Giant cell arteritis (GCA), also known as temporal or cranial arteritis, is the most common primary vasculitis of the elderly in the Western world (1). The prevalence among whites older than 50 years is 10–25/100,000 (1,2). Multisystem involvement is often the case, affecting the ocular and cerebral circulations. Because of the variability and sometimes paucity of disease manifestations, the clinical diagnosis can be challenging, yet critical, in preventing the devastating complications of visual loss (3,4) and neurological deficits (3).
GCA affects medium- to large-sized arteries, with a preferential effect on branches of the internal and external carotid arteries (5). Blindness and stroke are feared complications because of inflammation-induced vascular occlusion (5). Headache, the most common symptom in patients with GCA, is believed to be due to inflammation of branches of the external carotid artery (6), and jaw claudication, which has been shown to be a specific symptom of GCA, is a consequence of ischemia to the masseter muscle supplied by the maxillary artery (7,8).
In 1990, the American College of Rheumatology (ACR) published a 5-point scoring system for diagnosing GCA that was initially developed for research purposes. A score of 3 or more had a sensitivity of 93.5% and specificity of 91.2% in distinguishing GCA from other vasculitides (9,10). With these criteria, it is possible to make the diagnosis of GCA without a temporal artery biopsy (TAB) (11). However, most experts recommend a TAB when GCA is suspected because the results of a TAB significantly increase the agreement between the clinical diagnosis and the ACR criteria (12,13).
The goal of our study was to discern clinical and laboratory findings that may improve the diagnostic yield of a positive TAB. We compiled and reviewed our 10-year experience at Duke University Medical Center and performed a review of the clinical studies published in the English literature to derive an algorithm that would be useful in determining the risk for GCA before obtaining a TAB.
The Institutional Review Board at Duke University Medical Center approved this retrospective study. The Department of Pathology electronic database (Cerner PathNet, North Kansas City, MO) was used to identify all patients who underwent a TAB at Duke University Medical Center from 2000 to 2009. Corresponding electronic and paper medical records were reviewed to gather clinical, demographic, pathologic, and laboratory information. Coexisting ocular and medical conditions were collected. Visual acuity measurements at presentation and at the last eye examination were recorded and the surgical specialty obtaining the TAB, length of the TAB specimen, and number of days on corticosteroid before the time of the TAB.
All TABs processed at Duke University Medical Center were cut into 6 to 10 cross sections, depending on the vessel length, and then the cross sections were prepared as ten slides containing 5 μm histological sections of every cross section. Each slide represented a “level” or “step section” with 50 μm between slides. Thus, 500 μm of each of the multiple artery cross sections was evaluated to minimize the possibility of a skip lesion being missed during histological evaluation. Seven slides were stained with hematoxylin and eosin, 1 slide with elastic stain to evaluate the internal elastic lamina, 1 slide with Masson trichrome stain to highlight scarring, and 1 slide with Congo red to evaluate for amyloid deposition.
Three inflammatory markers were measured: erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), and platelet count. Three methods were used to identify the upper limits of normal for the ESR:
- ESR > 50 as per ACR criteria
- The formula of Miller et al (14):
Female: (age + 10)/2.
- 3. The formula of Hayreh et al (15):
Male: 17.3+ (0.18 × age).
Female: 22.1 + (0.18 × age).
We defined an elevated CRP as >0.5 mg/dL and an elevated platelet count as >400,000/μL.
The association between symptoms or laboratory findings and a positive/healed TAB was determined by using contingency tables and chi-square analyses. Statistical significance was defined as P ≤ 0.05. Equivocal biopsies were excluded from the pathological analysis. Risk factors meeting statistical significance and odds ratios were tabulated.
With the aim of creating a numerical score that would improve the sensitivity of a TAB, 2 score models were created:
- Composite score for clinical suspicion calculated by adding 1 point for each item on the following list: 1) new onset headache, 2) scalp tenderness or abnormal superficial temporal artery (STA) on examination, 3) elevation of any inflammatory markers (ESR, CRP, or platelet count), 4) evidence of new onset clinical ischemia involving the anterior extracranial vasculature, 5) jaw claudication, and 6) signs and symptoms consistent with polymyalgia rheumatica (PMR). One point was subtracted for each sign or symptom that could be explained by an alternative condition (e.g., poorly controlled diabetes mellitus and evidence of ischemia, kidney disease, chronic rheumatologic condition, or central retinal artery occlusion in the setting of carotid artery occlusion).
