Impact of obesity on follicular fluid lipid composition and IVF/ICSI outcomes in Korean women: A lipidomic study.

According to the Health Insurance Review and Assessment Service-National Health Insurance Service of South Korea, the prevalence of obesity among women aged >20 years will increase by 27.8% by 2021, continuing the decade-long rise. The number of women treated for infertility in Korea is expected to increase from 121,038 in 2018–140,458 in 2022. A recent study revealed that female infertility in Korea was associated with obesity, with an odds ratio of 2.06 (95% confidence interval, 1.61–2.64) compared to women with a normal body mass index (BMI) (18.5 kg/m 2  ≤ BMI < 25 kg/m 2 ) [ 1 ]. Since South Korea has the lowest total fertility rate (0.78) among the Organization for Economic Co-operation and Development countries, the rising prevalence of obesity among reproductive-aged women, along with increasing infertility rates, poses a significant social challenge. In 2021, the American Society for Reproductive Medicine outlined the adverse effects of obesity on reproduction, including impaired ovulatory function, decreased responsiveness to ovarian stimulation, diminished oocyte quality, lower fertilization rates, increased miscarriage rates, and increased risks of maternal and fetal complications during pregnancy [ 2 ]. These findings align with those of other studies [ 1 , 3 – 6 ]. However, obesity has not been reported to significantly affect the clinical or embryological outcomes of in vitro fertilization (IVF) [ 7 – 8 ]. Despite the numerous studies on infertility and obesity, the underlying mechanisms remain unclear. Yong et al. reported that obesity affects the hypothalamic-pituitary-ovarian axis, oocyte maturation, embryonic development, and all stages of fetal development, leading to decreased female reproduction [ 9 ]. Lipotoxicity is the primary mechanism underlying this impairment, which causes mitochondrial damage in oocytes, an increase in reactive oxygen species, and endoplasmic reticulum stress. These conditions ultimately disrupt hormone production, the formation of cumulus-oocyte complexes (COC), fertilization, and embryonic development [ 10 – 15 ]. Furthermore, obesity causes chronic inflammation by increasing proinflammatory cytokine levels and gene activity [ 16 – 18 ]. As the follicular fluid (FF) represents the immediate microenvironment of the COC, its analysis provides insights into the effects of obesity on IVF outcomes. Alterations in the FF components can reflect in the metabolic levels and influence oocyte development, maturation, and embryo quality [ 19 – 21 ]. Lipid metabolism is essential for oocyte meiosis and folliculogenesis. It provides energy through β-oxidation in mitochondria and functions both as a mediator for cell signaling and as a structural component of cellular membranes and steroid synthesis [ 15 , 22 – 24 ]. Previous studies have revealed that alterations in FF lipid composition are associated with follicular growth and oocyte development, thereby affecting IVF outcomes [ 25 – 30 ]. Some studies have also sought to identify lipid metabolites in the FF as predictive markers of ovarian stimulation and follicular development [ 31 – 32 ]. However, few studies have investigated the composition of BMI-related lipid metabolites in the FF [ 12 , 22 , 26 , 33 – 35 ]. Triacylglycerols (TAG) and free fatty acids (FA) have been the focus of most studies on lipid metabolites [ 12 , 22 , 25 , 26 , 34 , 35 ]. While one study revealed no significant change in FF metabolic composition according to BMI, many others have demonstrated a positive correlation between maternal BMI and TAG/FA accumulation, which impairs oocyte development [ 12 , 22 , 25 , 31 ]. Additionally, animal in vitro maturation studies using lipid supplementation indicate that lipid metabolism with TAG and FA deposition influences oocyte maturation and development [ 36 – 38 ]. However, data on its effects on human oocyte maturation are limited. Therefore, a comprehensive lipidomics profiling of the FF was conducted to identify changes in metabolites and their effects on oocyte development and embryological outcomes in Korean women with infertility and obesity, defined as a BMI ≥ 25 kg/m 2 by the Korean Society for the Study of Obesity (KSSO) [ 39 ].

