Antidepressants During Pregnancy: Benefits, Risks, and Treatment Decisions
Updated: 3 days ago
Author: Angela Bogdanova · Published: October 18, 2026 · Editorial Policy
Research cluster: Cluster Perinatal and Reproductive Mental Health · Written by Artificial Sapiens
Antidepressants during pregnancy require an individualized decision that considers the severity and recurrence of the mental health condition, the benefits of effective treatment, the particular medication and dose, the timing of exposure, and the health of the pregnant patient and developing child. For many patients with significant depression or anxiety, continuing an effective antidepressant is a reasonable, guideline-supported option. For some with mild, stable illness, evidence-based psychotherapy and a carefully monitored medication change may be appropriate. The central clinical question is which treatment plan offers the best overall prospect of maternal and infant health in this pregnancy. The American College of Obstetricians and Gynecologists (ACOG) recommends against withholding or stopping medication for a mental health condition solely because of pregnancy or lactation, and emphasizes individualized care and symptom monitoring (ACOG, 2023; ACOG assessment guidance).
This chapter explains what evidence can establish, where uncertainty remains, and how a patient and clinicians can reach a documented, revisable decision. Its focus is antidepressant treatment during pregnancy; treatment during breastfeeding and specialist care for bipolar disorder are separate clinical questions. It is educational material for informed discussion with a treating team, rather than an individualized prescription, medication-switch instruction, or emergency assessment.
What antidepressants treat during pregnancy
Antidepressants are medications prescribed for defined clinical indications, including major depressive disorder and several anxiety-related conditions. Their name describes a primary therapeutic use, not a guarantee that every patient taking one has depression. Selective serotonin reuptake inhibitors (SSRIs) and serotonin–norepinephrine reuptake inhibitors (SNRIs) are common examples. An individual may also be treated for obsessive-compulsive disorder, panic disorder, posttraumatic stress disorder, or another condition, each with its own evidence-based treatment pathway.
Pregnancy can coincide with a first depressive episode, a recurrence of a previous condition, or sustained remission on a maintenance medication. These circumstances have different relapse risks and treatment objectives. A symptom such as tearfulness, fatigue, irritability, reduced appetite, or poor sleep needs interpretation in context. Pregnancy-related physiological changes overlap with psychiatric symptoms, and psychological distress alone does not establish a major depressive episode. A clinical diagnosis requires evaluation of duration, associated symptoms, functioning, context, differential diagnoses, and safety. Screening questionnaires identify people who need further assessment; they do not themselves diagnose an illness (ACOG, 2023, Screening and Diagnosis).
Before choosing an antidepressant, clinicians should also ask about episodes of elevated or irritable mood accompanied by decreased need for sleep, increased energy, impulsivity, or other evidence suggesting bipolar disorder. An antidepressant-centered treatment strategy can be inappropriate for bipolar I disorder and may contribute to mood destabilization in susceptible patients. That distinction has particular importance when assessing severe antenatal or postpartum mood symptoms (ACOG, 2023, Treatment and Management).
The clinical decision: compare complete treatment pathways
The choice is rarely between a medication with possible adverse effects and a pregnancy with zero other risks. Depression and anxiety occur within biological, psychological, social, and medical circumstances that can themselves affect sleep, nutrition, relationships, prenatal care, substance exposure, functioning, and safety. A sound assessment compares plausible pathways: continuing effective medication; beginning medication; psychotherapy alone; combined treatment; switching medication; gradual discontinuation with structured monitoring; or a more intensive specialist treatment when illness is severe.
For a patient whose previous recurrent episodes caused hospitalization, marked loss of functioning, suicidal thinking, or substantial inability to care for herself, preventing relapse is a substantial potential benefit. A patient with a single remote episode, long remission, robust supports, and ready access to psychotherapy may weigh the options differently. Previous response and adverse reactions often matter more clinically than an abstract ranking of drugs detached from the patient's history. NICE explicitly directs clinicians to review the benefits of treatment, background risks of the condition, previous response, possible harm from stopping or switching, and the patient's own preferences (NICE, CG192).
The decision is also dynamic. Preconception planning permits time to optimize treatment, but an unplanned pregnancy should prompt medication review rather than a reflexive interruption. Early pregnancy and late pregnancy present different exposure questions. A plan may change if symptoms recur, side effects appear, blood pressure changes, fetal concerns arise, or delivery approaches. Shared decision-making is an ongoing clinical process, with a named clinician responsible for follow-up.
What are the benefits of treating depression during pregnancy?
