Recurrent implantation failure
Repeated implantation failure is a complex situation in reproductive medicine, characterised by the absence of pregnancy after several transfers of embryos considered capable of implanting. Its definition is not uniform and varies from one team to another, which makes studies difficult to compare with one another. The work in which Arsène Mekinian has taken part concerns blood immune markers, a French cohort of women treated with immunomodulation and an ongoing controlled trial. This is research, it is not an offer of care.
After in vitro fertilisation, the embryo obtained in the laboratory is placed in the uterus: that is embryo transfer. For a pregnancy to begin, this embryo must implant in the lining of the uterus, the endometrium. It happens that several transfers of embryos judged to be of good quality follow one another without any pregnancy becoming established. It is this situation that is called repeated implantation failure, often abbreviated RIF from the English recurrent implantation failure. It is not a disease, it is a clinical situation: an observation of repetition, and the question of whether to look further.
Its definition is not uniform and must take account of several parameters, in particular maternal age, embryo characteristics, the number and type of transfers performed and the expected probability of implantation.
After exclusion of the main embryonic, anatomical, endocrine and gynaecological causes, the possible contribution of endometrial, inflammatory or immunological mechanisms constitutes an active field of research. The rest of this page describes what the work in which Arsène Mekinian took part actually measured, including where the result is negative.
A definition still evolving
There is no single international definition of recurrent implantation failure.
The first studies mainly used a predefined number of unsuccessful transfers. That is the position taken by his team's French multicentre study of 2021, which defines in its first sentence unexplained recurrent implantation failure as the absence of pregnancy after at least three transfers of good-quality embryos following in vitro fertilisation or ICSI 3. The CERTIFY trial takes up this threshold of three transfers, relying on the Istanbul criteria of embryo quality 8.
More recent approaches tend to incorporate the individual probability of implantation, so as to avoid applying a single threshold to biologically different situations. The 2023 ESHRE good practice recommendations, which are not his, thus decline to set a number of transfers and describe the situation by the decision to investigate: recurrent implantation failure is defined there as the situation in which the transfer of embryos considered viable has failed to result in a positive pregnancy test sufficiently often, in a given patient, to warrant considering further investigations or interventions, with a numerical reference point, a predicted cumulative chance of implantation above 60% 9. A committee opinion of the American Society for Reproductive Medicine published in August 2026, which is not his either, announces a revision of the diagnostic criteria and proposes an evidence-based evaluation; its full text is behind a paywall and only its abstract has been read, so that none of its recommendations can be cited in detail here 10.
Three definitions therefore coexist, and writing that recurrent implantation failure "is defined by three failures" would be inaccurate. This heterogeneity of definitions constitutes an important limitation for comparing studies and assessing diagnostic or treatment strategies.
Why study immunity?
Implantation results from complex interactions between the embryo, the trophoblast, the endometrium and the maternal immune system. Pregnancy requires the mother's immune system to tolerate a half-foreign embryo, and for some twenty years it has been asked whether a defect in this tolerance might explain certain failures.
Uterine NK cells, macrophages, regulatory T cells, cytokines, HLA molecules and KIR/HLA-C interactions take part in the regulation of the maternal-fetal interface and in early placental development. The team's most recent synthesis, a review in French published in January 2026 in La Revue de médecine interne where Arsène Mekinian is 4th of 5 named authors, describes the role of uterine NK cells, macrophages and dendritic cells at this interface, the balance between pro-inflammatory and anti-inflammatory cytokines, and the alterations of endometrial receptivity observed in endometriosis, adenomyosis and chronic endometritis. Its concluding sentence is that immunomodulatory treatments such as corticosteroids, intravenous immunoglobulin and intralipids are under study to improve pregnancy outcomes, but that their efficacy remains to be confirmed by large studies 7.
Variations in some of these parameters have been described in patients with implantation failure. Their clinical significance nevertheless remains difficult to establish.
In particular, it is essential to distinguish physiological mechanism, biological association, predictive biomarker and therapeutic target: demonstrating one does not necessarily establish the others.
Lines of research
The work in which Arsène Mekinian takes part seeks in particular to clarify several questions.
Immunological characterisation
Identifying possible immunological phenotypes associated with repeated implantation failure.
Biomarkers
Assessing the relevance of the circulating and endometrial markers proposed to characterise these patients.
