Correlation of blood lipid levels with the severity of polycystic ovary syndrome and its predictive value for pregnancy outcome

Abstract

Objective

To analyze the correlation between lipid levels and the severity of polycystic ovary syndrome (PCOS) and its predictive value for pregnancy outcome.

Methods

This retrospective study included 275 PCOS patients treated with ovulation induction therapy and 234 healthy controls (used only for baseline comparisons). Lipid levels were correlated with disease phenotype and sex hormones using Spearman/Pearson coefficients. Binary logistic regression and ROC curves assessed the predictive value of lipid levels for pregnancy failure.

Results

There were statistically significant differences between the two groups in glycemic indexes (fasting blood glucose (FBG), fasting insulin (FINS), homeostatic model assessment for insulin resistance (HOMA-IR)) and sex hormone indexes (testosterone (T), luteinizing hormone (LH), follicle-stimulating hormone (FSH), estradiol (E2), anti-Müllerian hormone (AMH)). The levels of total cholesterol (TC), triglycerides (TG), low-density lipoprotein cholesterol (LDL-C), and apolipoprotein B (Apo B) were significantly elevated in patients with PCOS and were closely correlated with the severity of the disease. In addition, these four lipid parameters were significantly positively correlated with T, LH, FSH, and AMH, and significantly negatively correlated with E2. Elevated levels of T, LH, TG, LDL-C, and Apo B were independent risk factors for pregnancy failure after ovulation induction treatment. TG assisted in predicting pregnancy failure after ovulation induction therapy in PCOS patients with an AUC of 0.861 (sensitivity 75.61%, specificity 85.53%); LDL-C assisted in predicting pregnancy failure after ovulation induction therapy in PCOS patients with an AUC of 0.868 (sensitivity 75.61%, specificity 83.55%); and Apo B assisted in predicting pregnancy failure after ovulation induction therapy in PCOS patients with an AUC of 0.836 (sensitivity 74.80%, specificity 86.84%).

Conclusion

Lipid levels were significantly correlated with the severity of disease in PCOS patients, and TG, LDL-C, and Apo B levels assisted in predicting the occurrence of pregnancy failure after ovulation induction therapy.

What does this study add to the clinical work

Elevated serum TG, LDL-C, and Apo B levels are significantly correlated with PCOS severity and serve as independent predictors of pregnancy failure following ovulation induction. Routine lipid screening is recommended for PCOS patients to assess fertility prognosis and guide preconception metabolic management.

Introduction

Polycystic ovary syndrome (PCOS) is the most common endocrine and metabolic disorder in women and the leading cause of anovulatory infertility and hyperandrogenemia [1]. It has been reported that about 15% of women of reproductive age worldwide suffer from PCOS, which is characterized by a combination of signs and symptoms, including polycystic ovaries (enlarged ovaries containing many antral follicles), irregular or oligomenorrhea, elevated androgen levels and insulin resistance, and obesity. Notably, a diagnosis of PCOS does not require the presence of all features; rather, only a subset of diagnostic criteria must be met [2]. Severe metabolic disorders can lead to long-term complications of PCOS, including diabetes mellitus, hyperlipidemia, cardiovascular disease, and even endometrial cancer, affecting women’s physical and mental health [3, 4]. The pathogenesis of PCOS is still unclear, and further research on the etiology of PCOS is needed to discover relevant biomarkers through appropriate screening, early and accurate diagnosis and condition assessment, and effective intervention to prevent the occurrence of long-term complications.

About 70% of PCOS patients have abnormal lipid metabolism, leading to a series of related clinical manifestations and complications, with elevated plasma total cholesterol (TC), triglycerides (TG), and low-density lipoprotein cholesterol (LDL-C), and lowered high-density lipoprotein cholesterol (HDL-C) in patients with PCOS being the typical manifestations [5]. Previous animal studies have found that implementation of a high-fat diet in prepubertal rats induces metabolic and ovarian alterations common in PCOS, thus suggesting a potential effect of hyperlipidemia on hormonal profiles, and in a high-fat diet-induced model of PCOS, ovarian changes such as increased follicular number and increased follicular wall thickness were demonstrated, suggesting that infertility is associated with dyslipidemia in PCOS [6]. Obesity has a significant effect on oocyte quality and early embryo growth, which is caused by adiposity-induced endoplasmic reticulum stress, mitochondrial dysfunction, and apoptosis [7]. Relevant clinical studies suggest that women with mild hypercholesterolemia have higher body mass index (BMI) and higher fasting insulin and IR levels than women with PCOS with normal cholesterol levels, and hypercholesterolemia is also associated with oxidative stress [8, 9]. In summary, abnormal lipid metabolism promotes pathophysiologic changes of hyperandrogenism, IR, oxidative stress, and infertility in PCOS patients. A study found that low levels of HDL and high levels of TG were prevalent in PCOS patients, and HDL predicted the development of metabolic syndrome (MS) in PCOS patients [10]. Thus, lipid abnormalities affect the pathologic development of PCOS.