- A weighted score for clinical suspicion was created by the addition of each statistically significant variable tested multiplied by its own relative risk (RR).
An area under the receiver operating curve (ROC) was created for each of the above scores, and the sensitivity and specificity were calculated.
The TAB pathology report was reviewed, and slides from all cases were re-reviewed in a masked fashion by one of the authors (A.D.P.) (see Supplemental Digital Content, Figure E1, http://links.lww.com/WNO/A145). Positive TABs were labeled as active or healed arteritis. Active GCA was defined as inflammation comprising lymphocytes, epithelioid histiocytes, occasional giant cells, and disruption/loss of the internal elastic lamina (16,17). Healed arteritis was identified and distinguished from age-related change by diffuse and marked intimal thickening, large areas of absent internal elastic lamina, asymmetric (eccentric) atrophy and fibrosis of the tunica media, adventitial or transmural scarring, and sometimes vascularization of the media or residual foci of lymphocytes (17,18). The presence of chronic perivascular inflammation adjacent to the STA was considered to be of no clinical significance when this was the sole abnormality in the biopsy sections (19). In a minority of cases, we could not distinguish healed arteritis from senescent changes, (18) and these were classified as suggestive TABs. Negative TAB exhibited a normal histological appearance and lacked any of the above histopathological findings (18).
Two hundred thirteen consecutive TABs were identified. Nine TABs were excluded from analysis for the following reasons: 4 did not yield the STA, 3 were inadequate specimens, 1 was performed at an outside institution and the original slides could not be retrieved for review. One biopsy was excluded because it was performed as part of a parotidectomy for a parotid tumor (the result of that biopsy was positive, but prebiopsy clinical and laboratory information was not available). A total of 204 TABs were analyzed. Forty-nine cases (24.0%) exhibited active or healed GCA (mean age, 74.8 ± 7.8 years), and 12 cases (5.9%) were suggestive of GCA (mean age 76.7 ± 6.5 years). One hundred and forty-three (70.0%) patients had a negative TAB (mean age 69.7 ± 11.0 years). Patients older than 65 years were more likely to have a positive TAB than those younger than 65 years (RR = 3.05, P = 0.0002). The youngest patient with a positive TAB was 54 years old (Table 1).
Among the 143 patients with a negative TAB, the most common final diagnoses were primary headache (17.8%) and an autoimmune disease other than GCA or PMR (8.5%). The remaining alternative diagnoses are detailed in Table 2. One patient had a TAB consistent with Takayasu arteritis. Patients with diabetes mellitus and kidney disease (creatinine >2.0 mg/dL or estimated glomerular filtration rate <30 mL/min) were less likely to have a positive TAB (RR = 0.331, P = 0.0007 and RR = 0.286, P = 0.0141, respectively).
Thirty-three patients with negative TAB would have met ACR criteria for GCA. Five patients were given the diagnosis of TAB-negative GCA; of these, 2 had bilateral TABs. All 5 of these patients had ESR >60 mm/h, 2 patients had jaw claudication, 1 patient had bilateral choroidal ischemia, 1 patient had double vision, 1 patient had constitutional symptoms, and 3 patients had headaches. In all patients, the symptoms improved with the institution of oral corticosteroids.
Twelve patients had a TAB that was considered suspicious for GCA. Three were not given the clinical diagnosis of GCA because another cause was identified (aortic hematoma, musculoskeletal headache, and age-related macular degeneration). Of the remaining 9 patients who were ultimately diagnosed with GCA, 2 of them had an optic neuropathy, 5 had headaches, 3 had jaw claudication, 1 had malaise, 1 had fever, and 4 had an elevated ESR.
Of the 49 patients with positive TAB, the 3 most common findings on presentation were headache (38/47), vision change (16/46), and elevated ESR (33/46). Of the 9 (18.4%) patients who did not complain of headaches, 2 had scalp tenderness, 6 had vision loss, 1 had PMR without vision loss, two had constitutional symptoms, and all had elevated ESR. Of the patients who had vision loss, 9 (56.2%) presented with vision of less than 20/100. Only 1 patient presented with bilateral visual loss.
The clinical sign that had the highest likelihood for positive TAB was jaw claudication (RR = 3.26; P = 0.0014). Isolated findings of headache, scalp tenderness, fever, weight loss, visual loss, and ischemia were not associated with a positive TAB (Table 1).
