Intravenous immunoglobulin and prednisolone to women with unexplained recurrent pregnancy loss after assisted reproductive technology treatment: a randomised, double-blind, placebo-controlled trial.

Recurrent pregnancy loss (RPL), defined as two or more pregnancy losses, affects a substantial and increasing proportion of couples of reproductive age, 1 and it is associated with considerable medical, emotional and socioeconomic burdens. 2 Known risk factors for RPL include uterine malformations, thyroid dysfunction, acquired thrombophilia and chromosomal anomalies. At least one risk factor is identified in only 40%–50% of couples, leaving the majority of patients with unexplained RPL (uRPL). Immunological anomalies are often suspected to be involved in the pathogenesis of uRPL. Therefore, investigation of immunological aberrations and the effects of immunomodulating treatments currently dominates research in uRPL. 2 Prednisolone and intravenous immunoglobulin (P&IVIg) are immunomodulatory treatments with distinct but complementary mechanisms of action. In pregnant RPL patients, IVIg affects the distribution and activity of peripheral blood leucocyte subsets, that is, suppressing natural killer (NK) cell activity, increasing T regulatory (Treg) cell percentages, neutralising antibodies and modulating Fc receptor-mediated effects. 3 These alterations are associated with improved live birth in some studies. 4 Prednisolone, a synthetic glucocorticoid mimicking endogenous cortisol, exerts broad anti-inflammatory and immunomodulatory effects. It differentially modulates the generation, differentiation and function of leucocyte subsets, with significant implications for immune tolerance and sensitisation. Key mechanisms include suppression of NK cell cytotoxicity, reduction of Thelper (Th1) and cytotoxic NK cell levels, promotion of tolerogenic subsets such as Th2 and Treg cells in peripheral blood or endometrium, 5 8 downregulation of major histocompatibility gene class II expression, reduction of immunoglobulin production by plasma cells, and enhancement of trophoblast outgrowth in vitro. 8 10 A Cochrane review suggested that prednisolone during ovarian stimulation may increase clinical pregnancy rates, 11 while IVIg may support pregnancy maintenance when administered around the time of implantation. 12 The rationale for combining P&IVIg lies in their superior clinical effects in several immune-mediated disorders, achieved by targeting complementary immunological pathways, which may produce greater benefit than either agent alone. 13 17 Systematic reviews of the therapeutic effect of IVIg in RPL have shown conflicting results. Egerup et al 18 found that IVIg compared with placebo had a borderline (p=0.06) significant therapeutic effect in reducing pregnancy loss rate after secondary RPL (RPL preceded by a birth after GW 24) but not after primary RPL, whereas Wang et al 19 found a beneficial effect (p=0.05) of IVIg over placebo on live birth rate (LBR) in the total RPL population, with a stronger effect when IVIg was administered before conception. Recently, a randomised controlled trial (RCT) found a significant effect of high dose IVIg starting in early pregnancy among patients with ≥4 consecutive pregnancy losses. 20 Meta-analyses of the effect of prednisolone in patients suffering from RPL or infertility have provided conflicting results. 1121 23 Our cohort pilot study of patients with RPL after assisted reproduction technology (ART) treatment demonstrated an encouraging 36.5% LBR when patients were treated with P&IVIg during their subsequent embryo transfer (ET). 24 Notably, both our pilot study population and the present trial population of women with RPL after ART, with a median of 3–4 prior pregnancy losses, are not representative of the general RPL population. These women frequently also have infertility-related factors and previous failed ART cycles, which together are associated with a lower LBR than that typically observed in unselected RPL cohorts. 24 We performed this placebo-controlled RCT in order to evaluate the effects of P&IVIg on the ongoing pregnancy rate (OPR) in patients with a history of uRPL after ART.

