Background
In vitro fertilisation is a widely used reproductive technique that can be undertaken with or without intracytoplasmic sperm injection. The endometrial scratch procedure is an in vitro fertilisation ‘add-on’ that is sometimes provided prior to the first in vitro fertilisation cycle, but there is a lack of evidence to support its use.
Objectives
(1) To assess the clinical effectiveness, safety and cost-effectiveness of endometrial scratch compared with treatment as usual in women undergoing their first in vitro fertilisation cycle (the ‘Endometrial Scratch Trial’) and (2) to undertake a systematic review to combine the results of the Endometrial Scratch Trial with those of previous trials in which endometrial scratch was provided prior to the first in vitro fertilisation cycle.
Design
A pragmatic, multicentre, superiority, open-label, parallel-group, individually randomised controlled trial. Participants were randomised (1 : 1) via a web-based system to receive endometrial scratch or treatment as usual using stratified block randomisation. The systematic review involved searching electronic databases (undertaken in January 2020) and clinicaltrials.gov (undertaken in September 2020) for relevant trials.
Setting
Sixteen UK fertility units.
Participants
Women aged 18–37 years, inclusive, undergoing their first in vitro fertilisation cycle. The exclusion criteria included severe endometriosis, body mass index ≥ 35 kg/m2 and previous trauma to the endometrium.
Interventions
Endometrial scratch was undertaken in the mid-luteal phase of the menstrual cycle prior to in vitro fertilisation, and involved inserting a pipelle into the cavity of the uterus and rotating and withdrawing it three or four times. The endometrial scratch group then received usual in vitro fertilisation treatment. The treatment-as-usual group received usual in vitro fertilisation only.
Main outcome measures
The primary outcome was live birth after completion of 24 weeks’ gestation within 10.5 months of egg collection. Secondary outcomes included implantation, pregnancy, ectopic pregnancy, miscarriage, pain and tolerability of the procedure, adverse events and treatment costs.
Results
One thousand and forty-eight (30.3%) women were randomised to treatment as usual (n = 525) or endometrial scratch (n = 523) and were followed up between July 2016 and October 2019 and included in the intention-to-treat analysis. In the endometrial scratch group, 453 (86.6%) women received the endometrial scratch procedure. A total of 494 (94.1%) women in the treatment-as-usual group and 497 (95.0%) women in the endometrial scratch group underwent in vitro fertilisation. The live birth rate was 37.1% (195/525) in the treatment-as-usual group and 38.6% (202/523) in the endometrial scratch group: an unadjusted absolute difference of 1.5% (95% confidence interval –4.4% to 7.4%; p = 0.621). There were no statistically significant differences in secondary outcomes. Safety events were comparable across groups. No neonatal deaths were recorded. The cost per successful live birth was £11.90 per woman (95% confidence interval –£134 to £127). The pooled results of this trial and of eight similar trials found no evidence of a significant effect of endometrial scratch in increasing live birth rate (odds ratio 1.03, 95% confidence interval 0.87 to 1.22).
Limitations
A sham endometrial scratch procedure was not undertaken, but it is unlikely that doing so would have influenced the results, as objective fertility outcomes were used. A total of 9.2% of women randomised to receive endometrial scratch did not undergo the procedure, which may have slightly diluted the treatment effect.
Conclusions
We found no evidence to support the theory that performing endometrial scratch in the mid-luteal phase in women undergoing their first in vitro fertilisation cycle significantly improves live birth rate, although the procedure was well tolerated and safe. We recommend that endometrial scratch is not undertaken in this population.
Trial registration
This trial is registered as ISRCTN23800982.
CORE (COnnecting REpositories)
CORE (COnnecting REpositories)