Keywords
IVF - endometrial scratch - endometrial trauma - infertility add on
The use of local endometrial trauma known as Endometrial Scratch (ES) to improve implantation
rates in women undergoing assisted conception was first described in 2003[1] and was initially sparked by a paper investigating the pattern of endometrial expression
of gap junction proteins. Repetitive endometrial sampling was performed for 12 patients
undergoing in vitro fertilization (IVF) and recurrent implantation failure (RIF),
interestingly 11 of these patients became pregnant.[2] The concept of using a mechanical stimulus to cause decidual endometrial changes
however goes back many decades where Leo Loeb (1869–1959), an experimental pathologist
and physician had demonstrated that “transitory neoplasms of the uterine mucosa, deciduomas
and placentomas, could be caused mainly by two factors, first a sensitization of the
uterine mucosa by the hormone of the corpus luteum, followed secondarily by the application
of mechanical stimuli, such as a wound or the introduction of a foreign body into
the uterine lumen”[3]. Over the last two decades the use of the controlled endometrial trauma has gained
significant popularity as an adjuvant treatment that for nearly every group of women
with fertility problems in attempt to improve the implantation potential of the endometrium.
While some medical procedures are often slow to be accepted widely into medical practice,
endometrial scratch is one particular situation where an interventional procedure
has been rushed into clinical practice at a speed which far exceeds the time necessary
for generation of good quality evidence. The procedure has become so popular that
a recent survey[4] showed that 83% of surveyed clinicians recommended an endometrial scratch. 77% for
recommended it for women undergoing IVF and ICSI, 3.6% for women suffering from recurrent
miscarriage and yet another 3.6% even recommended it for women trying naturally or
with intrauterine insemination. Amongst those performing it for IVF patients, the
majority recommended it for women with recurrent implantation failure (92%) but some
even recommended it for all women undergoing IVF (4%). Given how common this procedure
is, it is of extreme importance that we objectively evaluate the evidence for its
clinical use. This article will aim to summarize the current state of the evidence
for the use of endometrial scratch in clinical practice in different infertile groups
Proposed Mechanisms by which Endometrial Scratch May Improve The Endometrial Implantation
Potential?
Proposed mechanisms are discussed in more detail elsewhere in this edition (Adjuvant
therapy in ART, Part One), but include the potential release of inflammatory mediators
including uterine natural killer cells, leukemia inhibitory factor and interleukin
15,[5] macrophages and dendritic cells, tumor necrosis factor-α, interleukin-15, growth-regulated
oncogene-α and macrophage inflammatory protein 1B.[6] ES has also been shown to cause the modulation of several endometrial genes that
may be involved in membrane stability during the process of implantation such as bladder
transmembranal protein (UPIb) and adipose differentiation-related protein and mucin1[7] and through enhancement of endometrial angiogenesis through an effect on matrix
metalloproteinase-3 (MMP-3), plasminogen activator inhibitor-1 (PAI-1), insulin-like
growth factor binding protein 1 (IGFBP-1), and IL-1α.[8]
Endometrial scratch for Women Undergoing IVF
Since the earliest study by Barash et al,[1] several studies focusing mainly on women with recurrent implantation failure demonstrated
a significant increase in pregnancy rates by almost double.[9]
[10]
[11] However, conflicting evidence was provided by at least one randomized controlled
study of 156 participants[12] that suggested that the procedure was harmful with a significant reduction in pregnancy
rates [OR of clinical pregnancy rate of 0.30 (0.14, 0.63) p = 0.002].[8] Notably, this trial performed the ES procedure at the time of oocyte retrieval and
not in the month prior to the IVF cycle. Another RCT of 132 embryo transfer cycles
randomized to receive an endometrial scratch procedure or not,[13] was stopped prematurely after an unplanned interim analysis showed a trend toward
a lower clinical pregnancy rates in the endometrial scratch arm (23.5%) compared with
the control arm (35.9%), (hazard ratio = 0.43; 95% CI, 0.18–1.02; p = 0.0568). However, this study was underpowered and more recent studies with larger
sample size have not shown evidence of harm.[14]
Overall however the majority of earlier evidence pointed to an improvement in fertility
outcomes which is reflected in the 2015 Cochrane review that suggested a significant
improvement in live birth rates in women with recurrent implantation failure undergoing
endometrial scratch and classed the evidence as moderate quality.[15]
The Turning Tides
There have since been several key studies that should be pointed out. The first study
by Yeung et al, (2014) was conducted in an unselected population of women undergoing
IVF, of whom nearly 70% were having their first IVF cycle.[16] No significant differences were seen in the unselected population in ongoing pregnancy
rates, miscarriage rates, clinical pregnancy rates, implantation rates and multiple
pregnancy rates. The authors performed a subgroup analysis in women undergoing their
first embryo transfer (N = 209) and similarly found no difference in ongoing pregnancy rate while they noted
a significantly lower pregnancy rate in control women who had had previous treatment
failure (N = 91). However, the study was not powered for these subgroup analyses and therefore
no reliable conclusions can be drawn. Also, a mixture of protocols was used and there
were no restrictions regarding age or day of embryo transfer with most patients receiving
two embryo transfers.[16]
The second study by Lensen et al, (2019), included 1364 women who were randomized
to endometrial scratch or no intervention and found that endometrial scratch did not
result in an increase in live birth rate.[14] Although this study was somewhat interpreted as conclusive evidence against the
use of endometrial scratch,[17] a careful examination of the study population leads to a different conclusion.
