Twenty studies (n=2,233) were included in the review and 16 were included in the meta-analysis. Six studies used a randomised crossover design and the others used a parallel design. Nine studies were double-blinded, five studies were single-blinded and six studies were open label or did not report blinding. Three crossover design studies did not report a washout period. Sample sizes ranged from eight to 304.Study duration ranged from 10 days to six months.
PEG showed a significant increase in mean defaecations per week in comparison with placebo (SMD 1.98, 95% CI 1.16 to 2.81; 10 studies). There was evidence of significant statistical heterogeneity (I2=82.1%). Limiting the analysis to only high-quality studies removed the heterogeneity (SMD 2.34, 95% CI 1.76 to 2.91, I2=14.9%; seven studies).
PEG showed a significant increase in mean defaecations per week in comparison with lactulose (SMD 1.01, 95% CI 0.41 to 1.62, I2=53.5%; seven studies). The level of statistical heterogeneity was removed when only the four high-quality studies were considered (SMD 1.65, 95% CI 0.93 to 2.38, I2=0%). Individual studies showed significantly improved weekly defaecation rates with PEG in comparison with ispaghula (two studies) and tegaserod (one study).
Adverse events were mainly limited to gastrointestinal symptoms; it was difficult to determine whether these were treatment-related or due to the underlying constipation. Tolerability of PEG was generally good and similar to the other active agents.