Association of Intralipid Infusions With Birth Outcomes and First-Trimester Placental Biomarkers in Frozen Embryo Transfer Cycles: A Retrospective Cohort Study.

This study received prospective approval from the Monash IVF Institutional Review Board (IRB No. 2024_P07). It was conducted as a retrospective review of routinely collected clinical, laboratory and pregnancy outcome data from patients undergoing autologous FET cycles. Data were extracted and analysed confidentially, with patient identifiers removed before analysis and reporting. The study was conducted in accordance with applicable institutional governance requirements and ethical principles for human research, including the protection of participant privacy and confidentiality.

The authors have nothing to report.

A total of 729 autologous FET pregnancies resulting in delivery beyond 20 weeks' gestation and with available first‐trimester screening were included in the analysis. Of these, 136 (18.7%) pregnancies were exposed to Intralipid and 593 (81.3%) served as controls (Table  1 ). Baseline maternal and treatment characteristics according to Intralipid exposure ( n  = 729). Note: Continuous variables are presented as mean ± standard deviation and categorical variables as n (%). p values were calculated using Welch's t ‐test for continuous variables and uncorrected Pearson's χ 2 test for categorical variables. Total sample: n  = 729 after excluding records with missing or invalid birthweight. Abbreviations: BMI, body mass index; FSH, follicle‐stimulating hormone; HRT, hormone replacement therapy; ICSI, intracytoplasmic sperm injection; IVF, in vitro fertilisation; PGT‐A, preimplantation genetic testing for aneuploidy. p  < 0.05. Women receiving Intralipid were slightly older than controls (36.4 ± 4.6 vs. 35.0 ± 4.3 years) and had a higher mean body mass index (27.1 ± 6.3 vs. 26.2 ± 6.1 kg/m 2 ). They were also more likely to have no previous live birth and to have undergone multiple previous IVF treatment cycles. Treatment protocols differed between groups, with ultralong down‐regulation regimens more frequently used among women receiving Intralipid, consistent with the higher prevalence of adenomyosis within this cohort (Table  1 ). Birthweight, gestational age at delivery, gestation‐adjusted birthweight Z ‐scores and individual adverse pregnancy and birth outcomes were comparable between the Intralipid and control groups (Table  2 ). No statistically significant differences were observed for preterm birth, hypertensive disorders of pregnancy, gestational diabetes mellitus, foetal growth restriction, congenital anomalies, or stillbirth. Pregnancy outcomes and multivariable analysis according to Intralipid exposure. Note: Data are presented as mean ± standard deviation, median [interquartile range], n (%) or adjusted odds ratio (95% confidence interval), as appropriate. p values were calculated using Welch's t ‐test, Mann–Whitney U test, Pearson's χ 2 test or Fisher's exact test, as appropriate. p values were not calculated when event counts were fewer than 5. Birthweight and gestational age are presented descriptively and were not included in the composite pregnancy complication analysis. The stillbirth in the Intralipid group was associated with placental villitis and a three‐vessel cord with marginal insertion; the congenital anomaly was associated with a de novo 3.1‐Mb interstitial deletion involving chromosome 6p24.1–p23. Abbreviations: BMI, body mass index; CI, confidence interval; FGR, foetal growth restriction; GA, gestational age; GDM, gestational diabetes mellitus; HRT, hormone replacement therapy; ICSI, intracytoplasmic sperm injection; IQR, interquartile range; IUFD, intrauterine foetal death; OR, odds ratio; PGT‐A, preimplantation genetic testing for aneuploidy. Fisher's exact test used because expected cell counts were < 5. Mann–Whitney U test used for median comparisons. Panel B reports event‐level outcomes, whereas Panel C reports patient‐level complication burden. Therefore, total adverse outcome events may exceed the number of pregnancies with