VASCULITIS

Utility of Erythrocyte Sedimentation Rate and C-Reactive

Protein for the Diagnosis of Giant Cell Arteritis
Tanaz A. Kermani, MD,* Jean Schmidt, MD,†
Cynthia S. Crowson, MS,*,‡ Steven R. Ytterberg, MD,*
Gene G. Hunder, MD,* Eric L. Matteson, MD, MPH,* and
Kenneth J. Warrington, MD*

Objectives: To evaluate the utility of erythrocyte sedimentation rate (ESR) and C-reactive protein
(CRP) for the diagnosis of giant cell arteritis (GCA) and to determine the frequency of normal ESR
and CRP at diagnosis of GCA.
Methods: All patients undergoing temporal artery biopsy (TAB) between 2000 and 2008 were
identified. Only subjects with both ESR and CRP at the time of TAB were included. The medical
records of all patients were reviewed.
Results: We included 764 patients (65% women), mean age 72.7 (⫾9.27) years, who underwent
TAB. Biopsy was consistent with GCA in 177 patients (23%). Elevated CRP and elevated ESR
provided a sensitivity of 86.9% and 84.1%, respectively, for a positive TAB. The odds ratio of a
concordantly elevated ESR and CRP for positive TAB was 3.06 (95% CI 2.03, 4.62), whereas the
odds ratio for concordantly normal ESR and CRP was 0.49 (95% CI 0.29, 0.83). Seven patients
(4%) with a positive TAB for GCA had a normal ESR and CRP at diagnosis. Compared with GCA
patients with elevated markers of inflammation, a greater proportion of these patients had poly-
myalgia rheumatica symptoms (P ⫽ 0.008), whereas constitutional symptoms, anemia and
thrombocytosis, were observed less often (P ⬍ 0.05).
Conclusions: CRP is a more sensitive marker than ESR for a positive TAB that is diagnostic of
GCA. There may be clinical utility in obtaining both tests in the evaluation of patients with
suspected GCA. A small proportion of patients with GCA may have normal inflammatory markers
at diagnosis.
© 2012 Elsevier Inc. All rights reserved. Semin Arthritis Rheum 41:866-871
Keywords: giant cell arteritis, erythrocyte sedimentation rate, C-reactive protein, temporal artery biopsy,
polymyalgia rheumatica

G iant cell arteritis (GCA) is a granulomatous at 1% for women and 0.5% for men (3). Vision loss
vasculitis involving the aorta and its primary due to ischemic optic neuropathy is a dreaded compli-
and secondary branches (1). GCA is the most cation of GCA and therefore prompt diagnosis and
common form of systemic vasculitis in people over the treatment are essential (1).
age of 50 years, with an annual incidence of approxi- To date, there is no specific blood biomarker for GCA,
mately 19 per 100,000 people aged 50 years and above and histologic examination of a temporal artery biopsy
(2). The lifetime risk for developing GCA is estimated (TAB) specimen is considered the gold standard diagnos-
tic test. Laboratory findings in GCA are nonspecific but
most patients have elevated acute phase reactants at dis-
*Division of Rheumatology, Department of Medicine, Mayo Clinic, Rochester, MN.
ease onset. An elevated erythrocyte sedimentation rate
†Department of Internal Medicine, Amiens University Hospital, Amiens, France. (ESR) has been considered a hallmark of this disease and
‡Division of Biostatistics, Department of Health Sciences Research, Mayo Clinic, is 1 of the American College of Rheumatology (ACR)
Rochester, MN.
Address reprint requests to Tanaz A. Kermani, MD, 200 First Street SW, Roch- classification criteria for GCA (4). The C-reactive protein
ester, MN 55905. E-mail: [email protected]. (CRP) is an acute phase reactant synthesized in the liver
866 0049-0172/12/$-see front matter © 2012 Elsevier Inc. All rights reserved.
doi:10.1016/j.semarthrit.2011.10.005
T.A. Kermani et al. 867

