Diagnostic and prognostic utility of complete blood count-derived ratios in giant cell arteritis: a retrospective fast-track clinic cohort study

Early diagnosis and treatment of giant cell arteritis (GCA) can prevent complications like vision loss. Traditional inflammatory markers C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR) lack specificity. Complete blood count (CBC) components (neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), and monocyte-to-lymphocyte ratio (MLR)) are emerging biomarkers in autoimmune diseases, but their role in GCA remains unclear. To evaluate the clinical utility of CBC components in GCA diagnosis and prognosis. This retrospective study analyzed patients from the University of Washington’s GCA fast-track clinic (2017–2024). Biomarkers were assessed using ROC curves, Kaplan-Meier survival analysis, and Cox proportional hazards models. Predictive models were developed using LASSO, stepwise, and Firth’s penalized logistic regression, with data split into training and test sets. A total of 250 patients were included (119 GCA, 131 non-GCA). All CBC biomarkers were significantly associated with ESR and CRP, but correlation coefficients remained low, with the strongest correlation observed between monocytes and ESR (R = 0.55). CBC components showed moderate predictive ability for GCA diagnosis, vascular ultrasound (vUS) and temporal artery biopsy positivity, and mortality (area under the curve (AUC) range 0.5–0.694). Stepwise models integrating both CBC and inflammatory markers provided marginal improvements in predictive performance, with most models including one of each or, in some cases, NLR alone. A total of 23 patients died, with 13 deaths (10.9%) in the GCA group during a median follow-up of 2.45 years (1.17–42.25). In survival analyses, ESR and CRP lost significance after time-stratified adjustments, while PLR (hazard ratio (HR): 1.004, p = 0.018), NLR (HR: 1.082, p = 0.040), and MLR (HR: 3.655, p = 0.044) remained associated with mortality. The best mortality prediction model (AUC = 0.914) identified coronary artery disease, age at vUS, and time since first Rheumatology visit as key predictors. Diagnostic models reached an AUC of 0.920 based on vascular ultrasound, clinical suspicion, and American College of Rheumatology/European League Against Rheumatism classification, outperforming models based solely on inflammatory traditional inflammatory biomarkers or CBC parameters. Routine CBC-derived ratios showed independent associations with mortality and performed comparably to ESR and CRP for GCA diagnosis. Their integration into clinical models may offer a simple, low-cost strategy to support diagnostic and prognostic assessment, including in patients with normal inflammatory markers.