implantation of Lotus Valve System II) study demonstrated only a weak (nonsignificant) correlation between implantation depth and need for PPI with the repositionable Lotus valve (30). The fact that the present post hoc analysis from the RESPOND study did demonstrate a strong correlation is most plausibly due to the methodological setup of the study. It represents the largest study to date with systematic measurements of implantation depth and outnumbers the REPRISE-II study or any other study with different devices for that matter. Furthermore, by dichotomizing the implantation depth (i.e., ≤5 mm vs. >5 mm), the REPRISE-II study consequently lost statistical power to demonstrate the correlation with PPI (30). The present study reveals a linear relation between Lotus implantation depth and PPI requirement after correction for multiple confounders. The mean implantation depth of 6.67 mm leaves room for improvement by device alterations on the one hand and operator awareness on the other hand. Depth Guard (Boston Scientific, Marlborough, Massachusetts) is a recent alteration that may improve accuracy in Lotus positioning enabling higher implants (Figure 7). Ongoing studies will determine whether PPI rates will fall with Depth Guard.
In this study, repositioning predicted PPI as well. Lotus radial strength and interactions at multiple different locations within the LVOT during repositioning may promote conduction disorders.
With the self-expanding CoreValve, PVL was associated with too high or too low implants (31–34). In the REPRISE-II trial, high Lotus implants were associated with mild or moderate PVL (35). In contrast, our study suggests high Lotus implants are not associated with PVL. The adaptive seal of the outer skirt seems to cover periprosthetic spaces and adequately prevents PVL. Also the latest balloon expandable and self-expanding THV have a sealing fabric to help reduce PVL (8).
Periprocedural PVL assessment is essential for TAVR success. Current trends of performing TAVR under local anesthesia and mild or no sedation make transesophageal echocardiography less suitable for PVL assessment. Many centers now rely more on contrast aortography, TTE, and hemodynamic assessment. The present study showed an overall 77% concordance between the per-procedural semiquantitative Sellers grading by contrast aortography and core laboratory–evaluated pre-discharge TTE. Interestingly, contrast aortography seemed more sensitive to detect PVL because in 19% PVL was detected by aortography but not by pre-discharge TTE, though it is possible that this difference might be at least partially explained by improvement in PVL between the time of the aortogram and echocardiography. This signal for a higher sensitivity for PVL detection by immediate aortography as compared with pre-discharge TTE was also observed in the randomized CHOICE (Comparison of Transcatheter Heart Valves in High Risk Patients With Severe Aortic Stenosis) trial (36). Both modalities, contrast aortography and TTE, have their inherent pitfalls in the detection of PVL (37). First, contrast aortography is performed during the procedure, whereas TTE is performed several days thereafter. In many centers TAVR still occurs under general anesthesia, which by itself affects hemodynamics. Also, PVL might dissolve over days due to continued frame expansion and settling of the sealing skirt.
Second, PVL grading by TTE may suffer from artifacts such as acoustic shadowing of the stented frame and merging jets from multiple directions. With contrast aortography, power injection too close to the bioprosthesis may induce PVL or overlapping anatomical entities such as the spine or the descending thoracic aorta may influence the interpretation of the aortogram (38,39). New software tools, such as video-densitometric aortography, may further improve the reliability and reproducibility of contrast aortography (39,40).
This subanalysis of the RESPOND post-market study represents the largest cohort to evaluate contrast aortography after TAVR. However, the following limitations should be acknowledged. Although data collection and validation was rigorous and the population size was large, the observational design of this post hoc analysis must be acknowledged. Therefore, our findings may require confirmation by further study.
Acquisition of a post-implantation aortogram is part of the default workflow of a Lotus TAVR. Nevertheless, in 1 of 6 patients this aortogram was not available for final analysis due to logistical reasons. Aortograms were executed by the local TAVR teams according to local standards. Aortogram quality and coaxiality affect device implantation depth analysis. Despite study recommendations 30% of the aortograms were obtained in a noncoaxial projection and 2% were of insufficient quality for accurate PVL grading. Also, the decision to proceed with PPI was per treating physician’s discretion.
Even with more shallow implants PPI rate was relatively high after Lotus implantation and suggests improving the implantation depth would only partly solve the pacemaker issue with this device.
