Author + information
- Received May 12, 2008
- Revision received July 10, 2008
- Accepted August 14, 2008
- Published online October 1, 2008.
- Michael C. John, MPH⁎,
- Rainer Wessely, MD‡,
- Adnan Kastrati, MD‡,
- Albert Schömig, MD‡,
- Michael Joner, MD†,
- Mayu Uchihashi, BA⁎,
- Johanna Crimins, BA⁎,
- Scott Lajoie⁎,
- Frank D. Kolodgie, PhD†,
- Herman K. Gold, MD⁎,
- Renu Virmani, MD† and
- Aloke V. Finn, MD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Aloke V. Finn, Department of Internal Medicine, Emory University School of Medicine, Emory Crawford Long Hospital, 550 Peachtree Street NE, Atlanta, Georgia 30308
The abstract was presented at the American Heart Association Scientific Sessions, November 12–16, 2006, Chicago, Illinois.
Objectives We compared the healing and inflammatory responses of polymer-free bare-metal stents (BMS), polymer-free sirolimus-eluting stents (SES) and polymer-free sirolimus-eluting stents plus estradiol (SES+ED) to Cypher drug-eluting stents (CDES) in a rabbit model of overlapping stent placement.
Background Inflammatory responses to polymers and delayed healing remain important safety issues associated with CDES. Whether nonpolymeric stents that elute sirolimus or sirolimus and estradiol provoke less inflammation and heal better is unknown.
Methods Twenty-eight rabbits received 2 overlapping stents in each iliac artery: SES, SES+ED, BMS, or CDES, and vessels were harvested at 28 days for histology and scanning electron microscopy.
Results Although similar at nonoverlapping segments, neointimal thickness within the overlap site of CDES was significantly less than in SES, SES+ED, and BMS (0.07 ± 0.04 mm vs. 0.16 ± 0.03 mm, 0.14 ± 0.03 mm, and 0.15 ± 0.03 mm, p < 0.0001). Endothelialization was greater in SES, SES+ED, and BMS compared with CDES in nonoverlapping sections (80.0 ± 5.0% vs. 95.3 ± 5.0%, 97.5 ± 2.5%, and 96.7 ± 3.8%; p = 0.0028) and overlapping sections (85.8 ± 2.9% vs. 90.8 ± 6.3%, 89.2 ± 6.3%, and 48.3 ± 2.9%; p < 0.0001). The number of luminal eosinophils was also less in overlapping sections of SES, SES+ED, and BMS versus CDES but was similar in nonoverlapping sections.
Conclusions Polymer-free stents coated with SES or SES+ED result in less robust neointimal suppression but markedly improved arterial healing compared with CDES in the rabbit model.
The polymer-based, sirolimus-eluting Cypher (Cordis Corp., Warren, New Jersey) drug-eluting stent (CDES) has become a common treatment for patients with symptomatic coronary artery disease undergoing percutaneous coronary intervention. However, the long-term safety of CDES has been called into question owing to concerns about late stent thrombosis secondary to impaired arterial healing characterized by delayed re-endothelialization and persistence of fibrin (1,2). Emerging evidence suggests that drug delivery polymers may play an important role in the pathophysiology of impaired healing by provoking inflammatory cell infiltration and/or causing long-term drug sequestration within the arterial wall (3–5). Moreover, the versatility of CDES for interventions involving multiple stents such as overlap or bifurcations is likely limited by local arterial toxicity, which occurs when drug and polymer concentrations are substantially increased (6–8).
The ISAR (individualizable drug-eluting stent [DES] system to abrogate restenosis) polymer-free DES was recently developed to allow for dose-adjustable, on-site coating of stents with a microporous surface without the obligate use of a polymer (9). Stents coated with sirolimus using this technology demonstrate efficacy equal to that of polymeric DES in reducing restenosis in humans at 9-month follow-up (10), but whether they heal more efficiently and have more versatility than CDES is unknown. Additionally, estradiol, which in animal models prevents smooth muscle proliferation and enhances endothelialization after vascular injury (11), may be able to improve healing if delivered on the stent in combination with sirolimus.
The aim of this study was to compare the healing properties of polymer-free bare-metal stents (BMS) and polymer-free sirolimus-eluting stents (SES) or sirolimus-eluting stents plus estradiol (SES+ED) to those of commercially available CDES in an established rabbit model of overlapping stent placement.
