A 74-year-old woman has recent onset of daily bitemporal headache but is otherwise well. Her general physical examination results are normal and the erythrocyte sedimentation rate (ESR) is moderately elevated, at 64 mm/h. You wonder whether additional medical history or physical examination findings will modify your suspicion of possible temporal arteritis (TA) or whether the historical features alone warrant proceeding to temporal artery biopsy.
A 53-year-old man has a 1-month history of fever and fatigue and reports a single episode of transient partial loss of vision in one eye. You believe that TA is among the diagnostic considerations but suspect that he is too young for this diagnosis. You wonder if additional medical history, physical examination, or laboratory testing will change the probability of TA sufficiently to alter your decision about the role of temporal artery biopsy, rather than pursuing diagnostic evaluation for carotid artery stenosis or other considerations first.
Why Is This An Important Question to Answer With a Clinical Examination?
When faced with a patient with headache, fatigue, or other possible presenting symptom of TA, clinicians must be able to correctly and confidently establish the diagnosis to prevent irreversible vision loss and to minimize the inappropriate evaluation and treatment of alternative diagnoses. Although headache is the most common reason for clinical suspicion of TA, no single type of headache or other clinical presentation is specific for TA, and the disorder is among the diagnostic considerations for many symptom complexes in older individuals. Our review will analyze the diagnostic value of these varied symptoms and signs in predicting the likelihood of TA among patients for whom there is a clinical suspicion of disease.
The first known report of a patient with TA was by Hutchinson in 1890.1 His case was that of man who was referred because of “red streaks on his head” that were painful and prevented him from wearing his hat; these proved to be swollen temporal arteries, which over time became firm and pulseless. It was not until 1932 that Horton et al2 described the first 2 cases of pathologically confirmed TA; both patients had fever, weakness, anorexia, weight loss, anemia, leukocytosis, and painful tender temporal arteries. Thus, many of the characteristic features of this newly described disease were present in these first few patients. Headache was absent. In 1937, headache was recognized as a common feature,3 and in 1938 vision loss was first reported.4 In the modern era, however, clinicians are unlikely to treat patients with such advanced disease and the full array of untreated symptoms.
The mortality of patients with treated TA, during follow-up periods as long as 12 years, is the same as for age-matched individuals without TA. For example, Matteson et al5 studied 205 of the patients with TA who formed the initial cohort for the development of the American College of Rheumatology classification criteria. During a mean of 7 years of follow-up, the survival for patients with giant-cell arteritis was nearly identical to that of age-matched controls; the standardized mortality ratio was 1.03. Other authors have also observed that no excess mortality exists among patients with TA during periods ranging from 4.5 to 12 years.6, 7, and 8 Unrecognized (and therefore untreated) patients may have a higher mortality, but no such natural history studies of untreated patients exist in the modern era.
Although preventing death may not be among the benefits of early diagnosis of TA, timely diagnosis and treatment will prevent vision loss. A prompt decision regarding further evaluation (including referral for temporal artery biopsy) and early initiation of treatment are the primary rationales for improving the clinical prediction of the diagnosis. In addition, clinicians may avoid an extensive evaluation for other causes of symptoms by establishing a proper diagnosis. Because systemic corticosteroids have been the standard therapy for TA for decades, few studies have determined the long-term incidence of vision loss among untreated patients. Several studies, however, have demonstrated a substantial reduction in the incidence of vision loss after institution of corticosteroid therapy. Even among patients with complete unilateral vision loss, prompt recognition and corticosteroid therapy will decrease the risk of vision loss in the contralateral eye.
Aiello et al9 reviewed the Mayo Clinic experience of 245 patients diagnosed with TA who had a complete ophthalmologic examination at diagnosis or early in the course of treatment. The estimated 5-year probability of developing vision loss after initiation of corticosteroid therapy was 1%; that of additional vision loss in patients who already had vision loss was 13%. These observations and others emphasize the importance of the early diagnosis and treatment of TA and of the clinical examination in identifying patients at risk for catastrophic vision outcomes.10, 11, 12
Estimates of the prevalence of TA have been fairly constant. Using population data from Olmsted County, Minnesota, Salvarani et al13 estimated the age-adjusted incidence for individuals aged 50 years or older to be 24.2 per 100 000 women and 8.2 per 100 000 men. In another report, prevalence estimates increased by age and were 200 per 100 000 individuals aged 50 years and older, and 1100 per 100 000 individuals aged 85 years and older.14 These findings are similar to those observed in a Swedish population study, in which the average annual incidence of TA among individuals older than 50 years was 22.2 per 100 000 and the incidence increased with age.15 In this study of 665 patients with TA proven by biopsy, only 1 patient was younger than 50 years. Other investigators have reported similar incidences.16, 17 That TA is predominantly a disease of older individuals has importance because of the aging of our society. In the US 2000 census (http://factfinder.census.gov/servlet/ACSSAFFFacts?_submenuId=factsheet_1&_sse=on; accessed June 15, 2008), 35 million individuals (12.4% of the population) were aged 65 years or older and 9 million (3.3% of the population) were aged 80 years or older; these proportions are expected to increase.
The relatively low prevalence of TA does not diminish its importance to clinicians because of the morbidity resulting from overlooking this disorder. In fact, the higher prevalence of TA (1.5%) in one large autopsy series suggests that the disorder may be either unrecognized or clinically occult in many cases.18 The vision prognosis of occult TA is, of course, unknown, and series that describe the frequency of signs and symptoms include only patients with clinically evident TA.
