Performance of the modified 2022 ACR/EULAR giant cell arteritis classification criteria without age restriction for discriminating from Takayasu arteritis

Abstract

Objective: To evaluate the ability to discriminate giant cell arteritis (GCA) from Takayasu arteritis (TAK) according to the modified 2022 American College of Rheumatology/European Alliance of Associations for Rheumatology (ACR/EULAR) GCA classification criteria.

Methods: Patients enrolled in the Japanese nationwide retrospective registry were evaluated using the criteria with partial modification; wall thickening of descending thoracic-abdominal aorta were mainly diagnosed by contrast-enhanced computed tomography (CT) or magnetic resonance imaging instead of evaluating with positron emission tomography (PET)-CT. The discriminability of the criteria was evaluated using C-statistic (> 0.7: good ability).

Results: Newly diagnosed patients with GCA (n = 139) and TAK (n = 129) were assessed, and 23.3% of TAK were aged 50 years or older at onset. The sensitivity of the modified 2022 ACR/EULAR GCA classification criteria with a score ≥ 6 was 82.0%, 68.5%, and 32.1% in all GCA, GCA with large-vessel involvement, and GCA without cranial arteritis, respectively. The specificity of the modified criteria was 96.1% for the 129 TAK as controls. Five patients with late-onset TAK met the modified criteria, and four had cranial signs and symptoms, two had bilateral axillary artery involvement, and four had descending thoracic-abdominal aorta involvement. The discriminability of the criteria was good (C-statistic: 0.986, 95% confidence interval [CI]: 0.976-0.996) and remained good after excluding age (C-statistic: 0.927, 95% CI: 0.894-0.961). The discriminability of a set of large-vessel lesions (bilateral axillary artery and descending thoracic-abdominal aorta) and inflammatory markers was markedly decreased with poor C-statistic value (C-statistic: 0.598, 95% CI: 0.530-0.667). Discriminability was improved after adding polymyalgia rheumatica (PMR) (C-statistic: 0.757, 95% CI: 0.700-0.813) or age (C-statistic: 0.913, 95%CI: 0.874-0.951) to the set of large-vessel lesions. In GCA patients with a score ≤ 5, 52% had bilateral subclavian and/or axillary artery involvement.

Conclusion: The modified 2022 ACR/EULAR GCA classification criteria well performed in classifying GCA and TAK without PET-CT in routine clinical practice. A set of items included in the modified GCA classification criteria had good discriminative ability for GCA and TAK, even when age was excluded. However, age restriction or PMR was required to distinguish GCA without cranial lesions from TAK.

Keywords: Bilateral axillary artery; Bilateral subclavian artery; Classification criteria; Descending thoracic-abdominal aorta; GCA; TAK.

Fig. 1

The discriminating ability for GCA and TAK of a combination of clinical parameters was evaluated by C-statistic. Gray highlighting indicates that the item is included in the model 1–7 (A). ESR of ≥ 50 mm/hr or CRP of ≥ 1.0 mg/dL was included in all models. Age at onset ≥ 50 years¹, positive temporal artery biopsy², cranial signs and symptoms (new headache, scalp tenderness, abnormal examination of the temporal artery, sudden visual abnormalities, and jaw claudication)³, PMR, descending thoracic-abdominal aorta involvement⁴, bilateral axillary artery involvement⁵, bilateral subclavian and/or axillary artery involvement⁶ were selected as clinical parameters. The ROC curve was drawn to calculate the area under the ROC curve as a C-statistic. The C statistic represents the ability to discriminate GCA from TAK in model 1 (B), model 2 (C), model 3 (D), model 4 (E), model 5 (F), model 6 (G), and model 7 (H). A value for the C-statistic greater than 0.7 indicates a reasonable and good ability to discriminate between patients with GCA and TAK. CI, confidence interval; ROC, receiver operating characteristic; TAK, Takayasu arteritis