Stent coating protocol
The ISAR polymer-free stent platform used to prepare the SES and SES+ED DES consists of a 316-liter, stainless steel, microporous stent in a disposable coating cartridge (Yukon DES, Translumina, Hechingen, Germany). Coating was performed as previously described (9) with either SES alone (1% concentration) or with SES+ED, both at 1%. The release kinetics for the SES- and SES+ED-coated ISAR stents have been published (9,12). The 1% SES solution used in this study was selected on the basis of results from a dose-finding study in humans (13). The ISAR BMS (Yukon, Translumina) served as control stents.
The protocol was approved by the Institutional Animal Care and Use Committee of the Massachusetts General Hospital, and all experiments were conducted according to the National Institutes of Health “Guide for the Care and Use of Laboratory Animals.” Twenty-five New Zealand White rabbits (3.7 to 4.1 kg) successfully underwent stent implantation under anesthesia with inhaled isoflurane for 28-day analysis by light microscopy (LM) or scanning electron microscopy (SEM). Using standard fluoroscopy, balloon injury was performed in both iliac arteries. Subsequently, animals were randomly allocated to receive a pair of overlapping SES, SES+ED, BMS, or CDES in 1 iliac artery and a different pair of overlapping stents in the other iliac artery. All stents were deployed at the area of injury at nominal pressure. Mean percentage of overlap was 53.1 ± 6.9% for CDES, and 49.0 ± 2.6%, 51.1 ± 3.2%, and 52.7 ± 4.7% for SES, SES+ED and BMS, respectively (p = NS). The CDES were 3.0 × 18 mm. All ISAR stents were 3.0 × 16 mm. The strut thickness of the ISAR stents is 87 μm, and that of CDES is 140 μm.
An additional 5 animals received balloon injury and single stent placement in each iliac artery, as described in preceding text, for organoid culture (OC).
All animals were pre-treated with aspirin, 40 mg orally (approximately 10 mg/kg), 24 h before stenting, and aspirin was continued until sacrifice. In addition, heparin (150 IU/kg) was administered intra-arterially before catheterization procedures.
Tissue harvest and processing
Twenty-eight days after stenting, animals assigned for LM or SEM were re-anesthetized, and follow-up angiography was performed to verify stent patency and position. Euthanasia was accomplished with an overdose of Beuthanasia-D given intravenously under deep anesthesia. Stented arteries were perfusion fixed in-situ with 10% buffered formalin. Arteries were prepared for LM as previously described (3).
Animals with stents for OC were reanesthetized in the same way at 14 days after stent placement.
Arteries were perfused in-situ with Ringer's lactate solution and placed in culture conditions.
Scanning electron microscopy
Stented arteries were harvested at 28 days, longitudinally cut, and analyzed by SEM, as previously described (14).
The 14-day stent implants using SES, SES+ED, and BMS were used for OC. (The CDES were not tested owing to limited availability and cost of devices.) Intact specimens were maintained in serum-free Dulbecco's modified Eagle's medium kept in humidified 95% air/5% CO2 at 37°C. The media from each artery were collected after 48 hours, concentrated, and analyzed using a Cytokine Array kit (RayBio Human Cytokine Array III, RayBiotech, Inc., Norcross, Georgia). Membrane blots were imaged with a Chemi-doc (Bio-Rad, Hercules, California) and analyzed using image analysis software (Quantity One, Bio-Rad).
All arterial segments were examined blindly. Computerized planimetry was performed on all stented arterial sections, as previously described (14). Percent luminal stenosis was calculated with the following formula: neointimal area divided by the external elastic lamina times 100. Fibrin deposition, the number of giant cells around stent struts, and heterophil/eosinophil infiltration were quantitated and expressed as the number of struts surrounded by fibrin, the percent of struts surrounded by giant cells, and the total number of heterophils/eosinophils lining the lumen of the vessel, respectively.
Light microscopy and SEM data are expressed as mean ± SD. For multiple group comparisons among arteries stented with different devices, we utilized a 1-way analysis of variance. If the variance ratio test (F-test) was significant, a more detailed post hoc analysis of differences between groups of arteries was made using a Tukey-Kramer honest significance difference test. For comparisons of overlapping to nonoverlapping segments within groups and for analysis of cytokine array between groups, a paired Student t test was utilized. The normality of distribution was tested using the Wilk-Shapiro test. A p value < 0.05 was considered significant.
