Color Doppler ultrasonography (CDUS) can detect inflammation in the vessel wall. No studies have evaluated the examination of the common carotid artery by CDUS in the diagnostics of giant cell arteritis (GCA). Our aim was to evaluate the combination of CDUS examination of the temporal, axillary, and common carotid arteries in the diagnosis of GCA.
Patients ages ≥50 years who were referred to our department between April 2010 and October 2012 and suspected to have GCA were consecutively examined. A positive clinical evaluation for GCA 6 months after the first evaluation by 3 rheumatologists was considered as the gold diagnostic standard. All patients underwent CDUS of the temporal, axillary, and common carotid arteries. A biopsy of the temporal artery was performed for most patients.
A total of 88 patients were assessed. Forty-six patients were diagnosed to have GCA by the defined gold standard. Forty-eight patients had a positive CDUS of the temporal artery. Forty-six patients diagnosed with GCA had a positive CDUS of the temporal, common carotid, and axillary arteries (100% sensitivity) and 4 patients had a positive CDUS without having GCA (91% specificity). Among the 39 GCA patients that underwent a biopsy, vasculitis was observed in 26 patients (66%), yielding a sensitivity of 67% and a specificity of 95%.
CDUS of the common carotid, axillary, and temporal arteries had an excellent sensitivity and high specificity to diagnose GCA. CDUS has the potential to replace biopsy in ordinary clinical care without compromising on sensitivity and specificity.
Giant cell arteritis (GCA) is the most common form of primary systemic vasculitis and occurs in the middle-aged and elderly population at a mean age of ∼70 years ([1, 2]). The clinical presentation is heterogeneous and may be dominated by a new onset of headache, polymyalgic symptoms, or systemic features (). The diagnosis of GCA is based on clinical symptoms and findings, inflammatory markers, and the existence of inflammation in the temporal artery biopsy, as reflected in the American College of Rheumatology (ACR) 1990 criteria for the classification of GCA (). The presence of inflammation in the vessel wall is the definite proof of vasculitis and is therefore considered as the gold diagnostic standard in GCA (). However, even a positive temporal artery biopsy with the presence of inflammation in the vessel wall is not pathognomonic for GCA. The presence of vasculitis in the temporal artery has also been reported occasionally in, for example, polyarteritis nodosa (PAN) and granulomatosis with polyangiitis (Wegener's) (GPA) ([6-8]). The diagnosis of GCA may be a clinical challenge due to clinical heterogeneity and because a negative temporal artery biopsy may be explained by, for example, a patchy pattern of artery inflammation in the artery wall () or by the biopsy being taken from an area with no presence of inflammation (). Sophisticated imaging methods like color Doppler ultrasonography (CDUS), computed tomography, positron emission tomography, and magnetic resonance imaging have revealed that the distribution of the inflammation is not limited to the temporal artery, but may also include large vessels such as the aorta and branches from the aortic arch ([11, 12]).
CDUS is a promising noninvasive technique to diagnose vasculitis not only at the temporal artery, but also at the large vessels. Studies evaluating CDUS as a diagnostic tool in GCA have reported a sensitivity ranging from as low as 10% to as high as 87% ([13-18]), bringing into question the diagnostic validity of the method for clinical use. During recent years, resolution of ultrasonography images has improved substantially. Until now, no studies have examined the role of CDUS of the common carotid artery in the diagnostics of GCA.
Our aim was to evaluate 1) if CDUS examination of the common carotid artery in addition to the temporal and axillary arteries could raise the sensitivity of CDUS, 2) if the presence of a halo sign alone is a sensitive marker in the diagnosis of GCA, and 3) the sensitivity and specificity of CDUS of the temporal/axillary/common carotid arteries compared with the biopsy of the temporal artery in GCA patients.
Box 1. Significance & Innovations
Color Doppler ultrasound of the common carotid artery in addition to temporal and axillary arteries yielded a sensitivity of 100% in the diagnostics of giant cell arteritis (GCA).
The halo sign alone is a sensitive marker in the diagnostics of GCA.
Color Doppler ultrasound has a better sensitivity and a comparable specificity to the biopsy of the temporal artery.
PATIENTS AND METHODS
Patients ages ≥50 years with a clinical suspicion of GCA or in whom GCA was a potential differential diagnosis who were referred either from general practitioners or other hospital departments were routinely examined at our department between April 2010 and October 2012. The data were systematically registered as part of ordinary clinical care and practice. Furthermore, patients had to have at least one of the following criteria: C-reactive protein level <5 mg/dl, new-onset headache, jaw claudication, fever, pain in the hip and shoulder girdles, temporal artery tenderness, or recent visual impairment.
