Vitamin A in Pregnancy: Teratogenicity Threshold Beats RDA

By Erin Rose · Updated July 19, 2026 · Methodology

Educational summary of published research — not medical advice, and not a substitute for prenatal care. Vitamin A dosing in pregnancy should be discussed with your obstetric provider before starting or continuing any supplement. This statement has not been evaluated by the FDA. This product is not intended to diagnose, treat, cure, or prevent any disease.

What Rothman 1995 actually found, in detail

Rothman et al. 1995 screened 22,748 pregnant women via maternal serum alpha-fetoprotein screening or amniocentesis between October 1984 and June 1987. Among the babies born, 339 had birth defects, 121 of cranial-neural-crest origin — the specific defect pattern (affecting structures like the face, skull, and heart that derive from cranial neural crest cells during early embryonic development) that the study's primary endpoint tracked against maternal preformed vitamin A intake. That primary endpoint was met: a positive, dose-dependent relationship. Women consuming more than 10,000 IU/day of supplemental preformed vitamin A had a prevalence ratio of 4.8 (95% CI 2.2–10.5) for cranial-neural-crest defects compared to women taking 5,000 IU/day or less; the wider >15,000 IU total-intake comparison carried a similar ratio of 4.8 (95% CI 1.7–7.3). The authors estimated approximately 1 in 57 infants born to women taking more than 10,000 IU/day had a malformation attributable to the supplement. The apparent threshold sat near 10,000 IU/day, and the excess risk was concentrated in intake before the 7th week of gestation — a critical, honest detail, because many women don't know they're pregnant that early.

This is an observational cohort, not a randomized trial — it cannot be an RCT for ethical reasons, since researchers cannot randomize pregnant women to a suspected teratogen. Dietary and supplement intake was self-reported/interview-based, and confounding by indication is possible, though the authors controlled for major covariates. No industry funding was disclosed in the abstract. Despite those honest limitations, this is the load-bearing citation for the entire vitamin A cluster, and it's the reason every page here carries a pregnancy caveat inline, not just in a footer.

Why the timing window matters as much as the dose

The excess risk in Rothman's cohort was concentrated in intake before the 7th week of gestation — the period of cranial neural crest cell migration and early organogenesis, and also the period during which many pregnancies are not yet clinically confirmed. That timing detail changes the practical guidance: this isn't advice that applies only once a pregnancy test is positive. Anyone who could become pregnant — not just someone actively trying to conceive or already pregnant — is inside the relevant risk window if they're taking a high-dose preformed-A product regularly. That's why the practical takeaway on every page in this cluster is phrased as "pregnant or could become pregnant," not "pregnant."

What to actually take instead: the RDA, food-first, prenatal-range

None of this means vitamin A should be avoided in pregnancy — it's an essential nutrient, and the pregnancy RDA of 770mcg RAE/day (rising to 1300mcg RAE/day during lactation, reflecting colostrum and milk vitamin A demand) is real and worth meeting. The practical path is a standard prenatal vitamin formulated at RDA-range doses, plus ordinary dietary intake — not a separate high-dose preformed-A supplement layered on top. Two specific stacking traps are worth naming explicitly: concentrated fish-liver-oil and cod-liver-oil products deliver preformed vitamin A at doses people don't always register (see the cod liver oil cluster for exact figures — some servings already reach 8–25% of the adult UL before a prenatal vitamin is even added), and liver consumption in excess, since liver is one of the most concentrated natural preformed-A sources. Someone taking a prenatal vitamin, a separate cod-liver-oil supplement, and eating liver regularly could reach the teratogenicity threshold without any single source looking dangerous in isolation.

What this evidence does not support

The practical takeaway

If you're pregnant or could become pregnant: meet the 770mcg RAE/day pregnancy RDA through a prenatal vitamin formulated at RDA-range doses and ordinary food, not a separate high-dose preformed-A product. Check the vitamin A content on any fish-liver-oil, cod-liver-oil, or standalone vitamin A product before combining it with a prenatal vitamin, and raise anything above RDA-range with your obstetric provider before starting it. This is the one page in this cluster where the guidance is closest to unambiguous: the downside of under-shooting the RDA is real but modest and correctable; the downside of exceeding roughly 10,000 IU/day of supplemental preformed A in early pregnancy is a documented, dose-dependent rise in serious birth defects.

Frequently asked questions

Related

• Vitamin A: what the evidence shows
• Vitamin A dosage guide — the full RDA-to-UL range
• Best vitamin A — which products are dosed appropriately
• Cod Liver Oil — a concentrated preformed-A source to count carefully

Sources