All 30 animals (SES, n = 14 arteries [8 LM, 3 SEM, 3 OC]; SES+ED, n = 14 arteries [8 LM, 3 SEM, 3 OC]; bare-metal ISAR DES [BMS], n = 18 [10 LM, 4 SEM, 4 OC]; CDES, n = 13 arteries [10 LM and 3 SEM]) were in good health for the duration of the study, and all stents were widely patent at follow-up angiography without incidence of dissection, thrombosis, stent migration, or aneurysm formation.
Histology and SEM
Scanning electron microscopy of 28-day stents demonstrated almost complete (i.e., 90% to 100%) endothelial coverage of arteries implanted with nonpolymer-based ISAR BMS within the overlapping and nonoverlapping segments of the stent, irrespective of the presence of drug coating (Table 1,Fig. 1A). The CDES demonstrated significantly decreased endothelial coverage compared with SES, SES+ED, and BMS within the nonoverlapping segments (80.0 ± 5.0% vs. 95.3 ± 5.0%, 97.5 ± 2.5%, and 96.7 ± 3.8%, respectively; p = 0.0028) and the overlapping segments (48.3 ± 2.9% vs. 85.8 ± 2.9%, 90.8 ± 6.3%, and 89.2 ± 6.3%, respectively; p < 0.0001) (Fig. 1B). In contrast to all nonpolymer-based ISAR stents, the degree of endothelialization was significantly less in overlapped than in nonoverlapped CDES sections (p = 0.018).
In arteries implanted with CDES, bare stent wires were occasionally observed within the nonoverlapping segment and more frequently at the overlap site along with adherent white cells and platelets (Fig. 1A). Moreover, the partially endothelialized luminal surface between stent struts was characterized by loose endothelial cell junctions and scattered inflammatory cells. In contrast, SES, SES+ED, and BMS showed a confluent layer of endothelial cells throughout the stent and little or no evidence of inflammatory cell infiltration (Fig. 1A insets).
The percentage of struts surrounded by fibrin in the nonoverlapping segments of CDES was considerably increased in comparison with SES, SES+ED, and BMS (25.2 ± 15.3% vs. 3.2 ± 3.9%, 6.4 ± 6.7%, and 2.8 ± 5.0%, respectively; p < 0.0002). Overlapping segments of CDES only exacerbated this trend, with nearly one-half of the stent struts in the overlap area exhibiting some degree of fibrin deposition versus SES, SES+ED, and BMS (47.8 ± 20.9% vs. 11.5 ± 10.8%, 15.7 ± 12.9%, and 7.90 ± 9.77%; p < 0.0001, respectively). Fibrin deposition was significantly greater in overlapped CDES than in nonoverlapped CDES (p = 0.0019) and in SES+ED (p = 0.02) (Fig. 2 and Table 1). Remarkably, there was no difference between BMS and polymer-free DES.
The number of heterophils/eosinophils on the luminal surface of SES, SES+ED, and BMS was negligible (≤1 cell) in the nonoverlapping segment and only marginally higher in CDES (p = NS) (Table 1). Overlapping stents resulted in a small nonsignificant increase in these luminal inflammatory cells in the SES, SES+ED, and BMS (6.5 ± 0.57, 4.2 ± 3.6, and 2.3 ± 1.4 cells, respectively), whereas overlapping CDES showed a substantial rise in the number of luminal eosinophils to 34.7 ± 17.6 cells (p = 0.0001 vs. nonoverlapped CDES). In overlapped CDES sections, eosinophils were also significantly higher than were seen in overlapped SES, SES+ED, and BMS (p < 0.0001).
Giant cells were observed throughout the stent in all stent groups, with significant increases between nonoverlapping and overlapping sections observed only in the case of CDES (p = 0.007) (Table 1, Fig. 2).
There were no significant differences in neointimal thickness within nonoverlapping areas between CDES and SES, SES+ED, and BMS (0.09 ± 0.03 mm vs. 0.10 ± 0.03 mm, 0.07 ± 0.01 mm, and 0.09 ± 0.03 mm; p = NS), although the CDES group showed decreased neointimal thickening versus the SES, SES+ED, and BMS groups at the overlap site (Table 1, Fig. 3). The SES+ED tended to have less neointimal thickness and percent stenosis as compared with SES alone within both the nonoverlapping and overlapping segments (Table 1).
The internal elastic lamina and external elastic lamina areas were greater in the CDES group than in any other stent group, both in overlapping and nonoverlapping areas. These indexes of positive remodeling did not translate into changes in medial area, however, as no differences in this measure were noted among the groups (Table 1).
