In this real-world retrospective cohort study, we found that the pre-test probability of a patient referred with suspected GCA (although not necessarily requiring a TAB) being found to have GCA is 13.9%. This highlights the need for ongoing specialist referrals with suspected GCA cases from general practitioners. Moreover, we found a significant difference in the absolute neutrophil count, in addition to WCC, platelet count, ESR and CRP, between GCA and suspected GCA cases.

Traditional serum inflammatory biomarkers such as ESR, CRP and platelets have established roles in the assessment of patients with suspected GCA [4, 5]. However, absolute neutrophil count has rarely been reported as a diagnostic marker [6, 12]. A recent study profiling peripheral blood leukocytes in biopsy-positive GCA patients demonstrated an elevated neutrophil presence throughout the course of the disease [6]. Other studies have suggested neutrophil-to-lymphocyte ratio (NLR) as a significant predictor of a positive biopsy in patients with suspected GCA [13]. Our study further highlights a statistically significant increase in the absolute neutrophil count within the GCA-confirmed cases (p = 0.02), which reinforces the involvement of neutrophils in the pathogenesis of GCA. However, the lack of difference in haemoglobin in this study is in contrast to other studies that have proposed that 20 to 50% of patients with GCA have normochromic normocytic anaemia [14]. Therefore, composite scores involving multiple laboratory parameters may improve the performance of such individual tests, although further research with large cohorts is required to have clinical validity.

While there is no consensus on the optimal length of a temporal artery biopsy specimen, a longer segment has previously been suggested to avoid a false negative GCA diagnosis due to the segmental nature of the disease. Specifically, some studies have proposed a 15 mm cut-off point [8]. “While the number of TAB-positive cases increased to 66.7% (from 33.3%) with >15 mm length segment, it was found to be not a statistically siginificant different in the mean length between TAB-positive and TAB-negative samples”. Nonetheless, these results are in keeping with recent data which highlight that a longer specimen length is not associated with a greater diagnostic yield [15]. The results depend more on adequate temporal artery histopathological sections and technical processing in the laboratory [16].

The current guidelines suggest prompt initiation of high-dose glucocorticoids when GCA is suspected combined with an urgent specialist referral for diagnostic confirmation [10]. Previous surveys have shown that up to one-third of primary care providers do not routinely initiate treatment routinely in suspected GCA cases [2]. Additionally, barriers to timely referral include variations in specialist referral pathways along with delays in accessing specialist care [2]. In the current study, two patients were on more than a month of steroid therapy prior to referral to our clinic. This may decrease the diagnostic yield of TAB as it is valuable within 4 weeks of starting high-dose corticosteroids and highest within the first 2 weeks [17].

We encourage the primary care providers to commence corticosteroids if tertiary care is not accessible immediately to prevent threatening complications of GCA. If accessible, they should be immediately referred to the nearest tertiary centre with specialist services to treat and investigate GCA. Symptoms that should raise high suspicion of GCA include sudden vision loss, morning stiffness in shoulder or neck, jaw claudication, new temporal headache, or scalp tenderness. Additionally, a rise in the inflammatory markers as described in the current study should also raise suspicion of the diagnosis.

There are limitations of the current study due to the retrospective nature of the design. For instance, some results did not reach significance due to our small sample size. Additionally, unequal sample sizes between the GCA-confirmed and non-GCA cases may have led to a decrease in statistical power of the results.