Discussion

A number of factors characteristic of the posterior circulation may produce a higher rate of treatment-related complications. The large number of perforating arteries originating along the basilar trunk and apex, complex aneurysm anatomy, and aneurysm location all influence this risk profile. The current literature evaluating the use of flow diverters in this aneurysm subgroup purports a higher complication rate than that associated with anterior circulation aneurysms.

In their review of seven patients with large or giant fusiform vertebrobasilar aneurysms, Siddiqui et al2 reported poor outcomes (including four deaths) in five patients. A critical evaluation of these cases reveals that the majority were dolichoectatic aneurysms involving long segments of the basilar artery. Multiple devices were placed in all cases with an average of 4.8 per patient.2 All but one patient presented with neurological deficits. Poor outcomes were the result of aneurysm rupture or brainstem infarction.2

These results cannot be applied to the entire cohort of patients with posterior circulation aneurysms. Dolichoectatic aneurysms of the basilar artery carry a dismal prognosis with a mortality rate approaching 35% at 5 years.2 ,5 The sheer length of these lesions requires the deployment of multiple devices, as the authors describe in their study, with long segments that remain unopposed to the construct.2 Compromise of perforating arteries, the requirement for multiple devices, the frequent coexistence of thrombus within the aneurysm fundus, and the presence of neurological instability secondary to mass effect or prior brainstem infarction further compound the potential for poor outcomes.

In a prospective case registry conducted at three Australian neurointerventional centers, Phillips et al3 reported 32 patients with posterior circulation aneurysms treated with the PED. Three permanent complications (3/32, 9.4%) occurred in this cohort and all involved perforator infarctions in patients with basilar aneurysms. Interestingly, all three infarctions occurred after the placement of a single PED, suggesting that coverage of a perforating artery ostium is not the only etiology of stroke in this subset of patients. An additional two patients had asymptomatic intracranial hemorrhages.3

These results led the authors to conclude that “a higher clinical perforator infarction rate may be associated with basilar artery PEDs relative to the internal carotid artery”, and that PEDs should only be used when more conventional endovascular or surgical techniques are not feasible.3

Perforator infarctions probably occur for three distinct reasons.6 In the acute setting, coverage over a perforating artery ostium by one or more tynes of the device may either sufficiently narrow or occlude the artery. Alternatively, acute thrombus within the device may migrate distally through a perforating artery. In the subacute or chronic setting, delayed infarctions may occur as a result of neointimal proliferation over the perforator ostia.

In an analysis of 12 patients with basilar aneurysms treated with the Silk device (Balt, Montmorency, France), Kulcsar et al6 specifically assessed branch vessel and perforator patency after flow diverter placement. Of the 22 branch vessels that were bridged by the Silk device, only one posterior cerebral artery (PCA) occluded at the time of treatment. This was managed with intra-arterial tirofiban administration and produced no neurological sequelae.

Nonetheless, four other complications occurred in this series, including one case of thrombosis through the construct 12 h after treatment. This patient presented in poor neurological condition (modified Rankin Scale score (mRS) of 4) and required the placement of an additional Enterprise stent (Codman, Raynham, Massachusetts, USA) for complete coverage of the aneurysm neck. The patient worsened as a result of treatment to an mRS score of 5.

Three patients presented with delayed ischemic insults, including two thalamic infarcts at 7 months and 4 weeks after treatment and one pontine infarct 5 weeks after treatment. This suggests that neointimal hyperplasia may progressively narrow or occlude perforating arteries. The authors astutely recognized that perforator patency is virtually impossible to ascertain during treatment, given that angiography often fails to opacify these small vessels. The potential of neointimal proliferation underscores the need for diligent antiplatelet therapy and patient compliance.

While they do not specifically mention the morphology of the aneurysms treated (ie, dolichoectatic or saccular), they state that the majority of patients were symptomatic, previously treated endovascularly, or deemed to be at high risk of aneurysm rupture. Certainly, this higher risk cohort would be expected to develop acute or subacute complications more frequently.

In their series of 251 aneurysms treated with the PED (19 of which involved the posterior circulation), Saatci et al7 described a single patient who developed thrombosis of the right PCA after jailing it with a PED extending from the left PCA to the basilar trunk. The patient sustained a brainstem infarction and subsequently died. The authors did not specifically state whether any complications other than the PCA thrombosis occurred in this subset of patients. We are unsure why major posterior circulation branches would be more prone to thrombosis after PED jailing than branches of the internal carotid artery. Such complications are more likely to be related to issues with antiplatelet therapy, aneurysm selection, or device problems such as vessel dissection or PED malposition.