Effective depression treatment may relieve persistent sadness, loss of interest, guilt, hopelessness, anxiety, poor concentration, and other symptoms that interfere with daily life. It can restore the capacity to work, sleep, eat regularly, maintain relationships, engage with prenatal care, and plan for delivery and postpartum needs. The benefits are experienced by the pregnant person directly and can support family functioning and continuity of healthcare. Antidepressants have evidence of benefit for several depressive and anxiety disorders outside pregnancy; perinatal guidelines incorporate those treatment benefits while weighing pregnancy-specific exposure data (ACOG, 2023).
Untreated or undertreated psychiatric illness may be associated with adverse obstetric outcomes, but observational associations should not be mistaken for a simple causal chain. Severity of illness, smoking or other substance exposures, poverty, trauma, comorbid conditions, and access to care can influence both psychiatric treatment and pregnancy outcomes. Medication decisions should therefore consider the immediate, demonstrable burden of symptoms while interpreting epidemiological risk estimates with their methodological limitations.
A particularly important benefit is relapse prevention in patients with recurrent major depression. In a prospective observational study of women with a prior major depressive episode who were well at conception, depressive relapse occurred in 68% of those who discontinued antidepressants and 26% of those who maintained medication during pregnancy. This influential cohort establishes a substantial risk signal in that selected clinical population; it is not a universal prediction for every pregnant patient, and treatment continuation was not randomly assigned (Cohen et al., 2006). Later studies and contemporary clinical decisions must account for differences in baseline illness severity, treatment history, and the reasons people continue or stop medication.
What are the major categories of risk?
A clinician can organize the evidence by outcome and gestational timing. First-trimester research often concentrates on congenital malformations, including cardiac defects, because organ formation is underway. Later-pregnancy studies more often assess premature birth, growth, maternal bleeding, persistent pulmonary hypertension of the newborn (PPHN), and signs of neonatal adaptation following delivery. Long-term studies examine learning, behavior, and neurodevelopment. Miscarriage, stillbirth, and other rare outcomes introduce additional problems of confounding and statistical precision.
For any reported association, four questions matter: What was the actual absolute frequency? What comparison group was used? How carefully did investigators account for psychiatric illness and other confounders? And how directly does the finding apply to this specific medication, dose, trimester, and patient? A finding described as a relative increase can correspond to a small difference in the number of affected infants. Conversely, a small average effect may be clinically important in an individual at elevated baseline risk.
Baseline risk and the meaning of absolute numbers
Congenital anomalies arise in pregnancies without antidepressant exposure as well as in exposed pregnancies. MotherToBaby uses a general background estimate of approximately 3 in 100 births with a birth defect when explaining medication research; the rate for a particular defect, population, or definition is different. If a study reports a 20% relative increase from a hypothetical 1% baseline, the corresponding absolute risk is 1.2%, an increase of two events per 1,000. That numerical illustration is a calculation, not a risk estimate for any named antidepressant (MotherToBaby, Sertraline).
Whenever possible, counseling should state the absolute risk in an appropriately matched population and the degree of confidence in that estimate. A risk ratio without its baseline obscures important meaning. Case reports can identify possible signals but cannot establish the frequency or causation of a complication.
First-trimester exposure and congenital malformations
Large epidemiological studies have not established a substantial overall increase in major congenital malformations for commonly prescribed SSRIs as a class. Some earlier reports implicated particular heart defects or specific drugs, but the strength of these associations often weakened after investigators accounted for characteristics associated with depression and treatment. In a large US Medicaid cohort, Huybrechts and colleagues found that the initially elevated cardiac-malformation association with antidepressant exposure attenuated after progressively controlling for confounding. The study does not prove identical safety for every drug and every defect; it illustrates why comparing all medicated pregnancies with all unmedicated pregnancies can mislead (Huybrechts et al., 2014).
Paroxetine deserves a separate discussion because some studies and historical regulatory warnings identified possible first-trimester cardiac-malformation associations. Current decisions depend on the magnitude and consistency of the evidence, previous clinical response, alternatives, and consequences of switching. Someone considering a new treatment may prefer a drug with broader reassuring reproductive data. Someone stable for years on paroxetine, after unsuccessful alternatives, should discuss the issue with clinicians before changing treatment. Exposure that has already occurred cannot be undone by abrupt cessation; individualized fetal assessment may be discussed with the obstetric team (NICE, CG192; MotherToBaby paroxetine resource).
The absence of a large demonstrated increase for common SSRIs is an evidence-based reassurance with limits: rare anomalies, less frequently used agents, medication combinations, and high-dose exposures are often studied with substantially less precision. A report of no statistically significant association also does not prove that the true effect is exactly zero.
Pregnancy loss, prematurity, fetal growth, and hypertensive complications
Research on miscarriage and stillbirth is difficult to interpret because underlying psychiatric disorders, medical comorbidities, past pregnancy history, and factors leading to treatment can be associated with outcomes. The appropriate comparison is often between people with similar psychiatric histories who differ in exposure, rather than between medicated patients and everyone else. Even careful observational adjustment cannot remove all unmeasured confounding.
