Implantation of an embryo within the endometrial cavity is a key step in assisted reproductive techniques (ART). It has been suggested that intentional endometrial injury, such as endometrial biopsy or curettage, prior to embryo transfer improves the chances of implantation and further development thereby increasing the likelihood of live birth. The effectiveness and safety of this procedure is, however, still unclear.
To assess the effectiveness and safety of endometrial injury performed prior to embryo transfer in women undergoing ART.
We searched the Cochrane Menstrual Disorders and Subfertility Group (MDSG) Specialised Register, Cochrane Central Register of Controlled Trials (CENTRAL), DARE, MEDLINE, EMBASE, CINAHL, LILACS, and ClinicalTrials.gov. The final search was performed in November 2011.
Randomised controlled trials comparing any kind of intentional endometrial injury prior to embryo transfer in women undergoing ART with no intervention or with a simulated (mock) procedure that could not cause endometrial injury.
Data collection and analysis
Data were extracted by one author and checked by a second. Assessment of the risk of bias was performed independently by two authors. We contacted and corresponded with study investigators as required. Data were analysed using Peto odds ratio (OR) and a fixed-effect model.
A total of 591 women from five trials were included. Clinical pregnancy was reported by all five studies and the combined analysis for clinical pregnancy per woman showed substantial heterogeneity (I² = 95.9%). We therefore conducted a planned subgroup analysis including four trials in the subgroup 'injury in the previous cycle' and one trial in the subgroup 'injury on the day of oocyte retrieval'. In the subgroup 'injury in the previous cycle' the procedure was performed within one month before starting ovulation induction in all four trials. This intervention resulted in a significant increase in the odds of live birth (2 trials, Peto OR 2.46; 95% CI 1.28 to 4.72; I² = 0%) and clinical pregnancy (4 trials, Peto OR 2.61; 95% CI 1.71 to 3.97; I² = 0%). The odds of miscarriage per clinical pregnancy (1 trial, Peto OR 1.13; 95% CI 0.17 to 7.45) and multiple pregnancy per clinical pregnancy (1 trial, Peto OR 0.87; 95% CI 0.23 to 3.30) were very imprecise, limiting any meaningful conclusion. No study reported pain or bleeding. In the subgroup 'injury on the day of oocyte retrieval', the intervention resulted in a significant reduction in the odds of clinical pregnancy (1 trial, Peto OR 0.30; 95% CI 0.14 to 0.63) and ongoing pregnancy (1 trial, Peto OR 0.28; 95% CI 0.13 to 0.61). The single trial in this subgroup did not report any adverse effect outcomes.
Endometrial injury performed prior to the embryo transfer cycle improves clinical pregnancy and live birth rates in women undergoing ART. It is advisable not to perform endometrial injury on the day of oocyte retrieval because it appears to significantly reduce clinical and ongoing pregnancy rates. There is insufficient evidence regarding the effect of endometrial injury on multiple pregnancy or miscarriage and none on adverse events such as pain and bleeding.