Meta-analysis of the Placebo and Nocebo effects associated with Placebo treatment in randomized trials of lipid-lowering therapies

Abstract

Abstract Background Randomized controlled trials (RCTs) of lipid-lowering therapy (LLT) in which the control groups received placebo without background LLT offer unique insights into the placebo and nocebo effects of lipid-lowering RCTs. Methods and results Embase and Medline were searched for hyperlipidaemia RCTs with placebo-controlled arms. Placebo arms with background LLT were excluded. A single arm meta-analysis of proportions was used to estimate major adverse cardiovascular events (MACE) and adverse events (AE). A meta-analysis of means was used to estimate the pooled mean differences of total cholesterol (TC), low-density lipoprotein (LDL), high-density lipoproteins (HDL) and triglycerides (TG). A total of 40 RCTs and 37 668 placebo-treated participants were included. The pooled mean changes for TC, LDL, HDL, and TG were −0.019 mmol/L, −0.028 mmol/L, 0.013 mmol/L, and 0.062 mmol/L respectively among placebo-treated participants, indicating a modest placebo effect. The pooled average nocebo effect among placebo-treated participants was 42.62% for all AEs and 3.38% for musculoskeletal-related AEs, 11.36% for gastrointestinal-related AEs, and 6.62% for headaches. Placebo-treated participants in secondary prevention RCTs had a far higher incidence of these nocebo effects than primary prevention RCTs: any AEs (OR 6.76, 95% CI: 5.56–8.24, P < 0.001), and gastrointestinal-related AE (OR 1.23, 95% CI: 1.00–1.51, P = 0.049). No differences in nocebo effects were found between the placebo arms of statin and non-statin trials. Conclusion Our meta-analysis of placebo-treated participants in RCTs with no background LLT indicate a modest placebo effect but prominent nocebo effect of musculoskeletal, headache, and gastrointestinal symptoms that was greatest among secondary prevention RCTs. These findings may inform the design of future LLT RCTs.