ABSTRACT
Giant cell arteritis (GCA) is often categorised as ”active” or ”healed” on temporal artery biopsy (TAB). The purpose of this study was to compare the initial clinical presentation of patients with GCA according to active versus healed arteritis on TAB. A retrospective chart review was performed for patients with biopsy-proven GCA (BP-GCA) at a single academic medical institution from a previously reported cohort. The arteritis on TAB was categorised as “active” or “healed” based on the pathological reports. Demographic information, clinical presentation, past medical history, and test results were collected from the date of TAB. These baseline characteristics were entered into the GCA Risk Calculator. Of 85 patients with BP-GCA, 80% had active and 20% had healed disease according to histopathology. A higher percentage of those with active arteritis had ischaemic optic neuropathy (ION) (36% versus 6%, p = .03), elevated erythrocyte sedimentation rates (92% versus 63%, p = .01), elevated C-reactive protein levels (79% versus 46%, p = .049), GCA risk score > 7.5% (99% sensitivity, 100% versus 71%, p < .001), higher mean GCA risk calculator scores (neural network p = .001; logistic regression p = .002). Patients with healed arteritis were less likely to have visual manifestations than the active arteritis group (38% versus 71%, p = .04). Patients with active vasculitis on biopsy had higher rates of ION and elevated inflammatory markers, as well as higher predictive scores from the GCA risk calculator. Further research is needed regarding correlation of biopsy findings and risk of complications or relapses.
KEYWORDS: Giant cell arteritis, temporal arteritis, vasculitis, temporal artery biopsy, pathology
Introduction
Giant cell arteritis (GCA) is the most common systemic vasculitis in the elderly and can lead to permanent vision loss if left untreated.1 Temporal artery biopsy (TAB) is often relied upon as the gold standard for diagnosis of GCA by ophthalmologists, with a positive TAB result indicative of GCA reported as either ”active” or ”healed.” Pathologically, active arteritis is defined by chronic inflammation of the vessel wall, which is usually granulomatous and includes multinucleated giant cells, involving the muscular media and internal elastic lamina.2 Healed arteritis, on the other hand, is typically characterised by post-inflammatory alterations that are suggestive of resolving inflammation, such as thinning, scarring, or breaks in internal elastic lamina.2 Unlike GCA with active arteritis, GCA with healed arteritis remains largely unstudied and is poorly understood clinically. Despite the pathological differences between active and healed arteritis, only a few studies published in the rheumatology literature have examined the differences between these two groups with regard to clinical presentation, treatment, and outcomes.3,4 To our knowledge, only one study from the ophthalmology literature has examined this issue.5 As a result, it remains largely unclear whether or not a pathological diagnosis of healed arteritis is associated with differences in clinical presentation.
The objective of this study was to look for differences in clinical presentation between patients with healed and active arteritis at the time of biopsy (i.e. at baseline).
Methods
Study design and population
We performed a retrospective review of the medical records of patients found to have biopsy-proven GCA (BP-GCA) in our previous study, which examined the incidence of GCA in white and black patients.6 The present study was approved by the institutional review board of the Johns Hopkins School of Medicine. As this was a retrospective study with minimal risk, the need for informed consent was waived. Additionally, this study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guidelines.
The study cohort consisted of the 92 BP-GCA patients as previously reported.6 In summary, these patients were referred for TAB from the Wilmer Eye Institute and other departments of the Johns Hopkins Health System, including satellite sites. A patient was categorised as having BP-GCA if histological examination of the temporal artery showed evidence of arteritis. The categorisation of healed versus active arteritis was based on the pathological report, using criteria described below, and all cases were reviewed by a pathologist at the time of diagnosis. Patients with indeterminate biopsy findings were considered to be negative for GCA and excluded from the study. We also excluded patients who were never put on steroids, or who were on a steroid course of less than 1 month, monitored without recurrence, and deemed clinically not to have GCA.
Active arteritis cases included all those with chronic inflammation involving the internal elastic lamina and/or muscular media. Changes in chronic, ”healed” arteritis include thinning, scarring and/or neovascularisation of the muscular media, as well as significant (generally more than one-third of the circumference) loss of the internal elastic lamina. Limited adventitial lymphocytic (non-granulomatous) inflammation was not considered sufficient for a diagnosis of active GCA if it was focal and did not involve the muscular media.
