Intended for healthcare professionals
Original research•

Prediction of clinical pregnancy after in vitro fertilisation-embryo transfer in infertile patients with polycystic ovary syndrome

Abstract

Background Polycystic ovary syndrome (PCOS) affects a patient’s ovarian function and reduces a patient’s fertility. This study aimed to evaluate predictors of clinical pregnancy after in vitro fertilisation-embryo transfer (IVF-ET) in infertile PCOS patients.

Methods The study included infertile patients with PCOS who received IVF-ET from January 2017 to March 2024 at the Reproductive Medicine Center of the General Hospital of Northern Theater Command. This study analysed the predictors of clinical pregnancy after IVF-ET in infertile patients with PCOS and analysed the predictive effect of these factors on clinical pregnancy.

Results A total of 425 patients with PCOS who met the criteria were included in the study. Multivariate logistic regression analysis found that luteinising hormone (LH)/follicle-stimulating hormone (FSH) (OR=0.577, 95% CI 0.422 to 0.789; p=0.001), free thyroxine (FT4) level (OR=1.107, 95% CI 1.009 to 1.214; p=0.031), number of embryos transferred (OR=1.781, 95% CI 1.238 to 2.562; p=0.002), number of oocytes retrieved (OR=1.071, 95% CI 1.016 to 1.130; p=0.011), optimal embryo rate (OR=21.347, 95% CI 2.647 to 172.131; p=0.004) and level of antral follicles (OR=1.444, 95% CI 1.131 to 1.845; p=0.003) were independent predictors of clinical pregnancy after ET in PCOS infertile patients. The predictive model constructed based on these predictors had an area under the curve of 0.732 (95% CI: 0.683 to 0.782) for clinical pregnancy.

Conclusions Our study found that LH/FSH, FT4 level, number of embryos transferred, number of oocytes retrieved, optimal embryo rate and level of antral follicles were significantly associated with clinical pregnancy rate in PCOS infertility patients after assisted reproductive therapy. Combining these variables can effectively predict clinical pregnancy.

What is already known on this topic

  • Polycystic ovary syndrome (PCOS) compromises ovarian function and fertility. Previous studies have examined PCOS’s impact on reproductive outcomes, yet specific predictors of clinical pregnancy after in vitro fertilisation-embryo transfer (IVF-ET) in infertile PCOS patients remain unclear.

What this study adds

  • This study identified independent predictors of clinical pregnancy in infertile PCOS patients undergoing IVF-ET, such as luteinising hormone/follicle-stimulating hormone ratio, FT4 levels, embryo and oocyte numbers, optimal embryo rate and antral follicle count. A predictive model based on these factors showed good accuracy (area under the curve=0.732).

How this study might affect research, practice or policy

  • The findings may help clinicians optimise IVF-ET protocols for PCOS patients, guide future research on improving reproductive outcomes and enhance the efficiency and success of assisted reproductive therapies by incorporating these predictors into clinical decisions.

Introduction

About 3%–10% of women of reproductive age in the world are affected by polycystic ovary syndrome (PCOS).1 PCOS is closely related to reproductive and metabolic diseases in patients, often leading to anovulatory infertility, miscarriage, ovarian hyperstimulation syndrome, gestational diabetes, gestational hypertension, preterm birth, abnormal body weight and other serious health problems.2 3 The study found PCOS and secondary diseases affect a patient’s ovulatory function, oocyte quality and endometrial receptivity4 and may ultimately lead to infertility.

Treatment of infertility in PCOS patients includes both non-pharmacological and pharmacological treatments, with lifestyle changes being the first-line non-pharmacological treatment.5 Clomiphene citrate is a first-line ovulation induction drug for infertility in women with PCOS.6 Letrozole is an aromatase inhibitor that causes lower E2 levels. Letrozole can be used as second-line therapy in women who are Clomiphene citrate-resistant or have failed treatment without other factors of infertility.7 Gonadotropin therapy is another second-line treatment option for infertility in PCOS patients.7 In the absence of an absolute indication for in vitro fertilisation/intracytoplasmic sperm injection (IVF/ICSI), IVF/ICSI can be used as a third-line treatment for women with PCOS and anovulatory infertility who have failed first-line or second-line ovulation induction therapy.8 IVF-embryo transfer (ET) is a common treatment option for women with anovulatory infertility and PCOS in China. However, the clinical pregnancy rate is low. Identifying predictors of clinical pregnancy in PCOS infertile patients has important implications for the management of assisted reproduction in patients. There are many factors affecting clinical pregnancy, including female age, gonadotrophin days, cycle treatment options, single/double intrauterine insemination, sperm straight-line velocity, sperm deformity index, endometrial thickness, the rate of transferrable embryos per oocytes received and number of embryo transfers.9–11

Numerous factors influence clinical pregnancy, and these can be broadly categorised to enhance clarity.