There was no difference in mean injury score between stent groups at the nonoverlapping segment (SES, 0.78 ± 0.26; SES+ED, 0.81 ± 0.30; BMS, 0.98 ± 0.26; CDES, 1.02 ± 0.15; p = NS). However, mean injury scores were significantly different at the overlap site between CDES and SES (1.15 ± 0.13 vs. 0.74 ± 0.24; p = 0.001), CDES and BMS (1.15 ± 0.13 vs. 0.93 ± 0.32; p = 0.05), and SES and SES+ED (0.74 ± 0.24 vs. 1.00 ± 0.22; p = 0.03) (Table 1). When corrected for the degree of injury, differences in intimal thickness remained significant for CDES versus SES and CDES versus BMS but not for CDES versus SES+ED (Table 1).
Using OC and protein array technology, we investigated the effect of SES, SES+ED, and BMS on cytokine and growth factor expression in arteries 14 days after stent placement. Relative protein expression of SES (n = 3 arteries) and SES+ED (n = 3 arteries) were normalized to the corresponding BMS (n = 4 arteries) and expressed as a ratio (SES/BMS or SES+ED/BMS). Figure 4 illustrates the relative expression levels of various proteins involved in vascular healing and inflammation: interleukin (IL)-2, IL-4, IL-8, IL-10, macrophage-derived chemokine, macrophage colony-stimulating factor, vascular endothelial growth factor (VEGF), and insulin-like growth factor. Paired statistical analyses showed no difference between SES/BMS and SES+ED/BMS in expression levels of IL-2, -4, -8, and -10. However, VEGF and macrophage colony-stimulating factor exhibited significantly increased protein expression in SES+ED/BMS versus SES/BMS (p ≤ 0.05), and there was a strong trend toward up-regulation of macrophage-derived chemokine and VEGF-1 in SES+ED/BMS compared with SES/BMS (p = 0.07 for both).
Although polymer-based CDES have been shown to be effective in preventing in-stent restenosis in humans, this approach is associated with delayed healing and increased risk of late stent thrombosis (1,2). We have shown for the first time that nonpolymeric microporous ISAR stents that eluted either SES or SES+ED have improved endothelial coverage and reduced inflammatory responses when compared with polymer-based CDES in a rabbit iliac model of overlapping stent placement. However, intimal suppression was not as robust as with CDES, especially at overlapping segments. These findings underscore the important role that polymers play in both neointimal suppression and impaired healing.
In our previous work in the rabbit as well as in the current study, we examined the histologic response to overlapping CDES placement and found evidence of dose-dependent arterial impairment of healing when comparing overlapping with nonoverlapping segments (3). In both studies, overlapping segments demonstrated significant increases in heterophils/eosinophils, fibrin, and impaired endothelialization. In contrast, nonpolymeric ISAR DES showed no increases in these variables with the exception of percent fibrin for SES+ED. Neointimal formation was similar when comparing nonoverlapping sections of CDES with those of nonpolymeric stents, although CDES clearly suppressed intimal thickness more effectively at overlapping segments, even when corrected for the injury score, except in the case of SES+ED. The lack of effect of CDES on reducing intimal thickness at nonoverlapping segments versus BMS was unexpected, given that others have shown efficacy in this model when comparing CDES with its corresponding BMS (BxVelocity, Cordis Corp., Warren, New Jersey) (15) and might be due to the significantly thinner strut thickness of the ISAR stents (87 μm) compared with the CDES (140 μm), a factor known to be important in determining restenosis rates (16).
In contrast to the findings of Wessely et al. (9), who implanted SES-coated ISAR single stents in the porcine coronary model, we found significantly decreased endothelial coverage in CDES compared with polymer-free DES and BMS in the nonoverlapping segment. Part of the explanation for this may lie in the methods used to detect endothelialization (i.e., SEM in this study vs. LM). Scanning electron microscopy allows en-face examination of the whole stent surface, whereas LM is limited to cross-sectional analysis. Another explanation may lie in the differential response to injury of the porcine versus rabbit models, with the pig re-endothelializing at a quicker pace than the rabbit (17).
