Cerecyte Coil Trial
Angiographic Outcomes of a Prospective Randomized Trial Comparing Endovascular Coiling of Cerebral Aneurysms With Either Cerecyte or Bare Platinum Coils
Background and Purpose—We report the primary outcome of the Cerecyte Coil Trial, a randomized trial to determine whether polymer-loaded Cerecyte coils compared with Micrus bare platinum coils improved the proportion of patients with angiographic occlusion of the aneurysm at 6 months when assessed by a core laboratory. The secondary objectives were to compare the clinical outcomes and retreatment rates in the 2 groups.
Methods—Five hundred patients between 18 and 70 years of age with a ruptured or unruptured target aneurysm were randomized to be treated with either Cerecyte or bare platinum coils in 23 centers worldwide. Two hundred forty-nine patients were assigned to Cerecyte coils and 251 to bare platinum coils. Analysis was by intention to treat.
Results—Four hundred ninety-four patients were eligible for analysis. Four hundred eighty-one patients underwent coil treatment of their aneurysm, 227 patients with recently ruptured aneurysms and 254 with unruptured aneurysms. Four hundred thirty-three follow-up angiograms were assessed by the core laboratory; 127 of 215 (59%) and 118 of 218 (54%) in the Cerecyte and bare platinum groups, respectively, fulfilled the trial prespecified definition of success, namely that the treated aneurysm showed complete angiographic occlusion, had stable neck remnant, or improved in angiographic appearance compared with the end-of-treatment angiogram (P=0.17). Late retreatment was performed in 25 of 452 (5.5%) patients, 17 (7.7%) Cerecyte versus 8 (3.5%) bare platinum (P=0.064; range, 4–34 months). The clinical outcomes did not differ between the groups.
Conclusion—There was no significant difference at 6 months in the angiographic outcomes between Cerecyte coils and bare platinum coils when assessed by the core laboratory.
- angiographic occlusion
- cerebral aneurysm
- randomized controlled trials
- ruptured cerebral aneurysm
- unruptured cerebral aneurysm
The use of detachable platinum coils to treat intracranial aneurysms is now well established following the publication of the results of the International Subarachnoid Aneurysm Trial (ISAT),1 which showed improved clinical outcomes at 1 year compared with neurosurgical clipping of ruptured aneurysms. The long-term follow-up data from ISAT have shown a very low level of delayed rebleeding from treated aneurysms.2 Since coiling was introduced, there has been a drive to improve the angiographic outcomes of treated intracranial aneurysms in the belief that angiographic results represent a surrogate marker for the risk of delayed hemorrhage. This led coil manufacturers to develop modified coils in an effort to improve the angiographic outcomes. Modified coils, either coated with polyglycolic acid or containing additional poly(lactic-coglycolic acid) fiber, or coated with a hydrogel, are all described as “bioactive.” The objectives of these modifications were to increase the inflammatory response at the neck of the aneurysm, to promote organization of clot in the aneurysm and thereby aid “healing” at the neck, or to improve packing, all with the theoretical purpose of reducing the risk of aneurysm recurrence, rebleeding, and the need for retreatment.
Several case series and anecdotal reports3–7 suggested that polyglycolic acid-loaded Cerecyte coils would improve the angiographic outcome in patients undergoing endovascular coil treatment of ruptured and unruptured intracranial aneurysms. With support from Micrus Endovascular (San Jose, CA) as the sponsoring company, the Cerecyte Coil Trial (CCT) was designed as a prospective multicenter randomized controlled trial to determine whether Cerecyte coils, compared with Micrus bare platinum coils, improved the proportion of patients with stable angiographic occlusion, stable neck remnant, or improved occlusion of the aneurysm at 6 months when assessed by an independent core laboratory (ISRCTN82461286). The secondary objectives included whether there was any difference in the safety of Cerecyte coils compared with bare platinum coils and to determine if there was any significant difference in retreatment rates between the 2 groups. The clinical outcomes of the trial at posttreatment, discharge, and 5 to 7 months after treatment have been previously reported.8 This article reports the results of the blinded core laboratory analysis of the 6-month follow-up angiograms (first follow-up) obtained between 5 and 7 months after initial treatment.
