Is Duplex Ultrasonography Useful for the Diagnosis of Giant-Cell Arteritis?

Context Color duplex ultrasonography detects dark halos around the temporal arteries of some patients with giant-cell arteritis. However, is ultrasonography any better than clinical examination for detecting biopsy-proven giant-cell arteritis? Contribution Patients with biopsy-proven giant-cell arteritis had tender temporal arteries, with or without altered pulses (sensitivity, 67% [95% CI, 38% to 88%]), more often than they had halos on ultrasonographic examination (sensitivity, 40% [CI, 16% to 68%]). Abnormal results on physical examination increased the probability of a positive biopsy result more than did halos 1 mm or greater in thickness (likelihood ratios, 47.3 [CI, 6.5 to 342.4] vs. 5.7 [CI, 2.0 to 16.2]). Implications Ultrasonography is not better than a careful physical examination for detecting biopsy-proven giant-cell arteritis. The Editors In 1990, the American College of Rheumatology developed classification criteria for giant-cell arteritis (1). In the traditional format, five criteria were chosen. The presence of at least three criteria was associated with a sensitivity of 93.5% and a specificity of 91.2%. The five criteria are age at disease onset of 50 years or older, headache with no history of headache, the presence of a temporal artery abnormality (tenderness on palpation or decreased or absent pulsation), erythrocyte sedimentation rate of 50 mm per hour or greater, and an artery biopsy specimen showing vasculitis with a predominance of mononuclear cells or granulomatous inflammation. Although these criteria were developed as classification criteria for research, clinicians often apply them for the diagnosis of giant-cell arteritis. However, their diagnostic utility remains controversial (2), and positive results on biopsy remain the gold standard for the diagnosis of temporal arteritis. A positive biopsy result can predict the likelihood of giant-cell arteritis in more than 90% of cases (3). Schmidt and colleagues (4) recently proposed that color duplex ultrasonography has a role in the diagnosis of temporal arteritis. These authors showed that evidence of a dark halo around the lumen of the temporal arteries, which may be due to edema of the artery wall, was the most specific sign of temporal arteritis on ultrasonography. They suggested that when typical signs of giant-cell arteritis were present, a clear halo allowed the diagnosis of giant-cell arteritis without performing a temporal artery biopsy. However, this study did not compare ultrasonography with physical examination of temporal arteries. It also did not clearly evaluate whether ultrasonography may be useful in patients with polymyalgia rheumatica or systemic illness in the absence of definitive clinical signs of giant-cell arteritis (5). In our study, we attempted to assess the usefulness of temporal artery ultrasonography for the diagnosis of giant-cell arteritis in a prospective cohort of patients with suspected giant-cell arteritis or polymyalgia rheumatica diagnosed over a 22-month period. We also compared the diagnostic accuracy of ultrasonography with that of physical examination of temporal arteries. A positive result on temporal artery biopsy was considered the gold standard for diagnosis of giant-cell arteritis. Methods From January 1998 to October 1999, 86 consecutive patients with suspected giant-cell arteritis or polymyalgia rheumatica were seen in the Departments of Internal Medicine, Rheumatology, Ophthalmology, Neurology, and Infectious Diseases of Reggio Emilia Hospital, Reggio Emilia, Italy. The study protocol was reviewed and approved by the Ethics and Research Committees of Reggio Emilia Hospital. Before entering the study, each patient was informed of the nature and purpose of the study and provided informed consent. Figure 1 is a flow diagram of the study. All patients had a clinical diagnosis of polymyalgia rheumatica or giant-cell arteritis and received a careful physical examination of the temporal and occipital arteries. Right and left common superficial temporal arteries with their parietal and frontal rami and right and left occipital arteries were examined for the presence or absence of tenderness on palpation and decreased or absent pulsation. The patients then had ultrasonographic evaluations performed by two ultrasonographers who were blinded to the clinical diagnosis. Temporal artery biopsy was the third step. It was performed in all patients at the site targeted by the ultrasonographer, usually where a halo was seen or where peak systolic velocity was reduced. The pathologist who read the biopsies had no knowledge of the clinical or ultrasonographic findings. None of the patients had been treated with corticosteroids before ultrasonography and temporal artery biopsy. The final diagnosis of giant-cell arteritis, polymyalgia rheumatica, or other unrelated disorder was made after histologic findings from the temporal artery biopsy became available and the