When to Start vs. Discontinue Psychiatric Medication in Pregnancy

The Default Problem

Clinicians and patients often approach psychiatric medication in pregnancy with an implicit default: medications should be avoided unless there is a compelling reason to use them. This default is understandable but clinically inaccurate.

The correct default is: make an individualized risk-benefit assessment. There is no blanket answer that applies to all patients, all medications, or all severity levels. A patient with a single mild depressive episode several years ago presents a different calculus than a patient with recurrent severe depression who is currently stable on an SSRI.

This article provides the framework for making that assessment.

---

The Components of the Risk-Benefit Analysis

Any medication decision in pregnancy requires weighing four things:

1. Risk of the medication to the fetus/neonate. This includes teratogenicity (malformation risk in the first trimester), fetal effects during ongoing pregnancy, neonatal adaptation effects at delivery, and potential long-term neurodevelopmental effects.

2. Risk of untreated psychiatric illness to the patient. Psychiatric illness affects maternal function, safety, prenatal care utilization, nutrition, and substance use risk. These are direct patient risks, not secondary considerations.

3. Risk of untreated psychiatric illness to the fetal/infant outcome. Maternal depression and anxiety elevate cortisol; cortisol crosses the placenta. Untreated prenatal depression is associated with preterm birth, low birth weight, and adverse neonatal neurodevelopmental outcomes. Maternal suicidality is a direct fetal risk.

4. Severity of the current psychiatric presentation. A patient with mild subsyndromal anxiety is in a fundamentally different risk-benefit position than a patient with severe recurrent major depression and a prior hospitalization.

---

The Starting Decision

When to start medication in pregnancy

Strong indication:

• Moderate-to-severe depression with functional impairment
• Active suicidal ideation
• Anxiety disorder with significant impairment (panic disorder, severe GAD) that is not responding to non-pharmacological treatment
• Psychiatric history with documented severe recurrence risk (multiple prior episodes, prior hospitalization, prior postpartum deterioration)
• OCD with significant impairment
• Bipolar disorder currently managed pharmacologically (continuation, not initiation, but same principle applies)

Consider medication with concurrent therapy:

• Moderate depression without active suicidality but with functional impairment
• Anxiety with moderate functional impairment in a patient with prior medication response
• Patient preference, after informed discussion of risks and non-pharmacological alternatives

Non-pharmacological treatment first:

• Mild depression or anxiety with minimal functional impairment
• Patient with strong preference against medication
• Adequate access to perinatal mental health therapy

The decision should never be "medication vs. nothing." If medication is deferred, what is the treatment plan? If the answer is "watch and wait," that is a clinical decision that requires explicit documentation of the rationale and close follow-up.

First-line agents for new starts in pregnancy

For depression and anxiety: sertraline 50 mg with titration to response. CBT or IPT concurrent with or instead of medication depending on severity and preference.

For OCD: sertraline or fluoxetine (often require higher doses than depression treatment; up to 200 mg sertraline may be needed for adequate OCD response).

---

The Continuation Decision

The patient who is already on medication

A patient who presents in early pregnancy while stable on psychiatric medication presents a continuation decision, not a new-start decision.

The default for a stable patient on effective medication is continuation, not discontinuation. Changing the medication situation of a stable patient -- by discontinuing or switching agents -- introduces relapse risk without a clear clinical benefit in most cases.

When continuation is supported:

• The patient is clinically stable and functioning
• The medication has an acceptable perinatal safety profile
• The indication remains active (the patient still has the underlying condition)

When modification is warranted:

• The medication has a safety profile that is meaningfully worse than available alternatives (paroxetine, some mood stabilizers)
• The dose can be reduced without sacrificing clinical stability
• A switch to a more studied agent can be accomplished without destabilization

When switching is not indicated:

• The patient is stable on fluoxetine and sertraline is "preferred" -- the advantage of sertraline is not sufficient to warrant switching a stable patient
• The patient has a prior documented sertraline failure but is stable on something else

The "I stopped before I found out I was pregnant" situation

This is common. A patient who stopped her medication before discovering she was pregnant presents with a question: restart?

The answer depends on severity of prior illness and her current state. If she has a history of moderate-to-severe depression with prior recurrence, restart is almost always the correct clinical decision. If she has a history of a single mild episode years ago that is not active, watchful monitoring with early reinstatement at symptom onset may be appropriate.

---

The Discontinuation Decision

When discontinuation might be appropriate

Discontinuation during pregnancy is appropriate when:

• The indication is no longer active (true remission, not medication-maintained stability)
• Taper can be accomplished without significant recurrence risk
• A supervised, gradual taper is clinically feasible
• The patient's history does not include rapid recurrence after medication discontinuation

Discontinuation is not appropriate when:

• The patient is on medication because the medication is what is keeping her stable
• The history includes rapid recurrence or severe episodes following prior discontinuation
• The patient is stopping because she read something alarming online rather than because of a clinical indication to stop

How to taper when discontinuation is indicated

Gradual taper over 4 to 8 weeks minimum. Monitor closely during taper and for 4 to 8 weeks post-discontinuation. The highest recurrence risk for perinatal depression is the postpartum period; stopping medication in the third trimester to "avoid NAS" and then relapsing postpartum is a worse outcome than managing NAS.

Agents with significant discontinuation syndromes (paroxetine, venlafaxine): taper slowly. Abrupt discontinuation is not appropriate for these agents even when discontinuation is clinically indicated.

---

The Postpartum Consideration

For patients who discontinued medication during pregnancy with plan to restart postpartum, reinstatement timing is clinically important.

When to restart: The postpartum period is the highest risk period for PMAD onset, not after symptoms develop. For patients with significant psychiatric history, reinstatement at delivery or immediately postpartum (rather than waiting for symptoms) is appropriate.

New onset postpartum: Treat. The same considerations that drive treatment during pregnancy apply postpartum, with the addition of breastfeeding safety considerations for the specific agents.

---

Documentation

Every decision about psychiatric medication in pregnancy should be documented with:

• The clinical indication
• The patient's current severity level
• The risk-benefit rationale for the decision made
• The alternatives considered
• Patient education provided
• Follow-up plan

This documentation protects the prescriber and ensures continuity of care across the multiple providers who will see the patient during the perinatal period.

---