“It is difficult for people like me to decide as the research is not done. No research is conclusive, nothing is conclusive for you to make a firm decision. It is emotional, you are desperate, so you think ‘I’ll take that risk’, but really you could be throwing money into nothing. There are different opinions because there isn’t anything conclusive out there. You rely on the trials to do the job, but they don’t. It gets to a point where you lose hope, and question everything.” Fiona Recurrent Implantation Failure
“Lianne, please can I ask you a question, what do you think to the ERA? The RCT on blastocysts? Discussed with consultant and she asked me to read this article to help make my mind up whether I have it!” Kelly Implantation Failure
Keywords
Complier Average Casual Effect (CACE)
Endometrial Receptivity Analysis (ERA)
Food and Drug Administration (FDA)
Human Fertilisation and Embryo Authority (HFEA)
National Library of Medicine (NLM)
Personalised Embryo Transfer (PET)
Pre-implantation Genetic Testing for Aneuploidy (PGT-A)
Randomised Controlled Trial (RCT)
Recurrent Implantation Failure (RIF)
Scientific and Clinical Advances Advisory Committee (SCAAC)
Window of Implantation (WOI)
Summary
There is no moderate, high-quality or any evidence in terms of an properly designed RCT that establishes ERA® with a personalised embryo transfer as an effective intervention that increases the likelihood of achieving a baby for most fertility patients, or for people that suffer recurrent implantation failure.
Although a statement is made by the study authors of a potential utility for ERA for improved cumulative live birth rate, peer review of the RCT under question by the HFEA unfortunately identified that the study suffered a series of biases. This means that the findings are considered unreliable. The secondary outcome cumulative live birth rate should be considered an exploration and communicated with caution.
It could be that ERA can improve birth rates, but there is no well-designed substantial evidence from an RCT to prove this currently. Despite much effort to try to conduct a trial, by Simon et al 2020 (A 5-year multicentre randomised controlled trial comparing personalised, frozen and fresh blastocyst transfer in IVF. Reprod. Biomed. Online. 2020; 41 (567-415)), a new RCT that is designed and implemented correctly is required for claims to be substantiated.
What is ERA®? An Outline
Endometrial Receptivity Analysis (ERA)evaluates a woman’s endometrial receptiveness in synergy with blastocyst implantation. ERA is classified as a treatment add-on. This is because ERA is not considered a standard method to evaluate endometrial status and guide the blastocyst transfer.
So, why a biopsy?
Genetic material is taken from endometrial tissue, and correlated with a panel of data, and evaluated by a computational predictor to determine the endometrial receptivity to blastocyst status. This is possible as the expression of genes change during the phases of a menstrual cycle. Therefore, specific and identifiable key gene expressions are present. ERA analyses 248 biological functions in relation to the window of implantation (WOI). A displaced window of implantation is estimated to occur in 30% of women. ERA may be incorporated with other tests and investigations rather than as a standalone option during an IVF journey.
To ERA with guided transfer or not to ERA?
Traffic light rating system: Green, Amber, or Red
In the UK, the Human Fertilisation and Embryo Authority (HFEA) currently uses a traffic light system. The traffic light system helps inform and guide the public, including health professionals and their patients, as well as researchers, about the:
-
- quality and rating of the evidence base in relation to an RCT (critical appraisal of the design and findings).
-
- cost of additional treatment and the likelihood of increased birth rates.
The traffic light system implies Green/go, Amber/wait, and Red/stop. See Table 1.
| Table 1: Outline of the HFEA RCT Traffic Light Rating System | |
| Traffic Light Rating | Requirement(s) and meaning |
| Green | More than one high-quality RCT study that shows the procedure is effective at improving the chances of having a baby for most fertility patients to be recommended for routine use. |
| Amber | Conflicting evidence from RCTs. One or more studies show a significant outcome and other studies show no difference. Amber means that the evidence is not conclusive, and further research is required. An amber rating also implies that the intervention should not be recommended for routine use. |
| Red | There is no evidence or no strong evidence from RCTs to show that the intervention is effective at improving the chances of having a baby for most fertility patients. |
| *HFEA’s traffic light system has been under development to improve the rating and grading of evidence according to HFEA meeting transcripts. | |
Currently there are zero Green-rated add-ons, seven add-ons are classified as Red and the remaining five are Amber.