Multiple metabolic pathways involve lipids, such as steroid hormone biosynthesis, sphingolipid, and fatty acid metabolism. In addition to the four lipid metabolism indices mentioned above, the lipid profile also includes small and dense low-density lipoprotein (sd-LDL-C), apolipoprotein A1 (Apo A-I), and apolipoprotein B (Apo B) [11–13]. Apo A-I is the main apolipoprotein of HDL-C, while Apo B is mainly found in atherogenic lipoprotein particles such as LDL-C, sd-LDL-C, and celiac disease [14]. Several clinical studies have shown that Apo B is a major risk factor for coronary atherosclerotic cardiovascular disease and plays an important role in lipid regulation [15, 16]. Previous studies have shown that TG and Apo B levels are positively correlated with BMI in patients with PCOS [17]. In addition, the ratios of TG/HDL and Apo B/Apo A were associated with clinical features of PCOS such as insulin resistance, obesity, and elevated androgenic and hepatic enzymes [12, 18]. Elevated fasting and postprandial plasma TG and Apo B residues in patients with PCOS provide evidence of an early clinical cardiovascular disease risk, and these markers are highly associated with impaired insulin metabolism and hyperandrogenemia [19]. A growing body of research suggests that dyslipidemia may contribute to adverse pregnancy outcomes by causing oxidative stress [20]. Currently, there are few studies on the relationship between lipid levels and PCOS severity and pregnancy prognosis. The aim of this study was to analyze the relationship between lipid levels and the severity of PCOS and its predictive value for pregnancy outcome, with the aim of providing a new reference for the clinical diagnosis and treatment of patients with PCOS.

Materials and methods

Study population

Retrospectively selected 355 PCOS patients admitted to our hospital between January 2020 and May 2023, and finally selected 275 PCOS patients as the subjects of this study according to the inclusion and exclusion criteria, and 234 healthy women who had a physical examination during the same period were selected as the control group. The control group was used exclusively for baseline comparisons and was not included in other parts of the analysis. In this study, all PCOS patients were treated with ovulation induction therapy, and general baseline data and 1-year follow-up records after outpatient treatment were collected to count the pregnancy status of the patients. The study was reviewed and approved by the Ethics Committee of our hospital in accordance with the Declaration of Helsinki.

Inclusion and exclusion criteria

The diagnosis of PCOS was in accordance with the 2003 Rotterdam criteria [21], with ultrasound criteria updated according to the 2023 ESHRE International Evidence-based Guideline for the assessment and management of polycystic ovary syndrome [22]. The following alterations were included: (1) oligo-/anovulation: defined as menstrual cycles > 35 days or fewer than 8 spontaneous ovulations per year; (2) clinical manifestations and/or biochemical signs of hyperandrogenism, where hyperandrogenism is based on elevated levels of total testosterone (> 0.7 ng/mL) or free testosterone index (> 5) measured in the laboratory. If neither total nor free testosterone is elevated, measuring dehydroepiandrosterone sulfate (DHEAS) and androstenedione levels can also be used to support the diagnosis of PCOS. Clinical features of hyperandrogenism may be acne and hirsutism (excluding other conditions that can lead to hyperandrogenism, such as congenital adrenal hyperplasia, androgen-producing tumors, and Cushing’s syndrome); and (3) polycystic ovarian morphology (PCOM) on ultrasound: a follicle number per ovary (FNPO) of ≥ 20 in at least one ovary should be considered the threshold for defining PCOM in adults. In addition, serum anti-Müllerian hormone (AMH) levels may be used as a complementary marker for PCOM in adults. A diagnosis of PCOS is made when the patient meets any two of the three criteria above.

Inclusion criteria

(1) PCOS diagnosed according to the Rotterdam criteria with specific blood parameters (as described above) and ultrasound findings, and the first time of PCOS treatment; (2) complete clinical data; (3) married infertile women aged 22–38 years; (4) normal sexual life for 1 year or more, ovulation dysfunction of the female partner, and normal sexual ability and semen examination of the male partner.

