Article Text


Original research
The occurrence of neointimal hyperplasia after flow-diverter implantation is associated with cardiovascular risks factors and the stent design
  1. Jildaz Caroff1,
  2. Marta Iacobucci1,
  3. Aymeric Rouchaud2,3,
  4. Cristian Mihalea1,
  5. Fernando Mota de Carvalho4,
  6. Victor Erwin D Jocson5,
  7. Vanessa Chalumeau1,
  8. Valerio Da Ros6,
  9. Robert M King7,
  10. Rose Arslanian7,
  11. Léon Ikka1,
  12. Nidhal Ben Achour1,
  13. Jacques Moret1,
  14. Laurent Spelle1
  1. 1Department of Interventional Neuroradiology, NEURI Centre, Bicêtre Hospital, Le Kremlin-Bicêtre, France
  2. 2Department of Radiology, Limoges University Hospital, Limoges, France
  3. 3CNRS, XLIM, UMR 7252, University of Limoges, Limoges, France
  4. 4Department of Radiology, Neuroradiology Section, Hospital Antonio Prudente, Fortaleza, Brazil
  5. 5The Medical City, Makati Medical Centre, Asian Hospital and Medical Centre, Manila, Philippines
  6. 6Dipartimento di Diagnostica per Immagini e Radiologia Interventistica, Policlinico Universitario di Roma ’Tor Vergata', Rome, Italy
  7. 7New England Center for Stroke Research, Department of Radiology, University of Massachusetts Medical School, Worcester, Massachusetts, USA
  1. Correspondence to Dr Jildaz Caroff, Department of Interventional Neuroradiology, NEURI Center, Bicêtre Hospital, Le Kremlin-Bicêtre 94270, France; jildaz.caroff{at}


Background Neo-intimal hyperplasia (NIH) is frequently observed after flow-diverter stent (FDS) implantation. Although mostly asymptomatic, this vascular response can sometimes lead to delayed ischemic strokes. This study intended to evaluate the factors potentially influencing the rates of NIH following FDS treatment.

Material and Methods All aneurysm treatments performed with a Pipeline embolization device (PED) or a SILK stent from May 2011 to May 2015 were collected in a prospectively maintained database. Patient demographics, clinical, and angiographic outcomes including both digital subtraction angiography and C-arm cone-beam CT were registered. Two blind reviewers rated the presence of NIH on a binary scale (present/absent).

Results From 148 patients, 63 datasets were available for analysis. Inter-reader agreement was excellent (Kappa=0.88). NIH was positively correlated with smoking, dyslipidemia, and high blood pressure, but not with aneurysm characteristics. At early follow-up (<12 months), NIH was more frequently associated with the use of the SILK stent (68%) rather than the PED (38%): P<0.02. At long-term follow-up, the NIH rate in the total population dropped from 55% to 26% with no more significant difference between the two stents. The complete occlusion rate as seen in early follow-up was higher in the SILK group with 76% vs 65% but without statistical significance (P=0.4).

Conclusion NIH is a dual-vessel reaction after FDS implant. When planning a treatment in locations at risk of ischemic complications if severe NIH would occur, then the stent design should be considered. However, minimal NIH might also be needed as it is involved in aneurysm healing. Before treatment patients should be recommended best medical management of their cardiovascular risks factors to prevent an excessive NIH reaction.

  • aneurysm
  • angiography
  • flow diverter
  • stenosis

Statistics from


Neointimal hyperplasia (NIH) is a vessel-wall reaction frequently depicted after flow-diverter stent (FDS) treatment for intracranial aneurysm,1–3 and may lead to delayed ischemic strokes.2–5 C-arm cone-beam CT (VasoCT) has been validated for measurement of in-stent tissue growth.1

The goal of this study is to evaluate if the FDS design and cardiovascular risks factors affect the vessel-wall reaction during follow-up using high resolution cone-beam CT.6

Material and methods


Between May 2011 and May 2015, 148 consecutive patients treated with a Pipeline embolization device (PED) (Medtronic, Irvine, California, USA) or a SILK stent (Balt, Montmorency, France) were retrospectively reviewed to assess the incidence of NIH using conventional digital subtraction angiography (DSA) and cone-beam CT multiplanar reconstructions. National Commission for Data Processing and Freedom approval and informed written consents were obtained.