complications. Multivariable logistic regression adjusted for maternal age, body mass index, parity, plurality, cycle type, intracytoplasmic sperm injection use and preimplantation genetic testing for aneuploidy status. Multivariable linear regression for birthweight outcomes adjusted for maternal age, body mass index, parity, plurality, foetal sex, gestational age, cycle type, intracytoplasmic sperm injection use and preimplantation genetic testing for aneuploidy status, where applicable. In Panel E, Group 1 refers to the Intralipid group and Group 2 refers to the no Intralipid group. p  < 0.05. Patient‐level complication burden was assessed separately because individual pregnancies could contribute more than one adverse outcome event. Overall, 29/136 (21.3%) women in the Intralipid group and 92/593 (15.5%) controls had at least one pregnancy complication, while the composite adverse pregnancy outcome, defined as two or more complications, occurred in 7/136 (5.1%) and 8/593 (1.3%), respectively. Multivariable linear regression demonstrated no independent association between Intralipid exposure and birthweight or gestation‐adjusted birthweight Z ‐scores after adjustment for maternal age, body mass index, parity, plurality, foetal sex, cycle type, ICSI use and PGT‐A status (Table  2 ). Similarly, in the fully adjusted multivariable logistic regression model, Intralipid exposure was not independently associated with the composite adverse pregnancy outcome (adjusted odds ratio [aOR] 2.08, 95% confidence interval [CI] 0.73–5.96; p  = 0.171). Independent predictors of the composite adverse pregnancy outcome were multiple pregnancy (aOR 7.17, 95% CI 1.28–40.17; p  = 0.025), programmed artificial cycles (aOR 6.10, 95% CI 1.40–26.60; p  = 0.016) and ultralong down‐regulated HRT cycles (aOR 9.15, 95% CI 2.05–40.84; p  = 0.004). Maternal age, body mass index, parity, ICSI and PGT‐A were not independently associated with the composite adverse pregnancy outcome. The number of Intralipid infusions administered varied according to individual clinical circumstances, with the distribution of Intralipid treatment exposure summarised in Table  3 . Distribution of Intralipid infusions among women receiving treatment ( n  = 136). Note: Data are presented as n (%). All 136 women who received Intralipid therapy had complete infusion data available. Median absolute PAPP‐A concentrations were significantly lower among women receiving Intralipid than controls (1021.0 [IQR 614.6–1690.8] vs. 1213 [IQR 730–2047] IU/L; p  = 0.009). Median PAPP‐A MoM values were likewise lower in the Intralipid group (0.82 [IQR 0.53–1.31] vs. 0.96 [IQR 0.61–1.48]; p  = 0.018). No significant differences were observed in absolute or corrected concentrations of free β‐human chorionic gonadotrophin (β‐hCG) or PlGF (Table  4 ). First‐trimester placental biomarkers according to Intralipid exposure. Note: Data are presented as median (interquartile range). p values were calculated using the Mann–Whitney U test. Abbreviations: β‐hCG, beta‐human chorionic gonadotrophin; IQR, interquartile range; MoM, multiples of the median; PAPP‐A, pregnancy‐associated plasma protein‐A; PlGF, placental growth factor. p  < 0.05. Following multivariable adjustment, Intralipid exposure remained independently associated with lower absolute PAPP‐A concentrations (adjusted geometric mean ratio 0.82, 95% CI 0.71–0.95; p  = 0.0069) and PAPP‐A MoM values (adjusted geometric mean ratio 0.86, 95% CI 0.76–0.98; p  = 0.0211). In contrast, no independent associations were identified between Intralipid exposure and free β‐hCG or PlGF concentrations. Additional multivariable analyses were performed to examine whether first‐trimester placental biomarkers were associated with foetal growth and placental pregnancy complications. Log‐transformed PAPP‐A concentrations were not independently associated with birthweight Z ‐scores (adjusted effect estimate 0.011, 95% CI −0.098 to 0.121; p  = 0.836), and first‐trimester biomarker concentrations were not independent predictors of placental or hypertensive pregnancy complications (all p  > 0.10).