Exclusion criteria

(1) History of ovarian surgery or radiotherapy, endometriosis, uterine fibroids (> 4 cm), endometrial polyps (> 1 cm), uterine malformations, intrauterine adhesions (Asherman’s syndrome grade II or above), various hereditary diseases, and history of acute pelvic inflammatory disease within the past 6 months; (2) history of hormonal medication in the last 3 months; (3) infertility caused by central diseases, hyperprolactinemia (prolactin > 25 ng/mL), thyroid dysfunction (TSH < 0.5 or > 5.0 mIU/L), and adrenal function abnormalities (17-hydroxyprogesterone > 2 ng/mL); (4) combination of severe diabetes mellitus (HbA1c > 8.0%), cardiovascular diseases (NYHA class III or above), cerebrovascular diseases (history of stroke or TIA), severe liver dysfunction (ALT or AST > 3 times upper limit of normal), renal dysfunction (serum creatinine > 1.5 mg/dL), immune diseases requiring systemic therapy, and hematologic diseases (hemoglobin < 8 g/dL or platelet count < 50,000/μL).

Clinical data collection

Baseline data were collected from all study participants, including age, body mass index (BMI), blood pressure (systolic blood pressure (SBP), diastolic blood pressure (DBP)), blood glucose index (fasting glucose (FPG), fasting insulin (FINS), insulin resistance index (HOMA-IR)), sex hormone index (testosterone (T), luteinizing hormone (LH), follicle-stimulating hormone (FSH), estradiol (E2), anti-Müllerian hormone (AMH)), and lipid levels [triacylglycerol (TG), total cholesterol (TC), high-density lipoprotein cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-C), small, dense low-density lipoprotein (sd-LDL-C), apolipoprotein A1 (Apo A-I), and apolipoprotein B (Apo B)]. During the subjects’ natural menstrual cycle or during the period of progesterone-induced withdrawal bleeding in amenorrhea (follicular phase: days 3–5 of menstruation), after a 10 h overnight fasting period, a 5 mL venous blood sample was drawn the next morning. An electrochemiluminescence immunoassay analyzer (Cobas 8000, Roche, USA) was used to measure the levels of FINS, T, LH, FSH, E2, and AMH, and FPG was measured by the glucose oxidase assay and the HOMA-IR was calculated as HOMA-IR = FINS × FPG/22.5. A fully automated biochemical analyzer (AU5800, Beckman Coulter, USA) was used to detect TC, TG, LDL-C, HDL-C, sd-LDL-C, Apo A-I, and Apo B levels.

Disease assessment

The degree of disease was assessed according to the Rotterdam Conference [23], including anovulation or sporadic ovulation (O), elevated androgen levels with biochemical manifestations (HA), and polycystic changes in the ovaries (P). Type A was O + HA + P, type B was O + HA, type C was HA + P, and type D was O + P. For correlation analyses, these phenotypes were converted into an ordinal disease severity score based on the number of Rotterdam features present (Type D = 1, Type C = 2, Type B = 3, Type A = 4). Spearman’s rank correlation coefficient was then calculated between this score and lipid parameters.

Treatment and ovulation stimulation protocol

All patients received standardized ovulation induction therapy as previously described [24]. The stimulation protocol was as follows: Clomiphene citrate, 50 mg/day, was taken orally from the third day of menstrual cycle or withdrawal bleeding of progesterone for 5 consecutive days. On the 8th day of menstruation, human menopausal urinary gonadotropin (75 IU/d) was injected intramuscularly. Follicular monitoring was performed every 2–3 days starting from day 10 using transvaginal ultrasound. The duration of gonadotropin administration was adjusted according to follicular development. When the leading follicle diameter reached ≥ 18 mm and urine LH test was positive, human chorionic gonadotropin (10,000 IU) was administered intramuscularly as a trigger injection. Timed intercourse was advised 24–36 h after the trigger injection.

Determination of pregnancy

Serum human chorionic gonadotropin (HCG) levels were measured on the 14th day after ovulation as previously described [25], and a biochemical pregnancy was defined when HCG levels were > 25 mIU/mL. Clinical pregnancy was confirmed when the presence of the gestational sac was indicated by transvaginal ultrasonography on days 30–35 after ovulation. The follow-up records of patients 1 year after outpatient treatment were collected to count the pregnancies. The primary outcome of this study was pregnancy failure after ovulation induction therapy. Pregnancy failure was defined as: (1) failure to achieve clinical pregnancy within 12 months after trigger injection; or (2) miscarriage within ≤ 12 weeks after clinical pregnancy confirmation. Stillbirth was not included in this definition.