This single-centre, double-blind, parallel-group, placebo-controlled RCT was conducted at the RPL Centre of Western Denmark, located at Aalborg University Hospital—a public tertiary care unit. The protocol and statistical analysis plan (SAP) were published before the interim analysis was performed. 25 The trial was performed in accordance with the Declaration of Helsinki and Good Clinical Practice (GCP) guidelines. The trial oversight and monitoring were outsourced to an independent Data Monitoring Committee (the Danish GCP unit) that conducted systematic on-site inspections after the randomisation of approximately every 10th participant to ensure data accuracy and that the investigators complied with Danish legislation and International Council for Harmonisation of Technical Requirements for Pharmaceutical for Human Use guidelines for GCP. Written and oral informed consents were obtained from all participants before randomisation. Each participant could enter the trial only once. Patients undergoing ART—defined as IVF, ICSI or frozen embryo/blastocyst transfer—were screened for eligibility by the principal investigators (PIs) if they had experienced ≥2 consecutive pregnancy losses before gestational week (GW) 10+0 in ART-induced pregnancies with their current partner or an egg/sperm donor. Exclusion criteria included body mass index (BMI) ≥35 kg/m 2 ; age ≥41 years; significant uterine malformation(s); known parental chromosomal abnormality; ≥2 previous pregnancies with confirmed fetal aneuploidy; anti-Müllerian hormone (AMH)<4 p m o l / l ; trial drug contraindications; treatment with drugs interacting with trial drugs; comorbidities that require particular caution when receiving the trial drugs; clinical or biochemical indication for levothyroxine or heparin treatment during pregnancy including hypothyroidism and antiphospholipid syndrome; and prior treatment with IVIg. An AMH level <4 pmol/L was not an exclusion criterion for women receiving egg donation. Women utilising sperm or egg donation or preimplantation genetic testing (PGT) in the index ART cycle were eligible only if these methods had been used in previous ART cycles leading to ≥2 pregnancy losses. Interventions or add-ons that had not been used in at least one prior unsuccessful ART cycle were not permitted during trial participation. The participant’s ART treatment was determined exclusively by her affiliated fertility clinic. Use of IVF add-ons in Danish IVF clinics, that have not been proven safe and effective, is generally limited. Potentially eligible patients were provided with information about the trial by the PIs during their first appointment in the RPL Centre after the diagnostic work-up. Thus, all baseline characteristics were collected during the diagnostic work-up prior to achievement of oral and written informed consent and randomisation. Each participant could enter the trial only once. Participants were randomly assigned in a 1:1 ratio to P&IVIg or placebo according to a computer-generated randomisation list, which included randomly selected block sizes, created by the Hospital Pharmacy North Denmark Region (HPNDR) ( figure 1 ). Consecutively numbered, opaque, sealed paper bags containing the type of trial tablet medication associated with the randomisation list were subsequently prepared by authorised personnel at the HPNDR. Once consent was obtained and the start of the next ART cycle approached, the PIs enrolled the participant, assigned the trial ID numbers in chronological sequence, and handed over the sealed numbered paper bag with a tablet container and a leaflet with sections for recording adverse events (AEs) and daily medication intake to monitor treatment compliance and safety. The randomisation number corresponded to the participant’s trial ID number. The filled-in leaflet and tablet container were returned to the PIs after treatment cessation, and the remaining tablets were counted by personnel not involved with the trial or its participants. A copy of the randomisation list was held by the laboratory at the Department of Clinical Immunology (DCI) preparing the intravenous trial medicine, which occurred approximately 2 hours prior to infusion. DCI personnel, who had no contact with participants or RPL clinic personnel and no access to clinical data, received standardised email orders from a PI with the participant’s allocation number, initials, weight and infusion date. The personnel involved in participant inclusion, treatment administration, outcome assessment and data analysis were blinded to group assignment and had no access to the randomisation list. The prednisolone and placebo tablets, along with their containers and paper bags, were identical in appearance. One tablet contained 5 mg prednisolone, or 85 mg lactose monohydrate, 86 mg potato starch, 8.1 