Twelve placebo controlled trials of orlistat (5,540 patients), six of sibutramine (1495) and four head-to-head comparisons (348) were eligible. The authors reported that the trials were generally of similar quality and that attrition was an important limitation. There was some attrition in all trials with rates often unbalanced and up to 54% of patients dropped out. Randomisation procedures and allocation concealment were unclear as they were poorly reported in the primary literature.
Compared with placebo, orlistat significantly reduced both systolic (WMD -1.9 mmHg, 95% CI -2.7 to -1.1mmHg) and diastolic blood pressure (WMD -1.5mmHg, 95% CI -2.2 to -0.8mmHg). Sibutramine illustrated no significant difference for systolic blood pressure (WMD 0.5mmHg, 95% CI -1.1 to 2.1mmHg) but significantly elevated diastolic blood pressure (WMD 1.7mmHg, 95% CI 0.7 to 2.6mmHg). Effect magnitudes were small.
Effect estimates from direct head-to-head comparisons were also small, non significant and less precise than placebo treatment comparisons. Interaction tests were not presented for diabetes subgroups, but 95% confidence intervals overlapped substantially suggesting no difference between subgroups. There was no substantial heterogeneity in any of the analyses except diastolic blood pressure for the orlistat-placebo comparison.