An Oral Option, a Placenta That Doesn't Screen Everything Out
The trial data on short-term safety are reassuring. What isn't yet answered is what a drug that crosses the placenta means twenty years from now for the child who never got a vote in this decision.
Rachel M., a 31-year-old woman at 27 weeks in her second pregnancy, teaches yoga classes three evenings a week and has kept most of her usual schedule through this pregnancy, modifying poses as needed rather than stopping. She was diagnosed with gestational diabetes at her routine screening five weeks ago, started on a structured diet and daily walking program, and has kept a careful glucose log since. Her fasting values have run consistently in the low 100s, above the 95 mg/dL target, and her one-hour postprandial readings average around 155 — elevated, but not dramatically so, and diet and exercise alone haven't brought her fully to target over five weeks of real effort.
She's asked directly whether an oral option exists rather than insulin injections, a question the team takes seriously rather than treating as simply a comfort preference. The MiG trial, a large randomized comparison of metformin against insulin in gestational diabetes, found broadly comparable glycemic control and short-term perinatal outcomes between the two. What that trial didn't fully resolve is what happens years later: metformin crosses the placenta in a way insulin does not, and longer-term follow-up of metformin-exposed pregnancies has shown some signal toward higher childhood BMI in exposed offspring — a finding that hasn't replicated with full consistency across every cohort studied, which is exactly what makes it a genuinely open question rather than a settled one in either direction.
A modest elevation, and a question that won't be answered until years from now
I'd lead with insulin. It doesn't cross the placenta, and the offspring metabolic-programming signal seen in some metformin follow-up cohorts, while inconsistent, is real enough that I'd rather not introduce it when maternal glycemic control is essentially comparable either way. This is an elective choice, not an emergency — that's exactly when I'd default to the more conservative option.
I'd weigh insulin's own certain maternal costs more heavily against an offspring signal that hasn't replicated consistently. The MiG trial, at real scale, found comparable short-term perinatal outcomes, and insulin means injections, real hypoglycemia risk, and a genuine burden on top of an already busy pregnancy — costs that are common and certain, against a longer-term risk that's modest and inconsistently observed.
I'm not dismissing the offspring question — it's genuinely unresolved, and I'd tell her that directly rather than downplay it. But "unresolved and inconsistent" isn't the same weight as "insulin's injections and hypoglycemia risk," which are certain, not theoretical.
I'd frame this around her specific numbers rather than a blanket preference. Her fasting and postprandial values are only modestly above target, not the kind of significant elevation where metformin's dose ceiling and GI tolerability limits would likely leave her under-treated — that's the scenario where I'd push harder toward insulin regardless of the offspring question.
For a patient closer to target, like her, metformin's simplicity is a reasonable first step, with a clear plan to add or switch to insulin if her values don't come down within a defined window. That keeps the offspring-risk values question live for her to weigh, without committing her to insulin on physiologic grounds she doesn't actually meet yet.
Metformin started with titration, a two-week reassessment scheduled, and an explicit plan to add insulin if targets aren't met by then. Rachel was given a direct, honest explanation of the offspring follow-up uncertainty rather than a reassurance either way, and said she preferred knowing the real state of the evidence to a simpler answer.
Not agreed: whether the two-week reassessment window is the right length, or whether it should be shorter given how close she already is to needing correction. The maternal-fetal medicine specialist would have preferred one week, given her continued preference for insulin's more predictable dose-response; the endocrinologist held to two weeks as enough time to judge a real trend. Left as the working plan, not fully resolved.