Article Text

Download PDFPDF

Case series
Stent and flow diverter assisted treatment of acutely ruptured brain aneurysms
  1. José E Cohen1,2,
  2. J Moshe Gomori2,
  3. Ronen R Leker3,
  4. Sergey Spektor1,
  5. Hosni Abu El Hassan1,
  6. Eyal Itshayek4
  1. 1 Department of Neurosurgery, Hadassah-Hebrew University Medical Center, Jerusalem, Israel
  2. 2 Department of Radiology, Hadassah-Hebrew University Medical Center, Jerusalem, Israel
  3. 3 Department of Neurology, Hadassah-Hebrew University Medical Center, Jerusalem, Israel
  4. 4 Department of Neurosurgery, Rabin Medical Center, Petah Tikva, Israel
  1. Correspondence to Professor José E Cohen, Department of Neurosurgery, Hadassah-Hebrew University Medical Center, Jerusalem 91120, Israel; jcohenns{at}


Objective We present our experience with stent techniques in the management of acutely ruptured aneurysms, focusing on aneurysm occlusion rates, intraprocedural complications, and late outcomes.

Methods We retrospectively reviewed the clinical records of patients treated by stent techniques during the early acute phase of aneurysmal rupture, from June 2011 to June 2016. Patients who underwent stenting for the management of unruptured aneurysms, or in a delayed fashion for a ruptured lesion, were excluded.

Results 47 patients met inclusion criteria, including 46 with subarachnoid hemorrhage (SAH). There were 27 men and 20 women, mean age 38 years (range 23–73). They harbored 71 aneurysms, including 56 treated in the acute phase. Aneurysmal dome and neck width averaged 4.7 mm (range 1.7–12.1) and 3.2 mm (range 1.5–7.1), respectively. Single stent techniques were used in 39 patients and dual stent techniques in 17. External ventricular drains (EVDs) were placed before embolization in 35 patients (92%) and after in 3. Intraprocedure thromboembolic complications due to a hyporesponse to antiplatlets in 4 patients (8.5%) were successfully managed with intra-arterial antiplatelet agents. In 45 surviving patients (96%), there was complete aneurysm occlusion at the 9–12 month follow-up in 26/29 aneurysms treated by stent-assisted coiling (90%), in 2/3 aneurysms treated by flow diverter-assisted coiling (66%), and in 19/22 aneurysms treated by flow diverter alone (86%); 42/45 patients (93%) presented with a modified Rankin Scale score of 0–2.

Conclusion Stenting techniques in ruptured aneurysms can be performed with good technical success; however, procedural thromboembolic complications related to the antiplatelet strategy merit investigation. EVD placement before stenting must be considered.

  • aneurysm
  • platelets
  • hemorrhage
  • device
  • subarachnoid
View Full Text

Statistics from


Stenting devices have greatly expanded the therapeutic spectrum of endovascular techniques for the management of brain aneurysms; however, they have largely been reserved for patients with unruptured aneurysms1 2 because of concerns regarding the use of antiplatelet therapy in cases of acute subarachnoid hemorrhage (SAH). Thromboembolic and hemorrhagic complications are reported to occur in 25% of patients with stent insertion for the management of ruptured aneurysms compared with only 4% in cases of unruptured lesions.3

Several authors have nevertheless presented their experience with the use of stents in patients with ruptured aneurysms,3–9 and a systematic review of stent assisted coiling in ruptured aneurysms concluded that the technique can be performed with a high degree of technical success, although adverse events appear more common.4 In recent reports, we found many variations in the strategies for managing these patients (table 1). Early external ventricular drain (EVD) placement was not frequently discussed, suggesting there are no special considerations in this population. There were many approaches to antiplatelet type, dose, and timing, and assessment of the effects of antiplatelet medications was rare. Given the relatively high rates of hemorrhage, thrombotic complications, and mortality that have been reported, guidelines in these areas are urgently needed.

Table 1

Recent literature presenting results from neuroendovascular management of acutely ruptured intracerebral aneurysms.

We have developed a protocol for stent-based treatment of acutely ruptured aneurysms based on a low threshold for EVD placement before stenting and preprocedure administration of aspirin and clopidogrel with platelet testing to ensure that the response falls within a safe range to limit the risk of both hemorrhage and stent-related thrombus formation. We present our experience with stent techniques in the management of acutely ruptured aneurysms, focusing on aneurysm occlusion rates, intraprocedural complications, and late outcomes in a series of patients managed according to our protocol for ventriculostomy and antiplatelet therapy.


