Clinical Practice
Selegiline Patch for TRD: A PMHNP Prescribing Protocol
The selegiline patch (EMSAM) is the MAOI with no diet rules at 6 mg. A PMHNP protocol for candidates, washouts, dosing, tyramine, and monitoring in TRD.
Read More →
TL;DR
- Lamotrigine clearance climbs roughly 80% in the first trimester and about 200% by the second, so a stable preconception dose quietly becomes a subtherapeutic one.
- Sertraline and escitalopram are the preferred first-line SSRIs; sertraline’s relative infant dose in breast milk is about 0.5%.
- ACOG advises against withholding or stopping psychiatric medication on the basis of pregnancy or lactation status alone.
- Valproate stays off the table: about 1-2% neural tube defect risk and malformation rates reported near 11% with first-trimester monotherapy.
- Postpartum psychosis hits 1-2 per 1,000 deliveries, usually inside the first two weeks, and is an emergency rather than a follow-up.
- The highest-yield visit is the one before conception, not the one at eight weeks gestation.
What’s in this guide
Lamotrigine clearance rises by roughly 200% by the second trimester. A patient who walked into preconception counseling stable on 200 mg can be functionally unmedicated by week 20 while filling the exact same prescription every month.
Nothing about her chart looks wrong. The dose is unchanged, adherence is perfect, and the relapse arrives anyway.
That gap is the whole problem with perinatal prescribing. Pregnancy does not simply add a teratogenicity question to an otherwise familiar decision. It rewrites the pharmacokinetics, moves the monitoring schedule, and compresses the window in which a missed change turns into a hospitalization.
Roughly 10% of pregnant patients and nearly 22% of postpartum patients meet criteria for depression, yet only about one in five receives adequate treatment. Most of that gap is not patient refusal. It is clinician hesitation, and it usually shows up as a reflexive taper the moment a pregnancy test turns positive.
This is the protocol side of that decision: what to start, what to monitor, what to avoid outright, and which two weeks demand a different level of vigilance than anything else in outpatient psychiatry.
The risk math nobody does out loud
The instinct to stop a medication in pregnancy feels conservative. It is not. It simply moves the risk somewhere that is harder to see on a consent form.
Untreated perinatal depression is associated with preterm birth, impaired bonding, substance use, disengagement from prenatal care, and suicide, which remains a leading cause of maternal death in the first postpartum year. Those outcomes rarely get written into the risk-benefit conversation with the same specificity as a malformation percentage, which makes the comparison dishonest before it starts.
The ACOG clinical practice guideline on treatment and management of perinatal mental health conditions is unusually direct about this. It recommends against withholding or discontinuing medication on the basis of pregnancy or lactation status alone, and it endorses SSRIs, SNRIs, and TCAs for postpartum depression rather than treating pharmacotherapy as a last resort.
Put in plain terms: discontinuation is an active clinical decision with its own side effect profile. It deserves the same documentation as a new prescription.
“The question I hear from new grads is always ‘is this medication safe in pregnancy.’ The better question is ‘what happens to this patient if I take it away.’ Once you ask it that way, most of these decisions get simpler, and a lot of them stop being tapers.”
— Lindsay Hill, DNP, PMHNP-BC
A practical framing helps here. Quantify the relapse risk from the patient’s own history, name the fetal risk with an absolute number rather than a relative one, and write both in the note. A patient who has had two hospitalizations off medication is not in the same decision as a patient with one mild episode six years ago.
Starting an SSRI in pregnancy
For a medication-naive patient, sertraline and escitalopram are the preferred starting agents. The reasoning is boring and that is the point: the largest exposure datasets, the most reassuring lactation figures, and the fewest interaction problems.
The efficacy data are better than most clinicians assume. In a sertraline trial in postpartum depression, 59% of patients on 50 to 200 mg over six weeks achieved clinical response against 26% on placebo, with remission rates of 53% versus 21%. That is a meaningful effect size in a population routinely told to try sleep hygiene first.
If a patient is already stable on a different agent, including paroxetine, the default is continuation rather than a switch. Swapping a working medication mid-pregnancy exposes the fetus to two drugs instead of one and gambles a known response on an unknown one.
Dosing and monitoring
Start at standard doses and titrate to remission, not to a number that feels cautious. Plasma volume expands substantially through the second and third trimesters, and hepatic metabolism shifts, so some patients need a dose increase in late pregnancy to hold the response they had at twelve weeks. Under-dosing to feel safe produces fetal exposure without maternal benefit, which is the worst available outcome.
Track response with the Edinburgh Postnatal Depression Scale or the PHQ-9 at every visit and chart the score. A 2026 review of updated perinatal depression pharmacotherapy guidance reinforces treating to remission on a validated instrument rather than to clinical impression.
