Key words
vascular - ultrasound color Doppler - vasculitis - temporal artery
Introduction
Giant cell arteritis (GCA) is the most common idiopathic vasculitis syndrome, affecting
medium and large-sized arteries. Although the involvement of large arteries throughout
the body is increasingly recognized, the only direct evidence for the diagnosis remains
the histological finding of characteristic vascular inflammation in biopsy samples
[1]
[2]. Due to the segmental nature of the inflammatory process in GCA (“skip lesions”),
biopsies – most commonly performed on temporal arteries (TA) – can provide false-negative
results, hampering the sensitivity of this diagnostic procedure [3].
In 1997, Schmidt et al. introduced color duplex ultrasound (CDU) as a noninvasive
method for differentiating inflamed from non-inflamed TA segments [4]. In this seminal publication, the halo sign was demonstrated to provide excellent
accuracy for the diagnosis of GCA. Subsequent studies reported less convincing results,
pointing at relevant operator dependency (experience, equipment) of the diagnostic
approach as established by Schmidt et al. [5]
[6]. In search of a simple, robust and less operator-dependent ultrasound-based diagnostic
tool, we used a prospective cohort of individuals with suspected GCA to compare the
diagnostic accuracy of the halo sign and a novel vascular ultrasound phenotype – namely
visibility vs. loss of visibility of the TA upon transducer-imposed artery compression.
Patients and Methods
Consecutive patients with a clinical suspicion of GCA were included in the study between
3/2009 and 9/2011 at the University Hospital Basel, Switzerland. The study was conducted
as a nested project of the IRB-approved prospective Basel Riesenzell-Arteritis Kohorte
(‘BARK’) study. TA biopsy was performed if deemed indicated by the referring physician.
The biopsy side was guided by CDU in the case of a positive finding. Definite diagnosis
after extensive clinical, laboratory and imaging evaluation was established by a team
of independent and experienced rheumatologists/immunologists not involved in CDU according
to ACR criteria [1]
[7], and patients were accordingly grouped as GCA vs. non-GCA.
Bilateral CDU evaluation of the TA was performed in all study participants, using
an iU22 duplex ultrasound (Philips, Best, Netherlands) with a linear 5 – 17 MHz broadband
transducer. The B-scan and Doppler settings were adjusted for best assessment of the
arteries. The TA (trunk, frontal and parietal branch) was scanned in the cross section;
the longitudinal section was used to detect stenoses. Blinded to the patients’ diagnosis,
a team of 3 senior physicians – each with more than 6 years of CDU experience – performed
the ultrasound examinations. One physician performed the entire examination in any
given patient. The order of examination (halo sign vs. compression sign) was left
to the discretion of the examiner.
Compression sonography was conducted as follows: slight pressure was applied via the
transducer to occlude the TA. Pressure was applied until the lumen was occluded and
no arterial pulsation remained visible by B-mode. The contact area of the transducer
with the scalp was positioned so that osseus structures did not impede the compression
of the artery. The compression sign was considered negative (absence of evidence for
vasculitis) if the arterial wall did not contrast with the surrounding tissue upon
compression. If, however, the arterial wall remained visible upon compression, i. e.,
echogenicity different than the surrounding tissue due to vessel wall inflammation,
the compression sign was considered positive (evidence of vasculitis) ([Fig. 1]). The halo sign, stenoses or occlusion was searched for as described elsewhere in
detail [4].
Fig. 1 The compression sign. a Loss of visibility of the temporal artery (lumen and wall) upon transducer-imposed
compression (= compression sign negative/lack of evidence of vasculitis). b Visibility of the temporal artery wall maintained upon transducer-imposed compression
(= compression sign positive/evidence of vasculitis). Arrows indicate temporal arteries.
Abb. 1 Das „Kompressions-Zeichen“. a Temporalarterie (Lumen und Wand) ist unter vollständiger Kompression nicht mehr nachweisbar
(= negatives „Kompressions-Zeichen“/kein Hinweis auf Vaskulitis). b Weiterhin sichtbare Gefäßwand unter vollständiger Kompression der Temporalarterie
(= positives „Kompressions-Zeichen“/Nachweis einer Vaskulitis). Pfeil weist auf Temporalarterie.
Statistical analysis was performed using SPSS/PC (version 19.0, SPSS Inc, Chicago).