We retrospectively reviewed our prospective hospital database to identify patients who presented to our department between April 2011 and April 2016 with acute SAH secondary to aneurysmal rupture, and underwent implantation of stents or stent-like devices during the acute phase of hemorrhage (first 3 days). Patients who underwent delayed stenting or stenting for the management of unruptured aneurysms were excluded. Our institutional review board approved the study design with a waiver of informed consent (0003–12-HMO).

Criteria for EVD placement

SAH was confirmed by CT imaging. Hunt and Hess score as well as Fisher grade were recorded. The need for an EVD was assessed based on clinical indications (mainly altered consciousness) and neuroradiological data (ventricular dilatation, hemorrhage size, risk of hydrocephalus). In acute cases where the use of stents was anticipated, an EVD was preventively placed in most patients, with the exception of those with unaltered sensorium or minor hemorrhage (Fisher grades 1–2). This strategy is based on evidence that EVD placement in patients on double antiplatelets is associated with increased risk of hemorrhagic complications.10 11 If permanent shunt placement was required, we placed the ventricular catheter of the shunt through the same tract and burr hole as those used for the original EVD. All conscious patients gave informed consent for the procedure.

Antiplatelet strategy

Before treatment, patients received dual antiplatelet therapy. The regimen includes a loading dose of 300 mg aspirin and either two doses of 150 mg clopidogrel or a single dose of 300 mg clopidogrel, orally or per nasogastric tube. In patients with EVDs, antiplatelet medications were administered 1 hour after insertion. Platelet function was tested 2 hours after 300 mg of clopidogrel was administered in all patients (Verify Now Platelet Reactivity Test; Accriva Diagnostics, San Diego, California, USA). Borderline responders received an additional 150 mg of clopidogrel, and platelet function was retested 2 hours later.

The target P2Y12 receptor inhibition level for neurostents was 60–240 P2Y12 reaction units (PRU), a range known as the ‘acceptable’ clopidogrel therapeutic window,12 13 with PRU >240 considered a clopidogrel hyporesponse. For the more thrombogenic flow diverter stents, we limited our target P2Y12 receptor inhibition to 60–207 PRUs, with PRU >207 considered a clopidogrel hyporesponse. This later limit was borrowed from the previous experience in the coronary artery stenting field.14

Patients who were hyporesponders and non-responders to clopidogrel were managed with either ticagrelor (loading dose 180 mg, posteroperative maintenance 90 mg/12 hours) or prasugrel (loading dose 40 mg, postoperative maintenance 5–10 mg/day). We checked platelet function again in all patients, on both drugs, 2 hours after administration of the loading dose, prior to the intervention. In general, hyperresponders were prescribed lower maintenance doses and aspirin was reduced (75 mg or 81 mg/day or alternating days) or discontinued.

When the decision to implant a stent was made during a diagnostic procedure in a patient who was not receiving antiplatelet therapy, a loading dose of eptifibatide (Integrilin) was administered, followed by a maintenance dose of the intravenous antiplatelet for 6 hours. Aspirin (300 mg loading dose) and clopidogrel (300–600 mg loading dose) were administered per nasogastric tube immediately after the procedure. This strategy was not applied to patients with intracerebral hemorrhage.

Neuroendovascular procedure

All procedures were performed under general anesthesia. A standard right common femoral approach with a 6 F catheter was used for all cases. Patients were fully heparinized with intravenous administration of a 70 unit/kg bolus dose and repeated bolus doses of heparin to maintain an activated clotting time of twice the normal range throughout the embolization procedure. Rotational angiography was used to accurately ascertain the anatomy of the aneurysm, afferent artery, and any vessels derived from the aneurysm neck, as well as to correctly determine the required stent size.

Neuroendovascular devices were selected at the discretion of the endovascular neurosurgeon, and included stents (LEO Baby (Balt Extrusion, Mt Montmorency, France), Solitaire (eV3, Irvine, California, USA), and Enterprise (Cordis, Miami Lakes, Florida, USA)), stent-like devices/flow diverter stents (Pipeline (Medtronic, Minneapolis, Minnesota, USA) or p64 (Phenox, Bochum, Germany)), and intra/extra-aneurysm stent-like, neck-bridging implants, such as the Solitaire stent used in a waffle cone technique or the pCONus (Phenox). In some cases requiring regular stents, and in all cases where it was indicated, the pCONus was delivered first and then the device was crossed with the coiling catheter. Some of the stent and flow diverter cases were treated with various jailing techniques.

