Recently, it was reported that botulinum toxin type B complex (BoNT/B) (NeuroBloc(R), Elan Pharmaceuticals) can produce an adequate therapeutic response in patients with antibody induced failure of botulinum toxin type A complex (BoNT/A) therapy. We wanted to study whether this effect is transient or sustained. For this, 10 consecutive patients (6 males, 4 females, age 54.6 +/- 14.3 years, duration of illness 15.8 +/- 7.0 years) with complete BoNT/A therapy failure and BoNT/A antibody titres in excess of 10mU/ml in the mouse diaphragm assay (MDA) received BoNT/B in an initial dose of 12370 +/- 1804MU. After the first BoNT/B application the Toronto Western Spasmodic Torticollis Rating Scale (TWSTRS) improved from 20.1 +/- 3.0 to 11.9 +/- 3.4. In all patients systemic anticholinergic side effects occurred. Three patients had stable continuous responses to two, three and five subsequent BoNT/B applications. Six patients showed complete secondary therapy failure to the second or third subsequent BoNT/B applications. Side effects did no longer occur. In four of them the BoNT/B doses were doubled without producing any therapeutic benefit or any side effects. In five of them MDA testing was performed and revealed BoNT/B antibody titres in excess of 1mU/ml. One patient lost half of her initial BoNT/B responsiveness indicating partial secondary BoNT/B therapy failure. This partial therapy failure was seen on two consecutive application series and has not proceeded to complete therapy failure so far. BoNT/B seems to be only temporarily effective in the majority of patients with BoNT/A antibody induced therapy failure. Whether the formation of BoNT/B antibody points to a high antigenic potency of BoNT/B, to an increased immunoreactivity in BoNT/A antibody carriers or whether it is due to the large amount of protein applied in BoNT/B therapy needs to be studied.