Short-term Metformin Protects Against Glucocorticoid-Induced Toxicity in Healthy Individuals: A Randomized, Double-Blind, Placebo-Controlled Trial

Abstract

Objective: Glucocorticoids (GCs) are potent anti-inflammatory drugs, but strategies to prevent side effects are lacking. We investigated whether metformin could prevent GC-related toxicity and explored the underlying mechanisms.

Research design and methods: This single-center, randomized, placebo-controlled, double-blind, crossover trial compared metformin with placebo during high-dose GC treatment in 18 lean, healthy, male study participants. The trial was conducted at the University Hospital Basel, Basel, Switzerland. Participants received prednisone 30 mg/day in combination with metformin or placebo for two 7-day periods (1:1 randomization). The primary outcome, change in insulin sensitivity, was assessed using a two-sided paired t test. Before and after each study period, we conducted a mixed-meal tolerance test, blood metabolomics, and RNA sequencing of subcutaneous adipose tissue biopsy specimens.

Results: Metformin improved insulin sensitivity as assessed by the Matsuda index (n = 17; mean change -2.73 ± 3.55 SD for placebo, 2.21 ± 3.95 for metformin; mean difference of change -4.94 [95% CI, -7.24, -2.65]; P < 0.001). Metabolomic and transcriptomic analyses revealed that metformin altered fatty acid flux in the blood and downregulated genes involved in fatty acid synthesis in adipose tissue. Metformin reduced markers of protein breakdown and bone resorption. Furthermore, metformin downregulated genes responsible for AMPK inhibition and affected glucagon-like peptide 1 and bile acid metabolism.

Conclusions: Metformin prevents GC-induced insulin resistance and reduces markers of dyslipidemia, myopathy, and, possibly, bone resorption through AMPK-dependent and -independent pathways.

Graphical abstract

Figure 1

Main study end points comparing the placebo and metformin study groups after a 7-day GC challenge. Changes between baseline and after GC treatment are shown for mixed-meal stimulated glucose calculated as the AUC (A), stimulated insulin (B), Matsuda index (C), HOMA-IR index (D), stimulated GLP-1 (E), and fasting GDF-15 (F). Data are presented as mean changes (bars indicate minimum and maximum values). *Adjusted P ≤ 0.05, **adjusted P ≤ 0.01, ***adjusted P < 0.001.

Figure 2

Comparison of plasma metabolome and adipose tissue transcriptome between metformin and placebo groups after a 7-day GC challenge. A: Plasma metabolomic changes in 17 participants after treatment with metformin and placebo in the fasting state. B: Plasma metabolomic changes in the same participants in the postprandial state. In both panels, yellow points represent metabolites significantly reduced with metformin; purple points represent metabolites significantly increased with metformin. C: Differential gene expression in white adipose tissue from 10 participants under metformin and placebo treatment. Orange points represent genes downregulated with metformin, and green points represent upregulated genes. The dotted line represents the adjusted P-value cutoff of 0.05.