Post-procedural medication included a daily lifetime dose of oral 100 mg acetylsalicylic acid and 75 mg clopidogrel, as well as 10 mg prasugrel or 90 mg ticagrelor twice a day for 3 months after neurostent implant or 6–9 months for flow diverter stent implant. In every case, postprocedural antiplatelet dose was modified if necessary based on monthly evaluation antiplatelet inhibition rates.

Clinical events

Clinical and angiographic events that occurred during or after the procedure were recorded. A neurological assessment (modified Rankin Scale (mRS)) was performed before the procedure, after treatment, at discharge, and at the follow-up visits.

Radiological follow-up

Preprocedure standard Towne’s as well as lateral and rotational angiographic images were acquired for all patients to determine the most suitable projection for visualization of the afferent artery, the aneurysm, and any branches that should be preserved. Control angiography was performed at the end of the procedure. The angiographic projection used to perform the embolization procedure was replicated on all follow-up angiography. Assessment of aneurysm occlusion at the end of the coiling procedures was based on the modified 4 point Raymond–Roy (mRRC) Scale1 (complete occlusion=mRRCI, neck remnant=mRRCII, residual contrast filling seen between coil loop interstices=mRRCIIIa, and contrast filling seen between coil loop interstices and aneurysm wall=mRRCIIIb). Aneurysm changes were documented on serial angiography (ie, any increase or decrease in the size of the aneurysm or neck remnants) and graded using the same scale.

Occlusion of aneurysms treated with flow diverter stent implant only was evaluated by a 5-point simple measurement of aneurysm residual after treatment (SMART) scale,15 which describes the location of residual flow within the aneurysm in the venous phase (grade 1=patent aneurysm with diffuse inflow; grade 2=residual filling of the aneurysm dome (saccular) or wall (fusiform); grade 3=residual neck only (saccular) or only intra-aneurysmal filling with former boundaries covered (fusiform); grade 4=complete occlusion). Initial angiographic follow-up was performed 3–6 months after coiling, with a further follow-up at 9–12 months.15 16


A total of 47 patients (27 men (57.5%) and 20 women (42.5%)); mean age 38.6 years (range 23–73) met  inclusion criteria. They presented 71 aneurysms, including 56 that were treated during the stage of acute hemorrhage. Patient details and procedural results are summarized in table 2. Representative cases are illustrated in figures 1–3.

Figure 1

Acutely ruptured wide necked left anterior communicating artery (AcomA) aneurysm in a middle-aged patient, Hunt and Hess 3 and Fisher grade III. The patient underwent preprocedure ventriculostomy. Following administration of a loading dose of aspirin and clopidogrel (600 mg), platelet evaluation showed 105 P2Y12 reaction units (Verify Now). Complete occlusion was achieved using a waffle cone technique with assisted coiling. The patient achieved a modified Rankin Scale score of 1 at the 90 day follow-up. (A) Head CT shows a diffuse basal subarachnoid and intraventricular hemorrhage. (B) Three-dimensional reconstruction of rotational angiogram of the left internal carotid artery (ICA) and (C) magnified angiographic image of the ICA show the irregularly shaped wide necked AcomA aneurysm. (D) Road map image obtained during intra-aneurysm stent placement. (E) Road map image obtained during first coil placement in waffle cone assisted embolization. (F) Immediate post-embolization angiographic image shows complete occlusion of the treated aneurysm.

Figure 2

This middle-aged patient presented with a ruptured wide necked basilar tip aneurysm and associated subarachnoid hemorrhage (Hunt and Hess 2, Fisher grade III). The patient underwent preprocedure ventriculoscopy. A loading dose of aspirin (300 mg) and prasugrel (60 mg) were administered. Platelet evaluation showed 126 P2Y12 reaction units. Complete occlusion of the aneurysm was achieved using Y stent assisted coiling techniques. At the 90 day follow-up, the patient achieved a modified Rankin Scale score of 1 at the neurological evaluation. (A) Head CT showing a diffuse basal subarachnoid hemorrhage and enlarged temporal horns. (B) CT angiography reconstruction, (C) three-dimensional reconstruction of a rotational angiogram, and (D) magnified view of a left vertebral artery angiogram show the irregularly shaped wide necked basilar tip aneurysm. (E) Radioscopic image obtained after Y stent assisted coiling. (F) Immediate post-embolization angiogram shows complete aneurysm occlusion of the basilar tip aneurysm.

