Background Giant Cell Arteritis (GCA) is a common vasculitis with significant risk of blindness. Although temporal artery biopsy (TAB) is currently performed for diagnosis, the sensitivity can be as low as 39%; ultrasound is emerging as a more effective and cheaper test. Vascular endothelial growth factor (VEGF) is present in inflamed arterial wall in GCA and elevated circulating levels have been reported in GCA compared to controls. Measurement of VEGF is less invasive than TAB, and cheaper than TAB or ultrasound.
Objectives 1) Analyse performance of VEGF as a diagnostic marker for GCA
2) Combine VEGF with ESR and CRP to assess performance
3) Assess local expression of VEGF in TAB
Methods We used samples from patients recruited to a large multi-centre study comparing temporal artery biopsy with ultrasound for diagnosis of suspected giant cell arteritis. We randomly selected 26 patients with TAB-positive GCA and 26 controls (negative TAB and final diagnosis not GCA). Most patients were receiving high doses of glucocorticoid therapy at recruitment. Serum VEGF concentration was measured at weeks 0, 2 and 26 by ELISA (Quantikine R&D systems). Week 0 values were plotted on a Receiver Operating Characteristic (ROC) curve to identify the most effective cut-off for an abnormal result. Sections of TAB were stained by immunohistochemistry for VEGF expression (Polyclonal Rabbit anti-human VEGF, AbCAM).
Results Patients were significantly older than controls (mean age 76 SD 7.3 vs 66 SD 11.3, p=0.003). but had similar gender (62% females each). The mean (95% CI) baseline VEGF concentration was 873 pg/ml (631–1110) in GCA vs 476 pg/ml [328–625] in controls (p=0.017). This difference between groups was not observed at any other timepoint. The area under the ROC curve for VEGF concentration at week 0 was 0.73. The optimal cut off was 713pg/ml, providing a sensitivity of 65% and a specificity of 88%. These values compare favourably with reported performance of other available tests such as ESR and CRP (sensitivity 86% and 84%; specificity 30% and 29% respectively), TAB (sensitivity 40%, specificity 100%), and ultrasound (sensitivity 54%, specificity 81%).
VEGF concentration was combined with ESR and CRP values to assess if this improved its performance. High ESR was defined as ≥22 in men, ≥29 in women, and high CRP was ≥8. These results are summarised in Figure 1. ESR+CRP LOW had a 94% Negative predictive value to exclude GCA. ESR/CRP High + VEGF High had 95% Positive predictive value to diagnose GCA. For 27% of patients a further diagnostic test is needed.
VEGF expression by immunohistochemistry in TAB sections from 3 patients with GCA was higher than 2 controls, implying that VEGF is upregulated in GCA and may signal locally to drive the vasculitic process. Furthermore, serum concentration appears to reflect local concentration in the temporal artery; therefore measuring serum VEGF concentration is a valid surrogate for local inflammation.
Conclusions In this study, serum VEGF was a useful test to support a diagnosis of GCA and potentially avoid the cost of TAB or ultrasound in over 60% of patients. Further studies will evaluate this in a larger cohort of suspected cases of GCA. Furthermore, VEGF could potentially be effective as a prognostic marker, or as a therapeutic target.
Disclosure of Interest None declared