Reviewing the Pathophysiology Behind the Advances in the Management of Giant Cell Arteritis

The lifetime risk of developing GCA is 1% for women and 0.5% for men [9] with a reported worldwide incidence ranging between 1 in 100,000 and 30 in 100,000 [10]. This appears to be related to a number of factors, including age, latitude, genetic susceptibility and ethnic background [9‐22]. Amongst these well-established risk factors, age is widely regarded as the most significant. In a cross-sectional study conducted in northern Spain, a near 20-fold increase in the incidence of GCA was reported between those in their 6th decade of life as compared to those in their 9th decade. Similar trends have also been reported by studies conducted in the UK [12], Scandinavia [13, 14], Italy [15, 16] and the USA [17‐20].

GCA most commonly affects those of northern European and Scandinavian descent [21]. This in part seems to be attributable to genetic factors; of particular interest have been the major histocompatibility complex (MHC) genes [22‐26]. Most recently, a large genome-wide association study in GCA was conducted on 1651 cases and 15,306 controls from six countries of European ancestry [24]. This revealed the human leucocyte antigen (HLA) class II gene loci as those with the strongest association to GCA susceptibility and specifically highlighted DRB1*04, DQA1*03 and DQB1*03 alleles as those contributing most significantly to this signal. This data builds on the previous body of evidence describing a strong association between HLA-DRB1*04 carrier status and an increased risk of GCA [25]. Interestingly, a higher population frequency of the HLA-DRB1*04 allele has been reported in more northerly latitudes [26], providing insight into the potential determinants of the geographic variability of GCA incidence.

MHC class I genes have also been implicated as risk factors in GCA, with a strong association being described between the propensity for GCA and the HLA-B gene locus [24, 26, 27]. Further, genotype interrogation of the MHC class I chain-related gene A (MICA), which encodes putative stress-inducible heat shock elements [26], has been shown to confer susceptibility to GCA in a northern Spanish cohort [27]. This appeared to be independent of linkage disequilibrium with HLA-B and exhibited a cumulative risk effect when concurrent with the HLA-B15 haplotype or the HLA-DRB1*04 allele [27].

Of the non-MHC genes found to predispose to GCA, the strongest evidence is in support of the interleukin-12β (IL-12β) gene [22]. Genotyping by immunochip array performed on a cohort compromising of 997 Spanish and Italian cases and 2775 controls reported the largest signal outside the MHC region to be the single nucleotide polymorphism (SNP) rs755374 in the IL-12β gene locus [23]. This relationship was confirmed in a further four cohorts of European ancestry. Remarkably, this gene is traditionally associated with an increased risk of Takayasu’s arteritis [23]. Other non-MHC genes, such as TNF-α microsatellite [28] and interleukin-6 (IL-6) [29], appear to be of emerging significance. However, better powered studies of more diverse populations are needed to validate their importance as risk genes.

GCA is a T cell driven disease characterised by the formation of vessel wall granulomas, intimal hyperplasia and end-organ ischaemia [21, 30]. It is a disease of immune dysregulation and, as such, emerging data not only sheds light on the importance of Th1 and Th17 pathways but also on the role of innate immunity in propagating and sustaining the inflammatory milieu (Fig. 1) [30‐35].

Giant cell arteritis is renowned for its insidious onset, heterogenous presenting features and constitutional symptoms, which often contribute to diagnostic uncertainty and delay in treatment [34]. Up to 40% of patients present atypically [53] and the pattern of disease involvement ranges from cranial involvement such as anterior ischaemic optic neuropathy (Fig. 2), retinal arterial occlusions (Fig. 2) and choroidal infarction (Fig. 3) to extracranial such as symptoms of polymyalgia rheumatica or large vessel vasculitis (Table 1); they can be mutually exclusive or a combination of the spectrum. This condition can therefore present to many specialists which can further compound a diagnostic delay [34]. Many studies have examined the presenting features, and among those with the highest positive predictive value include scalp tenderness and/or jaw claudication (Fig. 4) [54, 55]. The natural history of the disease varies, and because of the advent of targeted treatment disease stratification is required, with clear definitions of disease states (Table 2) [8].

Securing a diagnosis of GCA can be challenging, and requires a prompt, thorough history, full clinical examination and a combination of investigations. Laboratory findings in GCA typically reflect a pro-inflammatory picture with a prolonged ESR, increased plasma viscosity, raised CRP, thrombocytosis, raised alkaline phosphatase and a normochromic normocytic anaemia [56]. CRP is the most sensitive individual serological marker [57] and should always be checked. An isolated raised ESR with normal CRP should prompt a search for alternative diagnoses (e.g. myeloma, connective tissue disease).

Given that the standard treatment confers significant risk of morbidity most agree that a confirmatory investigational test is required. Temporal artery biopsy (TAB) allows for a histological diagnosis, abrogates diagnostic doubt and provides justification for the use of prolonged courses of GC and the morbidity which accompanies their use. Histologically, the inflammatory infiltrate in GCA is predominantly composed of histiocytes/macrophages and CD4-positive lymphocytes, with variable involvement of the three arterial layers (Fig. 5), often showing only segmental/sectorial inflammation in parts of the arterial wall. The lymphocyte population usually includes lesser numbers of CD8+ cytotoxic lymphocytes and even small clusters of CD20+ B lymphocytes (Fig. 6). Focal fusion of histiocytes and formation of multinucleate giant cells are seen (Fig. 7), but these are not required for the diagnosis. Morphologically, four different inflammatory patterns can be differentiated with the panarteritic pattern being the most common (69%), followed by the concentric bilayer pattern (18%) (Fig. 6) with involvement of intima and adventitia, and least common, two adventitial patterns, with and without local invasion of the media (7% and 6%, respectively) [58]. The inflammation typically results in breaks, segmental loss and reduplication of the elastic lamina. This is best seen on elastin stain that highlights disruption, segmental discontinuation and focal reduplication of elastica (Fig. 8). This may cause critical luminal narrowing and aneurysm formation. When the active phase of the disease has vanished, inflammatory cell infiltrates may be absent but intimal-medial scarring and injury to the elastic lamina often remains, allowing a diagnosis of ‘healed arteritis’. A recent report links the CD68+ (cluster of differentiation 68+) macrophage marker in TABs with those who are refractory to initial GC tapers, and those eventually placed on immunomodulatory therapy [59].

