The authors concluded that there was evidence to suggest therapeutic aquatic exercise had a potentially beneficial effect for patients who experienced chronic low-back pain and pregnancy-related low-back pain. The reliability of the authors' cautious conclusion is uncertain due to potential for language bias, poor-quality studies and a lack of reporting of review methods.

To evaluate the effectiveness of therapeutic aquatic exercise for the treatment of low back pain.

PEDro, CINAHL, PUBMED, Cochrane Central Register of Controlled Trials (CENTRAL) and SPORTDiscus databases were searched to July 2007 for articles published in English. Search terms were reported. Reference lists of retrieved articles were scanned for additional articles.

Randomised controlled trials (RCTs), clinical non-randomised trials and quasi-randomised controlled trials that evaluated all types of therapeutic aquatic exercise for adults (over 18 years) who experienced low-back pain compared to a control group were eligible for inclusion. Studies of post-surgery patients, spa therapy and balneotherapy (non-active) were excluded. The primary outcomes of interest were: assessments of pain using Oswestry Disability Index, McGill Pain Questionnaire and subjective assessment scales (such as visual analogue scale); and number of work days lost as a direct result of low-back pain.

Some of the interventions in included studies were aquatic exercise, deep-water walking, progressive aquatic programmes, aquatic fitness and other active aquatic therapy. Comparison groups included normal prenatal advice, land exercise, static traction techniques, body weight relief rehabilitation and stretching and no intervention. Duration of treatment varied between groups (details reported in the review). Age of participants ranged from 18 to 74 years. Participants included pregnant women with lower back pain and pelvic pain. Symptoms of other participants included non-specific lower back pain, disc degeneration, leg pain, acute lower back pain and back and leg pain. Duration of symptoms ranged from less than 90 days to more than 12 months (where reported).

The authors did not state how many reviewers performed the selection. Studies accepted for inclusion were further assessed by two reviewers independently. Disagreements were resolved through recourse to a third reviewer or referral to a senior professor or university sports faculty member.

Validity was assessed using the PEDro scale to evaluate randomisation, blinding and follow-up. Dichotomous responses to 10 criteria within the PEDro scale were added together to give a total score of between 0 and 10. Studies that scored 6 or more points were considered to be of high methodological quality; those that scored less than 6 points were considered to be of low methodological quality. Studies were further assessed using the SIGN 50 assessment. Studies were classified into three categories: low bias where all or most criteria were fulfilled; moderate bias where some criteria were fulfilled; and high bias where few or no criteria were fulfilled.

The authors did not state how many reviewers assessed validity.

Data were extracted from individual studies on differences in pain scores between baseline and follow-up and between intervention and control groups. Extracted data were used to calculate standardised mean difference (MD) and 95% confidence intervals (CIs).

The authors did not state how many reviewers extracted data.

Due to heterogeneity between studies and small sample size, the authors used a narrative synthesis to combine the studies. Additional data were reported in tables.

Seven trials (n=1,007) were included in the review. Methodological quality was considered low (one trial scored 6 points, one scored 5 points, two scored 4 points and three scored 2 points on the PEDro scale). Three trials reported adequate randomisation techniques; two reported quasi-randomisation techniques; the other two did not report the method of randomisation. No trials reported blinding of patients, one trial reported blinding of therapist and three trials reported blinding of outcome assessors. No trials conducted intention-to-treat analysis. Sample sizes ranged from 30 to 390.

In two trials of pregnant women (n=719), the aquatic exercise group reported a reduction in pain (p=0.03 and p=0.04) and a reduction in the number of sick days related to low back pain (p=0.09 and p=0.03) compared to control groups. The remaining five trials (n=288) reported no statistically significant differences in effect between therapeutic aquatic exercise groups and control groups for pain scores. No studies reported negative effects on low-back pain due to therapeutic aquatic exercise.

There was evidence to suggest that therapeutic aquatic exercise had a potentially beneficial effect for patients who experienced chronic low-back pain and pregnancy-related low-back pain.

The review question was clear with explicit inclusion criteria. Several relevant databases were searched. The authors included only studies published in English, which may have resulted in loss of some relevant data. Validity was assessed using published criteria and the results of the assessment were reported. The potential for reviewer error and bias in the review process could not be assessed as the methods used to select studies for inclusion, extract data and assess validity were unclear. A narrative synthesis was appropriate given differences between studies in terms of interventions, participants and outcomes. In general, sample sizes were small (particularly when studies of pregnant women were excluded) and studies were of poor methodological quality. The reliability of the authors' cautious conclusion is uncertain due to the potential for language bias, poor-quality studies and a lack of reporting of review methods.

Practice: The authors did not state any implications for practice.

Research: The authors stated that further methodologically robust research in a clinical setting was needed to substantiate use of therapeutic aquatic exercise for treatment of chronic low-back pain.

ERASMUS IP Grant Agreement no. 2006-2151/001-001 SO2-21CIEU; Pr. no. 27945-IC-1-2005-1-BE-ERASMUS-IPUC-6; LIKES-Research Centre for Health and Sports Science, Jyvaskyla, Finland.

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.