mg talc, 3 mg gelatine and 0.9 mg magnesium stearate, respectively. Albumin was selected as the IVIg comparator because of its similar appearance and physical properties, such as foam formation. To mask the slight colour difference, ultraviolet-protected yellow ethylene vinyl acetate bags and infusion kits were used. Each participant began treatment with one tablet daily, starting within the first 3 days of the menstrual cycle prior to the upcoming ET, and continued until the day of ET. From the day of ET until GW 8+0, the dosage was increased to two tablets daily. Tablet treatment was then tapered by administering one tablet daily for 4 days followed by complete cessation at GW 8+4. In the case of a negative pregnancy test or pregnancy loss, tapering was accelerated. If a pregnancy loss occurred, a cytogenetic analysis using array-CGH was performed on products of conception whenever possible ( figure 2 ). The first intravenous infusion of IVIg or albumin was administered within a window from five working days prior to and two working days after ET. Each infusion contained approximately 4 mL/kg of either 10% IVIg (Privigen, CSL Behring) (0.4 g/kg) or 5% human albumin (CSL Behring), with volume adjusted by pre-pregnancy body weight (BW): 250 mL for BW≤70 kg, 300 mL for BW 70–85 kg and 350 mL for BW≥85 kg. Approximately 2 weeks after ET, plasma human chorionic gonadotropin (hCG) levels were measured. If the plasma-hCG concentration exceeded 5 IU/L, the measurement was repeated within 24–48 hours, and if an increase of plasma-hCG of ≥30% per 24 hours and a level ≥50 IU was registered, the patient continued trial treatment. Additional intravenous infusions were administered in GW 5, 6 and 7, respectively, with the same dosage as the prepregnancy infusion. A plasma-hCG measurement or ultrasound scan was performed prior to each infusion to confirm the viability of the pregnancy. A blood sample was collected right before the first and the third infusion (and for non-pregnant participants; at the corresponding timepoint) for prespecified immunological analyses and establishment of a biobank. 25 After GW 8, participants were offered routine monitoring at the RPL clinic in addition to standard Danish antenatal care, which begins with a first trimester risk assessment (FTRA) including a nuchal translucency scan between GW 11+4 and 13+6. The preplanned primary outcomes were defined as the OPR at the FTRA among all participants and among those who conceived (defined as a plasma-hCG >50 IU), respectively, in an intention-to-treat (ITT) (including all randomised participants) and a per-protocol (PP) analysis (restricted to participants with no major protocol deviations). Furthermore, these outcomes were preplanned to be reported with and without exclusion of participants with aneuploid miscarriages. The primary endpoints were the relative difference, the absolute risk reduction and the adjusted relative risk for the OPRs. The secondary objective was to evaluate the treatment safety, including an assessment of AEs and reactions (AEs/ARs), obstetric complications and perinatal outcomes. The registration of AEs/ARs was censored up until 6 months after the last intravenous infusion in non-pregnant participants and until delivery in pregnant participants. The maternal AEs/ARs included both pregnancy-related and non-pregnancy-related; expected and unexpected; and serious and non-serious AEs/ARs (SAEs/SARs). The latter two categories were defined in accordance with EUR-Lex-CT-3 (Directive 2001/20/EC). A detailed list of all secondary outcomes is presented in the SAP. 25 Obstetric complications included the incidence of early (<GW 12+0) and late pregnancy loss (GW 12+0 to 23+6), stillbirth, pre-eclampsia, gestational hypertension, gestational diabetes mellitus and instrumental delivery. The perinatal outcomes included birth weight, gestational age at birth, sex and the incidence of congenital deformity, abnormal fetal karyotype, preterm birth (<GW 37+0), low birth weight (<2500 g), small for gestational age (birth weight below 2 SDs ≈ 22% of the Danish mean birth weight for gestational age and sex) and admission to a neonatal care unit. The primary and secondary outcomes were collected through online questionnaires and medical records. The participants received an online, follow-up questionnaire right after the FTRA and after the due date, respectively, which collected data on the obstetric and perinatal outcomes. All pertinent clinical data were collected and managed by the PIs using an electronic case report form (e-CRF) in Research Electronic Data Capture; a secure web-based platform hosted at Aalborg University Hospital. 