Statistical analysis

Sample size estimation was performed using G*Power 3.1.9.7 (University of Düsseldorf, Germany). The final sample size of 275 patients was determined to be adequate for the primary analyses: for multivariate logistic regression, a minimum of 10 events per predictor variable is recommended, and our study achieved 123 pregnancy failures with 5 predictors in the final model (events-per-variable ratio = 24.6), well exceeding this threshold. For correlation analyses, this sample size provides > 95% power to detect correlations of r ≥ 0.20 at α = 0.05. Therefore, the sample size included in the study was in accordance with the requirements of each statistical test. SPSS27.0 statistical software (SPSS, Inc, Chicago, IL, USA) and GraphPad Prism 9.5 software (GraphPad Software Inc., San Diego, CA, USA) were used for statistical analysis and graphing of data. The Kolmogorov–Smirnov test was used to test for normal distribution, and measures that conformed to normal distribution were expressed as mean ± standard deviation; independent samples t test was used for comparison between two groups; one-way ANOVA was used to analyze between multiple groups; Tukey’s multiple comparisons test was used for post hoc analysis. The correlation coefficient r was calculated using Pearson's correlation coefficient; non-normally distributed measures were expressed as the median (interquartile spacing), and comparisons between two groups were made using the Mann–Whitney U test, and multiple groups were analyzed using the Kruskal–Wallis test, and Dunn’s multiple comparisons test was used for post hoc analysis; Spearman correlation coefficient was used for correlation analysis, and the correlation coefficient r was calculated. Count data were expressed as cases and percentages, and the Chi-square test was used for between-group comparisons. A binary logistic regression model was established to analyze the influencing factors of pregnancy failure after ovulation induction therapy in PCOS patients, and ROC curves were plotted to assess the predictive value of lipid levels on pregnancy failure after ovulation induction therapy. MedCalc software was used to compare the AUC of multiple indexes, and Delong’s test was used. P was a two-sided test, and the difference was considered to be statistically significant at P < 0.05.

Results

Comparison of baseline data

Retrospectively selected 355 PCOS patients admitted to our hospital between January 2020 and May 2023, and finally selected 275 PCOS patients as the subjects of this study according to the inclusion and exclusion criteria, and 234 healthy women who had a physical examination during the same period were selected as the control group. The control group was used only for baseline comparisons and was not included in subsequent correlation analyses, logistic regression, or ROC curve analyses. There was no statistical difference between the two groups in terms of age, BMI, SBP, and DBP (all P > 0.05), while the comparison of glycemic indexes (FBG, FINS, HOMA-IR) and sex hormone indexes (T, LH, FSH, E2, AMH) between the two groups showed statistically significant differences (P < 0.05), as shown in Table 1.

Table 1.

Comparative analysis of the clinical data of the two groups of patients

Clinical information	Control subjects(N = 234)	PCOS group(N = 275)	t	P
Age (years)	28.5 (26,31)	28 (25,31)	0.215	0.830
BMI (kg/m2)	23.05 ± 2.58	23.10 ± 2.11	0.274	0.784
SBP (mmHg)	118.73 ± 10.45	119.23 ± 9.16	0.579	0.563
DBP (mmHg)	75.93 ± 6.10	76.81 ± 6.85	1.511	0.131
Glycemic index
FBG (mmol/L)	4.60 ± 0.52	5.20 ± 1.58	5.638	< 0.001
FINS (mIU/L)	8.27 ± 1.80	15.94 ± 3.55	29.881	< 0.001
HOMA-IR	1.70 (1.30,2.11)	3.56 (2.68,4.55)	16.494	< 0.001
Sex hormone index
LH (IU/L)	5.36 ± 1.04	10.95 ± 3.03	26.906	< 0.001
FSH (IU/L)	6.30 ± 1.69	7.61 ± 1.47	9.365	< 0.001
E2 (pmol/L)	171.42 ± 24.78	134.17 ± 21.20	18.277	< 0.001
T (μg/L)	0.49 ± 0.13	0.87 ± 0.21	23.938	< 0.001
AMH (ng/mL)	3.86 ± 1.05	8.44 ± 2.32	27.881	< 0.001

Adopted The Kolmogorov–Smirnov test was used to test for normal distribution, and measurements that conformed to normal distribution were expressed as mean ± standard deviation, and comparisons between the two groups were made using the independent samples t test. Measures that were not normally distributed were expressed as median (interquartile spacing), and the Mann–Whitney U test was used for comparison between the two groups

