Antidepressants in Pregnancy

The calculation NICE actually asks for

The most common mistake in this area is treating "stop the medication" as the cautious option. NICE guideline CG192 does not frame it that way. Recommendation 1.4.7 asks clinicians to discuss "the background risk of harm to the woman and the fetus or baby associated with the mental health problem and the risk to mental health and parenting associated with no treatment" alongside "the risks or harms to the woman and the fetus or baby associated with each treatment option".

For antidepressants specifically, CG192 asks clinicians to take into account "the risks associated with switching from or stopping a previously effective medication", and to "seek advice from a specialist (preferably from a specialist perinatal mental health service) if there is uncertainty about specific drugs".

Two things follow. Stopping abruptly is not recommended. And the decision should involve someone with perinatal expertise, not be made alone at 11pm after reading a patient information leaflet.

Why the leaflet in the box frightens people

Almost every antidepressant patient information leaflet contains cautious pregnancy wording, because manufacturers write to protect against liability rather than to guide clinical decisions. That wording is not a statement that the drug is known to be harmful; usually it reflects the absence of randomised trials, which do not get done in pregnancy for ethical reasons.

The UK's specialist resource for this is Bumps, produced by the UK Teratology Information Service, which describes itself as providing "freely available, reliable, evidenced-based and accurate information about the use of medicines in pregnancy", with leaflets "based on the scientific material UKTIS produces for healthcare professionals" incorporating "information from all relevant published scientific studies". It is drug-specific, updated, and written to be read by patients. It is a far better source than the paper in the box.

What is actually known, and what is not

The honest summary is that the evidence for the commonly used antidepressants in pregnancy is observational, and observational evidence in this field is very hard to interpret because depression itself is associated with the outcomes being studied. When a study finds an association between an antidepressant and an outcome, disentangling the drug from the illness it treats is the central difficulty, and it is rarely fully solved.

This page therefore names classes rather than assigning them risk scores. The drug groups involved are SSRIs, SNRIs, tricyclic antidepressants and mirtazapine, and CG192 discusses them under those headings. There are no doses on this page, and there is no ranking of "safest" antidepressant, because the right choice is dominated by which one has worked for you before. CG192 recommendation 1.4.12 says as much when it asks clinicians to choose the drug with the lowest risk profile "taking into account a woman's previous response to medication".

Where NICE does give a hard instruction, it is about a different drug class: recommendation 1.2.3 excludes valproate for mental health treatment in women of childbearing potential unless other options have failed and a pregnancy prevention programme is in place.

Around birth: what to expect for your baby

Babies exposed to antidepressants late in pregnancy are observed after birth for a self-limiting adaptation syndrome: jitteriness, irritability, feeding difficulty or fast breathing in the first days. It is usually mild and usually resolves without treatment, but it is the reason a baby may be kept for observation rather than discharged in a few hours.

This is worth knowing in advance for one reason: if nobody warns you, an anxious few days of watching your newborn get monitored can be read as confirmation that you made the wrong decision. It is not; it is the plan working as intended. Ask before birth what your unit's observation period is, so it does not arrive as a surprise.

Stopping, if that is the decision

If you and your clinician decide to stop, it is done as a planned reduction, not abruptly, and ideally before conception rather than in the first trimester when much of the relevant organ development has already happened. Rapid discontinuation causes discontinuation symptoms in its own right, which are easily mistaken for relapse and for pregnancy symptoms at the same time.

Agree in advance what you will do if symptoms return, including whether restarting is the plan and who authorises it. A relapse plan written while you are well is worth more than one improvised while you are unwell.

Talking therapy is part of the answer

The Royal College of Psychiatrists sets out what perinatal mental health services in the UK offer, including specialist mental health midwives for milder anxiety or low mood and priority access to talking therapies for people who are pregnant or have recently given birth. Priority access is a real entitlement and it is under-claimed. Ask specifically.

NICE guideline NG201 places mental health enquiry inside routine antenatal care, so booking is the natural moment to raise it, and ACOG's 2023 clinical practice guideline on screening and diagnosis of perinatal mental health conditions sets out the equivalent structured screening approach used in the US. The NHS pages on antidepressants describe the drug groups and their usual effects outside pregnancy.

Breastfeeding

Feeding decisions are made separately from the pregnancy decision, and the answer differs by drug. CG192 points clinicians to specialist advice and to lactation medicines information services rather than offering a blanket rule. If you plan to breastfeed, ask for that specific answer before birth, because it occasionally changes which antidepressant is preferred.