Although Intralipid is increasingly used as an adjunctive therapy for women with RIF and RPL, published studies have focused predominantly on reproductive efficacy, with limited evaluation of maternal and neonatal safety after an ongoing pregnancy has been achieved. Systematic reviews have highlighted the limited quality and scope of the available evidence and have called for larger, well‐designed studies evaluating both efficacy and safety outcomes following Intralipid exposure [ 20 , 21 , 22 ]. In this retrospective cohort of autologous FET pregnancies, Intralipid administration was not independently associated with adverse pregnancy outcomes after adjustment for baseline clinical and treatment characteristics. Although women receiving Intralipid had a greater unadjusted burden of pregnancy complications, this appeared to reflect the higher‐risk clinical profile of women selected for treatment rather than an independent association with Intralipid exposure. In contrast, Intralipid administration was independently associated with lower first‐trimester PAPP‐A concentrations, without corresponding differences in foetal growth, PlGF, free β‐hCG, or clinically apparent placental complications. The higher burden of pregnancy complications observed on unadjusted analyses may reflect confounding by indication. Women receiving Intralipid represented a more complex infertility population, being older, more likely to have undergone multiple previous IVF cycles and more frequently treated using ultralong down‐regulation protocols because of adenomyosis. These factors may reflect a higher‐risk reproductive and treatment context, in which underlying infertility diagnoses, adenomyosis and ART‐related treatment factors have been associated with adverse obstetric, neonatal and offspring outcomes [ 23 , 24 , 25 , 26 ]. This highlights the importance of evaluating adjunctive therapies within the broader context of patient selection and treatment characteristics. A notable biochemical finding was the significantly lower first‐trimester PAPP‐A concentrations among women receiving Intralipid. PAPP‐A contributes to first‐trimester screening algorithms for foetal aneuploidy and pre‐eclampsia; therefore, lower maternal serum concentrations may increase calculated screening risk and lead to additional investigations or counselling. Low first‐trimester PAPP‐A has traditionally been associated with impaired placentation and increased risk of foetal growth restriction, pre‐eclampsia and other adverse obstetric outcomes [ 16 , 17 , 27 ]. However, in the present cohort, lower PAPP‐A concentrations were not accompanied by lower birthweight Z ‐scores, increased foetal growth restriction, hypertensive disorders of pregnancy or other clinically apparent placental complications. These findings suggest that the observed PAPP‐A difference may represent a biochemical association rather than clinically evident placental dysfunction, although replication in larger prospective cohorts is warranted. The biological mechanism underlying the association between Intralipid exposure and lower PAPP‐A concentrations remains uncertain. The association with PAPP‐A MoM persisted after log‐transformed multivariable regression adjustment, whereas the fully maternal‐characteristic corrected MoM comparison was attenuated in univariable analysis. Residual confounding by indication therefore cannot be excluded given the retrospective observational design. PAPP‐A is expressed in both placental and non‐placental tissues, including vascular tissues, which may explain why changes in circulating PAPP‐A do not necessarily correspond to clinically significant placental dysfunction [ 28 ]. Whether the observed difference reflects a transient biochemical effect, altered maternal production or altered placental physiology requires further investigation. This study has several limitations. Its retrospective design introduces the potential for selection bias and residual confounding. Although multivariable adjustment accounted for important clinical and treatment‐related factors, residual confounding by indication remains a limitation. Women receiving Intralipid represented a higher‐risk reproductive population; however, reliable retrospective classification according to RIF, RPL and other treatment indications was not consistently possible because these variables were not systematically recorded in a standardised manner across the study period. The relatively small number of adverse pregnancy outcomes limited statistical power to detect modest differences in uncommon obstetric complications. In addition, this was a single‐centre study, and variation in Intralipid protocols between fertility centres may limit generalisability. Routine clinic follow‐up, which relies in part on patient‐reported information, may also have under‐ascertained some obstetric outcomes, as foetal growth restriction was identified more frequently using gestation‐ and sex‐adjusted birthweight Z ‐scores than through routine outcome data. Linkage with population‐based perinatal registries would provide more complete and objective ascertainment of maternal and neonatal outcomes. Despite these limitations, the present study has several important strengths. To our knowledge, no previous study has comprehensively evaluated both detailed obstetric outcomes and first‐trimester placental biomarkers following Intralipid exposure during FET cycles. Previous studies have largely focused on reproductive efficacy, with limited and inconsistent reporting of maternal, foetal and neonatal safety outcomes, highlighting an important evidence gap that the present study begins to address [ 20 , 21 , 22 ]. Additional strengths include obstetric follow‐up, adjustment for clinically relevant confounders, and integration of first‐trimester placental biomarkers with objective measures of foetal growth. In conclusion, Intralipid administration during FET cycles was not independently associated with adverse pregnancy outcomes after adjustment for measured confounders. Although Intralipid exposure was associated with significantly lower first‐trimester PAPP‐A concentrations, this biochemical finding was not accompanied by evidence of impaired foetal growth or placental complications. As adjunctive therapies in assisted reproduction continue to be used despite limited long‐term safety data, these findings provide reassuring observational data regarding the obstetric safety of Intralipid, while highlighting the need for larger registry‐based studies to confirm these observations [ 1 , 2 , 4 , 21 , 22 ]. Expansion of ANZARD to routinely capture adjunctive ART therapies, together with linkage to Australian and New Zealand perinatal outcome registries, should be considered to improve outcome ascertainment and long‐term safety evaluation.

