Author + information
- Received February 3, 1999
- Revision received January 20, 2000
- Accepted March 29, 2000
- Published online August 1, 2000.
- Dietmar Bänsch, MDa,* (, )
- Dirk Böcker, MDb,
- Jürgen Brunn, MDb,
- Max Weber, MDb,
- Günter Breithardt, MD, FACC, FESCb and
- Michael Block, MDb
- ↵*Reprint requests and correspondence: Dr. Dietmar Bänsch, St. Georg Hospital, Lohmühlenstr. 5 D-20099 Hamburg, Germany
This retrospective study was undertaken to provide data on occurrence, significance and therapy of ventricular tachyarrhythmia (VT) clusters (VTCs) in patients with dilated cardiomyopathy (DCM) and implantable cardioverter defibrillators (ICDs).
Data on the clinical significance of VTCs are lacking in patients with DCM and ICDs.
Baseline characteristics of 106 consecutive patients with DCM and ICDs were prospectively collected, and chart reviews and episode data retrospectively analyzed. A VTC was defined as ≥3 sustained VTs/24 h.
During a mean follow-up of 33 ± 23 months, 73 patients (68.9%) had recurrent VT or ventricular fibrillation (VF), 43 patients (40.6%) suffered only single VTs and 30 patients (28.3%) experienced 52 clusters of VTs. Actuarial survival free of VT or VF was 44.6%, 33.0% and 26.5%, and survival free of VTC was 77.3%, 72.2% and 67.1% after one, two and three years, respectively. Independent predictors of VT clusters were heart failure before ICD implantation (p = 0.033), presenting monomorphic VT (p = 0.044), EF <0.40 (p = 0.014) and inducible mVT, especially with right bundle branch block and superior axis configuration (p < 0.001). Survival free of recurrent VTCs was 50.8%, 38.1% and 19.0% after one, two and three years, respectively. Once a VTC had occurred, only 56.7%, 46.4%, 30.9% and 15.5% of patients survived and were not transplanted after one, two, three and four years, respectively. Survival was even more reduced if a VTC was associated with cardiac decompensation: 65.6% and 21.9% after one and two years, respectively.
Despite antiarrhythmic intervention, clusters of VTs occur and recur frequently in patients with DCM. They signify impaired survival, especially if they are associated with cardiac decompensation, and may be a harbinger of progressive myocardial deterioration rather than a primarily arrhythmic problem. The benefit of ICD therapy may therefore be low in these patients.
Implantable cardioverter defibrillators (ICDs) have a high success rate in terminating ventricular tachycardias (VT) and ventricular fibrillation (VF) (1–3). Some studies, however, have demonstrated impaired survival in patients with frequent episodes of ventricular tachyarrhythmias (4–8). In contrast, one study, mainly on patients with coronary artery disease (CAD), revealed that these patients did not have a worse outcome than patients without frequent VTs (9). The significance of VTCs in patients with dilated cardiomyopathy (DCM) could not be determined from these studies because of the small numbers of patients in each study population. Furthermore, definitions of clusters varied between studies, and it remains to be demonstrated which type of VT or shock cluster is of clinical significance (5–10).
Although frequent episodes are one of the most demanding clinical problems during ICD patients’ follow-up, data on management of VTCs in patients with ICDs are rare and inconsistent. Besides, existing data generally refer to few patients, and only with CAD (7–9,11–13).
Current therapeutic options for incessant and clusters of VTs are antiarrhythmic medication (11,12,14), radiofrequency ablation (15–17), anxiolytic therapy (11) and finally external cardiopulmonary support and heart transplantation for intractable incessant VTs (18,19). With the exception of amiodarone, antiarrhythmic therapy has yielded frustrating results in patients with VTCs (11,12,14).
The purpose of the present study was to analyze the incidence, permissive factors and prognostic significance of VTCs in patients with DCM.
The assumption that a VT or VF had occurred was based on ICD memory data and clinical information. The techniques used to differentiate among VT or VF, supra-VTs and oversensing have been described previously (20).
Various definitions of VTCs have been suggested (5–8,10). For sensitivity, a VTC was defined as at least three different sustained episodes of VT or VF terminated by ICD intervention during 24 h. Repetitive ineffective shocks were not included.
