A systematic review and meta-analysis published Thursday 14 May 2026 in The Lancet Psychiatry offers what accompanying news coverage calls the strongest synthesis yet on whether antidepressants taken shortly before or during pregnancy cause autism spectrum conditions or attention-deficit/hyperactivity disorder (ADHD) in offspring. Pulling together 37 prior studies covering more than 600,000 pregnancies with antidepressant exposure and roughly 25 million without, the authors report that crude associations largely attenuate once confounders—especially familial psychiatric burden and indication for treatment—are modelled transparently.
What the unadjusted numbers looked like
According to CNN’s reading of the paper, maternal antidepressant use during pregnancy initially tracked with a 35% higher odds of childhood ADHD and a 69% higher odds of autism without intellectual disability. Paternal use during the partner’s pregnancy showed a similar impossible-to-ignore pattern—46% for ADHD and 28% for autism—a clue that ambient genetics and shared environment, not placental drug transfer alone, were doing much of the explanatory work.
What happened after adjustment
Once the research team layered in covariates capturing parental psychopathology, family history, and related social determinants, most effect sizes shrank toward the null; CNN cites an autism signal of roughly 15% elevated odds after adjustment—still a fractional change that epidemiologists will debate for causal interpretation versus residual confounding. For SSRIs, the dominant first-line class, the article characterises the post-adjustment picture as essentially reassuring, aligning with ACOG-style guidance that untreated perinatal depression carries concrete obstetric and neonatal harms.
Why co-senior author Wing Chung Chang stresses vulnerability, not pills
Dr Wing Chung Chang of the University of Hong Kong told CNN by email that the dose–response pattern “strongly suggests” elevated neurodevelopmental diagnoses in exposed families “are largely driven by the underlying vulnerability” rather than pharmacology itself. He couples that statistical point with a clinical one: more than one in ten pregnant people worldwide experience depression, per WHO figures cited in the same reporting—so policymakers should avoid panic labeling that pushes cold-turkey tapers without supervision.
Policy echo chamber: FDA SSRI hearings in 2025
CNN notes the July 2025 FDA expert panel that floated tougher SSRI pregnancy label language on autism, miscarriage, and malformations—hearings APA, ACOG, and others criticised as “alarmingly unbalanced” for underweighting maternal morbidity when medicines are withheld. The Lancet Psychiatry drop therefore lands in a charged regulatory moment: evidence tables will matter for whether Risk Evaluation and Mitigation Strategies expand or whether Black Box revisions stay evidence-tied.
Nuance: tricyclics and ADHD signal
CNN relays commentary from Dr Jonathan Alpert (Montefiore) highlighting that, even after adjustment, amitriptyline and nortriptyline—older tricyclic antidepressants with anticholinergic load—retained a statistical link to childhood ADHD. The commentary stresses unknown causality versus indication severity (clinicians reach for TCAs when illness is often more refractory). Pending prospective pharmacoepidemiology, the practical message is individualised titration, not class-wide ban.
Motor, speech, and intellectual disability endpoints
The review reportedly found no meaningful association between antidepressant classes and motor disorders, speech and language delays, or intellectual disability—broadening reassurance beyond the headline ADHD/autism dyad. Dose stratification remained underpowered because contributing cohorts rarely logged milligram-day metrics, a limitation Chang flagged for future IPD meta-analyses.
Fathers on the radar
The manuscript explicitly argues paternal mental health is an understudied vector for offspring neurodevelopment, CNN notes—pushing perinatal psychiatry toward “a field of three” (mother, infant, father). That framing matters for employer parental leave policy and for insurance parity laws covering couples therapy alongside medication management.
Dose, duration, and design limits
The CNN summary notes that most source cohorts lacked granular dose or trimester-limited exposure windows, preventing the meta-analysis from adjudicating whether first-trimester-only use differs from continuous treatment through delivery. Because every included study is observational, Mendelian randomisation or sibling-comparison designs—better shields against shared confounding—remain complementary projects rather than substitutes. Readers should interpret pooled odds ratios as association metrics, not bench toxicology.
Clinical bottom line for shared decision-making
Neither Newsorga nor lay summaries replace obstetric–psychiatric co-management. The Lancet Psychiatry evidence, as relayed by CNN, supports the current consensus: continue SSRIs when symptoms are moderate-to-severe, taper slowly if discontinuing, and document reasoning in the chart. Neonatology should still watch for neonatal adaptation syndrome with late-third-trimester serotonergic exposure—an orthogonal but real pharmacology issue distinct from autism risk.
What research must do next
- Trimester-specific exposure windows tied to fetal MRI substudies.
- Polygenic risk scores overlaid on Nordic registry pedigree data.
- Breastfeeding pharmacokinetics for sertraline versus venlafaxine PK curves.
- Health equity audits so Black and Indigenous birthing people—often under-treated—are not left out of precision messaging.
Bottom line
The May 2026 Lancet Psychiatry mega-review does not support a causal story that routine antidepressants in pregnancy spark autism or ADHD; apparent risks largely mirror family psychiatric load. SSRIs emerge comparatively clean, while TCA signals deserve nuanced follow-up. Clinicians should pair these data with compassionate counselling about relapse risk—because maternal survival matters too.