Endovascular Therapy for Intracranial Giant Cell Arteritis

Giant cell arteritis (GCA) is an idiopathic, large vessel vasculitis which may cause flow-limiting stenosis of the intracranial arteries. GCA affects patients older than 50 years of age, with an estimated incidence of 10–30/100,000 [1]. Beyond the classical presenting symptoms of scalp tenderness, headache, and jaw claudication, the most feared cranial manifestation of GCA is visual loss seen in in 8–34% [2]. Ischemic stroke is a well-described but relatively rare complication in GCA, occurring in an estimated 2–7% [3, 4]. The majority of these ischemic complications are attributable to extracranial arterial disease resulting in thromboembolism or watershed ischemia [5]. By contrast, intracranial vessel luminal narrowing is rare [6]; however, for the subset of patients who develop severe intracranial manifestations of GCA, mortality despite maximum medical therapy may be as high as 58% [7].

Histopathologically, GCA manifests with inflammatory changes in all three arterial wall layers with granulomatous inflammation and “giant cells” at the intima-media junction, resulting in circumferential stenosis and hypermetabolism on 18F-FDG PET (Fig. 1a, b). Mononuclear cell migration is followed by destruction of elastic lamina and proliferation of medial smooth muscle cells (SMC) resulting in intimal hyperplasia and narrowing of the vessel lumen (Fig. 1c), accounting for ischemic symptoms throughout the body. In later phases, the inflammatory infiltrate dissipates, and SMC proliferation is replaced with fibrosis. Large vessel involvement was once thought to be uncommon in GCA but is now recognized in 2/3 of patients including the involvement of the brachiocephalic (47.5%), cervical carotid (35%), and subclavian arteries (42.5%) [8]. More recent work using high-sensitivity MR vessel-wall imaging also found higher than expected rates of intracranial arterial involvement (up to 40%) [9].

Medical therapy with aggressive glucocorticoids and adjuvant therapy with other immunosuppressive agents remains standard of care in GCA [2, 10]. Aspirin has been shown to reduce intracranial ischemic events in GCA via both platelet inhibition and a proposed secondary mechanism of suppression of interferon gamma signaling, but its efficacy has not been proven in randomized trials [11, 12]. Despite the overall efficacy of medical therapy, a subset of patients develop refractory neurologic deficits and cerebral infarcts.

Percutaneous transluminal angioplasty (PTA) was first reported as a treatment for GCA involving the axillary arteries. Although data are limited, PTA appears safe and efficacious in peripheral vessels but its durability is uncertain [13]; restenosis is estimated to occur in 33% after single session treatment and 18.4% after repeat treatment [14]. Reports of post-PTA stent placement in large vessel vasculitides are even more sparse [15]. The stenoses of intracranial GCA are appealing targets for endovascular treatment; however, the risk of spontaneous dissection in this population is considerable with an estimated incidence of 25% [16]. Given the uncertain efficacy and high potential risk, endovascular treatment of intracranial GCA is rarely described in the medical literature.

The purpose of this study was to investigate the indications, safety profile, and clinical efficacy of endovascular treatment for medically refractory intracranial GCA. To this end, the authors present the technical experience of a single case of severe and refractory intracranial GCA which required seven endovascular treatments that included intra-arterial calcium channel blocker infusion, which has not previously been reported in this setting. Lastly, the authors present a systematic review of the existing literature on endovascular management of GCA.

The authors performed a systematic review of EMBASE, Medline, and Web of Science online databases through January 2021 in accordance with PRISMA guidelines [17]. Studies which reported any endovascular treatment (angioplasty, stent, intra-arterial drug infusion) to address symptoms attributed to confirmed GCA (diagnosed in accordance with EULAR-ACR criteria). Studies of non-GCA vasculitides were excluded. The literature search strategy is included as Supplemental Fig. 1. Duplicate references were removed, and references were manually reviewed to identify additional studies of interest.

The clinical and technical details of a single patient treated for medically refractory GCA treated at our institution were reviewed in detail and technical aspects about the endovascular treatment, including decision thresholds for treatment, are presented and discussed in depth. After each treatment with PTA, post-procedure blood pressure control was achieved with intravenous nicardipine infusion with target systolic pressure ≤ 120 mm Hg.

Intracranial arterial stenosis due to GCA is potentially a life-threatening condition. When available anti-inflammatory and immunosuppressive therapies are ineffective, endovascular treatment can prevent catastrophic cerebral infarction. Emerging evidence has resulted in the inclusion of endovascular therapy in the revised European League Against Rheumatism (EULAR) guidelines for the first time (a 4C recommendation), noting that unless emergent, such procedures should be deferred until disease activity is stabilized [30]. Accordingly, the authors believe that practicing neurointerventionalists should understand the risks, indications, and therapeutic considerations of treating intracranial GCA.

