This part evaluated the clinical efficacy and metabolic effect of acupuncture combined with IVF on PCOS undergoing IVF from the perspectives of oocyte quality, embryo quality, and follicular fluid metabolism.

The FF was collected on the day of oocyte retrieval from PCOS. FF in the first and last tube may affect the experimental results due to mixing organic solvent rinse. Therefore, yellowish, clear, and bloodless FF were collected after removing the first and last FF mixed thoroughly, and centrifuged at 12,000 rpm, temperature of 4 °C for 10 min. Supernatants were then collected and frozen in an ultra-low temperature freezer at −80 °C waiting to be metabolite extracted.

16S rRNA gene sequencing was carried out on all fecal samples, and metagenomic sequencing was implemented on a randomly chosen subset that constituted one-tenth of the entire sample. Total genomic DNA was extracted from fecal samples using the SDS method, with concentration and purity assessed via 1% agarose gels electrophoresis. DNA was diluted to 1 ng/μL in sterile water based on measured concentrations. The V3-V4 regions of the 16S rRNA gene were amplified using specific primers 341F (5′-CCTAYGGGRBGCASCAG-3′) and 806R (5′-GGACTACNNGGGTATCTAAT-3′) (Qin et al., 2024). The PCR amplifications utilized Phusion® High-Fidelity PCR Master Mix, with 10 ng of template DNA input. PCR products were quantified and qualified by mixing equal volumes with 1X loading buffer containing SYBR Green and running electrophoresis on 2% agarose gel. The resulting PCR amplicons were cleaned up using Qiagen gel extraction kits and sequencing libraries were assembled with TruSeq® DNA PCR-free kits, incorporating index codes. The libraries’ quality underwent examination with a Qubit 2.0 fluorometer and the Agilent Bioanalyzer 2,100 system. Finally, libraries were sequenced on the Illumina NovaSeq platform, yielding 250 bp paired-end reads.

The raw data obtained were concatenated and filtered, followed by denoising using DADA2. Sequences with abundances of less than 5 were filtered out to yield the final amplified sequence variants (ASVs). After obtaining the ASVs, alpha diversity and β diversity analyses were conducted for the ASVs to provide insights into species richness and evenness within samples, and differences in community structure between different groups (Li et al., 2020). Species exhibiting significant differences between and within groups before and after treatment were screened using the Linear discriminant analysis Effect Size (LEfSe) method.

For library construction, genomic DNA was fragmented into brief segments, followed by end repair, A-tailing, adapter ligation, purification, and PCR amplification. Qualified libraries were pooled based on concentrations and data output, while size distribution was evaluated via bioanalyzer. Then sequenced on Illumina platform with PE150 reads.

DIAMOND software was used to align Unigenes with those in the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. From the alignment results of each sequence, the Best Blast Hit results were selected for subsequent analysis. According to the alignment results, the relative abundance at different functional levels was calculated. Based on the abundance table at each taxonomy level, combined with principal component analysis of dimension reduction, metabolic pathway comparative and LEfSe on inter-group functional difference was analyzed.

The FF samples were centrifuged after using 80% methanol aqueous and mass spectrometric grade water. The supernatant was collected and analyzed using ultra-high performance liquid chromatography combined with high-resolution mass spectrometry, specifically liquid chromatography-mass spectrometry technology for separation. Data files generated by the instrument were imported into CD3.1 search software for processing, and search comparisons were conducted using mzCloud, mzVault, and Masslist databases to ultimately obtain metabolite identification and relative quantification results. For quality control purposes, QC samples were also established during the processing of the samples.

Identified metabolites were annotated using KEGG database. In the multivariate statistical analysis part, metabolomics data processing software metaX was used to transform the data and partial least squares discriminant analysis (PLS-DA) was performed. Stem plots were made using the R package ggplot2 to visualize the data to screen for metabolites. Correlations between differential metabolites (Spearman correlation coefficients) were performed using the corrplot software package in R language to draw correlation plots. Enrichment analysis was performed using the KEGG database to investigate the function and metabolic pathways of metabolites.