Finally, we compared procedural contrast aortography with pre-discharge TTE. PVL may have changed within this time window and this may have impacted the comparison between the 2 modalities. Contrast aortography suffers from inherent limitations such as challenging determination of PVL mechanism and location. Also, it requires potentially harmful administration of intravascular contrast, especially in case of repeated aortograms. Nevertheless, apart from aortic regurgitation assessment aortography is essential in contemporary TAVR practice to determine coronary patency and device position.
In this post hoc analysis of the RESPOND study PPI was highly correlated with Lotus implantation depth, whereas PVL was not. Higher Lotus implantation may reduce need for PPI.
WHAT IS KNOWN? Implantation depth as determined by contrast aortography was shown to correlate with need for PPI with self-expanding and balloon-expandable THVs.
WHAT IS NEW? With Lotus TAVR, contrast aortography allows assessment of PVL and determination of device implantation depth, which in turn correlates with need for PPI. This contributes to feasibility of Lotus TAVR under local anesthesia.
WHAT IS NEXT? Operator awareness and new device alterations may reduce PPI rates through higher device implantations.
The authors thank Prof. E. Boersma (Erasmus Medical Center) and P. Lam (Boston Scientific) for critical statistical review.
Dr. Wöhrle has received research grant support from Boston Scientific. Dr. Hildick-Smith has served as a proctor and consultant for Boston Scientific, Medtronic, and Edwards Lifesciences. Dr. Bleiziffer has served as a proctor and consultant for Medtronic. Dr. Abdel-Wahab has served as a proctor for Boston Scientific. Dr. Gerckens has served as a consultant and proctor for Medtronic and Boston Scientific. Dr. Brecker has served as a consultant for Medtronic and Boston Scientific. Dr. Bapat has served as a consultant for Edwards Lifesciences, Medtronic, Abbott Vascular, Boston Scientific, and 4tech. Dr. Modine has a served as a consultant for Boston Scientific. Mr. Nersesov and Dr. Allocco are full-time employees of Boston Scientific. Dr. Van Mieghem has received research grant support from Claret Medical, Abbott Vascular, Boston Scientific, Medtronic, and Edwards Lifesciences. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- confidence interval
- left ventricular outflow tract
- odds ratio
- permanent pacemaker implantation
- paravalvular leak
- right bundle branch block
- transcatheter aortic valve replacement
- transcatheter heart valve
- transthoracic echocardiography
- Received August 15, 2017.
- Revision received October 2, 2017.
- Accepted October 10, 2017.
- Reardon M.J.,
- Van Mieghem N.M.,
- Popma J.J.,
- et al.
- Nishimura R.A.,
- Otto C.M.,
- Bonow R.O.,
- et al.
- van Gils L.,
- Tchetche D.,
- Latib A.,
- et al.
- Meredith Am I.T.,
- Walters D.L.,
- Dumonteil N.,
- et al.
- Falk V.,
- Wohrle J.,
- Hildick-Smith D.,
- et al.
- Kappetein A.P.,
- Head S.J.,
- Genereux P.,
- et al.
- Piazza N.,
- Lauzier P.,
- Mylotte D.
- Sterne J.A.,
- White I.R.,
- Carlin J.B.,
- et al.
- Oguri A.,
- Yamamoto M.,
- Mouillet G.,
- et al.
- Popma J.,
- Dibu G.,
- Hughes G.C.,
- et al.
- Yoon S.H.,
- Lefevre T.,
- Ahn J.M.,
- et al.
- Piazza N.,
- de Jaegere P.,
- Schultz C.,
- Becker A.E.,
- Serruys P.W.,
- Anderson R.H.
- Khawaja M.Z.,
- Rajani R.,
- Cook A.,
- et al.
- Piazza N.,
- Onuma Y.,
- Jesserun E.,
- et al.
- Petronio A.S.,
- Sinning J.M.,
- Van Mieghem N.,
- et al.
- Binder R.K.,
- Webb J.G.,
- Toggweiler S.,
- et al.
- Urena M.,
- Mok M.,
- Serra V.,
- et al.
- Husser O.,
- Pellegrini C.,
- Kessler T.,
- et al.
- Dumonteil N.,
- Meredith I.T.,
- Blackman D.J.,
- et al.
- Ali O.F.,
- Schultz C.,
- Jabbour A.,
- et al.
- Sherif M.A.,
- Abdel-Wahab M.,
- Stocker B.,
- et al.
- Blackman D.J.,
- Meredith I.T.,
- Dumonteil N.,
- et al.
- Pibarot P.,
- Hahn R.T.,
- Weissman N.J.,
- Monaghan M.J.