Here is the current state of evidence and guidelines on metformin vs. insulin for GDM, as of mid-2025:
Metformin vs. Insulin in GDM: Latest Guidelines & Evidence
Major Guideline Positions (2024-2025)
| Organization | Recommendation |
|---|
| ADA (updated Dec 2024) | Metformin and glyburide should NOT be used as first-line agents; both cross the placenta and may not achieve glycemic goals. Insulin is preferred. |
| Endocrine Society / European Society of Endocrinology (July 2025) | For T2DM on insulin in pregnancy, suggests against routinely adding metformin. Insulin remains the anchor therapy. |
| ACOG (2018, still referenced) | Insulin is the preferred pharmacological agent. Metformin is a second-line option when insulin is not feasible or is declined by the patient. Glyburide is no longer recommended. |
| NICE (UK) | Recommends metformin as first-line pharmacological therapy if lifestyle modification fails, with insulin added as needed - the notable outlier among major guidelines. |
| WHO | Endorses insulin as first-line, but allows metformin where resources or patient acceptability limit insulin use. |
| DIPSI/FIGO (2023) | Both insulin and metformin are supported, especially in resource-limited settings. |
The bottom line from the
2025 ADA vs. Endocrine Society side-by-side comparison: US and international endocrinology bodies consistently place insulin first, while public health-oriented guidelines (NICE, WHO) are more permissive of metformin given real-world barriers to insulin use.
What the Recent Evidence Shows
Short-term Maternal & Neonatal Outcomes: Largely Equivalent
A 2024 systematic review of RCTs (
Berti et al., PMID 39669300) concluded:
- Insulin alone is not superior to metformin ± insulin on key outcomes.
- No significant differences in macrosomia, neonatal hypoglycemia, or cesarean delivery rates.
- However, the certainty of evidence is very low across all assessed outcomes, and more RCTs are needed.
A 2024 meta-analysis (
Wu et al., PMID 38124287) similarly found comparable maternal and neonatal outcomes between the two agents.
The Critical Caveat: ~50% Metformin Treatment Failure
Up to 50% of women started on metformin require supplementary insulin to achieve glycemic targets. This "add-on" design is common in practice - metformin reduces total insulin requirements but rarely works as monotherapy throughout pregnancy.
Long-term Offspring Effects: Where the Debate Lives
This is the most actively studied and contested area:
| Study | Finding |
|---|
| Dutta et al. 2024 (meta-analysis, PMID 38876183) | Children at 5-11 years born to metformin-treated mothers had similar BMI, body composition, obesity rates, lipids, and developmental outcomes vs. insulin-exposed offspring. No adverse long-term effects detected. |
| Rawat et al. 2024 (meta-analysis, PMID 39395229) | Children at 5-9 years in the metformin group had more fat-free mass and greater abdominal fat volume (subcutaneous + visceral by MRI), but lower fasting glucose and triglycerides. No major differences in other cardiometabolic outcomes. |
| MiTy Kids follow-up (longer-term MiTy trial data) | Some signal of increased adiposity in metformin-exposed offspring, raising concern about programming effects - though not consistently replicated. |
The concern is metformin's placental transfer - it reaches the fetal circulation and could affect fetal metabolism and adipose programming. Results are inconsistent enough that neither harm nor safety has been definitively established beyond age 9-11.
Practical Clinical Takeaway
When to use insulin:
- Preferred first-line per ADA (2024), ES (2025), ACOG
- Always required in T1DM; strongly preferred in T2DM and GDM when pharmacotherapy is needed
- When metformin fails (fasting glucose persistently >95 or postprandial >140/120 mg/dL)
- If contraindications to metformin exist (renal impairment, risk of preterm birth, fetal growth restriction)
When metformin is acceptable:
- Patient declines insulin or has significant barriers to insulin use
- Resource-limited settings (NICE, WHO, FIGO)
- As add-on to insulin to reduce insulin dose requirements
- Metformin starting dose: 500 mg/day for 1 week, then 500 mg twice daily, up to 2,500-3,000 mg/day maximum
Glyburide: No longer recommended by ADA or ACOG - associated with increased neonatal hypoglycemia, NICU admissions, and macrosomia compared to insulin.
GLP-1 receptor agonists: Currently not recommended in pregnancy by ADA or ACOG due to lack of long-term safety data in humans, despite increased exposure early in pregnancy in women using them for T2DM/obesity.
Summary
| Parameter | Insulin | Metformin |
|---|
| Guideline preference (ADA/ACOG/ES 2025) | First-line | Second-line / alternative |
| Crosses placenta | No | Yes |
| Short-term maternal/neonatal outcomes | Gold standard | Comparable in most RCTs |
| Treatment failure rate | N/A | ~50% need insulin added |
| Long-term offspring data | Reassuring | Mixed signals on adiposity; no definitive harm to age 11 |
| NICE/WHO position | First-line | First-line (NICE) / acceptable alternative (WHO) |
The core tension is between short-term equivalence and long-term uncertainty from placental transfer. Until larger RCTs with follow-up beyond age 11 are available, US/international endocrinology guidelines maintain insulin as the preferred agent, while allowing metformin as an informed, patient-centered alternative when insulin is not feasible.