Some analyses report a modest association between prenatal antidepressant use and preterm birth. In a large Swedish cohort using sibling comparisons, first-trimester antidepressant exposure remained associated with preterm birth, whereas associations with being small for gestational age and later autism or ADHD attenuated under the sibling design. The sibling-adjusted odds ratio for preterm birth was 1.34 (95% CI 1.18–1.52). That result deserves attention, yet design assumptions and residual differences between pregnancies remain relevant. It should not be translated into an individual prediction that medication necessarily causes premature delivery (Sujan et al., 2017).
Some research has raised concerns about hypertensive disorders and other obstetric outcomes with particular antidepressant exposures. Medication class, baseline hypertension, metabolic health, disease severity, and timing matter. Obstetric monitoring should be based on the patient's overall clinical risk rather than a generic claim that all antidepressants cause preeclampsia. A new blood pressure elevation, severe headache, visual changes, or other obstetric warning signs require prompt medical evaluation regardless of psychiatric treatment.
Later pregnancy: newborn adaptation and respiratory outcomes
Poor or delayed neonatal adaptation
After exposure to some serotonergic antidepressants late in pregnancy, newborns may experience transient signs such as jitteriness, irritability, changes in tone, feeding difficulty, abnormal temperature regulation, or respiratory problems. Terminology varies across studies: neonatal adaptation syndrome, poor neonatal adaptation, delayed adaptation, and sometimes neonatal withdrawal. These labels describe overlapping observations, and biological mechanisms may include serotonergic effects as well as changes following delivery. Most reported cases are self-limited, while some infants need observation or respiratory support (MotherToBaby, Sertraline).
A 2024 population-based study of 280,090 term infants observed delayed neonatal adaptation in 11.2% of infants exposed to SSRIs after 20 weeks compared with 4.4% in unexposed infants. Its composite outcome included a low five-minute Apgar score, resuscitation at birth, or NICU admission for respiratory support. The adjusted odds ratio was 2.14 (95% CI 1.96–2.32), with variation by type and dose. Those figures refer to a defined hospital-based outcome in one study and should not be used as the universal frequency of mild jitteriness or all neonatal symptoms. Residual confounding remains possible even with extensive adjustment (Cornet et al., 2024).
The practical response is a coordinated delivery and newborn plan. The obstetric and pediatric teams should know the medication, dosage, other exposures, and anticipated monitoring needs. A trial of routinely tapering all antidepressants immediately before birth to avoid neonatal symptoms has no established benefit that overrides the potential maternal risks of relapse or discontinuation; an individualized plan is preferable. Parents should receive a clear explanation of what staff will assess and what follow-up is needed.
Persistent pulmonary hypertension of the newborn
PPHN is a serious but uncommon disturbance of circulation after birth. Some studies suggest an association with SSRI exposure later in pregnancy. A large US study reported PPHN in 20.8 per 10,000 infants without late antidepressant exposure and 31.5 per 10,000 with late SSRI exposure before full adjustment. After restricting analyses to mothers with depression and adjusting extensively, the overall association attenuated to an odds ratio of 1.10 (95% CI 0.94–1.29); one narrower outcome definition retained a modest association. The researchers concluded that any increased absolute risk was small and less pronounced than early reports suggested (Huybrechts et al., 2015).
An individual consultation should recognize both the seriousness of PPHN and the low absolute event frequency. Risk is influenced by other perinatal factors, and the observational evidence does not permit a simple personal risk calculation from a medication name alone. Neonatal respiratory distress requires immediate pediatric assessment whatever its cause.
Bleeding near delivery
Serotonergic antidepressants may impair platelet-mediated hemostasis, providing a plausible mechanism for increased bleeding. Obstetric guidelines and drug-safety communications discuss a possible modest increase in postpartum hemorrhage with SSRI or SNRI exposure near delivery. The magnitude differs across studies and populations; obstetric history, anemia, placenta-related conditions, uterine factors, anticoagulant use, and delivery circumstances contribute strongly to an individual's risk. The appropriate action is an obstetric bleeding-risk plan rather than abrupt unsupervised antidepressant cessation (NICE, CG192).
Long-term child development: autism, ADHD, and related outcomes
Observational studies initially reported higher frequencies of autism spectrum disorder and attention-deficit/hyperactivity disorder among children exposed prenatally to antidepressants. Such comparisons are particularly vulnerable to genetic and family-level confounding because psychiatric illness, medication treatment, and some developmental diagnoses are related to familial factors. Large studies using comparison groups of individuals with depression, high-dimensional adjustment, and sibling analyses have generally weakened or eliminated associations that were prominent in crude comparisons.