Interface between endometrium and immunity
Studying the interactions between endometrial receptivity, inflammation and the immune system, in particular in situations such as endometriosis, adenomyosis or chronic endometritis. Two reviews by the team thus shift the question from the blood towards the endometrium. The first, published in March 2021 in the American Journal of Reproductive Immunology, concerns endometriosis associated with infertility and brings together the data on cytokines and autoantibodies in serum and peritoneal fluid: it finds in a few uncontrolled studies an increase in pregnancy rates on corticosteroids and on TNF alpha antagonists, and concludes explicitly that there are no properly designed trials; Arsène Mekinian is there the 12th and last of 12 named authors 5. The second, published in August 2021 in the Journal of Clinical Medicine, reviews the microRNAs of the normal and pathological endometrium and refers to the expected existence of signatures specific to recurrent implantation failure as a diagnostic avenue to be assessed; he is 4th of 8 named authors there 6.
Immunomodulation
Assessing in controlled studies whether certain immunomodulatory strategies can alter the chances of implantation and, above all, the obstetric outcomes. The team's founding text on this subject is a review with meta-analysis published in July 2016 in the American Journal of Reproductive Immunology, of which Arsène Mekinian is 1st of 9 named authors 1. It deals together with recurrent miscarriage and implantation failure, and reviews corticosteroids, progesterone, intralipids, TNF alpha antagonists, G-CSF, hydroxychloroquine, intravenous immunoglobulin and endometrial scratching, from searches conducted in Medline, Embase and the Cochrane Library. The figures it reports are those of the literature assembled by the authors, and not those of a series of their own: a modest benefit of progesterone on live birth, with an odds ratio of 1.38 (95% confidence interval 1.07 to 1.77) and strong heterogeneity between studies (p = 0.01; I² = 78%); in 200 women treated with intralipids for recurrent miscarriage and implantation failure, a pregnancy rate of 52%. The most favourable results cited in this review, those of TNF alpha antagonists and of G-CSF, concern series of recurrent miscarriage and not implantation failure, which makes it impossible to transpose them. The authors' conclusion is that the efficacy of these drugs remains to be demonstrated and that it remains to be determined which subgroups of patients might derive a benefit from them.
The work on circulating NK cells illustrates the current limits of this approach: differences have been reported between populations in certain studies, but their measurement in blood has not demonstrated a value reproducible enough to guide individual management on its own. The team's clearest piece of work is, from this point of view, a negative result, and it concerns the test most often offered to the women concerned: the measurement of NK cells in the blood. In a case-control study followed by a meta-analysis of the literature, published in August 2019 in Archivum Immunologiae et Therapiae Experimentalis, 115 women were recruited consecutively from December 2015 to October 2017 in three university hospitals: 54 with recurrent miscarriage, 41 with recurrent implantation failure and 20 controls who had had at least two term births. The percentages of NK cells (CD3−CD56+) and of T-cell large granular lymphocytes (CD8+CD57+) measured before conception do not differ between cases and controls, and do not differ either between the women who will subsequently miscarry and those who will have a live birth. The meta-analysis of the literature does find higher proportions of NK cells than in controls in recurrent miscarriage (mean difference 3.47; 95% confidence interval 2.94 to 4.00; p < 0.001) and in implantation failure (mean difference 1.64; confidence interval 0.82 to 2.45; p < 0.001), but with high heterogeneity between studies. The authors conclude that the proportion of blood NK cells in the preconception period does not reflect the risk of implantation failure or of miscarriage and should not serve as an indicator for management. Arsène Mekinian is the 16th and last of 16 named authors of this work 2.
From observational data to the controlled trial
Several immunomodulatory treatments have been used or studied in repeated implantation failure: corticosteroids, intralipids, intravenous immunoglobulin and TNF antagonists, among others.
French cohorts have reported associations between some of these strategies and the achievement of a pregnancy. The main one is the French multicentre cohort of 2021, published in the Journal of Reproductive Immunology, whose figures circulate the most and whose limits must be read together with the result. It is a multicentre observational study, indexed as such, and not a trial. Sixty-four women of mean age 36 plus or minus 3 years were recruited consecutively in university departments for unexplained recurrent implantation failure. Their 340 embryo transfers gave 68 clinical pregnancies and 18 live births. A clinical pregnancy was recorded after 56% of the transfers performed on intralipids and 50% of the transfers performed on prednisone, against 5% of the untreated transfers (p < 0.001). Across all transfers, clinical pregnancies were more frequent after treated transfers than after untreated transfers, 44% against 9% (p < 0.001), with an odds ratio of 8.13 (95% confidence interval 4.49 to 14.72; p < 0.0001). Cumulative pregnancy rates did not differ between transfers on corticosteroids and transfers on intralipids. The conclusion written by the authors themselves is that well-designed randomised studies, in a properly defined population of women, are needed to confirm these preliminary data. Arsène Mekinian is the 15th and last of 15 named authors, and corresponding author 3.
