Intralipid infusion at time of embryo transfer in women with history of recurrent implantation failure: A systematic review and meta‐analysis

Recurrent implantation failure (RIF) affects 10% of couples undergoing assisted conception, often due to poor endometrial receptivity. We conducted a systematic review and meta‐analysis to evaluate the effectiveness of Intra‐venous intralipid (IVI) in improving pregnancy rates in women with history of RIF using.


Introduction
More and more couples rely on assisted reproductive technology (ART) to become pregnant worldwide, with an annually rising numbers of cycles performed. 1 Still, optimizing the implantation process following embryo transfer remains a clinical challenge with 10% of couples undergoing ART affected by recurrent implantation failure (RIF). 2, 3 RIF, defined as the repeated implantation failure in spite of the transfer of good quality embryos following ART treatment, 4 continues to be a clinical dilemma contributing to increased morbidity and anxiety in affected couples. Several factors are thought to increase the risk of RIF such as embryo quality, maternal thrombophilia, uterine abnormalities, and endocrine disorders. 4,5 Still, we do not have a comprehensive understanding of the implantation process despite extensive research. 6 Recently, our increased understanding of the implantation window characteristics governed by subtitle yet highly regulated pro-and antiinflammatory responses 7,8 highlighted the potential effect of aberrant immune responses on implantation and pregnancy outcomes. 9 The role of immunotherapy to regulate the maternal immune system has received much attention recently, with a presumed effect in improving endometrial receptivity and the chances for successful conception. 10 Mothers with history of recurrent pregnancy loss and implantation failure seem to have an impaired immunological response driven by an increased activity of uterine natural killer (uNK), macrophages, and T1 helper cells elevating cytokine production and cytotoxicity in the endometrium. [11][12][13] However, effective screening and diagnostic methods are lacking which limits the process of patient and treatment matching. Various immunological therapies have been evaluated as ART add-on treatments, but evidence of their value to the general population remains unclear. 14 Intravenous intralipid (IVI), a fat-based emulsion of soybean oil, glycerine, phospholipids, egg, and polyunsaturated fatty acids, has been evaluated in several trials as a potential intervention to suppress the maternal immunological response, 15, 16 which seems to be mediated by reduced monocyte endothelial adhesion and pro-inflammatory cytokine generation. 16 Although its overall mechanistic effect on the immune system remains unclear, several studies support its role in downregulating the uNK cell macrophages pro-inflammatory mediators especially T1 helper cells. 15 This seems to be mediated by the interaction between the intralipid fatty acids of and the uNK cell peroxisome proliferator-activated receptors which decrease uNK cytotoxicity through reduced secretion of interferon-γ, demonstrated both in vivo and in vitro. 17,18 As such, the value and safety of selectively using IVI in women with the history of RIF to improve their reproductive outcomes need careful evaluation. We aimed to examine this hypothesis by systematically reviewing the literature for randomized trials evaluating the effectiveness of IVI to increase the chance of successful conception post ART treatments.

Materials and Methods
We performed this systematic review using a prospectively registered protocol (CRD42019148517) and reported in accordance with the PRISMA guidelines. 19

Literature search
We searched major electronic databases (MEDLINE, EMBASE, and Cochrane CENTRAL) for any randomized trials meeting our inclusion criteria from inception until September 2020.
We used the following MeSH search terms and combined them using the Boolean operators AND/OR to screen for relevant trials ("intralipid," "pregnancy outcomes," "pregnancy," "biochemical pregnancy," "clinical pregnancy," "ongoing pregnancy," "live birth," "birth," "pregnancy loss," "implantation failure," "miscarriage," "reproductive technology," "ART," "assisted reproductive technology," "IVF," "in vitro fertilisation," "ICSI," "intracytoplasmic sperm injection," and "embryo transfer"). No search filters or language restrictions were applied. We manually searched the bibliographies of any relevant articles to identify any missing studies. We contacted two groups of authors 20, 21 seeking further information before confirming eligibility for inclusion in our review.

Study inclusion and data extraction
Two independent reviewers (M.P. Rimmer and N. Black) completed the study selection and inclusion process in two stages. Initially, we screened titles and abstracts to identify potentially relevant trials and then reviewed the full texts against our inclusion criteria. We included all randomized trials evaluating the use of IVI during IVF/ICSI in women with history of RIF. Any discrepancies were resolved in consensus with a third reviewer (B.H. Al Wattar). We excluded non-randomized studies, animal studies, and review articles. We extracted data in duplicate (M.P. Rimmer and N. Black) using a piloted electronic data extraction tool reporting on the number of participants, inclusion and exclusion criteria, treatment protocols, clinical pregnancy or implantation rate (defined as biochemical testing or ultrasound), and live birth or ongoing pregnancy beyond 12 weeks' gestation.

Risk of bias
Two reviewers (M.P. Rimmer and N. Black) assessed the methodological quality of the included RCTs using the Cochrane risk of bias tool. 22 We evaluated each study in five domains: randomization and sequence generation, allocation concealment, blinding and outcome assessment, completeness of outcome data, and selective outcome reporting.

Statistical analysis
We conducted a pairwise effectiveness meta-analysis using a random-effects model and adjusted using restricted maximum likelihood. 23 We reported on dichotomous outcomes using risk ratio (RR) and 95% confidence interval (CI). We assessed the heterogeneity in included trials using the I 2 statistics. We conducted all statistical analyses using Stata, version 14 (StataCorp, College Station, TX, USA).