and is a sensitive marker of systemic inflammation. Al- ceiver operator characteristic (ROC) curves were used to
though CRP is often used in conjunction with the ESR determine the optimal cutoff value for CRP.
for the initial evaluation of patients with suspected GCA, In the second portion of the study, we compared the
few studies have evaluated CRP performance in subjects clinical features of patients with GCA confirmed by TAB
undergoing TAB (5,6). who had normal ESR and CRP to those of patients with
We sought to evaluate the performance of ESR, CRP, elevated markers at the time of diagnosis. We used the
and a combination of the 2 tests for the diagnosis of GCA Fisher’s exact test to compare categorical variables be-
among all patients undergoing TAB. Second, we sought tween patients with normal ESR and CRP at the time of
to evaluate the frequency of normal ESR and CRP at the diagnosis to patients with either test elevated. Rank sum
time of GCA diagnosis in patients with a positive TAB tests were used to compare continuous variables between
and describe the clinical characteristics of these patients. the 2 groups.

PATIENTS AND METHODS RESULTS

The study was approved by the Mayo Clinic Institutional A total of 1106 subjects underwent 1174 TAB at our
Review Board. Using the Mayo Clinic database, a data institution between January 1, 2000 and December 31,
retrieval specialist identified all patients who underwent 2008. TAB findings were consistent with GCA in 242
TAB at our institution between January 1, 2000 and De- patients (21.9%).
cember 31, 2008.
Performance of ESR and CRP
Among Patients Undergoing TAB
Data Collection
Of the 1106 patients who underwent TAB, 764 (69.1%)
The medical records of all patients who underwent TAB
had both ESR and CRP testing within 6 weeks before
were reviewed. Data regarding demographic information,
TAB and were included for this portion of the study. The
date of biopsy, and laboratory findings, including com-
study cohort therefore included 587 patients (77%) with
plete blood count, ESR, and CRP before the biopsy, were
a negative TAB and 177 patients (23%) with a positive
abstracted from the medical records. Pathology reports
TAB for GCA. Of the 177 patients with positive TAB,
from all subjects undergoing TAB were reviewed to iden-
165 patients (93.2%) met 1990 ACR classification crite-
tify patients in whom TAB was interpreted as showing
ria for GCA. Twelve patients with positive TAB who did
changes consistent with GCA. Patients with a positive
not meet 1990 ACR classification criteria included 8 pa-
TAB were all treated for GCA. In the subset of patients
tients with other cranial symptoms suspicious for GCA
with biopsy-positive GCA, we also abstracted the date of
and 4 patients with symptoms of limb claudication and
first symptom onset, date of diagnosis of GCA, symptoms
evidence of large-artery stenosis. We compared baseline
at disease onset, glucocorticoid use at the time of labora-
demographics and the distribution of ESR and CRP in
tory testing (if applicable), and use of nonsteroidal anti-
patients with a positive TAB for GCA to those with a
inflammatory medications at the time of laboratory test-
negative TAB (Table 1). At baseline, patients with a pos-
ing.
itive TAB for GCA were, on average, 2.4 years older than
ESR was measured by Westergren method. Based on
patients who had a negative TAB. The median ESR and
Mayo Clinic clinical laboratory reporting, an ESR value
CRP were also significantly higher in patients with a pos-
of ⱕ22 mm/h in men or ⱕ29 mm/h in women was
itive TAB compared with those with a negative biopsy.
considered normal regardless of age. A normal CRP was
Additionally, a significantly greater proportion of patients
defined as a value of ⱕ8 mg/L in all patients regardless of
with a positive TAB had a concordantly elevated ESR and
age.
CRP (81%) compared with subjects with a negative bi-
opsy (59%).
Exclusions The sensitivity, specificity, positive and negative-pre-
Subjects who did not have both ESR and CRP testing dictive values of ESR, CRP, and a combination of the 2
available in the 6 weeks before TAB were excluded from for a positive biopsy result are provided in Table 2. An
the study. elevated CRP provided a slightly higher sensitivity for a
positive biopsy than an elevated ESR. The specificity of
the ESR and CRP individually was low, whereas the com-
Statistical Analysis bination of elevated ESR and CRP provided modest spec-
Descriptive statistics were used to summarize data. For ificity (41%). Both a normal ESR and a normal CRP had
the first portion of the study, sensitivity, specificity, pos- excellent negative-predictive values, with the CRP slightly
itive and negative-predictive value of ESR, CRP and the outperforming the ESR. An elevated ESR was associated
combination of the 2 for a positive TAB were calculated. with increased odds ratio (OR) for a positive TAB (OR
Logistic regression was used to evaluate the association 2.22; 95%CI 1.43, 3.46) as was elevated CRP (OR 2.94;
between the acute phase reactants and positive TAB. Re- 95%CI 1.83, 4.71). The highest OR for positive TAB was
868 ESR and CRP for the diagnosis of giant cell arteritis