Aneurysm morphology and presentation are critical factors to consider when selecting posterior circulation aneurysms for treatment with the PED. In their report of seven patients with posterior circulation aneurysms, Chalouhi et al8 treated only one fusiform aneurysm that exerted brainstem compression. This patient failed to improve clinically and demonstrated no change in the morphology of the aneurysm after treatment. No major complications occurred in this series. The authors argue that perhaps ‘symptomatic’ fusiform aneurysms, specifically those compressing the brainstem, should not be treated with flow diversion due to the frequent need for multiple telescoping devices. This factor may add to procedural morbidity.8

Within their series of 31 patients with posterior circulation fusiform aneurysms, Raphaeli et al treated four with flow diversion (Silk).9 Two complications occurred in this cohort including one death and one case of delayed transient hemiparesis. The patient who died had symptoms of brainstem compression secondary to a giant, partially thrombosed vertebral artery aneurysm which was treated with a single flow diverter. This case, as well as those of Chalouhi et al, suggests that aneurysm enlargement from further thrombosis compounds brainstem compression producing either devastating neurological sequelae or death. The argument against the use of this technology in the setting of such clinical instability is clear.

Basilar artery thrombosis, both acutely and in a delayed fashion, is another source of complications in this patient subset. Acute thrombosis is probably produced through a number of mechanisms including lack of response to antiplatelet medication, malposition or lack of apposition of the stent, or arterial injury at the time of PED deployment. Careful assessment of appropriate response and compliance with the antiplatelet regimen as well as the exclusion of unfavorable (dolichoectatic or circumferential) aneurysms will lessen the likelihood of this life-threatening complication.

More troubling are reports of delayed thrombosis. Klisch et al10 described two patients with fusiform basilar trunk aneurysms who developed thrombosis after clopidogrel was stopped when follow-up angiography at 1 year demonstrated slight residual filling of both aneurysms. The authors postulate that flow diverters remain more thrombogenic and undergo neoendothelialization more slowly when the devices are within the thrombosed portion of these fusiform aneurysms. When encountering residual filling in this situation, the authors argue that placing additional devices through the construct is preferable to altering the antiplatelet regimen.10

Fiorella et al11 reported a patient with a V4 aneurysm who developed thrombosis 23 months after treatment with PED-supported coil embolization. The authors suggest that ‘over-exuberant neointimal hyperplasia’ occurred within the construct as a result of an aggressive biological response to the polyglycolic acid-impregnated coils.11 This patient's antiplatelet regimen had also been modified in the months preceding his demise.

These reports underscore our lack of understanding of the need and duration of antiplatelet therapy, as well as uncertainty regarding the interaction of these devices with coils and other endovascular hardware.10 ,11 Furthermore, these cases involved circumferential aneurysms, a subgroup that is not included in our present analysis and that is not optimally treated with flow diversion.10 ,11

Other complications have been described in patients with posterior circulation aneurysms treated with the PED. Colby et al12 reported a sixth cranial neuropathy in a patient with a giant vertebrobasilar aneurysm. Fischer et al13 reported a ‘severe’ cerebellar hemorrhage after treatment of a fusiform basilar trunk aneurysm. Like the complications discussed above, these complications occurred in patients whose aneurysms were not ideally suited for treatment with this device.12 ,13

Distal wire tip fracture occurred in a case reported by Briganti et al.14 Following removal of the fractured wire, the patient developed thrombosis of the SCA requiring emergency thrombolysis. Fortunately, the patient eventually made a complete recovery.14 In our case, the distal wire tip was retained in a small perforator arising from the right PCA. We did not retrieve the fractured portion. The patient awoke normally and did not manifest any late clinical sequelae.

Our series demonstrates that, with proper patient and aneurysm selection, the PED can be used safely and effectively. We encountered five complications after use of this device in this setting, only one of which was permanent. Including the distal wire tip fracture, the other transient complications involved a clinically silent SCA thrombosis, small CVA, and a urethral tear from traumatic Foley catheter placement. Our single permanent complication was an intracerebral hemorrhage after external ventricular drain replacement. This event left the patient disabled and was probably exacerbated by antiplatelet therapy. Other than improper aneurysm selection, these complications represent the main categories accounting for patient morbidity: issues pertaining to antiplatelet therapy, perforating or branch vessel thrombosis, and non-PED-related procedural difficulties.

The utility of this device has been demonstrated in the management of V4 aneurysms, a large proportion of which are dissecting and prone to re-rupture.4 Eight V4 aneurysms were treated in our series. Of the seven patients undergoing angiographic follow-up, five had complete aneurysm occlusion and two had near-complete occlusion.

Dissecting aneurysms proximal to the V4 segment and distal along the course of the basilar artery are also amenable to PED treatment. De Barros Faria et al reported the use of the PED for the treatment of 21 dissecting aneurysms of the posterior circulation involving the V4 segment, vertebrobasilar junction, basilar trunk, and the posterior inferior cerebellar artery. The utility and safety of this device for this particular subset of dangerous aneurysms represents an important therapeutic advance.15 Despite the added potential morbidity associated with antiplatelet therapy, use of the PED in the setting of SAH secondary to a posterior circulation dissecting aneurysm should be considered when other conventional endovascular and microsurgical treatment options are either not possible or fail.