was the major malformation rate about 4% after first-trimester z-drug exposure in both data sets?
“In both data sets, the rate of any major congenital malformation was 4% after an exposure”
a guest, Dr. Amanda Corrier, presenting for the Psychopharmacology Institute said on Psychopharmacology and Psychiatry Updates, 2026-10-18, at 05:17. hear it in context → the speaker cited: a recent JAMA Psychiatry study of insurance claims.
the episode: Are Z-Drugs Safe in the First Trimester of Pregnancy? (at 05:17). we link the show's own page; we do not re-host audio.
first, the part that holds
the study is real and says what the speaker says it says on the question that matters: in a JAMA Psychiatry cohort of 4.3 million pregnancies drawn from Medicaid and MarketScan insurance claims, first-trimester z-drug exposure was not linked to more major congenital malformations after adjustment (pooled relative risk 1.01, 95% CI 0.95-1.08).
what we found when we went and looked
we read the published abstract of the study (JAMA Psychiatry 2026;83(2):162-171). it covers 2,506,106 Medicaid and 1,775,473 MarketScan pregnancies, with 11,652 and 10,862 exposed in the first trimester, 92.1% of them to zolpidem. the adjusted pooled relative risk for any major malformation was 1.01 (0.95-1.08), which backs the "not significantly different after adjustment" part. the specific "4% in both data sets" figure is not in the published abstract. a search snippet credits roughly 4.0% (Medicaid) and 4.2% (MarketScan) in exposed pregnancies to an earlier conference abstract, but that page blocked us and we could not read it, so we can't confirm the number. the same snippet puts unexposed risks lower before adjustment, at about 3.3% and 3.8%.
assertion by assertion
| assertion | verdict | what the record says |
|---|---|---|
| a recent JAMA Psychiatry study used Medicaid and MarketScan insurance claims to look at first-trimester z-drug exposure and major congenital malformations | confirmed | JAMA Psychiatry 2026;83(2):162-171 (doi 10.1001/jamapsychiatry.2025.3874). It used Medicaid 2000-2018 and MarketScan 2003-2020, with 4,281,579 pregnancies, and defined exposure as at least 1 dispensing of zaleplon, eszopiclone or zolpidem in the first trimester. It covers liveborn infants only. [1] |
| in both datasets the rate of any major congenital malformation after exposure was about 4% | unverifiable | the published abstract gives no absolute risks, and we could not see the full results tables. a search snippet attributes 4.0% (Medicaid) and 4.2% (MarketScan) to a 2024 conference abstract by the same research program, but that page returned an access error and we did not read it. the number is plausible, but we have not verified it from the primary record. [1] |
| that rate was similar to unexposed pregnancies and not significantly different after adjustment | confirmed | adjusted pooled relative risk 1.01 (95% CI 0.95-1.08), using propensity score fine stratification weights. The study's conclusion is no meaningful elevation in overall malformation risk. One caveat: the abstract reports only the pooled adjusted estimate, not an adjusted estimate for each database, and the unread conference snippet suggests the crude exposed rates were somewhat higher than the unexposed ones before adjustment. [1] |
| popular framing: the study shows z-drugs carry no malformation signal of any kind in early pregnancy | overstated | the adjusted pooled relative risks were raised for neural tube defects (1.62, 0.96-2.74), abdominal wall defects (1.46, 0.89-2.38) and tetralogy of Fallot (1.45, 0.86-2.46). The authors call these imprecise, driven by the Medicaid cohort and not replicated in MarketScan. That is no consistent signal, which is weaker than proof of no risk. [1] |
| context: malformations are not the only pregnancy outcome studied for z-drugs | confirmed | a 2025 review describes earlier studies that linked zolpidem to preterm birth, low birth weight and small-for-gestational-age births but not to congenital abnormalities, and notes that underlying maternal conditions may explain those findings. This study did not look at those outcomes. [2] |
both directions
for the speaker: the primary record backs the conclusion: an adjusted pooled relative risk of 1.01 with a narrow confidence interval, based on more than 22,000 exposed pregnancies in two independent insurance populations, results that held across sensitivity analyses, and an author conclusion of no meaningful elevation in overall malformation risk. the speaker also named the journal, the data type and both databases correctly. a roughly 4% figure for exposed pregnancies in both cohorts matches what a search result attributes to the investigators' earlier conference abstract.
against: the 4% number does not appear in the published abstract, and we could not read the source it reportedly comes from. "similar to unexposed" fits the adjusted estimate, but the unread snippet suggests exposed pregnancies had somewhat higher crude rates (about 4.0% vs 3.3% in Medicaid, 4.2% vs 3.8% in MarketScan), so the similarity depends on the confounding adjustment. the abstract reports adjusted results pooled across databases, not each database's adjusted estimate. three rare defects, neural tube defects most of all, had raised but imprecise pooled estimates, driven by Medicaid.
what this cannot say
we read the abstract and key points but not the full-text tables, so we could not check the per-database absolute risks or the adjusted estimates for each database. the study includes only liveborn infants, so malformations ending in miscarriage or termination are missed. exposure is a pharmacy dispensing, not confirmed ingestion. 92% of exposure was zolpidem, so the results say little about zaleplon or eszopiclone on their own. claims-based malformation coding misses some defects. the study looked at malformations only, not preterm birth, growth or neonatal effects. our own check rests on one primary paper plus one review.
sources, each one fetched and read
- PubMed record PMID 41432941, abstract of the same JAMA Psychiatry study (National Library of Medicine E-utilities) What we read there: Z-drug exposure in 11,652 (0.5%) of 2,506,106 Medicaid pregnancies and 10,862 (0.6%) of 1,775,473 MarketScan pregnancies; 92.1% zolpidem; pooled adjusted RRs raised for abdominal wall defects 1.46, tetralogy of Fallot 1.45 and neural tube defects 1.62, "imprecisely estimated, driven by the Medicaid cohort and not replicated in the MarketScan cohort." Fetched 2026-10-19 (HTTP 203). https://pubmed.ncbi.nlm.nih.gov/41432941/
- How do you Zzz during pregnancy? A brief review of Z-drug use and management of insomnia during pregnancy. Journal of Clinical Sleep Medicine. 2025;21(8):1499-1502 What we read there: "there is no clear association between zolpidem use and major birth defects"; a Taiwan study tied zolpidem to preterm and cesarean births, low birth weight and small-for-gestational-age births but not congenital abnormalities, without adjusting for concurrent psychiatric disorders. Fetched 2026-10-19 (HTTP 200). https://pmc.ncbi.nlm.nih.gov/articles/PMC12320687/
speakers are described by role, not named. one sentence is quoted under fair use, with attribution and a link to the original at the timestamp; we do not re-host audio or video. nothing on this page is a reason to start, stop or change a medication. if you have the file behind anything marked unverifiable, the research board is where people can answer. adam