The aim of the study was to assess the effectiveness of surgical treatment for degenerative lumbar spinal stenosis (LSS) as compared with non-operative measures. Four university hospitals contributed, after agreement on study protocol, surgical rationale and non-operative procedures (For details, see Spine2007;32:1–8). Ninety-four patients were randomized into a surgical or nonoperative treatment group, 50 and 44 patients, respectively. Surgery comprised undercutting laminectomy of the stenotic segments, in 10 patients augmented with transpedicular instrumented fusion. The primary outcome was based on assessment of functional disability using the Oswestry Disability Index (ODI, scale 0–100). Intensity of leg and back pain (scales 0–10), as well as self-reported walking ability, were recorded at randomization and at follow-ups at 6, 12, 24 months and on average 6 years after the randomization. At the 2-year follow-up, back and leg pain scales and ODI had improved more in the surgical than the nonoperative group (p-values for global difference < 0,01). At the 6-year follow-up the mean difference in ODI in favor of surgery was 9.5 (95% confidence interval 0.9–18.1). However, the intensity of pains did not any-more differ between the two treatment groups at the 6-year follow-up. Walking ability did not differ between the treatment groups at any time point. Of the 44 patients in the nonoperative group, 4 had been subjected to surgery within two years after randomization because of persistent symptoms.

We conclude that surgical treatment improves functional ability in lumbar spinal stenosis. We emphasize that improvement also occurs after nonoperative measures. We recommend starting treatment with non-operative measures during a 2-year surveillance period, as during this period only 10 per cent of the patients will need surgical intervention.

Since neurological claudication is a major symptom in lumbar spinal stenosis (LSS), walking distance is commonly used as a measure of the severity and surgical outcome in LSS. The aim of this study was to compare self-reported and treadmill-measured walking distances in a trial, in which 94 patients with moderate LSS had been randomized into conservative and surgical treatment. Among the 44 patients in conservative treatment, the treadmill-measured walking distance was more reproducible after 6 months than the self-reported distance; the intraclass correlation coefficients were 0.75 and 0.41, respectively. Among all the patients at baseline, the agreement between self-reported and measured walking distance was satisfactory (intraclass correlation 0.57), although male patients overwhelmed their performance by 200 meters. Such a shift was not found in women. For walking distance categorized as < 400, 400–1249 and ≥ 1250 meters, there was a fair agreement between self-report and treadmill (weighed kappa 0.42). However, when the analysis was restricted to those whose walking distance was restricted to < 1250 meters, the corresponding agreement was poor (intraclass correlation 0.26). The self-reported walking distance was closely correlated with Oswestry index at baseline (r = 0.26), and changes in these outcomes from randomization to the follow-up of 6 months showed a strong correlation with each other (r = 0.37). We conclude that walking distance is a fundamental element of disability in LSS. Self-reported walking distance seems to be an appropriate clinical tool, but its limited precision in relation to treadmill-measured distance must be considered, when walking ability is severely restricted.