A 2022 study spanning more than three million pregnancies, including 145,702 exposed pregnancies, found no increased risk attributable to antidepressant use in its most extensively controlled analyses for autism, ADHD, or several other neurodevelopmental diagnoses. The sibling comparison for any neurodevelopmental disorder yielded a hazard ratio of 0.97 (95% CI 0.88–1.06). This is substantial reassurance about major outcomes while retaining the usual limitations of prescription-based observational data and long-term follow-up (Suarez et al., 2022).
The evidence therefore supports a carefully stated conclusion: a causal increase in autism or ADHD from commonly used antidepressants during pregnancy has not been demonstrated by the strongest available population studies. Scientific follow-up remains important for less-studied drugs, rare outcomes, and outcomes that routine diagnostic records capture poorly. Parents should receive developmental care and assessment based on a child's needs, rather than exposure alone being treated as a prediction of impairment.
How the evidence is produced—and why studies disagree
Randomized trials that deliberately assign pregnant people to stop an effective antidepressant or continue it for the purpose of studying rare fetal outcomes are difficult to conduct ethically and practically. Much of the pregnancy safety evidence therefore comes from registries, insurance claims, prescription records, clinical cohorts, and pooled observational studies. A prescription fill may not mean the medication was taken; exposure dates, dose changes, and adherence can be incomplete. Rare outcomes are sensitive to coding errors, small cell counts, and multiple comparisons.
Confounding by indication arises when the illness requiring treatment is itself associated with the measured outcome. Confounding by severity occurs when patients with the most serious illness are more likely to receive continued or high-dose medication. Family and genetic factors may affect psychiatric treatment and child development. Exposure timing can produce another distortion: a first-trimester analysis cannot be used automatically to infer third-trimester neonatal risk. Live birth–restricted cohorts may also miss pregnancies ending earlier, changing apparent associations.
Researchers address these problems with active comparators, restriction to individuals with a similar diagnosis, propensity scores, negative controls, sibling comparisons, and sensitivity analyses. Each method answers a somewhat different question and carries its own assumptions. A sibling comparison reduces confounding shared by siblings but can introduce bias from factors that differ between pregnancies. A high-quality clinical synthesis weighs convergent patterns, absolute frequencies, precision, biological plausibility, and consistency across designs. Risk is best labeled established, suggestive, limited, or unresolved for each outcome separately.
Which antidepressants are commonly considered?
Sertraline
Sertraline is a commonly used SSRI for depression and several anxiety disorders. It has extensive pregnancy exposure data and is frequently considered when choosing a first antidepressant in a pregnant patient with no history of successful treatment on another agent. Large observational samples have not identified a consistent, substantial increase in overall congenital anomalies attributable to sertraline after accounting for relevant confounders. Late-pregnancy neonatal adaptation remains a counseling topic. The current MotherToBaby review reports more than 25,000 exposed pregnancies and distinguishes conflicting crude malformation associations from studies comparing patients with similar mental health conditions (MotherToBaby, Sertraline).
Escitalopram and citalopram
Escitalopram and citalopram are SSRIs used for depression and anxiety. They are familiar options in perinatal practice, with relatively substantial reproductive evidence compared with many newer agents. Selection should consider prior response, tolerability, potential drug interactions, cardiac or other medical factors, and the trimester-specific evidence. Late-pregnancy neonatal adaptation is a relevant discussion for these medications as for other SSRIs. A patient who has maintained robust remission on one of them may derive meaningful benefit from treatment continuity (ACOG, 2023).
Fluoxetine
Fluoxetine has a long history of pregnancy use and a comparatively long elimination half-life. Its pharmacology may affect dose changes, drug transitions, and the persistence of neonatal exposure. The available reproductive data are substantial, though neonatal adaptation and other outcome-specific questions remain. Long-standing effective treatment can be a compelling reason not to switch purely on the basis of pregnancy status. A new prescription should be assessed according to the individual's symptoms, previous treatment, comorbid conditions, and preferences (NICE, CG192).
Paroxetine
Paroxetine is associated with clinically relevant discontinuation symptoms in some patients, and historical studies reported possible specific cardiac-malformation signals with first-trimester exposure. These considerations can favor an alternative for someone starting therapy, yet they do not create a universal directive to switch a patient who is stable. Abrupt cessation may be especially difficult. Any change requires an individual review of prior efficacy, the particular pregnancy stage, psychiatric relapse risk, and the available evidence (NICE, CG192; MotherToBaby, Paroxetine).