Comparisons between patients categorized as GCA vs. non-GCA were made using unpaired
Student’s t-tests for independent samples or Mann-Whitney-U-tests and Chi-square tests,
as appropriate. All hypothesis testing was two-tailed.
Results
Patient Characteristics
The mean age of the 80 study participants was 72 years (± 10 years SD), and 69 % of
the patients were female. All study participants were referred for vascular ultrasound
based on the clinical suspicion of GCA. In 43/80 study participants (54 %), the final
diagnosis was GCA. The remaining 37 patients (46 %) in the non-GCA group were diagnosed
with polymyalgia rheumatica (n = 14), rheumatoid arthritis (n = 3), carcinoma (n = 4),
amaurosis fugax (n = 2), anterior ischemic optic neuropathy (n = 2), small vessel
vasculitis (n = 3), and retinal detachment, pneumonitis, enthesitis, neuralgia of
the trigeminal nerve, ulcerative proctitis, diabetes mellitus, syncope, coronary artery
disease, and familial Mediterranean fever (n = 1 in each case).
In 53/80 study participants (66 %), a TA biopsy was performed. In the 43 patients
with GCA, 38 underwent a TA biopsy, among them 20 showed active vasculitis. One biopsy
revealed only a venous segment and was therefore not conclusive. In the 37 patients
categorized as non-GCA, 15 underwent a TA biopsy. None of them showed active vasculitis.
At the time of CDU examination, steroid therapy was initiated in 47/80 study participants
(59 %), and at the time of biopsy in 45/53 individuals (85 %). Patient characteristics,
including the median number of days participants were treated with steroids and the
median number of days CDU was performed before TA biopsy in each subgroup, are summarized
in [Table 1].
Table 1
Patient characteristics.[1]
|
GCA (n = 43)
|
Non-GCA (n = 37)
|
p-value
|
|
female n (%)
|
25 (58)
|
30 (70)
|
n. s.[2]
|
|
age years (mean±SD)
|
75 ± 8
|
68 ± 11
|
n. s.2
|
|
clinical presentation n (%)
|
|
|
|
|
fever
|
7 (16)
|
5 (14)
|
n. s.[3]
|
|
vision disorder
|
11 (26)
|
11 (30)
|
n. s.3
|
|
headache
|
26 (60)
|
15 (42)
|
n. s. 3
|
|
jaw claudication
|
20 (47)
|
3 (8)
|
< 0.0013
|
|
tenderness on scalp
|
13 (30)
|
6 (17)
|
n. s. 3
|
|
prominent TA
|
13 (30)
|
4 (11)
|
0.0283
|
|
polymyalgia
|
16 (37)
|
17 (47)
|
n. s. 3
|
|
weight loss
|
21 (49)
|
3 (8)
|
< 0.0013
|
|
claudication
|
1 (2)
|
2 (5)
|
n. s. 3
|
|
laboratory values
|
|
|
|
|
ESR mm/h
|
76 (61.95)
|
42 (11.60)
|
0.003[4]
|
|
CRP mg/l
|
65 (36.114)
|
13 (6.47)
|
0.0044
|
|
Lc x109/l (mean±SD)
|
10.2 ± 2.8
|
9.3 ± 3.4
|
n. s.3
|
|
Hb g/l (mean±SD)
|
116 ± 22
|
118 ± 26
|
n. s.3
|
|
Tc x109/l (mean±SD)
|
436 ± 171
|
347 ± 171
|
n. s.4
|
|
biopsy performed n (%)
|
38[5] (88)
|
15 (41)
|
< 0.0013
|
|
on steroids before CDU (days)
|
1 (0.3)
|
2 (0.5)
|
n. s.4
|
|
on steroids before biopsy (days)
|
4 (2.7)
|
4 (2.8)
|
n. s.4
|
|
days between CDU and biopsy
|
3 (1.5)
|
2 (1.5)
|
n. s.4
|
1 TA = temporal artery, ESR = erythrocyte sedimentation rate, CRP = C-reactive protein,
Lc = leucocytes, Hb = hemoglobin, Tc = thrombocytes, CDU = color duplex ultrasound.
2 Student’s t-test.
3 Chi-square-test.
4 Mann-Whitney-U-test.
5 One biopsy revealed only a venous segment. ESR and CRP, on steroids before CDU and
biopsy, respectively, as well as days between CDU and biopsy are given as median (25 – 75
percentile).
