Twenty-seven RCTs (44 articles) were included (n=440), 15 of which used a crossover design.
The studies were small (mean n=15) and none described adequate allocation concealment. The groups were similar at baseline in 20 studies, 23 used placebo control, 11 specified eligibility criteria, 23 blinded the participants, 20 blinded the clinicians, 24 reported point estimates and variability, and 18 used intention-to-treat analysis. Drop-out rates were low, with 98% of participants completing the study protocol.
Body composition: when the studies were pooled, lean body mass increased significantly more in the intervention groups than in controls (WMD 2.1 kg, 95% CI: 1.3, 2.9, p<0.01; 11 RCTs). There was no statistically significant difference between the groups in changes in weight (9 RCTs) or fat mass (10 RCTs).
Strength: muscle strength was measured after 42 or 84 days' treatment. When the studies were pooled, there was no statistically significant difference between the intervention group and controls in measures of quadriceps and biceps strength (2 RCTs, n=40), nor was any statistically significant difference found between the groups across seven other muscle groups (1 RCT).
Basal metabolism: daily basal metabolic rate was significantly higher in the intervention group than in controls (WMD 141 kilocalories/day, 95% CI: 69, 213, p<0.001; 7 RCTs). Resting respiratory exchange ratio/respiratory quotient and resting heart rate were significantly lower in the intervention group (respectively: WMD -0.02, 95% CI: -0.03, -0.01, p<0.001; 7 RCTs and WMD 3.8 beats per minute, 95% CI: 0.2, 7.4, p<0.05; 11 RCTs). The only significant statistical heterogeneity found was for the outcome of resting heart rate (p=0.01; I2= 55%): when only studies with a duration of more than 14 days were included, the heterogeneity was no longer evident (I2=22%).
Exercise capacity: 6 RCTs measured exercise capacity but were not pooled because of differences in the exercise interventions used. Three RCTs measured lactate levels after a single dose of growth hormone; these were not pooled because of clinical heterogeneity. Two individual RCTs reported a significantly higher lactate level in the intervention groups compared with controls (p<0.001). Exercising levels of fatty acids and glycerol were both significantly higher in the intervention groups in all relevant studies (p<0.05; 3 RCTs). Heart rate was significantly increased (p<0.02) in two of the 4 RCTs that measured this outcome, and maximum inspiratory pressure at rest was significantly increased (p<0.05) in the single relevant study. None of the relevant studies found a statistically significant difference between the groups for the outcomes of exercising respiratory rate quotient or respiratory quotient (4 RCTs) and VO2max (2 RCTs), bicycling speed, exercising energy expenditure, or power output (1 RCT each).
Adverse events: there was a higher rate of adverse events in the intervention group than in controls, especially soft tissue oedema (44% versus 1%; 8 RCTs) and fatigue (35% versus 0%; 4 RCTs).
Sensitivity analyses were performed. The results of 3 meta-analyses were sensitive to the removal of a single study: there was a statistically significant increase in weight in the intervention group when one of 9 RCTs was removed, and the outcomes of fat mass and resting heart rate became non significant when a single RCT was removed from each.
There was little indication of publication bias.