The same racial and ethnic labels from our previous study were used in the present study.5
Clinical data elements
Demographic information, clinical presentation, past medical history, and test results were collected from each patient’s medical record. All data were collected prior to formal diagnosis of GCA, defined as date of TAB and subsequent pathology report.
Visual symptoms collected at baseline included the presence of transient monocular vision loss, ischaemic optic neuropathy (ION; including both anterior and posterior ION), central retinal artery occlusion (CRAO), and diplopia (either transient or cranial nerve palsy). Non-visual cranial symptoms collected included the presence of headache, superficial temporal artery tenderness, jaw claudication, and scalp tenderness. The presence of polymyalgia rheumatica (PMR) and unintentional weight loss were also noted.
Baseline test results collected included erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), and platelet count prior to initiation of steroids. Test results of patients initiated on steroids prior to formal diagnosis of GCA were collected from the closest date prior to initiation of steroid use. Elevated ESR was defined as levels being above the upper limits of normal (ULN) for age (age/2 in men and (age+10)/2 in women). Elevated CRP was defined as being above 2.5 × the ULN given by the laboratory.7 Elevated platelet counts were defined as being greater than 400,000/µL (400 K).
GCA risk calculator
The GCA Risk Calculator was developed and validated by Ing et al.8 Their two models, the neural network and logistic regression diagnostic prediction models, were created for patients with suspected GCA. According to their models, patients with a 7.5% observed probability cut-off point have a low risk of GCA, with a statistical sensitivity of 99%.
Baseline symptoms for each patient were inputted into their online calculator to determine GCA risk scores, and patients with a mean risk of 7.5% or greater were noted. Of note, 15 patients from this study’s population were used as part of their model.
Statistical analysis
Baseline demographic information, clinical presentation, past medical history, test results, and calculated GCA risk score were summarised and compared between the active and healed groups. T test was used to compare continuous variables. Chi-square test was used to compare categorical variables. A significance level of .05 was used. All analyses were performed in R software (R core team, Vienna, Austria).9
Results
Study participants
Of the 92 biopsy-proven patients from our previous study, five were excluded due to incongruity with clinical diagnosis & course as well as two for diagnosis of relapsed GCA rather than the initial presentation; leaving 85 patients for analysis, including 71 white (84%), 12 black (14%), one (1%) unknown, and one (1%) other race. Seventy-one percent were women, and the median age at the time of diagnosis was 76 years of age (range 53–94). The patient demographics are shown in Table 1.
Table 1.
Baseline demographics of patients with biopsy-proven giant cell arteritis, overall and separated by biopsy classification.
| Overall | Active | Healed | p value | |
|---|---|---|---|---|
| Mean age (median, IQR) | 76 (53, 94) | 77 (53, 94) | 70 (61, 90) | .5 |
| Female (%) | 60/85 (71%) | 45/68 (66%) | 15/17 (88%) | .2 |
| Unintentional weight loss (%) | 33/83 (40%) | 27/66 (41%) | 6/17 (35%) | .9 |
| Polymyalgia rheumatica (%) | 31/83 (39%) | 26/66 (39%) | 6/17 (35%) | 1 |
| Diabetes mellitus (%) | 17/82 (21%) | 11/65 (17%) | 6/17 (35%) | .1 |
| White race | 71/85 (84%) | 60/68 (88%) | 11/17 (65%) | .09 |
IQR = interquartile range.
p < 0.05 considered statistically significant.
Data regarding timing of steroids and laboratory testing were available for 75 patients. The mean time from laboratory testing to the start of steroids was 6 days (standard deviation [SD] 20 days). Four patients (5%) had laboratory tests drawn more than 30 days prior to steroids); only three patients (4%) were started on steroids prior to laboratory testing.
Data regarding timing of steroids and biopsy were available for 83 patients. Mean biopsy time from the start of steroids was 6 days after (SD 9 days). Patients on steroids for more than 2 weeks prior to biopsy made up 8% of the studied cohort.
At least one visual manifestation was present at baseline in 55% of the cohort, including ION (30%), transient monocular vision loss (25%), diplopia (23%), and CRAO (5%). At least one non-visual cranial manifestation was present at baseline in 88% of the population, including headache (79%), jaw claudication (51%), superficial temporal artery tenderness or reduced pulse (35%), and scalp tenderness (35%). Thirty-nine percent of patients had PMR and 40% had unintentional weight loss. Presenting cranial and visual signs and symptoms are shown in Table 2.