Patients and baseline examinations
We retrospectively analyzed data from 106 consecutive patients with DCM (diffuse left ventricular dysfunction on angiography and exclusion of CAD and other causes, ejection fraction [EF] <0.60), who received an ICD, with the capability of documenting at least four consecutive episodes, at the University of Münster (21). All patients underwent an electrophysiologic study (EPS) off antiarrhythmic drugs or on amiodarone before ICD implantation. Programmed ventricular stimulation was performed according to a protocol presented elsewhere (22).
Variables for risk prediction of VTs, VTCs, shock clusters and final events were prospectively collected in an ICD database, which included patients’ baseline characteristics and implantation data (Table 1).
Patients visited the ICD outpatient clinic routinely every three months and were encouraged to schedule additional visits if the first shock, a cluster of shocks or syncope had occurred.
Mean ± standard deviation was used for continuous variables with normal distribution. In this case, the t-test was applied to compare means. Median and range or 95% confidence interval (CI) was used to describe continuous variables without normal distribution. In this case, the Mann-Whitney U test was applied. Frequency distributions between groups of patients were tested by the chi-square test. A two-tailed p value of ≤0.05 was regarded as significant.
The probability of freedom of major events was calculated according to the Kaplan-Meier method and measured from the date of implantation, first VT or first VTC to the event (23). Differences between pairs of actuarial curves were tested by Mantel-Haenszel log-rank test. Cox regression analysis was performed on patients’ baseline characteristics to investigate and compare the influence of different variables (24). Because the inclusion of drug regimes would have violated two strong assumptions of the Cox model (the effects of different variables must be constant over time; second variables must be independent and additive), drugs were excluded from the Cox model and variables that showed interdependence were not used together in the model. Statistical analysis was done with SPSS (Statistical Package for Social Sciences, version 6.0) for Windows.
Indications for ICD implantation were VF in 48 patients (45.3%), polymorphic VT in 3 patients (2.8%), monomorphic VT in 49 patients (46.2%), nonsustained VT in 4 patients (3.8%) and syncope plus inducible VT or very low EF in 2 patients (1.9%). Prophylactic indications for ICD implantation were excluded from analysis (Tables 1 and 2). ⇓In 54 patients (50.9%), only shocks were given and in 52 patients (49.1%), additional antitachycardia pacing was programmed. The mean detection cycle length for ICD therapy was 337 ± 50 (median 300, range 290–500) ms.
The mean follow-up was 32.5 ± 23 months. Seventy-three patients (68.9%) had recurrent VTs/VF, 65 patients (61.3%) suffered shocks, 43 patients (40.6%) experienced single VTs, 30 patients (28.3%) suffered 52 clusters of VTs and 27 (25.5%) patients had shock clusters. Seventeen patients had a single VTC, 13 patients had recurrent clusters of VTs and 6 had more than 2 recurrent VTCs. A median of 19 VTs (4 to 440) occurred during clusters. A median of 4 (0 to 42) VTs were terminated with shocks, 22 clusters of VTs (42.3%) consisted of ≤3 VTs terminated by shocks, 30 (57.7%) of >3 VTs terminated by shocks. Clusters lasted for 2 (1 to 14) days, 24 (46.2%) lasted just 1 day, 28 (53.8%) more than 1 day. Most VTCs consisted of monomorphic VTs (n = 45, 86.5%), 31 clusters of VTs (59.6%) were slower than 200 beats/min, 14 (26.9%) were faster than 200 beats/min. Four VTCs (7.7%) were polymorphic VT or VF clusters and 2 clusters (3.8%) consisted of VTs with various morphologies.
Twenty-one patients (19.8%) died during follow-up; 12 patients (11.3%) underwent heart transplantation. One was transplanted immediately after a VTC and died during the procedure. Sixteen deaths (15.1%) were classified as cardiac, 3 as non-cardiac (2.8%), and the cause of death in 2 patients (1.8%) remained unknown.
Two patients died of intractable incessant tachycardia and one from a VT below the detection rate. In two patients death occurred shortly after two shocks, which terminated fast VTs promptly. In these patients death was not supposed to be caused by the tachycardias or by therapy exhaustion. Thus, in only three patients was death believed to be due to ventricular arryhthmias. In nine patients no arrhythmia was documented in the final printout after death. In these cases death was due to terminal heart failure. Actuarial freedom of major events is presented in Table 3.
Patients in NYHA class III (n = 46), patients with a history of heart failure before ICD implantation (n = 40) or with an EF below 0.40 (n = 69) had a significantly higher risk of VTC than patients in NYHA class I (n = 11) or II (n = 49, Fig. 1) without a history of heart failure (n = 66) or an EF above 0.40 (n = 66).