Pathophysiology of delayed healing
Current-generation CDES elute sirolimus from a nonbiodegradable polymer matrix composed of polyethylene-co-vinyl acetate (PEVA) and poly n-butyl methacrylate (PBMA), which remains permanently in contact with the arterial wall. Although this polymer has been reported to provoke chronic eosinophilic infiltration of the arterial wall suggestive of hypersensitivity, a causal relationship between inflammation and thrombosis has been observed in a very small number of patients (2,4). We observed substantial increases in luminal heterophils/eosinophils with CDES as compared with nonpolymeric DES and BMS only at overlapping segments, although we found delayed endothelialization and increases in fibrin in both overlapped and nonoverlapped segments. This finding suggests that increases in inflammatory cells cannot by themselves be responsible for delayed healing. We observed this same effect in human pathology specimens from patients who formerly received either polymer-based DES or BMS implanted at similar times. Although there was increased fibrin accumulation and significantly less endothelialization in DES versus BMS, inflammation scores were similar (2). This observation suggests that other mechanisms such as drug and drug-release kinetics, which are affected by polymeric systems, may also be important in the pathophysiology of delayed healing.
Although the rationale behind polymeric DES platforms is to achieve controlled drug release, and therefore impressive reductions in intimal formation, there are important risks associated with this strategy resulting from local cellular toxicity and/or drug overdose. Even though the total dose of sirolimus is more than 2 times higher on the SES and SES+ED than on the CDES (CDES = 1.4 μg sirolimus/mm2 vs. ISAR 1% SES stent = 3.1 μg/mm2), amounts well above that needed to effectively inhibit smooth muscle and endothelial cells proliferation (18,19), we demonstrated significantly improved healing in ISAR stents as compared with CDES. Wessely et al. (9,15) demonstrated that ISAR rapamycin-coated stents elute two-thirds of their drug in the first week and nearly all of the loaded dose by 21 days, whereas CDES stents elute only 68.4% at 28 days. Moreover, in the pig coronary artery, tissue levels of rapamycin peak at 3 days with the ISAR system and at 14 days with the CDES. Aside from the biological reaction to the polymer itself, our data underscore the important role release kinetics play with regard to both the healing characteristics and the antirestenotic efficacy of these 2 stent systems.
An alternative approach to accelerate healing is local codelivery of compounds that are known to inhibit smooth muscle proliferation along with others known to accelerate endothelial regrowth after vascular injury, such as the hormone estradiol (11,20). Prior work in animal models has shown that estradiol promotes stent re-endothelialization and neointimal inhibition, although Adriaenssens et al. (12) recently reported no improvement in neointimal inhibition when rapamycin-coated stents were layered with 17-β-estradiol in humans. Our data demonstrate no additional effect on endothelial coverage within overlapping and nonoverlapping segments of SES+ED as compared with SES, although measures of neointimal formation (i.e., intimal thickness and percent stenosis) tended to be lower with SES+ED, especially in the overlapping segment when corrected for injury. The explanation as to why we saw no differences in endothelial coverage may lie in the time point chosen. Given the relatively short time course of healing in the rabbit, with near-complete endothelialization in all ISAR stents at 28 days, any improvements in coverage would be difficult to detect in this model at this time point.
Organoid culture of ISAR DES
Cytokines and growth factors play a major role in the cascade of biological events leading to vascular inflammation and healing. We found that the addition of ED to the SES-coated nonpolymeric stent caused a trend toward up-regulation in each of the 8 proteins examined, a trend that reached statistical significance for macrophage colony-stimulating factor and VEGF. Macrophage colony-stimulating factor is known to facilitate monocyte survival, monocyte-to-macrophage conversion, and macrophage proliferation (21). Monocyte-derived cells are thought to play an important role in wound healing, especially in formation of highly vascular scaffold tissue (22), although their role in the process of re-endothelialization after vascular injury has not been well characterized. Vascular endothelial growth factor is the most important angiogenic factor involved in migration and proliferation of endothelial cells, with higher levels correlating with healing versus quiescent endothelial surfaces (23,24). Our data mirror that of Concina et al. (25), who also demonstrated up-regulation of VEGF (along with enhanced endothelialization) in ED-treated rat after carotid balloon injury. In aggregate, these data suggest that the addition of ED to SES may have had a favorable effect on vascular healing, which may have been detectable had an earlier time point been chosen for pathology analysis.
Impairment of vascular healing is one of the major drawbacks of current polymer-based DES. Our data point to the role polymers play in provoking local inflammation as well as allowing for controlled release of drug. Both mechanisms likely play a role in the pathophysiology of delayed healing that underlies all cases of late stent thrombosis at autopsy. The ISAR system demonstrates the advantages of a nonpolymeric system from the standpoint of inflammation and arterial healing but also shows its drawbacks in terms of antirestenotic efficacy, because drug release cannot be tightly controlled. Although we did not see any difference in percent stenosis in the nonoverlapping segments of CDES and the ISAR DES, the differential response of these 2 systems has been seen in clinical trials. In the randomized trial of a nonpolymer-based rapamycin-eluting stent versus a polymer-based paclitaxel-eluting stent for the reduction of late lumen loss, the late loss of the ISAR stent was 0.48 mm, far below the 0.18 mm reported for the Cypher in the SIRIUS (Sirolimus-Eluting Stent in De Novo Native Coronary Lesions) trial (10,26).