  • Patient factors: These include female age, which is a well-established determinant of reproductive potential as it is associated with a decline in ovarian reserve and oocyte quality over time. Baseline biochemical parameters such as hormone levels (eg, follicle-stimulating hormone (FSH), luteinising hormone (LH), oestradiol) also fall under this category, as they reflect the patient’s endocrine status and can impact ovarian response and embryo implantation.

  • Stimulation-related factors: The duration of gonadotrophin treatment (gonadotrophin days) is a key factor in this group. It reflects the intensity and duration of ovarian stimulation, which in turn affects the number and quality of oocytes retrieved. Additionally, the choice of cycle treatment options (eg, long-acting GnRH, modified long-acting GnRH, GnRH antagonist regimen, luteal phase superovulation, other superovulation) can influence the ovarian response and the subsequent reproductive outcomes.

  • Sperm factors: Sperm characteristics, such as sperm straight-line velocity and sperm deformity index, are crucial as they determine the fertilising capacity of the sperm and can impact the likelihood of successful fertilisation and subsequent embryo development.

  • Embryo and endometrial factors: Endometrial thickness is an important parameter as it reflects the receptivity of the endometrium to embryo implantation. The rate of transferable embryos per oocytes received and the number of embryos transferred are also key embryo-related factors that directly influence the chances of clinical pregnancy.

Despite the multitude of factors affecting clinical pregnancy, there is a dearth of studies in China that have comprehensively analysed the predictors of clinical pregnancy following IVF-ET in PCOS infertile patients. The primary objective of this study was to identify the independent predictors of clinical pregnancy after IVF-ET in PCOS infertile patients and to assess the predictive performance of models constructed based on these predictors on clinical pregnancy.

Methods

Study design and patients

This retrospective study included infertile PCOS patients who underwent IVF-ET at the Reproductive Medicine Center of Northern Theater Command General Hospital from January 2017 to March 2024, with inclusion criteria being: (1) age ≥20 and <40 years; (2) confirmed PCOS diagnosis meeting the 2003 Rotterdam criteria (established by ESHRE/ASRM, requiring at least two of: clinical/biochemical hyperandrogenism, oligo-ovulation/anovulation or polycystic ovaries on ultrasound) and (3) undergoing first-cycle IVF-ET treatment with signed informed consent. Exclusion criteria comprised: (1) severe hypertension, diabetes, thyroid disorders, congenital adrenal hyperplasia, androgen-secreting tumours, Cushing’s syndrome, or other endocrine disorders mimicking PCOS; (2) reproductive tract anomalies, masses, or uterine abnormalities; (3) genetic/chromosomal disorders in either partner; (4) severe male factor infertility (oligozoospermia/asthenozoospermia/teratozoospermia); (5) cancelled IVF-ET cycles or (6) incomplete clinical/follow-up data. The study was approved by the hospital’s ethics committee [Approval Number: Research Ethics Review No. (2025) 002].

Assisted reproduction protocols

Patients received ovulation induction with any of the following five superovulation regimens: long-acting GnRH, modified long-acting GnRH, GnRH antagonist regimen, luteal phase superovulation, other superovulation. During superovulation, clinicians assess follicle development based on vaginal ultrasonography combined with serum hormone levels. IVF or ICSI insemination was performed 4–6 hours after oocyte retrieval, and fertilisation was observed 16–20 hours after fertilisation. 72 hours after fertilisation, the quality of embryos was evaluated, and 1–2 embryos were selected for ET. Serum β-HCG levels were measured 2 weeks after ET in patients receiving luteal support, and B-mode ultrasound was performed 33–35 days after transplantation. Patients with a visible gestational sac and yolk sac were regarded as having a clinical pregnancy.

Data handling and missing data

In this study, efforts were made to minimise data loss through strict data collection and quality control procedures. However, in the case of missing data, the following methods are adopted. Just delete the data.

Variables and definitions

The variables of interest in this study included general patient characteristics, biochemical parameters, IVF-ET-related variables and outcome variables. The diagnosis of PCOS can refer to the Rotterdam standard12 diagnosis, when any two of the following three conditions are met. The patient presented with (1) clinical or biochemical hyperandrogenism, (2) menstrual irregularities and anovulation and (3) ultrasonographic evidence of polycystic lesions.

