“…BD patients were then split according to active and inactive disease where possible: we arbitrarily considered patients with vascular involvement [20,37,38] as having active disease and pooled them with active disease patients proper [20,22,23,25,26,40]: FMD was significantly lower with moderate heterogeneity (SMD ¼ À1.17, 95% CI: À 1.45 to À 0.89, p ¼ 0.0001, I 2 ¼ 37.5%) (Fig. 2B).…”