Anxiety and insomnia grip roughly one in five pregnancies. Clinicians know it, mothers know it, and yet the standard advice has long leaned heavily toward caution: avoid sedatives where possible, manage symptoms without medication, and wait. The fear, never quite spoken aloud, was that benzodiazepines and Z-hypnotics (the drugs most commonly prescribed for both conditions) might reach the developing fetal brain and leave something altered. A new study of nearly four million South Korean children now offers the most rigorous challenge yet to that fear. And it largely finds it wanting.
The research, published in The BMJ, followed 3.8 million children born between 2010 and 2022, tracking whether those whose mothers took sedative medications during pregnancy were more likely to develop psychiatric or neurodevelopmental conditions. ADHD, autism, schizophrenia, mood disorders, intellectual disability, behavioural disorder: twelve conditions in all, assessed over up to fourteen years of follow-up.
At first glance, the numbers looked worrying. Children born to mothers who had taken benzodiazepines or Z-hypnotics were more likely, by age 13, to have received a psychiatric diagnosis: about 19% compared with roughly 14% in unexposed children. A gap like that tends to make headlines. But the researchers, based at Sungkyunkwan University in Seoul, knew the raw comparison was almost certainly misleading. Mothers who need sedatives during pregnancy are, by definition, more likely to have anxiety disorders, sleep disorders, and mood conditions of their own, conditions with well-established genetic components. The children of anxious mothers may be more prone to anxiety not because of any pill, but because anxiety runs in families.
Untangling the Drug from the Family
To disentangle the two, the team used a method that epidemiologists often call a sibling controlled design. The logic is elegant and rather ruthless. Find families where one child was exposed in utero and another wasn’t. Compare those siblings. Suddenly the shared genetic background, the same household, the same parenting environment. All of it held constant. What varies is only the pregnancy, and the drug.
When the researchers made that comparison (33,370 exposed siblings against 37,574 unexposed ones from the same families), the apparent risk collapsed. The adjusted hazard ratio for overall psychiatric disorders dropped to 0.99, meaning essentially no difference at all. No significant elevation was found for any individual disorder. The original 5-percentage-point gap, it seems, reflected the kinds of families these children were born into, not what their mothers had taken during pregnancy.
As the linked editorial put it, the study “offers a compelling example of how observational research can generate reliable estimates of prenatal drug safety.” That kind of methodological rigour matters here, because the stakes for getting the answer wrong run in both directions. Unnecessarily frightening pregnant women away from effective treatment carries its own costs: untreated anxiety and insomnia are not benign. They’re associated with preterm birth, low birth weight, and poorer outcomes for both mother and child. The question was never really whether sedatives were risk-free. It was whether the risk, if any, outweighed the harm of going without.
A Window That Still Warrants Watching
The study didn’t give sedatives a completely clean bill of health. In certain subgroups, the point estimates remained modestly elevated even after accounting for family factors, though none reached statistical significance. Exposure across both halves of pregnancy was associated with a sibling-controlled hazard ratio of around 1.35 for benzodiazepines and 1.44 for Z-hypnotics. Women who took Z-hypnotics for 30 days or more saw a hazard ratio of 1.31. These numbers sound alarming, perhaps, but their confidence intervals were wide enough to include 1.0, meaning no difference, meaning chance can’t be ruled out as an explanation. Still, the researchers flagged them as worth investigating further.
There’s a plausible biological story behind the late-pregnancy signal, if it turns out to be real. From around week 20 of gestation, the fetal brain enters a period of rapid synaptogenesis, with new connections forming at a pace that won’t be matched again until early childhood. During this window, the signalling molecule GABA shifts from being excitatory to inhibitory, a transition that plays a central role in shaping neuronal circuitry. Benzodiazepines and Z-hypnotics both act on the receptor GABA binds to. Sustained activation of that receptor during a sensitive period, the thinking goes, could conceivably interfere with how those circuits settle. Could. The evidence doesn’t yet say it does.