Statistical analysis was conducted using IBM SPSS Statistics 27.0 software, and chi square test was used for count data; Exploratory analysis was first conducted on the quantitative data. If the data conforms to normal distribution and homogeneity of variance, t-test was used. Among them, inter group comparison was conducted using two independent sample t-test; Two paired sample t-test was used for comparison before and after treatment within the group; The Mann Whitney test was used for data that does not follow a normal distribution. All tests were conducted using a two-sided test, with a statistical threshold set at p-value < 0.05. Concurrently, correlation analyses were conducted between key species and clinical parameters to investigate the potential clinical impacts of various species.

PCOS has systemic metabolic abnormalities, including obesity (Wang et al., 2023), hyperandrogenism (Livadas et al., 2014), hyperinsulinemia (Greenwood and Huddleston, 2019), lipid metabolism disorders (Saha et al., 2008), and elevated AMH and LH levels (Homburg et al., 2013) in this research and previous studies. Obesity is a common feature in PCOS. While the average BMI in our study did not meet the criteria for obesity, individuals with PCOS exhibited significantly higher weight and BMI compared to the healthy control group, aligning with findings from previous studies (Yu et al., 2022), including one conducted in China (Zhou et al., 2020). In addition, the waist circumference and WHR in PCOS were increased, which linked to reproductive outcomes in previous research (Xia et al., 2024). The 2023 “International Evidence-Based Guidelines for Polycystic Ovary Syndrome” indicated that AMH can serve as an alternative to transvaginal ultrasound for diagnosis PCOS (Teede et al., 2023a), consistent with the clinical manifestations of elevated AMH and LH levels in PCOS (Teede et al., 2023a, 2023b; Guo et al., 2022). In glucose and lipid metabolism, PCOS exhibited higher FINS, HOMA-IR, LDL-C, and APOB levels compared to HC (Guo et al., 2022), with studies linking elevated LDL-C to lower ovulation rates (Cai et al., 2022).

Significant differences were observed in gut microbiota composition, abundance, diversity, and KEGG metabolic pathways in PCOS compared to HC., with increased fructose and mannose metabolism (ko00051) (Hanna et al., 2025) and suppressed amino acid metabolism (ko00250) pathways (Yang et al., 2022). Consistent with prior studies, The alpha diversity and beta diversity analyses indicated lower microbial diversity and richness in PCOS (Yu et al., 2022; Zou et al., 2023; da Silva et al., 2024), associated with PCOS pathogenesis (Corrie et al., 2021). s_Lachnospiraceae, s_Blautia_sp., and g_Escherichia-Shigella are specific gut microbiota of PCOS. The s_Blautia_sp. in Chinese PCOS is more abundant than in HC (Yang et al., 2024) and negatively correlated with LH and AMH. S_Blautia_sp., as an anaerobic Gram-positive bacterium commonly found in the gastrointestinal tract of mammals (Murphy et al., 2025), is associated with the production of primary bile acids such as bile acid (Islam et al., 2011; Yu et al., 2022). Bile acid, as a common secondary metabolite, regulates IL-22 to impact ovarian function and hormone secretion and plays an important role in the pathogenesis and treatment of PCOS (Qi et al., 2019). This research found a positive correlation between FBG and s_Lachnospiraceae. In the letrozole-induced PCOS mouse model, both s_Lachnospiraceae and s_Blautia_sp. (Kelley et al., 2016) were increased, which may be closely related to the pathogenesis of PCOS. Both this research and other studies have found the abundance of g_Escherichia-Shigella increased in PCOS, with g_Shigella potentially making the host more susceptible to metabolic disorders and inflammation, thereby contributing to the pathogenesis of PCOS (Chu et al., 2020; Yu et al., 2022).