Allocation to the coil group was by block randomization stratified by a ruptured or unruptured target aneurysm, maximum aneurysm lumen size, and neck size to ensure balance between the 2 coil types. The total number of patients randomized was 500 (Figure). The trial manager visited and audited center records to monitor compliance with the CCT protocol. There are a total of 6 exclusions from angiographic data analysis, 2 patients due to consent issues, one exceeded the age limit, and 3 due to stent use during the initial procedure as a “rescue measure,” which was contrary to the trial protocol.
The main purpose of the core laboratory was to classify consistently the completeness of angiographic aneurysm occlusion after endovascular coiling as dictated by the trial protocol. “Success” was defined, based on the review of the first follow-up angiogram after treatment, as the target aneurysm having complete angiographic occlusion, had improved, or there was no change (eg, remained complete occlusion or a stable neck remnant) in angiographic appearances when compared with the end of treatment angiogram. Any target aneurysm retreated during the study period was defined by the trial protocol as having failed to meet this definition of success.
A senior neuroradiologist (A.J.F.), based at Sunnybrook Health Sciences Centre, Toronto, Canada, with >30 years of experience of core laboratory work, reviewed the submitted angiograms, working independently of both the coordinating center and the sponsoring company. The core laboratory was blinded to the coil allocation as well as the imaging reports from the participating centers. The first follow-up angiogram was reviewed with reference to the original end of treatment angiogram. If a late follow-up angiogram was performed at 12 to 24 months, a second follow-up review was done with reference to the first follow-up.
Angiographic information from the CDs relating to the treatment and follow-up angiograms were stored on a password-protected Access database (Microsoft Office Access, 2003) designed by the coordinating center with guidance from the core laboratory. Aneurysms were consistently named by assessment of the aneurysm origin and its relation to the parent artery and branches. The angiographic information recorded by the core laboratory also included aneurysm characteristics, size, neck width, and an assessment of the degree of aneurysm occlusion. There was also a free-text facility for comments.
Each center, as part of the set-up arrangements, submitted to the core laboratory an anonymized DICOM-compatible CD to ensure that the images were in a readable format that could be accessed and reviewed on the PACS research workstation (IMPAX 6; Agfa HealthCare, Mortsel, Belgium). During the analysis process, a few centers submitted CDs without universal DICOM directories that could not be accessed by the research workstation and therefore were viewed on a MacBook Pro laptop using Osirix (Apple Inc, Cupertino, CA) imaging software; 42 cases were viewed using this method.
The centers were asked to use specified fiducials (Radionics Europe NV, Gent, Belgium) during the angiographic procedures to enable standardized measurements of the aneurysms to be carried out. In cases in which fiducials were not used, the core laboratory used various methods to acquire a comparable measurement: (1) from single fiducials; (2) from submitted 3-dimensional images with annotated millimeter measures; (3) using a template device developed for the International Study of Unruptured Intracranial Aneurysms (ISUIA)9; or (4) “embedded” millimeters on the workstation from the site angiography machine.10
The degree of aneurysm occlusion was categorized by the core laboratory as complete, neck remnant, or sac filling/incomplete adapted from the Raymond scale.11 An additional category (overlap coil/neck) was created and used by the core laboratory for cases in which the posttreatment images submitted by the centers showed the neck region of the aneurysm overlapped by coils or the parent artery making it difficult to assess if the aneurysm was completely occluded. Any visible aneurysm neck remnant or sac demonstrated on the posttreatment angiograms was measured as millimeter depth and width with depth used as the primary measure.
At first follow-up, the trial protocol requested that, where possible, an intra-arterial angiogram (digital subtraction angiogram) should be carried out; however, the clinical management of the patient would take precedence. Therefore, a small proportion of patients (<8%) had a follow-up MR angiogram instead of digital subtraction angiogram. This meant that for the primary outcome, the core laboratory analyzed 402 of 433 (92.8%) digital subtraction angiogram and 31 of 433 (7.2%) MR angiogram follow-ups.
The first follow-up images were viewed with reference to the procedure and posttreatment angiogram. An assessment was made by the core laboratory as to whether the aneurysm occlusion was complete, had improved, or not changed in appearance, for example, stable neck remnant as per the primary outcome. Although not specified in the trial protocol, the core laboratory classified any deterioration in angiographic appearance that measured ≥2 mm compared with the posttreatment angiogram was considered to represent a major recurrence.