Endovascular treatment

All endovascular treatments were performed using a biplane, flat-panel detector angiography system (Allura Xper 20/10; Philips Healthcare, Best, The Netherlands). Conventional 2D DSA and VasoCT were routinely performed on the day of the procedure to confirm good wall apposition, and in the follow-up angiograms at 6, 18, and 42 months after treatment to check for any in-stent stenosis. Non-routine additional angiograms were obtained at the discretion of the physician and when unexpected clinical or angiographic findings were seen during follow-up. Informed consent was obtained from each patient.

Endovascular procedures were performed with the patient under general anesthesia and systemic heparinization. Patients were premedicated with clopidrogrel (75 mg daily) and aspirin (160 mg daily) starting 7 days before the procedure. Thrombocyte inhibition levels were confirmed using the VerifyNow P2Y12 assay (Accumetrics, San Diego, CA, USA) and four hematological laboratory tests. In cases of resistance a loading dose of plavix was given prior to intervention. The choice of stent used for treatment was dependent on the operator’s choice based on anatomy and personal experience. The dual antiplatelet therapy was maintained for 3 months then discontinued, while aspirin was continued until 12 months after treatment. In cases of severe NIH the dual antiplatelet therapy was maintained until the following DSA.

Images analysis

A set of images was selected by a qualified interventional radiologist. It contained the best DSA and unsubtracted images with VasoCT images at the time of treatment, first and last follow-up. Blinded from the clinical data, two experienced interventional neuroradiologists rated the NIH on a 3-point scale described as follows: 0=absence of NIH; 1=minor NIH (<25%); and 2=mild and severe NIH (>25%). From this the scale was also analyzed as binary, describing the presence or the absence of NIH. Differences in image interpretation were then resolved by consensus.

Inclusion criteria

The inclusion criteria in this study were: the use of a PED or SILK FDS; and at least one available early follow-up angiogram performed between 3 and 12 months (VasoCT quality had to be sufficient for NIH analysis).

Cardiovascular risks factors

Smoking was defined by the status at the time of the treatment. High blood pressure and dyslipidemia were considered if the patient was at least treated with one drug for those conditions.

Statistical analysis

Statistical analysis was performed using R version 3.4.1 (R Foundation for Statistical Computing, Vienna, Austria). NIH evaluation reported by two raters and agreement was assessed using Cohens' weighted Kappa (κ) scores for both binary and 3-point scales. Comparison of categorical variables was obtained with Fisher’s exact test. We performed univariate analyses using the Mann–Whitney U test for the comparison of continuous variables. A P-value of P<0.05 was considered statistically significant.


Patients' demographics

From 148 patients, 63 datasets were available for analysis. Basic demographic data and aneurysmal characteristics are summarized in tables 1 and 2. No significant difference was observed between the PED and SILK groups.

Table 1

Demographic data and aneurysmal characteristics

Table 2

Aneurysms' location

Imaging analysis

Inter-reader agreement was excellent (κ=0.89 and 0.81 for binary and 3-points scales respectively).

At early follow-up (<12 months) NIH was more frequently associated with the usage of the SILK stent 68% rather than the PED (38% for both readers) with a P value =0.02 (table 3).

Table 3

Correlation between NIH and demographic data, aneurysm characteristics, and stent type

When present, the two readers rated the NIH as minimal (<25% stenosis) in 33% for the SILK population and 50% for the PED population.

The presence of NIH was not correlated with gender, age, or aneurysms' characteristics. But it was significantly associated with cardiovascular risk factors (smoking, high blood pressure, dyslipidemia), all with P<0.01.

