Article Text

Contemporary endovascular and open aneurysm treatment in the era of flow diversion
  1. Marcus D Mazur,
  2. Philipp Taussky,
  3. Min S Park,
  4. William T Couldwell
  1. Department of Neurosurgery, Clinical Neurosciences Center, University of Utah, Salt Lake, Utah, USA
  1. Correspondence to Dr William T Couldwell, Department of Neurosurgery, University of Utah, Salt Lake, Utah 84132, USA; neuropub{at}


Clinical outcomes have improved considerably over the last decade for patients with ruptured and unruptured aneurysms. Modern endovascular techniques, such as flow diversion, are associated with high aneurysm occlusion rates and have become a popular treatment modality for many types of aneurysms. However, the safety and effectiveness of flow diversion has not yet been established in trials comparing it with traditional aneurysm treatments. Moreover, there are some types of aneurysms that may not be appropriate for endovascular coiling, such as wide-necked aneurysms located at branch points of major vessels, large saccular aneurysms with multiple efferent arteries, dolichoectatic aneurysms, large aneurysms with mass effect, when there are technical complications with endovascular treatment, when patients cannot tolerate or have contraindications to antiplatelet therapy or in the setting of a subarachnoid haemorrhage. For these cases, open cerebrovascular surgery remains important. This review provides a discussion on the current trends and evidence for both flow diversion and open cerebrovascular surgery for complex aneurysms that may not be suitable for coiling. We emphasise a continued important role for surgical treatment in certain situations.

  • open cerebrovascular surger
  • endovascula
  • flow diversion
  • aneurysm
  • uptured
  • unruptured
  • subarachnoid hemorrhage
  • bypass
  • clipping
  • revascularization

Statistics from

Improvement in patient outcomes after subarachnoid haemorrhage

The last decade has seen a considerable improvement in clinical outcomes after aneurysmal subarachnoid haemorrhage (SAH). Mortality has decreased, functional outcomes have improved and there is a greater likelihood of discharge to home.1 Clinical outcomes are better for patients with aneurysms treated with surgical clipping and for those undergoing endovascular coiling.1 A recent analysis of the US Nationwide Inpatient Sample reported that inpatient mortality decreased from 32.2% in 2002 to 22.2% in 2010, while discharge to home increased from 28.5% to 40.8%.1 Similar trends have been observed in several other countries as well.2 The reasons for improved clinical and functional outcomes are likely multifactorial: protocol-driven treatment, the availability of advanced invasive monitoring and complex procedures to treat SAH-associated vasospasm (eg, angioplasty), experienced vascular and endovascular surgeons and specialised nursing.3 Better recognition and aggressive management of the sequelae of aneurysmal SAH, such as hydrocephalus, vasospasm and delayed ischaemic deficits, are also likely important. Modern endovascular treatment may further improve clinical outcomes as new modalities, such as flow diversion, are potentially approved for aneurysms that, historically, have been challenging to treat using open surgical techniques or coiling.

Advent and efficacy of flow diversion

Endovascular flow diversion is a paradigm shift for aneurysm treatment. Whereas coil embolisation aims to completely occlude the aneurysm at the time of treatment, flow diversion is an entirely endoluminal treatment that requires no device to enter the aneurysm itself. Tightly braided flow-diverting stents (FDSs) have low porosity and provide 30%–35% wall coverage, which directs blood flow along the parent artery and away from the aneurysm. Aneurysm occlusion occurs in a delayed fashion, over several weeks or months, as endothelialisation occurs, neointimal growth proceeds across the aneurysm neck and the parent artery remodels around the FDS.

Early experience with flow diversion came from the treatment of large and giant wide-necked, unruptured aneurysms of the internal carotid artery (ICA); the pipeline embolization device (PED, Covidien, Plymouth, Minnesota, USA) was approved by the US Food and Drug Administration for this application. Before the development of flow diversion techniques, large symptomatic aneurysms of the carotid artery could be managed with an extracranial-to-intracranial (EC-IC) carotid artery bypass or, if tolerated, carotid sacrifice. These technically challenging operations (eg, the petrous-to-supraclinoid interpositional bypass) may require working in a deep, narrow corridor, retracting the temporal lobe and arresting blood flow temporarily, all of which increase the risk of morbidity. Results are operator dependent, and significant experience is required to achieve high bypass patency rates with acceptable morbidity. Similarly, coil embolisation for large and giant wide-necked aneurysms is associated with high rates of recurrence (40%–60%) and rebleeding (1.9 %/year).4 On the other hand, the Pipeline for Uncoilable or Failed Aneurysms (PUFs) trial demonstrated high occlusion rates after flow diversion with low morbidity.5 Complete occlusion had occurred in 86% of patients 1 year after PED placement, a rate that increased to 93% at 3 years.5 6 Adverse neurological events occurred in 5.6% of patients, and there were no aneurysm recurrences.5