As the gold diagnostic standard, we used the positive clinical evaluation for GCA at 6 months after the initial evaluation performed by 3 rheumatologists (GM, DMS, and GH). The 3 physicians made their diagnostic decision without knowing the results of the CDUS findings. Each assessed patient was mainly judged by 1 rheumatologist and not by all 3.
Clinical findings such as new-onset headache, jaw claudication, fever, malaise, thickness or tenderness of the temporal artery, weight loss, vision disturbances, and pain in the hip and shoulder girdles were recorded. A thorough laboratory evaluation was also performed. Patients were also classified according to ACR 1990 criteria for the classification of GCA ().
Biopsy of the temporal artery
A unilateral biopsy of the temporal artery was performed when possible and mainly in all patients with a strong clinical suspicion of GCA. We did not perform a US-guided biopsy. In all patients, the biopsies were performed up to 10 days after the introduction of corticosteroid treatment in the outpatient surgery clinic of our hospital. A pathologist experienced in the diagnostics of vasculitis (RB) evaluated the biopsy specimens.
US evaluation of the arteries
CDUS was performed by a rheumatologist experienced in vascular ultrasound (APD) using a Siemens Acuson Antares US system (Siemens Medical Systems, Ultrasound Group). A high-resolution multidimensional linear array transducer VFX 13-5 MHz (Siemens Medical Systems) with a Doppler frequency above 6.5 MHz was used. The VFX 13-5 gives a high frequency range up to 13 MHz and was designed for the high-resolution imaging of superficial structures. The system settings were as follows: dynamic range 45–50 dB, wall filter low or general, depth 1.5 cm (temporal artery) or deeper according to the depth of the vessel examined (axillary and common carotid), pulse repetition frequency 2,250 Hz (temporal artery) or 3,500 Hz (axillary and common carotid), and flow persistence 1 or 2. Standard frequencies of 13.0 MHz for B-mode and 9.0 MHz for color-mode scanning with the focus at 5 mm were used. For the temporal artery, the US system was optimized to detect the lowest signal possible by selecting a pulse-repetition frequency of 2,250 Hz, a low wall filter, and a high priority.
Examination of the superficial temporal arteries (the common, frontal, and parietal branches) and the axillary and common carotid arteries in longitudinal and transverse planes was performed in all patients. The CDUS examination was performed prior to or within 1 day after the introduction of corticosteroid treatment in all patients, except for 2 patients who had been receiving corticosteroids for 1 week and 1 patient who had been receiving corticosteroids for 4 weeks. The presence of a hypoechoic ring (eccentric or circumferential) around the vessel wall (halo sign) in any part of the temporal, axillary, or common carotid arteries in both the transverse and longitudinal views was considered as a positive CDUS finding (Figures 1 and 2).
For comparisons between groups, the independent samples t-test was used for continuous variables and the chi-square test was used for categorical variables. For nonparametric numerical variables, Wilcoxon's signed rank test was used. The statistical analyses were performed using SPSS, version 17. P values less than 0.05 were considered to be significant.
The study was approved by the Institutional Review Board, Hospital of Southern Norway Trust. Due to the retrospective design of the study, informed consent was not obtained.
A total of 88 patients (54 women and 34 men) suspected to have GCA were examined. The mean age of the examined patients was 72 years in men and 70 years in women. A diagnosis of GCA based on an overall judgment of the 3 rheumatologists was made in 46 patients (13 men and 33 women; P = 0.036). Among them, 14 patients (30%) had a previous diagnosis of polymyalgia rheumatica (PMR) and developed signs and symptoms creating suspicions of GCA during followup. At the time of diagnosis, the prednisolone doses in these patients were <10 mg. Among the 42 non-GCA patients, 12 had headaches/migraines, 10 had PMR, 8 had infections, 3 had other forms of vasculitis (2 had GPA and 1 had PAN), 3 had rheumatoid arthritis, and 6 had other diagnoses (1 dissection of the thoracic aorta, 1 embolism of the temporal artery, 2 generalized atherosclerosis, and 2 malignancy). The baseline characteristics for the GCA patients and those with other diagnoses are shown in Table 1.
Table 1. Baseline characteristics of GCA patients and patients with a diagnosis other than GCA*
GCA patients (n = 46)
Patients with other diagnoses (n = 42)
Values are the number unless otherwise indicated. GCA = giant cell arteritis; CRP = C-reactive protein; ESR = erythrocyte sedimentation rate; ALP = alkaline phosphatase.
