Home/Issues/Issue 05/Study

Pooled analysis of 14 randomised trials in women having IVF or ICSI Publication date verified at source

The effects of oral antioxidants on outcomes of assisted reproductive technology: a systematic review and meta-analysis of randomized controlled trials

AI narration, generated on first listen
Journal
Frontiers in Endocrinology
Authors
Qiang R
Institution
Beijing University of Chinese Medicine
Published
22 July 2026
Source
PMID 42558317 · DOI 10.3389/fendo.2026.1891276
Design
Systematic review and meta-analysis restricted to randomized controlled trials, with GRADE. Women undergoing IVF or ICSI with embryo transfer.
Sample
Women undergoing IVF or ICSI with embryo transfer

What a mature oocyte is

A mature oocyte is an egg that has reached the stage at which it can be fertilised. Clinics count them after egg collection because immature eggs cannot be used.

Why they ran it

Success rates in assisted reproduction remain limited, and oxidative stress has been identified as one mechanism behind poorer oocyte and embryo quality.

Existing reviews mostly cover one antioxidant at a time, reach inconsistent conclusions, or fold in observational studies with a high risk of bias, so a synthesis restricted to randomized trials was still missing.

Drawn from the paper's introduction.

The review searched four databases and ended up with 14 randomised trials covering 1,190 participants, 637 on an antioxidant and 553 on placebo or standard care. The supplements varied widely: melatonin, N-acetylcysteine, coenzyme Q10, astaxanthin, resveratrol and nanomicelle curcumin. So did the patients, spanning polycystic ovary syndrome, poor ovarian response, diminished ovarian reserve, endometriosis, advanced age and unexplained infertility. The pooled outcomes were clinical pregnancy, mature oocyte count, high quality embryo count, live birth and miscarriage. Nine of the 14 trials used a placebo, the other five simply gave one group nothing extra.

Clinical pregnancy came out at a risk ratio of 1.29, interval 1.07 to 1.55, across 12 trials and 989 people, with no measurable heterogeneity between them. Mature oocytes rose by 2.99, interval 1.82 to 4.16, across nine trials, though heterogeneity there ran to 80.9 percent. High quality embryos rose by 1.08, interval 0.69 to 1.47, from four trials. Live birth came in at 1.41, interval 1.05 to 1.88, pooled from three trials, one of which reported ongoing pregnancy at 12 weeks rather than an actual birth.

Leave one out testing held for pregnancy, oocytes and embryos, but not for live birth. Removing the 2018 coenzyme Q10 trial pushed that estimate to an interval of 0.90 to 1.75, crossing no effect. Trim and fill on clinical pregnancy imputed three missing studies and trimmed the effect from 1.29 to 1.20. And although oxidative stress is the entire explanation offered for why any of this should work, not one pooled outcome is a redox measurement. Every pooled number is a count of pregnancies, eggs, embryos or births.

The numbers

undefinedundefined
undefinedundefined
undefinedundefined
undefinedundefined

Why this might happen

Proposed by the authors This is the explanation the authors offer in their discussion. This study did not test it.

The authors argue that low levels of reactive oxygen species are needed for normal signalling but excess production tips into oxidative stress, which they describe as compromising oocyte quality and endometrial receptivity and driving DNA fragmentation in the developing embryo.

They add that assisted reproduction itself, through ovarian stimulation and prolonged handling of gametes, exposes tissue to non physiological conditions that accelerate reactive oxygen species build up, so supplementation is framed as restoring balance.

For the pregnancy result they say only that they speculate it is a cumulative consequence of better oocyte maturity and embryo quality, and that the specific associations remain to be elucidated.

Drawn from the paper's discussion.

What this does not show

  • The live birth result does not survive a standard robustness check. Dropping the 2018 coenzyme Q10 trial moved the pooled confidence interval to 0.90 to 1.75. The finding hangs on one of only three contributing trials.
  • Nothing about oxidative stress was actually measured in the pooled results. The review's whole rationale is redox balance in the oocyte, yet no antioxidant capacity, reactive oxygen species or oxidative damage marker was pooled from any of the 14 trials.
  • One of the three live birth trials did not report live births. The 2022 advanced age trial contributed ongoing pregnancy confirmed at 12 weeks, which the authors accepted as a stand in for live birth in the pooled estimate.
  • Miscarriage could not be pooled at all. Only two trials reported it, so the review described them separately: 15.7 against 18.6 percent in one and 8.7 against 12.5 percent in the other, with both intervals crossing no effect.

Where this leaves us

Pooling randomized trials only, oral antioxidants taken alongside IVF now look more likely to be associated with higher clinical pregnancy rates and more mature oocytes. That refines the earlier single agent reviews that disagreed with each other, while undercutting agent specific claims, since no individual antioxidant's pregnancy subgroup reached significance on its own.

Women undergoing IVF or ICSI with embryo transfer in 14 randomized trials totalling 1190 participants, across mixed diagnoses and five countries.

Large multicentre randomized trials using standardized regimens and stratified by patient characteristics, with live birth and miscarriage as the primary endpoints rather than afterthoughts.

Caveats worth holding

  • Five of the 14 trials gave the control group nothing rather than a placebo, so those participants knew what they were getting.
  • The pooled supplements are chemically unrelated to one another, and the review itself says the result should be read as an overall trend rather than evidence for any single one.
  • Twelve of the 14 trials were run in Iran, Turkey or China, so the clinic populations are narrow.
  • Doses, formulations and durations varied enormously, from six weeks to twelve and from a few milligrams to hundreds, and the review does not settle on any single regimen.

Newsletter

Each issue by email, when the newsletter launches. Leaving your address puts you on the list, nothing is sent yet.