The clinical characteristics of the 68 patients included in this study are presented in Table 3 . Group A (obese group, BMI ≥ 25 kg/m 2 ) comprised 41.2% (n = 28) of the participants, whereas Group B (non-obese group, BMI < 25 kg/m 2 ) comprised 58.8% (n = 40). The mean BMIs for the two groups were significantly different (Group A: 27.78 ± 2.12 kg/m 2 ; Group B:22.00 ± 2.37 kg/m 2 ; p  < 0.001). The mean total serum cholesterol levels were significantly different between groups (Group A: 213.78 ± 36.64 mg/dL; Group B: 195.24 ± 0.78 mg/dL; p  = 0.031). Additionally, the total cholesterol levels were positively correlated with the BMI ( p  = 0.002); a unit-increase in BMI corresponded to an estimated increase of 3.601 mg/dL in the serum cholesterol levels. Other variables, including age, cause of infertility, antral follicle count, basal follicle-stimulating hormone levels, AMH levels, infertility duration, and number of previous IVF attempts, were not significantly different between the groups. Other cycle variables, including stimulation protocol, total gonadotropin dose, duration of stimulation, estradiol level on the triggering day, type of trigger, and fertilization type, did not show significant differences between the groups ( Table 4 ). BMI, body mass index; PCOS, polycystic ovary syndrome; AFC, antral follicle count; AMH, anti- Müllerian hormone; FSH, follicle-stimulating hormone; IVF, in vitro fertilization; n, number of cycles; NS, not significant ( p  > 0.05). Values are presented as mean ± standard deviation n, number of cycles; BMI, body mass index; GnRH, gonadotropin-releasing hormone; Ovi, oviderel; Deca, decapeptyl; IVF, in vitro fertilization; ICSI, intracytoplasmic sperm injection. Values are presented as mean ± standard deviation or as number (%). A total of 159 of the 230 lipids ( Table 2 and Additional File 1, S1– S5 Tables ) in the FF samples were quantified. The assay was validated by confirming that lipid species with coefficient of variation <30% in the QC sample exceeded 70% (83.0%) of the total lipid species [ 41 ]. Significant differences were observed between the two groups in all neutral lipid compositions ( Figs 2 and 3 ). Group A exhibited markedly higher concentrations of TAG than Group B (Group A: 132,443.3 ± 71,011.6 pmol/mL; Group B: 92,298.1 ± 41,084.8 pmol/mL; p  = 0.0102; Fig 2 ). Similarly, DAG levels were also elevated in Group A relative to Group B (Group A: 1,788.4 ± 903.0 pmol/mL; Group B: 1,198.0 ± 636.9 pmol/mL; p  = 0.0014; Fig 2 ). However, MAG levels (16:0) demonstrated a negative correlation with the BMI (Group A: 115,531.2 ± 16,058.7 pmol/mL; Group B = 126,500.0 ± 17,281.2 pmol/mL; p =  0.0149; Fig 3A ). The bar graphs show the differences in lipid concentrations between the BMI groups. The lipid concentration (pmol/mL) was normalized to the volume. Different letters indicate significant differences compared with the control group [(*) p  < 0.05, (**) p  < 0.01], as determined using the t-test. Data are shown as mean ± standard deviation (n = 28 and n = 40). BMI, body mass index; LPC, Lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; PC, phosphatidylcholine; PE, phosphatidylethanolamine; MAG, monoacylglycerol; DAG, diacylglycerol; TAG, triacylglycerol; SM, sphingomyelin; CER, ceramide; CE, cholesteryl ester; AC, acylcarnitine; Group A, obese group; Group B, non-obese group. The bar graphs show the lipid subgroup concentration differences between the BMI groups and (A) MAG, (B) DAG, and (C) TAG. The lipid concentration (pmol/mL) was normalized to the volume. Different letters indicate significant differences compared with the control group [(*) p  < 0.05, (**) p  < 0.01], as determined by the t-test. The data shown as mean ± standard deviation (n = 28 and n = 40). BMI, body mass index; MAG, monoacylglycerol; DAG, diacylglycerol; TAG, triacylglycerol. Additionally, a notable difference was observed in the composition of neutral lipids between groups. In Group A, the compositions of TAG and DAG were higher at 53.02% and 0.72%, respectively, compared to Group B, which had compositions of 41.19% and 0.54%, respectively. Conversely, Group A exhibited lower MAG content (46.26%) than Group B (57.27%). Further analysis was conducted on the DAG lipid chains, covering five DAGs (32:2, 32:1, 34:0, 36:5, and 36:1), all showing statistical significance, with mean concentrations positively correlated with the groups ( Fig 3B ). Additionally, the analysis of 37 TAGs revealed significant differences in TAGs 48:0, 50:4, 50:3, 50:2, 50:1, 50:0, 52:5, 54:4, 52:3, 52:2, 52:1, 54:6, 54:5, 54:4, 54:3, 54:2, and 56:7 ( Fig 3C ). Additionally, total AC levels were higher in group A than in group