Since the positive results of the PUFs trial were described, the use of flow diversion has increased dramatically. One single-centre, retrospective study in USA reported that 88% of unruptured ICA aneurysms were treated with flow diversion.7 Moreover, flow diversion use has expanded to off-label conditions, such as ruptured or blister-like aneurysms and aneurysms located beyond the ICA and in the posterior circulation. An international retrospective study (IntrePED) evaluating the safety of FDSs in a real-world setting reported that 8.4% of aneurysms treated were already ruptured. In addition to those treated in the ICA, aneurysms were also located in the posterior circulation in 10.2% of patients, the middle cerebral artery (MCA) in 4.7%, the anterior communicating artery (AcomA) in 1.3% and the anterior cerebral artery (ACA) in 1.0%.8 Even with wide variation in aneurysms, the results of IntrePED and other studies demonstrate that flow diversion has higher rates of complete occlusion, similar rates of adverse events and lower rates of recurrence when compared with coil embolisation.9

It is important to note that the safety profile for FDS is still being determined. Although results of initial non-randomised studies of FDS use have generally been favourable, a recent randomised controlled trial comparing flow diversion with standard treatment options (observation, coil embolisation, parent vessel occlusion or surgical clipping) suggests that flow diversion may not be as safe and effective as initially thought.10 The Flow Diversion in the Treatment of Intracranial Aneurysm Trial (FIAT), conducted in Canada, was stopped prematurely because a high proportion of patients allocated to flow diversion did not reach the primary outcome of angiographic occlusion at 3–12 months and were dead or dependent at 3 months because of procedural complications. These complications included device migration, arterial rupture, delayed aneurysm rupture, distal intraparenchymal haemorrhage and delayed primary vessel occlusions. More high-quality studies are needed to establish the types of aneurysms that should be treated with flow diversion. Although the popularity of flow diversion is increasing, coil embolisation remains a suitable treatment for many saccular aneurysms and future trials are needed to compare directly the safety and efficacy of these two modalities.11

Many neurointerventionalists reserve FDS for aneurysms that are difficult to treat by endovascular coil embolisation or open surgical techniques (eg, cavernous ICA aneurysms, fusiform aneurysms of the anterior circulation and dorsal variant blister aneurysms). However, it will be important that larger studies with longer follow-up are performed to establish the safety profile and long-term outcomes of using flow diversion in various settings. For example, despite early hopes for a safer alternative treatment modality, poor results with high morbidity and mortality for large dolichoectatic aneurysms of the basilar artery were documented.12 These aneurysms often required the placement of multiple stents, and complications arose from brainstem infarctions caused by covering perforating arteries or delayed rupture. Infarction may result from direct coverage of the perforator ostia by the tines of the device, by migration of acute thrombus or by neointimal growth across the ostia.13 Clinical deterioration of patients with brainstem compression from giant, partially thrombosed posterior circulation aneurysms treated using flow diversion has been reported.14 15 In contrast to these findings, results appear more promising for non-dolichoectatic aneurysms of the posterior circulation.13 16 Nevertheless, the potential risks of flow diversion must be considered when new indications for FDS are investigated, such as aneurysms located at bifurcation sites (ie, basilar apex or MCA bifurcation).

Thromboembolic and haemorrhagic risks of flow diversion

Antiplatelet therapy is required with flow diversion to prevent thromboembolic complications and in-stent stenosis until vessel remodelling occurs and epithelial growth covers the stent lumen. Thromboembolic or haemorrhagic complications occur in 4% of cases of flow diversion.17 Such complications include ischaemic stroke due to device occlusion, thromboemboli, ipsilateral intraparenchymal haemorrhages or aneurysm rupture. Others have noted morbidity and mortality rates between 8% and 10%.18

At our institution, patients are treated with clopidogrel (75 mg/day) and aspirin (325 mg/day) for 1 week before FDS deployment. The P2Y12 level is checked to confirm adequate antiplatelet affect. For patients in whom these agents are therapeutic (P2Y12<192 PRU, defined by our institution laboratory), clopidogrel and aspirin are continued for 6 months, then aspirin monotherapy is continued for life. For patients who are clopidogrel resistant (P2Y12≥192 PRU), prasugrel is administered. However, platelet testing is controversial and its benefits have yet to be proven definitively.19 Thromboembolic complications, however, have been reported in clopidogrel-resistant patients,20 and altering drug therapy before deployment is relatively simple. Thus, we continue to check P2Y12 levels in our patients.