A retrospective series devoted to intralipids alone, published in September 2020 in the European Journal of Obstetrics and Gynecology and Reproductive Biology, gives the measure of the numbers actually treated. Women with unexplained recurrent miscarriage and implantation failure were included from 2015 to 2018 in three French university hospitals. Of 187 women treated for these two situations, 26 received intralipids, of median age 36 years (29 to 43). Among these 26 women, 16 had a history of recurrent implantation failure, with a median age of 37 years (29 to 43) and a median of 9.5 embryo transfers (3 to 19): a clinical pregnancy occurred in 9 of them (56%) after a transfer performed on intralipids, and 5 of these pregnancies (55%) resulted in a live birth. In the subgroup of the 10 women with recurrent miscarriage, with a median of 5 previous miscarriages (4 to 8), 7 live births (70%) were observed on intralipids, more often than in the 20 untreated women (p = 0.02), with no difference in age, in number of previous miscarriages or in associated treatments between the two groups. Arsène Mekinian is the 18th and last of 18 named authors 4.
These studies are, however, observational, with small numbers and without randomisation, and therefore do not make it possible to attribute the results with certainty to the treatment given.
The data from the available randomised trials moreover call for caution regarding the empirical use of immunomodulators. The only randomised trial of sufficient size published to date in this indication is negative, and it is not his: a double-blind placebo-controlled randomised trial conducted in eight fertility centres in China, with inclusions from November 2018 to August 2020 and published in JAMA in 2023, randomised 715 women with at least two unsuccessful transfers and aged under 38 years at the time of oocyte retrieval: 357 received 10 mg of prednisone per day and 358 a placebo. Live birth occurred in 37.8% of the women in the prednisone group (135 of 357) against 38.8% in the placebo group (139 of 358), an absolute difference of −1.0% (95% confidence interval −8.1 to 6.1; relative risk 0.97; confidence interval 0.81 to 1.17; p = 0.78). Two secondary outcomes are unfavourable to prednisone: biochemical pregnancy losses 17.3% against 9.9% (relative risk 1.75; confidence interval 1.03 to 2.99; p = 0.04) and preterm deliveries 11.8% against 5.5% (relative risk 2.14; confidence interval 1.00 to 4.58; p = 0.04) 11.
The two must be read together: prednisone is the treatment of half the treated cycles of the 2021 French cohort, where it is associated with 50% clinical pregnancies after transfer against 5% without treatment 3, and this difference observed without randomisation is not found on live birth when allocation is random. A high clinical pregnancy rate in an uncontrolled series is not proof of efficacy.
What is at stake now is moving from these observational signals to prospective controlled trials, with a homogeneous definition of the patients and clinically relevant outcomes.
CERTIFY
CERTIFY, NCT05930613, is a phase III randomised, double-blind, placebo-controlled trial, sponsored by AP-HP. It also carries the numbers EudraCT 2021-005309-28 and internal number APHP200031.
It assesses certolizumab pegol, a TNF antagonist, in women with unexplained repeated implantation failure.
The study aims to determine whether a targeted immunomodulation strategy can improve implantation and subsequent pregnancy outcomes in a precisely defined population. The registered protocol provides for 161 participants, women aged 18 to 40 years meeting the definition of recurrent implantation failure by at least three transfers of good-quality embryos according to the Istanbul criteria, and four subcutaneous injections of 400 mg of certolizumab or of placebo, five weeks and one week before embryo transfer and then three and seven weeks after it in the event of pregnancy. Blinding covers the participant and the investigator. The primary outcome is clinical pregnancy, defined by the presence of cardiac activity on ultrasound at five weeks of pregnancy; live birth, from 22 to 40 weeks, and miscarriage before 12 weeks are secondary outcomes. The actual start of the trial is dated 30 November 2023 and the end is estimated in April 2029. One participant in two receives a placebo.
