Abstract

Lumbar spinal stenosis is a common cause of back pain that can also give rise to pain in the buttock, thigh or leg, particularly when walking. Several possible treatments are available, of which surgery appears to be best at restoring function and reducing pain. Surgical outcome is not ideal, and a sizeable proportion of patients do not regain good function. No accepted evidence-based approach to postoperative care is known—a fact thathas prompted this review.

To determine whether active rehabilitation programmes following primary surgery for lumbar spinal stenosis have an impact on functional outcomes and whether such programmes are superior to ‘usual postoperative care’.

We searched the following databases from their first issues to March 2013: CENTRAL (The Cochrane Library, most recent issue), the Cochrane Back Review Group Trials Register, MEDLINE, EMBASE, CINAHL and PEDro.

We considered randomised controlled trials (RCTs) that compared the effectiveness of active rehabilitation versus usual care in adults (> 18 years of age) with confirmed lumbar spinal stenosis who had undergone spinal decompressive surgery (with or without fusion) for the first time.

Two review authors independently extracted data from the included trials by using a predeveloped form. We contacted authors of original trials to request additional unpublished data as required. We recorded baseline characteristics of participants, interventions, comparisons, follow-up and outcome measures to enable assessment of clinical homogeneity. Clinical relevance was independently assessed by using the five questions recommended by the Cochrane Back Review Group (CBRG), and risk of bias within studies was determined by using CBRG criteria.

We pooled individual study results in a meta-analysis when appropriate. For continuous outcomes, we calculated the mean difference (MD) when the same measurement scales were used in all studies and the standardised mean difference (SMD) when different measurement scales were used. Whenreported means and standard deviations of the outcomes showed that outcome data were skewed, we log-transformed data for all studies in the comparison and performed a meta-analysis on the log-scale. Results of analyses performed on the log-scale were converted back to the original scale. We used a fixed-effect inverse variance model to measure treatment effect when no substantial evidence of statistical heterogeneity was found. When we detected substantial statistical heterogeneity, we used a random-effects inverse variance model.

The primary outcome measure was functional status as measured by a back-specific functional scale. Secondary outcomes included measures of leg pain, low back pain and global improvement/general health. We considered statistical significance and clinical relevance of outcomes. We used the GRADE approach to assess the overall quality of evidence for each outcome on the basis of five criteria, for which evidence was ranked from high to very low quality, depending on the number of criteria met.

Our searches yielded 1,726 results, and a total of three studies (N = 373 participants) were included in the review and meta-analysis. All studies were deemed to have low risk of bias; no study had unacceptably high dropout rates. Also, no unacceptably unbalanced dropout rates, unacceptably low adherence rates or non-adherence to the protocol or clearly significant unbalanced baseline differences were noted for the primary outcome.

Outcomes in the short term (within six months postoperative)

Evidence of moderate quality from three RCTs (N = 340) shows that active rehabilitation is more effective than usual care for functional status (log SMD -0.22, 95% confidence interval (CI) -0.44 to 0.00, corresponding to an average percentage improvement (reduction in standardised functional score) of 20%, 95% CI 0% to 36%) and for reported low back pain (log MD -0.18, 95% CI-0.35 to -0.02, corresponding to an average percentage improvement (reduction in VAS score) of 16%, 95% CI 2% to 30%). In contrast, evidence of low quality suggests that rehabilitation is no more effective than usual care for leg pain (log MD -0.17, 95% CI -0.52 to 0.19, corresponding to an average percentage improvement (reduction in VAS score) of 16%, 95% CI 21% worsening to 41% improvement). Low-quality evidence from two RCTs (N = 238) indicates that rehabilitation has no additional benefit on general health status as compared to usual care (MD 1.30, 95% CI -4.45 to 7.06).

Outcomes in the long term (at 12 months postoperative)

Evidence of moderate quality from three RCTs (N = 373) shows that rehabilitation is more effective than usual care for functional status (log SMD -0.26, 95% CI -0.46 to -0.05, corresponding to an average percentage improvement (reduction in standardised functional score) of 23%, 95% CI 5% to 37%), for reported low back pain (log MD -0.20, 95% CI -0.36 to -0.05, corresponding to an average percentage improvement (reduction in VAS score) of 18%, 95% CI 5% to 30%]. Evidence of moderate quality (N = 373) and for leg pain (log MD -0.24, 95% CI -0.47 to -0.01, corresponding to an average percentage improvement (reduction in VAS score) of 21%, 95% CI 1% to 37%). In contrast, evidence of low quality from two studies (N = 273) suggests that rehabilitation is no more effective than usual care with respect to improvement in general health (MD -0.48, 95% CI -6.41 to 5.4).

None of the included papers reported any relevant adverse events.

Evidence suggests that active rehabilitation is more effective than usual care in improving both short- and long-term (back-related) functional status. Similar findings were noted for secondary outcomes, including short-term improvement in low back pain and long-term improvement in both low back pain and leg pain, although limited impact was observed in relation to improvements in general health status. The clinical relevance of these effects is medium to small. Our evaluation is limited by the small number of relevant studies identified, and further research is required.