Characteristics of included studies
Our electronic search identified 318 potentially relevant citations; of these, we evaluated 10 articles in full against our inclusion criteria. We excluded three nonrandomized studies, one study that did not report on women undergoing assisted conception and another study reporting on the same cohort of women ( Figure 1). We included a total of five trials in our meta-analysis (n = 843 women). 20,21,[24][25][26] Two trials were from Egypt and one was from each of the United Kingdom, India, and Saudi Arabia. All trials used 20% IVI solution at the time of embryo transfer but there were variations in the volume given and the duration of the protocol. Two trials administered the first dose of intralipid prior to or at the time of oocyte retrieval. Four trials administered a second dose of intralipid following a positive pregnancy test and one continued intralipid administration fortnightly until the end of the first trimester. All trials used a normal saline infusion or no intervention (routine care) as a comparator ( Table 1). All trials included women with previous implantation failure although the inclusion threshold varied across included trials. Four trials excluded women with other known causes for RIF while one did not report clear exclusion criteria. 20

Risk of bias
There was an overall moderate risk of bias of the included studies with four studies (80%) showing low risk of bias for randomization and three (60%) for allocation concealment. Three (60%) studies had low risk for missing outcome data while two (40%) studies    ; thus, a meta-analysis was not possible for those two outcomes. One trial reported two cases of congenital middle ear anomalies in the intralipid group (n = 2/14, 14%), but no statistical testing was reported. 20

Summary of findings
Our meta-analysis suggests that compared to routine care, providing IVI at the time of embryo transfer could increase the rate of clinical pregnancy and live birth in women with the history of RIF. The effect estimates were consistent for both outcomes which could suggest a persistent benefit through both the implantation phase and later till the first trimester of the pregnancy. Such effect could be explained by an improvement in the endometrial receptivity driven by

FIGURE 2
Risk of bias in included randomized trials on the effectiveness of intralipid during assisted conception in women with history of recurrent implantation failure a favorable immunological response to the invading trophoblast during the implantation phase. None of the included studies reported serious adverse events in women receiving IVI. Similar trials in the literature reported minor side effects such as a flushing feeling, nausea, and headache. 27 Two cases of congenital abnormalities were reported in one trial of relatively small sample size. While it is difficult to ascertain causality due to statistical limitation, we highlight the importance of reporting fetal anomalies in any future studies planning to evaluate IVI use pre-pregnancy.

Strength and limitations
We used a prospectively registered protocol and a standardized methodology to conduct this review.
We assessed the risk of bias in all included studies and extracted data in duplicate. We contacted two groups of authors 20, 21 to ensure study eligibility for inclusion. Still, our findings are not without limitations. While all trials used 20% IVI solutions, there were variations in the used doses, frequency of IVI administration, and in the ART protocols which could increase the performance bias. For example, three trials used a diluted form of IVI in normal saline 20,24,25 while the other two used a concentrated form of IVI prior to its use. 23,27 Due to the small number of included studies, we were unable to explore this further using a meta-regression. The comparison varied among included trials (normal saline vs no intervention) with limited information on blinding efforts (as IVI is of white color). As the reported outcomes are objective (pregnancy and live births), we perceive little placebo effect on the quality of reporting, although performance bias is possible. Finally, we were unable to investigate the effect of potential effect modifiers such as the process of ovulation stimulation, the number and quality of transferred embryos and the use of additional add-on treatments. Such effect could be best evaluated with individual patient data meta-analysis; however, given the relatively small number of women included in this meta-analysis, such an approach is unlikely to yield significant findings pending the conduct of larger trials in the future. Our findings are, therefore, important to harmonize the standardize future trial conduct to yield more efficient evidence synthesis.

Implications for future practice
The value of several ART add-ons has been heavily debated with concerns regarding the ethicality of offering such additional treatments without clear evidence of benefit. 28 Our meta-analysis of five randomized trials offers a pragmatic overview on the potential role of IVI to help a specific group of women affected by RIF. However, given the limitations and the quality of included trials, we conclude that adopting the use of IVI in ART practice is at present immature and offering it a-la-carte to all couples undergoing ART should not be adopted until larger RCTs are published with a clear persistent beneficial effect. This imprecision in effect size is largely driven by the variation in treatment protocols and use of IVI across included studies. There is a need to focus future research work on evaluating the perceived immunological effect on IVI on the endometrium at the implantation phase within larger cohorts of women with history of recurrent pregnancy loss. More specifically, there is a need to delineate this perceived impact in both subgroups of RIF and recurrent pregnancy loss in the first trimester. As the mechanistic effect of IVI on the maternal innate immunological system remains unclear, there is also a need for more discovery research to aid the stratification and selection of women who might benefit the most of IVI. 29 Novel tests are now available to aid the identification of women with nonreceptive endometrium such as microarray technology which could offer more precise evaluation of the effectiveness of IVI in future studies, 30 but they remain of limited use in clinical practice. Several methodological improvements are required to enhance the quality of future studies on IVI such as standardizing the definition and management pathways for women with RIF 3 and following more transparent reporting using an established core outcome set for fertility treatments. 31 We were unable to evaluate any long-term protective effect of IVI later on in pregnancy. As it is argued to promote healthy implantation and placentation, IVI might reduce the risk of placenta mediated pregnancy complications (e.g. pre-eclampsia) which are more common in women undergoing ART. 32 Planned longterm follow-up of randomized cohort is essential to precisely evaluate such protective effects and aid more robust evidence synthesis on ART add-ons. Despite our meta-analysis results suggesting IVI intralipid increased both clinical pregnancy rate and live birth, the small number of studies makes it challenging to draw firm conclusions as to the benefit of intralipid.
In conclusion, evidence to support the use of IVI at the time of embryo transfer in women with a history of RIF is limited; more research is needed before adopting it in clinical practice.