For moderate-to-severe postpartum depression needing rapid onset, zuranolone offers a 14-day oral course at 50 mg daily. Worth knowing: brexanolone, the infusion that preceded it, was withdrawn from the market in January 2025, so any protocol or lecture slide still routing patients to it is out of date.
Mood stabilizers split three ways
Bipolar disorder in pregnancy sorts into three very different conversations depending on the agent, and collapsing them into one “mood stabilizers are risky” statement is how patients end up unmedicated.
Valproate: off the table
First-trimester valproate monotherapy carries a neural tube defect risk of roughly 1-2% and major congenital malformation rates reported as high as about 11%, more than triple background. There is also a dose-dependent neurodevelopmental signal. For any patient who could become pregnant, valproate should not be a maintenance choice unless safer options have genuinely failed and that reasoning is in the chart.
Lamotrigine: safe drug, moving target
Lamotrigine has the most reassuring malformation data of the anticonvulsant mood stabilizers across thousands of exposures. The risk is not teratogenicity; it is pharmacokinetic drift.
Apparent clearance rises about 82% in the first trimester and roughly 200% in the second and third compared with prepregnancy. Modeled trough concentrations fall from about 4.3 mg/L in nonpregnant patients to roughly 1.5 mg/L by the third trimester on an unchanged dose. Some patients lose nearly 90% of their level.
The protocol: draw a preconception or early-pregnancy reference level while the patient is well, repeat every four weeks, and increase the dose 20-25% when the level falls meaningfully below that reference. Then reverse it. Clearance returns to baseline within roughly two weeks of delivery, so a patient left at her third-trimester dose becomes toxic exactly when she is least able to report it. Write the postpartum taper into the plan before she delivers. If you are re-titrating from scratch, the same rash-risk rules apply as in any other patient — the Lamotrigine Rash: A 2026 PMHNP Titration Protocol covers that ramp.
Lithium: smaller number than you were taught
The Ebstein anomaly figure most clinicians carry is inflated. Current estimates put the absolute risk at roughly 1 in 1,000 to 1 in 2,000 first-trimester exposures, and a large cohort study of lithium use in pregnancy and cardiac malformations found an adjusted risk ratio of 1.65 for any cardiac malformation. Real, but small in absolute terms, and frequently smaller than the relapse risk of stopping.
Monitoring tightens considerably: levels at least monthly, more often near term, with a plan for the rapid fluid shifts around delivery. The Modern Lithium Monitoring: The 2026 PMHNP Protocol covers the baseline schedule this builds on.
Lactation: read the RID, not the package insert
Most package inserts say some version of “caution should be exercised,” which is not a clinical instruction. The usable number is the relative infant dose, or RID: the weight-adjusted percentage of the maternal dose the infant receives through milk. Below 10% is generally treated as compatible with breastfeeding.
Sertraline sits around 0.5%. Most breastfed infants of mothers taking it have undetectable or barely detectable serum levels, and follow-up studies out to five years have not shown developmental effects. The NIH’s Drugs and Lactation Database entry for sertraline lists it as a preferred antidepressant during breastfeeding, and it is free to look up in front of the patient.
That last point matters more than the number. A patient who has been told by three people to “ask your doctor” will believe a database she can see on your screen. Pulling up LactMed during the visit converts an anxious conversation into a two-minute one.
Two practical cautions. First, a low RID does not license a high dose without monitoring; watch the infant for sedation, poor feeding, or irritability, especially in preterm or medically fragile babies. Second, “pumping and dumping” around doses is almost never indicated for SSRIs and mostly serves to undermine both milk supply and the patient’s confidence in the plan.
For quick chairside dosing and monitoring references across the rest of your formulary, the free PMHNP prescribing cheat sheet is built for exactly this kind of lookup.
What most people get wrong
The common assumption is that the critical decision point is the positive pregnancy test. It is not. By the time a patient tests positive, organogenesis is already underway and the clinician is reacting rather than planning.
The decision that actually changes outcomes happens earlier, at the visit where a patient of reproductive age is started on anything. Roughly half of pregnancies in the United States are unplanned, which means the maintenance regimen chosen on an ordinary Tuesday is, statistically, the regimen that will be in place at conception.
That reframes the choice. Selecting valproate for a 28-year-old because it worked well in the hospital is not a neutral decision deferred to a future obstetric conversation. It is a first-trimester exposure decision made two years early.
The second thing clinicians get wrong is treating discontinuation as the cautious path. Stopping a medication produces a documented relapse risk, and relapse in pregnancy carries its own obstetric consequences. If a taper is genuinely indicated, it should be a gradual, monitored process rather than an abrupt stop — the same principles in the Hyperbolic Antidepressant Tapering: 2026 PMHNP Protocol apply, with tighter follow-up.
The practical fix is small. Ask every patient of reproductive age about pregnancy plans at the visit where you choose a maintenance agent, and write the answer down. That single question moves the whole decision upstream, where you still have options.
