Does Antidepressant Use During Pregnancy Cause Autism or ADHD? A Huge Analysis of 25 Million Pregnancies

A Meta-Analysis Covering 25 Million Pregnancies Disproves a Fear That Has Persisted for Years

37 Studies, 25 Million Pregnancies and a Decade of Methodological Progress

Previous meta-analyses on this topic were conducted approximately a decade ago and had important limitations: few studies and insufficient control for "confounding by indication." The new meta-analysis has both a much larger data pool and a much more advanced statistical adjustment approach:

Alarming Numbers, Then an Association That Disappears

The essence of the study lies in comparing raw, unadjusted data with adjusted data:

At first glance, RR 1,35 and RR 1,69 look concerning, especially a 69% higher autism risk, which would make a dramatic headline. But the strength of the study lies in explaining why these numbers arose. Once maternal mental health and familial and genetic influences were accounted for, these associations weakened or lost statistical significance. The most commonly used antidepressants, particularly selective serotonin reuptake inhibitors (SSRIs) , were not found to be associated with autism or ADHD after adjustment.

Three Indirect Pieces of Evidence Proving the Drug Is Innocent

The cleverest aspect of the study is the three "negative control" comparisons used to separate the drug's direct fetal effect from other explanations. If the increased risk really arose from the drug's fetal effect during pregnancy, the risk should not have been seen in these three scenarios. But it was seen, showing that other underlying factors, rather than the drug, are responsible:

Let's unpack this logic. If a drug truly harms the fetus: (1) the father's drug use should not affect the child, because the fetus is not exposed to the father's drug; yet the risk was still observed. (2) The mother's use before pregnancy followed by discontinuation should not affect the child, because the fetus was not exposed to the drug during pregnancy; yet the risk was still observed. (3) If there were a genuine drug effect, a high dose should have created more risk than a low dose, a dose-response relationship being a classic indicator of causality; but no significant difference was found. All three negative controls point to the same conclusion: the drug is not responsible; the genetic and environmental background shared by the family is.

Amitriptyline and Nortriptyline: the Only Remaining Signal

The study does not simply say that "every antidepressant is innocent"; there is a nuance. Even after adjustment, among the older-generation tricyclic antidepressants, amitriptyline and nortriptylineremained associated with elevated risk in some analyses: amitriptyline with ADHD, RR 1,74, and amitriptyline and nortriptyline together with autism, RR 2,02.

However, the authors also give a careful warning here: these drugs are usually prescribed for more severe or treatment-resistant depression as their indication. In other words, even this remaining association may arise from a more severe underlying mental health condition that requires the drug, rather than from the drug itself: again, confounding by indication. Nevertheless, the authors, taking a cautious approach, imply that when possible care should be taken to avoid recommending these two specific tricyclics during pregnancy.

In an independent commentary accompanying the study, Lisa Vitte, Emmanuel Devouche and Gisèle Apter from the University of Rouen in Normandy offered this assessment: the study confirms and expands current knowledge about antidepressant use during pregnancy; these drugs should continue to be used because they protect maternal mental health and do not harm fetal development. The commentators described this as having "a noteworthy impact after many conflicting and controversial studies."

An honest reading must also state the limitations. The studies lacked data on some important variables such as socioeconomic status, lifestyle risk factors and low birth weight. Few studies examined antidepressant use in specific trimesters or at precise doses/dose changes. In addition, women prescribed antidepressants generally have more severe depression than those not prescribed them, so some residual bias may remain even after controlling for mental health status. These limitations do not change the direction of the findings, but they limit their degree of certainty; indeed, the certainty of evidence in the raw analysis was graded "very low."

This article is intended for medical news/information and does not provide individual treatment guidance. The use, initiation, dose adjustment or discontinuation of antidepressants must be carried out under the supervision of a physician, a psychiatrist and/or an obstetrician-gynecologist. No one taking antidepressants during pregnancy or while planning pregnancy should stop or change their medication on their own based on this article. As the study also emphasizes, abruptly stopping treatment is a risk in itself because of the risk of relapse and the serious consequences of untreated depression. This meta-analysis synthesizes observational studies; it provides strong evidence but is not absolute proof of causality. The association remaining in some analyses for tricyclic antidepressants, amitriptyline and nortriptyline, is a nuance to discuss with your physician. People concerned about mental health during pregnancy should consult a healthcare professional for both their own health and their baby's health. This article has sought to use neutral, nonstigmatizing language about childhood neurodevelopmental disorders, autism and ADHD. All data have been verified from the original Lancet Psychiatry publication, Chan et al., 14 May 2026, and related institutional statements.