B (Group A: 4878.1 ± 1179.3 pmol/mL; Group B: 4302.1 ± 1216.0; p  = 0.0222; Fig 2 ). Chain analysis revealed significant differences in C2:0, C3:0, C16:0, C18:0, and C18:1 levels, all of which were higher in Group A than in Group B ( Fig 4 ). These bar graphs show the concentration differences between the BMI groups and the AC subgroup. The lipid concentration (pmol/mL) was normalized to the volume. Different letters indicate significant differences compared with the control group [(*) p  < 0.05, (**) p  < 0.01], as determined by the t-test. The data shown as mean ± standard deviation (n = 28 and n = 40). BMI, body mass index; AC, acylcarnitine. Of the phospholipid family, only plasmenyl PC demonstrated a statistically significant inverse relationship with the BMI (Group A: 1,6093.3 ± 5323.7 pmol/mL; Group B: 1,8121.9 ± 4797.2 pmol/mL; p  = 0.0423; Fig 2 ). Further analysis of plasmenyl PC chains revealed that plasmenyl 34:0, 36:2, 40:7, and 40:4 showed significantly higher levels in Group B. No significant differences were observed between the groups in other lipids, such as sphingolipids (sphingomyelin and ceramide) or sterol lipids (cholesterol and cholesteryl ester) ( Fig 2 ). Adjusted p -value analysis revealed that, among the total lipids, only DAG showed a significant difference between groups. Among the total lipid chains, only AC 16:0 differed significantly between groups (adjusted p < 0.05). The lipid and lipid chain concentration data were not clearly separated between the obese and non-obese groups in the PCA ( Figs 5A and 5B ). When analyzing only the lipids and lipid chains that showed a significant difference in abundance between the groups ( p  < 0.05), some differences were observed. The variance explained by the first principal component increased significantly from 36.7% to 38.9% and from 26.8% to 41.0% for lipids and lipid chains, respectively, indicating that these discriminatory lipids or lipid chains could help distinguish between the groups ( Figs 5C and 5D ). The PCA plots show that the distributions of (A) all lipids, (B) all lipid chains, (C) lipids, and (D) lipid chains were significantly different between the two groups. The light green and red areas represent the 95% confidence intervals for each group. PCA, principal component analysis. Correlation analysis revealed a significant positive correlation between BMI and TAG, DAG, and AC levels (adjusted p  < 0.05; Fig 6 ). Further analysis of AC chains showed that AC 3:0, 16:0, 16:1, 18:0, and 18:1 levels were significantly positively associated with the BMI (adjusted p  = 0.014, 0.006, 0.019, 0.029, and 0.014, respectively; Fig 7 ). These scatter plots show a positive correlation between the BMI and (A) TAG, (B) DAG, and (C) AC levels, respectively. The red line represents the linear regression fit of each data point and the gray area indicates the 95% confidence interval of the regression fit. The correlation coefficient (r) was calculated using Spearman’s correlation method. BMI, body mass index; TAG, triacylglycerol; DAG, diacylglycerol; AC, acylcarnitine. Scatter plots show a positive correlation between the BMI and (A) AC 3:0, (B) AC 16:0, (C) AC 16:1, (D) AC 18:0, and (E) AC 18:1. The red line represents the linear regression fit of each data point and the gray area indicates the 95% confidence interval of the regression fit. The correlation coefficient (r) was calculated using Spearman’s correlation method. BMI, body mass index; AC, acylcarnitine. After adjusting for maternal age and AMH, which had the most significant impact on embryological outcomes, notable differences were observed in the numbers of fertilized oocytes, CL-stage embryos, and MO-stage embryos between Groups A and B. The total number of oocytes retrieved during the COH cycle was comparable between groups. However, in comparison to Group B, Group A had significantly fewer fertilized oocytes (Group A: 5.07 ± 4.16; Group B: 6.65 ± 4.61; p  = 0.043), as well as fewer CL-stage (Group A: 4.86 ± 4.26; Group B: 6.63 ± 4.61; p  = 0.016) and MO-stage (Group A: 4.00 ± 4.51; Group B: 6.05 ± 5.14; p  = 0.024) embryos. Nevertheless, Group A exhibited lower values across various embryological outcomes, including the number of mature oocytes retrieved, oocyte maturation rate, fertilization rate, CL-embryo formation rate, MO-embryo formation rate, number of total BL-stage embryos, BL formation rate, number of usable embryos, and usable embryo rate; however, these differences were not statistically significant ( Table 5 ). Data are presented either as mean ± standard deviation or as percentages. Abbreviations: IVF, in vitro fertilization; BMI, body mass index; AMH, anti-Müllerian hormone.