26 27 The sample size calculation was based on the OPR following ET among all randomised participants. We anticipated an absolute increase in the OPR from 12% in the placebo group 24 to 40% in the treatment group, 28 based on findings from a similar population. Using an α -level of 5% and a β -level of 20%, the required sample size was calculated to be 37 participants in each group adhering to the treatment protocol. To account for major protocol deviations, the enrolment plan included one extra participant per identified case of major deviation. 25 The full SAP describing all preplanned analyses in detail is accessible as supplemental material to the protocol. 25 A data monitoring committee continuously assessed the e-CRF and source data to verify collected data. A preplanned interim analysis was conducted by independent statisticians after 38 participants with no major protocol deviations had completed the trial treatment to ensure participant safety by comparing the rates of SARs and SAEs between the two allocation groups. Stopping guideline was a difference of ≥3 SAEs between groups. The results were presented to the PIs in a blinded manner and since the stopping criterion was not met, the trial continued. The sample-size calculation was not adjusted in the interim analysis because the analysis was conducted solely as a preplanned safety assessment focusing on the number of women experiencing SARs, SAEs and suspected unexpected serious adverse reactions (SUSARs) in the two allocation groups. The analyses after complete data collection were carried out by the PIs and independent statisticians in accordance with the SAP using Stata/MP V.18·0 (StataCorp). A p ≤ 0.05 was considered statistically significant. The primary outcomes were analysed in the ITT and PP groups, while the analyses of secondary outcomes were performed only in the PP group. The overall incidence of AEs/ARs and of specific AEs/ARs experienced by ≥2 participants in at least one group was reported. The perinatal outcomes were compared excluding twins. For more details on pre-planned analyses and reporting of results, we refer to the SAP. 25 The categorical outcome variables were compared between groups using a χ 2 test or Fisher’s exact test when less than five participants were expected in at least one group. An adjusted relative risk for achieving a normal, viable pregnancy at the FTRA between groups was estimated using Poisson regression with adjustment for relevant confounders including BMI, smoking and age. For continuous secondary outcomes, data distribution was evaluated using QQ-plots and histograms in accordance with the SAP to guide the choice between parametric and non-parametric methods. Continuous data variables were compared between groups using an unpaired two-sample Mann-Whitney U test or t-test depending on data distribution. Preplanned subgroup analyses included OPR among pRPL and sRPL; and among women with and without autoantibodies, separately. The trial was registered at ClinicalTrials.gov ( NCT04701034 ) and on WHO (WHO unique trial number: U1111-1273-8585) before the first participant in the trial was included.

From 28 January 2021 to 23 February 2024, 80 women were equally randomised to P&IVIg or placebo, and final follow-up ended on 18 April 2024. The PP analyses included 74 (92.5%) participants, as 3/40 (7.5%) participants in each group had major protocol deviations ( figure 1 ). The leaflet, diary and count of remaining tablets after completing trial participation revealed full compliance, that is, no missing tablet and infusion administrations in the PP population. The demographics and baseline characteristics of the participants in the ITT analysis were well balanced between the two allocation groups ( table 1 ), and the parameters were also equally distributed in the PP population and among those who conceived ( online supplemental tables 1-2 ). In the ITT P&IVIg group, 18/40 (45.0%) and in the placebo group 14/40 (35.0%), respectively, had received prednisolone treatment in at least one previous unsuccessful ART cycle. In the PP population, 68 participants received the first infusion before ET, whereas the remaining six participants received the first infusion within 24 hours after ET. Characteristics of the ART treatment used in the index ART cycle were comparable between the two allocation groups ( table 2 ). A failed ART cycle was defined as a negative pregnancy test after an embryo transfer. Some couples had >1 indications for ART. ART, artificial reproductive technology; BMI, body mass index; PCOS, polycystic ovarian syndrome; PGT-M, preimplantation genetic testing for monogenic diseases; P&IVIg, prednisolone and intravenous immunoglobulin. ART, artificial reproductive technology; FET, frozen embryo/blastocyst transfer; ICSI, intracytoplasmic sperm injection; IVF, in vitro fertilisation; PGT-A/M, preimplantation genetic testing for aneuploidy/monogenic disease; P&IVIg, prednisolone and intravenous