The choice of anaesthesia in routine knee arthroscopy varies considerably. Concerns about local anaesthesia include the fear that it will take longer to perform surgery and that the anaesthesia will be inadequate, leading to patient discomfort. In this study, data from all patients (n = 6519) who had undergone a knee arthroscopy at St Göran Hospital Artro Clinic, in Stockholm, Sweden, during a 3.5 year period, between January 1993 and July 1996, were reviewed. Of these 6519 primary arthroscopies, 4101 were performed under local anaesthesia and 2418 under general anaesthesia. The purpose of the study was first to identify those arthroscopies that could not be successfully performed because the local anaesthesia was inadequate, and second, to investigate if arthroscopy under local anesthesia was associated with an increased number of rearthroscopies compared to general anaesthesia. The total number of rearthroscopies, performed within 180 days from the primary arthroscopy, was 214. Of these 214 rearthroscopies, 146 were due to a new indication for surgery and 30 were due to persisting clinical symptoms (true rearthroscopies). The remaining 38 rearthroscopies were due to an incomplete examination (because of patient discomfort) in a primary procedure where local anaesthesia was used. Of the 30 true rearthroscopies, 19 originated from the 4101 primary arthroscopies performed under local anaesthesia (0.46%) and 11 originated from the 2418 primary arthroscopies performed under general anaesthesia (0.45%). It is concluded that 0.9% of the primary arthroscopies performed under local anaesthesia could not be performed safely due to patient discomfort. There was no difference in the frequency of rearthroscopy between the arthroscopies performed under local anaesthesia compared to those performed under general anaesthesia.