The study included different subgroups undergoing IVF to maintain a pragmatic approach
but in doing so also introduced potential methodological problems. First, the title
implied that the findings pertain to all women undergoing IVF when this is not the
case. The study combined a mixture of patients with different prognostic potential.
The study had two main subgroups, the first is women with recurrent implantation failure
and the second is women who have had a maximum of one previous cycle. These are two
very different populations with different prognostic outcomes. The study therefore
did not include a specific group of women having their first IVF cycle and was not
powered specifically for this group. Furthermore, there is no report on the live birth
rate for this group. The findings therefore cannot be generalized to include woman
having their first IVF cycle.
Patients undergoing IVF represent a large heterogeneous group of women and the results
of any IVF study is subject to several inherent causes for bias that lie within the
underlying pathology and the characteristics of the studied population. It is therefore
of upmost importance to target the clinical question to a very specific and homogenous
group of participants. In this case, the heterogeneity is further increased and the
power is further compromised by including both fresh and frozen IVF cycles. Fresh
and frozen embryo transfers are different populations and it is not ideal to combine
them within one analysis. The dynamics of the endometrium are different in fresh and
frozen cycle and combining the two introduces further cause for bias. There is currently
a large UK national randomized control study (the E-Freeze study) which is looking
at the possible increased implantation potential in women having frozen embryo replacement
cycles compared with fresh cycles and until the results are published it is uncertain
how the combination of these two groups could have influenced the results.[18]
Furthermore, the study did not exclude potential clinical confounders where the endometrium
may have been compromised. For example, 10.9% of women in the scratch group and 12.8%
of women in the control group had ovulatory disorders. Although the nature of the
ovulatory disorder is not specified, polycystic ovarian syndrome can be associated
with hyperandrogenism and an adverse effect on the endometrium. Furthermore, patients
with endometriosis were included and the use of an ultra-long protocol for ovarian
stimulation suggests that some women may have had severe endometriosis. Severe endometriosis
can be associated with an adverse effect on the endometrium and therefore it would
have been ideal to exclude these two groups from the analysis or perform a sensitivity
analysis. It is also noted that 7.2% of women with the intervention and 4.2% of women
in the control group received a short (flare) protocol which is often used in women
who have a low ovarian reserve and this may again have influenced results. Similar
potential causes for bias include the variability in the number of embryos transferred,
the day of embryo transfer and the phase of the cycle where the scratch was performed.
The conclusions of this study therefore should be limited mainly to women having recurrent
implantation failure.
Regarding women undergoing first time IVF treatment. Our recently published randomized
controlled study[19] was powered only to this particular group and attempted to minimize bias and heterogeneity
by including only women predicted to have a good response with no significant pathologies
that may affect the endometrium and women expected to have a single blastocyst transfer.
Similar to the findings of Lensen[14] however, we found no evidence of improvement in live birth rates with the use of
endometrial scratch [ES (n = 523) 37.1%, Control (n = 525) 38.6%; 95% C.I. −4.4% to 7.4%, p = 0.621].
Addressing the middle part of the spectrum, is the recently published study from the
Netherlands that focused on women who had only one previously unsuccessful IVF cycle[20] and again found no evidence of a significant improvement in live birth rate [ES
(n = 465) 23.7%, Control (n = 461) 19.1%; R.R. 95% C.I. 096–1.59]. Neither studies demonstrated any increased
in adverse effects and miscarriage rates.
Although this study recommends further research into endometrial scratch for the IVF
population, when viewed in the context of the two other studies,[14]
[19] it is difficult to support this recommendation. All three studies have examined
different populations yet have produced similar results. It seems therefore that the
use of endometrial scratch should no longer be offered for any population of women
undergoing IVF.
Endometrial Scratch for Women Undergoing Other Fertility Treatments
The potential role of endometrial scratch has also been examined in women undergoing
other fertility treatments such as intrauterine insemination (IUI) with conflicting
and inconclusive results. The two main problems with the current evidence is that
the studies have been relatively small and therefore potentially underpowered and
inconclusive[21]
[22]
[23]
[24] and consequently the overall quality of the evidence has been poor. At least two
meta-analyses have systematically analyzed the evidence. The more recent meta-analysis
included eight trials with a total of 1,871 IUI cycles and found that the clinical
and ongoing clinical pregnancy rates were more than doubled in women having the endometrial
scratch prior to IUI, but this was limited to those having the procedure in the early
follicular phase of the treatment cycle rather than in the preceding cycle.[25] The earlier Cochrane review included nine trials with a total of 1512 women. Although
the study found similar findings i.e., a potential improvement in clinical pregnancy
rates with endometrial scratch, the quality of the evidence was classed as very poor
and therefore the results were found to be inconclusive.[4] Similarly, the evidence for the use of endometrial scratch to improve pregnancy
rates in women with unexplained infertility or those trying to conceive naturally
is far from conclusive and the evidence is poor.[4]
Conclusion
Endometrial scratch is an example of an add-on treatment that has been rapidly adopted
into the field of reproductive medicine without the backing of good evidence. It is
probably the simplicity of the technique that led to its rapid adoption into the field
and was initially supported by a plethora of studies that were limited by inherent
bias as a result of small sample sizes and heterogeneity. With the gradual emergence
of newer studies that have been large enough to attain adequate power and homogenous
enough to address some of the inherent heterogeneity in infertility populations, there
is now little doubt that this technique offers no benefit for women undergoing IVF
treatment. While evidence for other infertility populations is awaited, the evidence
that we now have makes it difficult for any clinician to continue to offer this treatment
in the hope that it will increase the chances of a live birth.
