Few IVF Add-Ons Have Any Proven Benefits, Study Shows

Many popular in-vitro fertilization (IVF) add-ons have little proven benefit, but an evidence-based website helped patients better understand the treatments so they could make an informed decision, a pair of companion studies found.

In a systematic review and meta-analysis of randomized controlled trials (RCTs), most IVF add-ons had limited or uncertain benefit for patients and much of the evidence was low quality, according to Sarah Lensen, PhD, of Royal Women’s Hospital and the University of Melbourne in Australia, and colleagues.

And in a parallel-group, single-blinded, randomized trial, patients showed a higher understanding of IVF add-ons after viewing a novel evidence-based website co-designed by patients and clinicians to support informed decision making compared with control content generated from the internet, Lensen and colleagues reported.

Both papers were published in The Lancet Obstetrics, Gynaecology, & Women’s Health.

IVF is increasingly popular, as one in seven couples struggle to conceive. But even with IVF, the probability of having a baby is only 30% to 40% per cycle. Many patients turn to IVF add-on treatments — which are mostly new and understudied — that aim to improve chances of pregnancy but cost more.

“Research integrity is a major concern in the field of infertility, with analyses suggesting up to 50% of RCTs in reproductive medicine are either unregistered or raise trustworthiness concerns,” Lensen and team wrote, adding that this is especially true in research about IVF add-ons.

In an editorial linked with both papers, David Barad, MD, of the Center for Human Reproduction in New York City, wrote that these two studies address how evidence about IVF add-ons is weak and the information patients are given about them is often poor.

He argued that the challenge is more than misinformation, but rather “the interaction between scientific uncertainty, commercial incentives, and underinvestment in the systems needed to generate and appraise reliable evidence.”

“Claims should be judged against trustworthy trials, uncertainty should be communicated clearly, not concealed beneath optimistic language, and patient information should be treated as part of the clinical intervention, not as a marketing accessory,” Barad wrote, calling for more investment in large trials and research-integrity safeguards.

Evidence of Safety and Effectiveness

For the systematic review and meta-analysis, authors searched MEDLINE, Embase, and the Cochrane Central Register of Controlled Trials from inception through January 2026 for prospectively registered RCTs that compared IVF add-ons to no treatment or placebo among IVF patients. Nearly half of the studies had serious trustworthiness concerns and were excluded, as were cluster RCTs. Ultimately, 85 trials were included, and the primary outcome was live birth.

Researchers found that the only add-on that may be associated with an increased chance of live birth was endometrial scratching (OR 1.20, 95% CI 1.02-1.41, P=0.02), which had 19 RCTs contributing to moderate certainty.

Four add-ons seemed to have no effect on live birth rates:

• EmbryoGlue: OR 1.12, 95% CI 0.91-1.37, P=0.29; seven RCTs; low-certainty evidence
• Preimplantation genetic testing for aneuploidy: OR 1.16, 95% CI 0.91-1.47, P=0.23; five trials; moderate certainty
• Endometrial receptivity testing: three trials; meta-analysis not undertaken; moderate certainty
• Corticosteroids: OR 0.95, 95% CI 0.70-1.27, P=0.71; two trials; moderate certainty

The remaining add-ons (physiological intracytoplasmic sperm injection [PICSI], acupuncture, intralipid infusions, and platelet-rich plasma [PRP] injection to the ovary or uterus) did not have enough high-quality evidence to make a determination. Based on what evidence there is, PICSI appeared to lower the chance of miscarriage, and both endometrial scratching and EmbryoGlue may increase the probability of pregnancy and live birth, the authors wrote.

Many studies included in the review used IVF methods not used now, which the authors noted as a limitation. They also called for more well-designed RCTs to definitively determine benefits and harms.

Evidence-Based Website Led to Better Understanding

A total of 1,217 Australian patients undergoing IVF were recruited via social media from March 19 through April 6, 2025 and randomly assigned to either a co-designed evidence-based website (n=606) or to control information generated from high-ranking Google search results (n=611). Participants were then randomized to receive study or control information on one of three IVF add-ons chosen to represent different evidence levels: EmbryoGlue, which has some evidence of benefit; intralipid infusion, which has unclear evidence; or endometrial receptivity testing, which has evidence of no benefit.

The primary outcome was gist comprehension regarding the effect of the three extras on the chance of pregnancy, miscarriage, and live birth as measured by a five-item questionnaire using a 0 to 10 scale, with higher scores indicating higher comprehension.

Most participants were women between ages 31 and 40 who had previously undergone IVF and were planning future IVF treatment. These patients reported frequent use of clinic websites and social media for information about add-ons (which may overstate benefits), though about three-quarters reported looking at scientific journals.

Those randomized to the evidence-based website information had substantially higher mean gist comprehension than the control group (respectively 6.1 vs 2.2; mean difference 3.8, 95% CI 3.5-4.2; standardized mean difference 1.42, 95% CI 1.27-1.57, P<0.0001). They also had better understanding about the cost of add-ons, the quality of evidence, and what target populations they are most appropriate for. While the content on the evidence-based website was longer, participants said it was appropriate.

The researchers said study limitations included that their control content replicated top search results which could change over time, that patient characteristics were self-reported, and that they didn’t measure long-term comprehension.