The clinical manifestations of TA are a direct consequence of local (or “arteritic”) and systemic inflammatory disease. Localized arterial inflammation, particularly in the smaller branches of the external carotid artery, cause endovascular damage, vessel stenosis, and occlusion, ultimately leading to tissue ischemia or necrosis. Examples of localized arteritic symptoms include jaw claudication, caused by involvement of the masticatory muscles, and vision loss caused by involvement of the ophthalmic or posterior ciliary arteries. The particular cytokine profile may contribute to the ischemic and prominent constitutional features, such as malaise, fever, or weight loss.19, 20
How to Elicit the Signs and Symptoms
The myriad signs and symptoms in patients with TA require familiarity with the most common ones, recognizing that many patients will demonstrate few symptoms and have a normal physical examination result. Headache, jaw claudication, vision complaints, polymyalgia rheumatica (PMR), and constitutional features in a patient older than 55 years are among the most common symptoms. A high index of suspicion will lead the clinician to pursue these features because they may not be part of routine history-taking. Headache quality (typically severe and throbbing; less often sharp, dull, or burning), location (may be diffuse or localized but is bitemporal in half of cases), and onset (typically acute) are key features to assess; however, the headache of TA is often nonspecific in character.21 Headache may actually be due to scalp tenderness, reported by the patient as pain when combing the hair or putting on a hat. The headache is a new headache that is either recent in onset or different from previous headaches among patients with a history of chronic headaches. The duration of the headache before seeking medical attention is commonly 2 to 3 months. Jaw claudication refers to pain in the proximal jaw near the temporomandibular joint that develops only after a brief period of chewing, especially food requiring vigorous mastication, such as steak or a bagel.
Clinicians must distinguish jaw claudication from other causes of jaw pain in elderly persons, such as disorders of the temporomandibular joint (in which pain begins right away with chewing) or ill-fitting dentures. Vision complaints commonly include sudden monocular blindness, but clinicians should ask patients about a stuttering onset of vision loss, amaurosis fugax, a field cut, or diplopia. As an inflammatory polyarthritis with tendon or bursal involvement, PMR typically causes abrupt onset of morning stiffness involving the neck, shoulders, and hips, with referred pain to the proximal arms and thighs; this explains the prominent myalgias.22 Although neoplasm and infection may be highly suspected in the older patient with fever, anorexia, weight loss, and malaise, systemic inflammatory disease such as TA may also cause these symptoms.
The physical examination result is frequently unremarkable in patients with TA, but the detection of certain abnormalities may increase the suspicion of disease. The patient's temperature and general appearance are important first steps. Abnormalities of the temporal arteries, including tenderness, reduced or absent pulsation, erythema, nodularity, or swelling, may be detected by light palpation just anterior and slightly superior to the tragus of the ear; following the pulse anteriorly along the temples and comparison with the contralateral side helps detect findings that may be remarkably focal. Scalp tenderness, usually near the temporal arteries, may also be evident by light palpation. The scalp and tongue should be inspected for ischemic or necrotic skin changes. The funduscopic examination, ideally with pupillary dilation, may reveal a pale or swollen disc (evidence of ischemic optic neuropathy)23 or retinal artery occlusion, whereas vision field testing may demonstrate a field cut. Joint examination may reveal reduced range of motion in the shoulder or hip because of pain or more distal synovitis, particularly of the wrist.
Search Strategy and Quality Review
We performed a MEDLINE search of English-language articles published between January 1966 and July 2000. Search terms included “temporal arteritis,” “giant cell arteritis,” “clinical features,” “diagnosis,” “diagnostic tests,” “sensitivity and specificity,” “medical history taking,” “physical examination,” “signs and symptoms,” and “erythrocyte sedimentation rate.” We identified additional references by the use of a previously published search strategy in The Rational Clinical Examination series.24 This strategy combined 10 exploded Medical Subject Headings (“physical examination,” “medical history taking,” “professional competence,” “sensitivity and specificity,” “reproducibility of results,” “observer variation,” “diagnostic tests, routine,” “decision support techniques,” “Bayes theorem,” “mass screening”) and 2 text-word categories (“sensitivity and specificity” and “physical examination”), and intersected with “temporal arteritis.” We identified additional articles, including those predating MEDLINE, through a hand search of the bibliographies of retrieved articles, previous reviews, monographs, and textbooks. Both authors independently reviewed all retrieved articles to determine their eligibility for our review and included only those articles in which agreement existed that the study had met our inclusion criteria. We sought no unpublished studies.
The purpose of our review is to determine the value of individual clinical features in predicting the likelihood of positive results from temporal artery biopsy. Eligible studies were, therefore, those in which the authors provided a detailed list of clinical features for patients suspected of having or confirmed to have TA. We excluded articles with limited data on clinical features and those with fewer than 7 patients with positive temporal artery biopsy results. Many early studies classified patients as having TA according to either the authors’ own clinical criteria alone or the presence of positive biopsy results. When a study considered both groups of patients as having TA, we required that at least 90% of included patients had undergone temporal artery biopsy and had had a positive result.
We classified each article by the pathologic criteria used to determine the presence of positive biopsy results and by the referral source for recruitment of patients. In some cases, authors published clinical data on the same or overlapping series of patients in more than 1 article. In these cases, if we could not determine with certainty that no overlap existed between the patients in these studies, we excluded all studies except for the report with the largest number of patients. Of 114 studies retrieved using our search strategy, 41 were eligible for our review. Twenty-one studies included patients with both positive and negative temporal artery biopsy results; these form the core of our review.
We determined whether the authors required any predetermined published clinical criteria for patient inclusion, such as the American College of Rheumatology criteria25 for the diagnosis of TA, or other criteria. When studies used such criteria to classify patients as having TA with positive biopsy results or TA with negative biopsy results, we considered a positive biopsy result to be the true reference standard and considered only those patients with such results to have the disease. In our analysis, we included only those clinical features that were cited by at least 2 studies.