Table 2.
Presenting cranial and visual signs and symptoms of patients with biopsy-proven giant cell arteritis, overall and separated by biopsy classification.
| Overall | Active | Healed | p value | |
|---|---|---|---|---|
| Headache (%) | 66/84 (79%) | 52/67 (78%) | 14/17 (82%) | .9 |
| Superficial temporal artery tenderness or reduced pulse (%) | 29/83 (35%) | 22/66 (33%) | 7/17 (41%) | .7 |
| Jaw claudication (%) | 43/84 (51%) | 38/67 (57%) | 5/17 (29%) | .07 |
| Scalp tenderness (%) | 29/83 (35%) | 21/66 (32%) | 8/17 (47%) | .3 |
| No non-visual cranial manifestations (%) | 10/85 (12%) | 8/68 (12%) | 2/17 (12%) | 1 |
| Transient monocular vision loss (%) | 21/83 (25%) | 19/66 (29%) | 2/17 (12%) | .3 |
| Ischaemic optic neuropathy (%) | 25/83 (30%) | 24/66 (36%) | 1/17 (6%) | *.03 |
| Central retinal artery occlusion (%) | 4/82 (5%) | 4/66 (6%) | 0 (0%) | .7 |
| Diplopia (%) | 19/83 (23%) | 16/66 (24%) | 3/17 (18%) | .8 |
| No visual manifestations (%) | 38/85 (45%) | 26/68 (38%) | 12/17 (71%) | *.04 |
*p < 0.05 considered statistically significant.
Eighty-six percent of the patients had elevated ESR levels, 72% had elevated CRP levels, and 28% had elevated platelet counts.
The mean risk for GCA using the GCA risk calculator was 51% using the neural network model, and 43% under the logistic regression model. A risk of 7.5% or greater in the neural or logistic regression model (sensitivity 99%) was present in 94% of the patient population. Laboratory and GCA calculator scores are shown in Table 3.
Table 3.
Presenting laboratory findings and composite risk score of patients with biopsy-proven giant cell arteritis, overall and separated by biopsy classification.
| Overall | Active | Healed | p value | |
|---|---|---|---|---|
| High ESR (%) | 66/77 (86%) | 56/61 (92%) | 10/17 (63%) | *.01 |
| High CRP (2.5× ULN) (%) | 50/69 (72%) | 44/56 (79%) | 6/13 (46%) | *.049 |
| High platelet count (above 400,000/µl) (%) | 21/85 (28%) | 19/61 (31%) | 2/14 (14%) | .3 |
| Normal ESR, CRP, & platelet count (%) | 15/85 (18%) | 9/68 (13%) | 6/17 (35%) | .08 |
| Mean GCA risk score LR (SD) | 43 (25) | 47 (24) | 25 (22) | *.002 |
| Mean GCA risk score NN (SD) | 51 (24) | 55 (22) | 33 (25) | *.001 |
| GCA risk score>7.5% (99% sensitivity) (%) | 73/78 (94%) | 61/61 (100%) | 12/17 (71%) | *< .001 |
CRP = C-reactive protein; ESR = estimated sedation rate; GCA = giant cell arteritis. LR = logistic regression; NN = neural network; SD = standard deviation.
*p < .05 considered statistically significant.
Active versus healed biopsy
Of the 85 patients, 17 (20%) were diagnosed with healed arteritis, while 67 (80%) had active arteritis based on the pathological reports. There were no significant differences in gender or race between the two groups.
A higher percentage of patients with active arteritis versus healed arteritis had ION (36% versus 6%, p = .03). Patients with healed arteritis were more likely to have no visual manifestations (71% versus 38%, p = .04) compared with patients with active arteritis. Patients with active arteritis also had higher rates of elevated ESR (92% versus 63%, p = .01) and elevated CRP levels (79% versus 46%, p = .049). Additionally, patients with active arteritis on biopsy had a higher mean GCA risk score using the neural network model (55% versus 33%, p = .001) and the logistic regression model (47% versus 25%, p = .002). A higher proportion of patients with active arteritis had a GCA risk score > 7.5% (99% sensitivity) in either model compared with patients with healed arteritis (100% versus 71%, p < .001). Furthermore, there were trends towards more jaw claudication (57% versus 29%, p = .07) and at least one elevated inflammatory marker (87% versus 65%, p = .08) among patients with active arteritis. There were no significant differences in any of the other baseline characteristics collected.