A further follow-up between 12 and 24 months was requested as part of the trial but was not mandatory if it was not part of routine care in the center and if available, these data were collected.
A sample size of 250 in each arm of the study was calculated at a significance level of P=0.05 and 80% power to determine whether Cerecyte coils improve the proportion of patients with angiographic occlusion, improvement, or no change, of aneurysm occlusion at 5 to 7 months by 50%, from a rate of 75% to 87.5% as set out in the “definition of success” and allowing for a 5% attrition rate. All analyses were by intention to treat. The primary objective was compared with a one-sided probability value of 0.05 by Fisher exact test. Other comparisons were made using Fisher exact test or χ2 test with a 2-sided probability value of 0.05, as appropriate.
Compliance With Allocated Treatment
There was an excellent compliance with the treatment allocation, 222 of 247 (89.9%) receiving all Cerecyte coils and 233 of 250 (93.2%) receiving Micrus bare platinum coils. This has been reported previously.8
Angiographic Image Availability for Core Laboratory Review
Of the 494 patients who were eligible for inclusion in the CCT angiographic analysis, 13 of 494 (2.6%) patients had a failed procedure in which no coils were placed. In respect of the procedural and immediate posttreatment imaging, there were 16 of 494 (3.2%) patients in whom images could not be reviewed by the core laboratory for technical reasons. Therefore, the core laboratory reviewed 465 of 494 (94.1%) immediate posttreatment angiograms (Figure). Five patients had died after the end of treatment and before the first follow-up, as reported previously.8 There were 22 of 465 (4.7%) patients who did not have a follow-up angiogram carried out and the core laboratory was unable to review a further 5 follow-up angiograms for technical reasons. Thus, of the potentially eligible patients at first follow-up, the core laboratory reviewed 433 of 465 (93.1%) follow-up angiograms.
Immediate Posttreatment Assessment
The results of the analysis of 465 immediate posttreatment angiograms are shown in Table 1. There was no difference in the number of aneurysms that were completely occluded at the end of treatment between the 2 groups with 63 of 234 (26.9%) Cerecyte assessed as complete aneurysm occlusion compared with 59 of 231 (25.5%) in the bare platinum group. The numbers for the other occlusion categories used were also comparable; 114 of 234 (48.7%) Cerecyte versus 116 of 231 (50.2%) bare platinum were classed as having a neck remnant. Those assessed as showing sac filling or incomplete occlusion, which included incomplete luminal clotting of the coiled aneurysm, were 30 of 234 (12.8%) in the Cerecyte and 27 of 231 (11.7%) bare platinum, respectively. Aneurysms categorized as having coils overlapping the aneurysm neck were 27 of 234 (11.5%) in the Cerecyte group and 29 of 231 (12.5%) in the bare platinum group. This category is described in the “Methods” section. A sensitivity analysis regarding the core laboratory category of “overlapping coil/neck” showed that there is no statistical difference between Cerecyte and bare platinum coils, in the protocol definition of “success” at first follow-up, if these 56 cases were added to either the complete occlusion or neck remnant category at the end of the procedure.
Primary Outcome at First Angiographic Follow-Up
Of a possible 465 patients potentially eligible for angiographic assessment, the core laboratory reviewed 433 first follow-up angiograms (Figure). There were 127 of 215 (59%) aneurysms treated with Cerecyte compared with 118 of 218 (54%) treated with bare platinum coils that met the criteria for the definition of “success” as per the primary objective, that is, they were completely occluded, showed no change, or had improved in angiographic appearance when compared with the posttreatment angiogram (Table 2). There was no statistical significance between the 2 groups (P=0.17).
Angiographic Retreatment and Rebleeding Rates
In the CCT protocol, the retreatment of the target aneurysm was deemed to be a trial failure. After retreatment, any further statistical analysis of aneurysm occlusion was censored. The decision to retreat an aneurysm was made at the centers and the core laboratory was blinded to any additional treatment. The centers did not have to retreat the aneurysm using the initial coil allocation; it was at the clinician's discretion. There were 452 of 494 patients who were eligible for angiographic follow-up at the participating centers, of which 25 of 452 (5.5%) aneurysms went on to be retreated over the study period. Of the aneurysms that had been originally treated with Cerecyte coils, 17 of 222 (7.7%) were retreated versus 8 of 230 (3.5%) originally treated with bare platinum. This did not reach statistical significance (P=0.064). The median time to retreatment was 6 months with an interquartile range of 7 to 11.5 months (range, 4–34 months). Statistical analysis of these patients showed that there were no adverse procedural events associated with retreatment that worsened clinical outcome. There was no significant variance in the proportion of patients retreated between centers in different geographic regions with 11 of 187 (5.9%) and 14 of 265 (5.3%) of North American and non-North American centers, respectively.
