Giant cell arteritis (GCA) is “the prime medical emergency in ophthalmology.” No other condition has the potential to produce such rapid, profound, irreversible bilateral visual loss, even when managed appropriately. Ophthalmologists should be acutely aware of the 3 critical junctures in the management of GCA to prevent blindness: early diagnosis, adequate initial corticosteroid dosage, and appropriate duration of treatment. Histologic confirmation of the diagnosis with a temporal artery biopsy makes an important contribution to all 3 of these stages. A positive tissue diagnosis provides the clinician with the conviction to maintain high dosages of corticosteroids for a prolonged period, even if significant steroid-induced side-effects develop in the patient. Conversely, biopsy confirmation of GCA is important for preventing unnecessary corticosteroid therapy for long duration in patients who do not actually have the disease. Despite the overwhelming and seemingly obvious need to perform a temporal artery biopsy as part the GCA work-up and management plan, there is a school of thought that disputes the contribution of this investigation. The role of the temporal artery biopsy in GCA is challenged continuously, with up to one third of clinicians surveyed not recommending a temporal artery biopsy for the diagnosis of GCA.

Part of the explanation for this is the American College of Rheumatologists (ACR) clinical criteria for diagnosing GCA, delineated in 1990. A patient is considered to have GCA if they meet 3 of the following 5 criteria: (1) age 50 years or older, (2) new-onset localized headache, (3) temporal artery tenderness or decreased temporal artery pulse, (4) erythrocyte sedimentation rate (ESR) of at least 50 mm/hour, and (5) abnormal artery biopsy specimen characterized by mononuclear infiltration or granulomatous inflammation. These criteria were developed simultaneously with criteria for other vasculitides such as Wegener granulomatosis and polyarteritis nodosa. The criteria were validated through multicenter data collected by rheumatologists who retrospectively provided their diagnosis along with the clinical, laboratory, and histologic data that were available from the patient charts. Histologic confirmation was not required for the diagnosis of any of the vasculitides studied. In fact, the data set included some patients classified as having GCA who never underwent a biopsy or who had negative biopsy results. Instead, the gold standard in developing the ACR GCA criteria was considered the diagnosis of the referring rheumatologist.

Although the ACR criteria originally were developed as classification criteria for research, they are now used widely as the reference standard for the clinical diagnosis for GCA. This is particularly concerning for several reasons. The ACR criteria were not established to differentiate patients who have vasculitis from those who do not have vasculitis for diagnostic purposes, but rather to distinguish a specific type of vasculitis among patients with various vasculitides. In this context, the ACR criteria have a reported sensitivity of 93.5% and a reported specificity of 91.2% for the classification of GCA compared with other vasculitides. However, the ACR classification criteria do not have the same positive and negative predictive values in patients who may not have a previously established diagnosis of vasculitis because the capacity of any set of diagnostic criteria to identify a patient with GCA is better when disease prevalence is high, as in rheumatology clinics, rather than when disease prevalence is low, such as in ophthalmology or general medical clinics. Furthermore, if the criteria are applied to a population in which the prevalence of GCA is low, the false-positive rate also will be high.

The limitations of the ACR criteria to identify GCA are confirmed in this issue of the Journal by Murchison and associates, who compared these diagnostic criteria with temporal artery biopsy results in a group of patients from ophthalmologic clinics. One fourth of the patients with positive biopsy results would not have met sufficient clinical ACR criteria for the correct diagnosis had they not undergone a temporal artery biopsy and might have experienced severe visual consequences without the histologic confirmation and treatment. This is not surprising given that approximately 20% of patients with GCA present with only ophthalmologic symptoms—this form of occult GCA originally was highlighted 50 years ago by Simmons and Cogan. Furthermore, Murchison and associates found that 28.3% of patients met the ACR clinical criteria, but had negative biopsy results. In this group, the tissue diagnosis prevented the potential complications of long-term steroid treatment. The corticosteroid treatment of all of the patients who had negative temporal artery biopsy results ceased immediately; thus, 10% of the total cohort were spared inappropriate exposure to this potentially debilitating treatment.

It has been argued that a limitation of performing a temporal artery biopsy is the possibility of false-negative results, which occur most commonly if the tissue removed is too small, if there are so-called skip areas, or if bilateral biopsy is not performed. Addressing these issues will identify most cases of biopsy-negative GCA. Murchison and associates observed no subsequent visual loss in their cohort of biopsy-negative patients that could be attributed to GCA-related causes. Ten percent of their patients underwent bilateral temporal artery biopsies. This suggests that the false-negative rate of biopsy in fact is low. However, given that a second temporal artery biopsy has been shown to identify cases missed by a unilateral biopsy, it should be considered in patients with a high clinical index of suspicion for GCA. In an ophthalmologic setting with a clinical presentation of visual loss, the diagnosis of biopsy-negative GCA should be made with caution after extensive review of the pathologic results and clinical information.

There are additional consequences to using the ACR criteria for research into GCA. Misclassification of patients as GCA could hamper research attempts to identify new treatment strategies. With the advancing area of biologics, it is not an unrealistic expectation for more focused therapies for GCA to become available. Excellent laboratory research has implicated the NOTCH signaling pathway, the role of adaptive immunity, and T-cell pathologic features in vascular injury in GCA. However, if the clinical trials using agents that target these pathways resort to the ACR criteria, then a possible reason for failure of these agents may be inclusion of patients with diseases other than GCA in the trials.

There are promising new developments for the future of diagnosis of GCA. It is now well-established that C-reactive protein should be obtained in addition to ESR because of its higher specificity. The presence of thrombocytosis provides further contributory information, although in isolation is no more helpful than ESR or C-reactive protein data. Other biological markers are under investigation. Fibrinogen is particularly interesting because it often is at a normal level in patients with elevated ESR for reasons other than GCA. Interleukin-6 may mirror disease activity. Other potential biological markers that may have some role include alkaline phosphatase, haptoglobin, complement, and orsomucoid, although their usefulness has yet to be validated. An area that warrants more focused research is immunohistochemical analysis of the biopsy sample itself. There have been few advances in the assessment of the pathologic characteristics of the biopsy specimen that have influenced the interpretation of the temporal artery biopsy by the pathologist. We now appreciate that the length of the biopsy is critical and that there is significant contraction before and with formalin fixation. Markers of epithelioid histocytes, CD-68, and HAM-56 may assist in indeterminate cases and may identify evidence of granulomatous inflammation in patients already receiving corticosteroid therapy. Further studies are required to evaluate whether it is appropriate to use such techniques widely.

Although these new insights into GCA have the potential to modify clinical practice, the study by Murchison and associates confirms the opinion of many experts, that for the time being, GCA can be diagnosed best by histopathologic examination of temporal artery biopsy.