Antidepressants in Pregnancy And The Risk To Offspring: What This Meta-Analysis Does — And Does Not — Show
Why This Question Needs a Careful Reading
Few medication questions are as emotionally charged as antidepressant use during pregnancy. The patient is not only thinking about symptom control. She may be thinking about fetal development, future child health, stigma, guilt, family pressure, and the fear of making an irreversible decision. A headline suggesting a link between antidepressants and autism or ADHD can therefore have consequences far beyond academic debate. This is also a field where simple claims are unusually risky. Saying “antidepressants cause neurodevelopmental disorders” is not supported by the best current evidence. Saying “there is no risk to discuss” is also too casual. Pregnancy care always involves weighing risks: medication exposure, untreated or relapsing psychiatric illness, obstetric factors, past illness severity, relapse history, previous response to treatment, and the patient’s own values.
The Lancet Psychiatry systematic review and meta-analysis on maternal and paternal antidepressant use before and during pregnancy is important because it tries to address the central problem that has complicated this literature for years: confounding (1). Women who take antidepressants during pregnancy differ from women who do not in many ways that can also be associated with child neurodevelopment. Depression, anxiety, genetic liability, family environment, sleep disruption, smoking, substance use, health care contact, socioeconomic stress, and illness severity can all shape risk or diagnosis.
The most useful reading of this meta-analysis is therefore balanced. It is broadly reassuring about commonly used antidepressants and offspring ADHD or autism risk after accounting for confounding. It does not make individual pregnancy decisions automatic. It reduces one major fear, but it does not remove the need for individualized psychiatric and obstetric care.
What The Meta-Analysis Actually Studied
The authors conducted a systematic review and meta-analysis of studies examining maternal or paternal antidepressant use before or during pregnancy and neurodevelopmental outcomes in offspring. They searched Embase, MEDLINE, PsycINFO, and Web of Science from database inception to May 14, 2025.
The review included 37 studies, covering 648,626 antidepressant-exposed pregnancies and 24,967,806 unexposed pregnancies. Outcomes included neurodevelopmental disorders overall, attention-deficit/hyperactivity disorder, autism spectrum disorder, intellectual disabilities, motor disorders, and speech and language disorders.
This is a large evidence synthesis, but size alone does not solve the problem. The included studies were observational. They can identify associations, but they cannot randomly assign pregnant patients to untreated illness, medication continuation, or medication discontinuation. That means causal interpretation depends on how well studies account for confounding.
The meta-analysis is especially valuable because it did not only compare exposed and unexposed pregnancies. It also examined maternal antidepressant use before pregnancy, paternal antidepressant use around conception or during pregnancy, antidepressant classes, specific agents, dose, and analyses that attempted to reduce confounding by indication.
Paternal exposure is particularly useful as a negative-control comparison. A father’s antidepressant use during a partner’s pregnancy does not expose the fetus to the medication through the placenta. If paternal antidepressant use is also associated with ADHD or autism in offspring, that pattern points away from a simple intrauterine drug effect and toward shared genetic, familial, psychiatric, or environmental factors.
Prepregnancy maternal exposure works similarly. If a mother’s antidepressant use before pregnancy is associated with later child neurodevelopmental diagnosis, even when the exposure did not occur during fetal development, the association may reflect underlying maternal illness or familial liability rather than medication exposure during pregnancy.
These comparison strategies do not make the evidence perfect. They do, however, make the causal question more sophisticated than a basic “exposed versus unexposed” contrast.
The Initial Signal: Small Associations before Full Adjustment
In the broad pooled analysis, prenatal antidepressant use was associated with a modestly increased risk of neurodevelopmental disorders overall. The association was seen for ADHD and autism spectrum disorder, while no clear association was found for intellectual disabilities, motor disorders, or speech and language disorders. These initial numbers are the kind of findings that often generate alarming headlines. A relative risk can sound frightening when presented without context. For example, a pooled association with ADHD or autism does not automatically mean that antidepressants caused those diagnoses. It also does not tell an individual patient what will happen in her pregnancy.