A HFEA Red-Rated Randomised Controlled Trial ERA
A “5-year multicentre randomised controlled trial comparing frozen and fresh blastocyst transfer in IVF” used to classify ERA as Red found a potential utility for ERA based on the trial’s secondary outcome. The study’s authors state:
-
- statistically significant improvement in pregnancy, implantation, and cumulative live birth rates: indicating the potential utility of personalised embryo transfer guided by the ERA test at the first appointment.
An independent scientific review of the 5-year multicentred RCT found the primary outcome unreliable due to the trial suffering a range of major issues, including around 40% of the women not receiving the randomly assigned interventions for the primary outcome. The independent review describes that the second outcome cumulative live birth was promising but far from conclusive.
The Scientific and Clinical Advances Advisory Committee (SCAAC) on review decided on a Red recommendation for ERA with guided transfer as, although they state that the study outcomes are promising, they cannot be certain of this trial’s reliability (Figure 1).
Figure 1: HFEA Summary of ERA Red Rating
Unreliable Trial?
The cumulative live birth result was deemed to be a very good finding according to the study authors even though they applied a per protocol analysis.
What is per protocol analysis?
Per protocol analysis excludes people from their randomly allocated group in a randomised controlled trial when they should not be excluded. Per protocol is a controversial method of analysis in an RCT, and not recommended by expert methodolgists. This is because the per protocol violates randomisation and can introduce bias. Intention to treat, which keeps the randomly assigned people in their groups, should be the primary method to analyse data from an RCT. Per protocol may occur in a trial that encountered problems such as non-compliance to the study protocol and/or when analysis explores the effect of a proportion of people only that let’s say, complied with the study protocol. This exploratory method can over-estimate the effect of treatment.1
Fixing the target population vs. random
Fixing and selecting a new target population post randomisation to a set of preferred data casts doubt over the RCT’s integrity. In other words, it is some sort of trial without the randomisation gold standard key feature. See Table 2 below for more detailed information.
Other researchers had concerns about this trial’s reliability. Published communications in Reproductive BioMedicine Online highlight troubles associated with the study findings and claims. Table 2 outlines the concerns raised and responses to the ERA multicentre study findings.
| Table 2: ERA 5-year Multicentre Trial Debate | ||
| Concerns raised via peer review: Unreliable trial, invalid and misleading results. | Response to concerns raised by study authors: Our interpretation is enabled as randomisation is not compromised. | My follow-up |
| There are substantial departures from the original trial registration. It is not clear when or why these departures occurred. | The initial trial registration was amended early on due to recruitment problems. This amendment was approved and updated on clinicaltrials.gov. | The trial was amended due to recruitment problems. However, the methods and results of the trial are what is under question. The trial specified the primary outcome as live birth rate and listed 8 secondary outcomes. Clinical trial registration did not have the data manager pre- specify intended method(s) of analyses. |
| The first embryo transfer per protocol analysis had only 58% of women originally enrolled remaining. This is misleading because it disregards the benefits of randomisation. Findings are likely to not be a result of ERA with personalised transfer; rather, differences in prognosis. | The primary outcome live birth was limited by unexpectedly high patient dropout. However, the secondary outcome cumulative birth rate had no differences in the study groups, and sample size was balanced. Therefore, the effects of randomisation were not negated. | Per protocol method violates and compromises randomisation. Per protocol is an exploratory investigation and experts state that per protocol should not be used to inform whether an intervention is effective or not.2 |
| There are substantial imbalances in the number of embryo transfers for the primary outcome (282 PET vs. 248 Control). Therefore, the authors’ conclusions of benefit from PET (ERA) are misleading. The secondary outcome finding cumulative live birth is invalid. | There are not substantial imbalances in the group. The total number of embryo transfers per arm did not differ statistically. Those that critique us may wish to see our supplementary table. Our interpretations are enabled. | Exploratory interpretations are enabled, but the results should be exploratory only and not presented as an outcome of the RCT.2 Randomisation has been violated by introduction of per protocol and overriding randomisation which controls for known and essential unknown differences – assessing statistically only for what is known does not account for what is unknown. |