Diagnostic Performance of Vascular Ultrasound
34/43 patients categorized as GCA (79 %) had a positive compression sign and a positive
halo sign, i. e., there was a 100 % congruency between the compression and halo sign.
None of the patients in the non-GCA group had a positive compression or halo sign.
Based on these findings, the sensitivity and specificity and negative and positive
predictive value of the compression and halo sign for the GCA and non-GCA groups were
79 % and 100 %, and 80 % and 100 %, respectively.
All 20 patients with active vasculitis on TA biopsy showed a positive compression
and halo sign. In 10 patients with a positive compression and halo sign, the TA biopsy
was negative. None of the patients lacking the compression and halo sign had active
vasculitis on TA biopsy.
Stenosis or occlusion of the TA was detected infrequently (12 % and 7.5 % among individuals
in the GCA group, respectively, none in the non-GCA group, respectively). Of note,
stenosis or occlusion was only observed in patients with a positive compression and
halo sign. These CDU findings thus did not provide additional diagnostic information.
Discussion
Vascular ultrasound as a noninvasive method able to monitor the TA (as well as many
other arteries) in their entirety – hence overcoming the issue of inflammatory “skip
lesions” – is widely accepted as a diagnostic tool of choice in patients with suspected
GCA. Indeed, CDU has been shown to have a high sensitivity and excellent specificity
in patients with suspected GCA assessed in experienced centers [1]
[4]. However, other studies (including meta-analyses) indicated that the diagnostic
performance of currently established CDU criteria is highly operator-dependent and
is thus susceptible to producing misleading artifacts if not applied by trained experts
using appropriate equipment [5]
[8]
[9]. Indeed, as exemplified in [Fig. 2], a thickened vessel wall (halo sign) can easily be missed or wrongfully diagnosed
if the respective color gain is chosen inappropriately high or low.
Fig. 2 Inadequate color-gain may impact interpretation when searching for a halo.Identical
setting on CDU scanner in all panels, except for color-gain. a and b Healthy individual: color-gain too low – mimicking a halo sign a. Appropriate color-gain (b). c and d Individual with biopsy-proven temporal arteritis: appropriate color-gain demonstrating
a halo (c). Over-steered color-gain covering the thickened arterial wall (d).
Abb. 2 Nicht korrekter Color-Gain beeinflusst Interpretation des „Halo-Zeichens“. Identisches
Apparate-Setting auf allen Bildern mit Ausnahme des Color-Gain. a und b Gesundes Individuum: Color-Gain zu tief – imitiert ein „Halo-Zeichen“ a. Korrekter Color-Gain (b). c und d Patient mit bioptisch nachgewiesener Temporalarteritis: Color-Gain korrekt mit Nachweis
eines „Halo-Zeichens“ (c). Color-Gain übersteuert – Überdeckung der verdickten Gefäßwand (d).
Against this background, we tested a new and simple diagnostic ultrasound criterion
– termed ‘compression sign’ – relying on the B-mode modality only and thus avoiding
the above-mentioned drawbacks of CDU. The diagnostic performance of the compression
sign was assessed on the basis of the groups GCA and non-GCA established based also
on ACR criteria by a team of independent, experienced physicians. Notably, the diagnostic
performance of the compression sign and the halo sign was identical, and in all cases
in which the TA biopsy was positive, the compression sign and the halo sign were positive
as well. The fact that the compression sign was negative in 9/43 study participants
in the GCA group is consistent with the fact that extensive inflammation of the TA
is not always present in GCA patients [10]
[11]. Extended vascular ultrasound examination – as previously reported by our group
– or additional imaging modalities may thus have to be considered as part of the diagnostic
work-up in individuals with suspected GCA [1]
[12]
[13].
There are several limitations in this study. First, the number of patients included
is relatively small. Second, although the examiner was unaware of the diagnosis at
the time of CDU, the compression sign was not tested independently from the halo sign
and thus such a bias cannot be completely excluded. Third, the study design also does
not allow for interobserver comparison for the two signs individually.
In conclusion, in this cohort of individuals with suspected GCA, the halo sign and
the compression sign were equal with respect to their diagnostic performance. The
simplicity of the compression sign suggests a clinical potential warranting further
evaluation. Clearly, further prospective studies are needed to investigate this new
ultrasound finding in different patient cohorts and study centers.
Funding
Swiss National Science Foundation Grant (SNSF 31 003A_135 677) to CH.
Acknowledgments
We thank all contributors to the ‘BARK’ cohort study.
