Five studies were included with a total of 176 participants (71 people participated in water-based exercise training and 54 in land-based exercise training; 51 completed no exercise training). All studies compared supervised water-based exercise training versus land-based exercise training and/or no exercise training in people with COPD (with average forced expiratory volume in one second (FEV1) %predicted ranging from 39% to 62%). Sample sizes ranged from 11 to 53 participants. The exercise training programmes lasted from four to 12 weeks, and the mean age of participants ranged from 57 to 73 years. A moderate risk of bias was due to lack of reporting of randomisation, allocation and blinding procedures in some studies, as well as small sample sizes.
Compared with no exercise, water-based exercise training improved the six-minute walk distance (mean difference (MD) 62 metres; 95% confidence interval (CI) 44 to 80 metres; three studies; 99 participants; moderate quality evidence), the incremental shuttle walk distance (MD 50 metres; 95% CI 20 to 80 metres; one study; 30 participants; high quality evidence) and the endurance shuttle walk distance (MD 371 metres; 95% CI 121 to 621 metres; one study; 30 participants; high quality evidence). Quality of life was also improved after water-based exercise training compared with no exercise (standardised mean difference (SMD) -0.97, 95% CI -0.37 to -1.57; two studies; 49 participants; low quality evidence). Compared with land-based exercise training, water-based exercise training did not significantly change the six-minute walk distance (MD 11 metres; 95% CI -11 to 33 metres; three studies; 62 participants; moderate quality evidence) or the incremental shuttle walk distance (MD 9 metres; 95% CI -15 to 34 metres; two studies; 59 participants; low quality evidence). However, the endurance shuttle walk distance improved following water-based exercise training compared with land-based exercise training (MD 313 metres; 95% CI 232 to 394 metres; two studies; 59 participants; moderate quality evidence). No significant differences were found between water-based exercise training and land-based exercise training for quality of life, as measured by the St George's Respiratory Questionnaire or by three of four domains of the Chronic Respiratory Disease Questionnaire (CRDQ); however, the fatigue domain of the CRDQ showed a statistically significant difference in favour of water-based exercise (MD -3.00; 95% CI -5.26 to -0.74; one study; 30 participants). Only one study reported long-term outcomes after water-based exercise training for quality of life and body composition, and no significant change was observed between baseline results and six-month follow-up results. One minor adverse event was reported for water-based exercise training (based on reporting from two studies; 20 participants). Impact of disease severity could not be examined because data were insufficient.