We classified the quality of evidence in each study by 2 methods. First, we developed our own criteria that focused on the diagnostic criteria (Table 49-1). This step was necessary to distinguish studies that used biopsy result as a reference standard from those that used established clinical criteria. In addition, we graded the quality of each study with a classification scheme for levels of evidence adapted from that previously developed for The Rational Clinical Examination series (see Table 1-7).26 In this scheme of levels 1 through 5, the highest levels of evidence we found were in level 3 studies.
Table 49-1Temporal Arteritis Diagnostic Criteria Quality Score |Favorite Table|Download (.pdf) Table 49-1 Temporal Arteritis Diagnostic Criteria Quality Score
|Score ||Diagnostic Quality |
|1 ||Patients require biopsy confirmation to be classified as having temporal arteritis. |
|2 ||Patients are classified according to the presence of predefined established clinical criteria for temporal arteritis and on biopsy results. |
|3 ||All patients meet predefined established clinical criteria for temporal arteritis. The authors consider patients with negative biopsy results to have temporal arteritis if they meet established clinical criteria for temporal arteritis. |
|4 ||A series of consecutive patients with temporal arteritis proven by biopsy. No controls or patients with negative biopsy results. |
|5 ||No use of established clinical criteria. Patients do not require biopsy confirmation to be classified as having temporal arteritis. |
|6 ||The investigators require the presence of a particular symptom (eg, visual problems) in all patients with temporal arteritis. |
27, 28 Uncertainty in these measures is reflected in the broad 95% CIs around the estimates.
Twenty-one studies that met our inclusion criteria included patients with both positive and negative temporal artery biopsy results and form the basis of our review (Table 49-2). These studies reported clinical findings on a total of 2680 patients, 1050 of whom had positive temporal artery biopsy results. The overall prevalence (prior probability) of positive biopsy results among patients with a clinical suspicion of TA in these studies was 39%. All but 4 of the studies were retrospective chart reviews. Eleven of the studies were of the highest quality (study quality 1) according to our predetermined criteria, and 19 of the studies included all patients who had a temporal artery biopsy during the study period.
Table 49-2Characteristics of Studies That Include Patients With Both Positive and Negative Temporal Artery Biopsy Results |Favorite Table|Download (.pdf) Table 49-2 Characteristics of Studies That Include Patients With Both Positive and Negative Temporal Artery Biopsy Results
|Study, y ||Diagnostic Quality/Level of Evidencea ||Study Type ||No. of Patients ||Positive Biopsy Results, No. (%) ||Referral Sourceb,c ||Pathologic Criteria Used to Establish Positive Biopsy Resultsc ||Comments |
|Gabriel et al,29 1995 ||1/3 ||Retrospective ||525 ||172 (33) ||All ||Achkar et al30 || |
|Hayreh et al,31 1997 ||1/3 ||Prospective ||363 ||106 (29) ||All ||Author || |
|McDonnell et al,32 1986 ||1/3 ||Retrospective ||250 ||42 (17) ||Specialty ||Author || |
|Hall et al,33 1983 ||1/3 ||Retrospective ||134 ||46 (34) ||All ||Not stated || |
|Fernandez Herlihy,34 1988 ||1/3 ||Retrospective ||107 ||29 (27) ||All ||Author ||Omitted group C patients with equivocal biopsies |
|Chmelewski et al,35 1992 ||1/3 ||Retrospective ||98 ||30 (31) ||All ||Author || |
|Fauchald et al,36 1972 ||1/3 ||Retrospective ||94 ||61 (65) ||All ||Not stated ||Comparison group patients all had PMR |
|Stuart,37 1989 ||1/3 ||Retrospective ||75 ||14 (19) ||All ||Allsop and Gallagher38 || |
|Kent and Thomas,39 1990 ||1/3 ||Retrospective ||70 ||8 (11) ||All ||Not stated || |
|Roth et al,40 1984 ||1/3 ||Retrospective ||51 ||7 (14) ||All ||Not stated || |
|Bevan et al,41 1968 ||1/4 ||Retrospective ||37 ||28 (76) ||All ||Author ||Arteritis and giant cells pooled as biopsy-result positive |
|Duhaut et al,42 1999 ||2/3 ||Prospective ||292 ||207 (71) ||All ||McDonnell et al32 ||All patients >50 y old, ESR >40 mm/h, response to 72 h of corticosteroids |
|Baldursson et al,43 1994 ||2/3 ||Retrospective ||133 ||127 (96) ||All ||ACR || |
|Gonzalez et al,44 1989 ||2/4 ||Retrospective ||21 ||10 (48) ||All ||Not stated ||All patients met clinical criteria for GCA |
|Genereau et al,45 1999 ||3/3 ||Retrospective ||37 ||19 (51) ||All ||ACR || |
|Vilaseca et al,46 1987 ||3/4 ||Retrospective ||103 ||45 (44) ||All ||Allsop and Gallagher38 || |
|Gur et al,47 1996 ||3/4 ||Retrospective ||39 ||30 (77) ||Specialty and PCP ||Banks et al48 ||All patients met ACR criteria for GCA |
|Brittain et al,49 1991 ||5/4 ||Prospective ||31 ||15 (48) ||Not stated ||Not stated || |
|Hedges et al,50 1983 ||5/5 ||Retrospective ||91 ||28 (31) ||All ||Author ||Patients excluded if adequate chart documentation of history-taking was absent |
|Skaug et al,51 1995 ||6/3 ||Retrospective ||98 ||13 (13) ||Specialty ||Not stated ||All patients had eye complaints |
|Dixon et al,52 1966 ||6/4 ||Prospective ||31 ||13 (42) ||Specialty ||Author ||All patients had PMR |
Precision of the Medical History and Physical Examination for Temporal Arteritis
No study that met our inclusion criteria evaluated the precision (ie, interobserver or intraobserver variation) of the medical history and physical examination for the diagnosis of TA. Most of the studies cited in this review are retrospective chart reviews and did not use standardized instruments for eliciting signs and symptoms across different observers. We therefore restrict our discussion to the accuracy of clinical findings.