| The conclusions of this study should be based on the intention-to-treat analysis of the first embryo transfer. | Transparency is important in reporting data from a trial. We embrace this. Readers can interpret the results at different levels, and we also promote clinical informed decision-making regarding ERA and personalised transfer. | If transparency is important, the study should communicate the limitations, the issues around per protocol and advise caution regarding the findings in the outcome. Correcting for bias in the primary outcome and by taking on per protocol analysis produces further bias. |
| Personalised embryo transfer. | Per protocol is described as a flawed method to ascertain the effectiveness of treatments. According to experts, the ITT approach should always remain the primary method of analysis. In certain cases, an alternative approach (data depending) that better estimates the effect of the treatment compared to per protocol is the complier average causal effect (CACE). CACE, if applicable, can maintain the initial randomised assignment. | |
|
*Statistical Analyses Reported in Multicentre Trial: Multivariable regression analysis with a binomial endpoint (live birth Yes/No) demonstrated homogeneity of the key baseline variables and the absence of bias towards our final endpoint.
|
||
Euploid Blastocysts and ERA
A different RCT, “Endometrial Receptivity Analysis vs. Standard Timing of Transfer
using Genetically Tested Euploid Blastocysts”, investigated the use of ERA.
The authors of this study found that there was no statistically significant difference to support the routine use of ERA when women have genetically tested euploid blastocysts. Note though, that pre-genetic testing for aneuploidy (PGT-A) has a Red traffic light rating, as there is not an RCT that proves PGT-A is effective at improving the birth rates for most fertility patients.
Important participants in this study were also excluded from initial eligibility. Women were not allowed to take part at all if they had a history of recurrent implantation failure (RIF). The study authors outline the limitation of the findings for patients that experience RIF.
Igenomix, which sponsored the 5 year-multicentre trial, responded to the emergent issues on RIF (defined in this study as more than three failed IVF cycles with good quality embryos transferred):
“Regarding the recent publications, it is important to take into account that the analysed population are patients without recurrent implantation failure, which is the population of patients to whom this test is indicated.”
What Next?
If you have no implantation failure of endometrial origin, and you have genetically tested euploid blastocyst(s), you may find that ERA is not considered to be a suitable option for you as one RCT showed no increased likelihood of having a baby when using ERA. Yet, if you have recurrent implantation failure due to endometrial origin you may find that ERA is still considered and recommended (even with euploid blastocysts) as an appropriate clinical treatment given your fertility history when in consultation with your IVF healthcare provider. Well, that’s what several of my patients and I have found.
Although there are claims of ERA and PET improving cumulative birth rates based on a RCT findings (See Figure 2), you should know that the evidence is rated Red (See Table 1).
Figure 2: Engagement of Study Findings to IVF Specialists and Fertility Patients.
Igenomix have recently issued an official statement regarding ERA:
“As experts in the field, we are firmly dedicated to sharing the correct data and providing clear guidance on the appropriate use of our services.
The credibility of the ERA test stems from its extensive scientific validation, with numerous publications confirming its effectiveness. Since 2011, over 200,000 women worldwide have undergone the ERA test provided by Igenomix.”
Many IVF centres offer ERA. They may have observed clinical utility and positive outcomes incorporating ERA with guided transfer when faced with complex patient histories. Clinician’s may not have advanced knowledge of research methodology and impact on RCT outcomes.
Unfortunately for clinicians and their patients, decision-making around the limitations of evidence base in the fertility sector, and the traffic light system means that some people are still faced with challenging and difficult decisions regarding their treatment strategy.