The paucity of data on endovascular therapy for intracranial GCA is a testament to the anticipated risk of intervention in this population. Our review identified reports of only 14 patients in the literature. By synthesizing these cases, we identified several important themes. First, endovascular treatment carries significant procedural risks but may benefit patients with refractory, flow-dependent neurologic symptoms, or recurrent stroke despite maximal medical therapy. In our experience, asymptomatic dissection occurred in 1/8 (12.5%) symptomatic stenosis treated with PTA compared with 10.5% from the literature review. Intraprocedural dissection is relatively common in PTA of intracranial arteries for atherosclerotic disease and despite slow balloon inflation and proper balloon sizing, it occurs in 14–18% of cases [31]. The use of stents following PTA likely conveys an additional risk to PTA alone, although the marginal risk is difficult to estimate [32]. Contemporary estimates of intracranial PTA with or without stenting for atherosclerotic lesions show substantial morbidity, with the non-stroke periprocedural complication rate ranging from 6.4–13.4% and the periprocedural stroke rate of 9.5% (7–12%) [33]. The risk of PTA and stenting in GCA is presumably higher due to vessel wall inflammation and predisposition to spontaneous dissection in extracranial vessels in these patients [16]. GCA stenoses are histologically associated with overexpression of platelet-derived growth factor, a key driver of intimal hyperplasia, which adds to the theoretical risk of delayed stenosis of intracranial stents in this population [34]. Despite these risks, we noted a substantial repeat treatment rate (28.5% of patients) with the PTA only strategy compared with 0% in stented patients. The risk of a repeat procedure must therefore be weighed against the risks of stent-related complications and DAPT.

We also encountered two (25.0%) delayed, postprocedural intraparenchymal hemorrhages in the same territory as PTA. In both instances, immediate postprocedure Dyna-CT showed no hemorrhage, but hemorrhage occurred within 24 h. The mechanism in both cases was therefore attributed to rapid reperfusion (within subacute infarcts in one case and de novo infarct in the other). Reperfusion hemorrhage is a well-described but infrequent complication of PTA in intracranial atherosclerotic disease (ICAD) and reported in 3.3% of cases from the SAMMPRIS trial. [35] Unlike ICAD which affects both large and small vessels, GCA is an isolated large vessel arteritis with presumably normal microvascular flow and intact autoregulatory capacity. Rapid correction of flow-limiting stenosis, particularly when stenoses are present in additional vascular territories, is a plausible mechanism that may account for the higher rate of postprocedural ICH in our case, as has been reported in cervical ICA intervention [36]. Prompt reduction of blood pressure after reperfusion may be appropriate for EVT in GCA, as this technique is associated with lower rates of spontaneous intracerebral hemorrhage (ICH) and better functional outcomes in acute stroke intervention [37]. The authors believe that periprocedural blood pressure control should be tailored on a case-by-case basis, generally targeting a systolic pressure ≤ 100 mm Hg. Novel post-PTA angiographic biomarkers such as early draining vein and capillary blush may also aid in blood pressure target selection [38]. Although not entirely clear from the present study, PTA without adjunctive stenting showed lower rates of postprocedural ICH in one prospective study [32]. Both of these strategies should be considered when endovascular treatment of GCA is undertaken.

This review also highlights that the majority of patients who required endovascular treatment (66.7%) had at least one pre-existing cardiovascular risk factor. This suggests that co-morbid GCA and cardiovascular disease synergistically increase stroke risk, and that such patients require optimization of statins, antihypertensives, and antiplatelet agents. The use of antiplatelet agents in GCA with either visual symptoms or stroke has a contentious history, as evident in the inconsistent use of antiplatelet agents in the current literature review. The 2009 European League Against Rheumatism (EULAR) guidelines recommended empiric low dose aspirin in patients diagnosed with GCA based on retrospective data from two large studies [10]. A 2014 meta-analysis on the effect of antiplatelet and anticoagulation on ischemic events in GCA found no protective benefit for patients prior to diagnosis but found a marginal event reduction when antiplatelet and/or anticoagulants were instituted after diagnosis (odds ratio, OR 0.318 (0101–0.996), p = 0.049) [39]. Consequently, the 2018 update to the EULAR guidelines reversed the original recommendation, citing the results of subsequent studies which showed no benefit to prophylactic antiplatelet therapy [30]. More recently, the 2021 American College of Rheumatology guidelines provided a conditional recommendation for aspirin only for patients with flow-limiting vertebral or carotid artery stenosis [40]. In our review, most patients were managed with a single antiplatelet agent although treatment was highly variable (Table 2). Aspirin, compared with other antiplatelet agents, offers the additive benefit of inhibiting interferon gamma signaling, an important mediator of disease severity in GCA and a pathway which is not directly inhibited by glucocorticoids alone [41]. Dual antiplatelet therapy (DAPT) has not been well-studied in this population. Currently, the 2022 American Academy of Neurology Clinical Practice Guidelines recommend DAPT as first-line therapy for symptomatic intracranial atherosclerosis to prevent ischemic stroke [42]; however, given the pathogenetic differences between radiographic stenoses in intracranial atherosclerosis and GCA, it is unclear if the protective effects of DAPT seen in intracranial atherosclerosis translate to intracranial GCA. Moreover, the risk of gastrointestinal hemorrhage (as in our case) or surgical complications related to bowel perforation in tocilizumab-treated patients are important factors when DAPT is considered [19].

