- Metformin reduces type 2 diabetes (T2D) development over 15 years in at-risk patients, regardless of diagnostic method.
- Patients with higher baseline fasting plasma glucose (FPG) or HbA1c and women with gestational diabetes mellitus (GDM) history benefit most.
Why this matters
- Decision to use metformin for T2D prevention requires informed benefit-risk balance.
- Analysis of 15-year follow-up data for Diabetes Prevention Program participants, randomly allocated to metformin 850 mg twice daily (n=1073) or placebo (n=1082), 1996-2001.
- Funding: NIH; additional multi-industry support.
- For T2D with metformin vs placebo, based on FPG or 2-hour glucose results:
- HR, 0.83 (95% CI, 0.73-0.93), and
- Risk difference (RD) −1.25 cases/100 person-years (−2.01 to −0.49).
- Using HbA1c for T2D diagnosis:
- HR, 0.64 (0.55-0.75), and
- RD, −1.67 cases/100 person-years (−2.24 to −1.10).
- Metformin effect greater at higher baseline FPG (Pinteraction=.02 for RD and P=.0004 for HR).
- Metformin vs placebo with GDM history:
- HR, 0.59 (Pinteraction=.03) vs nonsignificant 0.94 in parous women without, and
- RD, 4.57 vs 0.38 cases/100 person-years, respectively (Pinteraction=.01).
- HRs almost identical for baseline HbA1c:
- But significant difference by RD (−1.03 and −3.88 cases/100 person-years, respectively, Pinteraction=.001).
- Post hoc HbA1c analysis.
- Small numbers for some subgroups.