Accuracy of Symptoms for the Diagnosis of Temporal Arteritis
Among the studies that included data on patients both with positive and negative temporal artery biopsy results, 14 symptoms were cited by at least 2 studies (Table 49-3). A limitation of our approach is that authors reported some findings much more frequently than others. However, our review incorporates the full extent of the published experience and presumably these reports include all of the major clinical features. Only 2 symptoms had LRs of sufficient power to be useful to clinicians. Jaw claudication had the highest LR+ (4.2), which is consistent with the traditional clinical teaching that jaw claudication, although somewhat insensitive, is a relatively specific feature for TA. When we pooled the sensitivity data from all eligible studies, including those studies that reported only patients with positive temporal artery biopsy results,20, 25, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, and 71 jaw claudication was present in only 34% of patients with disease (Table 49-4).
Table 49-3Summary Likelihood Ratios for Symptoms Among Patients With Suspected Temporal Arteritis |Favorite Table|Download (.pdf) Table 49-3 Summary Likelihood Ratios for Symptoms Among Patients With Suspected Temporal Arteritis
|Symptom/References ||No. of Patients With Data on Variablea ||LR+ (95% CI) ||LR– (95% CI) |
|Jaw claudication29, 31-35, 37, 39, 40, 42, 44, 45, and 46, 50, 51, 52 ||2314 ||4.2 (2.8-6.2) ||0.72 (0.65-0.81) |
|Diplopia33, 34, 42, 50, 51 ||703 ||3.4 (1.3-8.6) ||0.95 (0.91-0.99) |
|Temporal headache36, 42 ||386 ||1.5 (0.78-3.0) ||0.82 (0.64-1.0) |
|Weight loss31, 34, 36, 37, 39, 41, 42, 46, 47 ||1417 ||1.3 (1.1-1.5) ||0.89 (0.79-1.0) |
|Anorexia34, 37, 39, 41, 42, 46 ||674 ||1.2 (0.96-1.4) ||0.87 (0.75-1.0) |
|Fatigue31, 33, 37, 39, 41, 42, 44, 46 ||1095 ||1.2 (0.98-1.4) ||0.94 (0.86-1.0) |
|Fever29, 31, 34, 35, 36, 37, 40, 41, and 42, 46, 47 ||1708 ||1.2 (0.98-1.4) ||0.92 (0.85-0.99) |
|Any headache29, 31, 32, 33, 34, 35, 37, 39, 40, 41, 42, 43, 44, 45, 46, and 47, 50, 51 ||2475 ||1.2 (1.1-1.4) ||0.7 (0.57-0.85) |
|Arthralgia33, 34, 37, 39, 40, 44, 46, 52 ||582 ||1.1 (0.86-1.4) ||1.0 (0.92-1.1) |
|Any vision symptom29, 32, 33, 34, 35, 36, and 37, 39, 40, 41, and 42, 44, 45, 46, 47, 51, 52 ||2083 ||1.1 (0.93-1.3) ||0.97 (0.9-1.0) |
|Polymyalgia rheumatica29, 34, 35, 37, 39, 40, 42, 44, 45, 47, 50 ||1383 ||0.97 (0.76-1.2) ||0.99 (0.83-1.2) |
|Myalgia31, 36, 39, 40, 46 ||681 ||0.93 (0.81-1.1) ||1.1 (0.87-1.3) |
|Unilateral vision loss32, 50 ||341 ||0.85 (0.58-1.2) ||1.2 (1.0-1.3) |
|Vertigo34, 36, 44 ||212 ||0.71 (0.38-1.3) ||1.1 (0.93-1.2) |
Table 49-4Summary Sensitivity of Symptoms Among All Patients With Positive Temporal Artery Biopsy Resultsa |Favorite Table|Download (.pdf) Table 49-4 Summary Sensitivity of Symptoms Among All Patients With Positive Temporal Artery Biopsy Resultsa
|Variable ||No. of Studies ||Sensitivity (95% CI) |
|Any headache ||32 ||0.76 (0.72-0.79) |
|Temporal headache ||8 ||0.52 (0.36-0.67) |
|Weight loss ||19 ||0.43 (0.35-0.53) |
|Fever ||26 ||0.42 (0.33-0.52) |
|Fatigue ||19 ||0.39 (0.28-0.52) |
|Myalgia ||8 ||0.39 (0.23-0.56) |
|Any vision symptom ||35 ||0.37 (0.30-0.44) |
|Anorexia ||12 ||0.35 (0.23-0.48) |
|Polymyalgia rheumatica ||30 ||0.34 (0.28-0.41) |
|Jaw claudication ||35 ||0.34 (0.29-0.41) |
|Arthralgia ||13 ||0.30 (0.21-0.40) |
|Unilateral vision loss ||11 ||0.24 (0.14-0.36) |
|Facial pain ||4 ||0.17 (0.12-0.23) |
|Bilateral vision loss ||7 ||0.15 (0.07-0.27) |
|Vertigo ||4 ||0.11 (0.05-0.19) |
|Diplopia ||14 ||0.09 (0.07-0.13) |
More surprising was the finding that diplopia was the next most predictive symptom, with an LR+ of 3.4. Although the presence of diplopia substantially increases the likelihood of disease, the absence of diplopia does not significantly modify the probability of disease (LR–, 0.95) because of its low sensitivity (9% among all studies). We derived this value from 5 studies that evaluated this feature; previous reviews and textbooks have not emphasized the importance of diplopia. No other symptom had an LR+ exceeding 2. This includes features often thought to be useful to clinicians, such as fever, PMR, vision loss, and temporal headache. The LR– of all 14 symptoms was near 1. In other words, the absence of any particular symptom did not rule out TA or make the disorder substantially less likely. Patients with positive temporal artery biopsy results had a mean duration of symptoms of 3.5 months before diagnosis; this was 1.5 months (95% CI, 0.4-2.5 months) shorter than those with negative biopsy results, emphasizing the relatively acute onset of symptoms of biopsy-proven TA and that a longer duration of symptoms makes a positive temporal artery biopsy result less likely.
