Valproate's Teratogenicity in a Treatment-Refractory Pregnancy
A single patient, 9 weeks into an unplanned pregnancy, already exposed to valproate for weeks before the pregnancy was known. It is the only agent that has ever controlled her illness — and the most teratogenic mood stabilizer available.
L.F. is a 36-year-old ICU nurse with a nine-year history of bipolar I disorder that has proven genuinely difficult to treat. Lithium caused a tremor severe enough to interfere with her job and persistent nausea that never resolved with dose adjustment; quetiapine caused sedation she couldn't safely work night shifts through; lamotrigine, tried for eight months at an adequate dose, still let through a breakthrough manic episode that cost her a hospitalization. Valproate is the only agent that has kept her fully stable, for the past four years without a single relapse. She presents today at 9 weeks gestation with an unplanned pregnancy, discovered only in the last week — she has been taking her usual valproate dose throughout the entire window since conception, without knowing she was pregnant.
Valproate is the most teratogenic of the major mood stabilizers, with well-documented risk for neural tube defects and, at higher doses, for broader neurodevelopmental effects including reduced IQ and elevated autism-spectrum risk. Some of that risk picture is already fixed by the time she is sitting in this appointment: the neural tube closes by roughly six weeks of gestation, well before she knew she was pregnant, so switching agents today cannot undo whatever exposure already occurred during that window. But the first trimester isn't over — craniofacial, cardiac, and limb development continue for weeks yet, and continuing valproate through the rest of it adds real, ongoing risk on top of whatever has already happened. Against that sits the fact that nothing else has kept her well, and an uncontrolled manic or depressive relapse during pregnancy carries its own serious risks to her and to the pregnancy — this is not a case where stopping the dangerous drug is simply the safer choice on every axis at once.
Weighing an already-fixed exposure against ongoing risk
I want to be precise about the timeline, because it changes what today's decision can actually accomplish: the neural tube defect risk from her exposure so far is already set, switching today does not reverse it. What switching does do is reduce her cumulative exposure for the anomalies still developing through the rest of the first trimester — that's a real benefit, just not the one she may be picturing.
High-dose folic acid, at least 4mg daily, should start today regardless of what happens with the valproate itself — it has some evidence for reducing neural tube defect risk even this late, though less than if started preconception.
I don't want the teratogenicity data to crowd out the fact that she is genuinely treatment-refractory — three prior agents failed her for real, distinct reasons, not from lack of trying. An uncontrolled relapse during this pregnancy is not a mild alternative risk; it carries its own documented harms to maternal and fetal outcomes, and I don't have a confident substitute to offer her today.
I'm not arguing to keep her at her current dose unchanged. I'm arguing against an abrupt full stop specifically, given how poorly she has tolerated changes to this regimen in the past.
The risk from valproate is dose-dependent, which gives a real middle path here: taper toward the lowest dose that still holds her, rather than treating this as a binary stay-or-stop decision. That reduces ongoing exposure meaningfully without repeating the abrupt discontinuation pattern that triggered her last hospitalization.
Divalproex sodium ER continued at her current dose for now, with a gradual, monitored taper plan toward the lowest dose that maintains her stability, rather than an abrupt stop. High-dose folic acid started today. A detailed anatomy ultrasound was scheduled for later in the second trimester specifically to screen for the anomalies still under active risk.
No clean resolution was reached, and the team was explicit about that with her directly: continuing any dose of valproate carries real ongoing risk, and reducing or stopping it carries real relapse risk that has already put her in the hospital once before. The taper plan reflects a judgment about which risk is more manageable to monitor closely, not a claim that either risk has been eliminated.