Table 1 Baseline and Laboratory Characteristics Among 764 Subjects Undergoing a Temporal Artery Biopsy (TAB)
Variable TAB Negative (N ⫽ 587) TAB Positive (N ⫽ 177) P Value
Mean age (⫾SD), yr 72.1 (9.6) 74.5 (7.8) 0.008
Female sex, No. (%) 371 (63.2) 130 (73.4) 0.012
Median ESR (IQR), mm/h 44.0 (25, 75) 62.0 (36, 91) ⬍0.001
Median CRP (IQR), mg/L 24.0 (6, 73) 52.0 (27, 95) ⬍0.001
ESR elevated, No. (%) 414 (70.5) 149 (84.2) ⬍0.001
CRP elevated, No. (%) 408 (69.5) 153 (86.4) ⬍0.001
Elevated ESR and CRP, No. (%) 345 (58.8) 143 (80.8) ⬍0.001
Normal ESR and CRP, No. (%) 110 (18.7) 18 (10.2) 0.006
Data presented as median with 25th and 75th percentile values.
SD, standard deviation; ESR, erythrocyte sedimentation rate; CRP, C-reactive protein; IQR, interquartile range.

observed in subjects who had an elevation of both ESR ESR and CRP at diagnosis. This included 130 women
and CRP (OR 3.06; 95% CI 2.03, 4.62). Conversely, if (73.4%) and 47 men (26.6%) with a mean age at diagno-
ESR and CRP were both normal, the OR for positive sis of 74.5 (⫾7.8) years. Table 3 summarizes the symp-
biopsy was reduced (OR 0.49; 95% CI 0.29, 0.83). toms at diagnosis in all 177 patients.
We performed a sensitivity analysis in which we in- The ESR and/or CRP were elevated in 159 (89.8%)
cluded all subjects with TAB who had either ESR or CRP patients, whereas 18 patients (10.2%) had a normal ESR
tested. The results of this analysis for ESR or CRP were and CRP at the time of diagnosis. There was good con-
similar to that reported in Table 2. When all subjects who cordance between ESR and CRP, as both were either
had an ESR performed were included, elevated ESR had a elevated or normal in 163 patients (92.1%). However, 14
sensitivity of 86% (193/225), specificity of 27% (217/ patients (7.9%) had discordant results; of these, CRP was
810), positive-predictive value of 25% (193/786), and elevated with a normal ESR in 9 patients.
negative-predictive value of 87% (217/249). For CRP, Eleven of the 18 subjects with normal ESR and CRP
the sensitivity of elevated CRP was 87% (155/178), spec- were on glucocorticoids at the time of testing. There-
ificity was 31% (183/599), positive-predictive value was fore, 7 (4%) patients had a normal ESR and CRP even
27% (155/571), and negative-predictive value was 89% in the absence of glucocorticoid use. Median ESR for
(183/206). Based on estimates derived by modeling an these 7 patients was 13 mm/h (range, 3-25) with me-
ROC curve and including all subjects who underwent dian CRP 4.4 mg/L (range, 1.3-8.0 mg/L). Although
TAB and had CRP testing performed, the optimal cutoff not statistically significant, the 7 patients with normal
value for CRP was 26.9 mg/L, yielding a sensitivity of ESR and CRP tended to be younger, have a longer
75% and specificity of 51% (Fig. 1A). Similarly, estimates duration of symptoms, and had fewer constitutional
derived by modeling an ROC curve for all subjects with symptoms than the patients with elevated ESR and/or
TAB and ESR testing, the optimal cutoff value for ESR CRP (Table 4). Of concern, visual symptoms were
was 56 mm/h, which gave a sensitivity of 66% and spec- present in 3 of the 7 patients (43%) with vision loss
ificity of 55% (Fig. 1B). occurring in 1 patient (14.3%). A greater proportion of
patients with normal ESR and CRP at the time of
Normal ESR and CRP at Diagnosis in GCA diagnosis had symptoms of polymyalgia rheumatica
(Table 4). These patients also had a significantly lower
The subset of 177 patients with GCA was used to evaluate