immunoglobulin; PP, per-protocol. All participants with a normal, viable fetus at the FTRA successfully gave birth, as all pregnancy losses happened in the first trimester. Thus, the OPR was identical to the LBR. In the ITT group, the rate of positive pregnancy tests (p-hCG>50 IU) was 30.0% in the P&IVIg group and 35.0% in the placebo group (p=0.63). The crude RR of OPR between the two allocation groups including all randomised participants was 1.67 (95% CI 0.67 to 4.15), the absolute risk reduction was 10.00% (95% CI: −7.39% to 27.39%), and the adjusted RR of OPR was 1.86 (95% CI 0.70 to 4.93). The number needed to treat (NNT) for one additional live birth was 10. In the PP population, the crude RR of OPR was 1.33 (95% CI 0.51 to 3.46), the absolute risk reduction was 5.40% (95% CI −12.39% to 23.21%), the adjusted RR was 1.35 (95% CI 0.49 to 3.72) and the NNT was 19. When the analysis was confined to participants who conceived after ET in the ITT population, the OPR was higher in the P&IVIg group than in the placebo group (p=0.05). Thus, in this analysis, the crude RR was 1.94 (95% CI 1.01 to 3.75), the absolute risk reduction was 40.47% (95% CI 7.06% to 73.89%); the adjusted RR of OPR was 2.13 (95% CI 0.99 to 4.59), and the NNT was three for one additional live birth. Six (66.7%) pregnancy losses were confirmed by ultrasound while three (33.3%) were biochemical. Cytogenetic analysis using array-CGH was conducted in five clinical pregnancy losses. Three cases showed fetal aneuploidy (trisomy 15, 16 and 22, respectively), one case had placental mosaicism (20%–30% monosomy X) and one case was euploid (46,XX). Excluding the three pregnancy losses with confirmed clinically significant chromosomal anomalies (the modified analysis), the OPR in the ITT population was 90.90% in the P&IVIg group and 50.00% in the placebo group (absolute risk reduction: 40.90%, 95% CI 7.91% to 73.90%) ( table 3 ). As a post hoc analysis, the OPRs in PP patients with <5 and ≥5 pregnancy losses were analysed, and no significant differences were found, but numbers were too small to draw any firm conclusions ( online supplemental table 3 ). The ongoing pregnancy rates were identical to the live birth rates. Modified analysis excluded pregnancy losses with confirmed embryonal aneuploidy. ITT, intention-to-treat; OPR, ongoing pregnancy rate; P&IVIg, prednisolone and intravenous immunoglobulin; PP, per-protocol; RPL, recurrent pregnancy loss. Among participants who achieved a live birth, the gestational age at delivery and birth weight were non-significantly higher in the P&IVIg group. The incidence of obstetric complications was not significantly different between the two allocation groups ( table 4 ). One participant in the placebo group, who had a single ET, gave birth in GW 35+2 to a pair of twins, which were excluded from the comparisons of perinatal outcomes but not the analysis of obstetrical complications. Perinatal outcomes shown among all singleton births. Obstetric complications shown among all births. Among patients in the PP group who conceived. P&IVIg, prednisolone and intravenous immunoglobulin. The incidence of AEs/ARs did not differ between the two groups ( table 5 ). Two SAEs occurred in the active treatment group and one in the placebo group. One participant experienced vaginal haemorrhage at GW 11+4 and was admitted to the hospital for observation overnight. The fetus was not affected. This event was interpreted during the blinded data assessment as related to pregnancy and not trial drugs. One participant developed signs of chorioamnionitis, including fever and foul-smelling amniotic fluid, during delivery in GW 40+0. A caesarean section was performed, and both the mother and her otherwise healthy baby were treated with antibiotics for 12 days postdelivery. They had no sequelae on discharge. The events were reported as a SAE but deemed unrelated to the trial drugs before breaking the blind. One participant in the placebo group gave birth to a girl who required resuscitation and experienced seizures immediately after delivery. An MRI scan of the brain showed signs of ischaemia. The girl is now 12 months old, developing normally, and has had no seizures or need for medication. No withdrawals or loss to follow-up occurred due to AEs/ARs, and no maternal or perinatal deaths were observed. Specific adverse events and reactions with a minimum 2 cases in at least one group are listed and compared between groups. One patient admitted to hospital with vaginal haemorrhage in gestational week 11 and one delivery complicated by chorioamnionitis in mother and child without sequelae. One child with postpartum seizures and signs of intracerebral ischaemia by MR scan; the child is healthy with no need of medicine 14 months post partum. P&IVIg, prednisolone and intravenous immunoglobulin.