At last follow-up (average 29 months) the NIH presence rate in the total population dropped from 55% (35/63) to 26% (11/35), with no significant difference between the two stents (5/18 for the SILK and 4/17 for the PED; P=1).

Occlusion rates

At early follow-up (≤12 months) the complete occlusion rate was higher in the SILK group with 76% (26/34) vs 65% (19/29) in the PED group without reaching significance (P=0.4).

Clinical outcome

In only one case were some clinical symptoms linked to NIH, which consisted of transient amaurosis episodes 6 months after the implant. DSA and VasoCT revealed severe NIH inside a SILK stent at the level of the origin of the patent ophthalmic artery. Symptoms resolved with the reintroduction of double antiplatelet therapy, but the 40-month follow-up demonstrated the persistence of mild NIH (figure 1).

Figure 1

A sexagenarian patient was treated with a SILK stent for an ophthalmic aneurysm, (A) postoperative DSA. This patient was admitted 6 months later because of episodes of transient amaurosis. (B) DSA and (C) VasoCT depicted a severe NIH at the level of the origin of the patent ophthalmic artery (arrow). Symptoms resolved with the reintroduction of double antiplatelet therapy, but the 40-month follow-up demonstrated the persistence of NIH (D).



FDS have proven highly effective in treating intracranial brain aneurysms.7 The vascular response (ie, neo-intimal and neo-endothelial reaction) after stent implant contribute to the long-term healing of the aneurysms.8 9 The NIH results from smooth muscle cell (SMC) migration from the media to the intima following an endothelium injury during FDS implantation. The exact biological mechanisms underlying this wall reaction are still unclear and probably different from the atherosclerotic vessel reaction after coronary stent implant.10

Although frequently depicted and usually asymptomatic, an excessive NIH reaction can sometimes lead to delayed ischemic complications.3 5 11

Cardiovascular risks factors’ impact

In our study, smoking, high bloodpressure, and dyslipidemia were individually strongly correlated to the presence of NIH. In a large surveillance angiographic analysis after coronary stenting, Cassese et al demonstrated that patient-related predictors of restenosis were diabetes mellitus and a history of bypass surgery.12 The population and the underlying disease in our study are different from the population in the large cardiovascular trials.

In 2015 the AHA/ASA recommended that patients with unruptured intracranial aneurysms should be counseled regarding the importance of smoking cessation and undergo treatment for hypertension.13 The data from our study indicates that cardiovascular risks factor management could also be relevant for excessive NIH prevention.

Stent design effect

We found in our study that at early follow-up the SILK stent was more frequently associated with the occurrence of NIH than the PED (68% vs 38% respectively) and that it also tends to be more severe.

The results of our study cannot directly compare with previous reports because here we systematically used VasoCT to evaluate NIH, with this tool having a sensibility closer to histology than DSA to depict NIH.1 We also decided to target and analyze any minimal amount of NIH detectable.

Nevertheless, the results of this work focused on NIH detection seem to be in accordance with the results of larger clinical studies on FDS. In a SILK meta-analysis of 285 patients, authors found an intra-stent stenosis in 5% of patients during follow-up,14 whereas in a large meta-analysis of 1556 patients treated with PED it was described in less than 1.5% of cases.15

In a single center study John et al2 depicted intra-stent tissue growth in 41% of cases using PED, whereas with the SILK stent Essbaiheen et al16 reported a 44% rate and Lubicz et al17 up to 57%.

Cohen et al3 also reported a study focusing on in-stent proliferation rates after FDS implantation. Authors reported similar rates of in-stent stenosis for PED and SILK (39% and 38% respectively) but there were more cases of severe grade stenosis in the SILK patients.