The majority (74.5%) of TABs were performed on females; however, there was no statistical difference in TAB outcome between men and women. Out of 52 African American patients who underwent TAB, 3 had a positive TAB and 2 had a suggestive TAB (RR for whites vs African Americans was 4.61; P = 0.0013; OR = 6.59). No Hispanic or Asian patients underwent a TAB during the study period. One patient from the Arabian Peninsula had a negative TAB. Race was not recorded in 4 patients with a negative TAB.
The laboratory marker with the strongest association with positive TAB was an elevated platelet count greater than 400,000/μL (RR = 3.3; P = 0.0072) (Table 2). Elevated CRP had a RR of 1.8 (P = 0.037). ESR (>50 mm/h) or an elevated ESR after adjustment according to either the Hayreh or Miller formulas was not associated with a positive TAB (P > 0.18). When both ESR and CRP were elevated, the P value was 0.067. It should be noted that only 37.8% (77/204) subjects had a CRP before the TAB. When both ESR and platelet counts were elevated, RR was 2.01 (P = 0.017). Having any isolated laboratory value elevated, ESR, CRP, or platelet count was not associated with a positive TAB; however, when 2 or more inflammatory markers were elevated, the RR was 3.4 (P < 0.0001) (23/25 patients with elevated platelet count also had a high ESR, and 63/76 patients with a high CRP also had high ESR). An increase in ESR, CRP, and platelet count was present in 11 of 77 patients (14.7%) and was not associated with positive TAB. Two patients with positive TAB had normal ESR, CRP, and platelet counts.
The majority of TABs were performed by the ophthalmology service (136/204 [68%]) followed by otolaryngology (57/204 [28%]). Neurosurgery, vascular surgery, and general surgery were the other services that performed a TAB. The length of the STA did not vary greatly among the different subspecialties. The average length was 20.8 mm.
Seven sequential and 5 simultaneous TABs were performed. In 2 patients (16.6%), there was a discordant pathologic diagnosis. Upon re-review of the original slides, the first biopsy in one of the 2 cases (performed at an outside institution) was believed to be positive, making the discordance rate 8.3%.
Thirty-two patients with a positive TAB and 10 patients with suspicious TAB were on corticosteroid therapy before the TAB. Ten patients with a positive TAB had been on the steroids for more than 3 weeks, and 8 of them had been on the steroids for more than 2 months. Two patients with positive TAB had been on steroids for more than 1 year (Table 1).
A longer segment biopsy was not associated with a higher rate of positive TAB (mean biopsy length 21.6 ± 4.1 for a negative TAB vs 21.6 ± 6.8, P = 0.49 for a positive TAB, P = 0.49).
Both score models had a very high association with a positive or suspicious TAB (P < 0.001). For the nonweighted score, ROC was 0.8. Patients with a score of 1 (only 1 criterion suggestive of GCA) had 0% rate of a positive TAB (70 patients), those with a score of 2 had a 32.9% chance of a positive TAB (26/79), those with a score of 3 or higher had a 55.6% chance of a positive TAB (30/54) (score 3 = 19/39 [49%], score 4 = 6/9 [67%], score 5 = 5/7 [71.4%]), and with a score of 5 had biopsy negative GCA. None of our patients had more than 5 of these criteria.
For the weighted score, the formula was 3.05 (if age >65 years) + 4.61 (if race = white) + 3.26 (if jaw claudication present) + 1.8 (if CRP higher than 0.5 mg/dL) + 3.3 (if platelets >400,000/L) + 3.4 (if >2 laboratory results abnormal). ROC for the weighted score was 0.77. A cut-point score of 8.5 had a positive predictive value of 0.82 and a negative predictive value of 0.53 for a positive TAB.
In our series, the yield of a positive or healed TAB was 24%, consistent with the previously published data ranging from 13.7% to 38.4% (5,6,8,20–24). Based on a comprehensive analysis of the demographic, clinical, and laboratory data of all the patients in this study, we developed an algorithm to quantitate the risk of GCA and increase the likelihood of positive TAB (Fig. 1). Given the potentially devastating visual and systemic complications associated with GCA and the relatively few surgical complications associated with a TAB, the low yield of a positive TAB seems justified in patients suspected of having GCA. However, the relatively low yield of obtaining a positive TAB indicates a deficiency in the ability of the clinician to reliably ascertain whether a patient has GCA based on clinical and laboratory findings. At this time, alternative diagnostic tools, such as Doppler ultrasonography and magnetic resonance imaging, cannot reliably replace a TAB (25).
