Clinical Psychopharmacology · Paper A
Teratogenicity in pregnancy
This concept matters because prescribing in pregnancy means weighing drug risk against illness risk, and TERATOGENICCapable of causing a birth defect if taken during pregnancy.Read more → (capable of harming a developing baby) drugs differ hugely in how much risk they carry and when. This concept is about which psychiatric drugs carry the greatest risk in pregnancy, and when in pregnancy that risk falls. A woman of childbearing age newly diagnosed with bipolar disorder shows the stakes: valproate might be an effective mood stabiliser, but its high teratogenic risk usually rules it out before another option is even considered. Because of this, valproate should be avoided in any woman of childbearing potential unless strict pregnancy-prevention measures are in place.
The risk of structural malformations is greatest in the FIRST TRIMESTER, when the baby's organs are forming. Sodium valproate carries the highest risk, neural tube defects and significant developmental harm, and must be avoided in women who could become pregnant unless strict precautions are in place. Carbamazepine also raises neural tube defect risk. Lithium in the first trimester is classically linked to Ebstein's anomaly, a heart-valve defect. A woman on lithium who discovers she is six weeks pregnant shows why timing matters: the safest approach is planning medication choice BEFORE conception, not reacting once pregnancy is already confirmed. Take a woman with bipolar disorder planning a pregnancy while still on valproate, e.g. switching her to a lower-risk mood stabiliser before she conceives avoids exposing the fetus during the critical first-trimester window. Clinically, pre-conception counselling for any woman of childbearing potential on a psychiatric drug is what turns this risk from unavoidable into manageable.
This concept matters because prescribing in pregnancy is never a simple 'safe or unsafe' question, it is always a balance between the risk a drug poses to the developing baby and the risk that untreated psychiatric illness poses to both mother and baby. A fresh graduate needs a working framework for weighing that trade-off, not just a memorised list of banned drugs.
A TERATOGEN is any drug, chemical or exposure capable of causing a structural or functional abnormality in a developing baby. In psychiatry, several commonly used mood stabilisers and, to a lesser extent, some antidepressants and benzodiazepines, carry meaningful risk, and that risk varies enormously by drug and by exact timing in pregnancy.
The single most important idea is timing: the risk of structural malformation is not constant across pregnancy, it is concentrated in a specific early window, and by the time many women realise they are pregnant, that window may already be closing or passed. This is why medication planning BEFORE conception, not reacting once pregnancy is confirmed, is repeatedly emphasised as the safest strategy.
Organogenesis (the period roughly from week 3 to week 8, when the baby's major organs are forming) is when the risk of structural malformation is greatest, because a teratogen acting on a forming organ can permanently disrupt its structure. Before that, in the all-or-nothing period of roughly weeks 0 to 2, an exposure either damages so few cells that development recovers completely or is severe enough to end the pregnancy; there is little room for a partial structural defect. After organogenesis, later exposure mainly affects growth and function rather than gross structure, and exposure very late in pregnancy causes neonatal effects, problems in the newborn immediately after birth, rather than malformation. Sodium valproate carries the highest known risk among mood stabilisers, causing neural tube defects (failures of the spine or brain to close properly) and significant later neurodevelopmental harm, and must be avoided in any woman who could become pregnant unless strict pregnancy-prevention measures are in place. Carbamazepine also raises neural tube defect risk, though generally considered lower than valproate's. Lithium in the first trimester is classically linked to Ebstein's anomaly, a rare heart-valve malformation, though the absolute risk is smaller than older teaching implied. Among SSRIs, paroxetine carries a small extra risk of cardiac defects, which is why sertraline is often the preferred in women who may become pregnant. Benzodiazepines used near birth can cause 'floppy infant syndrome' in the newborn: (low muscle tone), poor feeding and slowed breathing, plus a withdrawal syndrome in the baby after delivery, a NEONATAL effect rather than a structural malformation.
For example, a woman of childbearing age is newly diagnosed with bipolar disorder. Valproate might be an effective mood stabiliser on efficacy grounds alone, but its high risk usually rules it out before any other option is even weighed, because of the severity and frequency of harm it can cause if she later becomes pregnant, planned or not. Now imagine a woman already established on lithium who discovers she is six weeks pregnant, already past the point of prospective planning: the safest approach at that stage is not to panic-stop the drug, since abrupt discontinuation risks a relapse of severe mood illness, but to urgently review risk and benefit with a specialist perinatal psychiatry team, who will weigh the modest structural risk already incurred against the danger of destabilising her mental state. A third scenario, a pregnant patient stable and well on paroxetine, illustrates the same balancing act from the other side: the small extra cardiac risk must be weighed against the very real risk of relapse from stopping or switching an effective, tolerated drug, not treated as an automatic reason to change medication.
It is easy to conflate ALL psychotropic drugs as equally risky in pregnancy, but risk varies hugely: valproate sits at the high-risk end, most SSRIs other than paroxetine sit toward the lower-risk end, and untreated severe depression or itself carries real risks to mother and baby, including poor engagement with antenatal care, poor nutrition and risk of self-harm. Floppy infant syndrome from late benzodiazepine exposure should not be confused with a structural malformation from first-trimester exposure; they occur through different mechanisms and at different points in pregnancy.
There is no test that 'diagnoses' risk prospectively; risk assessment instead rests on a careful pregnancy-timing history (last menstrual period, or an early dating scan), a full medication history, and specialist antenatal screening (such as detailed anomaly ultrasound scanning) where a known teratogenic exposure has already occurred.
NICE guidance on antenatal and postnatal mental health, alongside specialist perinatal psychiatry input, recommends avoiding valproate in any woman of childbearing potential unless strict pregnancy-prevention measures are documented and in place, and recommends planning medication choice before conception wherever the pregnancy is anticipated. BNF and Maudsley Prescribing Guidelines both stress using the lowest effective dose and, where possible, a single drug rather than combination therapy, since combining agents multiplies uncertainty about the risk. Decisions should always weigh the drug's risk against the risk of untreated illness, never assume 'stop everything' is automatically the safer choice.
If a stem gives a specific gestational week alongside a drug exposure, map that week onto the timeline: weeks 3 to 8 (organogenesis) point to structural malformation risk, while exposure right before delivery points to a neonatal effect such as floppy infant syndrome, not a structural defect.
risk in psychiatric drugs is highest during organogenesis (weeks 3 to 8), varies hugely by drug (valproate highest, most SSRIs other than paroxetine lower), and management always weighs that risk against the real danger of untreated illness.