Venlafaxine and duloxetine
Venlafaxine and duloxetine are SNRIs. They may be used when clinically indicated, particularly when a patient previously responded to them or other treatments were ineffective. Their noradrenergic as well as serotonergic actions mean blood pressure, adverse reactions, and discontinuation symptoms may merit attention. As with SSRIs, data on neonatal adaptation and bleeding close to delivery should inform obstetric planning. Neither a drug-class label nor an average study estimate substitutes for medication-specific counseling (NICE, CG192).
Bupropion
Bupropion acts through different neurotransmitter pathways and is sometimes used for depression or smoking cessation. MotherToBaby's January 2026 review reports that available evidence does not suggest an overall increase in birth defects, while discussing limited earlier signals for possible cardiac defects and their methodological weaknesses. Its review also addresses uncertainty in miscarriage research and the importance of avoiding abrupt medication changes. Bupropion may be a reasonable consideration in selected circumstances, but patient history and the quality of indication-specific evidence are decisive (MotherToBaby, 2026).
Mirtazapine, tricyclic antidepressants, and other agents
Mirtazapine and older tricyclic antidepressants may be options for particular clinical profiles, such as documented prior response or inability to tolerate other agents. Choice depends on sedation, weight or appetite effects, cardiac considerations, drug interactions, overdose safety, and the individual reproductive dataset. Newer antidepressants and less common combinations typically have fewer high-quality pregnancy exposures available for evaluation. Limited data should lead to proportionate uncertainty and specialist consultation rather than an automatic inference of harm. Medication combinations also require specific assessment; using more than one agent may sometimes be necessary, but polypharmacy should have a clear rationale (NICE, CG192).
Is one antidepressant the safest during pregnancy?
A universal ranking is scientifically misleading. A medication with a comparatively large, reassuring reproductive evidence base can be a sensible starting point when initiating treatment, while the drug that previously achieved sustained remission may offer the best overall benefit for another patient. Drug safety is also outcome-specific: an agent can have reassuring malformation data and still be associated with temporary neonatal adaptation effects near delivery. Treatment acceptability, availability, interactions, dosage, and the consequences of switching form part of safety.
ACOG identifies SSRIs as first-line pharmacotherapy for perinatal depression and anxiety disorders when medication is appropriate and often names sertraline or escitalopram as reasonable first-line choices when there is no prior pharmacotherapy history. That recommendation sits alongside an emphasis on previously effective treatment and individualized shared decision-making (ACOG, 2023). The first decision is whether medication is indicated; the second is which particular medication and follow-up plan best fit the clinical situation.
Continue, start, switch, or taper: four decision pathways
Continuing an antidepressant that is working
Continuation may be appropriate when a patient has recurrent or severe illness, previously relapsed after medication withdrawal, experiences significant impairment without treatment, or has only achieved remission on a particular drug. Clinicians should review the exact medicine, diagnosis, dose, comorbid medications, residual symptoms, pregnancy course, and side effects, then document why the anticipated benefits justify exposure. Continuing treatment is an active clinical decision, with monitoring and opportunity for revision. The goal is sustained remission and functioning, not exposure minimization at the expense of disease control (ACOG, 2023).
Starting treatment during pregnancy
New-onset depression warrants clinical assessment of symptom severity, function, suicidal thoughts, bipolar history, and contributing medical or psychosocial factors. For many patients with mild-to-moderate symptoms, psychotherapy is a suitable starting treatment when accessible and clinically appropriate. For moderate-to-severe episodes, substantial impairment, high recurrence risk, or inadequate response to psychotherapy, antidepressant treatment can be indicated. The patient should understand the expected timeline of symptom improvement, common adverse reactions, what to monitor, and whom to contact if symptoms worsen. NICE includes high-intensity psychotherapy, an antidepressant, and combined care among options according to severity and preferences (NICE, CG192).
Switching from an effective medicine
Switching has potential advantages when the current agent has a concerning, well-supported reproductive risk or poor tolerability. It also carries risks of withdrawal, pharmacologic interactions, loss of remission, side effects from the replacement, and a period of inadequate treatment. Many first-trimester exposures have already occurred by the time pregnancy is recognized, and a switch then may add a second medication exposure without erasing the first. Switching requires a clinician-designed transition and symptom monitoring; it should not be inferred from a generic list of preferred SSRIs (NICE, CG192).
Gradual discontinuation when clinically appropriate
Discontinuation may be considered after an individualized evaluation, especially for patients with milder, sustained remission who wish to prioritize a psychotherapy-based plan and can access close follow-up. Tapering is generally safer than abrupt cessation for reducing discontinuation symptoms, but tapering alone does not prevent recurrence of the underlying disorder. The specific approach depends on medication pharmacology, dose, duration of treatment, previous withdrawal experiences, and psychiatric history. A patient who elects to stop needs an explicit relapse-response plan and a way to restart care rapidly. In severe or recurrent illness, the consequences of stopping may outweigh the anticipated exposure reduction (NICE, CG192; Cohen et al., 2006).
