People with chronic obstructive pulmonary disease (COPD) often experience difficulty with performing upper limb exercise due to dyspnoea and arm fatigue. Consequently, upper limb exercise training is typically incorporated in pulmonary rehabilitation programmes to improve upper limb exercise capacity; however, the effects of this training on dyspnoea and health-related quality of life (HRQoL) remain unclear.
To determine the effects of upper limb training (endurance or resistance training, or both) on symptoms of dyspnoea and HRQoL in people with COPD.
We searched the Cochrane Airways Group Specialised Register of trials, ClinicalTrials.gov and the World Health Organization trials portal from inception to 28 September 2016 as well as checking all reference lists of primary studies and review articles.
We included randomised controlled trials (RCTs) in which upper limb exercise training of at least four weeks' duration was performed. Three comparisons were structured as: a) upper limb training only versus no training or sham intervention; b) combined upper limb training and lower limb training versus lower limb training alone; and c) upper limb training versus another type of upper limb training.
Two review authors independently selected trials for inclusion, extracted outcome data and assessed risk of bias. We contacted study authors to provide missing data. We determined the treatment effect from each study as the post-treatment scores. We were able to analyse data for all three planned comparisons. For the upper limb training only versus no training or sham intervention structure, the upper limb training was further classified as 'endurance training' or 'resistance training' to determine the impact of training modality.
Fifteen studies on 425 participants were included in the review, one of which was in abstract form only. Twelve studies were included in the meta-analysis across one or more of the three comparisons. The sample size of the included studies was small (12 to 43 participants) and overall study quality was moderate to low given the imprecision and risk of bias issues (i.e. missing information on sequence generation and allocation concealment as well as no blinding of outcome assessment and incomplete data).When upper limb training was compared to either no training or sham training, there was a small significant improvement in symptoms of dyspnoea with a mean difference (MD) of 0.37 points (95% confidence interval (CI) 0.02 to 0.72 points; data from four studies on 129 people). However, there was no significant improvement in dyspnoea when the studies of endurance training only (MD 0.41 points, 95% CI -0.13 to 0.95 points; data from two studies on 55 people) or resistance training only (MD 0.34 points, 95% CI -0.11 to 0.80 points; data from two studies on 74 people) were analysed. When upper limb training combined with lower limb training was compared to lower limb training alone, no significant difference in dyspnoea was shown (MD 0.36 points, 95% CI -0.04 to 0.76 points; data from three studies on 86 people). There were no studies which examined the effects on dyspnoea of upper limb training compared to another upper limb training intervention.There was no significant improvement in HRQoL when upper limb training was compared to either no training or sham training with a standardised mean difference (SMD) of 0.05 (95% CI -0.31 to 0.40; four studies on 126 people) or when upper limb training combined with lower limb training was compared to lower limb training alone (SMD 0.01, 95% CI -0.40 to 0.43; three studies on 95 people). Only one study, in which endurance upper limb training was compared to resistance upper limb training, reported on HRQoL and showed no between-group differences (St George's Respiratory Questionnaire MD 2.0 points, 95% CI -9 to 12; one study on 20 people).Positive findings were shown for the effects of upper limb training on the secondary outcome of unsupported endurance upper limb exercise capacity. When upper limb training was compared to either no training or sham training, there was a large significant improvement in unsupported endurance upper limb capacity (SMD 0.66, 95% CI 0.19 to 1.13; six studies on 142 people) which remained significant when the studies in this analysis of endurance training only were examined (SMD 0.99, 95% CI 0.32 to 1.66; four studies on 85 people) but not when the studies of resistance training only were examined (SMD 0.23, 95% CI -0.31 to 0.76; three studies on 57 people, P = 0.08 for test of subgroup differences). When upper limb training combined with lower limb training was compared to lower limb training alone, there was also a large significant improvement in unsupported endurance upper limb capacity (SMD 0.90, 95% CI 0.12 to 1.68; three studies on 87 people). A single study compared endurance upper limb training to resistance upper limb training with a significant improvement in the number of lifts performed in one minute favouring endurance upper limb training (MD 6.0 lifts, 95% CI 0.29 to 11.71 lifts; one study on 17 people).Available data were insufficient to examine the impact of disease severity on any outcome.
Evidence from this review indicates that some form of upper limb exercise training when compared to no upper limb training or a sham intervention improves dyspnoea but not HRQoL in people with COPD. The limited number of studies comparing different upper limb training interventions precludes conclusions being made about the optimal upper limb training programme for people with COPD, although endurance upper limb training using unsupported upper limb exercises does have a large effect on unsupported endurance upper limb capacity. Future RCTs require larger participant numbers to compare the differences between endurance upper limb training, resistance upper limb training, and combining endurance and resistance upper limb training on patient-relevant outcomes such as dyspnoea, HRQoL and arm activity levels.