What This Means for the Consultation Room
The editorial accompanying the study struck a note of cautious reassurance: these findings don’t license unrestricted prescribing, but they do shift the calculus. Clinicians who have long hesitated to treat severe anxiety or insomnia in pregnancy now have something more robust to work with than the vague and often overstated concern that something bad might happen to the baby’s developing brain. It turns out the picture is considerably more complicated, and probably considerably less grim, than the precautionary instinct had assumed.
Some limitations are worth holding in mind. Prescription records don’t always mean the medication was actually taken; benzodiazepines in particular are often prescribed on an as-needed basis. The follow-up period, though long by the standards of such research, may not be sufficient to capture conditions like schizophrenia or personality disorder, which tend to emerge in late adolescence or early adulthood. And the study wasn’t designed to look at physical outcomes like birth defects or preterm delivery, where some earlier research has raised concerns.
Even so, the finding that the study’s authors call “no substantial evidence” of psychiatric harm is the most credible answer to that question anyone has yet produced. The message to clinicians, as the researchers see it, is not to abandon caution but to calibrate it. When a pregnant woman can’t sleep, when anxiety is compounding the physiological weight of a pregnancy already underway, the risk of leaving that untreated deserves as much weight as the risk, now substantially qualified, of the drugs used to treat it.
Frequently Asked Questions
Does taking a sleeping pill during pregnancy harm the baby’s brain development?
Based on this large study of nearly four million children, prenatal exposure to common sedatives like benzodiazepines or Z-hypnotics does not appear to significantly increase the risk of psychiatric or neurodevelopmental disorders. The key insight came from comparing siblings, which controlled for shared family genetics and environment, and that comparison showed no meaningful difference in outcomes. Some signals in specific subgroups (longer use, late-pregnancy exposure) remain worth investigating, but none reached statistical significance.
Why did earlier studies seem to show a higher risk?
Earlier research didn’t adequately account for the fact that mothers who need sedatives during pregnancy are more likely to have anxiety, mood, or sleep disorders themselves, conditions with genetic links that can be passed to children regardless of medication. When this study compared siblings from the same families, the apparent risk essentially disappeared, suggesting the previous signal reflected family background, not the drugs.
Is it safe to take benzodiazepines throughout an entire pregnancy?
The study’s overall reassurance comes with some nuance: certain patterns of use, particularly across both halves of pregnancy or Z-hypnotics taken for 30 or more days, showed modestly elevated risk estimates, even if they didn’t reach statistical significance. Clinicians are advised to weigh those signals against the real risks of leaving severe anxiety or insomnia untreated, rather than treating sedatives as categorically off-limits.
What is it about the second half of pregnancy that might matter more?
From around week 20, the fetal brain undergoes rapid synaptogenesis, and a key signalling molecule called GABA shifts from an excitatory to an inhibitory role, a transition central to wiring neuronal circuits. Sedatives act on the receptor that GABA binds to, raising the theoretical possibility of interference during this window. The study’s data don’t confirm this mechanism causes harm, but it’s the biological rationale behind why researchers think later-pregnancy exposure deserves closer scrutiny.
Should pregnant women with anxiety stop worrying about taking prescribed sedatives?
The study provides meaningful reassurance, but it’s not a green light for unrestricted use. The stronger message is that untreated maternal anxiety and insomnia also carry risks, and clinicians should weigh both sides of the equation rather than defaulting to avoidance. Decisions about sedative use during pregnancy remain individual conversations, informed now by considerably stronger evidence than before.
ScienceBlog.com has no paywalls, no sponsored content, and no agenda beyond getting the science right. Every story here is written to inform, not to impress an advertiser or push a point of view.
Good science journalism takes time — reading the papers, checking the claims, finding researchers who can put findings in context. We do that work because we think it matters.
If you find this site useful, consider supporting it with a donation. Even a few dollars a month helps keep the coverage independent and free for everyone.