Duration of steroids prior to biopsy was not significantly different between the active and healed groups; likewise, the percentage of patients on steroids for more than 3 weeks, 2 weeks, or 1 week prior to biopsy was not significantly different between these two groups (Table 4).
Table 4.
Duration in days on steroids before biopsy.
| Active | Healed | p value | |
|---|---|---|---|
| Mean (SD) | 5 (8) | 8 (11) | .4 |
| Minimum/maximum | § −4/62 | § −1/34 | |
| >3 weeks duration n (%) | 1 (2%) | 2 (12%) | .3 |
| >2 weeks duration n (%) | 3 (5%) | 4 (24%) | .1 |
| >1 week duration n (%) | 14 (21%) | 5 (29%) | .5 |
SD = standard deviation.
*p < .05 considered statistically significant.
§negative number indicates steroids started after biopsy.
Discussion
Little is known about the distinction between active and healed TAB beyond histopathological differences. The potential implications of this pathological feature for clinical presentation, treatment, and outcomes of GCA remain unclear. In this study, we found that patients with active arteritis presented with higher rates of ION, higher rates of having at least one visual manifestation, and higher rates of elevated inflammatory markers, as well as higher predictive scores from the GCA risk calculator. However, while these differences existed, patients with healed arteritis still shared many features with patients with active arteritis. There was no statistical difference in mean age, rates of non-visual cranial manifestations, transient monocular vision loss, CRAO, diplopia, elevated platelet levels, weight loss, diabetes mellitus, or PMR. Ultimately, we found that while the two groups shared several similar baseline characteristics, patients with healed arteritis may present with milder symptoms than patients with active arteritis. Though this may suggest that patients should receive different treatment based on biopsy results, we still found that the incidence of ION in patients with healed arteritis was 6%, indicating that the risk of vision loss in this population is still present. Patients are treated based on their clinical presentations rather than their pathological diagnoses alone. To our knowledge, only one study in rheumatology3 and one study in ophthalmology5 have compared presenting symptoms in GCA patients with active versus healed biopsy. The ter Borg et al. study evaluated 44 patients and found that the only statistically significant differences between patients with healed and active arteritis at baseline were higher ESR and lower haemoglobin levels in the active arteritis groups.3 The Salter et al. study similarly evaluated 22 patients and found that the only statistically significant differences between the healed and active groups at presentation were the average CRP and average platelet count.5 Although they noted more than double the number of initial ocular ischaemic events in the active arteritis group (seven versus three) compared with the healed arteritis group, they did not report this as statistically significant, likely owing to small numbers. The study also commented that their active arteritis group was more symptomatic on initial presentation though this conclusion is limited by the study’s small sample size. Given the conflicting findings from these two studies, there is no consensus on whether clinical presentation differs when comparing these two histological subtypes.
Our study includes more patients than these two previous studies, and our results confirm several of their findings; a majority of baseline characteristics were not significantly different, such as the non-visual cranial and constitutional symptoms. In keeping with the aforementioned studies, we found that patients with active arteritis had higher rates of elevated ESR and CRP levels. However, contrary to these two small studies that found no statistically significant difference in visual manifestations, we found that patients with active arteritis on biopsy presented with higher rates of ION while patients with healed arteritis were more likely to have no visual manifestations. This may reflect a difference in sampling or inclusion criteria, but it is important to remember that the risk of vision loss in patients with healed arteritis still exists.
We also found that patients with active arteritis had higher GCA risk calculator scores and higher rates of a risk score > 7.5%, the cut-off threshold for 99%.8 In the development and validation of the GCA risk calculator, both healed and active arteritis were considered positive for GCA. As these two populations were grouped together, it remains untested whether the GCA risk calculator would be equally useful in patients with active and healed arteritis. Since patients with healed arteritis overall had lower risk calculator scores, this suggests that the risk calculator may not be as effective at capturing GCA patients with healed arteritis. This could potentially lead to underdiagnosis of GCA in these patients who are still at risk of ischaemic complications if clinicians rely primarily on the risk calculator.