Mean EF at baseline was not significantly different among patients with no VTs (0.38 ± 0.12), single VTs (0.36 ± 0.13) and VTCs (0.32 ± 0.13), although a tendency was obvious.
Patients with monomorphic VTs (n = 53, patients with monomorphic nsVTs included) presenting with tachycardias had a significantly higher risk of clusters of VTs than patients who presented with VF or polymorphic VTs (n = 51): after four years 78.3% of patients with VF or polymorphic VTs survived free of a VTC, as opposed to 47.5% of patients with monomorphic VTs (p < 0.001). Patients in whom a monomorphic VT could be induced had a significantly higher risk of a VTC than those in whom no monomorphic VT or only VF could be induced. Patients in whom a monomorphic VT with superior axis (left to right superior), mostly right bundle branch block (RBBB), had a significantly higher risk of a VTC than those in whom only monomorphic VTs with inferior axis (left to right inferior) could be induced (Fig. 2). The inducible VT in patients with VTCs was slower (311 ± 91 ms) than in patients with singular events (252 ± 90 ms, p = 0.051). No other baseline variable proved to be predictive of VTCs.
In particular, there was no difference between proportional freedom of VTCs in patients with singular VTs and more than one VT before ICD implantation, and with sinus rhythm, atrial fibrillation or a history of atrial flutter.
Cox regression analysis identified prior decompensation, EF below 0.40 and monomorphic VT as presenting or inducible VT, especially with RBBB and superior axis, as independent significant risk predictors of VTCs (Table 3). NYHA class III could be substituted for a history of heart failure before implantation and low EF, but did not provide additional significant information in the Cox model.
Causes of VTCs and intervention
In 18 clusters (34.6%), no clinical cause could be identified. Sixteen clusters (30.8%) were associated with heart failure and 18 clusters with “extrinsic” causes: 10 (19.6%) with diarrhea or low potassium levels (mean 3.6, range 3.0–4.0 mmol/l on the day of the VT cluster), 2 with fever and 2 with strain (operation/traveling). Two clusters were attributed to proarrhythmia (QT-time prolongation associated with antiarrhythmic medication): one—a polimorphic VT cluster—was associated with hyperkalemia (potassium level of 7.9 mmol/l), and one was assumed to be due to discontinuation of antiarrhythmic therapy.
The risk of a recurrent VTC was high: only 50.8%, 38.1% and 19% of patients survived free of a recurrent VTC after one, two and three years, respectively (Table 2). Survival free of a recurrent shock cluster was 58.0%, 37.3% and 37.3% after one, two and three years, respectively. The recurrence risk of shock clusters tended to increase with number of VTs documented during the first VTC (3 to 6 VTs: 28.6%, 7 to 12 VTs: 35.7%, 13 to 24 VTs: 33.3%, >24 VTs: 57.9%, p = 0.074) and the duration of the first VTC (1 day: 31.6%, >1 day: 65.2%, p = 0.030).
The risk of a recurrence was not related to the first cluster being a shock cluster or a cluster with antitachycardia pacing, although physicians more often decided in favor of an antiarrhythmic therapy if a VTC was a shock cluster. Antiarrhythmic intervention performed after 24 VTCs did not show any impact on VTC recurrence: the proportion free of recurrent VTCs was 75%, 59.6%, and 47.7% after intervention and 78.4%, 50.1%, and 37.8% without intervention after one, two, and three years, respectively (N.S.).
Sixteen clusters (30.8%) occurred while patients were on beta-blockers. Survival free of shock clusters was 90.0%, 82.1%, 75.8% and 60.7% if patients were discharged on a beta-blocker after one, two, three and four years, respectively (n = 25).
Six VTCs occurred while patients were on sotalol and 16 while patients were taking amiodarone and 3 with patients taking amiodarone and a beta-blocker. The risk of a recurrent VTC was 40% on sotalol and 57.1% on amiodarone, and one cluster recurred on amiodarone and a beta-blocker. Projected survival free of VTCs after prescription of a class III antiarrhythmic was 72.9%, 62.5% and 31.3% after one, two and three years, respectively.