As the analysis used existing clinical records and de‐identified data, individual patient consent was not required in accordance with the conditions of institutional approval.

Recurrent implantation failure (RIF) and recurrent pregnancy loss (RPL) remain significant challenges in assisted reproductive technology (ART). Immunomodulatory therapies, including intravenous Intralipid, are increasingly used as adjunctive therapy for selected women with RIF, RPL and presumed immune‐mediated reproductive failure, despite limited evidence for their efficacy and safety [ 1 , 2 , 3 ]. Professional organisations, including the European Society of Human Reproduction and Embryology (ESHRE) and the British Fertility Society, therefore advise against the routine use of Intralipid because current evidence is insufficient to support its efficacy and safety [ 2 , 4 ]. Intralipid is a soybean oil‐based lipid emulsion originally developed for parenteral nutrition. Its use in reproductive medicine is based on putative immunomodulatory effects, including modulation of natural killer cell activity and immune responses at the maternal–foetal interface, with potential effects on trophoblast invasion and implantation [ 5 , 6 , 7 ]. Limited experimental evidence, including animal data, suggests that Intralipid or lipid‐related metabolic perturbation may influence placental inflammatory and oxidative stress pathways, with potential downstream effects on trophoblast function and foetal growth, although direct pregnancy safety data remain sparse [ 8 , 9 , 10 , 11 , 12 ]. While previous studies have primarily examined implantation and live birth rates, little is known about the association between Intralipid exposure and placental development, first‐trimester placental biomarkers or subsequent pregnancy outcomes. First‐trimester placental biomarkers, including pregnancy‐associated plasma protein‐A (PAPP‐A), placental growth factor (PlGF) and free β‐human chorionic gonadotropin (β‐hCG), are routinely measured during first‐trimester aneuploidy screening and are increasingly incorporated into pre‐eclampsia screening algorithms. These biomarkers provide indirect measures of early placental function, and altered concentrations have been associated with impaired placentation and adverse pregnancy outcomes, including foetal growth restriction, hypertensive disorders of pregnancy and stillbirth [ 13 , 14 , 15 , 16 , 17 ]. Despite the increasing clinical use of Intralipid in ART, few studies have examined its association with first‐trimester placental biomarkers or subsequent pregnancy outcomes. We hypothesised that Intralipid exposure during the peri‐implantation period and early pregnancy would be associated with alterations in first‐trimester placental biomarkers, as indirect measures of early placental function, and with differences in foetal growth and subsequent pregnancy outcomes. The primary aim of this study was to evaluate the association between Intralipid exposure during frozen embryo transfer (FET) cycles and pregnancy outcomes, including birthweight and gestational age at delivery. Secondary aims were to examine the association between Intralipid exposure and first‐trimester placental biomarkers, including PAPP‐A, PlGF and free β‐hCG.

The authors declare no conflicts of interest.

