European Headache Federation recommendations for neurologists managing giant cell arteritis

The GCA headache phenotype is yet to be fully characterized in terms of frequency, severity and other associated characteristics. Vigilance is warranted in patients with new onset headache over the age of 50 years.

A comprehensive history and examination in all those suspected with GCA will help guide investigations.

There is currently no specific blood biomarker that can definitely diagnose GCA. The majority of those with GCA have elevated acute phase reactants at disease onset such as an elevated erythrocyte sedimentation rate (ESR) [32] or plasma viscosity, C-reactive protein (CRP) [32] and platelets [33]. However inflammatory markers have been reported as normal in between 1 and 10% [32, 34]. Additionally, few cases who have presented with visual loss within the literature have reported normal inflammatory markers [35, 36]. It is well known that those with cranial GCA typically have lower inflammatory markers than patients with extracranial manifestations (LV-GCA and PMR) [24, 37, 38]. Additionally, those with higher inflammatory markers tend to have fewer ischaemic events [39]. Other blood changes include anaemia and elevated liver enzymes.

Both standard treatment with GC [40] and indeed licensed targeted treatment [41] for GCA confers significant risk of morbidity, therefore a confirmatory investigational test should be performed to secure the diagnosis.

Colour Duplex Ultrasonography (CDUS) is a non-invasive imaging modality that assesses the arterial wall anatomy, the patency of the lumen and allows assessment of blood flow. In GCA, imaging the temporal arteries shows homogenous, hypoechogenic (dark), circumferential vessel wall thickening, which is known as the “halo sign” when pictured in cross-section (Fig. 1) [42]. On meta-analysis the sensitivity and specificity of a unilateral hypoechoic halo compared to positive TAB were respectively of 68% (95% Confidence Intervals (CI): 57–78) and 81%. The diagnostic value of any abnormality on CDUS, such as hypoechoic halo and/or stenosis and/or occlusion, compared to TAB had a sensitivity of 78% (95% CI, 64–87), a specificity of 79% [43].

Other features on examination include the compression sign when using b-scan mode, where a normal artery image is extinguished on compression whilst an artery with features of active vasculitis is not, providing superior inter-observer agreement [44]. This is important in distinguishing patients with artefactual changes which lead to the detection of false halos (such as the presence of extensive vessel tortuosity and/or atherosclerosis) from a true halo.

Major advantages of CDUS are that it is non-invasive and can sample the majority of main arterial territories which can be affected such as the common superficial temporal artery, frontal and parietal branches as well as axillary, carotid, subclavian and vertebral arteries [45, 46]. The Temporal Artery Biopsy vs Ultrasound in Diagnosis of Giant Cell Arteritis (TABUL) study, a cross-sectional prospective study of 381 patients investigated for GCA, has given the most comprehensive analysis of use of CDUS in GCA [45]; with this evidence CDUS has allowed for a significant reduction in the requirement for temporal artery biopsy [47].

¹⁸F-fluorodeoxyglucose positron emission tomography (¹⁸F-FDG-PET) has traditionally been technically difficult to interpret in cranial disease due to the brain activity and the size of the temporal vessels being assessed. Recently the GAP study [48], reported on a progressive protocol to assess the head, neck and thorax with a dedicated time-of-flight 18F-FDG-PET protocol and 1 mm CT reconstruction. They found that it was helpful early in the investigational pathway not only to distinguish masquerades of GCA but also for distinguishing active disease in the intracranial vessels (Fig. 2). However, major disadvantages of this modality remain such as the cost, radiation exposure, availability and validation with existing imaging modalities.

Temporal artery biopsy (TAB) allows for a histological diagnosis, but in some centres may be difficult to access readily and for the patient, it is an invasive procedure. When a TAB is positive it provides justification for treatments used in GCA, however when negative can create diagnostic uncertainty. The sensitivity of TAB ranges in the literature between 80 and 90% probably due to publication bias, however in a comparative trial, it was shown to have a poor sensitivity of 39% [45]. Negative biopsies can be a concern for the clinician, and a number of factors may contribute including the nature of the tissue obtained (i.e vein instead of artery biopsied), length of biopsy (preferably greater than 1 cm), histological interpretation and length of time on GC treatment resulting in a false negative biopsy [45, 49]. However, given the morbidity of treatment, some would still consider performing a biopsy at any point during the disease [50]. The utility of TAB has been recommended by some international bodies to be superseded by non-invasive imaging modalities such as CDUS [20]. Some centres have chosen to use TAB when the results of CDUS are either equivocal or not in keeping with the clinical presentation [51].

