Aims: To clarify the effectiveness of surgical intervention (S) as compared to non-operative measures (NO) in the treatment of patients with moderate lumbar spinal stenosis. Patients and methods. Four university hospitals agreed upon classiþcation of the disease, inclusion and exclusion criteria, radiographic routines, surgical principles and follow up protocols. Moderate stenosis was deþned as a longstanding (> 6 months) low back disorder with pain radiating to the lower limbs and buttocks, aggravated by walking. Narrowing of the dural sac was sagittally < 10 mm or the cross-sectional area < 75 mm2. 94 patients were randomized in the two groups. Results. Sex-adjusted average score of pain in the lower limbs (VAS scale, 0–10) before treatment was 6.5, and 6.3 in the S and NO groups, respectively. Between group difference favoured S group in the follow-up: at six month 2,4 (95% CI 0,9 to 3,8), and at 12 month 1,4 (95% CI 0,1 to 2,8). Sex-adjusted Oswestry score before treatment was 33,0, and 35,8 in the S and NO groups, respectively. Again, between group difference favoured S group: at six months follow-up 9.8 (95% CI 3,2 to 16,4) and at 12 months 12,3 (95% CI 5,2 to 19,5). Walking ability did not differ between the two groups: before treatment 32% of the patients in S group and 30% in NO group walked less than 400, at six month follow-up 23% in both groups, and at one year 18%, and 20% in the S and NO groups, respectively. Conclusions: The results indicate a favorable early effect on perceived outcomes after operative treatment, whereas measured walking ability remained largely unaffected.