The key issue is that the exposed and unexposed groups are not equivalent. Patients who use antidepressants during pregnancy may have more severe or recurrent depression, anxiety disorders, trauma histories, psychiatric comorbidity, sleep disturbance, substance-use risk, or family histories of neurodevelopmental conditions. They may also have more contact with health care systems, which can increase the likelihood that developmental concerns in children are detected and diagnosed.
This is confounding by indication: the reason a medication is prescribed may itself be associated with the outcome being studied. In perinatal psychiatry, it is one of the central interpretive problems. The medication and the underlying condition travel together, and separating their effects is difficult. That is why the initial association should be treated as a signal requiring analysis, not as proof of harm. It tells researchers where to look. It does not answer the causal question by itself.
What Happens after Accounting for Confounding
The most important finding of the meta-analysis is that the observed associations were attenuated or became statistically non-significant in analyses that accounted for confounding factors, including maternal psychiatric disorders, familial or genetic influences, and exposure misclassification.
This is the reassuring core of the paper. The initial associations with ADHD and autism appear to be largely explained by factors surrounding antidepressant use rather than by a direct medication effect on fetal neurodevelopment. Put differently, the risk signal becomes weaker when the comparison becomes fairer. Paternal exposure strengthens this interpretation. In the meta-analysis, paternal antidepressant use during pregnancy was associated with increased risk of ADHD and autism in offspring. That pattern is difficult to explain through direct fetal drug exposure. It is more consistent with shared genetic vulnerability, parental mental health, family environment, or other correlated factors.
Preconception exposure points in the same direction. Similar associations were found for antidepressant use before pregnancy. If a risk signal appears when exposure occurs before fetal development, then the signal cannot be attributed simply to medication acting on the fetus during gestation.
The results by antidepressant class and agent also require careful reading. Commonly used SSRIs and SNRIs did not show significant associations after confounding by indication was minimized. Signals were reported for some older tricyclic antidepressants, including amitriptyline and nortriptyline, but those findings should be interpreted cautiously. Such agents may be used in different clinical contexts, including more severe or complex illness, pain syndromes, insomnia, or treatment histories that differ from typical SSRI/SNRI use. The study itself rated the certainty of evidence as low to very low. Dose findings also did not support a simple causal story. Available analyses did not show a significant difference in autism risk between high-dose and low-dose exposure, although dose data were limited. If a medication were a strong causal driver of neurodevelopmental diagnosis, a clearer dose-response relationship would strengthen that claim. Its absence is reassuring but not definitive.
The best summary is precise: the meta-analysis does not support a causal association between commonly used antidepressants during pregnancy and offspring ADHD or autism after accounting for major confounding factors. It does not prove that every antidepressant, every dose, every timing pattern, and every patient circumstance is risk-free.
What The Findings Do Not Mean For Pregnant Patients
This study should not be used to tell every pregnant patient to continue antidepressants automatically. It should also not be used to frighten patients into stopping them. Both reactions miss the point.
The meta-analysis focused on neurodevelopmental disorders in offspring. It did not answer every question about antidepressant safety in pregnancy. Other outcomes remain part of clinical decision-making: miscarriage, preterm birth, fetal growth, neonatal adaptation syndrome, postpartum hemorrhage, persistent pulmonary hypertension of the newborn, breastfeeding considerations, and postpartum maternal relapse. These risks differ by medication, dose, timing, illness severity, and patient history. Nor does the paper erase the risks of untreated or relapsing depression and anxiety. For some patients, discontinuing medication during pregnancy may be reasonable, especially if symptoms have been mild, stable, and well supported by psychotherapy or other non-pharmacological care. But for patients with recurrent, moderate-to-severe, psychotic, suicidal, or functionally impairing illness, stopping an effective antidepressant can carry serious risk.
Maternal relapse is not a benign event. It can affect sleep, nutrition, prenatal care, substance use, medical adherence, relationships, bonding, occupational functioning, and suicide risk. It can also increase the risk of postpartum depression, which has consequences for both parent and child. The decision is therefore not “medication risk versus no risk.” It is medication risk versus illness risk, with uncertainty on both sides.
