The review concluded that there was moderate support for the use of exercise to improve HRQoL in stroke survivors. The authors' conclusion reflects the evidence presented but lack of reported review process and the possibility of publication bias mean that these results should be interpreted with caution.

To examine the effect of different doses and types of exercise on health-related quality of life (HRQoL) in stroke survivors.

MEDLINE, CINAHL, EMBASE and SPORTDiscus were searched up to March 2010 for articles published in English. Search terms were reported in a supplementary online table. Reference lists of relevant articles were manually searched for further articles.

Eligible randomised controlled trials (RCTs) included adult (over 19 years) stroke survivors and reported HRQoL where exercise was the primary intervention. HRQoL was defined as at least one physical and/or mental attribute associated with overall health status. Trials were excluded if the focus was therapy/rehabilitation such as body weight supported treadmill training, if other disability groups were included or if there was insufficient data on effect size estimation.

Most studies used a combination of aerobic, strength, balance and/or flexibility exercises. The mean duration of intervention was 10.4 weeks (range four to 12 weeks) and most were implemented three times a week with the average duration of session lasting 67.9 minutes (range 30 to 90 minutes). Five different HRQoL scales were used; the most frequently used scale was the Short Form-36, followed by the Stroke Impact Scale. Several studies monitored adverse events but none reported results. The mean age of participants was 66.9 years (range 61 to 72 years) and the mean time since stroke was 26.7 months (range 2.2 to 109 months).

The authors did not state how many reviewers performed the study selection.

Study quality was assessed with the PEDro scale which included one non-scored item for examining external validity and ten scored items for appraising internal validity and quality of statistical reporting.

The authors did not state how many reviewers performed the quality assessment.

Mean differences were extracted and the standardised mean difference was used to calculate effect size (ES). Hedges' g was used. For studies with more than one effect size, an average was calculated that represented the general effect of the intervention.

The authors did not state how many reviewers performed data extraction.

Summary effect sizes and associated 95% confidence intervals (CI) for the overall effect of exercise on HRQoL were calculated with a fixed-effect model; where statistical heterogeneity was found a random-effects model was used. The Q-statistic was used to assess heterogeneity and Ι² to measure degree of variation. Separate analyses were performed for the two subdomains of the HRQoL (physical and mental).

Subgroup analyses were conducted for time since stroke (less than six months versus at least six months) and intervention characteristics such as mode of exercise (aerobic, strength training or combined), weekly amount of exercise meeting USA Physical Activity Guidelines (less than 150 minutes versus over 150 minutes), intervention format (individual versus group) and intervention setting (clinic-based versus community-based).

Nine RCTs (426 participants, range 13 to 93) were included. Eight RCTs were considered to have been high quality (PEDro score of 5 or more) and one of fair quality (PEDro score of 4 or less).

A small to moderate significant effect on overall HRQoL was found in favour of exercise (ES 0.32, 95% CI 0.12 to 0.51; Ι²=0%; nine RCTs). No significant between group differences were found for overall HRQoL at follow-up (12 to 24 weeks post-intervention). A moderate significant effect was found for the subdomains HRQoL physical (ES 0.33, p=0.02) and HRQoL mental (ES 0.23, p=0.04) in favour of exercise. No significant between group differences were found at follow-up.

Subgroup analyses were reported; differences between subgroup comparisons did not reach statistical significance.

There was moderate support for the use of exercise to improve HRQoL in stroke survivors.

The review question was supported by clearly defined inclusion criteria. Several relevant databases were searched, but this search was restricted by language and only published peer reviewed papers were eligible for inclusion, raising the possibility of language and publication bias. The authors did not report whether steps were taken to minimise the likelihood of error and bias in the selection of studies, data extraction or validity assessment. Study quality was assessed with a validated scale although individual results were not reported.

Data were analysed in a meta-analysis and possible sources of heterogeneity assessed. The outcomes were based on largely self-reported measures which were subject to inherent biases.

The authors' conclusion reflected the evidence presented but lack of reported review process and the possibility of publication bias mean that these results should be interpreted with caution.

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

Research: The authors suggested that identification of effective strategies for sustaining the effects of exercise on HRQoL post-intervention were needed, particularly in community-based settings. The effects of exercise on HRQoL by time since stroke and setting required further study, as well as the precise doses of exercise tailored to specific HRQoL outcomes.

National Institute on Disability and Rehabilitation Research (USA).

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.