Accuracy of the Physical Examination for the Diagnosis of Temporal Arteritis
Findings on physical examination were more likely to influence the probability of positive temporal artery biopsy results than were historical features (Table 49-5). The presence of synovitis made positive temporal artery biopsy results significantly less likely (LR+, 0.41). The absence of any temporal artery abnormality also made disease substantially less likely (LR, 0.53). Scalp tenderness, a finding often thought to be specific for TA, did not perform well as a predictor of positive biopsy results. Among patients in whom TA was suspected, the frequency of scalp tenderness was similar in patients with and without the disease (LR+, 1.6).
Table 49-5Summary Likelihood Ratios for Signs and Demographics and Laboratory Data Among Patients With Suspected Temporal Arteritis |Favorite Table|Download (.pdf) Table 49-5 Summary Likelihood Ratios for Signs and Demographics and Laboratory Data Among Patients With Suspected Temporal Arteritis
|Variable/References ||No. of Patients With Data on Variablea ||LR+ (95% CI) ||LR– (95% CI) |
|Signs and Demographics |
|Beaded temporal artery42, 52 ||323 ||4.6 (1.1-18.4) ||0.93 (0.88-0.99) |
|Prominent or enlarged temporal artery36, 39, 42, 44, 52 ||508 ||4.3 (2.1-8.9) ||0.67 (0.5-0.89) |
|Absent temporal artery pulse41, 52 ||68 ||2.7 (0.55-13.4) ||0.71 (0.38-1.3) |
|Tender temporal artery36, 39, 40, 41, and 42, 50, 51, and 52 ||755 ||2.6 (1.9-3.7) ||0.82 (0.74-0.92) |
|Any temporal artery abnormality29, 31, 32, and 33, 37, 43, 46b ||1559 ||2.0 (1.4-3.0) ||0.53 (0.38-0.75) |
|Scalp tenderness31, 33, 34, 35, 52 ||923 ||1.6 (1.2-2.1) ||0.93 (0.86-1.0) |
|Optic atrophy or ischemic optic neuropathy40, 50 ||142 ||1.6 (1.0-2.5) ||0.8 (0.58-1.1) |
|Any funduscopic abnormality29, 35, 50, 52 ||745 ||1.1 (0.8-1.4) ||1.0 (0.92-1.1) |
|White race32, 35, 37, 40, 50 ||565 ||1.1 (0.99-1.2) || |
|Male sex29, 31, 32, 33, 34, 35, 36, and 37, 40, 41, 42, aand 43, 45, 46, 47, 49, 50, 51, 52 ||2565 ||0.83 (0.72-0.96) || |
|Synovitis29, 37, 46, 52 ||734 ||0.41 (0.23-0.72) ||1.1 (1.0-1.2) |
|Laboratory Data |
|ESR || || || |
|>100 mm/h35, 49, 50 ||220 ||1.9 (1.1-3.3) ||0.8 (0.68-0.95) |
|>50 mm/h35, 47, 49, 50 ||259 ||1.2 (1.0-1.4) ||0.35 (0.18-0.67) |
|Abnormal32, 37, 42, 46, 49, 50, and 51 ||941 ||1.1 (1.0-1.2) ||0.2 (0.08-0.51) |
|Anemia31, 32, 34, 35, 37, 46, 47, 49 ||1057 ||1.5 (0.82-2.9) ||0.79 (0.6-1.0) |
Abnormal findings on examination of the temporal artery increased the probability of positive biopsy results and predicted disease to a greater extent than any other variable. Beading, prominence, or enlargement of the temporal artery all conferred LR+s of greater than 4. A tender temporal artery also suggested an increased probability of positive biopsy results (LR, 2.6). An absent temporal artery pulse showed a trend toward a useful LR+; the value of 2.7 was, however, not statistically different from 1. The LRs for “any temporal artery abnormality” may underestimate their power. If eligible studies did not list clinical features separately for each patient, it was not possible to determine whether specific temporal artery abnormalities overlapped; in such cases, we made the most conservative calculation about the actual number of patients with any temporal artery abnormality.
Table 49-6). In each study, physicians would have referred patients for a temporal artery biopsy when they believed the diagnosis to be sufficiently likely to justify a biopsy. These patients represent a selected sample who often manifested several clinical features of interest, including those analyzed in this review. Patients who lacked features commonly considered suggestive of TA were presumably less likely to have a temporal artery biopsy. This verification bias makes the value of those few findings with the highest positive LRs even greater because they help predict biopsy results among patients with a significant clinical suspicion of disease.