The two-week window nobody screens for
Postpartum psychosis occurs after roughly 0.1% to 0.2% of deliveries in patients with no psychiatric history, and the rate is at least ten times higher than at any other point in a woman’s life. Onset is fast, usually within the first two weeks and often within the first few days.
It is a genuine emergency. Suicide and infanticide risk are both elevated, and the presentation can fluctuate hour to hour, which means a reassuring phone check at noon is not evidence of safety at 8 p.m.
The early signs are easy to normalize in a new parent: insomnia without the expected fatigue, rapid speech, confusion, or a conviction about the baby that does not track with reality. Waking every two hours is expected. Being unable to sleep when the baby sleeps, three nights running, is not.
Risk concentrates sharply in bipolar disorder and in anyone with a prior postpartum psychotic episode, where recurrence after a subsequent pregnancy runs around 31%. A study on prevention of postpartum psychosis and mania in high-risk women supports prophylaxis initiated immediately postpartum in that group, which is a planning conversation held during the third trimester, not a decision made in triage.
Three things belong in every discharge plan for a patient with bipolar disorder: a named support person instructed on what to watch for, an explicit sleep-protection plan for the first two weeks, and a direct route to you that does not involve a two-week callback queue. The same risk-assessment structure used elsewhere in practice applies here — the Suicide Risk Assessment: New PMHNP Confidence Playbook covers how to document that conversation.
Frequently asked questions
Is sertraline safe during pregnancy and breastfeeding?
Sertraline is the most commonly recommended first-line SSRI across both pregnancy and lactation. Its relative infant dose through breast milk is roughly 0.5%, far below the 10% threshold generally treated as compatible with breastfeeding, and most exposed infants have undetectable serum levels. ACOG explicitly advises against withholding or stopping medication on the basis of pregnancy or lactation status alone.
Do I need to change a patient’s lamotrigine dose during pregnancy?
Almost always, yes. Apparent lamotrigine clearance rises roughly 80% in the first trimester and about 200% by the second and third, so trough levels fall sharply on an unchanged dose. Check a preconception reference level, repeat every four weeks, and raise the dose 20-25% when the level drops meaningfully below that baseline.
What happens to lamotrigine after delivery?
Clearance returns to the prepregnancy baseline fast, sometimes within about two weeks. A patient left on her inflated third-trimester dose can become toxic during the same weeks she is sleep-deprived and hardest to assess. Plan the taper back toward the preconception dose before she delivers, not after she calls with dizziness.
Can a patient on lithium get pregnant safely?
Often, yes, with monitoring. The absolute risk of Ebstein anomaly with first-trimester lithium exposure is roughly 1 in 1,000 to 1 in 2,000 exposures. That is a real but small number, and for many patients the relapse risk after stopping lithium is the larger threat. Levels need checking more often as plasma volume expands, and dosing must be revisited immediately postpartum.
Which psychiatric medication should be avoided in pregnancy?
Valproate. First-trimester monotherapy carries a neural tube defect risk of about 1-2% and a major congenital malformation rate reported as high as roughly 11%, more than three times background. It should not be a maintenance choice for any patient who could become pregnant unless every safer option has genuinely failed and that discussion is documented.
When is postpartum psychosis most likely to appear?
Most cases declare themselves within the first two weeks after delivery, and often within the first few days. It affects roughly 1 to 2 per 1,000 deliveries in patients with no psychiatric history and is a true emergency because of suicide and infanticide risk. Any new confusion, insomnia without fatigue, or disorganized thinking in that window is an immediate evaluation, not a follow-up visit.
The bottom line
Perinatal prescribing is not a specialty carve-out that belongs to someone else. Any PMHNP with a panel of adults is already doing it, whether or not the conversation has happened out loud.
The clinical content is learnable in an afternoon: sertraline or escitalopram first line, lamotrigine levels every four weeks with a planned postpartum taper, valproate off the table, LactMed open on the screen, and the first two weeks postpartum treated as a monitored period rather than a gap before the six-week visit.
The habit is harder and matters more. Start asking every patient of reproductive age about pregnancy plans at the visit where you choose a maintenance medication, and put the answer in the note. It costs fifteen seconds and moves the entire risk conversation to a point where you still have good options.
Prescribing decisions are easier with backup
The Psych NP Fellowship is a 12-month clinical mentorship for new and early-career PMHNPs, built around the cases that do not have a clean algorithm.
This content is for educational purposes and does not replace individualized clinical judgment or supervision.
Lindsay Hill, DNP, PMHNP-BC is the founder of the Psych NP Fellowship, a 12-month clinical mentorship program for new and early-career psychiatric nurse practitioners. She is a published contributor to Psychiatric Times, past President of the Arizona APNA Chapter, and co-founder of the Psych NP Network.