The mandatory antiplatelet therapy required presents an increased risk of haemorrhagic complications. A recent review found that up to 80% of patients who experience a delayed aneurysm rupture or intraparenchymal haemorrhage have a poor clinical outcome or death.21 Nearly half of patients with a delayed rupture had giant aneurysms.21 The exact mechanism behind the delayed haemorrhagic complications has not been established, and haemorrhages have occurred in vascular territories remote from the FDS. However, most events occur within the first 30 days of treatment and are on the ipsilateral vascular territory as the FDS.

Role of open cerebrovascular surgery

Although the utilisation of endovascular techniques is increasing, national data indicate that surgical clipping is still performed on ruptured and unruptured aneurysms, particularly in younger patients, although at a lower rate than in years past. Recent data from the US Nationwide Inpatient Sample (2002–2010) showed that 59.1% of patients with aneurysmal SAH underwent surgical clipping compared with 40.9% that received endovascular treatments,1 but recent trends suggest that the proportion of aneurysms treated with surgical clipping is much lower today. Some authors consider surgical clipping to be more durable and believe it carries a lower risk of recurrence and rerupture than endovascular treatment (after coiling of aneurysms),22 although the perceived risk and management of incompletely treated but stable aneurysms varies widely.23 At some high-volume academic centres with low complication rates, surgical clipping remains the preferred treatment for some patients.24 Clipping is also more likely to be recommended than endovascular treatment in many countries outside USA and Western Europe, which may reflect disparities in available resources.25 Nevertheless, the trend is towards the increased use of endovascular treatment of aneurysms worldwide.

Despite these trends, open cerebrovascular surgery retains an important role. Open microsurgery may be necessary for many challenging clinical scenarios, such as wide-necked aneurysms located at branch points of major vessels (eg, basilar apex), large saccular aneurysms with multiple efferent arteries, dolichoectatic aneurysms, large aneurysms with mass effect, when there are technical complications with endovascular treatment, when patients cannot tolerate or have contraindications to antiplatelet therapy or in the setting of a SAH. Aneurysm size, location and morphology may also prohibit endovascular treatment. Aneurysms located on small or very tortuous vessels may not be accessible via microcatheters. Blister aneurysms, pseudoaneurysms and small saccular aneurysms may also have fragile domes or may simply be too small for endoluminal treatment, and open surgery may be the safest treatment option.

Complex clip constructs, parent artery occlusion with bypass, vessel reconstruction and aneurysm trapping with interposition grafting have been described.26–28 These surgical procedures can be technically demanding and are typically performed in high-volume cerebrovascular centres. Importantly, given the growth of endovascular treatment, aneurysms referred for surgical treatment are usually more complex than those treated in historical series.

Direct cerebral bypass operations are necessary for many complex aneurysms; such revascularisation techniques include EC-IC bypass with aneurysm trapping or with distal or proximal artery occlusion, intracranial-to-intracranial (IC-IC) bypass and excision of the diseased segment with direct vessel reconstruction (in situ bypass). Recent case series describe revascularisation for aneurysms that are not amenable to direct clipping or clip reconstruction, such as those with giant size, wide necks, dolichoectatic morphology, thrombosed and atherosclerotic lumens and perforating arteries arising from the parent vessel.29 30 Some authors prefer to perform a bypass when acute vessel sacrifice is needed29 because false-negative rates of balloon occlusion tests are significant.31 Large case series have had favourable outcomes in complex aneurysms treated with bypass when performed by experienced surgeons.27 29 32

In most instances, the type of bypass required—high-flow or low-flow—is guided by the aneurysm location and the available collateral circulation.27 Low-flow bypass grafts use distal scalp arteries of small calibre as donors, such as the superficial temporal artery (STA) or occipital artery, and may be suitable for some complex aneurysms of the MCA or posterior circulation. High-flow bypasses use large-calibre donor arteries, such as the proximal carotid or subclavian arteries, and are used to revascularise large territories when bypassing a complex aneurysm located in the proximal ICA.26 33 Some surgeons also prefer a high-flow bypass to revascularise the MCA territory, whereas others use single-barrel or double-barrel STA–distal MCA branches. The senior author prefers a submandibular–infratemporal bypass technique using a saphenous vein or radial artery interposition graft between the distal cervical external carotid artery and the recipient supraclinoid ICA or MCA because it enables a direct route to the recipient supraclinoid ICA or MCA and avoids the need for a subcutaneous tunnel that may compress the graft.26 32 33 However, preauricular and postauricular techniques for routing the graft are used extensively.28