Table 49-6Summary Sensitivity of Signs and Demographics and Laboratory Data Among All Patients With Positive Temporal Artery Biopsy Resultsa |Favorite Table|Download (.pdf) Table 49-6 Summary Sensitivity of Signs and Demographics and Laboratory Data Among All Patients With Positive Temporal Artery Biopsy Resultsa
|Variable ||No. of Studies With Data on Variable ||Sensitivity (95% CI) |
|Signs and Demographics |
|White race ||11 ||0.86 (0.62-0.97) |
|Any temporal artery abnormality ||16 ||0.65 (0.54-0.74) |
|Prominent or enlarged temporal artery ||6 ||0.47 (0.40-0.54) |
|Absent temporal artery pulse ||6 ||0.45 (0.26-0.66) |
|Tender temporal artery ||13 ||0.41 (0.30-0.52) |
|Male sex ||40 ||0.32 (0.29-0.35) |
|Any funduscopic abnormality ||6 ||0.31 (0.14-0.54) |
|Scalp tenderness ||13 ||0.31 (0.20-0.44) |
|Optic atrophy or ischemic optic neuropathy ||4 ||0.29 (0.10-0.57) |
|Beaded temporal artery ||3 ||0.16 (0.07-0.28) |
|Laboratory Data |
|ESR || || |
|Abnormal ||24 ||0.96 (0.93-0.97) |
|>50 mm/h ||14 ||0.83 (0.75-0.90) |
|>100 mm/h ||10 ||0.39 (0.29-0.50) |
|Anemia ||22 ||0.44 (0.34-0.54) |
TA is more common among women than men and among whites than blacks. The LRs do not reflect this observation, perhaps because referring physicians incorporated this knowledge into their decisions about which patients to refer for biopsy. However, if one pools the data from all eligible studies, including those that reported only patients with positive temporal artery biopsy results, TA was 2.1 times more common in women than men (Table 49-6). TA among black patients in published reports is restricted largely to small case series,66 and white patients constituted 86% of all patients with positive biopsy results.
Among patients referred for biopsy, the average age of those with positive results was 73 years; this was only 3.8 years (95% CI, 2.1-5.4 years) older than the average age of patients with negative results. Age was, however, a valuable criterion for predicting the likelihood of TA. When data for all eligible studies were reviewed, including those that reported only patients with positive biopsy results, 26 studies provided sufficient data to determine the age range of patients with biopsy-proven TA. Only 2 patients among a total of 1435 patients were younger than 50 years; this resulted in a sensitivity of 99% for the criterion of age older than 50 years. This outcome suggests that clinicians should consider TA only as a diagnostic possibility in a person younger than 50 years if multiple characteristic or high-probability features are present.
Accuracy of the Laboratory Evaluation for the Diagnosis of Temporal Arteritis
Although the primary purpose of this analysis was to determine the operating characteristics of the medical history and physical examination in diagnosis, clinicians usually obtain an ESR before determining which patients have sufficient likelihood of TA to justify a referral for biopsy. We therefore chose to evaluate the test characteristics of the ESR. The mean value for patients with disease was 88 mm/h; that for patients without disease was a mean of 10 mm/h lower (95% CI, 4-25 mm/h). This difference was not statistically significant.
Results of the ESR measurement were a valuable guide to clinicians; a low or normal level was more likely to rule out disease than a high value was likely to rule in disease. Previously, Miller et al72 had determined normal ESR values among 27 912 adults without apparent disease and suggested defining the upper limit of normal ESR as either age/2 (for men) or as (age + 10)/2 (for women). In our source studies, authors most commonly did not define “normal” ESR; it was not possible to determine whether these normal values were adjusted for age. With this caveat, a normal ESR made TA unlikely; the LR for a normal ESR was 0.2 (Table 49-5). When we separately analyzed the pooled data from all studies, only 4% of patients with positive temporal artery biopsy results and data on ESR had a normal value. If one uses a less strict cutoff point, even an ESR of less than 50 mm/h substantially reduces the probability of disease (LR, 0.35). This value is lower than the LR– of any symptom or sign.
In contrast to clinical lore, a high ESR was less useful in identifying those with TA among all patients referred for biopsy, which likely relates to the verification bias inherent in patient selection for the eligible studies because referring physicians would have had knowledge of the ESR before recommending a biopsy. Although an ESR of greater than 100 mm/h conferred an LR+ of 1.9, this value is less than the most useful symptoms and signs. In contrast, mean ESR values were similar for patients with and without positive temporal artery biopsy results.
Anemia was present in 44% of patients with biopsy-proven TA. This finding was present in a similar number of patients who had negative biopsy results. Mean hemoglobin levels were similar between patients with positive and negative biopsy results (11.6 g/dL vs 12.4 g/dL, respectively); the lack of anemia was not helpful in ruling out disease.
Are These Clinical Features Ever Normal?
The presence of particular symptoms or signs in patients with negative temporal artery biopsy results does not imply that these findings are “normal” or common in patients without disease. Rather, it suggests that other conditions that clinicians may initially confuse for TA have overlapping clinical features. The frequency of such findings in randomly selected individuals of the same age would likely be lower than the frequency among patients in this review with negative biopsy results.
Several studies have followed patients with negative biopsy results to determine their ultimate or correct diagnoses. Chmelewski et al35 reported the outcomes of 98 patients undergoing temporal artery biopsies during a 5-year period at their institution. Among the 68 patients with negative biopsy results, 15 proved to have neurologic disorders (including migraine, stroke, and optic neuropathy), 14 had PMR, 10 had other rheumatologic disorders (including vasculitis other than TA, rheumatoid arthritis, and CREST [calcinosis, Raynaud disease, esophageal dysmotility, sclerodactyly, telangiectasia] syndrome), and 4 had fever of unknown origin. Miscellaneous diagnoses included sinusitis, endocarditis, amyloidosis, and malignancy. In another biopsy series, Roth et al40 studied 33 patients with a clinical suspicion of TA but negative biopsy results. The most common diagnoses, in descending order, were joint disease (degenerative or rheumatoid), malignant lymphoma, arteriosclerotic carotid artery disease, diabetes mellitus, and ischemic optic neuropathy.
