Endometrial scratch
Live birth 26.1% with it, 26.1% without. It also hurts.
The idea is that deliberately injuring the lining of the womb with a small plastic tube in the cycle before transfer triggers a repair response that makes it more receptive. It is usually sold as a quick outpatient procedure to women after a failed transfer.
1,364 women undergoing IVF were randomized to a scratch or to nothing:
Adjusted odds ratio 1.00, with a confidence interval from 0.78 to 1.27. No difference in ongoing pregnancy, clinical pregnancy, multiple pregnancy, ectopic pregnancy or miscarriage.
The current systematic review pooled 24 randomized trials in 7,234 women, restricting the main analysis to the 8 studies at low risk of bias. Live birth odds ratio 1.12 (95% CI 0.98 to 1.28), moderate certainty. Translated: if baseline live birth is 27%, with scratching it is somewhere between 27% and 32%. The reviewers concluded they are uncertain whether it improves live birth, and that endometrial injury probably causes mild to moderate pain, at moderate certainty.
One thing that review did is worth knowing: it applied a research-integrity checklist to the studies it included. Part of the older literature suggesting that scratching helps is of questionable trustworthiness, and the reviewers said so.