Unplanned analysis ties higher DHA dose to fewer early preterm births in women with a gene variant

A gene variant carried by more than half the women in a US pregnancy trial may shape how much a DHA supplement helps, and the finding comes with two warnings that belong up front. The analysis was not one the trial was designed to do, and its clearest result rests on 17 early preterm births. Researchers at the University of Texas at Austin went back to the trial, which tested DHA, the omega-3 fat in most prenatal supplements, and asked whether its results differed by a single gene.

Women in the ADORE trial had been randomly assigned to take either 200 milligrams of DHA a day, the usual prenatal amount, or 1,000 milligrams. Among those who carried a common variant of the FADS gene, early preterm births, before 34 weeks of pregnancy, were more frequent on the lower dose and much less frequent on the higher one. Women without the variant had slightly more early preterm births on the higher dose, too few to be sure of, and more of their pregnancies ran to 40 weeks or longer. The first author is Yan Ning Li and the senior author J. Thomas Brenna, both at the university’s Dell Medical School. They reported this secondary analysis of the ADORE randomized trial in PLOS Medicine on October 8.

How much weight the DHA gene result can bear

The authors are direct about it. Genetic testing was anticipated when the trial was set up, they write, but “this was a post hoc analysis and the genotype-by-dose comparison was not a prespecified trial aim.” That matters because a question asked after the data are in is easier to answer by chance. Among the 761 participants there were 74 preterm births, 17 of them early, and once those were split by gene type and dose, some groups, such as women with two insertion copies, held only one to three early preterm births. A single birth moving from one column to another would shift those percentages noticeably.

Several features make it more than a fishing trip. Only one genetic variant was examined, chosen in advance on biological grounds, and it was genotyped blind to outcome by a laboratory that knew nothing about the women until the work was done. Within each gene type, the comparison between doses is still randomized, because the trial assigned doses by chance. And the women with and without the variant were similar on the things that most predict preterm birth, such as an earlier preterm birth, age and their DHA levels at the start.

The researchers also tested how easily the result could be overturned. The early preterm finding held up when they deliberately skewed the data against it. The preterm birth result was shakier: shifting the outcome of roughly three or four of the 31 carrier preterm birth cases on the low dose was enough to weaken it below their threshold. Their conclusion is that the finding “warrants replication in independent omega-3 supplementation trials.” Until that happens, it is a hypothesis with a good reason behind it.

What the ADORE trial had already shown about DHA dose

ADORE, short for Assessment of DHA on Reducing Early Preterm Birth, ran from 2016 to 2020 at medical centers in Kansas City and in Columbus and Cincinnati, Ohio. Women joined between 12 and 20 weeks of pregnancy and took capsules from bottles labeled only “black stripe” or “solid black”, so neither they nor the staff knew who had which dose.

The main results, published in 2021 by Susan Carlson’s group at the University of Kansas Medical Center, favored the higher dose without being decisive. Early preterm birth occurred in 1.7% of women on 1,000 mg and 2.4% on 200 mg, and the difference was larger in women who started the trial with low DHA status, at 2.0% on the high dose and 4.1% on the low one.

That fits the wider evidence. A 2018 Cochrane review of 70 randomized trials involving 19,927 women found early preterm birth at 4.6% versus 2.7%, without omega-3 and with it, and preterm birth before 37 weeks at 13.4% and 11.9%. The reviewers rated both results high-quality evidence. Omega-3s also probably increased prolonged pregnancy beyond 42 weeks, from 1.6% to 2.6%.

Not every trial agrees. In Australia’s ORIP trial of 5,544 pregnancies, 900 mg of omega-3 a day from early pregnancy did not lower early preterm birth: early preterm delivery occurred in 2.2% of the omega-3 group and 2.0% of the control group. A follow-up analysis of the same trial found that women who started with low omega-3 levels benefited, while those with a baseline status above 4.9% of blood fats had more early preterm births on the supplement, 2.20% against 0.97%. That is the puzzle the new analysis tries to explain.

What is the FADS gene variant, and why would it matter?