platelet count and higher hemoglobin at the time of
the clinical characteristics of patients who had normal
diagnosis compared with patients with an elevated ESR
or CRP (Table 4).
Table 2 Test Characteristics of ESR, CRP, and a
Combination of the 2 for Positive Temporal Artery
Biopsy DISCUSSION
Sensitivity Specificity PPV NPV Biomarkers of inflammation, particularly the ESR and
ESR 84.2% 29.5% 26.4% 86.1%
CRP, are often used in the evaluation of patients sus-
149/177 173/592 149/563 173/201 pected of having GCA. However, performance of both in
CRP 86.4% 30.5% 27.2% 88.6% the diagnosis of GCA is not well described (5,6). To our
153/177 179/587 153/561 179/202 knowledge, the current study is the largest to date that
ESR and CRPa 80.8% 41.2% 29.3% 87.7% evaluates these markers of inflammation in a cohort of
143/177 242/587 143/488 242/276 patients undergoing TAB for the diagnosis of GCA. Ad-
ESR, erythrocyte sedimentation rate; CRP, C-reactive protein; ditionally, we also assessed the frequency and clinical
PPV, positive predictive value; NPV, negative predictive value. characteristics of patients with biopsy-positive GCA who
aEvaluating subjects with both tests elevated
had normal ESR and CRP.
T.A. Kermani et al. 869

among subjects undergoing a TAB and also concluded

that elevated CRP was a better predictor of a positive
biopsy than ESR (5). However, although the latter study
included 3001 subjects undergoing TAB (459 with posi-
tive biopsy), the frequency of CRP testing was low and
only available in 20% of the patients (98 patients with
GCA and 493 subjects with a negative biopsy). In our
study, 66% of patients who underwent TAB had both
ESR and CRP measured at the time of the procedure.
We were able to evaluate the test characteristics of ESR
or CRP alone and the combination of the 2 to determine
which performed better and whether there is clinical util-
ity to testing both. As in previous reports, we found that
elevated CRP had a higher sensitivity than ESR for GCA
(86.4%). However, the specificity of either ESR or CRP
alone was low (approximately 30%). This improved mod-
estly to 41% when a combination of ESR and CRP were
used. Although the combination of ESR and CRP did not
improve sensitivity, the highest OR for a positive TAB
and the diagnosis of GCA was observed when both ESR
and CRP were concordantly elevated. Conversely, the
combination of normal ESR and CRP was associated with
decreased odds of having a positive TAB.
Although the majority of patients with GCA have ele-
vated markers of inflammation at clinical presentation,
normal ESR or CRP has been reported in a small propor-
tion of patients at diagnosis. The frequency varies be-
tween 4 and 14% depending on the study and the defini-
tion of normal values (7-9). In 1 study, 22.5% subjects
with polymyalgia rheumatica or GCA had a normal ESR
at diagnosis (10). In a population-based study of newly
diagnosed GCA, 5.3% of 167 patients had an ESR less
than 40 mm/h (7). A meta-analysis evaluating the diag-
nostic value of clinical findings among patients suspected
of having GCA estimated that about 4% patients with
GCA present with normal ESR (9). Few studies have
evaluated normal CRP at diagnosis, with estimates rang-
ing from 2% to 14% (8,11-13). In 1 study of 119 patients
with GCA, only 3 subjects (1.7%) had a normal CRP at
the time of GCA diagnosis, of whom 1 subject (0.8%) had
Figure 1 Receiver operator characteristic curve of CRP (A)
a normal ESR and CRP (8). Other studies which have
and ESR (B) in subjects undergoing temporal artery biopsy.