This retrospective cohort study evaluated the association between Intralipid exposure and pregnancy outcomes and first‐trimester placental biomarkers among women undergoing autologous FET cycles at a private assisted reproduction unit between 1 January 2017 and 31 December 2023. Participant selection, exclusions, pregnancy losses before 20 weeks' gestation, missing birthweight data and derivation of the final analytic cohort are summarised in the STROBE‐style flow diagram (Figure  1 ). Study cohort flow diagram showing identification of eligible autologous frozen embryo transfer pregnancies, application of exclusion criteria and allocation according to Intralipid exposure status. Eligible participants were women who achieved pregnancies continuing beyond 20 weeks' gestation and underwent routine first‐trimester screening. Pregnancies ending before 20 weeks' gestation and those with missing birthweight data were excluded. Birthweight was required to calculate gestation‐adjusted birthweight Z ‐scores and classify foetal growth restriction. Where women had more than one eligible pregnancy during the study period, only the first qualifying pregnancy was included to ensure independence of observations. Maternal age was recorded at the time of embryo transfer, and body mass index (BMI) was obtained at the commencement of the treatment cycle. FET protocols included natural cycles, ovulation induction using letrozole or follicle‐stimulating hormone (FSH), programmed hormone replacement therapy (HRT) cycles and ultralong down‐regulation regimens combined with either HRT or ovulation induction. Intralipid was administered to women with RIF, recurrent miscarriage or diffuse adenomyosis where the treating clinician considered a presumed immune‐mediated contribution to reproductive failure. As this was a retrospective study, the specific indication for treatment could not always be reliably determined from the medical records. Patients received 100 mL of 20% Intralipid, diluted in 900 mL of normal saline, administered intravenously over approximately 2 h, with routine clinical monitoring. The standard treatment protocol comprised an initial infusion within 24 h before embryo transfer, followed by a second infusion 1 week later. A third infusion, when administered, was routinely given at approximately 4–5 weeks' gestation following confirmation of a positive pregnancy test. Additional infusions were administered weekly to a small number of patients, according to individual clinical circumstances, at the discretion of the treating clinician. This treatment regimen was based on previous evidence demonstrating immunomodulatory effects lasting at least 7 days [ 5 , 18 ]. Maternal serum PAPP‐A, PlGF and free β‐hCG were measured between 10 and 13 + 6 weeks' gestation using the DELFIA Xpress analyser (PerkinElmer/Revvity). Multiples of the median (MoM) were calculated using proprietary first‐trimester screening software, which automatically adjusted for established maternal and pregnancy characteristics, including maternal age, ethnicity, maternal weight, gestational age, estimated date of delivery, crown‐rump length, plurality, assisted reproduction, previous obstetric history and other standard screening variables. Pregnancy and neonatal outcomes included gestational age at delivery, birthweight, hypertensive disorders, gestational diabetes, foetal growth restriction, stillbirth and congenital anomalies. Outcome data were obtained primarily through patient follow‐up and supplemented by hospital records where clarification was required. Birthweight Z ‐scores were calculated using the INTERGROWTH‐21st International Newborn Size Standards according to gestational age and infant sex [ 19 ]. Individual adverse pregnancy outcomes were tabulated as event‐level outcomes, while patient‐level complication burden was assessed according to the number of adverse pregnancy complications within each pregnancy. Statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). Continuous variables were assessed for normality. Normally distributed variables are presented as mean ± standard deviation (SD) and were compared using Student's t ‐test, whereas non‐normally distributed variables are presented as median (interquartile range [IQR]) and were compared using the Mann–Whitney U test. Categorical variables were compared using the χ 2 test or Fisher's exact test, as appropriate. Biomarker variables demonstrating skewed distributions were log‐transformed before regression analyses. For the multivariable logistic regression analysis, a composite adverse pregnancy outcome was defined as the occurrence of two or more adverse pregnancy complications within the same pregnancy. This outcome was chosen to reflect the cumulative burden of obstetric morbidity rather than isolated pregnancy complications. Covariates were prespecified a priori based on clinical relevance and included maternal age, BMI, parity, plurality (singleton vs. multiple pregnancy), cycle type, intracytoplasmic sperm injection (ICSI) use and preimplantation genetic testing for aneuploidy (PGT‐A). Maternal age and BMI were entered as continuous variables, and the assumption of linearity with the log odds of the outcome was assessed using restricted cubic splines. As no evidence of non‐linearity was identified, both variables were retained as linear terms in the final model. Multivariable linear regression was used to examine the association between Intralipid exposure and first‐trimester placental biomarkers after adjustment for maternal age, BMI, parity, plurality, foetal sex, cycle type, ICSI use and PGT‐A status. Results are presented as adjusted geometric mean ratios with 95% confidence intervals. Cycle regimen type was included as a prespecified covariate in multivariable analyses to account for differences between standard FET protocols and ultralong down‐regulation protocols used in higher‐risk clinical contexts. Following application of the study inclusion and exclusion criteria, missing data among variables included in the multivariable regression models were minimal (< 1%). Regression analyses were therefore performed using complete‐case analysis. Statistical significance was defined as a two‐sided p value < 0.05.