This is the first large-scale lipidomic study involving the FF to identify differences in metabolic profiles in the ovarian microenvironment associated with obesity in Korean women with infertility. We observed significant differences in the number of fertilized oocytes, CL-stage embryos, and MO-stage embryos, indicating altered embryonic development in the obese group. Analysis of the FF lipid composition provides valuable insights into the metabolic state and microenvironment of the COC. The accumulation of TAGs and DAGs, coupled with lower levels of MAGs and higher levels of ACs in the obese group, indicated inefficient lipid metabolism with insufficient energy production and potential lipotoxicity. This lipid imbalance may compromise early embryological outcomes in patients with obesity. Our results provide crucial insights into the importance of balanced lipid metabolism under conditions of increased energy demand during early embryonic development. This study provides a deeper understanding of the pathophysiology of lipid metabolism and obesity during early embryonic development. Mechanistic studies with larger sample sizes are needed to understand obesity-associated lipid metabolism and develop targeted interventions to improve fertility outcomes in women with obesity. Moreover, multicenter studies are necessary to enhance the external validity of our results and ensure their applicability to broader patient populations.

This was a single-center prospective cohort study. A total of 68 women without metabolic diseases (such as hyperlipidemia, hypertension, and diabetes mellitus) who underwent IVF/intracytoplasmic sperm injection (ICSI)-embryo transfer between February 2023 and September 2023 were included. The duration of infertility was defined as the period from the initial attempt at conception to the confirmation of clinical pregnancy as reported by the patient through self-examination. Women were divided into two BMI groups according to the KSSO definition. Group A (n = 28) was defined as the obese group with a BMI ≥ 25 kg/m 2 , whereas Group B (n = 40) was defined as the non-obese group with a BMI 15 mm during oocyte retrieval. A stereomicroscope and glass Pasteur pipette were used to retrieve the COC. Following oocyte retrieval, the FF was examined macroscopically. Follicular aspirates exhibiting severe viscosity, lack of clarity, or contamination with endometriotic cysts or the flushing medium were discarded. Only the uncontaminated FF samples were included in the analysis. All samples were preserved at 4 °C and transported to the laboratory on melting ice. This study was approved by the Institutional Review Board of CHA Gangnam Medical Center (IRB approval number: 2022-09-002). Written informed consent was obtained from all the patients. All patients underwent COH with gonadotropin-releasing hormone (GnRH) agonists or antagonists for pituitary suppression. Stimulation was personalized based on age, ovarian reserve test results, and previous ovarian responses to gonadotropins during stimulation cycles. Ovulation was induced using 250–500 μg of recombinant human chorionic gonadotropin (hCG) (Ovidrel®; Serono, Modugno, Italy) and 0.1–0.2 mg of GnRH agonist (Decapeptyl®; Ferring, Sweden) when at least three follicles with a diameter of at least 17 mm or at least two follicles with a diameter of at least 18 mm were observed in both groups. Oocytes were retrieved 34–36 h after hCG and GnRH agonist administration. Embryological outcomes were assessed as follows: Mature oocytes (metaphase II [MII] oocytes) were defined as those with the first polar body and were subsequently used for ICSI. The oocyte maturation rate was defined as the ratio of MII oocytes to the total number of oocytes retrieved. The presence of two pronuclei and a second polar body 16–18 h after insemination confirmed normal fertilization. The normal fertilization rate was calculated by dividing the number of normally fertilized oocytes by the total number of retrieved oocytes. Total cleavage (CL), morula (MO), and blastocyst (BL) embryo formation rates were determined by dividing the number of embryos at each stage by the number of normally fertilized oocytes. Usable embryos were defined as those available for fresh embryo transfer, pre-implantation genetic testing, or vitrification. The liquid chromatography-mass spectrometry (LC-MS) grade solvents (isopropanol, methanol, and water), ammonium acetate, butylated hydroxytoluene, chloroform, and methyl tert -butyl ether were procured from Sigma-Aldrich (St. Louis, MO, USA). The internal standards were procured from Avanti Polar Lipids (Alabaster, AL, USA) and Sigma-Aldrich (St. Louis, MO, USA) ( Table 1 ). This table presents the details of the internal standards from Avanti SPLASH Lipid MIX. Internal standards, crucial for ensuring the reproducibility of the analyses, were prepared at a final concentration achieved through a 100-fold dilution, as described