The effect of administering pregnancy hormone into the womb of subfertile women undergoing assisted reproduction

Review question

Does administering pregnancy hormone into the womb of subfertile women undergoing assisted reproduction provide any benefit?

Background

Subfertility affects 15% of couples and is defined as the inability to become pregnant naturally following 12 months of regular unprotected sexual intercourse. Assisted reproduction refers to procedures involving handling of both sperm and eggs in the laboratory to create embryos to be transferred into the womb (embryo transfer (ET)). Administering natural or synthetic pregnancy hormone into the womb of subfertile women undergoing assisted reproduction treatment is a novel approach that might increase the chance of having a baby.

Study characteristics

We evaluated 17 studies (4751 women) comparing administration of pregnancy hormone versus no hormone. The natural or synthetic hormone was administered at variable doses at different times before ET.

Key results

Live birth rates in women having day three ET with human chorionic gonadotropin administered into the uterus (IC-hCG) at a dose < 500 IU compared to women having day three ET without pregnancy hormone showed no benefit of the intervention and would be consistent with no substantive difference or disadvantage of indeterminate magnitude (very low-quality evidence: one study; 280 women). In a clinic with a live birth rate of 49% per cycle following day three ET, use of a pregnancy hormone dose < 500 IU would be associated with a live birth rate varying from 28% to 50%.

Live birth rate was increased in a subgroup of women having day three ET with a pregnancy hormone dose of 500 IU or greater compared to women having day three ET without pregnancy hormone (moderate-quality evidence: three studies; 914 women). At a clinic with a live birth rate of 27% per cycle, use of a pregnancy hormone dose of 500 IU or greater would be associated with a live birth rate varying from 36% to 51%.

Trial results show no substantive differences in live birth among women having day five ET with a pregnancy hormone dose of 500 IU or greater compared to women having day five ET without pregnancy hormone (moderate-quality evidence: two studies; 1666 women). At a clinic with a live birth rate of 36% per cycle, use of a pregnancy hormone dose of 500 IU or greater would be associated with a live birth rate varying from 29% to 38%.

We are uncertain whether administration of pregnancy hormone into the womb at any dose or time affected miscarriage (very low-quality evidence: 11 studies; 3927 women).

Evidence for clinical pregnancy among women having day three ET with a pregnancy hormone dose < 500 IU showed no benefit of the intervention and would be consistent with no substantive difference or disadvantage of indeterminate magnitude (very low-quality evidence: one study; 280 women).

The clinical pregnancy rate was increased in the subgroup of women having day three ET with a pregnancy hormone dose of 500 IU or greater compared to women having day three ET without pregnancy hormone (moderate-quality evidence: 12 studies; 2186 women).

Trial results show no substantive difference in clinical pregnancy among women having day five ET with a pregnancy hormone dose of 500 IU or greater compared to women having day five ET with no pregnancy hormone (moderate-quality evidence: four studies; 2091 women).

No randomised controlled trials (RCTs) investigated day five ET with a pregnancy hormone dose < 500 IU.

Other complications reported in the included studies were ectopic pregnancy (where the embryo develops outside the womb), heterotopic pregnancy (where embryos develop inside and outside the womb), foetal death, and triplets. Events were few, and insufficient evidence of very low quality does not permit us to determine whether there were differences between groups.

There should be further trials with live birth as the primary outcome to identify the groups of women who would benefit the most from this intervention.

Quality of the evidence

Evidence quality varied from very low to moderate depending on the outcome. The main limitations for the overall quality of the evidence were high risk of bias and serious imprecision.

Authors' conclusions: 

There is moderate quality evidence that women undergoing cleavage-stage transfer using an IC-hCG dose ≥ 500 IU have an improved live birth rate. There is insufficient evidence for IC-hCG treatment for blastocyst transfer. There should be further trials with live birth as the primary outcome to identify the groups of women who would benefit the most from this intervention. There was no evidence that miscarriage was reduced following IC-hCG administration, irrespective of embryo stage at transfer or dose of IC-hCG. Events were too few to allow conclusions to be drawn with regard to other complications.

Read the full abstract...
Background: 

Most women undergoing assisted reproduction treatment will reach the stage of embryo transfer (ET), but the proportion of embryos that can be successfully implanted after ET has remained small since the mid-1990s. Human chorionic gonadotropin (hCG) is a hormone that is synthesised and released by the syncytiotrophoblast and has a fundamental role in embryo implantation and the early stages of pregnancy. Intrauterine administration of hCG via ET catheter during a mock procedure around the time of ET is a novel approach that has been suggested to improve the outcomes of assisted reproduction.