Although some have supported use of ACR criteria to diagnose GCA, (10,12), the rate of a negative TAB in patients meeting the ACR criteria has ranged from 15% to 39% (12,16,26). In our study, 21% of TAB-negative patients met at least 3 of the ACR criteria. In one series, while all GCA patients with a positive TAB met at least 3 ACR criteria, there were 68.5% of the negative TAB patients who also met at least 3 of the ACR criteria (27). Conversely, Murchison et al (12) reported that 25.7% of their patients with a positive TAB would not have met the ACR criteria. All of our patients who had an ultimate diagnosis of GCA met at least 3 of the ACR criteria.
A TAB is considered the gold standard for confirming the diagnosing of GCA. A TAB is 100% specific for GCA but is not 100% sensitive (20,28). Niederkohr and Levin (28) estimated the sensitivity of a single TAB to be 87.1%. We found the sensitivity of a positive TAB was 91.4%, similar to the rate reported in the literature (29).
In our study, the most common mimickers of GCA were primary headache syndrome, PMR, and chronic disease (diabetes mellitus and renal failure). Rheumatologic diseases (rheumatoid arthritis, systemic lupus erythematosus), lymphoma, and vasculopathy also have been reported to mimic GCA (21). Similar to the results reported by Matthews et al (30), we found that patients with diabetes and kidney disease were less likely to have a positive TAB. This does not imply that patients with diabetes or kidney disease are protected from GCA, but clinicians should keep in mind that diabetes and kidney disease have vascular complications that may mimic GCA.
Among our patients with a positive TAB, the most common presenting symptoms were headaches and vision loss. A new onset severe headache has been reported in approximately 2/3 of patients with GCA (15). Vision loss, one of the most feared complications of GCA, has been reported in 14%–70% of cases (3,31). Among our patients, 42.3% complained of vision changes and of those, 50% had permanent vision loss in 1 or both eyes due to retinal or optic nerve ischemia. In our case series, the rate of permanent vision loss was 21.2% (Table 1).
Among our patients, the most common abnormal laboratory measure was an elevated platelet count, followed by a high CRP. Elevation of these 2 inflammatory markers was highly predictive of a positive TAB, similar to previously reported laboratory findings in GCA (22,23). Although an increase in any 2 laboratory markers was associated with a higher risk for a positive TAB, an increase in all 3 was not associated with a positive TAB. This is contrary to previously published studies (22,23), and we suspect the reason for this finding is that only about one-third of our patients had all three markers measured before TAB. Also, in concordance with previous studies (15,24), jaw claudication was associated with the greatest likelihood of GCA while none of the other symptoms (e.g., headache, vision loss, abnormal STA, PMR) were predictive of GCA as isolated markers.
In most cases, a unilateral TAB is often sufficient in suspected cases of GCA (32,33), but in 3%–5% of cases, there can be significant differences in the pathologic grade of disease from one side to the other in cases of bilateral TAB (30,31). The majority of our patients had a unilateral TAB, but of those patients who had a bilateral biopsy, the discordance rate was 8.3%. In an analysis of 11 pooled studies, Niederkohr and Levin (34) found the mean discordance rate for bilateral TAB to be 5.5%. Our discordance rate is higher than previously published studies, but our sample size of bilateral TAB was small. Based on our study and the low discordant rate published in the literature, we believe that a unilateral TAB is sufficient to exclude the diagnosis of GCA in patients for which there is a low clinical suspicion (33). Even with a high concordance in pathologic diagnosis, a statistically significant difference can also be observed in the objective histopathologic severity scale between the 2 sides (35). The single patient in our study with a discordant result on bilateral TABs was likely due to the inadequate sampling on the initial TAB (9-mm segment), although we cannot exclude the possibility of misinterpreting the first specimen (the histopathologic slides were not available for our review).
Although corticosteroid use decreases the inflammatory change in the vessel wall, it does not have an effect on scarring of the vascular media and surrounding the inner elastic lamina, thereby permitting a pathologic diagnosis (36,37). Even in patients who developed GCA with a history of PMR and chronic low-dose corticosteroid use, TAB remained positive in 88% of cases in one study (26). Two of our patients who had been on chronic steroid therapy for more than 1 year had a positive TAB.
