Purpose of the study: The principal objective of this prospective continuous observation study was to determine the incidence of perioperative and early postoperative complications secondary to surgical treatment of lumbar stenosis. Secondary objectives were to describe these complications and try to identify favouring factors.

Material and methods: All patients without major spinal deviation who underwent surgery for lumbar spine stenosis in our unit from January 1998 to January 2000 were included in the study. Minimal follow-up had to be six months. The series thus included 306 patients. Three categories of preoperative parameters were recorded: comorbid fractures, type of stenosis operated, type of surgery (simple release, release plus fusion, etc). In order to obtain an exhaustive data set, all complications were recorded on observation charts during hospitalisation and at follow-up visits at three, six and twelve months postoperatively. Complications were divided into four major categories: major complications, early or late infections, early or late mechanical disorders, neurological complications including meningeal disease and neurological disorders secondary to surgery. Data were explored with univariate analysis to determine the overall incidence of complications and the specific incidence for each category of complications and multivariate analysis with logistic regression to determine factor favouring development of complications.

Results: Overall incidence of complications secondary to surgery was 26.5%. Incidence of general, infectious, neurological and mechanical complications were 13, 4.5, 2.6, and 2% respectively. Incidence of complications considered to be serious and/or requiring reoperation was 12%. Factors influencing the development of complications were comorbidity, body mass index, duration of the operation, and reoperation.

Discussion and conclusion: The rate of complications reported in the literature have been very variable and have been established from retrospective reviews making comparison with our findings rather difficult. Our work pointed out the role of certain favouring factors which could be usefully examined in a larger series.