Psychiatric illness during pregnancy often returns when treatment stops
Pregnancy does not reliably protect against psychiatric disorders. Recent data place mood or anxiety symptoms in as many as 20% of pregnant women, and the risk rises sharply in patients with a prior history of illness who discontinue medication after conception. In one prospective cohort of women with major depression, 21 of 82 women who stayed on antidepressants relapsed during pregnancy, compared with 44 of 65 who stopped treatment. The difference was not subtle: discontinuation carried about a fivefold higher relapse risk.
The same pattern appears in bipolar disorder, where recurrence during pregnancy is common and clinically disruptive. In one study, 70.8% of women experienced at least one mood episode during pregnancy, and recurrence was higher among those who stopped mood stabilizers than among those who maintained them. The bedside lesson is plain. For many patients, abrupt cessation of psychotropic therapy can destabilize mood faster than clinicians or families expect, and the cost of relapse may extend beyond maternal suffering to prenatal care, substance use, and obstetric complications.
Why untreated maternal illness can be as consequential as drug exposure
The decision to continue, taper, switch, or stop treatment cannot be made on fetal exposure alone. Untreated depression and anxiety carry their own biologic and obstetric burdens. Women with active psychiatric illness are less likely to receive adequate prenatal care and more likely to use alcohol, tobacco, or other substances that worsen pregnancy outcomes. Studies have linked maternal depression with low birth weight, fetal growth restriction, preterm delivery, preeclampsia, operative delivery, and admission of the newborn to special care nurseries for problems such as respiratory distress, hypoglycemia, and prematurity.
Those associations matter because they force a more realistic clinical framing. Psychotropic treatment is not being weighed against a neutral baseline; it is being weighed against an illness state that may already be harming mother and fetus. The practical question is not whether exposure is risk-free, because no medication in this space is entirely risk-free. The real issue is whether treatment reduces a larger maternal-fetal threat than the medication itself introduces.
Placental transfer, teratogenic windows, and neonatal symptoms
All medications cross the placenta to some degree, and psychotropic agents are no exception. Risk assessment has to account for three separate possibilities: teratogenesis, neonatal toxicity, and long-term neurobehavioral effects. The first trimester deserves special attention because major organogenesis is concentrated in the earliest weeks of pregnancy, with neural tube closure occurring within the first four weeks of gestation and major cardiac development largely taking place by weeks four through nine.
The baseline incidence of major congenital malformations in the United States is estimated at 2% to 4%, which is the reference point every counseling conversation should carry. Against that background, several antidepressants have accumulated fairly reassuring human data. Fluoxetine, the best characterized antidepressant in the review, has not been linked to an increased rate of major congenital malformation across more than 2,500 cases. Prospective data on 531 infants with first-trimester SSRI exposure, mostly citalopram, also did not show an increased risk of organ malformation. Tricyclic antidepressants, including desipramine and nortriptyline, have similarly not shown a meaningful signal for major congenital anomalies in pooled analyses of more than 400 first-trimester exposures.
One antidepressant remains more controversial: paroxetine. Some earlier reports connected first-trimester exposure with cardiac defects, including atrial and ventricular septal defects, although other studies did not reproduce that signal. Independent meta-analyses have consistently failed to show an association with cardiovascular malformations, yet the FDA still moved the drug’s category from C to D. The disagreement reflects the challenge of separating a true teratogenic signal from confounding, small sample sizes, and the underlying illness itself.
Bupropion has also generated extensive discussion. Registry data from 517 pregnancies with first-trimester exposure found 20 major malformations, a 3.9% rate that sits within the expected background range. A retrospective cohort of more than 1,200 exposed infants did not show excess overall or cardiovascular malformation risk. Venlafaxine data from 150 first-trimester exposures were similarly reassuring, while studies of nefazodone, trazodone, and mirtazapine have not shown significant differences in congenital malformation rates compared with control groups. Even so, these newer agents need larger datasets before reproductive safety can be considered settled.
Neonatal symptoms are a separate issue from structural malformation. Late-pregnancy exposure to antidepressants has been associated with transient newborn syndromes marked by tremor, restlessness, increased muscle tone, and increased crying. These symptoms have been described in about 25% of infants exposed near delivery and typically resolve within one to four days without specific intervention. The clinical interpretation has to stay disciplined: some reports may reflect cautious observation rather than severe drug toxicity, and maternal mood at delivery is a major confounder because depression and anxiety themselves are associated with poor neonatal outcomes.
Counseling patients requires more than medication labels
The FDA’s shift from letter categories to the Pregnancy and Lactation Labeling Rule was meant to correct a blunt system that often misled more than it informed. The older A, B, C, D, and X framework compressed complex risk into a label that failed to capture timing, dose, indication, or the maternal consequences of stopping therapy. The newer labeling approach replaces that shorthand with more detailed information on maternal benefit, fetal risk, and how risk may change across pregnancy and lactation.
For clinicians, the value of that change lies in how counseling is actually done. A patient who is stable on antidepressant therapy and suddenly stops because of pregnancy may be taking a greater psychiatric risk than fetal risk, especially if her history includes severe depression or bipolar disorder. A patient with first-onset illness during pregnancy may need treatment before symptoms erode sleep, appetite, engagement in prenatal care, and safety. And for some patients, the plan may involve continuing the current medication rather than switching to a less effective drug simply because pregnancy has begun.
The evidence also argues against reflexive tapering of SSRIs before delivery. Some newborn symptoms have been reported with late-pregnancy exposure, but tapering has not been shown to improve neonatal outcomes, and dose reduction late in pregnancy may increase the risk of postpartum depression. That postpartum period is a high-risk window on its own, particularly for women who have been symptomatic during pregnancy.
What is still missing is a larger, better-controlled evidence base for several newer antidepressants and for long-term neurobehavioral outcomes after in utero exposure. Human data remain limited on whether prenatal psychotropic exposure alters cognition, behavior, or development later in childhood, especially for agents outside the best-studied SSRIs. Until those gaps close, treatment decisions will continue to depend on individualized risk assessment, illness severity, prior response, and close follow-up rather than any single category label or one-size-fits-all rule.