The outcome of laminectomy for the relief of symptoms resulting from degenerative lumbar stenosis is not well established. Eighty-eight consecutive patients who had had a laminectomy for degenerative lumbar stenosis between 1983 and 1986 were studied. Eight of the patients had had a concomitant arthrodesis. The follow-up evaluation included a review of charts and standardized questionnaires that were completed by the patients. One year postoperatively, five patients (6 per cent) had had a second operation and five still had severe pain. By the time of the latest follow-up, in 1989, fifteen (17 per cent) of the original eighty-eight patients had had a repeat operation because of instability or stenosis; twenty-one (30 per cent) of the seventy patients who were evaluated by questionnaire in 1989 had severe pain. The factors found to be associated with a poor long-term outcome, defined as severe pain or the need for a repeat operation, or both, included co-existing illnesses (such as osteoarthrosis, cardiac disease, rheumatoid arthritis, or chronic pulmonary disease) (p = 0.004), the duration of follow-up (p = 0.01), and an initial laminectomy involving a single interspace (p = 0.04). We concluded that the long-term outcome of decompressive laminectomy is less favorable than has been previously reported, and that co-morbidity and a single-interspace laminectomy are risk factors for a poor outcome.