Preconception Lipids Linked to Pregnancy Risks After IVF

In women who had singleton live births through IVF or intracytoplasmic sperm injection with fresh embryo transfer, blood lipid levels measured before conception were associated with three major pregnancy complications: gestational diabetes, preterm delivery and infants who were small for gestational age. A retrospective study of 1,036 women found that higher triglycerides were associated with preterm delivery after accounting for age, body mass index and glucose, while higher total cholesterol and triglycerides were linked to gestational diabetes. Total cholesterol, low-density lipoprotein cholesterol and high-density lipoprotein cholesterol were each associated with increased odds of having a small-for-gestational-age infant. These findings describe statistical associations in a selected group of successful IVF or ICSI pregnancies, not proof that changing lipid levels would prevent complications, and the researchers did not test whether treatment before conception improves outcomes.

The study was conducted at a university-based hospital in China, where 1,530 live births were recorded between January 2019 and January 2021. The investigators included women aged 20 to 45 years who had delivered a singleton live infant after a fresh IVF or ICSI embryo transfer. A singleton gestational sac had to be confirmed by ultrasound at five to seven weeks after transfer, and cases involving vanishing twin syndrome were excluded. Pregnancies ending before 28 weeks were not considered live births. Multiple pregnancies, missing lipid results, unavailable complication data and other missing information led to the exclusion of 494 women. The final cohort therefore describes only pregnancies that proceeded to a singleton live birth, not all women who began assisted reproduction, an important selection constraint when interpreting the results.

Blood was collected while participants were fasting on days two to four of the menstrual cycle, before assisted reproductive treatment. The researchers measured total cholesterol, triglycerides, low-density lipoprotein cholesterol and high-density lipoprotein cholesterol, with glucose and reproductive hormones also recorded. A woman was classified as having dyslipidemia if any one of four thresholds was met: high-density lipoprotein cholesterol below 1.0 mmol/L, low-density lipoprotein cholesterol at or above 4.1 mmol/L, total cholesterol at or above 6.2 mmol/L, or triglycerides at or above 2.3 mmol/L. These criteria were validated for a Chinese population. Laboratory assays showed intra-assay coefficients of variation below 1.53% and inter-assay coefficients below 3.55%, providing a measure of the precision of the hormone and metabolic assays.

Among the 1,036 participants, the mean age was 30.8 years and the mean body mass index was 21.7 kg/m². Overall, 180 women met the study definition of dyslipidemia and 856 did not. Women with dyslipidemia had a higher mean body mass index, 23.2 compared with 21.4 kg/m², and slightly higher mean glucose, 5.2 compared with 5.1 mmol/L. They also had higher total cholesterol, low-density lipoprotein cholesterol and triglycerides, together with lower high-density lipoprotein cholesterol. Their infertility diagnoses also differed, and basal follicle-stimulating hormone levels were lower, 5.8 compared with 6.8 IU/L. Because several baseline characteristics were unevenly distributed, the investigators treated the lipid-outcome relationships as exploratory and used multivariable regression to adjust for age, body mass index and glucose.

Across the full cohort, 535 births, or 51.6%, were by cesarean section. The study recorded gestational diabetes in 3.8% of participants, preeclampsia in 0.7%, preterm delivery in 7.6%, small-for-gestational-age infants in 6.8%, large-for-gestational-age infants in 15.6%, low birthweight in 6.4%, macrosomia in 5.7% and neonatal intensive care unit admission in 1.8%. Before statistical adjustment, women with dyslipidemia had higher rates of gestational diabetes, 11.7% versus 2.1%; preterm delivery, 13.9% versus 6.3%; and low birthweight, 10.6% versus 5.5%. The corresponding p-values were below 0.001 for gestational diabetes and preterm delivery and 0.012 for low birthweight. By contrast, the dyslipidemia and comparison groups did not differ significantly in the observed rates of cesarean delivery, preeclampsia, small- or large-for-gestational-age infants, macrosomia or neonatal intensive care admission.