In our first clinical scenario, the history of bitemporal headache and a modestly increased ESR would be among those factors that may lead a clinician to suspect TA. In this setting, one would seek the potential additional history of jaw claudication or diplopia and determine the presence of a prominent, tender, or beaded temporal artery. If present, these factors would substantially increase the likelihood of positive temporal artery biopsy results.
In the second scenario, TA is among the diagnostic considerations for transient partial monocular vision loss in the setting of a constitutional illness. The history in this case is sufficiently compelling to justify a temporal artery biopsy. Given the high prior probability and the poor performance of historical and examination features in excluding disease, an otherwise normal medical history and physical examination result would not sufficiently reduce the likelihood of TA to avoid the need for a temporal artery biopsy. A normal ESR would, however, reduce the likelihood of disease by a factor of 0.2 and should prompt consideration of alternative diagnoses.
Available data suggest that many of the clinical features commonly found in patients with the disease are unhelpful in predicting the likelihood of positive temporal artery biopsy results. Our study evaluates the predictive value of clinical features among patients who are already clinically suspected of having the disease, as determined by the clinicians who referred them for biopsy. Although we could not determine, from the primary studies, the factors that went into the decision to refer for biopsy, certain clinical features modified the likelihood of disease among these patients. It is likely that these same clinical factors would be useful to consider at initial evaluation, even before the decision to proceed to biopsy. In addition, the verification bias inherent in this analysis makes the significance of our results greater because they help to predict biopsy results even among patients who have a higher prior probability of disease than do unselected patients with any particular clinical feature.
When a medical history is taken in a patient with possible TA, jaw claudication and diplopia substantially increase the probability of positive biopsy results (LR+s, 4.2 and 3.4, respectively). No symptoms help rule out the diagnosis by their absence. Among physical examination findings, synovitis makes the diagnosis of TA less likely, whereas beaded, prominent, enlarged, and tender temporal arteries increase the likelihood of positive biopsy results. Beaded, prominent, or enlarged arteries confer the highest positive LRs of any clinical or laboratory feature and substantially increase the probability that a patient with suspected TA will have positive biopsy results. Although these findings increase the chance of having TA, they are variably sensitive, from 16% (beaded temporal artery) to 65% (any temporal artery abnormality).
The results of tests of ESR alter the likelihood of positive biopsy results. A normal ESR (LR, 0.2) or ESR less than 50 mm/h (LR, 0.35) makes positive biopsy results less likely, but setting the ESR threshold at 100 mm/h is less efficient because patients with an ESR less than 100 mm/h have an LR (0.8) that only slightly decreases the likelihood of disease. Among patients clinically suspected of having disease, those with an ESR greater than 100 mm/h have a modestly increased likelihood of biopsy-proven TA (LR, 1.9).
The clinician faced with a patient who may have TA has a difficult diagnostic challenge. The goal is to rule out other morbid conditions that may mimic TA, to avoid unnecessary evaluation, and to quickly and correctly identify and treat patients who do in fact have the disorder. Given the extreme difference in prevalence of TA between the general population (<1%) vs those referred for temporal artery biopsy (39%), we infer that clinicians are adept at identifying patients at high risk for disease. Many clinicians choose to treat patients they have referred for biopsy with corticosteroids, in the absence of contraindication, pending biopsy results. Although this strategy would appear particularly wise in the presence of a factor that we have shown predicts likelihood of disease, this approach deserves further study.
Our review of clinical series of patients with suspected TA does not allow a determination of the predictive value of selected combinations of clinical and laboratory features. In addition, it is not possible to determine from our data whether certain combinations of features would sufficiently increase the likelihood of disease that a clinician should treat presumptively for TA and not perform a biopsy at all. The morbidity of a prolonged course of corticosteroids, however, is such that most clinicians would favor confirmation of disease by biopsy even if the clinical probability is high.
Our analysis demonstrates that a limited number of clinical features substantially modify the probability of the diagnosis of TA among patients suspected of having the disease. Ultimately, the clinician must integrate multiple clinical factors to optimize diagnostic and therapeutic strategies for patients with suspected TA.
Author Affiliations at the Time of the Original Publication
Division of General Medicine and Primary Care (Dr Smetana) and Division of Rheumatology (Dr Shmerling), Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, Massachusetts.
We appreciate the expert advice offered by Stephanie Studenski, MD, and Kenneth Schmader, MD, during the preparation of our manuscript. We also wish to thank David Simel, MD, MHS, for his thoughtful guidance, review of the manuscript, and statistical advice.
J. Diseases of the arteries: on a peculiar form of thrombotic arteritis of the aged which is sometimes productive of gangrene. Arch Surg.[Archives of Surgery Full Text]
G. An undescribed form of arteritis of the temporal vessels. Proc Staff Mtg Mayo Clin. 1932;7:700–701.
T. Arteritis of the temporal vessels: report of seven cases. Proc Staff Mtg Mayo Clin. 1937;12:548–553.
G. Arteritis of the temporal vessels. Lancet.
GG. Long-term survival of patients with giant cell arteritis in the American College of Rheumatology giant cell arteritis classification criteria cohort
. Am J Med.
of giant cell arteritis including temporal arteritis and polymyalgia rheumatica. Acta Med Scand.
et al.. Giant cell arteritis in Lugo, Spain, is associated with low longterm mortality. J Rheumatol.
et al.. Survival in giant cell arteritis: a survey
of 87 patients [letter]. J Rheumatol.