IC-IC bypasses are technically challenging operations that are intended to reconstruct the cerebral circulation without extracranial donor arteries.27 IC-IC bypass grafts are less likely to be compromised by neck torsion and external compression; however, they put two vascular territories at risk for ischaemia during creation of the bypass. IC-IC bypass has been described for treating complex aneurysms of the ACA that are located in the interhemispheric fissure (side-to-side anastomosis of distal anterior cerebral arteries, usually A3-A3), too far distal from a suitable STA donor artery and where afferent and efferent vessels cannot easily be accessed for bypassing by a single surgical corridor.30 This bypass technique may be useful for complex posterior inferior cerebellar artery (PICA) aneurysms (PICA side-to side), for which harvesting of the occipital artery can be tedious.27 Because they do not require preparation of an extracranial donor vessel, IC-IC bypasses are useful in situations where an unanticipated need for revascularisation occurs during an operation.

Illustrative case scenarios

Modern endovascular techniques provide effective and minimally invasive treatment for a variety of aneurysms. Although coil embolisation is effective for many saccular aneurysms and the indications for flow diversion continue to expand, there remain clinical scenarios that ultimately may require open surgery. Here, we present several illustrative cases to demonstrate how modern endovascular and open cerebrovascular techniques can be used to manage some of the most challenging aneurysms. Flow diversion is a relatively new treatment modality with a growing evidence base. The current literature comprises a few randomised clinical trials and many small observational studies. We include the relevant evidence behind the decision making process for these cases. Currently, other than for unruptured large and giant ICA aneurysms, the use of flow diversion is considered to be off-label.

Symptomatic unruptured giant ICA aneurysm

Before the development of flow diversion techniques, giant unruptured aneurysms of the ICA were particularly difficult to treat. Giant ICA aneurysms often have wide necks that make endovascular coiling technically complicated, even with balloon or stent assistance. Chalouhi et al 4 reported high rates of recurrence (39%) and retreatment (33%) for large and giant aneurysms treated with coiling. A systematic review of giant aneurysms (≥25 mm) treated with coiling reported morbidity and mortality rates of 24% and 9%, respectively.34 Moreover, the coil mass could exert pressure on the cranial nerves and exacerbate neurological symptoms. Clip reconstruction, carotid occlusion and trapping and bypassing are other options, but can be technically demanding and carry higher surgical risks. One recent study of 141 giant aneurysms treated surgically, either directly by clipping or indirectly by parent artery occlusion with or without bypass, reported an occlusion rate of 77% but a surgical mortality rate of 13%.35 Conversely, the PUFs trial demonstrated that flow diversion for giant ICA aneurysms is a safer alternative (93% occlusion rate, 5% mortality rate), with many aneurysms that were incompletely occluded initially becoming completely occluded over time.5 6 Endoluminal reconstruction of the diseased segment leads to aneurysm occlusion and anatomical resolution of the mass effect over time with low rates of complications (6% of major stroke or death).6 The procedure is performed relatively quickly, and most patients are discharged home the following day. Large unruptured, wide-necked aneurysms of the ICA are currently the only types approved by the US Food and Drug Administration for flow diversion.

Case 1: A woman in her 60s with headaches and diplopia was found to have an unruptured partially thrombosed giant cavernous aneurysm (figure 1). She underwent placement of a FDS. One week before the procedure, she began daily dual antiplatelet therapy with clopidogrel (75 mg) and aspirin (325 mg). On the morning of the procedure, she underwent P2Y12 testing to verify effect of the drug (<192 PRU). She was given heparin during the procedure. A single PED was inserted with the distal end positioned in the communicating segment of the ICA and the stent covering the entire length of the aneurysm neck in the cavernous segment. An immediate postdeployment angiogram demonstrated increased contrast stasis in the aneurysm dome, good wall apposition of the device to the parent artery and no evidence of thromboembolic complication. After overnight observation in the intensive care unit, the patient was discharged home. A follow-up angiogram at 6 months demonstrated complete occlusion of the aneurysm with remodelling of the diseased segment of the ICA. The covered ophthalmic and posterior communicating arteries remained patent. The patient’s vision and diplopia gradually improved, and she was continued on lifelong aspirin monotherapy (81 mg).

Figure 1

(A) Partially thrombosed giant left cavernous aneurysm that was treated with a flow-diverting stent. (B) Six months after treatment, the aneurysm occluded and there was remodelling of the diseased internal carotid artery segment.