Figure 3

This middle-aged patient presented with an acutely ruptured wide necked right middle cerebral artery (MCA) and associated subarachnoid hemorrhage (Hunt and Hess 2, Fisher II). Ventriculoscopy was not performed. A loading dose of aspirin (300 mg) and clopidogrel (600 mg) were administered. Platelet assessment revealed 230 P2Y12 reaction units. pCONus assisted coiling was performed. Intraprocedural stent thrombosis was managed with selective infusion of intra-arterial antiplatelets and clopidogrel was switched to prasugrel. The aneurysm was completely occluded and the patient had a modified Rankin Scale score of 0 at the 90 day follow-up neurological examination. (A) Reconstruction image of the rotational angiogram of the right internal carotid artery shows an irregularly shaped wide necked right MCA aneurysm. (B) Angiographic anteroposterior view. (C) Unsubtracted angiographic image of the right internal carotid artery showing the intra-aneurysm deployment of a neck bridging device (pCONus). (D) Road map image obtained during the stent assisted coiling procedure. (E) Angiographic control obtained immediately after coiling, shows thrombotic occlusion of both main MCA branches. (F) Final angiographic image obtained after selective infusion of antiplatelet agent shows no signs of residual thrombotic occlusion. The patient was discharged without any neurological deficit.

Table 2

Characteristics and outcomes in 47 patients who underwent stenting procedures for the management of 56 acutely ruptured wide necked aneurysms

Average widths of the aneurysmal dome and neck were 4.7 mm (range 1.7–12.1) and 3.2 mm (range 1.5–7.1), respectively. The anterior communicating artery (AcomA) was the most common location (21/56 treated aneurysms, 38%). All treated aneurysms (ruptured and associated) are presented in table 2. SAH was present in 46/47 patients (98%), 16 presented with associated intraventricular hemorrhage (34%), 2 with associated intracerebral hemorrhage (4%), and 1 with a purely intracerebral hemorrhage (2%).

EVD and antiplatelet management

Ventriculostomy was placed in 38 patients (81%), including 35 (74%) with preprocedural EVD placement. There were no symptomatic hemorrhages associated with ventricular drain implantation.

Only 29/47 patients (62%) were considered responders to clopidogrel and 18/47 (38%) were hypo- or non-responders (table 2). Responders presented a mean 143 PRU (range 6–240); hypo/non-responders presented a mean 258 PRU (range 241–290). Among the 29 responders, 11 required a dose reduction due to hyper-response 3–4 weeks after the intervention. Among the 18 hypo- or non-responders, 11 received ticagrelor and 7 received prasugrel. All were found to be either responders or hyper-responders. For those receiving ticagrelor, the mean PRU was 87 (range 42–126) and for prasugrel the mean PRU was 82 (range 6–130). There were no cases of intracranial hemorrhage in these patients. This is in agreement with previous reports.17 18

Device implantation

A single stent or stent-like device was used to treat 39 aneurysms (70%); a single stent and assisted coiling in 19 aneurysms, and a single flow diverter stent without coiling in 17 aneurysms. Multiple devices were required in the treatment of 17 aneurysms (30%); 12 were treated by dual stent assisted coiling and 5 by double flow diverter stent implants. Details regarding embolization techniques are presented in table 2.

Immediate angiographic evaluation of treated aneurysms

In total, 31 aneurysms were treated with stent assisted coiling (19 with a single device, 12 with double device implantation). Stents were adequately deployed according to the standard technique for use of the device. mRRCI was achieved in 22 aneurysms (71%), mRRCII in 7 (23%), and mRRCIIIa in 2 (6%) who required a second coiling intervention. Of 22 aneurysms treated with a flow diverter implant, 19 (27%) presented SMART grade 3, 11 (50%) SMART grade 2, and 5 (23%) SMART grade 1 on immediate post-implant angiogram.

Clinical and angiographic follow-up

Two patients (4%) died. One presented with a left PcomA aneurysm and deep-seated purely intracerebral hemorrhage; the second patient was admitted with a proximal PcomA aneurysm and Fisher grade 4 hemorrhage.

Of the 45 surviving patients, 42 (93%) had a final mRS score of 0–2 at the 90 day follow-up, including 27 (57%) who reached mRS 0, 13 (28%) who reached mRS 1, and 2 (4%) who reached mRS 2. Three patients (6%) presented with mRS 3 at 90 days.