Dose, pharmacokinetics, and monitoring
Pregnancy changes blood volume, kidney function, gastrointestinal physiology, hepatic metabolism, and protein binding. These changes can affect concentrations of some medications, though the direction and magnitude differ by drug and individual. Dose selection should target meaningful symptom improvement or remission while avoiding unnecessary adverse effects. The lowest effective dose is a clinical goal; an ineffective dose offers exposure without adequate benefit. Automatic dose reduction solely because of pregnancy can therefore defeat the treatment's purpose (NICE, CG192).
Follow-up should assess depressive or anxiety symptoms, sleep, functioning, adverse effects, adherence, blood pressure when relevant, and emerging safety concerns. Validated measures such as the PHQ-9, EPDS, and GAD-7 can track symptom change, supplemented by a clinical interview. A screening result is one piece of information; the clinician evaluates diagnostic criteria, bipolar features, obstetric context, and safety. Medication reviews should cover over-the-counter agents, herbal products, alcohol and other substances, and interacting prescriptions. Suspected serotonin toxicity, severe allergic reactions, or sudden behavioral deterioration require urgent assessment.
Psychotherapy and nonmedication treatment
Evidence-based psychotherapy is an important treatment in its own right and a useful partner to medication. Cognitive behavioral therapy and interpersonal psychotherapy can address depressive symptoms, rumination, avoidance, relationship strain, practical stressors, and relapse prevention. Availability matters: a nominal therapy referral is not equivalent to a treatment that starts promptly, is culturally appropriate, and can be continued. Some patients need coordinated social assistance, housing or financial support, trauma-informed obstetric care, or treatment for insomnia and chronic pain. These factors influence clinical outcomes without replacing care for a diagnosed disorder.
For patients with significant illness, combined pharmacotherapy and psychotherapy may be preferable to an either-or framing. Clinical decisions should include patient experience, previous treatment response, symptom burden, and access barriers. Mindfulness, exercise, sleep support, and peer connection may aid well-being but should not be represented as substitutes for clinically indicated treatment of severe depression or imminent suicide risk (NICE, CG192).
Screening, diagnosis, and treatment response
Routine perinatal screening can help identify distress and possible mood or anxiety disorders. PHQ-9 and EPDS assess depressive symptom burden; GAD-7 can help characterize anxiety. Positive scores need follow-up questions about onset, persistence, functional impairment, pregnancy-related physical symptoms, trauma, substances, thyroid and other medical issues, and prior psychiatric episodes. A person can have clinically important distress without meeting full criteria for a disorder and still need support. Equally, someone with a low score but a convincing history of bipolar episodes, psychosis, or acute safety concerns needs individualized clinical attention (ACOG, 2023, Screening and Diagnosis).
Response should be measured over time rather than inferred from a single reassuring visit. Clinicians consider symptom remission, recovery of functioning, ongoing anxiety, adherence, tolerability, sleep, and the patient's perception of treatment. Plans should specify contact frequency, thresholds for modifying treatment, access to urgent support, and postpartum review. The period after childbirth deserves proactive planning because sleep disruption, rapid physiological change, and caregiving demands can coincide with symptom exacerbation.
Safety-critical conditions: suicidality, mania, and postpartum psychosis
Thoughts of suicide, a current plan or intent to self-harm, psychosis, severe confusion, or dangerous behavior require immediate clinical assessment. A patient with imminent risk should seek emergency medical services or the nearest emergency department and should not be left without appropriate safety support. A crisis response includes assessment, protection, and rapid psychiatric or obstetric coordination; routine antidepressant follow-up is insufficient. Questions about self-harm should be direct, compassionate, and repeated when symptoms change (ACOG, 2023, Screening and Diagnosis).
A history suggestive of bipolar disorder changes the risk–benefit calculation. Mania may present with markedly reduced need for sleep, elevated or irritable mood, agitation, grandiosity, impulsivity, and impaired judgment. Postpartum psychosis can emerge rapidly, sometimes with delusions, hallucinations, confusion, extreme insomnia, and fluctuating behavior; it is a psychiatric emergency requiring urgent specialist evaluation and commonly inpatient care. Antidepressant monotherapy is not a general treatment pathway for bipolar I disorder or postpartum psychosis. Care planning should identify these conditions before and throughout treatment (ACOG, 2023, Treatment and Management; NICE, CG192).
Planning birth and postpartum care
A pregnancy treatment plan should extend through delivery. The obstetric team needs an accurate medication list; neonatal clinicians need advance notice when late exposure creates possible adaptation needs; psychiatric and primary-care clinicians need a plan for continued prescribing and monitoring. Consider postpartum depression history, bipolar risk, anticipated sleep loss, available support, feeding preferences, and access to follow-up. Coordinated care minimizes abrupt handoffs and gaps in treatment.