Only 2 prior studies have evaluated the usefulness of Table 3 Symptoms at Onset in 177 Patients with GCA
CRP in the diagnosis of GCA among subjects undergoing Variable No. (%)
TAB (5,6). In a study evaluating the sensitivity and spec- Median duration symptoms (IQR), d 51.5 (18, 111.5)
ificity of different laboratory markers for the diagnosis of Headache 111 (62.7)
GCA, CRP (⬎0.5 mg/dL) was more sensitive (sensitivity Jaw claudication 78 (44.1)
100%) than ESR (sensitivity 97%) and the combination Scalp sensitivity 65 (36.7)
of the 2 provided the best specificity (97%) (6). However, Transient vision loss 37 (20.9)
the authors compared the test characteristics among pa- Permanent vision loss 18 (10.2)
tients with GCA to controls who were subjects seen in an Subjective fever 39 (22.0)
ophthalmology practice for other conditions (6). There- Fatigue 62 (35.0)
fore, this study did not provide information on the per- Anorexia 21 (11.9)
Subjective weight loss 54 (30.5)
formance of inflammatory markers in the setting of all-
Polymyalgia rheumatica 54 (30.5)
comers undergoing a TAB for suspected GCA. A more
recent study evaluated the performance of ESR and CRP GCA, giant cell arteritis.
870 ESR and CRP for the diagnosis of giant cell arteritis

Table 4 Clinical Manifestations of GCA Patients with Normal ESR and CRP to Those with Elevated ESR and/or CRP
Clinical Variable ESR and/or CRP Elevateda N ⫽ 138 ESR and CRP Normala N ⫽ 7 P Value
Female gender, No. (%) 98 (71) 6 (85.7) 0.40
Age at diagnosis, mean (⫾SD), yr 74.5 (8.0) 70.0 (5.9) 0.11
Symptom duration, median (IQR),b d 51 (18, 101) 112 (16, 167) 0.41
New headache, No. (%) 88 (63.8) 4 (57.1) 0.72
Jaw claudication, No. (%) 65 (47.1) 2 (28.6) 0.34
Scalp tenderness, No. (%) 50 (36.2) 2 (28.6) 0.68
Transient vision loss, No. (%) 30 (21.7) 2 (28.6) 0.68
Permanent vision loss, No. (%) 13 (9.4) 1 (14.3) 0.67
Fever, No. (%) 33 (23.9) 0 (0) 0.14
Fatigue, No. (%) 50 (36.2) 0 (0) 0.05
Anorexia, No. (%) 17 (12.3) 0 (0) 0.32
Weight loss, No. (%) 43 (31.2) 1 (14.3) 0.34
Polymyalgia rheumatica, No. (%) 35 (25.4) 5 (71.4) 0.008
Anemia, No. (%) 88 (68.8) 0 (0) ⬍0.001
Hemoglobin, mean (⫾SD), g/dL 11.6 (1.53) 13.4 (0.99) 0.002
Thrombocytosis, No. (%) 46 (36.2) 0 (0) 0.05
Platelet count, mean (⫾SD), ⫻109/L 424.2 (135.2) 270.9 (56.79) 0.001
SD, standard deviation; IQR, interquartile ratio; NSAID, non-steroidal anti-inflammatory drugs.
aSubjects on glucocorticoids were excluded.
bData presented as median with 25th and 75th percentile values.