The sensitivity of TAB can vary between 39% and 91% because of skip lesions, inadequate sample length and the initiation of GC prior to biopsy [60, 61]. The resultant high false negative rate could potentially lead to patients being continued on GC as a precautionary measure. The true false negative rate of TAB cannot be estimated with certainty as the clinical diagnosis of TAB-negative, TAUS-negative GCA cannot be confirmed diagnostically. Consequently, the European League Against Rheumatism (EULAR) recommendations on the diagnosis of large vessel vasculitis have been updated to include the use of non-invasive imaging techniques to assist or even supersede TAB in certain circumstances [62]. The first study examining the diagnostic application of ultrasound (USS) in GCA was conducted in 1997 by Schmidt et al. [63]. They assessed 30 patients for the presence of:

They found the halo sign to have a sensitivity of 73% and specificity of 100% when compared with the final diagnosis. Three meta-analyses have provided an evidence base for the role of USS in the diagnosis of GCA [64‐66]. Typically reported sensitivities range between 68% and 75% and specificities between 83% and 92%, when either final histological or clinical diagnosis is used as the comparator gold standard. It should be noted that the ability to detect a halo sign declines rapidly after the initiation of glucocorticosteroid (GC) with sensitivity falling to below 50% at 4 days [64, 65, 66]. The specificity of the halo sign is increased with the use of the ‘compression test’: a true halo will not disappear if the probe is used to compress the temporal artery [67].

The EULAR taskforce for imaging in large vessel vasculitis guideline recommended USS as the first-line imaging modality for predominantly cranial GCA in centres where experience is at hand, as the test is less invasive and thus well tolerated by elderly and often frail individuals, the result is not delayed and it is cost-effective [62]. Experience of USS for GCA has led to the reduction in numbers of TABs being performed at some centres [68, 69]. Additionally, USS of both the temporal artery and the axillary artery increases the yield of the diagnosis [68].

For large vessel GCA, other imaging techniques are emerging as clinically useful. ¹⁸F-fluorodeoxyglucose positron emission tomography (¹⁸F-FDG-PET) is usually combined with low-dose computed tomography (CT) assessing disease activity and extent of involvement [70]. Although large vessel imaging is sensitive to GC therapy, uptake can persist despite treatment and absence of clinical symptoms in some. Research areas in GCA imaging include super high-resolution magnetic resonance imaging (MRI) of the superficial and extracranial arteries demonstrating arterial wall thickening with peri-adventitial and mural contrast enhancement [71] and transdermal optical coherence tomography (OCT) of the superficial temporal artery [72].

Suspected GCA remains a medical emergency, and the number of cases investigated for GCA is rising [73]. Rapid access clinics, often with access to TA USS and combined ophthalmology/rheumatology clinical expertise, have been reported to improve patient outcomes [62, 68]. It should be noted that excluding GCA as a diagnosis is also important to prevent patients without disease being exposed to prolonged GC exposure unnecessarily.

Glucocorticosteroids remain the gold-standard treatment in GCA; it is a consensus-based approach rather than one supported by well-powered prospective studies [56]. Short-term side effects such as GC-induced psychosis can be challenging to treat and long-term GC use is associated with a multitude of adverse effects including an increased susceptibility to infections, osteoporotic fractures, diabetes, hypertension, glaucoma, gastric ulcer disease and mood disorders, amongst others [74]. A recent study by Proven et al. reported that 86% of patients developed at least one GC-related complication over a 10-year period [75]. These worrying findings in the context of an aged population with a high baseline burden of comorbidities have prompted research into GC-sparing agents in the form of traditional disease-modifying antirheumatic drugs (DMARDs) and, more recently, targeted biologic agents.

To date, only three randomised, placebo-controlled trials have examined the efficacy of methotrexate (MTX) as an adjunctive GC-sparing agent and these have reported conflicting results [76‐78]. A meta-analysis of these studies revealed that patients treated with MTX had a significant, though modestly reduced risk of relapse compared with placebo, and this on lower cumulative GC doses [79]. Notably, however, and despite the apparent GC-sparing effect of MTX, there was no significant difference in the rate of drug-related adverse events between the intervention and placebo-controlled groups [79]. These studies were conducted with low-dose MTX and in patients in whom GC titration from low doses had failed on several occasions, which may explain why the results were marginal.

Leflunomide has been shown to be effective in Takayasu’s arteritis and other T cell driven diseases such as rheumatoid arthritis. In the few retrospective case-series reported in the literature, statistically significant GC-sparing effects have been witnessed though these equate to minimal absolute reductions in cumulative GC doses [80, 81]. A paucity of well-powered prospective studies assessing leflunomide efficacy in GCA means it does not yet appear to have a role in GCA. However, a recent open-label prospective study conducted by Hočevar et al. [82] has shown some promise, supporting the need for well-powered randomised controlled trials (RCTs) to further interrogate the role of this drug.

Other conventional DMARDs such as azathioprine [83], cyclophosphamide [84] and mycophenolate [85] have not been shown to be superior to GC alone, whilst ciclosporin has been deemed ineffective [86]. Other adjunctive therapy, such as aspirin, has previously been supported in GCA, but clinical evidence is lacking [87].