BMI body mass index, SBP systolic blood pressure, DBP diastolic blood pressure, FPG fasting plasma glucose, FINS fasting insulin, HOMA-IR HOMA insulin resistance, LH luteinizing hormone, FSH follicle-stimulating hormone, E2 estradiol, T testosterone, AMH anti-Müllerian hormone

TC, TG, LDL-C, and Apo B levels were significantly elevated in patients with PCOS

We measured the levels of TC, TG, LDL-C, HDL-C, sd-LDL-C, Apo A-I, and Apo B by fully automated biochemical analyzer, and the results showed that the levels of TC, TG, LDL-C, and Apo B were significantly higher in the PCOS group than in the patients of the control group (all P < 0.001, Fig. 1A–C, G). There was no statistically significant difference in the levels of HDL-C, sd-LDL-C, and Apo A-I between the two groups (all P > 0.05, Fig. 1D–F).

Fig. 1.

Comparison of lipid levels between the two groups of patients. TC (A), TG (B), LDL-C (C), HDL-C (D), sd-LDL-C (E), Apo A-I (F), and Apo B (G) levels were detected using a fully automated biochemical analyzer; normal distribution test was performed using the Kolmogorov–Smirnov test, and the measurements conforming to the normal distribution were expressed as the mean ± standard deviation, and comparisons between two groups were made using the independent sample t test; non-normally distributed measurements were expressed as the median (interquartile spacing), and comparisons between two groups were made using the Mann–Whitney U test, ****. Comparisons between the two groups were made using the independent samples t test; non-normally distributed measurements were expressed as the median (interquartile spacing), and comparisons between the two groups were made using the Mann–Whitney U test, ***P < 0.001

Comparison of lipid levels in PCOS patients with different conditions

We then analyzed the lipid levels of PCOS patients with different conditions, and the results showed that the levels of lipids (TC, TG, LDL-C, and Apo B) in PCOS patients increased dependently with the aggravation of the condition. This was manifested in the following ways: the levels of TC, TG, LDL-C, and Apo B were significantly higher in type A PCOS patients than in type C and type D (all P < 0.001), and the levels of TC, TG, LDL-C, and Apo B were significantly higher in type B PCOS patients than in type D (all P < 0.05), as shown in Fig. 2. In addition, we further analyzed the levels of lipids in PCOS patients with the help of Spearman coefficients the correlation between lipid levels and disease severity. The results showed that the levels of TC (r = 0.309), TG (r = 0.356), LDL-C (r = 0.355), and Apo B (r = 0.314) were positively correlated with the severity of the disease in patients with PCOS (all P < 0.001).

Fig. 2.

Comparison of lipid levels in PCOS patients in each group. The Kolmogorov–Smirnov test was used to test for normal distribution, and measurements conforming to normal distribution were expressed as mean ± standard deviation, and comparisons between the two groups were made using the independent samples t test; *P < 0.05, ***P < 0.001

Significant correlation between lipid levels and sex hormones in PCOS patients

We used Pearson coefficient to analyze the correlation between TC, TG, LDL-C, and Apo B levels and sex hormone indexes (T, LH, FSH, E2, AMH) in patients with PCOS, and the results showed that TC, TG, LDL-C, and Apo B levels showed a significant positive correlation with T, LH, FSH, and AMH (all P < 0.001), and TC, TG, LDL-C, and Apo B levels were significantly negatively correlated with E2 (all P < 0.001); see Fig. 3.

Fig. 3.

Significant correlation between lipid levels and sex hormones in patients with PCOS. The correlation between TC, TG, LDL-C, and Apo B levels and sex hormone indices (T, LH, FSH, E2, AMH) in PCOS patients was analyzed using Pearson’s coefficient. r is the correlation coefficient

Ovulation induction outcomes

Among the 275 PCOS patients who underwent ovulation induction therapy, 265 (96.4%) received the standardized stimulation protocol as described. The remaining 10 patients required dose adjustments due to poor response. Following trigger injection, 248 patients (90.2%) showed evidence of ovulation based on ultrasound confirmation of corpus luteum formation. Clinical pregnancy was achieved in 152 patients (55.3%), with 123 (44.7%) experiencing pregnancy failure as defined. Among those who achieved clinical pregnancy, 113 (74.3%) resulted in live births, while 39 (25.7%) experienced early miscarriage. The detailed outcomes are presented in Table 2.

