A commonly offered add-on treatment for IVF fails to provide any benefit in a large randomized trial

The results of the study are presented today in Barcelona at the 34th Annual Meeting of ESHRE by Dr Sarah Lensen, a researcher from the University of Auckland, New Zealand.

The study was a large randomised trial performed at 13 fertility centres in five countries (New Zealand, UK, Belgium, Sweden and Australia) and involving more than 1300 women having IVF. One half were randomly assigned to endometrial scratching and the other to no adjuvant procedure. The scratch was performed with a Pipelle cannula, a small flexible plastic tube commonly used for performing uterine biopsy for a variety of indications. In IVF, where some studies have shown a benefit in outcome, it's been proposed that injury to the lining of the uterus causes an inflammatory response conducive to implantation following embryo transfer.

"Results from earlier studies have suggested a benefit from endometrial scratching in IVF," explained Lensen, "especially in women with previous implantation failure. However, many of these studies had a high risk of bias in their design or conduct and did not provide strong evidence. There was still uncertainty about the validity of a beneficial effect."

Women in the endometrial scratch arm of the study had a Pipelle biopsy between day 3 of the preceding cycle and day 3 of the IVF/embryo transfer cycle. Controls had no intervention. Results showed that clinical pregnancy rate in the endometrial scratch group was 31.4% and in the control group 31.2%; live birth rates were 26.1% in the former and 26.1% in the latter. The probabilities of pregnancy were still comparable after controlling for variables and sub-group analysis, which included patients with a history of implantation failure in IVF (defined as two or more unsuccessful embryo transfers). This group in earlier studies has appeared to gain particular benefit from the procedure.

This latest study also measured pain discomfort score associated with endometrial scratch and found "a moderate amount of pain and bleeding." This too, said the investigators, was further reason why endometrial scratch should be abandoned and removed from the list of IVF adjuvant options.

Endometrial scratch is a reportedly common add-on treatment in fertility clinics. A survey of clinics in Australia, New Zealand and UK performed by Lensen and colleagues in 2016 found that 83% of clinicians would recommend endometrial scratching prior to IVF, especially to women with recurrent implantation failure.(1)

"Our results contradict those of many studies published previously," said Lensen, "and, although our trial was the largest and most robust study undertaken so far, it can be difficult for one trial to change practice. However, there are other trials under way at the moment, including two large studies from the Netherlands and UK. Nevertheless, even based just on our results, I think clinics should now reconsider offering endometrial scratch as an adjuvant treatment."

Endometrial scratch: the story so far

1. See Lensen S, Sadler L, Farquhar C. Endometrial scratching for subfertility: everyone's doing it. Hum Reprod 2016; 31: 1241-1244.

2. Several (but not all) studies in the past reported a greater chance of pregnancy in women who had had some minor surgical procedure in the uterus, such as hysteroscopy (see for example, El Toukhy T, Sunkara SK, Coomarasamy A, et al. Outpatient hysteroscopy and subsequent IVF cycle outcome: a systematic review and meta-analysis. Reprod Biomed Online 2008; 16: 712-9.

3. This benefit was seen in a review of recent studies (of "intentional" endometrial injury) in women prior to IVF (see Nastri CO, Gibreel A, Raine-Fenning N, et al. Endometrial injury in women undergoing assisted reproductive techniques). However, conclusions were described as "uncertain."

4. Some studies have proposed that endometrial injury might help improve reproductive outcomes by increasing endometrial receptivity for an implanting embryo by the release of growth factors from the uterine tissue or by changes in its cell structure.

5. There are 20 or more randomised trials of endometrial injury published. What makes this more reliable than the others, says Sarah Lensen, is not that it had more reliable methods of randomisation, measured harms as well as benefit, and had a low risk of bias, but its large sample size -- at 1300 it is far bigger than in any other trials published to date.