To date, no result of the trial is published and no result is posted in the registry. No conclusion can therefore be drawn regarding the efficacy of certolizumab in this indication. The ClinicalTrials.gov record displays the status "recruiting", but it has not been updated since 11 December 2023 and its last verification by the sponsor dates from December 2023: it therefore does not prove that recruitment is continuing in September 2026.
That same record gives Arsène Mekinian as principal investigator; the page of the AP-HP clinical trials registry, consulted on 8 September 2026, names a coordinating investigator, Prof. Nathalie Chabbert-Buffet, and does not cite him. This page states what the two registries declare and does not decide between them; the exact wording of his role remains to be settled with him 8.
Multidisciplinary organisation
This work is developed within the department of internal medicine and clinical immunology of Saint-Antoine Hospital, AP-HP and Sorbonne Université, in collaboration with the obstetrics and gynaecology and reproductive medicine teams, in particular of Tenon Hospital, as well as with the immunology and biology teams.
The MEDLINE records in this field carry the form "service de médecine interne", the department of internal medicine of Saint-Antoine Hospital, attached to the medical and university department devoted to inflammation, immunopathology and biotherapies, designated DHU i2B until about 2019, then DMU 3iD and DMU i3 in the affiliations of his publications. That is the affiliation carried by Arsène Mekinian on his MEDLINE records in this field.
None of this work is the doing of an isolated department. The signatures bring together, on each of them, the department of obstetrics and gynaecology and reproductive medicine of Tenon Hospital, the department of obstetrics of Armand-Trousseau Hospital, and the laboratory haematology and immunology departments of Saint-Antoine Hospital, which perform the cytometry measurements. The published series carry the mention of three French university hospitals, in 2019 as in 2020, and the 2021 study is indexed as a multicentre study. This organisation makes it possible to combine reproductive medicine expertise with immunological assessment and clinical research.
The CERTIFY trial is sponsored by AP-HP. On the page of the AP-HP clinical trials registry consulted on 8 September 2026, the coordinating clinical research unit is that of Pitié-Salpêtrière, the coordinating centre is Tenon Hospital, the status displayed is "inclusions in progress" and the visible recruiting establishment is Saint-Antoine Hospital, a list of other sites being collapsed on the page. The ClinicalTrials.gov record, for its part, declares only one site, Saint-Antoine Hospital in Paris. The same AP-HP page states that the trial injections are given by nurses during internal medicine consultations.
Particularly complex situations may also be the subject of a multidisciplinary discussion, in particular within the national Infertility multidisciplinary team meeting of the FAI²R network. Arsène Mekinian states that he provides its coordination, according to the text sent by Arsène Mekinian on 9 September 2026; this meeting is described in none of his publications in this field, and the coordination he states he provides is named in no public source read.
The investigations and treatments described on this page come either under registered research protocols, with inclusion criteria published at the same time as the trial, or under decisions taken situation by situation with the assisted reproduction centre following the patient. This site takes no appointments and refers to no consultation.
What remains to be demonstrated
The participation of the immune system in implantation is established on the physiological level. On the other hand, the existence of a pathological immunological mechanism explaining certain repeated implantation failures remains difficult to characterise individually.
The definition itself is not settled. His team's 2021 article takes the absence of pregnancy after at least three transfers of good-quality embryos 3, the CERTIFY trial takes up this threshold relying on the Istanbul criteria 8, and the 2023 ESHRE recommendations by contrast decline to set a number of transfers 9. Three definitions coexist.
No isolated immunological biomarker makes it possible to select, in a sufficiently validated way, the patients likely to benefit from an immunomodulatory treatment. The 2019 work concludes that the proportion of blood NK cells measured before conception does not reflect the risk and should not serve as an indicator for management, and its authors call for large prospective studies to develop a reliable blood marker of immune dysregulation 2. The clinical value of several of the investigations proposed, in particular certain profiles of NK cells, cytokines or endometrial immune markers, remains to be established in reproducible prospective studies. Circulating and uterine NK cells must in particular be regarded as two distinct compartments: the 2019 work concerns those of the blood 2, whereas the 2026 review describes uterine NK cells as a central actor at the maternal-fetal interface without any threshold usable in practice following from it 7. The 2023 ESHRE recommendations, which are not his, classified the investigations and the interventions according to three colours, recommended, to be considered, not recommended in routine practice, with nineteen recommendations on investigations and thirteen on interventions 9.
