Metformin in Pregnancy and Lactation: What the Evidence Actually Shows

How Metformin Works: Mechanism Relevant to Pregnancy

Metformin lowers blood glucose primarily by suppressing hepatic gluconeogenesis and increasing skeletal muscle glucose uptake through AMP-activated protein kinase (AMPK) activation. It does not stimulate pancreatic insulin secretion, which means it carries no intrinsic hypoglycemia risk when used alone.

This mechanism matters during pregnancy for a specific reason. Pregnancy is a state of progressive insulin resistance, driven by placental hormones like human placental lactogen and tumor necrosis factor-alpha 1. Metformin addresses that resistance directly. Unlike sulfonylureas, it does not push already-stressed beta cells to produce more insulin. Unlike exogenous insulin, it is an oral tablet taken twice daily with food.

The drug is cleared renally without hepatic metabolism. Pregnancy increases glomerular filtration rate by 40% to 65%, which accelerates metformin clearance and often requires dose adjustment upward during the second and third trimesters 2. This pharmacokinetic shift is something clinicians must account for when monitoring glucose targets in pregnant patients already on metformin.

Metformin for Gestational Diabetes: The MiG Trial

The largest randomized trial comparing metformin to insulin for gestational diabetes mellitus (GDM) is the Metformin in Gestational Diabetes (MiG) trial, published in the New England Journal of Medicine in 2008 (N=751). Women with GDM between 20 and 33 weeks of gestation were randomized to open-label metformin (up to 2 to 500 mg daily) or insulin.

The primary composite neonatal outcome (neonatal hypoglycemia, respiratory distress, need for phototherapy, birth trauma, 5-minute Apgar score <7, or prematurity) occurred in 32.0% of the metformin group vs. 32.2% of the insulin group 2. That difference was not statistically significant. Women randomized to metformin gained less weight (0.4 kg less, P=0.01) and reported higher treatment satisfaction. The preterm birth rate was slightly higher in the metformin arm (12.1% vs. 7.6%, P=0.04), a finding that has prompted ongoing investigation but has not been consistently replicated across subsequent trials.

A critical practical finding: 46.3% of women randomized to metformin required supplemental insulin to achieve glycemic targets. Metformin alone was insufficient for nearly half the cohort.

"Women preferred metformin to insulin, and there were no evident adverse effects associated with metformin use in pregnancy," the MiG investigators concluded 2.

The 2024 ADA Standards of Care note that metformin "may be used for the treatment of GDM" but acknowledge that it crosses the placenta and that long-term safety data on offspring remain limited 3. ACOG Practice Bulletin 190 takes a similar position, stating insulin is the preferred agent but metformin is a reasonable alternative when patients decline insulin or when access is limited 4.

Pre-existing Type 2 Diabetes and Metformin Continuation

Women with pre-existing type 2 diabetes who become pregnant present a different clinical scenario than those diagnosed with GDM at 24 to 28 weeks. Many are already taking metformin at conception. The question becomes whether to continue it or switch to insulin.

No large randomized trial has directly compared metformin continuation vs. insulin switch in early pregnancy for pre-existing T2D. Observational data from the CPRD (Clinical Practice Research Datalink) in the UK, analyzing over 7,400 pregnancies exposed to metformin, found no increased risk of major congenital anomalies compared to pregnancies exposed to insulin alone 5. The adjusted odds ratio for major malformations was 0.84 (95% CI 0.60 to 1.19).

The NICE guideline NG3 on diabetes in pregnancy recommends that women with type 2 diabetes who become pregnant may continue metformin (alone or with insulin) if it was achieving good glycemic control before conception 6. NICE explicitly states: "Metformin can be used as an adjunct or alternative to insulin in the preconception period and during pregnancy."

Most endocrinologists add insulin when fasting glucose exceeds 95 mg/dL or 1-hour postprandial values exceed 140 mg/dL despite maximum-dose metformin. The combination approach reduces total daily insulin requirements, which may limit gestational weight gain.

PCOS, Metformin, and Early Pregnancy

Metformin has been used for over two decades to restore ovulation in women with polycystic ovary syndrome (PCOS). Once conception occurs, a common question arises: should metformin continue through the first trimester?

The rationale for continuation rests on the hypothesis that metformin reduces miscarriage rates in PCOS pregnancies. A 2010 meta-analysis of nine studies (N=1,008) found that first-trimester metformin use in women with PCOS was associated with a significant reduction in early pregnancy loss (OR 0.37 to 95% CI 0.26 to 0.53) 7. These were mostly observational studies and small RCTs.

The PREGmet trial (Vanky et al., 2010, N=274) randomized women with PCOS to metformin 2 to 000 mg/day or placebo from the first trimester through delivery. Metformin did not reduce the composite outcome of preeclampsia, preterm birth, gestational diabetes, or miscarriage in this population 8. Late miscarriage and preterm birth were numerically lower in the metformin group, but the trial was underpowered for these individual endpoints.

Current practice varies. Many reproductive endocrinologists continue metformin through weeks 10 to 12, then discontinue it. Others extend use if the patient develops GDM. No guideline society has issued a strong recommendation in either direction for PCOS-specific first-trimester continuation.

Placental Transfer and Fetal Exposure

Metformin is a small, hydrophilic molecule (molecular weight 129 Da) with no protein binding. These properties make placental transfer predictable. It crosses the placenta via organic cation transporters (OCTs), specifically OCT3 and MATE1 expressed on the syncytiotrophoblast.

Cord blood studies in the MiG trial showed that fetal metformin concentrations ranged from undetectable to levels equal to or exceeding maternal concentrations. The median cord-to-maternal plasma ratio was approximately 0.50, but individual values ranged from 0.01 to 2.00 9. This wide variability means some fetuses are exposed to substantial drug levels while others receive very little.