Symptomatic giant bifurcation aneurysm

Patients with aneurysms may present with neurological deficits from aneurysm pulsatility, mass effect or perianeurysmal oedema. Symptom improvement after aneurysm treatment may occur using either endovascular or open surgical methods.36 A recent multicentre study using historical controls reported rates of improvement for cranial neuropathies for unruptured ICA aneurysms treated with flow diversion compared with coil embolisation and surgical clipping.37 For ophthalmic nerve palsies, the authors reported 60% improvement after flow diversion compared with 69% for surgical clipping and 46% for coil embolisation. However, a meta-analysis comparing clipping with coiling in terms of recovery from oculomotor palsy found a benefit to clipping for ruptured aneurysms specifically (88% vs 56%, respectively; OR 5.1 (95% CI 1.08 to 25.1)).38 It is important to consider aneurysm composition when deciding on a treatment modality. The presence of coil mass, thrombus, atherosclerosis or calcifications in the aneurysm dome may prevent a clip from completely occluding the aneurysm neck. In this setting, performing coil embolisation would further add to the mass effect. Furthermore, a giant aneurysm located distally at a vessel bifurcation may not be appropriate for flow diversion. Thus, it may be necessary to surgically incise the aneurysm wall to remove the mass directly and deflate the aneurysm. For these unclippable symptomatic giant aneurysms located at branch points, a trapping and revascularisation operation may be required to effectively alleviate the mass effect and re-establish flow in the involved vasculature.33

Case 2: A man in his 60s presented with left-sided hemiparesis from mass effect from a giant unruptured calcified right ICA bifurcation aneurysm (figure 2).26 The aneurysm involved the proximal M1 and A1 branch arteries. Angiography with test occlusion of the right ICA revealed no significant flow across the AcomA. An aneurysm trapping and revascularisation operation was performed to preserve blood flow to the distal MCA and AcomA. A high-flow submandibular bypass was performed using a saphenous vein graft anastomosed end-to-side to a distal M2 branch. A side-to-side A3-A3 anastomosis was also performed to revascularise the distal right AcomA via the left anterior circulation. The aneurysm was trapped at the proximal M1, proximal A1 and ICA terminus distal to the take-off of the ophthalmic artery. The aneurysm was then decompressed to relieve the mass effect. Postoperative angiography demonstrated revascularisation of the distal MCA and AcomA, with patency of the interposition graft and the ophthalmic artery and exclusion of the aneurysm. After surgery, the patient’s left-sided hemiparesis improved.

Figure 2

(A, B) Giant right ICA terminus aneurysm with mass effect causing left-sided hemiparesis. (C) High-flow bypass using a saphenous vein interposition graft for an end-to-side anastomosis to an M2 branch. The distal right ICA continues to fill, but the trapped aneurysm is excluded from the circulation. (D) An A3-A3 side-to-side anastomosis has been performed. ICA, internal carotid artery.

Recurrent aneurysm after surgical clipping or endovascular coiling

Flow diversion is an attractive treatment for recurrent unruptured aneurysms after surgical clipping or endovascular coiling. Flow diversion obviates microsurgical re-exploration of a Sylvian fissure previously scarred from the initial clipping or the attempted placement of endovascular coils through existing stents and coil mass. Pretreatment with dual antiplatelet therapy can proceed on an elective basis, and the procedure can be scheduled when drug effect is therapeutic. Given that the delayed rupture of a previously treated unruptured aneurysm has been rarely reported,39 waiting a period of several weeks to months for aneurysm occlusion to occur is usually not a concern.

In contrast to the high occlusion rate when flow diversion is performed as an initial treatment (93%),6 flow diversion for recurrent aneurysms is less effective.40 Daou et al found that if a stent is already present, flow diversion for a recurrent aneurysm had an occlusion rate of 56%, with 10% of cases requiring retreatment.40 The technical difficulty may increase if the stent impedes advancement of the catheter and delivery of the device. Apposition between the FDS and the parent artery may be impeded, which increases the risk of endoleak and failure of aneurysm occlusion. The risk of thromboembolic complications (14%) may also increase because the prior stent can act as a nidus for clot formation, but more studies are needed to determine the safety profile of FDSs in this setting.40 Recanalised aneurysms may represent a particularly complex patient sample (eg, giant, wide-necked, multitreated) compared with those undergoing treatment for a de novo aneurysm, which may explain the increased rate of morbidity in the published literature.41

