Lithium's Perinatal Risk Profile: Teratogenicity in Pregnancy, Relapse After It
Two patients, one drug, two different points on the perinatal timeline. One is weighing lithium's revised, smaller teratogenicity risk during pregnancy; the other stopped lithium for that same fear and is now living through the postpartum relapse risk it left unprotected.
S.K. is a 31-year-old elementary school teacher, 14 weeks into her first pregnancy, which she and her husband planned carefully after two years of trying. She has bipolar I disorder, stable on lithium for five years with no manic or depressive episodes since starting it. She found out she was pregnant six weeks ago and has spent most of the time since agonizing over whether to stop lithium, after reading online that it causes a heart defect in the baby called Ebstein's anomaly — the figure she found cited was roughly 1 in 150, a number that has left her frightened enough that she skipped two doses last week on her own before calling the clinic.
The number she found is real, but it is also decades out of date. It comes from early lithium pregnancy registries, and more recent, larger studies place the actual absolute risk far lower — on the order of 1 in 1,000 to 1 in 2,000, an elevated risk relative to an unexposed pregnancy but a small absolute one, nowhere near what she has been carrying in her head for six weeks. Set against that revised, smaller number is a risk in the other direction that is not hypothetical for her specifically: she is only five years out from an illness that was, before lithium, severe enough to require hospitalization, and stopping her mood stabilizer in pregnancy carries a real risk of relapse during the pregnancy itself — with its own consequences for prenatal engagement, obstetric risk, and her ability to care for herself through the next six months.
At prenatal psychiatric consultation
The number she's been carrying around is roughly ten times higher than what current data actually support. That matters clinically, not just for her peace of mind — decisions made off an outdated risk figure aren't informed consent, they're decisions made off a wrong number.
I still want to be precise about what "far lower" means: the relative risk compared to an unexposed pregnancy is still real and still elevated. It's the absolute risk that's much smaller than she believes, not zero.
I want to name the risk on the other side just as concretely as the cardiac one: she was ill enough five years ago to need hospitalization, and stopping lithium now raises real risk of relapse during pregnancy itself, not just after it. Untreated bipolar illness in pregnancy is not a neutral, drug-free baseline — it carries its own documented risks to both of them.
She already skipped two doses on her own out of fear. That tells me the urgent first step is getting her an accurate number to act on, because the fear itself, based on outdated information, is already changing her behavior in a way that could hurt her.
If she continues lithium, a fetal echocardiogram around 16 to 18 weeks is the concrete next step — it directly screens for exactly the anomaly she is afraid of, which turns an abstract worry into an actual monitored plan rather than something she is just being asked to accept on faith.
Lithium resumed at her prior full dose after a detailed discussion of current, revised risk estimates, with a fetal echocardiogram scheduled around 16 to 18 weeks to screen directly for cardiac anomalies. Lithium levels to be checked each trimester given pregnancy-related changes in renal clearance.
R.M. is a 28-year-old woman, three weeks postpartum with her first child, managing largely alone while her husband is deployed overseas for another four months. She has bipolar I disorder and was stable on lithium before her pregnancy, but stopped it entirely at 6 weeks gestation after reading the same kind of frightening, outdated teratogenicity figures that brought S.K. into clinic today — she made that decision on her own, without the kind of risk discussion S.K. is having now, and carried the pregnancy to term with no mood-stabilizer coverage at all. The pregnancy itself was largely uneventful. The postpartum period has not been.
She is now in the single highest-risk window for relapse in the entire course of bipolar illness — the weeks immediately following delivery — with no mood-stabilizer coverage in her system, reduced sleep from a newborn, and a support system thinned by her husband's deployment. She has not had a full manic or depressive episode yet, but her mother, who is staying with her temporarily, describes three days this week of very little sleep, rapid talking, and grand plans to redo the nursery that don't match R.M.'s usual, more measured personality. The teratogenicity question that dominated her decision months ago is no longer live — the baby is here — but the relapse risk she was trying to avoid by stopping the drug never went away; it simply moved to a different, and in some ways more dangerous, part of the timeline. Her mother has been urging her to call the clinic for two days; R.M. kept putting it off, telling herself she was just adjusting to motherhood like everyone said she would be.
At an urgent postpartum visit
What her mother is describing sounds like early hypomania, and the postpartum window is exactly when relapse risk peaks for a patient off her mood stabilizer — this isn't a wait-and-see situation, restarting lithium now is the priority.
Agreed on restarting, and I want to flag the piece that's different from a routine restart: she's expressed interest in breastfeeding, and lithium does pass into breast milk at levels that need real monitoring, not a blanket "don't breastfeed" or a blanket "it's fine" — this needs to be its own explicit conversation with her, not an afterthought to restarting the drug.
That conversation shouldn't delay restarting lithium itself. The two decisions can move on separate tracks.
With her husband deployed and her mother staying only temporarily, the practical support for lab draws and close follow-up is thinner than it would be for most restarts. I'd build in a specific plan for who gets her to appointments over the next month, not assume it will happen on its own.
Lithium restarted with standard titration and a lithium level rechecked at one week. Her mother agreed to extend her stay by two additional weeks to help with appointments and support during titration. A separate, dedicated conversation about breastfeeding and lithium exposure was scheduled rather than folded into today's urgent restart.