Discussion
Towards a Green traffic light rating
According to the HFEA enquiries team, the RCT was assessed based on GRADE methodology (The Grading of Recommendations, Assessment, Development, and Evaluations) to aid their decision-making around this trial. The full GRADE report is omitted. Also, there was only one RCT for ERA selected at the time the Red rating was published, and two high-quality trials are strictly required for a Green go-ahead (See Table 1).
The HFEA policy of a strict requirement for two high-quality RCTs for add-ons is likely due to standards set for drug and biological licensure. The Food and Drug Administration (FDA) require substantial evidence that demonstrates and confirms a drug is safe, pure, and potent/effective. For non-pharmaceutical or non-biological licensure, the principles of robust substantial evidence and peer-review are important. The FDA does outline, though, that there are circumstances where two high-quality trials may not be required. These circumstances may include one high- quality RCT plus:
-
- confirmatory evidence in related populations, or
-
- supported by evidence of the drugs mechanism, or
-
- compelling results supported by additional data from the natural history of the disease.
It might be interesting for the HFEA to consider the policy requirement of two high- quality randomised trials in the future based on a case-by-case basis. To learn more, click here.
Traffic light system and new development
The HFEA traffic light system is developing to help rate evidence better. An enormous amount of work is conducted by the HFEA to raise awareness of evidence-based treatment and support decision-making. It seems to me, though, that what is also important is guidance for trialists designing RCTs vs. marking as Red or Amber only. It would be helpful at the outset to provide an advisory resource for research development, with the goal of producing high-quality research, to help trialists meet the standards expected by authorities.
Most fertility patients?
It is not clear what “most fertility patients” refers to (See Table 1). However, on questioning the HFEA most fertility patients’ refers to the exclusion of subpopulations of subfertilte people undergoing IVF. This is because a RCT’s population do not generally explicitly conduct research on specific subgroups such as older females, people who have suffered recurrent miscarriage or recurrent implantation failure. This definition needs to be reviewed, it discriminates unintentionally anyone who undergoes IVF and suffers from subfertility with a more complex medical history. Also aging fertility is part of the population of IVF treatments in the UK according to data. Let’s say, high-quality research was conducted for recurrent implantation failure – using this classification there would not be a recommendation for routine use to support this population.
Prespecified analysis
The National Library of Medicine (NLM) support team advises that there is not a specific section to prespecify statistical analysis. However, it states that a data provider can add the statistical analysis to the detailed description on registering a trial. To have a section or prompt may help reduce what is called outcome and reporting bias and help people reviewing the trial understand more about the context of a study. There is a recommendation that statistical considerations should be specified in advance to limit erroneous conclusions resulting from multiplicity.
New RCT designs
Trialists can improve the design of randomised controlled trials. Ultimately, the prevention of post randomisation bias requires addressing, as this negatively impacted the 5-year multicentre trial primary and secondary outcome.
My patients and I believe that the eligibility criteria at the design stage of a RCT can better reflect the population the ERA is stated to be indicated for which is RIF.
Post randomisation bias was introduced to the 5-year multicentre trial, for example, due to:
-
- substantial dropout of patients, non-compliance to the intervention under investigation (attrition bias).
-
- researcher(s) excluding participants (exclusion bias) and knowing information when assessing outcomes (non-blinded ascertainment bias).
-
- a per protocol option applied (violated randomisation) and introducing multiple analyses resulting in misleading claims.
I note that it is likely of great importance to consider sources of clustering present in individually randomised controlled trials and report on this. Clustering can present in RCTs when multicentred, and while clinicians and patients are grouped together. Not accounting for this can lead to effective treatments being discarded and the approval of ineffective treatments in the healthcare sector.
Regulatory policy is not immune to an RCT. An RCT can demonstrate the effectiveness of policies in practice.
P.S. I am not suggesting that fertility patients do not or do have ERA with guided transfer. I think for RIF it is currently worth considering in conjunction with your healthcare provider. Rather, I set out to help answer a couple of questions raised by my patients in terms of the evidence base and authority guidance, which turned out to be more complicated than expected! Hopefully a new trial and further debate will address the issues experienced and raised by the community and, if proven to be effective, the intervention is available to those that really need it.
Lianne Aquilina MSc Applied Health Research