LV-GCA often presents as an inflammatory syndrome and is only detected by imaging modalities such as: CDUS, computed tomography (CT) / CT angiography (CTA), magnetic resonance imaging/angiography (MRI/MRA) or ¹⁸F-FDG-PET. CDUS shows axillary and carotid artery circumferential homogenous hypoechogenic wall thickening, which is similar to the findings in the temporal arteries when affected [19]. However, CDUS lacks the ability to image the larger thoracic vessels.

¹⁸F-FDG-PET is usually combined with low-dose CT and has a role in assessing disease activity and the extent of involvement in extracranial GCA (case vignette 1 and 2, Figs. 2, 3 and 4) [52]. GCA with a reported sensitivity of 77% [53]. Increased FDG uptake in the vessel wall is the hall-mark of vasculitis in PET. In general, visual evidence of vascular uptake higher than tracer uptake in the liver is considered to be suspicious for LVV [54]. ¹⁸F-FDG-PET imaging can demonstrate involvement of the larger aortic vessels and increased reliability. Moreover, the increased uptake seems to persist longer after treatment initiation.

High-resolution (3 up to 7 Tesla) MRI to image the superficial and extracranial arteries in GCA demonstrated arterial wall thickening with peri-adventitial and mural contrast enhancement with sensitivities ranging from 50 to 89% [55]. Other imaging techniques that are being investigated include photoacoustic imaging [56] and optical coherence tomography of the superficial temporal artery [57].

Where there is a high clinical suspicion of GCA immediate high dose GC should be started [59]. The standard initial GC dose for GCA is between 40 and 60 mg oral prednisolone equivalent per day, depending on the patient characteristics, including pre-existing comorbidities and body weight.

For those with cranial ischaemic symptoms (such as visual loss secondary to GCA or stroke) high dose intravenous methylprednisolone 500 mg–1000 mg induction therapy for 3 days may be used, followed by reducing course of oral glucocorticoids.

The little evidence in the literature to guide the route of administration of glucocorticoids is conflicting. Mazlumzadeh et al [60] found that a 3-day induction of IV methylprednisolone allowed a more rapid weaning from oral prednisone than placebo and conferred a reduced the cumulative glucocorticoid dose at Week 78. However, Chevalet et al [61] showed no benefit for a single induction dose of IV methylprednisolone in reducing cumulative steroid dose at 1 year. Hayreh et al. [62] in his observational case series did not show any obvious benefit between IV or oral administration, although those treated with IV tended to have poorer visual acuity at diagnosis.

Worryingly, there is evidence that sight loss can still occur in nearly one third, despite the use of high dose IV methylprednisolone within the first 6 days [63].

Data from the GiACTA trial [64] reported that only 14% of patients on a short 6 month oral GC taper (without additional tocilizumab) achieved remission at 12 months, and 18% achieved remission at 12 months, if given a 12 month GC tapering regimen. The EULAR guidelines have recommended maintaining induction levels of 40-60 mg per day until there is clear resolution of symptoms and normalization of inflammatory markers. GC should be tapered to a target of 15–20 mg/day within 2–3 months and then to ≤5 mg/day after 1 year, with an aim to stop within 1–2 years if there is no relapse; this strategy would result in a projected cumulative dose of just over 6 g [59].

The prolonged use of GC is well documented and can exacerbate conditions such as hypertension, diabetes mellitus, heart failure, and osteoporosis. It can also cause depression [65]. In patients with GCA there is a high morbidity with 86% developing at least one side effect and 68% developing more than two [40]. Real world cumulative dose of GC can far exceed guideline recommendations, and one study showed a mean cumulative dose of 8.4 g, with more than one third accumulating a dose greater than 10 g [66]. This is likely because disease relapses are common, reported to occur in half of all patients, necessitating an increase in GC, or a delay in tapering the dose [67]. For those with a history of relapse there is an increased prevalence of osteoporosis [68]. The adverse event hazard ratio is increased by approximately 3% for every 1000 mg increase in the cumulative GC dose [69].

There have been three randomised, placebo-controlled trials investigating Methotrexate (MTX), between doses of 7.5–15 mg per week, as an adjunctive steroid-sparing agent [70‐72] Despite methodological limitations of each individual study, a meta-analysis using pooled individual patient data from these trials demonstrated a reduced risk of first relapse (HR 0.65, 95% CI 0.44 to 0.98, p = 0.04) and second relapse (HR 0.49, 95% CI 0.27 to 0.89, p = 0.02), a higher probability of glucocorticoid-free remission for at least 24 weeks (HR 2.84, 95% CI 1.52 to 5.28, p < 0.001) and a lower cumulative glucocorticoid dose (0.84 g less glucocorticoid use after 48 weeks in patients treated with MTX versus placebo). There was no significant difference in side effects between the intervention and placebo groups. On further review of the included studies a low dose of MTX had been used, compared with current practice and the patient selection included those in whom steroid titration from low doses had failed on several occasions [73]. Dumont et al. [67] recently reported a significant reduction in the duration of glucocorticoid treatment in half of their cohort with glucocorticoid-dependent disease who had received a steroid sparing agent (mainly but not exclusively methotrexate) of 36 [15—115] versus 61 [14—212] months (p = 0.008).