Statistical analysis

Kolmogorov-Smirnov test was used to check for normality of continuous variables. The normally distributed variables were expressed as mean±SD, and the skewed or unknown distributed variables were displayed as median and interquartile range (IQR). Categorical variables were presented as count (percentage). Univariable logistic regression was used to analyse the factors associated with clinical pregnancy. The factors with p<0.1 in univariable logistic regression were further analysed in a multivariate logistic regression model to screen the predictors of clinical pregnancy. Receiver-operating characteristic (ROC) curve and the area under the ROC curve (AUC) were used to evaluate discriminative ability feasibility of predictors. The clinical pregnancy probability was estimated using the nomogram. A calibration plot was used to evaluate calibrating ability. All tests were two-sided. A p<0.05 was considered significant. All statistical analyses were performed using the R software (V.4.2.0).

Results

Baseline characteristics of patients with PCOS

A total of 425 patients with PCOS who met the criteria were included in the study. The mean age of the patients was 30.56±3.47 years, and the median duration of infertility was 4.00 (2.00–5.00) years. Among them, 63.53% of the patients had primary infertility and 36.47% had secondary infertility. The median duration of PCOS for all patients was 4.00 (3.00–6.00) years, and all patients received assisted reproductive technology, of which 61.41% received IVF and 38.59% received ICSI. 0, 1, 2 and 3 embryos were transferred in 33 (7.76%), 147 (34.59%), 243 (57.18%) and 2 (0.47%) patients, respectively. The median number of oocytes retrieved was 18.00 (12.00–24.00), the median number of embryos retrieved was 6.00 (3.00–10.00), the median optimal embryo rate was 0.39 (0.28–0.52), the median left ovarian volume and right ovarian volume were 11.40cm3 (9.00–14.50) and 12.90cm3 (10.20–16.20), respectively. The proportions of patients with levels of antral follicles of 1, 2, 3, 4 and 5 were 0.24%, 46.12%, 31.53%, 13.41% and 8.71%, respectively. The distribution of other clinical features is shown in table 1.

Table 1
•
Baseline characteristics

Univariable logistic regression analysed the factors associated with clinical pregnancy

The study used logistic regression to analyse predictors of clinical pregnancy after ET in PCOS infertile patients. Univariate logistic regression analysis found that LH level (OR=0.927, 95% CI 0.887 to 0.970; p=0.001), LH/FSH (OR=0.644, 95% CI 0.498 to 0.832; p=0.001), E2 level (OR=0.988, 95% CI 0.977 to 1.000; p=0.043), number of embryos transferred (OR=2.048, 95% CI 1.500 to 2.796; p<0.001), number of oocytes retrieved (OR=1.033, 95% CI 1.012 to 1.055; p=0.002), number of embryos (OR=1.088, 95% CI 1.043 to 1.136; p<0.001), optimal embryo rate (OR=7.243, 95% CI 2.622 to 20.005; p<0.001) and level of antral follicles (OR=1.387, 95% CI 1.121 to 1.717; p=0.003) were all significantly associated with clinical pregnancy of PCOS infertile patients after ET. The remaining variables were not significantly associated with clinical pregnancy in PCOS patients (table 2).

Table 2
•
Univariate logistic regression analysis of predictors related to clinical pregnancy

Multivariable logistic regression analysed the predictors of clinical pregnancy

After adjusting for confounding factors, multivariate logistic regression analysis found that LH/FSH (OR=0.577, 95% CI 0.422 to 0.789; p=0.001), free thyroxine (FT4) level (OR=1.107, 95% CI 1.009 to 1.214; p=0.031), number of embryos transferred (OR=1.781, 95% CI 1.238 to 2.562; p=0.002), number of oocytes retrieved (OR=1.071, 95% CI 1.016 to 1.130; p=0.011), optimal embryo rate (OR=21.347, 95% CI 2.647 to 172.131; p=0.004) and level of antral follicles (OR=1.444, 95% CI 1.131 to 1.845; p=0.003) were independent predictors of clinical pregnancy after ET in PCOS infertile patients (table 3).

