This paper reports on the analysis of 332 otosclerosis revision operations. The results have been evaluated with reference to the type of the procedure at primary surgery, the alleged cause of failure and the applied technical solution. The need for revision surgery was found higher after primary total stapedectomy (3.4 per cent) than after partial stapedectomy (2.2 per cent) or stapedectomy (two per cent). The reason for revision varied according to the originally applied technique eg a migrated piston, a too short piston and a lateralized graft are almost exclusively found after total stapedectomies. The median hearing gain after revision of stapedectomy and partial stapedectomy was higher (20 dB and 18 dB respectively) than that after revision surgery for total stapedectomy (12 dB), but significantly lower than hearing gain after primary surgery (32 dB). Revisions yielded better results in the case of primary interventions with the use of a piston or pistonwire than in the case of primary interventions with a wire-type prosthesis. The risk for sensorineural loss (one per cent) was not higher than in primary surgery.