Does omega-3 in pregnancy prevent preterm birth? The evidence splits

Whether omega-3 supplements in pregnancy prevent preterm birth is a question where the best evidence genuinely disagrees with itself, and the honest answer is that it depends on which study you trust most. A large 2018 Cochrane review of 70 trials concluded that omega-3 supplementation lowered early preterm birth substantially [s1]; the single biggest trial since, published the following year, found no effect on early preterm birth at all [s2]. That split is the story, and it is the part a bottle of prenatal fish oil never mentions.

This piece reports what the trials found; it is not medical advice, and decisions about supplements in pregnancy belong with a clinician.

What the Cochrane review found

The 2018 review pooled 70 randomised trials involving 19,927 women at varying levels of pregnancy risk [s1]. Its headline results were favourable and rated high-quality. Preterm birth before 37 weeks fell from 13.4 percent to 11.9 percent, a relative risk of 0.89 (95% CI, 0.81 to 0.97) [s1]. The effect on the more serious outcome was larger: early preterm birth before 34 weeks fell from 4.6 percent to 2.7 percent, a relative risk of 0.58 (95% CI, 0.44 to 0.77) [s1]. Mean gestation was longer by 1.67 days (95% CI, 0.95 to 2.39), and low-birthweight births were modestly reduced (RR 0.90; 95% CI, 0.82 to 0.99) [s1].

The review was careful to flag a trade-off in the same breath: prolonged gestation beyond 42 weeks rose from 1.6 percent to 2.6 percent (RR 1.61; 95% CI, 1.11 to 2.33) [s1]. Pushing gestations longer prevents some early births but pushes a few past term, which carries its own risks — the intervention shifts the whole distribution rather than only trimming its dangerous tail.

What the biggest trial since found

Then came ORIP, a multicentre, double-blind randomised trial that set out specifically to test the early-preterm question the Cochrane review had answered optimistically. It enrolled 5,544 pregnancies in 5,517 women across six Australian centres, assigning them to daily capsules containing 900 mg of n-3 long-chain fatty acids or to a control oil from before 20 weeks until 34 weeks or delivery [s2].

The result was flatly null. Early preterm delivery — the primary outcome, defined as birth before 34 weeks — occurred in 61 of 2,734 pregnancies (2.2 percent) in the omega-3 group and 55 of 2,752 (2.0 percent) in the control group, an adjusted relative risk of 1.13 (95% CI, 0.79 to 1.63; P=0.50) [s2]. There was no signal of benefit, and if anything the point estimate pointed the wrong way. The trial also found no significant differences in post-term deliveries or adverse events [s2].

How to hold both results

A single large, well-designed trial finding nothing does not erase a meta-analysis of 70 — but it does complicate the confident version of the claim. Several things reconcile the two. ORIP was conducted in a generally well-nourished population; benefit from a nutrient supplement tends to be largest where baseline intake is lowest, so a homogeneous, adequately fed cohort is exactly where an effect can wash out. The Cochrane pooled estimate also drew on many small trials of mixed populations and risk levels, and small trials can inflate an effect that larger, stricter ones then fail to reproduce.

What survives both is narrow. The strongest reading is that any benefit is likely concentrated in women who are deficient or at elevated risk, that it is not the large, universal effect implied by supplement marketing, and that the prolonged-gestation trade-off is real. This is the same pattern that recurs across supplement research, where pooled biomarker-friendly findings shrink under hard clinical endpoints — a problem examined in Health Newspapers' look at supplement meta-analyses and their endpoints.

What it means for readers

For a general-population pregnancy, the evidence does not support omega-3 supplements as a reliable way to prevent preterm birth — the best-powered trial designed to show that effect did not [s2]. Where an individual is at higher risk or has low intake, the picture is less settled, and the pooled data leave room for benefit [s1]; that is a clinical judgement, not a shelf decision. It sits alongside the better-established pregnancy-nutrition questions, such as periconception folate, where the evidence is far less ambiguous.

The useful posture is skepticism toward any prenatal product that markets a single dramatic number. Here the dramatic number — early preterm birth falling from 4.6 percent to 2.7 percent (RR 0.58) — is real in one analysis and absent in the trial built to confirm it [s1] [s2]. Both facts are true, and only reporting both is honest.