CRD summary

The review concluded that exercise interventions reduced liver fat in people with non-alcoholic fatty liver disease, despite minimal or no weight loss. Due to the limited evidence of uncertain quality and variation between trials, the authors' conclusions should be considered with caution.

Authors' objectives

To evaluate the efficacy of aerobic exercise, progressive resistance training, or both, for the modulation of liver fat and alanine aminotransferase levels in adults with non-alcoholic fatty liver disease (steatohepatitis).

Searching

PubMed, CINAHL, AMED, Web of Science, Scopus, EMBASE, ProQuest 5000 and Science Direct were searched, without language restriction, for articles from inception to August 2011; search terms were reported. The reference lists of all retrieved papers were searched.

Study selection

Controlled trials that compared the effects of regular exercise (aerobic, progressive resistance, or both; three sessions or more) with a non-exercise control, were eligible for the review if they reported the change in liver fat, plasma alanine aminotransferase, or both, for adults (18 years or older) with non-alcoholic fatty liver disease or steatohepatitis. Liver fat had to be quantified by histological assessment or proton magnetic resonance spectroscopy or inferred by computed tomography or ultrasonography. Trials with a dietary control group were included if the diet was the same for both comparison groups. Observational studies and abstracts without adequate data were excluded.

In the included trials, the mean age of participants ranged from 36 to 69 years and the proportion of men ranged from 12% to 100%, where reported. Body mass index ranged from 27 to 44kg/m², where reported. Some trials only recruited obese or overweight patients or those with diabetes or metabolic syndrome. Trials either compared exercise versus control, or exercise plus diet versus control plus diet. Most trials had aerobic exercise interventions; some had progressive resistance training or load lifting exercise interventions. The frequency of exercise was generally three or five times per week, and the duration of the interventions ranged from two weeks to six months. The intensity of most of the exercise interventions was moderate to vigorous. No trials reported adverse events. In most trials, the change in liver fat was measured by proton magnetic resonance spectroscopy.

One reviewer performed the searches and the authors did not state how many reviewers selected trials for the review.

Assessment of study quality

Trials were assessed for quality using a modified Downs and Black checklist; the modified criteria included adequate description of control, whether sessions were supervised, and whether proton magnetic resonance spectroscopy was used to measure liver fat.

Two reviewers assessed quality, with disagreements resolved by a third reviewer or the criterion was considered to be negative.

Data extraction

The outcomes were extracted to calculate mean differences, with 95% confidence intervals. Where trials contained insufficient information, attempts were made to contact their authors for the missing data. Where the data in the included trials were only presented in graphs, estimates were made from these graphs. Means and standard deviations were calculated, using appropriate equations, if necessary.

Two reviewers independently extracted the data, with disagreements resolved by a third reviewer.

Methods of synthesis

Trial data were pooled and summary effect standardised mean differences were calculated, using a random-effects model. Subgroup analyses were undertaken for trials of only exercise interventions, and for those with dietary interventions for both the exercise and control groups.

Results of the review

Twelve controlled trials, with 439 patients (range 14 to 130), were included in the review. Eleven of the 12 trials were randomised. All of them specified their hypotheses, main outcomes, participant characteristics, interventions, main findings, variability estimates, and statistical tests, and all used accurate measures. Four trials did not provide an adequate description of the control or did not supervise the exercise sessions. Two trials attempted to blind researchers to the main outcome measures.

Ten trials assessed liver fat. There was no evidence of a significant overall effect of exercise on liver fat (SMD -0.05, 95% CI 0.27 to -0.38). In subgroup analyses, exercise alone was associated with a significant reduction in liver fat compared with control (SMD -0.37, 95% CI -0.06 to -0.69; six trials), but there was no evidence of a significant difference for exercise with diet versus control with diet. In four of the six trials of exercise alone, the change in body weight was less than 0.5kg with exercise.

Ten trials assessed alanine aminotransferase. There was no evidence of a significant overall effect of exercise on alanine aminotransferase levels overall (SMD -0.006, 95% CI 0.20 to -0.21), and in the two subgroup analyses with and without dietary interventions.

Authors' conclusions

Exercise interventions reduced liver fat in people with non-alcoholic fatty liver disease, despite minimal or no weight loss.

CRD commentary

The review addressed a clear research question, supported by appropriate inclusion criteria. A wide range of relevant sources was searched and no language restrictions were applied, minimising the chances of language and publication bias. Appropriate methods were used to assess trials for quality and to extract data, but only one reviewer selected studies, so reviewer error and bias cannot be ruled out. A basic tool was used to assess quality, making it difficult to evaluate the possible biases within the trials. In most of the included trials, the samples were very small.

The synthesis and choice of model and summary effect metric were appropriate, but variation between trials was not assessed. The authors acknowledged that there were differences in the exercise session duration, intensity, exercise dose and modality. Subgroup analyses were pre-specified to assess the effects of exercise alone versus control, and exercise plus diet versus control plus diet. The conclusion that exercise alone significantly reduced liver fat in one subgroup analysis might not be appropriate as interaction tests were not reported between subgroups; it was also based on limited data with few participants.

Due to the limited evidence of uncertain quality and variation between trials, the authors' conclusions should be considered with caution.

Implications of the review for practice and research

Practice: The authors stated that exercise levels should be assessed and exercise should be routinely prescribed for patients with non-alcoholic fatty liver disease.

Research: The authors stated that there was a need for randomised controlled trials to assess the relative importance of exercise dose. These trials should focus on exercise intensity, frequency and volume and investigate the relative effects of aerobic exercise, progressive resistance training or combined therapy. The effects on alanine aminotransferase should be investigated.

Funding

Support received from the Robert W. Storr Bequest and the National Health and Medical Research Council (NHMRC), Australia.

Bibliographic details

Keating SE, Hackett DA, George J, Johnson NA. Exercise and non-alcoholic fatty liver disease: a systematic review and meta-analysis. Journal of Hepatology 2012; 57(1): 157-166. [PubMed: 22414768]

Original Paper URL

http://www.journal-of-hepatology.eu/article/S0168-8278%2812%2900210-3/abstract

Indexing Status

Subject indexing assigned by NLM

MeSH

Exercise /physiology; Fatty Liver /physiopathology /therapy; Humans; Obesity /physiopathology /therapy; Resistance Training; Weight Reduction Programs

AccessionNumber

12012031926

Database entry date

15/11/2012

Record Status

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.