Breastfeeding presents a separate pharmacokinetic question because infant exposure through milk differs from placental exposure during gestation. It requires evidence about milk transfer, infant age and health, maternal medication response, and feeding goals. A decision about pregnancy medication should anticipate this discussion without treating lactation risks as interchangeable with fetal risks. Newborn observation after late exposure, support for feeding, and rapid assessment of significant respiratory or behavioral symptoms belong in the delivery plan (NICE, CG192).
Equity, access, and the ethics of consent
Perinatal medication decisions occur in systems with unequal access to specialist psychiatry, psychotherapy, adequate obstetric care, insurance coverage, paid leave, and childcare. Some patients encounter stigma for taking medicine; others encounter pressure to medicate when their preference is for a supported psychological approach. Informed consent requires understandable absolute risks, a realistic explanation of alternatives, accommodation of literacy and language needs, and attention to the patient's values. Family members can participate with the patient's agreement. Patient autonomy remains central to the decision (NICE, CG192).
Clinical discussion should avoid blame about infant outcomes or maternal symptoms. The aim is informed care for two interdependent health trajectories, supported by reliable information and practical follow-through. A decision made under uncertainty can still be rational and medically sound when the evidence, priorities, monitoring, and contingency plans are explicit.
A structured conversation with the treating team
A practical consultation begins by establishing the diagnosis, severity, past episodes, history of mania or psychosis, current symptoms, and any immediate safety issues. The patient and clinicians then review the exact medication and dose, how long it has worked, previous treatment failures, withdrawal history, and other exposures. They consider pregnancy stage, congenital-anomaly data, neonatal adaptation, bleeding and other relevant outcomes, along with the risks of relapse and undertreatment. Next comes a realistic assessment of psychotherapy access, preferred treatment, obstetric follow-up, and support after childbirth.
Before the visit ends, the team should document the selected pathway and the reasons for it, who will prescribe, when symptoms and side effects will be reviewed, what would prompt a change, and whom to contact urgently. The discussion can be revisited when evidence or circumstances change. For patients who recently learned they are pregnant while taking an antidepressant, the immediate practical step is to contact their prescribing and obstetric clinicians for a planned review rather than suddenly stopping treatment (ACOG, 2023; NICE, CG192).
Frequently asked questions
Should I stop my antidepressant as soon as I get a positive pregnancy test?
A positive test is a reason for timely clinician review, not automatic abrupt discontinuation. A sudden stop can cause discontinuation symptoms and may allow depression or anxiety to return. The optimal decision depends on diagnosis, recurrence history, effective treatment, gestational stage, and individual preferences. ACOG recommends against discontinuing psychiatric medication because of pregnancy alone (ACOG assessment guidance).
Is sertraline safe during pregnancy?
Sertraline has extensive reproductive data, and most studies do not identify a substantial increase in overall birth-defect risk attributable to the medication. Possible late-pregnancy neonatal adaptation effects still warrant counseling and delivery planning. The clinically relevant question is whether sertraline is a good choice for this patient and indication, compared with effective alternatives (MotherToBaby, Sertraline).
Can antidepressants cause birth defects?
Individual studies have reported signals for specific drugs and anomalies, but large analyses that account for psychiatric illness often find much smaller associations than unadjusted comparisons. Risk cannot be summarized accurately for the entire class with a single number. Drug, timing, outcome, and the quality of adjustment all matter (Huybrechts et al., 2014).
Do antidepressants cause autism or ADHD?
The strongest large observational analyses with extensive control of familial and clinical confounding do not support a causal increase in autism or ADHD associated with commonly used prenatal antidepressants. Ongoing research is appropriate, particularly for less-studied agents and outcomes (Suarez et al., 2022).
Is untreated depression always more dangerous than medication?
There is no universal numerical comparison that applies to every patient. Severe, recurrent, or suicidal depression can carry major immediate risk, and effective treatment may be crucial. Mild, sustained remission may allow other options with appropriate monitoring. The balance is individualized and should be reviewed as pregnancy progresses (NICE, CG192).
Should the dose be lowered in the third trimester?
A routine reduction without clinical reason may risk relapse and may not reliably prevent neonatal problems. Dose decisions should consider symptoms, tolerability, medication pharmacology, gestational changes, and the anticipated newborn care plan. A treating clinician should lead any adjustment (NICE, CG192).
What if I took an antidepressant before realizing I was pregnant?
Many pregnancies are recognized after some early exposure has already occurred. The useful next step is a medication-specific review of timing, dose, evidence, and standard prenatal assessment. Abrupt discontinuation cannot reverse past exposure and may worsen symptoms. Individual teratology counseling can help interpret a particular exposure (NICE, CG192).
