Identification of risk factors for permanent visual loss in patients with giant cell arteritis.

Objectives: permanent visual loss (PVL) is the most frequent ischemic complication of GCA. We aimed to evaluate whether clinical signs, symptoms, and blood test abnormalities at GCA diagnosis can predict PVL.

Patients and methods: retrospective, multicenter study of patients with biopsy-proven GCA. The whole cohort was randomly split into a derivation and a validation dataset. Multivariable logistic regression (MVLR) was used to develop a prediction model and a predictive score. The model's performance was determined through the area under the curve (AUC).

Results: 620 patients were included, 397 in the derivation cohort. PVL occurred in 20.3%. MVLR showed that amaurosis fugax (OR 5.86, 95%CI 3.41-10.07, p<0.001), jaw claudication (OR 2.48, 95%CI 1.51-4.07, p<0.001), and increasing age (OR 1.09, 95%CI 1.05-1.14, p<0.001) were independently associated to PVL. Fever was the only independent protective factor (OR 0.45, 95%CI 0.25-0.84, p=0.01). The optimum cut-off for age as a PVL predictor was 78 years (OR 2.4, 95%CI 1.62-3.59, p<0.001). No laboratory variables were independently associated with PVL. Our model showed an AUC 0.80 (95%CI 0.76-0.85) and good internal validity (AUC 0.82, 95%CI 0.76-0.87). A predictive risk score with a sensitivity of 64.9%, a specificity of 82.0%, and positive and negative predictive values of 46.4% and 90.01% is proposed.

Conclusions: The risk of developing PVL at GCA diagnosis may be estimated upon a detailed patient evaluation, including temporal arteries. Amaurosis fugax, jaw claudication, and older age can predict PVL. Fever is a protective factor. Inflammatory markers do not differentiate patients at risk of developing PVL.