This trial found that P&IVIg compared with placebo in women with RPL undergoing ART treatment did not significantly affect the reproductive outcome compared with placebo. However, it did improve reproductive outcome among patients who conceived after ART. The reproductive outcome, defined as OPR per ET in GW 13, was identical to the LBR since no late pregnancy losses and stillbirths occurred. These findings suggest that P&IVIg may not enhance endometrial receptivity but rather sustain implantation and support the continuation of ART-induced pregnancies, thereby reducing the risk of miscarriage. The comparable pregnancy rates between the two allocation groups suggest that factors affecting the probability of conception, such as endometrial receptivity, embryo quality and sperm quality, were equally distributed by randomisation. IVIg is a relatively costly and time-consuming treatment; however, with the LBR increasing from 42.9% to 83.3% among women with RPL, in couples with a previous long and expensive ART treatments, the benefits of P&IVIg may outweigh its disadvantages. This trial focused on immunomodulatory treatment during the window of implantation and early pregnancy based on the hypothesis that the level and composition of specific immune cells recruited to the decidualising endometrium during these phases are decisive in determining whether maternal immune tolerance or harmful sensitisation to the fetal alloantigens is elicited. 29 33 This hypothesis is based on findings from basic research studies on decidualisation and implantation in fertile women and patients with RPL. 29 33 Furthermore, previous RCTs on the efficacy of IVIg in RPL point towards a positive effect primarily when administered in very early pregnancy 20 34 or before implantation 28 35 when the analysis is confined to patients who conceived. In the RCTs by Coulam et al 35 and Stephenson and Fluker 28 , IVIg or saline was administered before conception. Among women who subsequently became pregnant, the LBRs were 80.0% vs 43.9% (p=0.25) and 62.0% vs 34.4% (p=0.03), respectively, emphasising the potential beneficial effect of IVIg when administered prior to implantation. However, the methodology of these RCTs 20 28 34 35 differs from that of the present trial, as most of the included RPL patients conceived naturally and neither of them received glucocorticoids. Therefore, further trials similar to the present RCT are needed to evaluate the credibility of the investigators’ hypothesis that immunomodulation improves pregnancy continuation rather than the probability of conception (fecundability). Most RCTs of IVIg efficacy in RPL patients implemented treatment protocols with only postconception IVIg infusions, 1234 36 41 and all except one 12 failed to demonstrate a significant impact of IVIg on the reproductive outcome. The non-significant findings in prior studies may be a consequence of not timing the treatment initiation to the critical stages during implantation and very early pregnancy. Immunomodulatory therapy may need to coincide with these critical stages in patients with immune-mediated RPL to be effective. If this critical period involves the first interaction between maternal immune cells and fetal trophoblasts, delaying treatment beyond this stage may be too late to influence the outcome. Human albumin was selected as the placebo to ensure complete blinding of both participants and clinical staff, as alternatives such as saline are visually distinguishable from IVIg due to differences in colour and foam formation. Previous RCTs have similarly prioritised optimal blinding over the minimal infection risk associated with albumin use. 35 3739 42 The trial design of our RCT was deliberately chosen to include only patients undergoing ART, as this allowed for the administration of IVIg immediately before implantation, thereby testing the hypothesis that immunomodulation at this critical timepoint is crucial. ART was not employed to any participants as a treatment of RPL per se but rather to manage infertility attributable to the factors outlined in table 1 . The rate of obstetric complications and adverse perinatal outcomes did not differ significantly between the two allocation groups. However, it is important to note that the trial was not powered to evaluate these outcomes, with only eight and six ongoing pregnancies in each group, respectively. The mean birth weight and gestational age appeared to be higher in the active treatment group. These findings align with those of our previous RCT. 40 Furthermore, no statistically significant or clinically relevant differences were observed in the incidence of overall and specific AEs or ARs between the two groups. The generalisability of our results is influenced by several factors. First, the patients were referred from several public and private fertility clinics across Denmark, indicating that the findings might have been independent of the ART protocols used and that the trial population was a geographically representative cohort. Moreover, age, BMI, tobacco use, and the distribution of ethnicity of the participants mirror that of the Danish population ( table 1 ). Moreover, the P&IVIg treatment protocol used in this trial closely mirrors the regimen that has been used in our RPL clinic for many years, demonstrating its feasibility for clinical practice. Additionally, the randomisation and blinding procedures employed (including the use of human albumin rather than saline as intravenous placebo) ensure that our results are robust and not influenced by selection or performance bias, further strengthening the general applicability of our findings. Noteworthy, the RPL patients included may not be fully representative of the broader RPL population, as several also presented with secondary infertility or implantation failure; hence, extrapolation should be made with caution. Consequently, the results may not be directly applicable to RPL patients who conceive naturally, since timing the administration of IVIg to coincide with the window of implantation could