Are antineutrophil cytoplasmic antibodies a marker predictive of relapse in Wegener's granulomatosis? A prospective study

Objectives. To investigate the predictive value of testing for antineutroph il cytoplasmic antibodies (ANCA) in 55 patients with systemic Wegener's gra nulomatosis (WG) included in a randomized, prospective trial comparing cort icosteroids and oral or pulse cyclophosphamide. Methods. All 55 patients received corticosteroids. A cyclophosphamide pulse of 0.7 g/m(2) was given at the time of diagnosis. After the first pulse, t he patients were assigned at random to receive either pulse or oral cycloph osphamide (2 mg/kg/day), independently of ANCA results. ANCA were sought us ing an immunofluorescence assay and an attempt was made to correlate them w ith relapse of WG. ANCA were monitored throughout the study. Results. At the time of diagnosis, ANCA were detected in 48 (87%) patients, with a cytoplasmic labelling pattern in 44 and a perinuclear pattern in fo ur. ANCA follow-up was available for 50 patients. ANCA disappeared in 34 pa tients and persisted in nine. For 79% of the patients, the clinical course improved with the disappearance of ANCA and deteriorated with their persist ence or increased titre. Among the patients who were initially ANCA-positiv e, 23 relapses occurred. Relapses were more frequent when ANCA remained pos itive or reappeared [13/19 ANCA-positive patients vs 3/29 ANCA-negative pat ients (P < 0.01)]. Nine relapses (39%) occurred in patients with persistent ANCA, and ANCA reappearance preceded relapse in eight (35%). The mean time between inclusion and relapse did not differ between the patients who beca me ANCA-negative and those who were persistently ANCA-positive (14.6 +/- 13 .2 vs 14.4 +/- 8.2 months). The mean time to ANCA disappearance was similar for the patients who relapsed and those who did not. Corticosteroids and p ulse or oral cyclophosphamide did not significantly modify the time to ANCA disappearance. Throughout the study, seven patients were ANCA-negative. Conclusion. Although ANCA positivity was associated with relapse, discordan ce between cytoplasmic ANCA and disease activity was not unusual. In the ab sence of clinical manifestations, ANCA titres alone can serve as a warning signal but not indicate whether to adjust or initiate treatment.