Background/Purpose: Giant cell arteritis (GCA) is the most common large vessel vasculitis. Failure to rapidly diagnose and treat patients with GCA can result in irreversible blindness. The American College of Rheumatology (ACR) developed classification criteria for GCA in 1990. However, these criteria are neither sensitive nor specific. Furthermore, these criteria do not provide a mechanism to quickly risk stratify referrals for GCA as the criteria rely in part on temporal artery biopsy (TAB) results, often not available at the time of referral. Triage, early treatment, and diagnosis would benefit from a GCA probability score. Laskou et al produced the GCA Probability Score (GCAPS) in 2019, which was subsequently used by Sebastian et al to successfully risk-stratify referrals to the Southend fast track clinic (FTC). However, this probability score requires further external validation. We report the results externally validating this probability score from our ultrasound-based FTC in the United States.

Methods: We included subjects referred to the FTC from November 2017 – December 2020. Only subjects with complete data to calculate the probability score were included. Subjects deemed very low risk were not offered an ultrasound for GCA and were not included in this study. All subjects evaluated by rheumatology received an ultrasound for GCA. Subjects received TAB at the discretion of the rheumatologist. A probability score was calculated as defined by Laskou et al, and compared to the low (< 9), intermediate (9-12), and high risk ( >12) probability score thresholds as defined by Sebastian et al. Diagnosis of GCA was also compared to the ACR 1990 classification of GCA criteria. A subject was determined to have GCA if this was deemed the most likely diagnosis 6 months after initial presentation.

Results: 166 subjects were referred to the FTC for suspected GCA during the specified time period. 121 had complete data to calculate the probability score (Table 1). 42 subjects were diagnosed with GCA. The median (interquartile range) probability score in the group diagnosed with GCA was 15 (5.75), in those not diagnosed with GCA was 11 (5), and the difference in probability score between the two groups was statistically significant (p=< 0.001). 3/27 subjects with low risk probability score, 12/42 subjects with intermediate risk probability score, and 27/52 subjects with high risk probability score were diagnosed with GCA. The low risk probability threshold was 92.9% sensitive and 30.4% specific (Table 2) for GCA. Of the low risk subjects diagnosed with GCA, one had a negative ultrasound for GCA and other diagnoses are being considered, one subject had negative TAB and ultrasound and was diagnosed based on clinical suspicion, and one subject had evidence of extra-cranial large vessel vasculitis on ultrasound. The high risk probability threshold was 64.3 sensitive and 68.4 specific. The 1990 ACR Classification of GCA Criteria were 52.4% sensitivity and 63.3% specific. Ultrasound was 62.5% sensitive and 97.3% specific.

Conclusion: In our cohort, the low risk probability score was sensitive, but not 100% sensitive, for GCA, and may serve as a reasonable screening tool. The high risk probability score threshold was not specific.

Clinical characteristics and outcomes of subjects evaluated by the fast track clinic for giant cell arteritis.

Sensitivity and specificity of diagnostic and screening modalities for GCA.

« Back to ACR Convergence 2021

ACR Meeting Abstracts - https://acrabstracts.org/abstract/validation-of-a-giant-cell-arteritis-probability-score/