In this review, patients with aggressive intracranial GCA typically did not respond to conventional glucocorticoid therapy and, 50% of cases progressed despite escalating therapy with immunomodulatory agents. This suggests that patients who develop refractory intracranial GCA differ from typical GCA patients but the reason for this is unclear. It is known that patients with more aggressive inflammatory responses (IL‑6, IL1-beta, and TNF-alpha) require longer duration of glucocorticoids to achieve remission [43]. It has also been observed that some elements of the inflammatory response, particularly neoangiogenesis, are associated with lower rates of ischemic complications and that patients who harbor VEGF mutations more frequently have severe ischemic complications [44]. Interestingly, patients who develop stroke typically do so following initiation of glucocorticoid therapy [3]. Together, these observations have led these investigators to hypothesize that inflammation-mediated neoangiogenesis is protective against ischemic events. Taken together, these data suggest that a select subset of GCA patients who show clinically severe disease with relatively milder than expected laboratory inflammatory markers may be particularly vulnerable to refractory large vessel stenoses and associated complications.

The use of IA calcium channel blocker (CCB) infusion is a novel treatment strategy for intracranial GCA. IA verapamil infusion is an established modality for vasospasm due to aneurysmal subarachnoid hemorrhage (aSAH) and serves as another potential rescue therapy in GCA when PTA is contraindicated or technically not feasible. The vasodilatory action of verapamil in aSAH vasospasm is likely multifactorial, but its principle mechanism is to induce vascular SMC relaxation via inhibition of L‑type calcium channel influx [45]. The pathogenesis of arterial luminal narrowing in GCA differs from SAH vasospasm but also appears to involve vascular SMC. A 4-phase model of vessel stenosis in GCA has been proposed, in which an inciting event stimulates activity in adventitial dendritic cells. Next, a cascade of inflammatory signals recruits and activates CD4+ cells and later, CD8+ cells and monocytes, the latter of which mature into macrophage-like multinucleated “giant” cells (MGCs) [46]. MGCs, via VEGF and PDGF signaling pathways [34], stimulate migration and proliferation of smooth muscle cells, resulting in vasoconstriction in the acute phase, followed by progressive intimal hyperplasia and eventually vaso-occlusive disease. Together, these observations implicate SMC relaxation as the common final pathway for the vasodilatory actions of CCB in both aSAH and the acute phase of GCA. The authors speculate that CCBs are unlikely to be effective in the later stages of GCA, which are characterized by fibroproliferative change. Nonetheless, drawing on decades of experience and level 1 evidence supporting the use of CCB in aSAH vasospasm, CCB may prove a valuable salvage therapy for patients with GCA and recurrent stroke, although future research is needed [47].

Several limitations of this review warrant additional discussion. First, the limited number of cases reported in the literature is insufficient to draw robust practice guidelines for neurointerventional care of GCA. Moreover, owing to publication bias, periprocedural complication rates may be underreported in this review. Heterogeneous reporting of inflammatory markers, clinical follow-up, and outcomes, as well as angiographic findings (i.e. degree of stenosis pretreatment and posttreatment) further limit this study. One major unanswered question is whether concomitant vessel stenting is advantageous or potentially harmful. The authors experience with variable durability of PTA and the need for multiple retreatments (restenosis rate of 77%) could be interpreted as evidence that the upfront risk of stenting may be favorable to the risks inherent to multiple retreatments. It should be noted that the present case differed from most in the literature review in that symptomatic stenosis and or infarct was seen in three large vessel territories, thus potential risk of collateral failure was considerable and influenced treatment decision-making in a manner different than cases with a single affected vascular territory. Finally, GCA stenoses are histologically associated with overexpression of platelet-derived growth factor, a key driver of intimal hyperplasia, which adds to the theoretical risk of stenting in this population [34].

In summary, endovascular management plays an important role in the management of severe and refractory intracranial GCA. PTA can be performed and repeated, if necessary, when stenting is deemed too risky. The risks of these mechanical interventions are considerable, and they should be used sparingly in patients with recurrent ischemic events, despite maximal medical therapy. Intra-arterial calcium channel blocker infusion offers a novel, less invasive treatment option for select patients.