Given this limitation, our data suggest that the ISAR system may still be a valid addition to current DES platforms, because it combines anti-restenotic efficacy with less interference with vascular healing.
This investigation of overlapping DES in an animal model using normal nonatherosclerotic arteries may have underestimated the effects of high doses of drug and polymer on the arterial wall, because atherosclerosis tends to intensify inflammatory responses. Current animal models used in the assessment of stents are limited in their ability to replicate human conditions, although results in the rabbit model have generally been representative of human responses, albeit with a different time course of healing.
We have shown for the first time that the nonpolymeric sirolimus-coated ISAR stents implanted in the rabbit iliac result in less impairment in arterial healing but also less effective antirestenotic efficacy when compared with CDES at 28 days. The CDES provoked significantly more fibrin, inflammation, and endothelial damage than did the SES and SES+ED DES coated using the ISAR system. The differences observed are likely due to a combination of reduced vascular inflammation and the shorter release kinetics and duration of tissue drug deposition of the ISAR compared with the CDES systems. As concerns mount regarding late stent thrombosis, the balance may lie in favor of nonpolymeric systems that allow for less robust intimal suppression but greater healing.
This study was supported, in part, by a grant from CVPath Institute, Inc., a nonprofit institute in Gaithersburg, Maryland, and by the Bavarian Research Foundation, Munich, Germany. Stents were provided by Translumina, Hechingen, Germany. Dr. Virmani receives company-sponsored research support from Medtronic AVE, Guidant, Abbott, W.L. Gore, Atrium Medical Corporation, Boston Scientific, NDC Cordis Corporation, Novartis, Orbus Medical Technologies, Biotronik, BioSensors, Alchimer, and Terumo. Dr. Vrimani is also a consultant to Medtronic AVE, Guidant, Abbott Laboratories, W.L. Gore, Terumo, and Volcano Therapeutics, Inc.
- Abbreviations and Acronyms
- bare-metal stent(s)
- Cypher drug-eluting stent(s)
- drug-eluting stent(s)
- individualizable drug-eluting stent system to abrogate restenosis
- light microscopy
- organoid culture
- scanning electron microscopy
- sirolimus-eluting stent(s)
- sirolimus-eluting stent(s) plus estradiol
- vascular endothelial growth factor
- Received May 12, 2008.
- Revision received July 10, 2008.
- Accepted August 14, 2008.
- American College of Cardiology Foundation
- Pfisterer M.,
- Brunner-La Rocca H.P.,
- et al.
- Joner M.,
- Finn A.V.,
- Farb A.,
- et al.
- Finn A.V.,
- Kolodgie F.D.,
- Harnek J.,
- et al.
- Virmani R.,
- Guagliumi G.,
- Farb A.,
- et al.
- Carter A.J.,
- Aggarwal M.,
- Kopia G.A.,
- et al.
- Hoye A.,
- Iakovou I.,
- Ge L.,
- et al.
- Kuchulakanti P.K.,
- Chu W.W.,
- Torguson R.,
- et al.
- Takano M.,
- Ohba T.,
- Inami S.,
- Seimiya K.,
- Sakai S.,
- Mizuno K.
- Wessely R.,
- Hausleiter J.,
- Michaelis C.,
- et al.
- Mehilli J.,
- Kastrati A.,
- Wessely R.,
- et al.
- Brouchet L.,
- Krust A.,
- Dupont S.,
- Chambon P.,
- Bayard F.,
- Arnal J.F.
- Adriaenssens T.,
- Mehilli J.,
- Wessely R.,
- et al.
- Hausleiter J.,
- Kastrati A.,
- Wessely R.,
- et al.
- Pache J.,
- Kastrati A.,
- Mehilli J.,
- et al.
- Finn A.V.,
- Nakazawa G.,
- Joner M.,
- et al.
- Marx S.O.,
- Jayaraman T.,
- Go L.O.,
- Marks A.R.
- Vinals F.,
- Chambard J.C.,
- Pouyssegur J.