Sir,

We read with interest the report by Dr Levy and colleagues1 about a case of giant cell arteritis (GCA) with normal C-reactive protein (CRP). The studies reviewed by the authors indicate that this is an unusual finding. However, the authors’ inadvertent omission of two recent articles evaluating laboratory predictors of a positive temporal artery biopsy is potentially misleading.2, 3 The study by Parikh et al4 used a much lower cut-off for normal CRP of 5 mg/l, which may be the reason for the high sensitivity of CRP reported in their study. In the study by Walvick and Walvick,2 a CRP cut-off of 5 mg/l yielded a sensitivity of 94.9%, which means that 5.1% had a falsely normal CRP (less than 5 mg/l). In our study of 764 patients who underwent temporal artery biopsy, the sensitivity of CRP for GCA was 86.4%.3 In other words, 13.6% patients had a normal CRP (less than 8 mg/l in our laboratory), a much higher percentage than previously reported. Therefore, normal CRP does not exclude GCA in a patient with high clinical suspicion such as the case reported by Levy and colleagues.1 We would also suggest that the case reported by Levy and colleagues1 had an elevated erythrocyte sedimentation rate (ESR) and therefore would more appropriately be considered ‘CRP-negative’ rather than ‘inflammatory marker-negative disease’. True ‘inflammatory marker-negative disease’ (ie, both ESR and CRP normal) is rare but was observed in 4% (seven patients) in our study.3 In summary, the currently available biomarkers for diagnosis of GCA (ie, ESR and CRP) are imperfect given the less than desired sensitivity and poor specificity. Additionally, while studies evaluating these biomarkers provide us with aggregate results about a group of patients, they remain suboptimal when considering an individual patient presentation. Regardless of laboratory evaluation, in patients with high clinical suspicion for GCA we believe a temporal artery biopsy should be pursued as was done by Dr Levy and colleagues1 to establish the diagnosis.