Other conventionally used steroid sparing therapies such as Azathioprine [74] [De Silva 1986], Cyclophosphamide [75] and Mycophenolate [76] have not been shown to be superior to GC alone. Ciclosporin was been deemed ineffective [77]. Use of anti-tumour necrosis factor (TNF) alpha therapy in GCA was disappointing [78]. Leflunomide has been investigated in an open-label prospective single centre study of 76 patients and shown some benefit in controlling disease, reducing the use of GC [79].

Aspirin is widely used to prevent ischaemic complications such as stroke and myocardial infarction in those with cardiovascular disease. The literature has conflicting reports of the protective use of aspirin in GCA, with some supporting its beneficial effects [80, 81] and others refuting it [82, 83]. Currently overwhelming clinical evidence of a reduction in ischaemic events is lacking [84]. The use of low-dose aspirin as an adjunctive treatment in GCA must acknowledge the recognised haemorrhagic risks associated with aspirin and concurrent GC use.

Tocilizumab (TCZ), a humanised monoclonal antibody to the IL-6 receptor, has been investigated in phase II [85] and phase III [64] RCTs. In the phase II study 30 patients with new-onset or relapsing disease were randomised to receive either weekly TCZ infusions or placebo, both with a tapering GC regimen. At week 12, 85% of the TCZ group had achieved clinical and biochemical remission as compared with only 40% of those in the placebo group (p = 0.03). At 1 year, 85% of the TCZ group had remained relapse-free versus a mere 20% of the placebo group (p = 0.001) Use of TCZ resulted in a significant reduction in steroid requirement [86].

The Giant Cell Arteritis Actemra (GiACTA) trial was a large phase III, multi-centre, double-blind, placebo-controlled study examining the efficacy of subcutaneously administered TCZ to induce and sustain remission to 12 months. GiACTA included newly diagnosed patients with GCA and those with refractory disease. Patients were enrolled and randomised to one of four arms; weekly or fortnightly TCZ (162 mg) combined with a 26-week prednisolone taper or; placebo plus a 26-week or 52-week prednisolone taper. At 12 months, those receiving TCZ were significantly more likely to have achieved sustained remission as compared with both the 26-week and 52-week steroid taper groups. This confirmed a significant steroid-sparing effect of TCZ (patients receiving TCZ had required half the cumulative glucocorticoid dose used in the placebo groups) [64].

The side effect profile of TCZ (Table 1) is well documented in the rheumatoid arthritis literature [40]. Within the older GCA population concern has been raised regarding the risk of concurrent diverticular disease, transient neutropenias, and elevations of triglycerides and deranged liver function tests [87].

The cost, side effect profile and inability to serially monitor the serum CRP for evidence of relapse means that there is a reluctance to prescribe TCZ for new onset patients, despite the trial evidence. However, when patients are deemed refractory i.e. those people with GCA who never achieve remission, this is typically a more accepted use of TCZ. Likewise, for those patients with glucocorticoid-induced side effects or patients with significant pre-existing comorbidities such as psychiatric disturbances, pancreatitis, or uncontrolled diabetes or hypertension, tocilizumab could be effective as a first line treatment [21, 59].

The 2-year open-label extension study from GiACTA analysed 215 of the original 251 randomised patients. Participants in clinical remission received no further treatment and those not in clinical remission received weekly TCZ (162 mg) and/or GC and/or MTX. Higher proportions of those originally assigned to weekly TCZ were treatment-free compared with those originally assigned to placebo. Importantly, the cumulative dose of GC was substantially lower (almost half) in patients assigned to weekly TCZ (2604 mg) versus patients given either 26 weeks of GC plus placebo (5006 mg) or 52 weeks of GC plus placebo (5322.5 mg). Interestingly, patients assigned to placebo plus only 26 weeks of GC required almost as much GC therapy as those assigned to 52 weeks of GC plus placebo, reflecting the poor control of disease when using 26 weeks of GC monotherapy [86].

Use of TCZ should be considered for patients with a confirmed diagnosis of GCA (based on imaging or TAB) who have either refractory disease, or those with co-morbid disease that could be significantly exacerbated by GC use.