Mental Health Medications During Pregnancy: What's Actually Safe

Why stopping abruptly usually isn't the safer option

Untreated depression and anxiety during pregnancy are associated with real risks of their own: poor engagement with prenatal care, inadequate weight gain or nutrition, increased substance use, preterm birth, and lower birth weight. Suicide and overdose are leading causes of pregnancy-associated death in the first year postpartum — this isn't a hypothetical risk being weighed against a hypothetical benefit, it's one real risk against another.

For bipolar disorder specifically, stopping a mood stabilizer abruptly carries a high risk of relapse during pregnancy and a sharply elevated risk of postpartum psychosis after delivery — a psychiatric emergency, not just a difficult mood episode. This is the single biggest reason bipolar medication decisions in pregnancy are usually about adjusting or switching rather than stopping.

None of this means every medication is automatically fine to continue without a conversation — it means the conversation should start from "what's the actual comparison here," not from an assumption that medication-free is the default safe choice.

Antidepressants (SSRIs and SNRIs) — generally reassuring

SSRIs (sertraline, escitalopram, fluoxetine, citalopram) and SNRIs (venlafaxine, duloxetine) are the most studied psychiatric medications in pregnancy, and the overall body of evidence is reassuring for major malformation risk — no consistent signal of a broad increase across this class. Sertraline in particular has an especially large safety dataset and is often a first choice when starting fresh during pregnancy for that reason, though continuing whichever SSRI already has you stable is usually preferable to switching just for the sake of switching.

The one specific exception worth naming: paroxetine has a modest but real signal for an increased risk of fetal cardiac malformations, most consistently seen with first-trimester use. It's generally not the first choice to start during pregnancy. If you're already taking it and stable, that's a real conversation about switching versus continuing — not an automatic reason to stop on your own.

A separate, unrelated issue: babies exposed to SSRIs/SNRIs late in pregnancy can show transient "neonatal adaptation" symptoms after birth — jitteriness, feeding difficulty, mild respiratory changes — usually resolving within one to two weeks without treatment. Hospitals plan for this by watching newborns a bit more closely after a known exposure; it's a monitoring issue, not a reason to stop the medication beforehand.

Benzodiazepines — situational, not automatically off the table

Occasional, as-needed use of a benzodiazepine (for acute anxiety or panic) is generally not a source of major alarm. Regular daily use is a more involved conversation, partly because of dependence and withdrawal considerations for you, and partly because of neonatal sedation or withdrawal symptoms if used close to delivery — something your care team can plan around with dose timing rather than by stopping outright.

Mood stabilizers — where the real, specific exceptions live

Valproate (Depakote, divalproex) is the clearest exception on this entire list: it's a well-established teratogen, with a meaningfully increased risk of neural tube defects and other major malformations, plus documented risks to a child's later cognitive development. It's generally avoided during pregnancy, and often avoided as a first-choice mood stabilizer in anyone who could become pregnant, specifically because unplanned pregnancies happen and the exposure window (the first few weeks, often before a pregnancy is even confirmed) is the highest-risk period.

Lithium carries a real but smaller-than-historically-believed increased risk of a specific heart defect (Ebstein's anomaly) with first-trimester exposure. Because the absolute risk is low and the relapse/postpartum-psychosis risk of stopping lithium in bipolar disorder is high, many people continue lithium through pregnancy with adjusted dosing and closer monitoring (including a specific fetal cardiac ultrasound) rather than stopping — this is genuinely an individualized, specialist-level decision, not a default in either direction.

Lamotrigine has one of the more reassuring safety profiles among mood stabilizers and is frequently a preferred option for mood stabilization during pregnancy for that reason, though its levels can shift substantially as pregnancy progresses and often need dose adjustment and monitoring.

Antipsychotics

Data on second-generation (atypical) antipsychotics in pregnancy has grown substantially and is generally reassuring for major malformation risk. The more common practical issue with this class in pregnancy is metabolic — an increased risk of gestational diabetes and excess weight gain with some agents — which is a monitoring conversation, not typically a reason to avoid the class altogether if it's what's keeping you stable.

How this actually gets decided

The right call depends on your specific medication, dose, diagnosis, history of relapse off medication, and how early in pregnancy you are — there's no single universal answer, which is exactly why a blanket "just stop" isn't good guidance and neither is a blanket "never change anything." If a pregnancy is planned, this is a conversation worth having beforehand, when there's time to adjust thoughtfully rather than react. If it isn't planned and you're already pregnant on medication, don't stop today and call tomorrow — reach out first.

Therapy alongside medication, or in some milder cases instead of it, remains a real option throughout pregnancy and is worth discussing regardless of what happens with medication.