Patients on beta-blockers seemed to fare better than patients on class III antiarrhythmic drugs as far as risk of clusters was concerned. This may only be because patients discharged on beta-blockers were better as far as LV function and frequency of tachycardias before ICD implantation was concerned. Because of this bias, the “effect” of antiarrhythmic drugs should be viewed with caution.
A class I antiarrhythmic drug was initiated after VTCs in three patients. All three had a recurrence; one patient died during a VTC. In six patients a beta-blocker was given after a VTC; one VTC recurred (16.7%). This may suggest a benefit of beta-blockers in these circumstances in line with the favorable outcome of patients discharged on beta-blockers.
Sotalol was given empirically in five patients, three had a recurrent VTC. Amiodarone was prescribed in seven patients, of whom two had a recurrence. Amiodarone and a beta-blocker were prescribed in one patient after a VTC; he had a recurrent VTC. Amiodarone was added to the medication after shock clusters in all cases: only one patient had a recurrent shock cluster during a mean follow-up of 29.4 months. In patients in whom the cluster recurred, number of VTs (19, 5-157 before intervention and 31, 4-64 after) and number of shocks (3, 0-23 before intervention and 12, 0-42 after) were increased, mainly in patients on sotalol. The duration of clusters was reduced (2, 1-7 vs. 1, 1-2; p = 0.076).
As stated in the method section, drugs at discharge were not included in the Cox model because they violated the assumption of continuity.
In two patients radiofrequency catheter ablation was performed. Both had a recurrence of a VTC. Only 2 of 30 patients with VTCs were listed for heart transplantation. They had no recurrent VTC. Two patients died during or shortly after VTCs. In two patients, other measures (sedation, etc.) were taken. In all 16 patients in whom VTCs were associated with heart failure, therapy was intensified and 5 had a recurrent VTC; 1 died during a VTC. In three patients, prescription of an angiotensin-converting enzyme inhibitor was the only therapeutic move; one had a recurrent VTC. In 12 patients, potassium was added to the medication. In 6 patients this was the only change in therapy; 6 had a recurrent VTC.
Among all VTCs, 78.4% led to hospitalization, which was significantly associated with the number of VTs and shocks: clusters with ≤3 shocks led to hospital admission in 50% of cases, whereas those with >3 shocks led to hospitalization in all cases (p = 0.006). Clusters with up to 6 VTs (n = 7) caused hospital admission in 42.9% of cases, whereas more than 6 VTs during one cluster (n = 45) led to hospitalization in 84.4% of cases. All clusters associated with heart failure (n = 16), 88.9% associated with extrinsic causes (n = 18), and only 50% of clusters with no definable cause resulted in hospitalization (p < 0.001).
Irrespective of the number of VTs during a cluster, survival in patients with clusters of VTs was significantly impaired in comparison with patients who had only single VTs or no VTs (Fig. 3). Among patients with single VTs, 85.4% survived four years after implantation, in contrast to 45.8% of those clusters of VTs (p < 0.004). Only 26% of patients with VTCs survived four years after ICD implantation and were not transplanted, compared with 78.2% of patients with single VTs (Fig. 3, p < 0.001). A total of 56.7%, 46.4%, 30.9% and 15.5% of patients survived without transplantation one, two, three and four years after first VTC. There was no significant difference in survival or survival free of heart transplantation after clusters of VTs with few or many VTs or shocks (two-year survival after 3 to 6 VTs was 42.9%, 6 to 12 VTs was 45%, and after >12 VTs 50%). If VTCs were associated with heart failure, survival was significantly reduced (65.6% and 21.9% after one and two years, respectively) compared with clusters of VTs associated with extrinsic causes (hypokalemia, diarrhea etc., 86.9% and 69.5% after one and two years, respectively, p = 0.0467)) or no cause (88.9% and 73.3%, after one and two years, respectively, p = 0.0450).
Definitions and clinical significance of VTCs
This study is the largest study on VTs and the first on VTCs in patients with DCM. In order to be sensitive, we used the very low inclusion definition of 3 VTs or more during 24 h from Hariman et al. (25), also used by Credner et al. (9). With this definition, almost 80% of clusters led to a hospitalization and were clinically significant. Most clusters actually consisted of many VTs, with the median being 19 (range 4 to 440), and only 13.5% of clusters consisted of <6 VTs. The probability of hospitalization because of clusters was associated with the number of VTs, the number of shocks during one cluster and concurrent heart failure.