At first follow-up angiogram, 36 of 215 (16.7%) aneurysms treated with Cerecyte were considered by the core laboratory to show major recurrence in angiographic appearance when compared with the end-of-treatment angiogram. Of these, 10 went on to be retreated. In the bare platinum group, 26 of 218 (11.9%) were considered by the core laboratory to show major recurrence in angiographic appearance when compared with the end-of-treatment angiogram (P=0.078). Of these, 4 went on to be retreated.
There was one in-hospital rebleed on the day of coiling in the Cerecyte group. This has been previously reported.8 One patient with an anterior communicating artery aneurysm who was initially treated with Cerecyte coils rebled 5 months after the initial treatment. The patient had a good clinical outcome after retreatment of the aneurysm by coiling. This was the only delayed rebleed reported in the follow-up period.
There were a total of 524 patient years of follow-up; 267 patient-years were in the ruptured group with the single delayed rebleed. No late rebleeding was observed in the unruptured aneurysm group in 257 patient-years of follow-up.
The primary hypothesis of the CCT, that Cerecyte coils produce improved angiographic occlusion rates at 6 months after endovascular treatment compared with bare platinum coils, has not been proven. There was no statistical significance between the 2 groups on any of the angiographic outcome measures. The rates of stable occlusion and retreatment did not differ significantly between the 2 coil types.
The number of aneurysms that were retreated was slightly more frequent in the Cerecyte group but this was not statistically significant. The overall retreatment rate over the follow-up period was low (5.5%).
Two recent randomized trials comparing “bioactive coils” with bare platinum coils have recently been reported: HydroCoil Endovascular Aneurysm Occlusion and Packing Study12 (HELPS, 499 patients) and the Matrix And Platinum Science trial13 (MAPS, 626 patients). The primary objectives of these 2 trials differed markedly when compared with CCT and each other. HELPS used a composite end point of major recurrence, defined as an aneurysm that theoretically could be retreated at 18 months plus procedure-related death or morbidity resulting in failure to obtain follow-up. Notably, for every 10 aneurysms in HELPS that “theoretically could have been” retreated, only one actually was retreated; this disconnect calls into question the clinical relevance of the “major recurrence” as used in HELPS. In CCT, of the 52 cases classified by the core laboratory as showing major recurrence (≥2 mm deterioration), only 14 were retreated.
The MAPS trial used target aneurysm recurrence as an end point, which was defined as rebleed, reintervention (retreatment), or death. In comparison, CCT used the angiographic outcome of completely occluded, stable, or improved angiographic appearance as the “definition of success.” The CCT “success” at first follow-up was 59% and 54% for Cerecyte and bare platinum, respectively, with HELPS reporting their primary end point as 64% and 60% for HydroCoil and bare platinum, respectively. However, these 2 trials used very different definitions of success and thus direct comparison is difficult or impossible and probably unwise. In HELPS, the subgroup of patients with recently ruptured aneurysms showed a modest benefit for HydroCoil over bare platinum. The MAPS trial showed no significant difference on the primary outcome between the Matrix and bare platinum coils. These results would suggest that overall there is no evidence, or only weak evidence with respect to the HydroCoil, that “bioactive coils” make any significant difference in angiographic outcome when compared with bare platinum coils.
The decision to retreat an aneurysm and by what method after coiling is to a large extent subjective.14 It depends on a variety of factors but particularly the opinion of the operator balancing treatment risks and the perceived risks of a delayed hemorrhage. The operator in making this decision may not be blinded to coil type in randomized studies. The aneurysm retreatment rates in CCT are similar to HELPS and lower than MAPS; however, this may be due to differing proportions of participants enrolled in North America. The MAPS data indicate that there was a substantial variation in reintervention rates depending on geographical region. In North America, 21% of ruptured aneurysms and 10.6% of unruptured aneurysms were retreated compared with 6.9% and 5.7%, respectively, for the rest of the world.13 Analysis of the CCT retreatment rate along geographical regions showed no such variations, although there was a smaller proportion of US centers, and the majority of North American enrollment in CCT was from Canada.