Follow-up angiography at 9–12 months in 45 surviving patients (54 treated aneurysms) confirmed complete aneurysmal occlusion (mRRC I) in 26/29 aneurysms treated by stent assisted coiling (90%), and a neck remnant (mRRCIIIa) in 3 (10%). Coiling was also used in 3/25 aneurysms that had been treated with flow diverter stents; 2 of them reached mRCCI and 1 mRCCIIIa. For aneurysms treated by sole flow diverter stent implant, 19 (86%) presented complete angiographic exclusion (SMART grade 4) and 3 (14%) presented a neck remnant (SMART grade 3).


There were no postprocedure hemorrhagic complications; however, intraoperative thrombotic complications were seen in 4/47 patients (9%), all treated by stent assisted coiling. In all four patients, occlusive thrombus was seen at the implanted stent site during (n=3) or after (n=1) the coiling procedure. All four patients were receiving aspirin and clopidogrel with PRU ranging from 197 to 240 (figure 3). These cases were managed by urgent selective intra-arterial pulse spray injection of diluted eptifibatide (mean dose 6 mg), and the thrombosis resolved. None of these four patients was treated with balloons before stent implant. The immediate response of these clots to intra-arterial infusion of antiplatelet agents (IIb-IIIa glycoprotein antagonist) leads us to hypothesize that the clots were fresh and platelet-rich, supporting the hypothesis that patients were under a suboptimal antiplatelet effect. There were no clinical consequences from the thromboembolic events and no aneurysmal ruptures during rescue treatment.

There were no infections in patients with EVD placement or in the four patients who had ventriculoperitoneal  shunts placed, despite use of the same burr hole for shunting.


In this series of 47 patients with acutely ruptured aneurysms, including 46 with aneurysmatic SAH, immediate total or near total aneurysm occlusion was achieved in up to 97% of patients treated with a combination of stents and coils, and there were even higher delayed occlusion rates in patients treated with flow diverter stents in the 45 surviving patients. Despite the routine administration of double antiplatelet therapy and heparin, there were no intra- or postprocedure hemorrhages. There were intraprocedural thromboembolic events in 8.5% of patients. All thrombotic complications occurred in cases treated by stent assisted coiling and none occurred in flow diverter implant cases. These events were successfully managed by intra-arterial infusion of glycoprotein IIb/IIIa inhibitors. At late follow-up, 42/45 surviving patients (93%) presented with mRS 0–2 and 3/45 reached mRS 3. No case of in-stent thrombosis was detected. Our results compare favorably with other reports.3–9 19–21 We attribute this to our protocol, which is based on a low threshold for early (preprocedure) EVD placement and a platelet function based antiplatelet regimen.

In recent years we adopted as a general rule that when we anticipate the need for stent implant due to aneurysm configuration, we have a low threshold for preprocedure EVD placement in patients with a reduced level of consciousness and those at high risk of developing hydrocephalus based on clinical and imaging findings (Fisher grade ≥2). In this series, 80% of patients underwent EVD placement, the vast majority prior to stenting. Only 20% of patients, mainly those with preserved sensorium and minor hemorrhages, were managed without ventriculostomy. In patients who required EVD placement or another surgical intervention after stent implant, we transiently (72 hours) discontinued clopidogrel or ticagrelor and performed the intervention under aspirin, with only a mild residual effect of the second agent (platelet inhibition of <40%).

For the management of antiplatelet therapy in the present study, we initially considered a PRU range of 60–240 PRU as an acceptable target for neurostents, and we defined a more conservative range of 60–207 PRU for flow diverter stents. An ‘optimal’ clopidogrel therapeutic window of 95–207 PRU that minimizes both ischemic and hemorrhagic complications after coronary artery stenting has been proposed,14 and a recent retrospective cohort study of patients undergoing treatment of cerebral aneurysms with flow diverter stents proposed an ‘acceptable’ clopidogrel therapeutic window of 60–240 PRU that could potentially minimize both thromboembolic and hemorrhagic complications related to flow diverter implantation.12 13 To our surprise, all four intraprocedural thrombotic complications occurred in patients treated with stent assisted coiling whose PRU ranged from 197 to 240. Based on this finding, we recently established a tighter protocol, accepting a range of 90190 PRU for all stenting interventions. The impact of our modified antiplatelet evaluation protocol is under evaluation.