above. Abbreviations: conc., concentration; PC, plasmenyl phosphatidylcholine; LPC, lysophosphatidylcholine; PE, phosphatidylethanolamine; LPE, lysophosphatidylethanolamine; CE, cholesteryl ester; SM, sphingomyelin; MAG, monoacylglycerol; DAG, diacylglycerol; TAG, triacylglycerol; GalCer, galactosylceramides. The Matyash method (with minor modifications) was used for all lipid extractions [ 40 ]. In brief, methanol (300 μL) containing 0.1% butylated hydroxytoluene and methyl tert -butyl ether (1 mL) was added to the FF (100 μL). After shaking for 1 h at room temperature, 250 μL of water was added, and the mixture vortexed for 10 min. Centrifugation was performed at 14,000 × g (4 °C, 15 min) for phase separation. The upper (220 μL) and lower (110 μL) phases were combined for analysis, and the solvent was evaporated. The sample was then reconstituted in 90 μL of methanol:chloroform (9:1, v/v) solvent with an internal standard mixture (10 μL) ( Fig 1 ). This mixture contained PC 33:1-d7, LPC 18:1-d7, PE 33:1-d7, LPE 18:1-d7, CE 18:1-d7, SM 36:2-d9, MAG 18:1-d7, DAG 33:1-d7, TAG 48:1-d7, cholesterol-d7, ceramide 42:1-d7, AC 2:0-d3, and galactosylceramides 30:1. LC-MS/MS (Shimadzu LCMS 8060 system, Shimadzu, Mass Spectrometry Based Convergence Research Institute, Kyungpook National University, Kyoto, Japan) was used to analyze all samples. Kinetex C18 column (100 × 2.1 mm, 2.6 μm particle size, Phenomenex, Torrance, CA, USA) was used to separate them. Mobile phases A and B consisted of 10 mM ammonium acetate in water/methanol (1:9, v/v) and 10 mM ammonium acetate in methanol/isopropanol (1:1, v/v), respectively. The gradient elution conditions were: 30% B (0 min), 95% B (15 min), 95% B (20 min), and 30% B (25 min). The sample injection volume was 1 μL, and the flow rate was maintained at 0.2 mL/min. The mass operating conditions were as follows: desolvation temperature, 250 °C; heat block temperature, 400 °C; spray voltage, 4 kV; drying gas (nitrogen) flow rate, 10 L/min; collision gas, argon; nebulizing gas (nitrogen) flow rate, 3 L/min; collision gas pressure, 230 kPa; and detector voltage, 1.82 kV. Selected reactions were monitored. Three samples were randomly selected from each group for the preliminary experiment, yielding 230 lipid species. These detected lipids were then analyzed. In the final analysis, 159 of 230 lipids ( Table 2 and S1 - S5 Tables ) in the FF samples were quantified after validating the assay by confirming the presence of lipid species with a coefficient of variation <30% of the quality control (QC) sample. The data for all detected peaks, including the peak areas and retention times, were exported to an Excel file for lipidomic analysis. The QC samples were prepared by mixing equal amounts of lipid extracts from all samples. Lipidomics data were analyzed using the Student’s t-test, Welch’s t-test, or Wilcoxon rank-sum test based on the underlying assumptions of the parametric tests, including normality and equality of variance. The analyses were performed using the stats R package. Principal component analysis (PCA) was performed using the prcomp function in the stats R package. Correlation analyses were performed using the rcorr function in the HMISC R package. The p-values were adjusted for multiple comparisons using the Benjamini-Hochberg method with the p.adjust function from the stats R package. PCA and scatter plots were generated using the ggplot2 package in R. Inter-group differences in clinical outcomes were assessed using the Student’s t-test, Welch’s t-test, or Wilcoxon rank-sum test for continuous variables, depending on the underlying assumptions of the parametric tests. Categorical variables are presented as percentages, and differences in these variables were analyzed using the Chi-squared test or Fisher’s exact test. Due to unequal sample sizes, analysis of covariance or Quade’s analysis of covariance was used to evaluate the differences between the BMI groups and embryological outcomes of IVF/ICSI, adjusting for potential confounders. Age and anti-Müllerian hormone (AMH) levels were selected as covariates based on clinical knowledge. P-values < 0.05 were considered statistically significant. All analyses were performed using R software version 4.4.1.

SRM, selected reaction monitoring; LC, liquid chromatography; MS, mass spectrometry; LPC, lysophosphatidylcholine; LPE, lysophosphatidylethanolamine; PC, phosphatidylcholine; PE, phosphatidylethanolamine (DOCX) SRM, selected reaction monitoring; LC, liquid chromatography; MS, mass spectrometry; MAG, monoacylglycerol; DAG, diacylglycerol; TAG, triacylglycerol. (DOCX) SRM, selected reaction monitoring; LC, liquid chromatography; MS, mass spectrometry; SM, sphingomyelin; CER, ceramide. (DOCX) SRM, selected reaction monitoring; LC, liquid chromatography; MS, mass spectrometry; CE, cholesteryl ester. (DOCX) SRM, selected reaction monitoring; LC, liquid chromatography; MS, mass spectrometry. (DOCX)