The other major difference between CCT and the other 2 trials was the use of stent-assisted coiling. This was a protocol violation in CCT but was allowed in HELPS and MAPS. Despite the increased use of stent-assisted coiling, the angiographic results between the studies appear broadly similar.
In CCT, during the 2-year follow-up period after treatment, there has been one delayed hemorrhage, which occurred in the Cerecyte group in a total of 524 patient-years of follow-up (267 years in the ruptured group). There was no delayed hemorrhage in the follow-up period of patients with ruptured aneurysms in HELPS. The recently published Barrow Ruptured Aneurysm Trial15 of randomized clipping and coiling reported no rebleeding from the coiled aneurysms in the first year of follow-up, and none had been reported up to 3 years after coiling. These rates of delayed recurrent subarachnoid hemorrhage do not differ from those observed in earlier studies, ISAT2 and the Cerebral Aneurysm Rerupture After Treatment (CARAT)16 study, which reported rates of 0.5% per patient follow-up year. This is at a level that would suggest that the observed clinical benefit of coiling compared with clipping would not be lost even over a substantial number of years. The rate of recurrent hemorrhage in CCT does not differ from those observed in these other studies.
Limitations of the CCT
In a small proportion of cases, the quality and completeness of the images sent to the core laboratory for review made assessment difficult or sometimes impossible. The lack of interobserver and intraobserver validation could also be seen as a major limitation of this trial. However, the objective of CCT set out in the protocol was not to provide a comparison with other studies or core laboratory analysis but to compare 2 different coil types whilst blinded to allocation. Having a single reviewer enabled us to have consistency in the reviews of the angiograms and thus we believe that the results have internal validity with respect to the questions we sought to address. Cloft et al17 and Tollard et al18 have examined intra- and interobserver variability in trials such as this and found poor interobserver correlations. They concluded that a simple dichotomous outcome or 3-point scale was the optimum surrogate marker in this type of trial. We have described in detail the core laboratory methodology to highlight the difficulties faced, which were not only technical in nature, but also semantic, for example, what is an overlapping coil/neck? We acknowledge that there are difficulties for a core laboratory when trying to determine the result on the end-of-treatment images. We recognized that some posttreatment angiogram images submitted did not show a view of the coiled aneurysm region clearing the neck from the coil mass and parent arteries. This made it difficult from these posttreatment images to know if an aneurysm was completely occluded. Although this made no difference between groups for the initial analysis, for the full assessment of whether they had improved could not be done but did show a balance of complete or neck remnants in both groups (Tables 3 and 4). It is likely that many studies and operators classify such cases as complete aneurysm occlusion, which may or may not be an accurate assessment. The failure to obtain follow-up angiography in 32 patients (7%) would not have affected the conclusion. Given the equivalence of primary outcomes between the allocated groups, even if the trial had complete follow-up or continued to a substantially larger sample size, it is extremely unlikely that any difference on the angiographic outcomes would have emerged, a futility situation.
The lack of complete 12- to 24-month follow-up is a potential limitation and the possibility of very late aneurysm reopening exists. The review of the additional data will be the subject of a further report.
The CCT along with HELPS and the MAPS trial have shown that modified or “bioactive coils” have either a marginal or no impact on angiographic outcome. This questions the rationale for their use, particularly at significantly increased financial cost. The risk of delayed hemorrhage after coiling remains exceptionally low. The clinical outcomes of coiling reported in both ruptured and unruptured aneurysms are significantly better than those seen in ISAT1 or reported from recent studies of aneurysm clipping.15 These studies provide the potential to carry out high-quality systematic reviews which, along with other randomized evidence already published, will provide reliable and valuable information on the current procedural and late outcomes for patients undergoing coil treatment of both ruptured and unruptured aneurysms. The ISAT data2 suggested that at long-term follow-up, new aneurysm formation or pre-existing aneurysms were more likely to bleed compared with a coiled aneurysm. Fundamental questions remain for the speciality: Is angiographic occlusion a reliable surrogate for lack of subsequent subarachnoid hemorrhage? Is the quest for a perfect angiographic result valid? What is an acceptable cost, both financial to healthcare systems and clinically for the patient, in terms of a potentially added procedural risk to achieve a perfect image?