With regard to the question whether or not healed arteritis on TAB is indicative of a more “benign” subset of GCA, our study, in accordance with Salter et al. and ter Borg et al., suggests that patients with healed arteritis may present with comparatively milder symptoms than patients with active arteritis. Based on these findings, both Salter et al. and ter Borg et al. suggest that patients with healed arteritis may require lower doses of steroids and appear to have better prognoses and fewer major complications than patients with active arteritis on TAB.3,5 However, these findings and suggestions are based on very small patient cohorts. As our findings show, individual patients with healed arteritis may nevertheless present with ION and CRAO.
Studies have shown that the distribution of GCA along the temporal artery is not uniform – active, healed, or normal artery segments may be found within the same patient.10,11 Diagnosing healed arteritis also has a larger degree of inter/intra-observer variability compared with diagnosing active arteritis on biopsy.12 In one study that re-examined 47 temporal artery specimens initially diagnosed as healed, only 32% were confirmed as healed when using standardised histological criteria. Four percent of these biopsies were reclassified as active arteritis, and 64% were reclassified as atherosclerotic or age-related changes.4 Despite the diagnostic uncertainty surrounding TABs and suggestion of a milder disease course with healed arteritis in our study and previous studies, we believe that it is important for clinicians to treat healed arteritis as aggressively as active arteritis in the context of clinical signs and symptoms supporting a diagnosis of GCA, until there is sufficient evidence regarding treatment outcomes in these groups to suggest alterations in therapy. With high clinical suspicion, clinicians should feel comfortable treating patients with healed biopsy with appropriate immunosuppression in order to prevent the most feared complications of GCA, which include permanent vision loss and stroke.
Further research is needed to help determine whether or not healed arteritis on biopsy influences treatment, outcomes, and risk of complications.
Our study was limited by its retrospective and single centre nature. While only a single centre, the Johns Hopkins Health System is a tertiary care centre and one of the major health care institutions in Maryland. Although the minority patient population is relatively large, its majority is still white (approximately 60%), and the results may therefore not be generalisable to other populations.
Most studies examining the impact of corticosteroid use on biopsy results have looked at positive versus negative results in comparison with clinical diagnosis of GCA. One study found an 82% positive rate in patients with a clinical diagnosis of GCA prior to corticosteroid therapy, 60% positive rate for steroids of less than 1 week, and only 10% among patients on steroids longer than 1 week.13 However, in another study, they did not identify a difference in the median duration of steroids among positive and negative biopsies or association of pathological findings compared with the duration of steroids, but did find an increased odds ratio of positive biopsy and inflammatory infiltrates among patients with less than 1 day of steroids compared with those on longer duration.14 Still, a third study showed declining rates of positive biopsy with duration (78% for<2 weeks of steroids, 65% for 2–4 weeks of steroids, 40% of those > 4 weeks of steroids).15 Yet a fourth study showed no association between duration of steroids but did find an association with cumulative steroid dose.16 While our study used different inclusion criteria (positive biopsy examining healed versus active arteritis), we did see a trend towards longer steroid duration in the healed biopsy group, although this did not meet statistical significance. However, the rate of long duration of steroids prior to biopsy was low in our cohort.
Furthermore, in our study, the pathologist’s initial diagnosis was used, instead of re-examining biopsy specimens. Use of the pathologist’s initial biopsy diagnosis is more representative of “real world” data and the typical clinical experience.
The current data show that GCA patients with active arteritis on biopsy results have significantly higher rates of visual manifestations (specifically, ION), elevated inflammatory markers, and predictive scores from the GCA risk calculator. As such, patients with healed arteritis may represent a distinct patient population that may present with comparatively milder symptoms. However, the risk of vision loss in patients with healed arteritis is still present and may warrant aggressive treatment in the appropriate clinical setting and with signs and symptoms of active disease. Ultimately, to prevent complications from GCA, physicians must rely on clinical features when determining how aggressively to treat patients with healed arteritis, weighing these against the potential side effects of immunosuppressive therapy. Future studies are needed to provide more data on steroid dosing for patients suspected of having GCA who are found to have healed arteritis on biopsy.
Funding Statement
This work was supported by the P30 Core Grant (P30EY001765).
Disclosure statement
No potential conflict of interest was reported by the authors.