evaluated CRP at diagnosis have included subjects with phase reactants. Previous studies have shown that a strong
polymyalgia rheumatica (PMR) and/or GCA, reporting a inflammatory response may be protective against cranial
prevalence of normal CRP to be between 1.2% and 20% ischemic complications including vision loss, possibly be-
(11-13). However, because these studies also included cause of increased alertness to the potential for visual loss in
patients with PMR, it is difficult to draw conclusions these patients on the part of the managing physician (15). It
regarding the subset of patients with GCA. is important to recognize that in the presence of clinical
In our study, the prevalence of normal ESR and CRP in symptoms or findings of ischemic eye disease, GCA should
the absence of glucocorticoid use was 4%. We used cutoff be considered even in the absence of elevated markers of
values defined by our laboratory for the normal ranges of inflammation.
these tests in our analysis, rather than other possible, gen- Although our study includes a large cohort of subjects,
erally higher, values for ESR and CRP. All of the patients several limitations need to be considered. This study was
with normal inflammatory markers underwent TAB be- conducted at a tertiary care referral center. We only in-
cause of clinical symptoms suggesting GCA. cluded cases with biopsy-positive GCA and therefore
Compared with patients with elevated acute phase re- these results may not be applicable to all patients with
actants and biopsy-confirmed GCA, the patients with GCA. There may also have been patients with clinically
normal markers had fewer constitutional symptoms. Ad- diagnosed GCA who had a negative TAB and were there-
ditionally, other laboratory abnormalities including ane- fore misclassified. Given our research question of interest,
mia and thrombocytosis were also less prevalent in those we only included patients with both ESR and CRP testing
with normal ESR/CRP, suggesting an overall muted in- available at the time of diagnosis. Also, the estimates for
flammatory response. This finding is similar to what has sensitivity and specificity of ESR and CRP are only appli-
been reported in a population-based study evaluating low cable to patients in whom a decision has been made to
ESR (⬍40 mm/h) at diagnosis (7). In that study, patients perform a TAB. Given the referral nature of our practice,
with a low ESR had fewer systemic symptoms (11%) than the prevalence of normal ESR and CRP in GCA patients
patients with GCA and elevated ESR (57%) (P ⫽ 0.02) may be higher than would be expected in the general
(7). Results from another study revealed a trend toward population. Additionally, there may be greater vigilance
less anemia and less fever in patients with GCA who had an for GCA at our center, leading to TAB even in atypical
ESR of ⬍50 mm/h compared with those with elevated ESR cases where markers of inflammation are normal.
(14). Although not statistically significant, we found that the In this study, CRP was a more sensitive marker than
duration of symptoms antecedent to TAB and diagnosis of ESR for GCA among subjects undergoing TAB. The
GCA was longer in patients with normal ESR and CRP combination of elevated ESR and CRP provided better
compared with patients with elevated acute phase reactants. specificity than either test alone and was associated with
This likely reflects a delay in diagnosis in these patients with greater odds of a positive TAB. Among GCA patients
atypical findings. Of concern is that vision loss (transient or with a positive biopsy, normal ESR and CRP were ob-
permanent) occurred in 3 of 7 patients with normal acute served in 4% of cases. These patients tended to have an
T.A. Kermani et al. 871