present challenges in these cases. Yet, the critical nature of this timing remains unknown. Although the findings are promising, it is crucial to acknowledge the need for further studies involving diverse populations and clinical settings to more thoroughly assess the external validity of our results. Additionally, exploring whether the treatment is more effective in subgroups with specific immunological abnormalities warrants investigation. This trial has some limitations. Sample-size calculation was challenging due to the absence of prior data from similar studies conducted in comparable populations, particularly regarding OPR under placebo or no additive treatment. The effect size was based on previously reported LBRs among the total study population 24 rather than specifically among those who conceived after ET. The anticipated 12% OPR in the placebo group was derived from Stephenson and Fluker, 28 reflecting ART outcomes 25 years ago. Their participants, described with repeated IVF failures, had experienced failed implantation, biochemical pregnanciesand early miscarriage, and had broadly similar reproductive histories as our trial population, but differences in ART protocols, laboratory techniques and patient selection over time introduce uncertainty in the expected effect size. It is not possible to determine the extent to which the treatment efficacy on OPR is attributable to IVIg, prednisolone or the combination, and whether these drugs act antagonistically, additively or synergistically with each other. However, 45% of the participants in the P&IVIg group had been treated with prednisolone in previous ART cycles without success, so the 83.3% LBR obtained after participation in this group may be attributed to the addition of IVIg. We suggest future trials should include four arms: one placebo, one combination therapy and one monotherapy for each drug. However, this would require a substantial number of participants. Moreover, embryonal aneuploidy is a major cause of implantation failure and pregnancy loss, but in many European countries, PGT for aneuploidies (PGT-A) is not routinely used due to its high costs and uncertain effectiveness. 43 44 In Denmark, PGT-A is offered by self-payment in only one private IVF clinic, accounting for less than 1% of ART cycles. 45 46 For trial feasibility and representativeness, we did not include a criterion that required the use of PGT-A-screened embryos in the index ART treatment. The recommendations on using immunomodulatory treatments to RPL patients are conflicting between international societies of reproductive medicine. The updated RPL guideline by ESHRE 2 included a conditional recommendation for IVIg to RPL patients with a history of four or more pregnancy losses based on the findings from one RCT. 12 Furthermore, despite prednisolone not being recommended, the ESHRE RPL guideline underlines that the current evidence points towards some beneficial effects of low-dose prednisolone and underscores the need for further investigations of its efficacy. Thus, both IVIg and prednisolone are considered to be potentially effective drugs to treat women with RPL where no other treatment has had a verified effect. The rationale for our combined immunomodulatory regimen was based on evidence from both corticosteroid and IVIg studies in reproductive immunology. Prednisolone 5 mg daily was administered before ET based on an RCT showing improved clinical pregnancy and lower cycle cancellation rates with 1 mg/day dexamethasone (≈5 mg prednisolone) before ET, 47 while keeping the dose low enough to avoid hypothalamic-pituitary-adrenal axis dysfunction or other corticosteroid-related side effects. 48 The prednisolone dose was increased from the day of ET onwards guided by prior studies in women with RPL, 49 50 aiming to provide sufficient immunomodulation to temper an exaggerated maternal inflammatory response to fetal antigens, without impairing implantation, decidualisation or the physiological inflammatory processes essential during the early days of implantation. 51 The selected IVIg dose of approximately 0.4 g/kg was similarly based on regimens commonly used in autoimmune disorders, balancing immunomodulatory efficacy with a low risk of ARs. IVIg was administered weekly to maintain a consistent immunomodulatory effect during the dynamic peri-implantation period, when immune tolerance and embryo–maternal communication are established. Although the plasma half-life of IVIg is 3–4 weeks, its effects on cytokine release, Fc receptor expression and NK-cell cytotoxicity are transient and dose-dependent, making weekly administration necessary to sustain adequate immune modulation throughout implantation and early placentation. While higher-dose IVIg over 3–5 consecutive days may be more comparable to regimens used in autoimmune diseases, 12 such a protocol was not feasible in our centre and was associated with unacceptable side effects in preliminary testing. Studies using longer dosing intervals (3–4 weeks) have generally failed to demonstrate efficacy in RPL. 2836 38 52 The combined regimen of prednisolone and weekly IVIg was therefore considered the most suitable strategy to achieve complementary immunomodulatory effects during the critical implantation window. In conclusion, our findings suggest that the P&IVIg treatment in RPL patients does not affect the likelihood of conception following ART. However, it may contribute to sustaining implantation and supporting continuation of pregnancy. Further research is needed to understand which specific interactions between maternal immune cells and embryonal antigens are critical for early and later stages of implantation. Additionally, it is essential to explore how the immune cells respond to immunomodulatory treatment to optimise immunotherapy protocols, including selection of target groups, timing of treatment initiation, dose and dosing interval, in order to achieve the desired alterations in maternal-fetal immune interactions. Future protocol modulations might also consider the incorporation of PGT-A.