References
- 1.Soulages A, Sibon I, Vallat JM, et al. Neurologic manifestations of giant cell arteritis. J Neurol. 2022;269(7):3430–3442. doi: 10.1007/s00415-022-10991-6. [DOI] [PubMed] [Google Scholar]
- 2.McDonnell PJ, Moore GW, Miller NR, Hutchins GM, Green WR.. Temporal arteritis. A clinicopathologic study. Ophthalmology. 1986;93(4):518–530. doi: 10.1016/S0161-6420(86)33706-0. [DOI] [PubMed] [Google Scholar]
- 3.ter Borg EJ, Haanen HC, Seldenrijk CA.. Relationship between histological subtypes and clinical characteristics at presentation and outcome in biopsy-proven temporal arteritis. identification of a relatively benign subgroup. Clin Rheumatol. 2007;26(4):529–532. doi: 10.1007/s10067-006-0332-0. [DOI] [PubMed] [Google Scholar]
- 4.Lee YC, Padera RF, Noss EH, et al. Clinical course and management of a consecutive series of patients with “healed temporal arteritis”. J Rheumatol. 2012;39(2):295–302. doi: 10.3899/jrheum.110317. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Salter M, Chacko J, Stanley J, Henry W, Brown H. The clinical characteristics, treatment, and outcomes of giant cell arteritis are dependent on histological subtype. Int J Ophthalmic Patho. 2018;7:1. doi: 10.4172/2324-8599.100214. [DOI] [Google Scholar]
- 6.Gruener AM, Poostchi A, Carey AR, et al. Association of giant cell arteritis with race. JAMA Ophthalmol. 2019;137(10):1175–1179. doi: 10.1001/jamaophthalmol.2019.2919. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hayreh SS, Podhajsky PA, Raman R, Zimmerman B. Giant cell arteritis: validity and reliability of various diagnostic criteria. Am J Ophthalmol. 1997;123(3):285–296. doi: 10.1016/S0002-9394(14)70123-0. [DOI] [PubMed] [Google Scholar]
- 8.Ing EB, Lahaie Luna G, Toren A, et al. Multivariable prediction model for suspected giant cell arteritis: development and validation. Clin Ophthalmol. 2017;11:2031–2042. doi: 10.2147/OPTH.S151385. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.R Core Team . R: a language and environment for statistical computing. R foundation for statistical computing. Vienna, Austria. 2021. https://Www.R-project.org/. [Google Scholar]
- 10.Albert DM, Ruchman MC, Keltner JL. Skip areas in temporal arteritis. Arch Ophthalmol. 1976;94(12):2072–2077. doi: 10.1001/archopht.1976.03910040732006. [DOI] [PubMed] [Google Scholar]
- 11.Durling B, Toren A, Patel V, Gilberg S, Weis E, Jordan D. Incidence of discordant temporal artery biopsy in the diagnosis of giant cell arteritis. Can J Ophthalmol. 2014;49(2):157–161. doi: 10.1016/j.jcjo.2013.12.008. [DOI] [PubMed] [Google Scholar]
- 12.Stacy RC, Rizzo JF, Cestari DM. Subtleties in the histopathology of giant cell arteritis. Semin Ophthalmol. 2011;26(4–5):342–348. doi: 10.3109/08820538.2011.588656. [DOI] [PubMed] [Google Scholar]
- 13.Allison MC, Gallagher PJ. Temporal artery biopsy and corticosteroid treatment. Ann Rheum Dis. 1984. Jun;43(3):416–417. doi: 10.1136/ard.43.3.416. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Jakobsson K, Jacobsson L, Mohammad AJ, et al. The effect of clinical features and glucocorticoids on biopsy findings in giant cell arteritis. BMC Musculoskelet Disord. Aug 24 2016;17(1):363. doi: 10.1186/s12891-016-1225-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Narváez J, Bernad B, Roig-Vilaseca D, et al. Influence of previous corticosteroid therapy on temporal artery biopsy yield in giant cell arteritis. Semin Arthritis Rheum. Aug 2007;37(1):13–19. doi: 10.1016/j.semarthrit.2006.12.005. [DOI] [PubMed] [Google Scholar]
- 16.González-López JJ, González-Moraleja J, Burdaspal-Moratilla A, et al. Factors associated to temporal artery biopsy result in suspects of giant cell arteritis: a retrospective, multicenter, case-control study. Acta Ophthalmol. Dec 2013;91(8):763–768. doi: 10.1111/j.1755-3768.2012.02505.x. [DOI] [PubMed] [Google Scholar]