Stimulant Use During Pregnancy for Preexisting ADHD
A pregnant patient with well-controlled ADHD faces limited safety data on her medication against a real, already-observed functional decline off it — a genuine tradeoff, not a simple stop-versus-continue default.
C.M., a 29-year-old woman, found out she was pregnant on a Tuesday and had her hand on the bottle of lisdexamfetamine before she'd finished reading the second test strip, not to take a dose but to decide whether to flush it. She is a structural engineer six weeks into a pregnancy she and her husband planned carefully, and has managed ADHD with a stimulant since college with what she describes as the only real success she's had staying employed in a field that punishes the kind of missed details her unmedicated brain reliably produces. She stopped the medication outright the day she found out, on instinct, and has spent the two weeks since running calculations at work with a level of anxious double-checking that has itself become its own new problem.
She is here because two weeks off her medication has made unmistakably clear what treating her ADHD was actually protecting against: a missed load calculation nearly went out to a contractor last week before a colleague caught it in review, the kind of error that in her field carries real professional and safety consequences, not just an inconvenience. She wants an honest accounting of what's actually known about stimulant exposure in early pregnancy, not reassurance and not alarm, because right now she is choosing between two real risks and has no way to weigh them without better information than "ask your doctor," which is what every online source has told her so far.
She is also candid about a detail she says she almost left out: the near-miss calculation wasn't a one-off. She has caught herself re-checking simple arithmetic three and four times this week, not because the numbers are hard but because she no longer trusts her own first pass the way she used to, and the anxious over-checking is beginning to slow her down almost as much as the original inattention once did.
It is, she says, the first time she has understood her medication as something that was doing more than helping her focus — it was letting her trust her own judgment enough to move at a normal pace, and that trust is what two weeks without it has actually cost her.
Which stimulant the malformation data actually implicate
I want to be precise about what the data actually say, because "limited" doesn't mean "clean," and because the signal in this literature is drug-specific rather than class-wide. Huybrechts and colleagues, in JAMA Psychiatry in 2018, pooled US Medicaid and Nordic registry data and found a small increase in cardiac malformations after first-trimester methylphenidate exposure — adjusted relative risk 1.28, confidence interval crossing 1, so genuinely modest and imprecise. For amphetamines, across more than five thousand first-trimester exposures, they found nothing: adjusted relative risk 0.96 for cardiac malformations. That is the drug she is already taking. What does apply to her is a separate finding — Cohen and colleagues, in Obstetrics & Gynecology the same year, found stimulants as a class carried roughly a 1.3-fold increased risk of preeclampsia, though those authors concluded plainly that the absolute increase was small enough that women with significant ADHD should not be counseled to stop treatment on that basis.
I'd weigh that signal against what two weeks of discontinuation has already shown us: a real near-miss error in a job where errors carry safety consequences, plus new anxiety that didn't exist before she stopped. Untreated ADHD in pregnancy isn't a null-risk baseline we're protecting the fetus by choosing — it has its own documented association with worse prenatal engagement and higher maternal stress, which carry their own, separately real risks to a pregnancy.
You're right that "the data are limited" undersells what's actually been measured — but I'd push back on where that lands us. You've just described a literature in which the agent she is already on is the one with the reassuring number, and I think that deserves to be said as plainly as the preeclampsia finding. My worry is that "there is a signal in this space" gets heard by a frightened patient as "your drug is the problem," when the specific study you're citing says close to the opposite about her specific drug.
Then the lever here isn't which stimulant — it's that the reflex to do something would actively make this worse. The intuitive move in a first-trimester ADHD consult is to switch to methylphenidate, because it reads as the older, better-characterized, more conservative agent. On the malformation question specifically, Huybrechts points the other way: that switch would move her from the drug with a null result onto the only one carrying a positive cardiac signal, and would do it during the exact exposure window the signal was measured in. Resume lisdexamfetamine at the lowest dose that restores her function. Cohen's preeclampsia finding is the real thing to act on, and it's a monitoring question, not a drug-choice one — blood-pressure surveillance through the pregnancy, not a substitution she has no evidentiary reason to make.
Agreed: lisdexamfetamine resumed at the lowest dose that restores her function, no substitution made, maternal-fetal medicine co-management continuing through pregnancy with blood-pressure surveillance for preeclampsia, and a detailed fetal cardiac ultrasound at the standard anatomy-scan window — offered less because her exposure warrants it than because she asked for it.
Not agreed: how firmly the amphetamine null result should be presented to her. The maternal-fetal medicine specialist wanted the confidence intervals said out loud, on the grounds that a woman who has spent two weeks interrogating her own judgment deserves the imprecision rather than a reassurance she might later feel was managed. The psychiatrist thought that framing risked handing an already-anxious patient a number to re-check nightly, the way she has been re-checking her arithmetic. Neither treated the other's read as wrong, and they left it to whoever sees her at the next visit.