Table 2.

Ovulation induction therapy outcomes in PCOS patients (n = 275)

Outcome	n (%)
Received standardized protocol	265 (96.4)
Required dose adjustment	10 (3.6)
Ovulation confirmed	248 (90.2)
Clinical pregnancy achieved	152 (55.3)
Pregnancy failure	123 (44.7)
No pregnancy within 12 months	84 (30.5)
Early miscarriage (≤ 12 weeks)	39 (14.2)
Live birth	113 (41.1)

Elevated TG, LDL-C, and Apo B levels are independent risk factors for pregnancy failure after ovulation induction therapy in PCOS patients

In order to accurately assess the effect of lipid levels on pregnancy after ovulation induction therapy in patients with PCOS, we used pregnancy within 1 year after outpatient treatment (0 = pregnant, 1 = not pregnant) as the dependent variable, and age, BMI, blood pressure (SBP, DBP), glycemic indexes (FBG, FINS, HOMA-IR), sex hormone indexes (T, LH, FSH, E2, AMH), and lipid indexes (TC, TG, LDL-C, HDL-C, sd-LDL-C, Apo A-I, Apo B) were included as independent variables for one-way logistic regression analysis, followed by multifactorial logistic regression analysis with the result of P < 0.05 as the independent variable. The results showed that elevated levels of T, LH, TG, LDL-C, and Apo B were independent risk factors for pregnancy failure after ovulation induction therapy (all P < 0.05, Table 3).

Table 3.

Analysis of risk factors for pregnancy failure after ovulation induction therapy in patients with PCOS

Variable	One-way analysis of variance			Multifactorial analysis
	P-value	OR value	95% CI	P-value	OR value	95% CI
Age (years)	0.280	0.968	0.912 ~ 1.027	–	–	–
BMI (kg/m2)	< 0.001	1.973	1.653 ~ 2.355	0.643	0.937	0.713 ~ 1.232
SBP (mmHg)	0.386	1.012	0.986 ~ 1.038	–	–	–
DBP (mmHg)	0.727	1.006	0.972 ~ 1.042	–	–	–
FBG (mmol/L)	0.006	1.243	1.063 ~ 1.454	0.327	0.587	0.203 ~ 1.701
FINS (mIU/L)	< 0.001	1.448	1.309 ~ 1.601	0.272	0.807	0.550 ~ 1.183
HOMA-IR	< 0.001	2.519	1.946 ~ 3.259	0.441	1.835	0.392 ~ 8.590
LH (IU/L)	< 0.001	1.703	1.491 ~ 1.945	0.034	1.217	1.014 ~ 1.459
FSH (IU/L)	< 0.001	2.218	1.765 ~ 2.786	0.951	1.010	0.725 ~ 1.409
E2 (pmol/L)	< 0.001	0.914	0.895 ~ 0.934	0.206	0.980	0.950 ~ 1.011
T (μg/L)	< 0.001	38.552	6.993 ~ 212.551	0.046	9.773	1.038 ~ 92.018
AMH (ng/mL)	< 0.001	1.621	1.410 ~ 1.863	0.218	0.878	0.714 ~ 1.080
TC	< 0.001	4.973	3.189 ~ 7.754	0.897	1.039	0.583 ~ 1.853
TG	< 0.001	36.198	9.288 ~ 141.068	0.048	7.901	1.022 ~ 61.060
LDL-C	< 0.001	17.221	8.808 ~ 33.669	0.045	2.773	1.022 ~ 7.522
HDL-C	0.875	1.079	0.416 ~ 2.796	–	–	–
sd-LDL-C	0.792	1.093	0.564 ~ 2.117	–	–	–
Apo A-I	0.784	1.192	0.340 ~ 4.182	–	–	–
Apo B	< 0.001	61.559	12.364 ~ 306.507	0.022	11.809	1.438 ~ 97.007

Lipid levels in PCOS patients assist in predicting pregnancy failure after ovulation induction therapy

Finally, we analyzed the predictive value of TG, LDL-C, and Apo B levels on pregnancy failure after ovulation induction therapy in PCOS patients by plotting ROC curves, and the results showed that the AUC of TG-assisted prediction of pregnancy failure after ovulation induction therapy in PCOS patients was 0.861 (95% CI 0.814 ~ 0.900), with a cut-off value of 1.7, and the sensitivity was 75.61% and specificity of 85.53% (Fig. 4A); LDL-C-assisted prediction of pregnancy failure after ovulation induction therapy in PCOS patients had an AUC of 0.868 (95% CI 0.823 to 0.906), a cut-off value of 3.38, a sensitivity of 75.61%, and a specificity of 83.55% (Fig. 4B); and Apo B-assisted prediction of pregnancy failure after ovulation induction therapy in PCOS patients had an AUC of 0.836 (95% CI 0.787 to 0.878), a cut-off value of 1.07, a sensitivity of 74.80%, and a specificity of 86.84% (Fig. 4C).