Coil embolisation has frequently been used to treat recurrent or residual aneurysms after surgical clipping with good results42; however, as noted by Owen et al,43 it is important to consider the morphology of the recurrent aneurysm when deciding treatment options. Insufficient coil compaction may not create enough space on the aneurysm neck for clip placement. The coil mass may resist closure of the clip blades and slide the clip down the neck and risk occluding the parent vessel. In some cases, an aneurysm clip may create a favourable neck for coil embolisation. In other cases, recurrent aneurysms may lack the spherical shape that facilitates coil embolisation. It is important to evaluate the aneurysm wall and parent artery itself for irregularities. A diseased aneurysm wall may be pathologic as evidenced by extruded coils or clear detectable regrowth. One study reported that 85.7% of previously coiled aneurysms that showed regrowth required multiple subsequent embolisation treatments.44 In these cases, flow diversion may provide an effective treatment in a single setting.

Case 3: A male smoker in his 50s with hypertension and two siblings with ruptured aneurysms underwent surgical clipping for an unruptured AcomA aneurysm. Postoperatively, there was a small neck remnant that was monitored with surveillance imaging. Ten years later, imaging demonstrated regrowth of the aneurysm to 4 mm (figure 3). Placement of a FDS was recommended. The patient was pretreated with clopidogrel and aspirin, and an appropriate antiplatelet effect was obtained. The aneurysm was located at the junction of the A1 and A2 segments of the dominant right ACA. A 2.75×14 mm PED was placed from the distal A2 to the proximal A1. The patient was continued on dual antiplatelet therapy for 7 months, when a follow-up angiogram demonstrated occlusion of the aneurysm without any thromboembolic complication. Aspirin monotherapy was continued for life. A CT angiogram 2.5 years later showed no evidence of aneurysm recurrence.

Figure 3

(A–C) Recurrent anterior communicating artery aneurysm (arrow) after surgical clipping. (D) A flow-diverting stent was placed from the distal A2 to the proximal A1, covering the neck of the recurrent aneurysm. (E) Angiogram obtained 7 months later demonstrated aneurysm occlusion.

Ruptured blister aneurysm treated with clip-wrapping and flow diversion

Although the indications for endovascular treatment of aneurysms are expanding, open vascular expertise remains important. Indeed, the modalities can be complementary. One scenario that may benefit from both open and endovascular treatment is a ruptured blister aneurysm. Lacking a true aneurysm neck and dome, blister aneurysms are created from a dissection or tear in the arterial wall, and only a small platelet plug and connective tissue maintain the integrity of the aneurysm. Most cases are diagnosed in the setting of SAH and affected patients tend to be younger than those with ruptured saccular aneurysms. Blister aneurysms have a dangerously high rate of growth and bleeding.45 They are also technically challenging to treat using open or endovascular techniques because of the fragility of the aneurysm wall, their small size and their ill-defined necks. Direct surgical clipping and primary coil embolisation must be approached with great trepidation. Because of these risks, alternative open and endovascular treatment strategies have been described, but evidence to guide decision making is lacking.

The antiplatelet therapy that is required for these patients presents an increased risk of haemorrhagic complications for ruptured aneurysms in the setting of acute SAH; however, flow diversion can be effective even in this setting, with rates of rerupture, bleeding complications and mortality that are relatively low, but higher than the complication rate for unruptured aneurysms.46 Flow diversion does not result in immediate aneurysm occlusion or complete dome protection. Until aneurysm thrombosis occurs, the aneurysm remains unsecured to an extent, and the risk of rerupture remains real because of the need for antiplatelet treatment. Moreover, patients are at risk for periprocedural bleeding complications should cerebrospinal fluid diversion be needed.47 Retrospective studies suggest that bleeding complications associated with antiplatelet therapy occur in 8% of patients with acute SAH48; however, most series describing the use of FDS in the setting of acute SAH have very low sample sizes and, thus, the risk and morbidity of bleeding complications may be under-reported. Larger studies with longer follow-up are needed to establish the safety profile and long-term outcomes of using flow diversion in the setting of acute SAH.