The occurrence of clusters predicted impaired survival in patients with DCM. Irrespective of the number of VTs, only 30.2% of patients survived and 15.5% survived without heart transplantation four years after the first cluster. Survival was significantly worse in patients in whom VTCs were associated with cardiac decompensation, as opposed to patients in whom an extrinsic cause or no cause of the VTC could be identified.
This result is in line with those of other studies. Wood et al. (4) found an association between the frequency of delivered therapies and shocks with total and cardiac mortality in a cohort of 401 patients, of whom 49 (12.4%) had DCM. However, frequent terminal shocks could not be excluded as a cause of this association, and details about patients and frequent episodes were not available (4). We found no significant association between frequency of shocks during clusters of VTs and survival. The study of Villacastin et al. (7) also revealed an increased death rate after multiple consecutive discharges: 7 of 25 patients (28%) with nonischemic cardiomyopathy and VT clustering died during follow-up, in contrast to 4 of 47 (8.5%) with previous myocardial infarction. However, clusters were defined as two or more discharges that occurred only a few seconds apart or were considered part of a single arrhythmic episode. This definition of VT clustering may in some cases represent an increase in defibrillation threshold or incessant VTs, rather than frequent recurrences of VTs. In a study by Fries et al. (8), total mortality and cardiac mortality was 24% to 29% in the group with and 4% in the group without short-term recurrent tachyarrhythmias (9). In their study 34 patients with clusters were presented; only 9 of these patients had a DCM and neither a survival analysis nor a comparison between different types of VTCs was presented.
The relation among heart failure, mortality and VTCs or inducible VTs requires explanation. On the one hand, it has been suggested that frequent VTs, and especially incessant VTs may cause congestive heart failure (26). Furthermore, animal (27) and echocardiographic studies (28–30) have suggested that left ventricular function, especially diastolic function and survival, may be impaired by shocks. This finding is in line with the fact that VTCs were often associated with heart failure and that heart failure was reversible in all but two cases. On the other hand, progressive scarring of the left ventricular free wall may increase susceptibility to re-entrant tachycardias (16,31–33). Therefore, clusters of VTs may be an additional powerful marker, rather than the cause, of cardiac deterioration in patients with DCM. This is supported by the fact that inducible VTs, VTCs and heart failure (NYHA) before ICD implantation were the only predictors of all-cause mortality, heart transplantation and sudden death (Table 3).
Risk prediction of VTCs
In our study, NYHA III heart failure before ICD implantation, low EF monomorphic VT as index tachycardia and the inducibility of a monomorphic VT, especially with superior axis, were the best predictors of an increased risk of VTCs in patients with DCM (Tables 2 and 3). One other study in 82 patients with ICDs, in whom 16 had VTCs, also found a significant difference in functional status and left ventricular function between patients with single episodes of VTs and patients with multiple consecutive discharges during a follow-up of 21 ± 19 months (7). In 9 of 16 patients with multiple discharges, functional class deteriorated during follow-up, which is in accordance with our finding that 30.8% of VTCs were associated with heart failure.
A predictive value of the presenting arrhythmia or the PVS for major events in patients with DCM has not yet been demonstrated, and the prognostic value of PVS in patients with DCM has been drawn into question (3,29,30). In our study, however, frequent recurrences of VTs could be anticipated with the results of PVS.
The risk of a recurrent VTC was high, irrespective of antiarrhythmic intervention. Only 19% of patients remained free of a second VTC after three years. The risk of recurrences was associated with the number of VTs during a VTC and the duration and cause of a VT cluster. Interestingly, the recurrence risk of VTCs was highest in patients in whom hypokalemia and associated conditions such as diarrhea and vomiting were noted (50%) or in patients in whom no clinical cause (52.9%) could be identified, and lowest in patients in whom VTCs were associated with cardiac decompensation (27%). This finding may be due to shorter follow-up because of high mortality in the latter group, but it also implies that patients with VTCs associated with extrinsic causes can hardly be secured against a recurrence by avoiding permissive factors.