Through further metagenomic analysis and KEGG annotation, we found that fructose metabolism (ko00051) in PCOS significantly increased, which is consistent with previous research (Hanna et al., 2025), correlated with FBG. HC exhibited higher energy metabolism pathways, including amino acid metabolism (ko00250) correlated with weight, BMI, and waist circumference, pyrimidine metabolism (ko00240), and general metabolism (ko00760). The previous research found the amino acid metabolism pathway (ko00250) decreased, and energy metabolism pathways such as carbohydrate digestion and absorption are decreased in PCOS compared to HC (Haudum et al., 2020), indicating energy metabolism dysregulation in PCOS.

This study demonstrated that, unlike the observed increase in body weight and BMI following IVF treatment alone, individuals with PCOS who underwent acupuncture in conjunction with IVF treatment experienced a reduction in body weight and BMI. The inter-group comparison of pre- to post-treatment differences was statistically significant (p < 0.05). Although no studies have consistently reported universal weight gain during IVF cycles in individuals with PCOS, a high BMI is associated with reduced oocyte quality (Raviv et al., 2020). The B group increased in weight, BMI, hip circumference, APOA1, and FINS after IVF, indicating exacerbated glucose and lipid metabolism disorders. And the A group showed decreased FBG after acupuncture combined with IVF treatment (p < 0.05), likely due to acupuncture improving IR by regulating the insulin signaling molecules (Huang et al., 2016). Interestingly, no significant differences in T or TG levels were observed between groups, though both showed increased levels post-treatment, possibly due to ovulation induction drugs like letrozole, which enhance LH and follicle-stimulating hormone, increasing natural T production (Bulow et al., 2022).

While the post-treatment alpha diversity and beta diversity of the two groups remained unchanged, LEfSe analysis showed the abundance of g_Escherichia-Shigella significantly decreased after acupuncture combined with IVF treatment compared to IVF treatment alone (d = 0.307, p = 0.017). Furthermore, our team previous studies and other previous researches have indicated that the abundance of g_Escherichia-Shigella is significantly elevated in the gut microbiota of PCOS (Liu et al., 2017; Yu et al., 2022) and is associated with the pathogenesis of PCOS, potentially serving as a potential microbial biomarker (Senthilkumar and Arumugam, 2025). The g_Escherichia-Shigella is linked to diarrhea (Devanga Ragupathi et al., 2018) and inflammatory bowel diseases (Zhao et al., 2023), inducing gut inflammation, affecting host metabolism and insulin sensitivity, and exacerbating PCOS symptoms, reducing oocyte quality (Corrie et al., 2023). This study found that g_Escherichia-Shigella was positively correlated with LDLC and TCH. Previous studies have shown that lipid metabolism improvements coincide with decreased g_Escherichia-Shigella abundance (Jin et al., 2021; Li et al., 2024), indicating acupuncture may having a positive effect on lipid metabolism by regulating the abundance of g_Escherichia-Shigella. Previous research has reported a negative correlation between the abundance of g_Escherichia-Shigella and the level of ghrelin (Liu et al., 2017), a mediator of the brain-gut axis, and a decrease in both ghrelin and peptide YY (PYY) levels in PCOS (Zhang et al., 2019b). These brain-gut axis mediators, including ghrelin and PYY, are believed to regulate appetite and promote mental health in women with PCOS (Konturek et al., 2004; Lang et al., 2015). However, the effects of acupuncture on the gut-brain axis and metabolism in PCOS remain to be investigated. The inflammatory pathways including IL-17 signaling (ko04657), PI3K-Akt signaling (ko04151), and Th17 cell differentiation (ko04659) suppressed in PCOS after the acupuncture combined with IVF treatment, indicating decreased systemic inflammation post-treatment, consistent with prior reports on the effects of acupuncture for PCOS (Cochran et al., 2024).