Reconstructing the parent vessel by clip wrapping can more securely bolster the aneurysm wall, thereby protecting the patient in the acute setting and allowing for treatment of the sequelae of the initial haemorrhage, that is, vasospasm and cerebrospinal fluid diversion.45 Nevertheless, clip wrapping does not effectively occlude the aneurysm and it does not prevent future rebleeding and regrowth.49 For definitive treatment, placement of a FDS can be performed after the need for cerebrospinal fluid diversion has been addressed and the patient can safely undergo dual antiplatelet therapy.50

Case 4: A man in his mid-30s with a history of uncontrolled hypertension presented with a severe headache and lethargy (Hunt and Hess Grade 3) and was found to have SAH from a ruptured blister aneurysm of the left distal ICA (figure 4). The blister aneurysm was treated with clip wrapping using Dacron mesh for external reinforcement. During his initial hospitalisation, the patient underwent lumbar drainage to help clear the subarachnoid blood and decrease the risk of vasospasm.51 He also had enlarged temporal horns and signs of early hydrocephalus on the presenting CT scan and was at risk for requiring permanent cerebrospinal fluid diversion. By treating the patient initially using open vascular surgery, we avoided the increased risk of haemorrhagic complications associated with flow diversion and dual antiplatelet therapy.52 Two months later, a follow-up angiogram demonstrated residual filling of the blister aneurysm, which was then treated with flow diversion using a PED for internal reinforcement at which point antiplatelet treatment could be safely initiated. Six months later, follow-up angiography demonstrated occlusion of the blister aneurysm.

Figure 4

(A, B) A 35-year-old man presented with subarachnoid haemorrhage from left internal carotid artery blister aneurysm. (C, D) Because of the high risk of rerupture, the patient underwent clip-wrapping of the ruptured blister aneurysm in the acute setting. A follow-up angiogram 2 months later showed persistent filling of the aneurysm. (E, F) Six months after placement of a flow-diverting stent, the blister aneurysm had occluded.

FDS migration treated with trapping and bypass

One of the disadvantages with flow diversion is that it is difficult, if not impossible, to retrieve malpositioned stents after deployment. A failed or malpositioned FDS can obstruct blood flow and be a source of thromboemboli. In some instances, FDS placement may paradoxically induce aneurysm growth and mural destabilisation if thrombus formation occurs at a rate faster than it is degraded.53 In cases of FDS failure, open surgery using bypass with or without parent artery occlusion is a bailout option.54 Although technically challenging, FDS removal can be achieved safely with microsurgery, and the vessel can remain patent after device removal.54

Case 5: A woman in her late 50s with right-eye vision loss underwent placement of a PED for treatment of an unruptured giant 3×2.2 cm aneurysm of the distal right ICA (figure 5). The PED was placed from the M1 segment into the cavernous segment, completely covering the neck of the aneurysm. Because initially there was inadequate apposition, balloon angioplasty of the PED was performed. This manoeuvre resulted in better wall apposition, but the device shortened, with the distal end of the PED protracting into the aneurysm dome. A second device was advanced in a telescoping fashion, but it could not be opened distally and was not deployed. During the procedure, the patient experienced a cervical carotid dissection that required stenting, so additional FDS placement was abandoned. Later, the patient failed a balloon test occlusion. After the procedure, the patient had headaches, vertigo and emesis, and the aneurysm was found to be partially thrombosed, causing concern for embolic complications. The patient was then given anticoagulation therapy with heparin and underwent a right pterional craniotomy for trapping and an external carotid artery-to-MCA bypass of the complex aneurysm. Once the cervical ICA was temporarily ligated and an aneurysm clip placed distal to the aneurysm neck on the ICA, the aneurysm was incised and the thrombus and PED were removed. The diseased vessel was ultimately trapped proximally by placement of a clip distal to the origin of the ophthalmic artery. A postoperative angiogram revealed patency of the bypass with filling of the native ICA terminating in a patent ophthalmic artery. The patient was discharged home, and at 6-month follow-up, her vision had improved.

Figure 5

(A, B) Giant right ICA unruptured aneurysm (3×2.2 cm). (C, D) Placement of a flow-diverting stent spanning from the distal middle cerebral artery to the proximal ICA was attempted; however, during stent placement, the device migrated with the distal end positioned within the aneurysm sac. Angiogram demonstrates persistent aneurysm filling after stent deployment. (E, F) The patient was treated with aneurysm trapping and bypass. The ICA aneurysmal segment was trapped with aneurysm clips, the aneurysm sac was opened, and the FDS was removed. A saphenous vein graft was used to create an end-to-end anastomosis between the external carotid artery and the right M2 segment (arrow). The proximal ICA supplies the ophthalmic artery. Panels (A–D), (F) reproduced from (Bowers et al 54). ICA, internal carotid artery.

Future of endovascular treatment

Flow diversion is often a short procedure55 that has a high likelihood of achieving complete occlusion even in challenging aneurysm cases. Almost all patients who undergo elective treatment are discharged the following day. Endovascular technology is advancing so rapidly that by the time outcomes data are published, the devices are already obsolete and their next iterations are already widely used. As such, the efficacy of some newer technologies is underestimated in the published literature. Indeed, recent evidence suggests that FDS are not as safe and effective as initially thought though more high-quality studies are needed to more definitively determine the appropriate indications for flow diversion.

