Could metformin benefit glucocorticoid-treated patients?

  • Pernicova I & al.
  • Lancet Diabetes Endocrinol
  • 25 Feb 2020

  • curated by Miriam Tucker
  • Clinical Essentials
Access to the full content of this site is available only to registered healthcare professionals. Access to the full content of this site is available only to registered healthcare professionals.

Takeaway

  • Metformin improved some metabolic parameters and clinical outcomes for glucocorticoid-treated patients with inflammatory disease.
  • Metformin did not affect visceral-to-subcutaneous fat area ratio.

Why this matters

  • Long-term glucocorticoids are prescribed for ≤3% of European adults.
  • Chronic glucocorticoid exposure can raise metabolic, infectious, and cardiovascular risks.

Study design

  • Randomized, double-blind, placebo-controlled, proof-of-concept, 12-week phase 2 trial of 53 adults without diabetes and with inflammatory disease treated with continuous prednisolone (≥20 mg/day for ≥4 weeks, ≥10 mg/day for subsequent ≥12 weeks, or cumulative dose-equivalent) randomly allocated to metformin (titrated up to 850 mg 3 times a day) vs placebo.
  • Funding: Barts Charity, Merck Serono.

Key results

  • No change in the primary outcome: visceral-to-subcutaneous fat area ratio over 12 weeks (0.11; P=.09).
  • Facial adiposity observed in 10% of metformin vs 52% of placebo groups (P=.007).
  • Dysglycemia present in 0% of metformin vs 33% of placebo groups (P=.009).
  • Metformin improved insulin resistance, β-cell function, liver function, fibrinolysis, carotid intima-media thickness, inflammatory parameters, and disease activity severity markers.
  • Metformin was associated with lower incidences of pneumonia (1 vs 7; P=.01), moderate-severe infections (2 vs 11; P=.001), and all-cause hospitalizations for adverse events (1 vs 9; P=.001).
  • More diarrhea with metformin (18 vs 8 events; P=.01).

Limitations

  • Small sample size.
  • Patient heterogeneity.
  • Relatively short treatment duration.
  • Possible selection bias.