The encouraging results of certain observational cohorts do not constitute a demonstration of therapeutic efficacy. In the 2021 cohort, sixty-four women, no randomisation, no control arm formed by random allocation: the treated transfers are compared with untreated transfers, often in the same women and at different moments of their course, which risks confusing the effect of treatment with that of time, of the number of attempts or of the selection of patients. The authors themselves describe their results as preliminary data and call for randomised studies 3. The same reservation applies to the 2020 series on intralipids, retrospective, concerning 16 women with implantation failure 4. And the only randomised trial of sufficient size published in this indication, in 715 women, is negative on live birth 11.
The ongoing trial says nothing yet. CERTIFY has posted no result, one participant in two receives a placebo, and the number actually included to date is given by no public source. The registry record has not been updated since 11 December 2023 and the AP-HP page displays no update date: the status "recruiting" cannot be presented here as an established fact in September 2026 8. Finally, Arsène Mekinian's exact role in this trial differs between two public sources, and this page does not settle it.
The research priorities are therefore to better define the patients, validate the biomarkers and assess the interventions in randomised trials, giving preference to clinically relevant outcomes, in particular live birth.
References
- Unexplained Recurrent Miscarriage and Recurrent Implantation Failure: Is There a Place for Immunomodulation? — American Journal of Reproductive Immunology, 2016 1st of 9 named authors. Review and meta-analysis of the literature.cited 117 times
- Proportion of Cytotoxic Peripheral Blood Natural Killer Cells and T-Cell Large Granular Lymphocytes in Recurrent Miscarriage and Repeated Implantation Failure: Case-Control Study and Meta-analysis — Archivum Immunologiae et Therapiae Experimentalis, 2019 16th of 16 named authors, corresponding author. Case-control study and meta-analysis.cited 24 times
- Unexplained recurrent implantation failures: Predictive factors of pregnancy and therapeutic management from a French multicentre study — Journal of Reproductive Immunology, 2021 15th of 15 named authors, corresponding author. Multicentre observational study, without randomisation.
- Intralipid therapy for unexplained recurrent miscarriage and implantation failure: Case-series and literature review — European Journal of Obstetrics & Gynecology and Reproductive Biology, 2020 18th of 18 named authors, corresponding author. Retrospective cohort and review of the literature.cited 19 times
- Endometriosis with infertility: A comprehensive review on the role of immune deregulation and immunomodulation therapy — American Journal of Reproductive Immunology, 2021 12th of 12 named authors. Review of the literature.
- Role of miRNAs in Normal Endometrium and in Endometrial Disorders: Comprehensive Review — Journal of Clinical Medicine, 2021 4th of 8 named authors. Review of the literature free full text
- Immunologie de la fécondation, de l'implantation et de la grossesse : interactions clés et perspectives thérapeutiques (article published in French; MEDLINE title in English: Immunology of fertilization, implantation and pregnancy: Key interactions and therapeutic perspectives) — La Revue de médecine interne, 2026 4th of 5 named authors. Review of the literature.
- Efficacy of Certolizumab in Women With Unexplained Recurrent Implantation Failure: a Double-blind Randomized Controlled Trial (CERTIFY) — ClinicalTrials.gov registry, 2023 Arsène Mekinian principal investigator according to the registry record. Phase 3 randomised double-blind placebo-controlled trial, sponsor AP-HP, 161 participants planned, no result posted, record updated on 11 December 2023.
- ESHRE good practice recommendations on recurrent implantation failure — Human Reproduction Open, 2023 ESHRE Working Group on Recurrent Implantation Failure, 10 named authors, Arsène Mekinian is not a signatory. Open access free full text
- Recurrent implantation failure: a committee opinion — Fertility and Sterility, 2026 Practice Committee of the American Society for Reproductive Medicine, consortium text with no named author, Arsène Mekinian is not a signatory. Paywalled at the publisher, only the abstract has been read.
- Prednisone vs Placebo and Live Birth in Patients With Recurrent Implantation Failure Undergoing In Vitro Fertilization: A Randomized Clinical Trial — JAMA, 2023 Sun Y et al., 30 named authors, Arsène Mekinian is not a signatory. Randomised trial, registry ChiCTR1800018783 free full text