In addition to flow diversion, intrasaccular aneurysm treatment is still common. Small-diameter balloons (HyperGlyde and HyperForm, Medtronic; Eclipse, Balt), stents (LVIS and LVIS Jr, Microvention) and microcatheters enable coiling of aneurysms located more distally and in more tortuous vessels than was possible before. New devices are designed to conform between wide aneurysms necks, covering the orifice as much as possible and even actually herniating into the aneurysm itself to assist coil occlusion.56 Dual-lumen balloons (eg, Sceptre, Microvention; Eclipse 2 L, Balt) may facilitate coiling of many wide-necked aneurysms by permitting balloon remodelling followed by deployment of a self-expanding stent through the wire channel. Next-generation coils are likely to have small diameters that fit in smaller microcatheters to facilitate delivery to distal branch vessels. Conversely, soft, larger-diameter coils available for use may decrease the total number of coils required for occlusion of an aneurysm. Bioactive substances are in development to create coils that are more thrombogenic to induce earlier connective tissue formation. Other devices, such as the PulseRider, are open-cell structures that are deployed outside the aneurysm at the site of the parent artery bifurcation and provide a more stable platform for coil deployment.

In an effort to eliminate the need for long-term antiplatelet therapy or anticoagulation, completely intra-aneurysmal flow disruption devices are being developed. The Woven EndoBridge (WEB, Sequent Medical) device is a woven nitinol self-expanding sphere that is inserted directly into the aneurysm dome. Other devices insert a porous membrane across the aneurysm neck from within the aneurysm itself through which additional coils can be inserted to promote occlusion. Tests are underway with the next generation of FDSs, which are coated with diamond-like carbon, polyzene-F or phosphorylcholine (Pipeline with Shield technology) to alter the surface of implants to render the devices less thrombogenic.

Another important issue in the deployment of more effective, newer endovascular technologies relates to the cost of the endovascular implant. For example, the cost of FDSs is prohibitive in many parts of the developing world. As with other medical technology, with time and increased production, the cost of these devices should decrease to enable wider application.

Future of open cerebrovascular surgery

As technology evolves and current trends continue, the proportion of aneurysms safely treated using endovascular methods will likely increase. This leaves only the most challenging aneurysms to undergo open surgical treatment. Already, the majority of open cerebrovascular operations are consolidated in a relatively small number of high-volume, mostly academic centres.57 The centralisation of aneurysm surgery is accompanied by better clinical outcomes58; however, a consequence of this trend is that there are fewer neurosurgeons trained to perform complex cerebrovascular operations. In much of North America and Europe, a traditional practice of open vascular neurosurgery is not possible as endovascular treatment is used to treat the predominance of intracranial aneurysms. Thus, patients who require open cerebrovascular surgery may have to travel long distances to obtain care and may not have access to appropriate care in a timely fashion. These limitations demonstrate the ongoing need for well-trained open cerebrovascular neurosurgeons. This shortage will be exacerbated in the near future with the ensuing retirement of the very experienced open vascular neurosurgeons who trained before the advent of endovascular therapy. One solution to this challenge will likely be in the continued training of neurosurgeons who have a combined vascular and skull base-focused surgical practice who will continue to perform complex cerebrovascular surgery for revascularisation following aneurysm trapping or tumour removal.


Over the last few years, clinical outcomes for aneurysms have improved in conjunction with advancements in endovascular treatments, such as coil embolisation and flow diversion. Although new technology has changed the method in which many complex aneurysms are now treated, endovascular techniques do not yet cover the full variety of complex aneurysms. Open cerebrovascular surgery maintains an important role for aneurysms that are not amenable to endovascular treatment, but as the popularity of endovascular techniques continues to increase, the number of aneurysms requiring open surgery is expected to decrease until only the most complex and technically challenging aneurysms undergo open treatment. This will create the challenge of training for and maintaining complex open vascular surgical techniques in the face of a smaller and declining surgical volume.


View Abstract


  • Contributors Each of the named authors has contributed significantly to the preparation of this manuscript. MDM predominantly drafted the manuscript. PT and MP provided additional details and made important revisions. WC oversaw the preparation and reviewed and revised the manuscript.

  • Competing interests PT is a consultant/proctor for Medtronic.

  • Provenance and peer review Commissioned; externally peer reviewed.

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.