Psychiatric Medication During Pregnancy and Breastfeeding

Pregnancy changes the decision, but it does not make untreated illness harmless

Finding out that you are pregnant while taking psychiatric medication can trigger an understandable urge to stop immediately. That can be dangerous. Abrupt withdrawal, relapse of depression or anxiety, mania, psychosis, insomnia, poor nutrition, substance use and loss of functioning can also harm the pregnant person and the baby.

ACOG recommends against withholding or discontinuing medication for a mental health condition because of pregnancy or breastfeeding status alone. NICE similarly advises weighing previous response, relapse risk, the risks of stopping or switching, fetal and neonatal effects, and the role of psychological treatment.

The right plan is diagnosis-specific and person-specific. A medicine that has kept severe bipolar disorder stable for years is a different decision from a first prescription for mild symptoms. The timing in pregnancy, dose, number of medicines, previous relapses, medical conditions and the availability of close follow-up all matter.

The comparison is treatment versus an active alternative, not medicine versus zero risk

• What is the diagnosis, its usual course, and the person's previous pattern of relapse?
• What happened during previous pregnancies or postpartum periods, if any?
• Which treatment has worked, which has failed, and what happened after earlier dose reductions or discontinuation?
• What is known about congenital malformations, pregnancy complications, neonatal adaptation and longer-term development for this medicine?
• Would switching expose the fetus to two medicines while risking relapse?
• Can one effective medicine be used instead of several, with the lowest dose that still treats the illness?
• What monitoring is needed during pregnancy, around delivery and after birth?

No pregnancy can be guaranteed to have a normal outcome. Background risks exist before medicine is considered, and observational studies are affected by illness severity, smoking, obesity, other medicines and healthcare differences. Good counselling uses absolute risks where reliable, acknowledges uncertainty and records the plan.

There is no single rule for all psychiatric medicines

The safest-looking drug on a population chart is not automatically the safest change for an individual. Previous response is a major part of the evidence. A poorly effective substitute can add exposure without controlling illness.

RID is useful, but it is a screening estimate, not a safety certificate

Relative infant dose, or RID, compares the estimated dose received through milk per kilogram of infant weight with the mother's dose per kilogram of maternal weight.

An RID below 10% is commonly used as a reassuring screening threshold. It is not a biological cliff and does not prove that a medicine is safe. The estimate can vary with milk sampling, dose timing and individual metabolism. It also does not capture an active metabolite, a long half-life, infant oral absorption, prematurity, illness, organ function, exclusive versus partial breastfeeding, or an observed adverse effect.

The most direct measure of exposure is the infant's blood concentration, but routine infant levels are not needed for most low-transfer medicines. They become more relevant when the medicine has substantial or variable transfer, the infant is premature or unwell, symptoms appear, or a specialist protocol requires them.

Illustrative ranges show why the number needs context

These are population estimates, not dose instructions. A low RID does not justify starting, stopping or switching a medicine without considering the mother's illness and the infant's health. A higher RID does not automatically make breastfeeding impossible, but it raises the level of expertise and monitoring required.

Observation is specific to the medicine and the infant's vulnerability

• Feeding, latch, sucking strength and whether feeds are being missed.
• Unusual sleepiness, difficulty waking, marked irritability or agitation.
• Weight gain, hydration and the number of wet nappies.
• Breathing pauses, limpness, tremor, abnormal movements or a new rash.
• Jaundice, persistent vomiting, diarrhoea or symptoms of infection.

Premature infants and infants with kidney, liver, cardiac or neurological illness clear medicines less efficiently and need a more conservative plan. Sedating medication also increases the practical risk of falling asleep while feeding. Avoid feeding in a position where accidental sleep could obstruct the baby's breathing, and involve another adult when sedation is significant.

For lithium, specialist plans may include infant lithium, kidney and thyroid tests. For lamotrigine, a concerning rash, apnoea, drowsiness or poor sucking may trigger infant levels and other tests. For clozapine exposure, sedation and white blood-cell monitoring are important. The paediatrician should know the full maternal medication list.

The best time to make a postpartum plan is before the crisis

• Review medication before conception where possible, without destabilising an effective regimen unnecessarily.
• Record what to do if symptoms return during pregnancy and who to contact out of hours.
• Plan medication monitoring as pregnancy changes drug clearance.
• Tell the obstetric and neonatal teams about late-pregnancy exposure so newborn adaptation can be observed.
• Discuss feeding preferences before delivery, including a backup plan if the infant is premature or unwell.
• For bipolar disorder or previous postpartum psychosis, make an explicit sleep, support and rapid-access postpartum plan.

The goal is not the smallest possible medication exposure at any cost. It is the lowest overall risk for mother and baby, including the risk of severe relapse. That balance can reasonably lead to continuing, adjusting, switching or, in selected cases, gradually stopping a medicine.