Among the 88 patients examined, 48 had a positive CDUS of the temporal arteries. Forty-four patients diagnosed with GCA had a positive CDUS of the temporal arteries. Two GCA patients with negative US of the temporal arteries had a positive US of the axillary and left carotid arteries (1 patient) and the left carotid artery (1 patient) (Table 2). In patients with a diagnosis other than GCA, the halo sign was observed in 2 patients with other forms of systemic vasculitis, in 1 patient with an infection (endocarditis), and in 1 patient with atherosclerosis.
Table 2. Distribution of large vessel involvement in the 17 large vessel GCA patients*
Large vessel involved
No. of patients with positive CDUS of the temporal artery
No. of patients with negative CDUS of the temporal artery
GCA = giant cell arteritis; CDUS = color Doppler ultrasound.
Axillary and carotid arteries
Large vessel affection
Seventeen patients (37%) with GCA (10 women and 7 men) had vasculitis of the large vessels. Two GCA patients without affection of the temporal artery (CDUS and biopsy negative) had only affection of the large vessels. The distribution of the large vessel involvement is shown in Table 2. None of the patients with a diagnosis other than GCA had a positive CDUS of the axillary and/or common carotid artery.
Sensitivity and specificity of CDUS
The sensitivity of CDUS of the temporal artery alone to diagnose GCA was 96% and the specificity was 90%, and the sensitivity and specificity of the combined CDUS examination of the temporal/axillary arteries were 98% and 91%, respectively. By adding the common carotid artery to the CDUS evaluation of the temporal and axillary arteries, the sensitivity rose to 100%. For all primary systemic vasculitides (GCA, PAN, and GPA), the specificity of CDUS was 95%.
Temporal artery biopsy
A biopsy of the temporal artery was performed in 58 of the 88 patients. The mean biopsy length after formalin fixation was 9.5 mm (range 4–16 mm). In the GCA diagnosis group, a biopsy was performed in 39 patients. Of the 7 patients with GCA who did not have a biopsy of the temporal artery performed, 1 did not consent, 3 did not receive a surgical appointment within 4 weeks, 1 had been receiving corticosteroids for 4 weeks at the diagnosis time, and 2 had surgeons who did not succeed in finding the artery. In the 39 GCA patients who had a biopsy performed, a positive biopsy was observed in 26 patients (66%). In 1 GCA patient, a peripheral nerve branch had been removed. In 1 patient with PAN, the biopsy of the temporal artery was positive with the presence of lymphocytic infiltration in the artery wall without giant cells. The sensitivity and specificity of the temporal biopsy in our series were calculated to be 67% and 95%, respectively.
All patients with a positive biopsy of the temporal artery also had a positive CDUS finding. Two GCA patients with a negative biopsy and negative CDUS of the temporal arteries had large vessel affection. The length of the temporal artery was a mean of 2.2 mm longer in GCA patients with a positive biopsy compared to GCA patients with negative biopsy (10.4 versus 8.2 mm; P = 0.04). The GCA patients with a positive biopsy had a significantly lower proportion of patients with a biopsy length <10 mm compared to GCA patients with a negative biopsy (44% versus 83%; P = 0.02).
ACR classification criteria for GCA
Among the 46 patients diagnosed with GCA, 41 patients also fulfilled the ACR classification criteria for GCA. In those with other diagnoses, 3 patients (1 with embolism of the temporal artery, 1 with PMR, and 1 with polyarthritis) fulfilled the ACR criteria. A sensitivity and specificity of 93% were calculated for the ACR classification criteria for GCA.
The main finding of this study was that the evaluation of the common carotid artery by CDUS in addition to temporal and axillary arteries raised the sensitivity to 100%, making CDUS an excellent modality to diagnose GCA. However, it is important to emphasize that US of the temporal artery alone in our study gave a sensitivity of 96% to diagnose GCA. CDUS of the common carotid artery is feasible and not time consuming because the vessel is superficial and can be investigated by high-resolution transducers. Interestingly, the affection of the carotid artery was observed in up to 27% of our GCA patients (Table 2); this observation is in accordance with a magnetic resonance angiographic study (). The investigation of the axillary and carotid arteries by CDUS seemed to be a useful method to detect vasculitis of the large vessels in patients who had both a negative CDUS and biopsy of the temporal arteries. Therefore, extending the US examination to also include large vessels, for example, the axillary and common carotid arteries, may further increase the sensitivity of the method to diagnose vasculitis. In our cohort, the prevalence of large vessel vasculitis among GCA patients was 37%, which is in agreement with previous US studies ([19, 20]).
The diagnostic value of CDUS in GCA is currently being validated in an international multicenter study, the Temporal Artery Biopsy Versus Ultrasound in Diagnosis of Giant Cell Arteritis (TABUL) study (). This observational, prospective study aims to validate CDUS as a diagnostic tool in GCA. In addition, the interobserver agreement and cost-effectiveness of CDUS instead of biopsy in the diagnosis of GCA will be examined. The data from this study are expected to be available in 2015.