overall muted inflammatory response but more fre- tors of a positive temporal artery biopsy. Ophthalmology
quently had PMR symptoms. Our findings highlight that 2011;118(6):1201-4.
6. Hayreh SS, Podhajsky PA, Raman R, Zimmerman B. Giant cell
the absence of a systemic inflammatory response as re-
arteritis: validity and reliability of various diagnostic criteria. Am J
flected in the currently routinely used ESR and CRP does Ophthalmol 1997;123:285-96.
not exclude GCA. TAB should be pursued in all patients 7. Salvarani C, Hunder GG. Giant cell arteritis with low erythrocyte
where clinical suspicion of GCA is high. Although ESR sedimentation rate: frequency of occurrence in a population-
and CRP are often helpful for the diagnosis of GCA, they based study. Arthritis Rheum 2001;45:140-5.
are nonspecific and better diagnostic markers are needed. 8. Parikh M, Miller NR, Lee AG, Savino PJ, Vacarezza MN, Corn-
blath W, et al. Prevalence of a normal C-reactive protein with an
elevated erythrocyte sedimentation rate in biopsy-proven giant
ACKNOWLEDGMENTS cell arteritis. Ophthalmology 2006;113:1842-5.
Dr Kermani was supported by the Vasculitis Clinical 9. Smetana GW, Shmerling RH. Does this patient have temporal
arteritis? JAMA 2002;287:92-101.
Research Consortium (VCRC), which has received 10. Ellis ME, Ralston S. The ESR in the diagnosis and management of
support from the National Institute of Arthritis and the polymyalgia rheumatica/giant cell arteritis syndrome. Ann
Musculoskeletal and Skin Diseases (U54AR057319), Rheum Dis 1983;42:168-70.
the National Center for Research Resources (U54 11. Kyle V, Cawston TE, Hazleman BL. Erythrocyte sedimentation
RR019497), and the Office of Rare Diseases Research. rate and C reactive protein in the assessment of polymyalgia rheu-
The VCRC is part of the Rare Diseases Clinical Re- matica/giant cell arteritis on presentation and during follow up.
Ann Rheum Dis 1989;48:667-71.
search Network (RDCRN).
12. Pountain GD, Calvin J, Hazleman BL. Alpha 1-antichymotryp-
sin, C-reactive protein and erythrocyte sedimentation rate in poly-
REFERENCES myalgia rheumatica and giant cell arteritis. Br J Rheumatol. 1994;
33:550-4.
1. Salvarani C, Cantini F, Hunder GG. Polymyalgia rheumatica and
giant-cell arteritis. Lancet 2008;372:234-45. 13. Myklebust G, Gran JT. A prospective study of 287 patients with
2. Kermani TA, Schafer VS, Crowson CS, Hunder GG, Gabriel SE, polymyalgia rheumatica and temporal arteritis: clinical and labo-
Matteson EL, et al. Increase in age at onset of giant cell arteritis: a ratory manifestations at onset of disease and at the time of diag-
population-based study. Ann Rheum Dis 2010;69:780-1. nosis. Br J Rheumatol 1996;35:1161-8.
3. Crowson CS, Matteson EL, Myasoedova E, Michet CJ, Ernste 14. Martinez-Taboada VM, Blanco R, Armona J, Uriarte E, Figueroa
FC, Warrington KJ, et al. The lifetime risk of adult-onset rheu- M, Gonzalez-Gay MA, et al. Giant cell arteritis with an erythro-
matoid arthritis and other inflammatory autoimmune rheumatic cyte sedimentation rate lower than 50. Clin Rheumatol 2000;19:
diseases. Arthritis Rheum 2011;63:633-9. 73-5.
4. Hunder GG, Bloch DA, Michel BA, Stevens MB, Arend WP, 15. Cid MC, Font C, Oristrell J, de la Sierra A, Coll-Vinent B, Lopez-
Calabrese LH, et al. The American College of Rheumatology Soto A, et al. Association between strong inflammatory response
1990 criteria for the classification of giant cell arteritis. Arthritis and low risk of developing visual loss and other cranial ischemic
Rheum 1990;33:1122-8. complications in giant cell (temporal) arteritis. Arthritis Rheum
5. Walvick MD, Walvick MP. Giant cell arteritis: laboratory predic- 1998;41:26-32.