In this article, we have reported the core laboratory assessments of angiographic occlusion; further work is needed to examine how this compares with the CCT operator assessment of the angiographic occlusion and the angiographic outcomes at 12- to 24-month follow-up.
Cerecyte coils compared with bare platinum coils did not improve 6-month angiographic outcomes in this randomized trial. Both coils types were safe and effective in preventing aneurysm rupture during the follow-up period, and retreatment rates were low for both devices.
Sources of Funding
The Cerecyte Coil Trial was funded by an unrestricted research grant from Micrus Endovascular, San Jose, CA, to the Oxford Radcliffe Hospitals National Health Service Trust, Oxford, UK, to coordinate the trial. Sunnybrook Health Sciences, Toronto, Canada, received funding from Micrus Endovascular to undertake the core laboratory case reviews and storage of CD imaging. Micrus Endovascular was not involved in the design, collection, analysis, or interpretation of the clinical or imaging data.
A.M. had a consulting and advisory agreement with Micrus and received travel/accommodation support for CCT investigator meetings and data presentation: travel and hotel support for attendance at scientific meetings. A.J.F. has received travel/accommodation support from Micrus for CCT investigator meeting and received an honorarium per case from Boston Scientific as chair of the MAPS trial clinical events committee. M.S. received travel/accommodation support from Micrus for site visits as part of the study protocol and meetings for study set-up and data presentation. S.C. received travel/accommodation support from Micrus Endovascular to present interim data at Val D'Isere (2008 and 2009). D.F.K. has received research support from Micrus Endovascular for a clinical trial; grants/grants pending from ev3, NFocus, Sequent, MicroVention, Cook, ArthroCare, and CareFusion; and payment for development of educational presentations, CareFusion and ev3. Travel/accommodation meeting support was received from MicroVention.
A.M. was Chief Investigator, M.S. was the Trial Coordinator, and A.J.F. conducted the Core Laboratory. We are grateful to all the CCT investigators and coordinators for their support and hard work in enrolling patients, completing data collection forms, and providing the core laboratory with CD imaging. We acknowledge the support of all the patients and their relatives who agreed to participate in the study. We also thank Julia Yarnold and Alison Clarke for their assistance at the coordinating center in Oxford and Narmatha Tharmaseelan and Sue Crisp at the core laboratory in Toronto.
- Received March 26, 2012.
- Revision received June 3, 2012.
- Accepted June 18, 2012.
- © 2012 American Heart Association, Inc.
- Molyneux AJ,
- Kerr RSC,
- Yu L-M,
- Clarke M,
- Sneade M,
- Yarnold JA,
- et al
- Molyneux AJ,
- Kerr RSC,
- Birks J,
- Ramzi N,
- Sneade M,
- Rischmiller J
- Bendszuz M,
- Solymosi L
- Bendszus M,
- Bartsch AJ,
- Solymosi L
- Butteriss D,
- Gholkar A,
- Mitra D,
- Birchall D,
- Jayakrishnan V
- Coley S,
- Sneade M,
- Clarke A,
- Mehta Z,
- Kallmes D,
- Cerkirge S,
- et al
- Forbes G,
- Fox AJ,
- Houston J III.,
- Wiebers DO,
- Torner J
- Fox AJ,
- Millar J,
- Raymond J,
- Pryor JC,
- Roy D,
- Tomlinson GA,
- et al
- Roy D,
- Milot G,
- Raymond J
Matrix and Platinum Science (MAPS). Abstract. Eighth Annual Meeting of the Society of Neurointerventional Surgery (SNIS); Colorado Springs, CO; July 25–28, 2011.
- Daugherty WP,
- Ehteshami Rad A,
- White JB,
- Meyers PM,
- Lanzino GL,
- Cloft HJ,
- et al
- Johnston SC,
- Dowd CF,
- Higashida RT,
- Lawton MT,
- Duckwiler GR,
- Gress DR
- Cloft HJ,
- Kaufmann T,
- Kallmes DF
- Tollard E,
- Darsaut TE,
- Bing F,
- Guilbert F,
- Gevry G,
- Raymond J