Age, sex, disease severity, and disease duration difference in placebo response: implications from a meta-analysis of diabetes mellitus

Background: The placebo response in patients with diabetes mellitus is very common. A systematic evaluation needs to be updated with the current evidence about the placebo response in diabetes mellitus and the associated factors in clinical trials of anti-diabetic medicine. Methods: Literature research was conducted in Medline, Embase, the Cochrane Central Register of Controlled Trials, and ClinicalTrials.gov for studies published between the date of inception and June 2019. Randomized placebo-controlled trials conducted in type 1and type 2 diabetes mellitus (T1DM/T2DM) were included. Random-effects model and meta-regression analysis were accordingly used. This meta-analysis was registered in PROSPERO as CRD42014009373. Results: Significantly weight elevation (effect size (ES) = 0.33 kg, 95% CI, 0.03 to 0.61 kg) was observed in patients with placebo treatments in T1DM subgroup while significantly HbA1c reduction (ES = − 0.12%, 95% CI, − 0.16 to − 0.07%) and weight reduction (ES = − 0.40 kg, 95% CI, − 0.50 to − 0.29 kg) were observed in patients with placebo treatments in T2DM subgroup. Greater HbA1c reduction was observed in patients with injectable placebo treatments (ES = − 0.22%, 95% CI, − 0.32 to − 0.11%) versus oral types (ES = − 0.09%, 95% CI, − 0.14 to − 0.04%) in T2DM (P = 0.03). Older age (β = − 0.01, 95% CI, − 0.02 to − 0.01, P < 0.01) and longer diabetes duration (β = − 0.02, 95% CI, − 0.03 to − 0.21 × 10⁻², P = 0.03) was significantly associated with more HbA1c reduction by placebo in T1DM. However, younger age (β = 0.02, 95% CI, 0.01 to 0.03, P = 0.01), lower male percentage (β = 0.01, 95% CI, 0.22 × 10⁻², 0.01, P < 0.01), higher baseline BMI (β = − 0.02, 95% CI, − 0.04 to − 0.26 × 10⁻², P = 0.02), and higher baseline HbA1c (β = − 0.09, 95% CI, − 0.16 to − 0.01, P = 0.02) were significantly associated with more HbA1c reduction by placebo in T2DM. Shorter diabetes duration (β = 0.06, 95% CI, 0.06 to 0.10, P < 0.01) was significantly associated with more weight reduction by placebo in T2DM. However, the associations between baseline BMI, baseline HbA1c, and placebo response were insignificant after the adjusted analyses. Conclusion: The placebo response in diabetes mellitus was systematically outlined. Age, sex, disease severity (indirectly reflected by baseline BMI and baseline HbA1c), and disease duration were associated with placebo response in diabetes mellitus. The association between baseline BMI, baseline HbA1c, and placebo response may be the result of regression to the mean.