We combined the results of our study with a review of the current literature to modify the algorithm published by Shmerling (38) for evaluating and treating suspected GCA (Fig. 1). The first step in the algorithm is to assess how many clinical and laboratory findings are consistent with GCA, which will determine the degree of clinical suspicion. If a patient meets less than 2 clinical or laboratory criteria, the likelihood of GCA is very low. If a patient meets 2 criteria, the likelihood of GCA is moderate (33%), and if a patient achieves 3 criteria, the likelihood is high (56%).
We recognize the limitations of our studies. Being a retrospective study, medical records reviewed may have had incomplete documentation of symptoms, signs, and laboratory results. A number of patients were not tested for CRP and platelet count, and the clinical examination of the STA was not always documented. Intravenous fluorescein angiography was performed on only 7 of the patients in our study. We were not able to assess the prevalence of choroidal ischemia and the sensitivity of this test in predicting a positive TAB. Also, some patients did not undergo a follow-up ophthalmologic examination at our medical center and, as a result, we were unable to assess the risk of vision loss after low-dose or high-dose corticosteroid treatment. Given the relatively low prevalence of GCA, a multicentered prospective study with various ethnic representations would be necessary for controlling these limitations.
1. Chew SS, Kerr NM, Danesh-Meyer HV. Giant cell arteritis. J Clin Neurosci. 2009;16:1263–1268.
2. Gonzalez-Gay MA, Vazquez-Rodriguez TR, Lopez-Diaz MJ, Miranda-Filloy JA, Gonzalez-Juanatey C, Martin J, Liorca J. Epidemiology of giant cell arteritis and polymyalgia rheumatica. Arthritis Rheum. 2009;61:1454–1461.
3. Caselli RJ, Hunder GG, Whisnant JP. Neurologic disease in biopsy-proven giant cell (temporal) arteritis. Neurology. 1988;38:352–359.
4. Gonzalez-Gay MA, Garcia-Porrua C, Llorca J, Hajeer AH, Branas F, Dababneh A, et al.. Visual manifestations of giant cell arteritis. Trends and clinical spectrum in 161 patients. Medicine (Baltimore). 2000;79:283–292.
5. Brack A, Martinez-Taboada V, Stanson A, Goronzy JJ, Weyand CM. Disease pattern in cranial and large-vessel giant cell arteritis. Arthritis Rheum. 1999;42:311–317.
6. ICHD-II. Available at: http://ihs-classification.org/en/
. Accessed December 14, 2014.
7. Brass SD, Durand ML, Stone JH, Chen JW, Stone JR. Case records of the Massachusetts General Hospital. Case 36-2008. A 59-year-old man with chronic daily headache. N Engl J Med. 2008;359:2267–2278.
8. Stone JH, Papaliodis GN, Dunbar MR, Stone JR. Case records of the Massachusetts General Hospital. Case 4–2010. A 53-year-old man with arthralgias, oral ulcers, vision loss, and vocal-cord paralysis. N Engl J Med. 2010;362:537–546.
9. Rao JK, Allen NB, Pincus T. Limitations of the 1990 American College of Rheumatology classification criteria in the diagnosis of vasculitis. Ann Intern Med. 1998;129:345–352.
10. Davies C, Frost B, Eshan O, McLain AD, Shandall A. Temporal artery biopsy… who needs one? Postgrad Med J. 2006;82:476–478.
11. Drehmer TJ, Khanna D, Markert RJ, Hawkins RA. Diagnostic and management trends of giant cell arteritis: a physician survey. J Rheumatol. 2005;32:1283–1289.
12. Murchison AP, Gilbert ME, Bilyk JR, Eagle RC Jr, Pueyo V, Sergott RC, Savino PJ. Validity of the American College of Rheumatology criteria for the diagnosis of giant cell arteritis. Am J Ophthalmol. 2012;154:722–729.
13. Danesh-Meyer HV. Temporal artery biopsy: skip it at your patient's peril. Am J Ophthalmol. 2012;154:617–619.
14. Miller A, Green M, Robinson D. Simple rule for calculating normal erythrocyte sedimentation rate. Br Med J (Clin Res Ed). 1983;286:266.
15. Hayreh SS, Podhajsky PA, Raman R, Zimmerman B. Giant cell arteritis: validity and reliability of various diagnostic criteria. Am J Ophthalmol. 1997;123:285–296.
16. Taylor-Gjevre R, Vo M, Shukla D, Resch L. Temporal artery biopsy for giant cell arteritis. J Rheumatol. 2005;32:1279–1282.
17. McDonnell PJ, Moore GW, Miller NR, Hutchins GM, Green WR. Temporal arteritis. A clinicopathologic study. Ophthalmology. 1986;93:518–530.
