Article Text
PDF
Statistics from Altmetric.com
Background
Patients receiving haemodialysis have poor physical strength,1 muscle metabolism2 and quality of life3 than the general population. Exercise training may improve these parameters for people requiring maintenance haemodialysis.
Aim
This systematic review with meta-analysis aimed to examine the impact of exercise training on peak VO2, strength, cardiac function, quality of life, adherence, safety and energy intake of patients receiving haemodialysis.
Searches and inclusion criteria
Randomised clinical trials of exercise interventions conducted in patients receiving haemodialysis were identified by systematic searches of Medline, Embase, Cochrane Central Register of Controlled Trials and CINAHL up to 2010 with no language restriction. The actual search strategy is not included. Search terms may not have included all outcome descriptors (eg, the primary outcome, peak VO2) or all possible terms used to identify patients receiving haemodialysis and, in particular, did not appear to include the terminology of the KDOQI kidney classification4 (eg, ‘dialysis’, ‘end stage kidney/renal disease/failure’, ‘chronic kidney disease’) excluding at least one eligible study.5 Included studies reported the post-training mean and SD of at least one primary or secondary outcome measures. Studies were excluded if they did not report an exercise intervention and desired outcome measures. Duplicate or likely duplicate reports were excluded.
Interventions
All exercise interventions were included with data collected on the type (aerobic training, strength, resistance training or combinations of these), delivery (intradialytic, interdialytic outpatient or interdialytic home) and other exercise parameters.
Main outcome measures
Primary outcome measures following exercise training were postexercise change in peak VO2, percentage change in left ventricular ejection fraction, heart rate variability, quality of life as measured by the SF-36 or Beck Depression Score (but not by other instruments), exercise adherence and safety (number of adverse events). Energy intake and muscle strength were implied primary outcomes. Exercise energy expenditure was a secondary outcome. Authors were contacted for missing information but no further information was forthcoming.
Statistical methods
Meta-analysis using fixed effects models was used to pool the results of individual randomised controlled studies within each outcome for each intervention type. Some analyses were deemed unsuitable for this approach although the criteria for suitability are not explicit. The review analyses the differences in changes from baseline between the intervention arms. A Cochran Q test was used to assess study heterogeneity and was reported for one analysis (changes from baseline in peak VO2).
Results
The search strategy identified 15 eligible randomised studies including 565 participants. Eight studies assessed aerobic training, two assessed resistance training and five assessed a combination of the two. Co-interventions included nandrolone injections in one factorial trial. One study compared home-based and intradialytic exercise. Another study compared intradialytic exercise, home-based exercise and usual care although the treatment of the three arms for comparison purposes is not made clear.
Prior to intervention, the peak VO2 values of the participants of the haemodialysis procedure represented 70% of values for an average 51-year-old sedentary non-obese adult. Exercise increased peak VO2 values (eight studies, 365 participants) by a mean difference (MD) of 5.22 mlO2/kg-1/min-1 (95% CI 3.86 to 6.59, p<0.00001) compared with controls. Subgroup analyses demonstrated that longer periods of intervention were more effective than shorter (greater or less than 6 months, p=0.002; greater or less than 3 months, p=0.04) and exercise on non-dialysis days was more effective than intradialytic exercise (p=0.03). There was no statistical difference between combined aerobic and strength training compared with isolated aerobic training.
Assessment of muscle strength was only reported in two studies. One study found an improvement in muscle mass but not strength and the other found an improvement in muscle mass and strength measures.
SF-36 scores were reported for two studies but not pooled for undefined reasons. Physical quality of life improved in one study but not in the second. Physical quality of life was reported in an additional study but not included in the systematic review as the authors had not reported the constituent components. There was no impact on depression in one study utilising the Beck Depression Score. Another validated depression score (Self-rating Depression Scale) was used in one study but not included in the systematic review.
There was a trend to improved left ventricular ejection fraction (mean difference between groups 4.6%, 95% CI −0.02 to 9.3%, p=0.06) in the two studies (97 participants) reporting this measure. Heart rate availability improved with exercise (mean difference 16.3, 95% CI 8.3 to 24.4, p<0.0001) in two studies (119 participants). Daily energy intake assessed by dietary recall increased with exercise compared with control (MD 270 kcal day-1, 95% CI 122 to 418 kcal day-1, p<0.0004).
Seven studies evaluated exercise eligibility, adherence and safety. Of the 903 potential participants approached, 25% did not meet eligibility criteria while a further 27% refused participation. Of those randomised, a further 15% withdrew from the studies. No deaths were reported during exercise training for any of the studies.
Study quality
In these randomised studies, blinding (presumably of outcome assessment) is reported as unclear or present in all but one study. Withdrawal rates are clear for all except two of the studies. Studies had a median quality score of 2 (range 1–3).
Limitations/considerations
The major limitations of this review relate to the small number of included trials and participants. The evidence base is underpowered, particularly for subgroup analyses. Indeed, apart from VO2, the analyses of outcomes were based on two to three studies. The large number of participants excluded, refusing or not completing the studies, limit the generalisability of the results. There was no measure of adherence to the study interventions other than study withdrawal. The safety data are limited to deaths reported during exercise. The authors suggest that exercise is most likely to benefit those with a peak VO2 of <17.5 mlO2/kg-1/min-1. Unfortunately, the participants in the studies did not generally fall into this category but rather represented healthier patients receiving haemodialysis.
The quality-of-life review would have been strengthened by analyses of the Physical Component Score (PCS) of SF-36 data using either the reported PCS or calculating it from the reported components. Not all validated depression instruments were eligible for inclusion in the review for reasons that are not apparent. The use of the fixed effect models rather than a random effects model may not be appropriate given the variation in interventions used.6
Clinical implications
The systematic review concludes that exercise training improves peak VO2 in patients receiving haemodialysis. The paucity of the evidence base in this population with high mortality and morbidity rates reinforces the authors' call for further research.
Footnotes
-
Competing interests None.
-
Provenance and peer review Commissioned; internally peer reviewed.