The A group had an oocyte maturation rate of 89.4% compared to 87.6% in the B group (p > 0.05). It suggested that acupuncture has a positive effect on the quality of oocytes in PCOS undergoing IVF. Mature oocyte rate refers to the proportion of mature oocytes in the number of oocytes retrieved, and is one of the evaluation indicators of oocyte quality. High-quality oocytes are the basis for the development of high-quality embryos, which have better developmental potential and higher pregnancy rates after transfer into the uterus (Gunther et al., 2022). For PCOS with low oocyte maturation rate, increasing oocyte maturation rate can improve the success rate of IVF (Xiang et al., 2021).

FF contains essential metabolites for oocyte development (Zhang et al., 2017). Non-targeted metabolomics analysis of FF revealed significant differences, with lipids and lipid-like molecules comprising 89.5% after acupuncture combined with IVF treatment versus 70.81% after IVF treatment only. Lipids and lipid-like molecules are closely related to the quality of oocytes in PCOS (Ding et al., 2022; Zhang et al., 2024). Analysis of the differentially expressed metabolites after intervention revealed that the relative content of boldione was significantly lower. Boldione and MetO were closely related to the development of oocytes in PCOS. Boldione, a precursor to anabolic steroids like T (Tarkowská, 2019), correlated with PCOS symptoms like hirsutism and acne, as well as elevated T levels. Studies have shown that elevated T levels in follicular fluid of PCOS can affect follicular cell development (Eini et al., 2022) and even promote local ovarian inflammation, leading to the pyroptosis of ovarian granulosa cells (Xiang et al., 2023). This study found that boldione is negatively correlated with the number of high-quality embryos at the cleavage stage in PCOS, potentially promoting ovarian inflammation to impact follicular cell development. Recent studies have found that boldione is related to the synthesis pathway of steroid-based drugs (Ke et al., 2024). MetO, a marker of oxidative stress due to methionine oxidation (Cabreiro et al., 2006), was negatively correlated with oocyte and embryo quality (Zhang et al., 2022). MetO had positive causal effects on PCOS risk in a bidirectional mendelian randomization study (Lei et al., 2025). These findings suggested that elevated MetO may induce increased oxidative stress in the ovary, thereby impairing oocyte quality in PCOS. KEGG enrichment analysis revealed the enrichment in steroid hormone biosynthesis, ovarian steroidogenesis, fatty acid biosynthesis, and amino acid biosynthesis. Steroids are involved in follicular development, ovulation, and pregnancy maintenance, as well as in regulating Gn secretion in the systemic circulation (Drummond, 2006). In animal models of anovulation exposed to androgens prenatally, the steroidogenic pathway was impaired in a phenotype-specific manner (Abruzzese et al., 2020), leading to dysfunctional steroid biosynthesis in PCOS. Previous studies also have suggested that steroid biosynthesis is dysfunctional in PCOS (Sander et al., 2011).

Multiple studies have shown that glucose metabolism disorders can lead to mitochondrial dysfunction and abnormal glycolysis in oocytes of PCOS, affecting the switch of oocyte energy metabolism and resulting in a decrease in oocyte maturation rate (Wang and Wu, 2020; Zhang et al., 2022). Dysregulation of lipid metabolism can cause biosynthetic metabolic dysfunction, leading to a decrease in fertilization rate and thus reducing embryo quality in PCOS (Sacks et al., 2018). Acupuncture had a positive effect on balancing glucose and lipid metabolism in PCOS undergoing IVF, regulating the relative abundance of key gut microbiota g_Escherichia-Shigella in PCOS, and improving systemic inflammatory metabolic pathways, which may be related to the improvement of oocyte and embryo quality in PCOS by acupuncture. A recent study has also demonstrated that the gut microbiota can ameliorate the reproductive phenotype, inhibit ovarian AMH expression, and modulate key metabolites in PCOS (Li et al., 2025). However, further research is needed to explore this further.