JM. Clinical features in patients with permanent visual loss due to biopsy-proven giant cell arteritis. Br J Rheumatol.
JL. Visual morbidity in giant cell arteritis: clinical characteristics and prognosis
for vision. Ophthalmology.
GG. The incidence
of giant cell arteritis in Olmsted County, Minnesota: apparent fluctuations in a cyclic pattern. Ann Intern Med.
et al.. Estimates of the prevalence
of arthritis and selected musculoskeletal disorders in the United States. Arthritis Rheum.
C. The epidemiology of biopsy positive giant cell arteritis: special reference to cyclic fluctuations. Rheumatology
S. Giant cell arteritis, temporal arteritis, and polymyalgia rheumatica in a Danish county: a prospective investigation. Arthritis Rheum.
JJ. Weyand CM. Correlation
of interleukin-6 production and disease activity in polymyalgia rheumatica and giant cell arteritis. Arthritis Rheum.
B. Goronzy JJ. Disease patterns and tissue cytokine profiles in giant cell arteritis. Arthritis Rheum.
et al.. Proximal bursitis in active polymyalgia rheumatica. Ann Intern Med.
B. Ocular manifestations of giant cell arteritis. Am J Ophthalmol.
et al.. The American College of Rheumatology 1990 criteria for the classification of giant cell arteritis. Arthritis Rheum.
R. Meta-analysis by the Confidence Profile Method: The Statistical Synthesis of Evidence
. San Diego, CA: Academic Press; 1992.
V. Fast∗Pro Software for Meta-analysis by the Confidence Profile Method
. 1.8 ed. San Diego, CA: Academic Press; 1992.
GG. The use of clinical characteristics to predict the results of temporal artery biopsy among patients with suspected giant cell arteritis. J Rheumatol.
SE. How does prior corticosteroid treatment affect the biopsy findings in giant cell arteritis? Ann Intern Med.
WR. Temporal arteritis: a clinicopathologic study. Ophthalmology.
GG. The therapeutic impact of temporal artery biopsy. Lancet.
L. Temporal arteritis: clinical aids to diagnosis [published correction appears in J Rheumatol
. 1989;16:260]. J Rheumatol.
B. Temporal arteritis and polymyalgia rheumatica: clinical and biopsy findings. Ann Intern Med.
RA. Temporal artery biopsy in suspected temporal arteritis: a five year survey
. N Z Med J.
P. Temporal artery biopsy in giant cell arteritis: a reappraisal. Am J Surg Pathol.
L. Temporal artery biopsy. Am Surg.
MJ. Clinical and biopsy findings in temporal arteritis. Ann Rheum Dis.
the. Groupe de Recherche sur l’Arterite a Cellules Geantes. Biopsy proven and biopsy negative temporal arteritis: differences in clinical spectrum at the onset of the disease. Ann Rheum Dis.
JT. Giant cell arteritis in Iceland: an epidemiologic and histopathologic analysis. Arthritis Rheum.
et al.. Temporal 67-gallium uptake is increased in temporal arteritis. Rheumatology
A. Clinical usefulness of temporal artery biopsy. Ann Rheum Dis.
Y. Clinical manifestations of temporal arteritis: a report from Israel. J Rheumatol.
GG. Immunohistologic and cytochemical studies of temporal arteritis. Arthritis Rheum.
J. Plasma viscosity or erythrocyte sedimentation rate in the diagnosis of giant cell arteritis? Br J Ophthalmol.
G. Clinical usefulness of biopsy in giant cell arteritis. Acta Ophthalmol Scand.
Y. Polymyalgia rheumatica and temporal arteritis. Ann Rheum Dis.
CM. Disease pattern in cranial and large-vessel giant cell arteritis. Arthritis Rheum.
J. Erythrocyte sedimentation rate in temporal arteritis. South Med J.
et al.. Association between strong inflammatory response and low risk of developing visual loss and other cranial ischemic complications in giant cell (temporal) arteritis. Arthritis Rheum.
E. Giant cell arteritis: a five-year review of biopsy proven cases in a teaching hospital. Med J Aust.
H. Temporal arteritis: the silent presentation and delay in diagnosis. J Intern Med.
NG. Ophthalmoplegia, ptosis, and miosis in temporal arteritis. Neurology.
B. Temporal arteritis in Israel: a review of 47 patients. J Rheumatol.
Glutz von Blotzheim
FX. Neuro ophthalmic complications of biopsy proven giant cell arteritis. Eur J Ophthalmol.
et al.. Permanent visual loss and cerebrovascular accidents in giant cell arteritis: predictors and response to treatment. Arthritis Rheum.
L. Temporal arteritis in Rochester, Minnesota, 1951 to 1967. Mayo Clin Proc.
KR. Presentation of occult giant cell arteritis. Arthritis Rheum.
RA. Temporal arteritis: a 14-year epidemiological, clinical and prognostic study
. Scott Med J.
et al.. Temporal arteritis in blacks. Ann Intern Med.
et al.. Trends in incidence
and clinical presentation of temporal arteritis in Olmsted County, Minnesota, 1950-1985. Arthritis Rheum.
NC. Temporal (granulomatous) arteritis: a histopathological study of 32 cases. Histopathology.
JT. A prospective study
of 287 patients with polymyalgia rheumatica and temporal arteritis: clinical and laboratory manifestations at onset of disease and at the time of diagnosis. Br J Rheumatol.
I. Observations on temporal arteritis. Acta Med Scand.
D. Simple rule for calculating normal erythrocyte sedimentation rate. Br Med J
(Clin Res Ed