Fig. 4.

ROC curve analysis of TG, LDL-C, and Apo B levels to predict pregnancy failure after ovulation induction therapy in PCOS patients. ROC curve analysis of the predictive value of TG (A), LDL-C (B), and Apo B (C) levels in predicting pregnancy failure after ovulation induction therapy in PCOS patients

Discussion

PCOS is a common gynecological endocrine disease characterized by persistent anovulation or oligoovulation, hyperandrogenic manifestations, or polycystic changes in the ovaries, often accompanied by insulin resistance and obesity [25]. Dyslipidemia, as one of the complications of PCOS, has a profound impact on the health of patients [26]. The aim of this paper is to investigate the correlation between lipid levels and the severity of PCOS, and to provide new ideas for subsequent clinical management and intervention strategies.

First, in a comparative study of participants, we found that TC, TG, LDL-C and Apo B levels were significantly elevated in patients with PCOS, which was largely consistent with the results of previous studies [17]. The causes of dyslipidemia are complex, including genetic factors, dietary irrationality, medication factors, obesity, lack of exercise, age, and gender [27]. For PCOS patients, hyperandrogenism, insulin resistance and oxidative stress, and inflammatory response are important triggers of dyslipidemia [28, 29]. Hyperandrogenemia is an important feature of patients with PCOS, and a previous study reported that PCOS patients without hyperandrogenemia had lower HDL-C levels [30]. Abruzzese G A et al. found that all patients with PCOS, independently of androgen status, exhibited high metabolic risk [31]. These studies suggest that androgens have an adverse effect on lipid metabolism, which is not directly affected by dyslipidemia in patients with PCOS. However, recent studies have produced different results. Gobl C S et al. found that hepatic adiposity was strongly associated with hyperandrogenemia and correlated with adverse metabolic risk profiles [32]. Obese PCOS patients have significantly lower sex hormone-binding globulin levels and significantly higher free testosterone levels [33]. Hyperandrogenemia in normal-weight women with PCOS is associated with preferential intra-abdominal fat deposition and an increased number of small SC abdominal adipocytes that can limit SC fat storage and promote metabolic dysfunction [34]. In addition, women with PCOS have adipose tissue dysfunction in the SC region, characterized by alterations in adipocyte size and leptin/lipocalin expression and secretion, likely associated with higher androgen concentrations [35]. All these findings suggest that hyperandrogenemia is associated with fat distribution. On the other hand, insulin resistance leads to increased insulin secretion, which in turn promotes fat synthesis and accumulation [36]. High insulin levels promote hepatic synthesis of cholesterol and triglycerides, as well as inhibit the oxidative breakdown of fatty acids, leading to elevated lipid levels [37]. In addition, insulin resistance may also affect the metabolism of LDL-C, resulting in smaller particles and higher density, forming small, dense LDL-C, which is more susceptible to oxidation and modification, thereby increasing the risk of atherosclerosis [38]. Apo B is the major apolipoprotein of LDL-C, and its elevated level is usually closely correlated with the elevated level of LDL-C [39]. In patients with PCOS, the synthesis and secretion of LDL-C are increased due to endocrine disorders, insulin resistance, and abnormalities in lipid metabolism, resulting in elevated Apo B levels. In addition, in one study, increased inflammation in PCOS was manifested by decreased levels of lipocalin and increased levels of adipokines [40], suggesting that dyslipidemia is closely related to inflammation. Subsequently, we analyzed the lipid levels of PCOS patients with different conditions, and the results showed that the levels of lipids (TC, TG, LDL-C, and Apo B) in PCOS patients increased dependently with the aggravation of the condition. In addition, the levels of TC, TG, LDL-C, and Apo B in PCOS patients were positively correlated with the severity of the disease. In view of the correlation between lipid levels and the severity of PCOS, lipid levels can be used as an auxiliary indicator to assess the severity of PCOS. Through regular monitoring of lipid levels, physicians can keep abreast of the progress of PCOS patients and adjust the treatment plan accordingly.