Objectives: 

To investigate whether intrauterine (intracavity) administration of hCG (IC-hCG) around the time of ET improves clinical outcomes in subfertile women undergoing assisted reproduction.

Search strategy: 

We performed searches on 9 January 2018 using Cochrane methods.

Selection criteria: 

We looked for randomised controlled trials (RCTs) evaluating IC-hCG around the time of ET, irrespective of language and country of origin.

Data collection and analysis: 

Two review authors independently selected studies, assessed risk of bias, extracted data from studies, and attempted to contact study authors when data were missing. We performed statistical analysis using Review Manager 5. We assessed evidence quality using GRADE methods. Primary outcomes were live birth and miscarriage; secondary outcomes were clinical pregnancy rate and complications.

Main results: 

Seventeen RCTs investigated the effects of IC-hCG administration for 4751 subfertile women undergoing assisted reproduction. IC-hCG was administered in variable doses at different times before the ET. hCG was obtained from the urine of pregnant women or from cell cultures using recombinant DNA technology.

Most studies (12/17) were at high risk of bias in at least one of the seven domains assessed. Common problems were unclear reporting of study methods and lack of blinding. The main limitations for evidence quality were high risk of bias and serious imprecision.

For analyses of live birth and clinical pregnancy, there was considerable heterogeneity (I² > 75%) and therefore we present subgroups for dosage and stage of ET. Exploration for sources of heterogeneity revealed two key prespecified variables as important determinants: stage of ET (cleavage vs blastocyst stage) and dose of IC-hCG (< 500 international units (IU) vs ≥ 500 IU). We performed meta-analyses within subgroups defined by stage of embryo and dose of IC-hCG.

Live birth rates among women having cleavage-stage ET with an IC-hCG dose < 500 IU compared to women having cleavage-stage ET without IC-hCG showed no benefit of the intervention and would be consistent with no substantive difference or disadvantage of indeterminate magnitude (risk ratio (RR) 0.76, 95% confidence interval (CI) 0.58 to 1.01; one RCT; 280 participants; I² = 0%; very low-quality evidence). In a clinic with a live birth rate of 49% per cycle, use of IC-hCG < 500 IU would be associated with a live birth rate ranging from 28% to 50%.

Results show an increase in live birth rate in the subgroup of women undergoing cleavage-stage ET with an IC-hCG dose ≥ 500 IU compared to women having cleavage-stage ET without IC-hCG (RR 1.57, 95% CI 1.32 to 1.87; three RCTs; 914 participants; I² = 0%; moderate-quality evidence). At a clinic with a live birth rate of 27% per cycle, use of IC-hCG ≥ 500 IU would be associated with a live birth rate ranging from 36% to 51%.

Results show no substantive differences in live birth among women having blastocyst-stage ET with an IC-hCG dose ≥ 500 IU compared to women having blastocyst-stage ET without IC-hCG (RR 0.92, 95% CI 0.80 to 1.04; two RCTs; 1666 participants; I² = 0%; moderate-quality evidence). At a clinic with a live birth rate of 36% per cycle, use of IC-hCG ≥ 500 IU would be associated with a live birth rate ranging from 29% to 38%.

Evidence for clinical pregnancy among women having cleavage-stage ET with an IC-hCG dose < 500 IU showed no benefit of the intervention and would be consistent with no substantive difference or disadvantage of indeterminate magnitude (RR 0.88, 95% CI 0.70 to 1.10; one RCT; 280 participants; I² = 0%; very low-quality evidence).

Results show an increase in clinical pregnancy rate in the subgroup of women having cleavage-stage ET with an IC-hCG dose ≥ 500 IU compared to women having cleavage-stage ET without IC-hCG (RR 1.49, 95% CI 1.32 to 1.68; 12 RCTs; 2186 participants; I² = 18%; moderate-quality evidence).

Results show no substantive differences in clinical pregnancy among women having blastocyst-stage ET with an IC-hCG dose ≥ 500 IU (RR 0.99, 95% CI 0.85 to 1.15; four RCTs; 2091 participants; I² = 42%; moderate-quality evidence) compared to women having blastocyst-stage ET with no IC-hCG.

No RCTs investigated blastocyst-stage ET with an IC-hCG dose < 500 IU.

We are uncertain whether miscarriage was influenced by intrauterine hCG administration (RR 1.04, 95% CI 0.81 to 1.35; 11 RCTs; 3927 participants; I² = 0%; very low-quality evidence).

Reported complications were ectopic pregnancy (four RCTs; 1073 participants; four events overall), heterotopic pregnancy (one RCT; 495 participants; one event), intrauterine death (three RCTs; 1078 participants; 22 events), and triplets (one RCT; 48 participants; three events). Events were few, and very low-quality evidence was insufficient to permit conclusions to be drawn.