In previous studies, a combination of a halo sign and stenoses has been recommended to be used to raise the sensitivity and specificity for the CDUS method ([14, 15]). Our study shows that the presence of a halo sign alone is a sensitive marker for the diagnosis of GCA; therefore, this makes the use of CDUS more feasible because it requires less training to detect a halo than stenosis.
In the present study, CDUS had a better sensitivity than the biopsy of the temporal artery and a comparable specificity to diagnose GCA. An advantage of the CDUS is that the method is noninvasive and thus easier to apply in all patients suspected to have GCA. This was clearly shown in our study because all GCA patients had a CDUS assessment, whereas only 39 patients had a biopsy performed. In addition, to perform a biopsy of the temporal artery in the real world is not without problems, as we clearly delineated in our cohort of GCA patients. Apart from being a noninvasive, feasible, and patient-friendly method, the advantage for using CDUS is that the whole length of the temporal artery with its branches can be examined. This is the main benefit to using CDUS because vessel wall inflammation may occur in a patchy pattern, explaining why inflammation can be missed in biopsies.
Our study confirms the crucial role of the biopsy length in the diagnosis of GCA. The length of the temporal artery biopsy is recommended to be at least 1 cm (). In the present study, 83% of GCA patients with a negative biopsy had a biopsy length <1 cm compared with 44% of GCA patients with a positive biopsy. To our knowledge, this is the first CDUS study in GCA patients confirming that the biopsy length should be <1 cm to raise the sensitivity of the biopsy to diagnose GCA.
In our study, 1 patient with a positive biopsy of the temporal artery and positive findings on US was diagnosed with PAN. This confirms previous reports that even the presence of inflammation in the vessel wall of the temporal artery is not pathognomonic for the diagnosis of GCA ([6-8]).
In a diverse group of studies, the sensitivity of CDUS to diagnose GCA has been reported to range from as low as 10% to up to 87% ([13-15, 17, 18, 22]). This could partly be explained by the quality of equipment that has been applied. In several studies, equipment with a B-mode frequency below 10 MHz was used, which is limited by relatively poor resolution. We used a high-frequency (13–5 MHz) B-mode multilinear transducer with a high frequency Doppler of 9 MHz, which is equipment that meets the high standards defined in the literature ([23, 24]). Of interest, in the 2 studies describing the lowest sensitivity and specificity of CDUS ([13, 22]), no data on Doppler frequency or adjustments were given. Furthermore, the method is also highly operator dependent and the experience of the person performing the CDUS is a critical factor. Our study highlights that CDUS performed with modern equipment by a trained rheumatologist is a sensitive method to diagnose GCA.
Interestingly, we found that patients with a positive CDUS and diagnoses other than GCA had other forms of systemic vasculitides (GPA and PAN), infection, or atherosclerosis. Vasculitis of the temporal arteries other than GCA has been described in pathologic ([6, 7]) and US ([16, 24]) studies. It appears that infections can also give intima-media complex thickness that presents as a halo sign in CDUS due to secondary vasculitis. The presence of typical vasculitic changes in the biopsies of temporal arteries in patients with infections (Lyme disease) has been described (). In addition, halo signs in patients with infections have also been observed in a US study (). Thus, the presence of a halo sign in a patient with inflammation does not exclude a possible infectious disease. A thorough clinical evaluation should be performed in order to exclude an infection in all patients presenting with inflammation signs and clinical symptoms suggesting an infection.
Our study has limitations. For example, the ultrasonographer (APD) was not blinded to the clinical and laboratory findings. This may have biased the results by overdiagnosing GCA in patients without the disease. However, the 6-month evaluation of the GCA patients, defined as the gold standard in this study, did not alter the initial diagnosis in any of the GCA patients. Another limitation of our study is that we were not able to explore the reliability of US in diagnosing vasculitis, because only 1 investigator performed the CDUS examinations.
In conclusion, the combination of the CDUS examination of the temporal, axillary, and carotid arteries had an excellent sensitivity to diagnose GCA. The presence of the halo sign alone in temporal arteries and large vessels was a reliable sign to diagnose GCA by CDUS. Therefore, CDUS has the potential to replace biopsy in the diagnostics of GCA without compromising on sensitivity and specificity.
All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Diamantopoulos had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study conception and design. Diamantopoulos, Haugeberg, Myklebust.
Acquisition of data. Diamantopoulos, Haugeberg, Hetland, Soldal, Bie, Myklebust.
Analysis and interpretation of data. Diamantopoulos, Hetland, Myklebust.