The authors concluded that evidence suggested that manual therapy combined with exercise may be of benefit in the treatment of lumbar spinal stenosis, but further research was needed. The heterogeneity, poor quality and doubtful applicability of most included studies and methodological limitations in the review mean that these conclusions may not be reliable.

To determine the effectiveness of manual therapy for lumbar spinal stenosis (LSS).

MEDLINE via PubMed (from 1966), CINAHL (from 1983), PEDro and Cochrane Central register of Controlled Trials (CENTRAL) were searched in October/November 2007. Search terms were reported. Reference lists of eligible studies were checked and experts in the field were consulted. The search was limited to studies in English.

Randomised controlled trials (RCTs) and observational studies (controlled trials, cohort studies, case series and case studies) of manual therapy for patients diagnosed with LSS or who had LSS symptoms were eligible for inclusion. Typical LSS symptoms were listed in the review. Manual therapy was defined as thrust or non-thrust manipulation, massage or other manual treatment. Studies of multi-modal interventions that included manual therapy were eligible. Comparators could include surgical interventions for LSS. Studies were required to report outcomes using a pain intensity scale, region specific disability scale, functional performance scale or global outcome rating. Studies of non-specific or non-LSS types of lower back pain were excluded, as were studies in which there was no manual contact between clinician and subject.

Mean age of participant groups in the included studies ranged from 63 to 70 years (where stated). Most studies did not report details of types and regimens of interventions used. Most participants who received manual therapies (such as spinal manipulation) also received non-surgical therapies (such as walking, flexion exercises, analgesia). Comparison groups (where relevant) received other conservative therapies or surgical treatment (such as laminectomy with or without fusion, where reported). In studies that compared non-surgical versus surgical treatments, non-surgical groups included patients who received a range of interventions that did not always include manual therapy. Many different outcome measures were used, including a variety of rating scales for pain, physical function and quality of life. Duration of follow-up ranged from end of treatment to 10 years.

The studies were selected by a single reviewer.

Published criteria (van Tulder 1997) were used to rate 19 quality items, which included randomisation and allocation methods, blinding, group equivalence, outcomes measures, compliance, withdrawals, use of intention-to-treat analysis and statistical reporting. The maximum possible score was 19. Studies that scored over 50% overall or for 10 items related to internal validity were considered to be of acceptable quality. Two reviewers independently conducted the assessment.

Detailed descriptive data were extracted from the included studies and presented in evidence tables, with event rates and p values for most outcomes. Data were reported as either between-group differences or within-group changes from baseline.

Two reviewers independently extracted data. Disagreements were resolved by consensus.

Studies were combined by narrative synthesis. The text was supported by a table organised by treatment comparison that summarised the level of evidence in favour of the intervention across all relevant studies. Levels of evidence were graded from 1 to 4 (denoted strong, moderate, limited or no evidence) using a modified version of published criteria (van Tulder 2003). Where over 75% of studies had statistically significant findings in the same direction, the evidence was regarded as consistent.

Nine studies (11 articles) were included in the review (n=446, range one to 148): one RCT (n=58); four cohort studies (n=377); two case series (n=9); and two case studies (n=2). Only one study (the RCT) was considered to be of acceptable validity; common methodological weaknesses in studies were poor reporting, lack of controls, small sample size and non-equivalence of groups at baseline.

There was generally consistent evidence of moderate strength (level 2 evidence) that manual physical therapy as a component of conservative intervention was associated with significant benefit: either significant improvement compared with other conservative interventions (one RCT) or significant improvement from baseline (one cohort study, three case series, two case reports).

There was no evidence of a statistically significant difference in outcomes associated with spinal manipulation as a component of conservative intervention compared to surgical intervention. Although surgical intervention improved outcomes for up to two years (compared to multiple non-surgical interventions including spinal manipulation), the effects decreased over time and at 10 years favoured the conservative group (two studies).

The authors concluded that evidence suggested that manual therapy combined with exercise may be of benefit in the treatment of LSS, but further research was needed.

The objectives and inclusion criteria of the review were clear in most respects. However, in some included studies manual therapy was administered only to a proportion of the non-surgical group and a wide range of cointerventions differed within and across studies, which confounded any effects that related specifically to manual therapy. Relevant sources were searched for studies. The restriction to studies in English meant that the review was prone to language bias. It was unclear whether attempts were made to retrieve unpublished studies; it appeared that publication bias was not assessed. Steps were taken to minimise the risk of reviewer bias and error by having more than one reviewer independently assess study validity and extract data. But, studies were selected by a single reviewer, which created a potential for selection bias. Suitable criteria were used to evaluate study validity, but method used to determine the strength of the evidence (levels of evidence) was used inappropriately, as it was designed solely for evaluating randomised controlled evidence. The evidence designated level 2 derived largely from tiny poor-quality studies and did not appear to merit this rating. The text did not quantify the clinical and/or statistical size of effects and no measures of variability were reported. The primary studies involved such a wide variety of designs, clinical conditions, interventions and outcomes measures that the value of combining them was questionable. The heterogeneity, poor quality, variability and doubtful applicability of most of the included studies and methodological limitations in the review mean that the authors’ conclusions may not be reliable.

Practice: The authors stated that manual therapy could be considered a plausible treatment for LSS in combination with other types of physiotherapy.

Research: The authors stated that future research should evaluate potential benefits of manual therapy for patients with LSS and investigate which types of manual therapy or multimodal approaches were most beneficial.

Not stated.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.