What does that exposure do? The available evidence is reassuring for structural teratogenicity. The National Birth Defects Prevention Study, a population-based case-control study, found no association between periconceptional metformin exposure and major birth defects 10. Animal reproductive studies in rats and rabbits at doses up to 600 mg/kg/day (roughly 2 to 6 times the maximum recommended human dose on a body surface area basis) showed no evidence of fetal malformations 11.

The concern is not teratogenicity. The concern is metabolic programming. If metformin suppresses hepatic glucose production in the fetus, what are the downstream consequences for metabolic development? This question drives the offspring follow-up studies discussed below.

Long-Term Offspring Outcomes: MiG TOFU and Beyond

The MiG TOFU (Metformin in Gestational Diabetes: The Offspring Follow-Up) study assessed 323 children at age 2 whose mothers participated in the original MiG trial. Children in the metformin-exposed group had larger mid-upper arm circumferences and greater subscapular and biceps skinfold thicknesses, indicating more subcutaneous fat deposition 12. Total body fat percentage, measured by bioimpedance, did not differ between groups.

Lead investigator Janet Rowan noted that "the children exposed to metformin had a pattern of growth that may be more insulin sensitive" 12. Storing fat subcutaneously rather than viscerally is generally considered metabolically favorable. But two years is a short follow-up window for metabolic programming effects.

A subsequent follow-up of MiG TOFU children at ages 7 to 9 (N=99 metformin, N=110 insulin) found that metformin-exposed children were heavier and had higher BMI, waist circumference, and waist-to-height ratio 13. This finding raised concern, though the small sample size and high attrition rate (only 209 of the original 751 were assessed) limit the strength of conclusions.

A Finnish RCT follow-up (Tertti et al., 2015) found no differences in motor, social, or linguistic development at 18 months between metformin-exposed and insulin-exposed offspring 14. Neurodevelopmental outcomes appear reassuring at the timepoints studied so far. Large registry-based studies following these children into adolescence and adulthood remain necessary.

Metformin and Breastfeeding: Milk Transfer Data

Metformin enters breast milk, but in very low quantities. The most cited pharmacokinetic study (Gardiner et al., 2003) measured metformin in breast milk and plasma of seven women taking 500 mg to 1 to 500 mg daily. The mean milk-to-plasma ratio was 0.35 (range 0.13 to 0.71). The estimated infant dose was 0.28% to 1.08% of the weight-adjusted maternal dose 15.

A threshold of <10% of the weight-adjusted maternal dose is the standard pharmacologic benchmark for drug compatibility with breastfeeding. Metformin sits well below this cutoff.

LactMed, the NIH database of drugs and lactation, states that "maternal doses of metformin up to 2 to 550 mg daily produce low levels in milk and would not be expected to cause adverse effects in breastfed infants" 16. No cases of infant hypoglycemia attributable to metformin in breast milk have been reported in the published literature.

The ABM (Academy of Breastfeeding Medicine) does not list metformin among drugs requiring interruption of breastfeeding. Blood glucose monitoring in the nursing infant is not routinely recommended when the mother is on standard metformin doses, though some clinicians advise it for premature or low-birth-weight infants.

Situations Where Metformin Is Not Appropriate in Pregnancy

Not every pregnant patient with hyperglycemia is a candidate for metformin. Specific contraindications and cautions apply.

Renal impairment changes the calculation significantly. The FDA label recommends against metformin when eGFR falls below 30 mL/min/1.73 m² and advises caution between 30 and 45 11. Preeclampsia with acute kidney injury, HELLP syndrome, or any condition involving tissue hypoxia (increasing lactic acidosis risk) should prompt immediate discontinuation.

Type 1 diabetes in pregnancy requires insulin. Metformin does not replace basal-bolus insulin therapy in this population and should not be used as monotherapy for T1D under any circumstances.

Severe hyperglycemia at GDM diagnosis (fasting glucose consistently above 110 mg/dL) predicts a high probability of metformin failure, based on MiG trial subgroup analyses. These patients are better served with insulin from the outset rather than trialing metformin and risking weeks of suboptimal glycemic control during a critical period of fetal growth.

Metformin should be held 48 hours before planned cesarean delivery and not restarted until renal function is confirmed stable post-operatively, following the same perioperative protocol used for any surgical patient on the drug 11.

Current Guideline Summary: Where Each Society Stands

The ADA 2024 Standards of Care classify insulin as the first-line pharmacologic agent for both GDM and pre-existing T2D in pregnancy. Metformin and glyburide are listed as alternatives, with the ADA noting that "metformin, when used to treat GDM, may slightly increase the risk of prematurity" 3.

ACOG Practice Bulletin 190 (2018, reaffirmed 2023) states: "Although glyburide and metformin are reasonable alternatives to insulin, insulin should remain the first-line pharmacologic therapy for GDM because it does not cross the placenta to a measurable extent" 4.

The Endocrine Society Clinical Practice Guideline on diabetes and pregnancy (2013) recommends insulin as the preferred treatment, with metformin as an option "only after a thorough discussion of the known risks and the limited information on the long-term safety for offspring" 17.

NICE NG3 is the most permissive guideline, recommending metformin as a first-line option for GDM alongside dietary modification, with insulin added if targets are not met 6. This reflects a healthcare system where access and adherence considerations weigh heavily in guideline development.

The maximum studied dose in pregnancy trials is 2 to 500 mg daily. Standard dosing starts at 500 mg once or twice daily with food, titrated over 1 to 2 weeks to the effective dose that achieves fasting glucose <95 mg/dL and 1-hour postprandial glucose <140 mg/dL.