18. Lie JT, Brown AL Jr, Carter ET. Spectrum of aging changes in temporal arteries. Its significance, in interpretation of biopsy of temporal artery. Arch Pathol. 1970;90:278–285.
19. Corcoran GM, Prayson RA, Herzog KM. The significance of perivascular inflammation in the absence of arteritis in temporal artery biopsy specimens. Am J Clin Pathol. 2001;115:342–347.
20. Bhatti MT, Tabandeh H. Giant cell arteritis: diagnosis and management. Curr Opin Ophthalmol. 2001;12:393–399.
21. Roth AM, Mislow L, Keltner JL. The ultimate diagnoses of patients undergoing temporal artery biopsies. Arch Opthalmol. 1984;102:901–903.
22. Parikh M, Miller NR, Lee AG, Savino PJ, Vacarezza MN, Cornblath W, Eggenberger E, Antonio-Santos A, Golnik K, Kardon R, Wall M. Prevalence of a normal C-reactive protein with an elevated erythrocyte sedimentation rate in biopsy-proven giant cell arteritis. Ophthalmology. 2006:113:1842–1845.
23. Walvick MD, Walvick MP. Giant cell arteritis: laboratory predictors of a positive temporal artery biopsy. Ophthalmology. 2011;118:1201–1204.
24. Salvarani C, Pipitone N, Versari A, Hunder GG. Clinical features of polymyalgia rheumatica and giant cell arteritis. Nat Rev Rheumatol. 2012;8:509–521.
25. Kawasaki A, Purvin V. Giant cell arteritis: an updated review. Acta Ophthalmol. 2009;87:13–32.
26. Narvaez J, Bernad B, Roig-Vilaseca D, Garcia-Gomez C, Gomez-Vaquero C, Juanola X, Rodriquez-Moreno J, Nolla JM, Valverde J. Influence of previous corticosteroid therapy on temporal artery biopsy yield in giant cell arteritis. Semin Arthritis Rheum. 2007;37:13–19.
27. Mari B, Monteagudo M, Bustamante E, Perez J, Casanovas A, Jordana R, Tolosa C, Oristrell J. Analysis of temporal artery biopsies in an 18-year period at a community hospital. Eur J Intern Med. 2009:20:533–536.
28. Niederkohr RD, Levin LA. A Bayesian analysis of the true sensitivity of a temporal artery biopsy. Invest Ophthalmol Vis Sci. 2007;48:675–680.
29. Hedges TR, Gieger GL, Albert DM. The clinical value of negative temporal artery biopsy specimens. Arch Ophthalmol. 1983;101:1251–1254.
30. Matthews JL, Gilbert DN, Farris BK, Siatkowski RM. Prevalence of diabetes mellitus in biopsy-positive giant cell arteritis. J Neuroophthalmol. 2012;32:202–206.
31. Huston KA, Hunder GG, Lie JT, Kennedy RH, Elveback LR. Temporal arteritis: a 25-year epidemiologic, clinical, and pathologic study. Ann Intern Med. 1978;88:162–167.
32. Boyev LR, Miller NR, Green WR. Efficacy of unilateral versus bilateral temporal artery biopsies for the diagnosis of giant cell arteritis. Am J Ophthalmol. 1999;128:211–215.
33. Hall JK, Volpe NJ, Galetta SL, Liu GT, Syed NA, Balcer LJ. The role of unilateral temporal artery biopsy. Ophthalmology. 2003;110:543–548.
34. Niederkohr RD, Levin LA. Management of the patient with suspected temporal arteritis a decision-analytic approach. Ophthalmology. 2005;112:744–756.
35. Diaz VA, DeBroff BM, Sinard J. Comparison of histopathologic features, clinical symptoms, and erythrocyte sedimentation rates in biopsy-positive temporal arteritis. Ophthalmology. 2005;112:1293–1298.
36. Font RL, Prabhakaran VC. Histological parameters helpful in recognizing steroid-treated temporal arteritis: an analysis of 35 cases. Br J Ophthalmol. 2007;91:204–209.
37. Zhou L, Luneau K, Weyand CM, Biousse V, Newman NJ, Grossniklaus HE. Clinicopathologic correlations in giant cell arteritis: a retrospective study of 107 cases. Ophthalmology. 2009;116:1574–1580.
38. Shmerling RH. An 81-year-old woman with temporal arteritis. JAMA. 2006;295:2525–2534.