Table 3
•
Multivariate logistic regression analysis of predictive factors for clinical pregnancy

Analysis of the predictive effect of predictive models on clinical pregnancy in PCOS infertile patients

Based on the variables with p<0.05 in multivariate logistic regression analysis, a predictive model for clinical pregnancy in infertile patients with PCOS, was constructed. The AUCs of LH/FSH, FT4 level, number of embryos transferred, number of oocytes retrieved, optimal embryo rate and level of antral follicles for predicting clinical pregnancy in PCOS infertile patients were 0.584 (95% CI 0.528 to 0.641), 0.564 (95% CI 0.508 to 0.619), 0.588 (95% CI 0.531 to 0.645), 0.585 (95% CI 0.529 to 0.641), 0.609 (95% CI 0.552 to 0.665) and 0.588 (95% CI 0.533 to 0.643). The combination of these risk factors was significantly better than any single variable in predicting clinical pregnancy, with an AUC of 0.732 (95% CI 0.683 to 0.782) (figure 1).

Figure 1
Request permissions

ROC curve of the predictors. AUC, area under the curve; FSH, follicle-stimulating hormone; LH, luteinising hormone; ROC, receiver operating characteristic curve; FT4, free thyroxine.

Nomograms predicting clinical pregnancy of PCOS infertile patients

Clinical pregnancy probability can be estimated with the nomograms (figure 2). In order to calculate the probability of clinical pregnancy of PCOS infertile patients, each parameter has a corresponding score on the point axis, and the sum of the scores is plotted on the ‘total point’ axis. The probability of clinical pregnancy is the value at a vertical line from corresponding total points.

Figure 2
Request permissions

Nomogram predicting clinical pregnancy. E2, oestradiol; FSH, follicle-stimulating hormone; LH, luteinising hormone; T, testosterone; FT4, free thyroxine.

Discussion

PCOS can affect the patient’s ovarian function, oocyte quality and even lead to infertility. In China, IVF-ET is often used as a treatment option for infertility patients with PCOS, but clinical pregnancy rates are low. Identifying the prognostic factors of clinical pregnancy in PCOS infertility patients can provide reference for improving the clinical pregnancy rate. The study found that LH/FSH was a risk factor for reduced likelihood of clinical pregnancy in PCOS infertile patients, while FT4 level, number of embryos transferred, number of oocytes retrieved, optimal embryo rate and level of antral follicles were significantly associated with increased likelihood of clinical pregnancy in PCOS infertile patients.13 Based on these variables, this study constructed a clinical pregnancy prediction model for PCOS infertile patients receiving assisted reproductive therapy. The results showed that the predictive performance of the model was better.

There are many factors that affect the clinical pregnancy rate after assisted reproductive treatment. FSH and LH are pituitary gonadotropins that are very important for fertility. Studies have shown that a high FSH: LH ratio on basal cycle day 3 can be used as an early biomarker of poor ovarian response.14 The baseline LH/FSH levels were a significant independent risk factor for live birth rates after fresh-ET in PCOS patients.15 For PCOS patients with day-3 LH/FSH ≥2, GnRH-agonist use during controlled ovarian hyperstimulation was significantly increased, with a non-significant increase in clinical pregnancy rate.16 The results of the above studies are inconsistent with the results of our study, in which the greater the LH/FSH ratio, the less likely the clinical pregnancy is in PCOS patients. The underlying reasons need to be explored in further clinical trials and may be related to the individual characteristics of PCOS patients.

In this study, FT4 level was an independent predictor of clinical pregnancy after assisted reproductive therapy in PCOS infertile patients. FT4 is essential for the normal development of the fetus. The main role of T4 is to regulate genes involved in myelination and neuronal/glial differentiation,17 and T4 is essential for the development and maintenance of normal physiological processes, especially the central nervous system, and contributes to fetal brain maturation throughout pregnancy.18 During pregnancy, the demand for maternal T4 production increases significantly. Total T4 concentrations must be increased by 20%–50% to maintain euthyroidism.19 The study found that total T4 levels in the second trimester were about 25% higher than those in the first trimester and 35% higher than those in non-pregnant women. However, there is still a lack of research on the correlation between serum FT4 levels and clinical pregnancy in PCOS patients after assisted reproductive therapy. We speculate that individual differences in FT4 levels among infertile patients with PCOS lead to differences in the extent of fetal development in these patients following assisted reproductive treatment, and ultimately affect clinical pregnancy rates. The specific reasons still need to be confirmed by clinical research.20