We further used Pearson coefficient to analyze the correlation between TC, TG, LDL-C, and Apo B levels and sex hormone indexes (T, LH, FSH, E2, AMH) in PCOS patients, and the results showed that the levels of TC, TG, LDL-C, and Apo B showed a significant positive correlation with T, LH, FSH, and AMH, and the levels of TC, TG, LDL-C, and Apo B showed a significant negative correlation with E2. It is suggested that the lipid differences in PCOS patients may be related to endocrine disorders. Endocrine disorders, especially elevated androgen levels, exist in PCOS patients. Elevated androgens may be caused by excessive adrenal cortical secretion or excessive production by follicular membrane cells, and these abnormalities may be associated with genetic factors, such as mutations in certain genes on the X chromosome [41]. Elevated androgens not only affect reproductive function, but may also interfere with fat metabolism, leading to abnormal lipid levels [42]. In addition, E2 has a regulatory role in lipid metabolism in women. It can lower LDL-C levels, inhibit hepatic enzyme activity to reduce HDL-C degradation, accelerate hepatic clearance of celiac disease, and promote bile acid secretion thereby accelerating cholesterol clearance by the body [43]. Although the direct role of LH, FSH, and AMH in lipid regulation is not as obvious as that of E2, they may indirectly affect lipid levels by influencing systemic metabolic status, hormone level balance, and so on.

In addition, our study showed that elevated levels of TG, LDL-C, and Apo B were independent risk factors for pregnancy failure after ovulation induction therapy, and all three aided in predicting the occurrence of pregnancy failure after ovulation induction therapy. A growing number of studies have shown that lipid levels are strongly associated with female reproductive outcomes [44]. Endometrial tolerance and early embryo implantation capacity may be affected by lipid changes during the window of implantation [45, 46]. In patients with PCOS, since they often have dyslipidemia, lipid levels may be an important factor affecting their pregnancy success after ovulation induction therapy. Previous studies have shown that elevated lipids are negatively associated with reproductive outcomes in women with PCOS with or without clomiphene combined with acupuncture to induce ovulation [47]. Therefore, by determining preconception lipid levels, physicians can assess the fertility risk of patients and take appropriate therapeutic measures to improve pregnancy success. For example, in patients with dyslipidemia, lipid levels can be improved by dietary adjustments, increased exercise, or the use of lipid-lowering drugs, which may improve pregnancy success.

We chose pregnancy failure as a composite endpoint because dyslipidemia is biologically plausible to impair both implantation and early embryo development. Combining failure to achieve clinical pregnancy and early miscarriage into a single outcome therefore captures the full spectrum of adverse reproductive events after ovulation induction in women with PCOS.

The findings of this study have important clinical implications for the management of PCOS patients. Dyslipidemia in PCOS not only affects reproductive outcomes but also serves as an important marker for cardiometabolic risk. The strong correlation between lipid levels and pregnancy failure suggests that lipid management should be an integral part of preconception care in PCOS patients. Early intervention strategies including lifestyle modifications (diet and exercise), omega-3 fatty acid supplementation, and when necessary, statin therapy (with appropriate contraception until conception is desired), may improve both reproductive outcomes and long-term cardiovascular health. Our results support routine lipid screening in all PCOS patients, particularly those planning pregnancy, with targeted interventions for those with TG > 1.7 mmol/L, LDL-C > 3.38 mmol/L, or Apo B > 1.07 g/L. This integrated approach to managing both metabolic and reproductive aspects of PCOS may lead to better overall patient outcomes.

In summary, we found that dyslipidemia was significantly associated with the severity of disease in patients with PCOS, and that TC, TG, LDL-C, and Apo B levels assisted in predicting the occurrence of pregnancy failure after ovulation induction therapy. However, limitations include the retrospective, single-center design, small sample size, and potential bias, and larger, multicenter studies are needed to validate this result. In addition, there is a considerable gap in fully elucidating the underlying mechanisms and therapeutic implications.

Author contributions

DTM and JT were involved in the conception and design, or analysis and interpretation of the data; YNW in the drafting of the paper, revising it critically for intellectual content; CCW in the final approval of the version to be published; and all the authors agree to be accountable for all aspects of the work.

Funding

No funding was received.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Conflict of interest

The authors declare no competing interests.

Ethical approval

The study was approved by the Ethics Committee of The First College of Clinical Medical Science, China Three Gorges University (Yichang Central People’s Hospital).

Informed consent

All the authors gave written informed consent.

Consent for publication

All the authors have agreed to publish.