This review concluded that preoperative whole-body exercise programmes (prehabilitation) could shorten hospital stay and provide postoperative physical benefits, but study quality was limited and there was significant risk of bias in the included studies. This conclusion was appropriately limited, and is likely to be reliable.

To assess the effects of preoperative exercise (prehabilitation) on perioperative and postoperative outcomes, for adults undergoing surgery.

Six databases, including MEDLINE and PEDro, were searched for articles from 1950 to August 2011. Search terms were reported in an online appendix. References of relevant studies were checked, and authors were contacted, to identify additional studies.

Prospective trials that assessed rehabilitation before surgery, which comprised whole-body exercise, for patients aged at least 18 years, were eligible for inclusion. There were no restrictions on the duration, type or intensity of the exercise, but the programmes had to specify these criteria.

Two-thirds of the included trials assessed orthopaedic patients, and the remainder assessed patients undergoing visceral surgery. Exercise was undertaken at a facility, at home, or both. Most trials assessed both aerobic and resistance or calisthenic exercise. The median trial duration was six weeks (range one to eight). Several trials incorporated elements such as dietary changes or patient education.

Two reviewers independently assessed the studies for inclusion; discrepancies were resolved by a third reviewer.

Two reviewers independently assessed the trials using the Cochrane risk of bias tool, adapted to award up to a total of 6 points; disagreements were resolved by a third reviewer.

Two reviewers independently extracted the outcome data, on an intention-to-treat basis. Effect sizes (Hedges' g) were calculated for continuous outcomes. A third reviewer resolved disagreements; authors were contacted to verify the data.

The effect sizes were pooled in a meta-analysis, weighted by inverse variance, provided that the outcome was reported consistently by at least three trials. Heterogeneity was evaluated using Cochran's Q and Ι². Publication bias was assessed in a funnel plot and using Begg's and Egger's tests.

Twenty-one studies, with 1,371 patients, were included in the review. Seventeen were randomised controlled trials (RCTs), one was a non-randomised controlled trial, and the others were case-control, uncontrolled or case studies. Sample sizes ranged from two to 246. The risk of bias was significant; one study had a low risk of bias on all criteria (score 6), and four scored 4 or more. Allocation concealment was adequate in six studies, and incomplete outcome data were addressed adequately in three studies.

The pooled effect size for length of hospital stay was -0.39 days (95% CI -0.76 to -0.029) based on nine studies. Heterogeneity was high (Ι²=82.85%). There was a suggestion of publication bias in the funnel plot.

The outcomes for preoperative physical function, postoperative quality of life, pain, functional task performance, fitness, satisfaction, and health care use were reported in a narrative, with mixed results. Full details were given.

Prehabilitation could shorten hospital stay and provide postoperative physical benefits, but study quality was limited and there was significant risk of bias in the included studies.

The review addressed a clear question, supported by reasonable inclusion criteria, except that the outcomes were not specified. The search was reasonably extensive, with no limitations. The authors used methods to reduce reviewer bias and error throughout the review. The quality assessment was reasonable, but the tool was designed for RCTs and not intended to be used to generate a summary score. There was some indication of publication bias in the funnel plot, but the utility of such plots for fewer than 10 studies is limited. The synthesis was appropriate.

The conclusions are reasonable and take account of the high risks of bias in the included studies.

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

Research: The authors stated that adequately powered RCTs, with consistent and clinically relevant outcome measures (suggestions in the paper), were needed.

No specific funding received.