Our study found that the number of embryos transferred, the number of oocytes retrieved and optimal embryo rate were independent predictors of clinical pregnancy in PCOS infertile patients. The number and quality of embryos transferred will undoubtedly affect the final clinical pregnancy rate. Earlier studies found that the number of embryos transferred and the average morphological score of the transferred embryos had significant predictive value for clinical pregnancy after ET.21 The quantity and quality of embryos transferred in a fresh IVF cycle are important factors for obtaining a live birth.22 The total number of high-quality embryos has been shown to be the most important predictor of cumulative outcomes for patients after IVF.23 A clinical study found that the transfer of additional embryos resulted in increased pregnancy rates regardless of embryo quality.24 Several studies have shown that antral follicle count, number of oocytes, total number of embryos and total number of high-quality embryos are predictors of clinical pregnancy after ET, and the higher the number of these parameters, the greater the likelihood of clinical pregnancy.25 26 A clinical study found that the number of oocytes, fertilised eggs and cleavage embryos was associated with live birth rates after IVF.27 Previous studies have shown that the number of oocytes retrieved is positively associated with an increased chance of a live birth.28–30 However, it should be noted that the greater the number of oocytes retrieved, the greater the stimulation of the ovaries, and the higher the risk of ovarian hyperstimulation syndrome. How to determine the optimal number of oocytes retrieved while obtaining clinical pregnancy in infertile patients with PCOS will be one of our next research goals.

The results of this study suggest that level of antral follicles is also an independent predictor of clinical pregnancy in PCOS infertile patients. Antral follicle count, one of the ovarian reserve marker parameters, is widely used to predict ovarian response during IVF treatment in assisted reproduction programmes. Antral follicle counts can not only aid in setting the initial dose and schedule of ovarian stimulation, but also predict subsequent live births.31 A clinical study found antral follicle count was shown to predict clinical pregnancy and live birth in patients over 40 years of age undergoing IVF treatment.32 An earlier clinical study found a significant correlation between the number of oocytes retrieved and antral follicle size of 5–10 mm, and that patients with more antral follicles of 5–10 mm had a higher clinical pregnancy rate.33 Our study also found that the level of antral follicles was significantly associated with an increased likelihood of clinical pregnancy after assisted reproductive therapy in PCOS infertility patients. However, some researchers hold a different view. Sahmay et al found no significant association between antral follicle count and clinical pregnancy rates in patients under 40 years of age. They concluded that antral follicle count cannot predict clinical pregnancy in these patients.34 Further clinical studies are needed to confirm whether there is a correlation between antral follicles and clinical pregnancy in PCOS infertile patients after assisted reproductive therapy.

We developed nomograms to assist clinicians in their assessment of the likelihood of clinical pregnancy following assisted reproductive therapy in infertile PCOS patients in clinical practice. The nomogram is a simple and intuitive statistical prediction model that generates numerical probabilities of clinical pregnancy. By adding the scores assigned to each factor, the probability of clinical pregnancy following assisted reproductive therapy in infertile patients with PCOS can be calculated. At the same time, clinicians can increase the likelihood of a patient’s clinical pregnancy by adjusting the level of modifiable factors. If a PCOS patient has a low probability of clinical pregnancy, clinicians may increase the patient’s probability of clinical pregnancy by increasing or decreasing certain modifiable factors, such as the number of oocytes retrieved or the number of embryos transferred.

The findings of our study have significant clinical implications. By recognising these prognostic factors, clinicians can better counsel PCOS infertile patients regarding their chances of achieving clinical pregnancy. For instance, patients with a lower LH/FSH ratio, higher FT4 levels, a greater number of embryos transferred, more oocytes retrieved, a higher optimal embryo rate and a higher level of antral follicles can be informed that they have a relatively better prognosis for clinical pregnancy. This information can help patients make more informed decisions about their treatment options and set realistic expectations.

Conclusions

This retrospective single-centre study of 425 infertile PCOS patients undergoing IVF-ET identified six independent predictors of clinical pregnancy: LH/FSH ratio, FT4 level, number of embryos transferred, number of oocytes retrieved, optimal embryo rate and antral follicle level and constructed a predictive model with an AUC of 0.732, which is more effective in predicting clinical pregnancy than any single variable.

These findings address the gap in comprehensive analysis of clinical pregnancy predictors for PCOS patients undergoing IVF-ET in China, providing reliable indicators and practical tools (eg, nomogram) for clinicians to assess patients’ pregnancy potential, optimise individualised IVF-ET protocols, and thereby improve the success rate of assisted reproductive therapy for PCOS-related infertility.

This study is limited by its single-centre retrospective design, which restricts the extrapolation of results to other populations; future research should conduct multi-centre prospective studies to validate the generalisability of the predictive model and the predictive value of antral follicle level, and explore the optimal number of oocytes retrieved to balance clinical pregnancy rate and the risk of ovarian hyperstimulation syndrome in PCOS patients.