The findings are most reassuring for patients taking commonly used antidepressants, especially SSRIs or SNRIs, who are worried specifically about ADHD or autism risk in offspring. The evidence suggests that earlier associations were probably influenced substantially by familial and clinical factors. That should reduce fear-driven discontinuation.
But reassurance should not become dismissal. A patient who is pregnant or planning pregnancy still needs an individualized medication review. That review should consider diagnosis, relapse history, previous attempts to stop medication, dose, side effects, comorbid anxiety or bipolarity, suicide history, obstetric risk, prior postpartum episodes, available psychotherapy, social support, and patient preference.
A calm clinical message would be this: do not stop antidepressants abruptly because of a headline. Also do not assume that every treatment plan must stay unchanged. Discuss the decision with clinicians who understand both perinatal psychiatry and obstetric risk.
Why Communication around This Evidence Matters
This topic is highly vulnerable to distorted communication. One exaggerated headline can make a patient feel that she has harmed her child by treating her depression. Another can lead someone to stop medication suddenly, without tapering, monitoring, or a relapse plan.
Abrupt discontinuation is not a neutral choice. It can cause withdrawal symptoms in some patients and can increase relapse risk, especially in those with recurrent or severe illness. In pregnancy, that relapse can occur at a time when access to timely psychiatric care may already be limited. The language used to describe this evidence therefore matters. It is better to say, “Initial associations were largely attenuated after accounting for confounding,” than to say, “Earlier studies were wrong.” It is better to say, “The evidence does not support a causal link for commonly used antidepressants and ADHD or autism,” than to say, “Antidepressants are proven harmless.”
Neurodevelopmental disorders are multifactorial. ADHD and autism are influenced by genetic liability, prenatal factors, perinatal factors, family environment, detection patterns, and broader social determinants. A single medication exposure rarely explains such outcomes by itself. The paternal and preconception findings in this meta-analysis are useful because they remind readers that family-level factors can create associations that look medication-related at first glance.
Good communication should also avoid moral pressure. Some patients need antidepressants during pregnancy. Some prefer and can safely use non-pharmacological approaches. Some may reduce dose or switch medication before conception. Some should not attempt discontinuation because the relapse risk is too high. The ethical task is not to push one universal decision. It is to help patients make informed choices without fear, shame, or oversimplified causal claims.
The Practical Takeaway
The Lancet Psychiatry meta-analysis is reassuring, especially for patients worried that commonly used antidepressants during pregnancy might directly cause autism or ADHD in offspring. The strongest interpretation is that the initial associations in observational studies are substantially explained by parental mental health, shared genetics, familial environment, and other confounding factors. At the same time, the paper does not make pregnancy medication decisions simple. It does not cover every safety outcome. It does not prove that every antidepressant exposure is risk-free. It does not eliminate the need to consider illness severity, relapse history, treatment response, dose, timing, and patient values.
For clinicians, the study supports more balanced counseling. For patients, it should reduce fear-driven discontinuation. For journalists and public-health communicators, it is a reminder that association is not causation, especially in perinatal medication research.
The practical message is not “take antidepressants no matter what” or “avoid antidepressants at all costs.” It is more careful: for many patients with moderate-to-severe depression or anxiety, continuing an effective antidepressant during pregnancy may be clinically appropriate, and the current evidence does not support the claim that common antidepressants independently cause ADHD or autism. The decision should be made with psychiatric and obstetric care teams, not by reacting to headlines.
References
- Chan, J. K. N., Zhong, A. H. F., Lam, J. Y. H., Wong, C. S. M., Solmi, M., Correll, C. U., & Chang, W. C. (2026). Maternal and paternal antidepressant use before and during pregnancy and offspring risk of neurodevelopmental disorders: A systematic review and meta-analysis. The Lancet Psychiatry, 13(6), 472–484. https://doi.org/10.1016/S2215-0366(26)00089-1
- A Practical, Plain-English Guide to ADHD Medications (2025, y the Shanghai Archives of Psychiatry blog team)