We know that different materials used in stent manufacture can induce different host responses leading to different levels of local inflammation. Those inflammatory cells will produce platelet-derived growth factor, a key signal in smooth muscle cell proliferation and subsequent NIH.18 Nickel ions are the most allergenic metal ions known and have proven to cause severe inflammation when tissue concentrations are high.19 Surface treatment is essential for the optimum bioperformance of nitinol, and the nickel content in the oxide layer over the FDS will impact the levels of nickel ion release.20 Aguilar Pérez et al published a series of patients treated with the p64 FDS made from nitinol and platinum similarly to the SILK stent.21 They reported a 29.6% rate of NIH. The difference from our results may come from the fact that we based our analysis on VasoCT which has proven to be able to depict smaller amounts of NIH than DSA.1

Long-term follow-up

At the late regression phase of NIH, a remodeling process has been described with a change of the SMC phenotypes from synthetic to contractile. The remodeling of the NIH is also related to a significant reduction in the SMCs' proteoglycan content with no change in SMC density.22

In the long-term follow-up of our study, a 29% decrease in the total NIH rate was observed and the significant difference in the NIH rates between the two different FDS was no longer present.

As previously reported2 3 16 no patient without evidence of NIH at the early follow-up developed NIH in the long-term follow-up. Also, no worsening of the vessel lumen caliber reduction due to NIH between early and late follow-up was depicted.

Dual impact of NIH

In the vast majority of cases, the NIH does not lead to any clinical complication.2 21 However, ischemic stroke is a rare but potential complication in cases of flow restriction related to severe stenosis and occlusions of covered branches.3 5 17

NIH is also very likely to be a key phenomenon in the physiopathology of aneurysm healing after treatment with a FDS.3 The neo-endothelialization and subsequent occlusion of the aneurysm depend on an underlying SMCs substrate.23 It is possible that a higher amount of NIH, especially at the level of the neck, would lead to higher occlusion rates. In our study, although it did not reach significance (probably due to the small population) we found a tendency toward more occlusion rates associated with the use of the SILK stent.

Antiplatelet therapies

Ticagrelor is a novel antiplatelet agent now being used in some centers (including ours) in association with aspirin. The pharmacodynamic effects of ticagrelor are more prompt, potent, and predictable compared with those of clopidogrel.24 It has been linked to inhibition of SMC proliferation and, hence, could be associated with reduced NIH.25 The impact of ticagrelor on NIH and occlusion rates should be the subject of another study.


Our study limitations include the retrospective design and single-center experience. The inclusion criteria required quality VasoCT follow-up imaging which can be difficult to obtain because of motion artifact with a patient under local anesthesia.

We encountered some loss of follow-up and perhaps patients with NIH could have benefited more frequently from DSA follow-up as this could have participated in over-estimating the NIH rate.

Another bias could be attributed to the FDS type selection. We believe that the SILK offers a better navigability so operators could be more prone to using it in case of tortuous anatomy which could have affected the FDS correct apposition and the vessel wall reaction by the end.


NIH is a dual vessel reaction after FDS implant. When planning a treatment strategy involving a FDS in locations at risk of ischemic complications if severe NIH would occur, then the stent design should be taken into account. However, minimal NIH might also be needed as it is involved in aneurysm healing.

Before treatment, patients should be recommended best medical management of their cardiovascular risks factors to prevent an excessive NIH reaction.


View Abstract


  • Contributors JC: designed the study, analyzed and interpreted the data, drafted the manuscript, and approved the final manuscript. All authors: data set creation and approved the final manuscript

  • Funding Partially supported by research grants from the Fulbright Program, the Philippe Foundation, and the French Society of Radiology (Bourse de Recherche SFR-CERF).

  • Competing interests JC: Educational scholarships from Medtronic Neurovascular and Microvention/Terumo. JM: consultant for Medtronic, Microvention, Stryker, and Balt. LS: consultant for Stryker, MicroVention, Medtronic, Balt.

  • Patient consent Obtained.

  • Ethics approval CNIL Paris.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement Data available upon request to the corresponding author.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.