The variant sits in the FADS gene cluster, which holds instructions for enzymes that build long-chain fats in the body. It comes in two forms, an insertion and a deletion. The insertion form boosts one of those enzymes and so increases arachidonic acid synthesis: the body makes more arachidonic acid, an omega-6 fat that it turns into prostaglandins, hormone-like compounds that help trigger labor. One of them, prostaglandin E2, is used clinically to induce labor.

DHA works against that pathway, competing with arachidonic acid for the same machinery. The researchers’ reasoning is that women who make more arachidonic acid may need more DHA to keep labor-starting signals in check, and that women who make less may not need extra, and could even be nudged toward longer pregnancies by it. The results fit that idea, but they do not prove the mechanism: the researchers measured the variant and blood fats, not prostaglandins in the womb.

The insertion form is common. In this US sample, 55% of the women carried at least one copy. Black participants were far more likely to carry the insertion form: 85% had at least one copy, compared with 47% of other women. Black participants were also more likely to give birth preterm, 16.1% compared with 8.1%. The team suggests the variant may account for part of that gap, alongside the social and environmental factors earlier research has documented.

The DHA dose results, gene type by gene type

For women carrying the insertion, the numbers moved in the direction the authors predicted. On 200 mg a day, 9 of 214 had an early preterm birth, 4.3%; on 1,000 mg, 2 of 205 did, 1.1%. The paper describes that as about 75% lower. Preterm births before 37 weeks fell from 14.5% to 9.3% in the same women, a drop of about a third.

For women with two deletion copies, the picture was different. Preterm birth was already lowest in this group and fell on the high dose, from 8.7% to 5.8%. Early preterm birth went the other way, from 1 in 151 on the low dose to 5 in 191 on the high dose, though with so few births the authors urge caution. Pregnancies reaching or passing the 40-week due date, which the authors count as prolonged, became more common on the high dose, 26% against 15%. That is a milder measure than the over-42-weeks figure in the Cochrane review, but the authors note that the 2024 guideline’s advice to stop DHA at about 37 weeks is meant to reduce the risk of prolonged gestation.

Put together, the effect of dose on early preterm birth and on long pregnancies clearly depended on the variant, by the authors’ statistical test. For preterm birth overall it did not: both genotypes reduced preterm birth similarly with high-dose DHA, each by about a third. When the researchers looked back at which women happened to get the dose their genotype would suggest, high for carriers and low for non-carriers, that pairing went with the highest share of births between 37 and 40 weeks. The authors call the variant a potential predictive biomarker for DHA dose response.

What the DHA gene analysis cannot show

The study cannot show that choosing a dose by gene works, because nobody was given a dose based on their gene; the matching was done afterwards on paper. Only about three-quarters of the women who finished ADORE agreed to genetic testing and had complete data, and although a weighting analysis suggested this did not skew the results, those who declined may differ.

The numbers in the rarest gene group, women with two insertion copies, are especially thin, a little over 50 per dose. The results come from three US medical centers and have not been tested in other trials or other countries. And the study looked at timing of birth; this analysis does not weigh the other benefits and harms of each dose.

Where this leaves DHA advice in pregnancy

Current guidance does not use genes. A 2024 clinical practice guideline, backed by several medical-scientific organizations, was co-written by Carlson. It recommends that women of childbearing age get a supply of at least 250 milligrams a day of DHA and EPA combined, from diet or supplements, with an extra 100 to 200 mg of DHA in pregnancy. For pregnant women with low intake or low blood levels it suggests about 600 to 1,000 mg a day, continued until approximately 37 weeks’ gestation. A position statement from the European Board and College of Obstetrics and Gynaecology likewise says low omega-3 levels are associated with higher risk of preterm labor.

The new analysis adds a possible reason why the same supplement looked helpful in some trials and not in others: the mix of gene types in each population. It is not, on its own, a basis for genetic testing or for any woman changing her dose, and the authors’ call for replication says as much. Decisions about supplements in pregnancy are made with a person’s own doctor or midwife.

If the result holds up in a trial designed to test it, a genetic test could one day help set the DHA dose. For now it is one unplanned analysis of 761 women that points a clear way for the next trial.