The multivariable analysis included 952 women with complete data and produced the study’s most prominent adjusted associations. For each one-unit increase in triglycerides, the adjusted odds ratio was 1.49, with a 95% confidence interval of 1.11 to 1.99, for preterm delivery. Total cholesterol was associated with gestational diabetes at an adjusted odds ratio of 2.36 per unit, with a 95% confidence interval of 1.74 to 3.21, while triglycerides had an adjusted odds ratio of 1.21 per unit, with a confidence interval of 1.00 to 1.46. For small-for-gestational-age infants, total cholesterol had an adjusted odds ratio of 1.48 per unit, low-density lipoprotein cholesterol 1.48 and high-density lipoprotein cholesterol 3.03; their respective 95% confidence intervals were 1.08 to 2.02, 1.02 to 2.14 and 1.19 to 7.73. These estimates are wide for some outcomes, especially the high-density lipoprotein association, underscoring the uncertainty surrounding their precision.

Not every lipid was related to every outcome, and the unadjusted pattern did not survive adjustment in the same form. Lower high-density lipoprotein cholesterol was initially associated with large-for-gestational-age infants and cesarean delivery, but neither relationship remained clearly significant in the multivariable model. Low birthweight showed borderline adjusted associations with triglycerides, with an odds ratio of 1.22 and a 95% confidence interval of 1.00 to 1.50. No significant adjusted associations were found for cesarean delivery, preeclampsia, large-for-gestational-age infants, macrosomia or neonatal intensive care admission. Preeclampsia was particularly uncommon, occurring in seven women, and the study lacked the statistical power to establish robust relationships for that outcome or for the 19 neonatal intensive care unit admissions. Spearman correlation analysis also linked triglycerides with preterm delivery, low birthweight, gestational diabetes, cesarean delivery and neonatal intensive care admission, but not with small- or large-for-gestational-age infants, macrosomia or preeclampsia.

The investigators proposed several mechanisms that could connect preconception lipid patterns with outcomes, but none was directly tested. Lipids are precursors for steroid hormones such as progesterone, estrogen and androgen, so abnormal circulating lipid levels might influence ovarian steroidogenesis, follicular development and oocyte quality before embryo transfer. A lipid profile may also reflect insulin resistance, altered glucose regulation or impaired endothelial function, all of which could affect placental blood flow and fetal growth. The authors noted that placental dysfunction is a possible contributor to adverse outcomes after assisted reproduction, while reduced placental oxygen delivery is associated with fetal growth restriction and preterm birth. The surprising positive association between high-density lipoprotein cholesterol and small-for-gestational-age birth remains unexplained; the authors suggested that higher high-density lipoprotein might reduce placental cholesterol transfer through low-density lipoprotein uptake, but this was a hypothesis rather than a measured mechanism.

The results are best viewed as hypothesis-generating because the study was retrospective, single-center and restricted to one relatively homogeneous population. The design did not include a prospective sample-size calculation, and multiple statistical comparisons increased the possibility of false-positive findings. Residual confounding cannot be excluded, particularly because the analysis adjusted for only three variables—age, body mass index and glucose—even though the groups differed in other characteristics, including infertility diagnoses and hormone levels. Selection bias is also substantial: the analysis excluded pregnancies that did not result in a live birth and did not represent women who became pregnant naturally, received frozen embryo transfers or delivered multiples. The low frequencies of gestational diabetes and preeclampsia further limited inference. Larger prospective cohorts, with attention to specific infertility diagnoses and interventions that modify lipids, would be needed to determine whether preconception lipid assessment can identify risk reliably or whether improving lipid levels changes pregnancy outcomes. The current data support an association worthy of further testing, not a causal or clinical claim that lipid screening or treatment prevents complications in IVF or ICSI pregnancies.

Subject of Research: Associations between preconception blood lipid levels and pregnancy complications following IVF or ICSI with fresh embryo transfer

Article Title: Association between preconception serum lipid levels and pregnancy complications in singleton live births from IVF/ICSI with fresh embryo transfers

Article References: Cai, W.-Y., Luo, X., Zhang, Q., Li, H., Yang, H., Wang, X., & Xu, J. (2026). Association between preconception serum lipid levels and pregnancy complications in singleton live births from IVF/ICSI with fresh embryo transfers. Frontiers in Nutrition, 13, Article 1915181. https://doi.org/10.3389/fnut.2026.1915181

Image Credits: AI Generated

DOI: 10.3389/fnut.2026.1915181

Keywords: assisted reproduction, preconception health, blood lipids, IVF, ICSI, gestational diabetes, preterm delivery, fetal growth, retrospective study, Association, between, preconception