Will my newborn need intensive care?
Most antidepressant-exposed newborns do not develop severe complications, but some experience transient adaptation difficulties, and the frequency depends on the study definition and exposure. The delivery team should know about late-pregnancy antidepressant treatment and arrange monitoring if indicated. A need for respiratory support or other clinical care is determined by the newborn's actual condition (Cornet et al., 2024).
Can psychotherapy replace antidepressants?
Psychotherapy can be a primary evidence-based treatment for selected depressive and anxiety presentations, especially when symptoms are less severe and treatment is accessible. Patients with severe or recurrent illness may need medication, sometimes combined with psychotherapy. A careful choice considers previous response and the consequences of inadequate treatment (NICE, CG192).
Clinical synthesis
Evidence-based decisions about antidepressants during pregnancy rest on the combination of psychiatric prognosis, maternal functioning, outcome-specific reproductive research, effective treatment history, and the patient's informed preference. Large studies offer meaningful reassurance about overall malformation risk and long-term neurodevelopment for commonly used SSRIs, while supporting attention to neonatal adaptation and selected obstetric outcomes. The certainty of each conclusion varies by medication and outcome. A treatment decision is strongest when it specifies the comparison, the absolute risk where known, the plan for monitoring, and the response to emerging symptoms. Effective maternal mental health care is an integral part of pregnancy care (ACOG, 2023).
Related Articles
Cluster Perinatal and Reproductive Mental Health — the central scientific overview of pregnancy, postpartum mental health, reproductive transitions, screening, treatment, and related research.
Perinatal OCD: What Is It? OCD During Pregnancy and Postpartum, Symptoms, Risk, and Treatment — the OCD-specific clinical pathway, including assessment of intrusive thoughts and treatment selection around pregnancy.
Eating Disorders During Pregnancy and Postpartum: Care and Risks — reproductive-context management of eating disorder symptoms, nutrition, psychological risk, and care coordination.
References
American College of Obstetricians and Gynecologists. (2023). Screening and diagnosis of mental health conditions during pregnancy and postpartum. Clinical Practice Guideline No. 4. Official guideline.
American College of Obstetricians and Gynecologists. (2023). Treatment and management of mental health conditions during pregnancy and postpartum. Clinical Practice Guideline No. 5. Official guideline.
American College of Obstetricians and Gynecologists. (n.d.). Assessment and treatment of perinatal mental health conditions. Official clinical resource.
Cohen, L. S., Altshuler, L. L., Harlow, B. L., et al. (2006). Relapse of major depression during pregnancy in women who maintain or discontinue antidepressant treatment. JAMA, 295(5), 499–507. https://doi.org/10.1001/jama.295.5.499.
Cornet, M.-C., et al. (2024). Maternal treatment with selective serotonin reuptake inhibitors during pregnancy and delayed neonatal adaptation: A population-based cohort study. Archives of Disease in Childhood – Fetal and Neonatal Edition, 109, 294–300. PubMed PMID 38071585.
Huybrechts, K. F., Bateman, B. T., Palmsten, K., et al. (2015). Antidepressant use late in pregnancy and risk of persistent pulmonary hypertension of the newborn. JAMA, 313(21), 2142–2151. https://doi.org/10.1001/jama.2015.5605.
Huybrechts, K. F., Palmsten, K., Avorn, J., et al. (2014). Antidepressant use in pregnancy and the risk of cardiac defects. New England Journal of Medicine, 370, 2397–2407. https://doi.org/10.1056/NEJMoa1312828.
MotherToBaby. (2026). Bupropion (Wellbutrin): Pregnancy and breastfeeding exposures. Official fact sheet.
MotherToBaby. (2025). Paroxetine (Paxil): Selected references and safety evidence. Official resource.
MotherToBaby. (2025). Sertraline (Zoloft): Pregnancy and breastfeeding exposures. NCBI Bookshelf.
National Institute for Health and Care Excellence. (2014; updated). Antenatal and postnatal mental health: Clinical management and service guidance (CG192). Official recommendations.
Suarez, E. A., Bateman, B. T., Hernández-Díaz, S., et al. (2022). Association of antidepressant use during pregnancy with risk of neurodevelopmental disorders in children. JAMA Internal Medicine, 182(11), 1149–1160. https://doi.org/10.1001/jamainternmed.2022.4268.
Sujan, A. C., Rickert, M. E., Öberg, A. S., et al. (2017). Associations of maternal antidepressant use during the first trimester of pregnancy with preterm birth, small for gestational age, autism spectrum disorder, and attention-deficit/hyperactivity disorder in offspring. JAMA. Official journal publication.