There is literature for flow diversion cases suggesting that high PRU values are not as relevant for thromboembolic complications.22 Historically, these devices were used in the vast majority of cases for management of unruptured aneurysms. Patients with ruptured aneurysms and acute SAH, such as those described here, are in a hypercoagulable state.8 As a result, PRU values may be more critical in this subset of patients. Additional research is needed to evaluate this issue.

Delgrado-Almandoz et al 13 found wide and dynamic variability in the initial response to clopidogrel therapy, and a frequent delayed conversion to clopidogrel hyperresponse in their prospective cohort study. They re-evaluated the antiplatelet dose every month and modified it to ‘center’ PRU levels in the optimal range. We routinely re-evaluate the antiplatelet effects monthly for the first 3 months after stenting. It is noteworthy that 11/29 patients (38%) who initially responded to clopidogrel required dose reduction during follow-up.

In contrast with our experience, among 45 patients who underwent stenting procedures for acutely ruptured aneurysms, Bechan et al 20 reported thromboembolic complications in 9 (20%) patients, with infarction in 4, and rebleeding from the treated aneurysm within 3–45 days of treatment in 5 (11%). Cai et al reported thromboembolic complications in 11.8% of patients and hemorrhagic complications in 10.8%.7

In a review of 17 articles describing outcomes in 339 patients with acute SAH who were treated by stenting techniques, Bodily et al 4 reported hemorrhagic complications in 27/339 patients (8%), including 44% due to intraprocedural rupture and a third related to EVD placement. Clinically significant thromboembolic events occurred in 6% of patients in the Bodily review, with no patient experiencing both thromboembolic and hemorrhagic events. The authors concluded that, in patients with SAH, the rate of adverse clinical events is higher for those undergoing stent procedures in comparison with those who are not stented.

The lack of hemorrhagic complications in our study may be due primarily to our heavy reliance on flow diverter stents. A second factor is routine EVD placement before the embolization procedure in patients with actual or anticipated hydrocephalus. In this way, we avoided placing ventriculostomy under double antiplatelet therapy and further reduced the overall risk of hemorrhage. If permanent shunt placement is required, we place the ventricular catheter of the shunt through the same tract and burr hole as those used for the original EVD. This last practice is considered safe and effective.23

Our series included the use of flow diverter stents without coiling in 22 patients. Unlike coil embolization, there is usually residual postprocedural filling within the aneurysm when flow diverters are used alone, with curative reconstruction occurring over a period of weeks. However, a protective effect seems to develop much earlier, since none of these patients rebled and all presented complete aneurysm exclusion at the 9–12 month follow-up. Similar results were observed by Rouchaud et al.19

Our study has limitations, including the single center retrospective design, relatively small sample, and use of diverse stenting techniques and devices. However, our primary focus was not on the efficacy of any specific technique, but on the role of antiplatelet strategy on neurosurgical management and procedure safety. Our preliminary experience supports the use of stent techniques on acute cases with a judicious adaptation preprocedure ventriculostomy placement and strict adherence to a carefully designed antiplatelet therapy policy in these patients.


In our experience, stent techniques have provided a high rate of technical success without intra- or periprocedural hemorrhage, and with good late neurological outcomes. The higher rate of intraprocedural thrombotic complications was successfully managed without clinical sequelae. We believe this good outcome is the result of careful adherence to a protocol based on a low threshold for preprocedure EVD placement and a platelet function based antiplatelet regimen.


The authors thank Shifra Fraifeld, a medical writer at the Hadassah-Hebrew University Medical Center, for her editorial assistance, critical review, and focus on data consistency during manuscript preparation.


View Abstract


  • Contributors Substantial contributions to the conception or design of the work: JEC, JMG, and RRL. Contribution to the acquisition, analysis, or interpretation of the data for the work: JEC, JMG, RRL, SS, HAEH, and EI. Drafting the work or revising it critically for important intellectual content: JEC, JMG, RRL, SS, HAEH, and EI. Final approval of the version to be published: JEC, JMG, RRL, SS, HAEH, and EI. Agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved: JEC, JMG, RRL, SS, HAEH, and EI.

  • Funding This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent Not required.

  • Ethics approval The study was approved by Hadassah-Hebrew University Medical Center institutional review board.

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

  • Data sharing statement Please direct requests for unpublished data that are relevant to this study to the corresponding author, JEC.

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.