The recurrence rate after antiarrhythmic intervention was only slightly and not significantly better than after no intervention: only 38% of clusters did not recur without intervention, as opposed to 48% with intervention. The risk of a recurrent VTC was only slightly lower if the antiarrhythmic medication was changed as opposed to maintained (47%) or escalated with a second antiarrhythmic (38%). This slight improvement was attributable to patients in whom amiodarone or a beta-blocker were prescribed after a VTC (9,13). This is in line with the findings on patients with CAD and VTCs reported by Credner et al. (9,34). The small benefit of antiarrhythmic therapy in patients with clusters of VTs is in line with a study from Dijkman, which reported that in seven of nine patients with postoperative VT storm, an average of 3.8 antiarrhythmic drugs were tried: only two patients could be stabilized with a combination of two drugs alone, three patients could be handled with overdrive stimulation, and four had to be sedated (11). Similarly, Kou et al. (14) showed that antiarrhythmic intervention offered no benefit after a single episode of VT in patients with ICD, but seemed to increase the risk of syncope during VTs. Antiarrhythmic drugs given on clinical grounds, i.e., without electrophysiologic testing, were no better in reducing the risk of recurrent tachycardias than no antiarrhythmic drugs (14). One reason may be that drugs that slow conduction may render the patient more susceptible to re-entrant tachycardias.
A retrospective analysis of complex events such as clustering of VTs is difficult from different points of view: inadequate episodes may be erroneously taken for clusters of VTs. Our analysis revealed that there was no difference between survival free of VTCs in patients with sinus rhythm and in patients with atrial fibrillation and flutter. Because atrial fibrillation is one of the strongest predictors and causes of inadequate episodes, we assume that differentiation of VTCs from other kinds of episode clustering was sufficiently achieved (35).
The differences of recurrent VTCs after interventions have to be viewed with caution, because there was no uniformity between interventions, and intervention depended on the treating physician’s preference.
The finding that VTCs are associated with impaired survival is an important observation. However, therapeutic conclusions should be drawn with caution as long as prospective confirmation is lacking.
Clusters of VTs occur frequently in ICD patients with DCM. Major predictors of VTCs were prior decompensation, EF below 0.40, and presenting or inducible monomorphic VT, especially with RBBB and superior axis configuration. In patients with DCM even clusters of as few as 3 VTs during 24 h anticipated impaired survival, as opposed to singular events, especially if they were associated with heart failure. Clusters of VTs together with monomorphic VTs, either as index or inducible tachycardia, may rather be seen as a useful additional criterion for the evaluation of a patient for a heart transplantation than as a primarily arrhythmic problem. Use of VT clustering as an additional criterion for listing should be prospectively tested.
- coronary artery disease
- dilated cardiomyopathy
- ejection fraction
- electrophysiologic study
- implantable cardioverter defibrillator
- programmed ventricular stimulation
- ventricular fibrillation
- ventricular tachycardia
- ventricular tachycardia cluster
- Received February 3, 1999.
- Revision received January 20, 2000.
- Accepted March 29, 2000.
- American College of Cardiology
- Block M, Breithardt G. Long-term follow-up and clinical results of implantable cardioverter defibrillators. In Zipes DP, Jalife J, editors. Cardiac electrophysiology: from cell to bedside. Philadelphia: WB Saunders, 1995, 1412–25.
- Wood M.A.,
- Stambler B.S.,
- Damiano R.J.,
- Greenway P.,
- Ellenbogen K.A.
- Kohno J.,
- Bhandari A.,
- Park Y.B.,
- Bolognese A.,
- Lerman R.D.,
- Cannom D.S.
- Villacastin J.,
- Almendral J.,
- Arenal A.,
- et al.
- Credner S.C.,
- Klingenheben T.,
- Mauss T.,
- Sticherling C.,
- Hohnloser S.
- Wolpert C.,
- Jung W.,
- Korte T.,
- et al.
- Tavernier R.,
- Derycke E.,
- Jordaens L.
- Strickberger S.A.,
- Man K.C.,
- Daoud E.G.,
- et al.
- Kottkamp H.,
- Hindricks G.,
- Chen X.,
- et al.
- Bänsch D.,
- Brunn J.,
- Castrucci M.,
- et al.
- Breithardt G.,
- Camm A.J.,
- Campbell R.W.F.,
- et al.
- Kaplan E.L.,
- Meier P.
- Cox D.R.
- Xie J.,
- Weil M.H.,
- Sun S.,
- Tang W.,
- Sato Y.,
- Jin X.,
- Bisera J.
- Runsio M.,
- Bergfeldt L.,
- Brodin L.A.,
- Ribeiro A.,
- Samuelsson S.,
- Rosenqvist M.
- Chen X.,
- Shenasa M.,
- Borggrefe M.,
- et al.
- Mary-Rabine L.,
- Albert A.,
- Pham T.D.,
- et al.