clinical course was assessed. Figure 1. Flow diagram of the study Simultaneous color Doppler and duplex ultrasonography were performed by using a 5- to 10-MHz linear probe (Aspen, Acuson Corp., Mountain View, California) along the course of the common superficial temporal arteries and their branches and occipital arteries. These vessels were examined as thoroughly as possible in a longitudinal and transverse plane to evaluate whether a halo was present around the lumen. The thickness of the halo was determined by taking the measurements between the inner echogenic line of the vessel wall, which represented the media-adventitia surface, and the outer echogenic line of the vessel wall, which represented the normal connective tissue. Two trained ultrasonographers independently performed the examination and were unaware of the patients' diagnoses. We calculated sensitivities, specificities, and likelihood ratios with 95% CIs (6). A positive likelihood ratio (defined as sensitivity/1 specificity) is the increase in the odds of having giant-cell arteritis when there is evidence on ultrasonography of a halo or when inspection reveals a temporal artery abnormality. A negative likelihood ratio (defined as 1 sensitivity/specificity) is the decrease in the odds of having giant-cell arteritis when there is no evidence on ultrasonography of a halo or when inspection does not reveal a temporal artery abnormality. All analyses were performed by using the SPSS statistical package, version 10 (SPSS, Inc., Chicago, Illinois). Results All 86 eligible patients were enrolled in the study (Figure 1). Table 1 shows patient characteristics. A clinical diagnosis of giant-cell arteritis was suspected in 26 patients. In 15 of these patients, the diagnosis was confirmed by positive biopsy results. Twenty (including all 15 patients with positive biopsy results) of the 26 patients met the 1990 American College of Rheumatology criteria for giant-cell arteritis (1). In the remaining 6 patients who had negative biopsy results and who did not meet the American College of Rheumatology criteria, a final diagnosis other than giant-cell arteritis or polymyalgia rheumatica was made after diagnostic procedures were completed and the follow-up period had ended. In these patients, the diagnoses were fever of unknown origin (2 patients), cancer (2 patients), polyarteritis nodosa (1 patient), and sepsis (1 patient). Table 1. Characteristics of the 86 Patients Enrolled in the Study Seventy-seven patients satisfied the criteria defined by Healey (7) for the diagnosis of polymyalgia rheumatica; however, 2 of these patients developed rheumatoid arthritis (8) during the follow-up period. Therefore, 75 patients had a final diagnosis of polymyalgia rheumatica, and 12 (16%) had associated histologic evidence of giant-cell arteritis. The median follow-up period for all patients was 13 months (range, 4 to 25 months), and the median length of biopsy specimens was 2.3 cm (range, 1.8 to 3.2 cm). Ultrasonographic Evidence of a Halo Of the 15 patients with histologic evidence of giant-cell arteritis, 6 (40%) had a hypoechoic halo around the lumen of the temporal arteries (Figures 1 and 2). The halo was also present in 15 of the 71 patients (21%) who had negative biopsy results. All 6 patients with positive biopsy results had a halo 1 mm in thickness or greater (median, 2 mm [range, 1 to 3 mm]). Therefore, assuming that the halo was significant only if it was at least 1 mm in thickness, the halo could be considered significant in only 5 of the 71 patients (7%) with negative biopsy results. One of these 5 patients satisfied the American College of Rheumatology criteria for giant-cell arteritis. This patient had temporal artery abnormalities; headache (with no history of headache); and fever, anorexia, and weight loss. The length of the biopsy specimen in this patient was 1.8 cm (the shortest biopsy specimen obtained), and biopsy results were probably false negative. A second biopsy was not performed. The other 4 patients did not present with any clinical manifestations of giant-cell arteritis. No evidence of a halo was found in occipital arteries, and biopsies of these arteries were not performed. Of the 20 patients who met the American College of Rheumatology criteria, 7 (35%) had a halo at least 1 mm in thickness. A halo was present in 14 of the 66 patients (21%) who did not meet the criteria of the American College of Rheumatology, and 4 of these 66 (6%) had a halo at least 1 mm in thickness. Ultrasonography was repeated 1 month after initiation of corticosteroid therapy in all 6 patients with histologic evidence of giant-cell arteritis and evidence of a halo. The halo had disappeared in all 6 patients. The rate of agreement between the two ultrasonographers was 100% for the initial diagnosis of halo (86 of the 86 tests). Figure 2. Color Doppler scans in giant-cell arteritis panel A black crosses panel B black crosses panel C black crosses panel D Temporal Artery Abnormalities Of the 15 patients with bi