Sexual function in Chinese women with different clinical phenotypes of polycystic ovary syndrome

Abstract

Objective: To investigate sexual function stratified according to four clinical phenotypes of polycystic ovary syndrome (PCOS) and its association with clinical and quality of life parameters, and to compare these with healthy controls in Chinese women with PCOS. Methods: A cross-sectional study was designed in 1000 PCOS women and 500 control women aged 18–45 years. PCOS women were grouped into four clinical phenotypes according to the Rotterdam Criteria. FSFI (Female Sexual Function Index), SF-12 (the 12-item short form health survey) and clinical and hormonal characteristics likely to affect sexual function were determined. Results: 809 PCOS women and 385 control women with complete parameters were evaluated after screening. Phenotype A had a lower total FSFI mean score (23.14 ± 3.22) compared with phenotype D and control group (p < 0.05). The control group had the highest total FSFI mean score (24.98 ± 3.78). For the percentage at risk of sexual dysfunction, phenotype A (87.5%) and phenotype B (82.46%) had a higher risk of female sexual dysfunction (FSD) than that in phenotype C (75.34%), phenotype D (70.56%) and control group (61.30%) (p < 0.05). SF-12 mental domain scores were significantly lower in phenotypes A and B compared with phenotypes C and control group (p < 0.05). Infertility treatment, bioavailable testosterone, psychological factors, age and waist circumference presented negative correlation with female sexual function. Conclusions: The risk of FSD in PCOS women seemed to be associated with PCOS clinical phenotypes. The classical PCOS phenotype with oligo-ovulation and hyperandrogenism had a higher risk of sexual dysfunction.

Keywords:

• Sexual function
• quality of life
• polycystic ovary syndrome
• clinical phenotypes
• correlation

Introduction

Polycystic ovary syndrome (PCOS) is a complex and heterogeneous endocrine disorder, with an estimated prevalence of 10% around the world and 5.6% in Chinese women aged 19–45 years [1]. Besides irregular menses, symptoms of androgen excess, infertility and adverse pregnancy outcomes, PCOS is often associated with long-term health risks such as elevated risks of metabolic syndrome, obesity, type 2 diabetes and future cardiovascular diseases [2, 3]. The impact of these physical and physiological characteristics on quality of life and sexual function has been studied [4–6]. In our center, we performed the first study in China to investigate sexual function face to face in 685 patients with PCOS [7]. About 80% of PCOS women were at high risk of female sexual dysfunction (FSD), higher than the risk in the general population in China of around 60% [8–10], and also higher than reported in some other countries, where prevalence ranges between 27.2 and 62.5% [11–13]. The status of sexual function in Chinese PCOS women is therefore serious, and current opinions on this topic in other countries are controversial through meta-analysis in recent years [14–16].

Sexual function is found to be associated with a wide range of factors, including cultural, medical, biological, psychological and relationship factors [17, 18]. Patients with PCOS may have multiple factors that can impair sexual function, such as hyperandrogenism (HA) and clinical signs of HA [19], insulin resistance (IR) and obesity [20–22], irregular menses and infertility [23, 24] and even unhealthy psychological factors [25, 26]. However, heterogeneity is an intrinsic characteristic of PCOS, from its definition to its phenotype. The spectrum is broad and extends from women with evident signs of hyperandrogenism and amenorrhea to those who do not present with hyperandrogenism or present with regular cycles. The relative contributions of insulin resistance, androgen excess and other factors, such as obesity, to the development of PCOS in the different PCOS phenotypes are also heterogeneous. In terms of diagnosis, treatment and long-term management, it has recently been argued that not all women with PCOS should be considered equal [27, 28]. As far as sexual function in PCOS women is concerned, the inconsistent opinions may possibly be explained by typing analysis in different phenotypes.

There is limited evidence comparing the sexual function of PCOS phenotypes. Therefore, the aim of this study was to investigate the sexual function status stratified according to four clinical phenotypes of PCOS, and its association with clinical and quality of life parameters, and to compare these with healthy controls in Chinese women with PCOS.

Materials and methods

Ethical approval

The study was approved by the Ethics Committee and Institutional Review Board of Beijing Obstetrics and Gynecology Hospital, Capital Medical University, People’s Republic of China (protocol no. 2019-KY-039-01).

Recruitment of the study population

This cross-sectional study was performed in 1000 consecutive women with PCOS, and 500 control women aged 18–45 years, who attended the Department of Gynaecological Endocrinology, Beijing Obstetrics and Gynecology Hospital, from 31 provinces of China during January 2021 to December 2021. All women coming during this defined recruitment time were included and were also willing to answer sensitive questions about sexual function during a direct face to face interview with the doctor or medical staff to confirm correct understanding of the questions. Informed consent was obtained from all patients. All study data was kept confidential.

Exclusion criteria were a history of neurologic and psychiatric disease, severe medical illnesses (hyperprolactinemia, thyroid dysfunction, adrenal dysfunction, diabetes mellitus, malignant diseases), pregnancy, previous history of ovarian surgery and medication that could interfere with the function of the hypothalamic-pituitary-gonadal axis within 3 months before entering the study, such as using oral contraceptives or ovulation induction drugs. Inclusion criteria were sexual activity and have stable relationship during the past 4 weeks, with normal desire in the partner, readiness to answer the questionnaires on sexual function and the 12-item short form health survey (SF-12) and other items with possible links to sexual life. PCOS women were grouped into four clinical phenotypes according to the Rotterdam Criteria (2004), A: oligo-ovulation or anovulation (OA) + hyperandrogenism (HA) + polycystic ovaries (PCO); B: OA + HA; C: HA + PCO; D: OA + PCO. Polycystic ovary was diagnosed by the presence of 12 or more 2–9 mm follicles in each ovary and/or increased ovarian volume (> 10 cm³). Oligo-anovulation was defined as oligomenorrhea (menstrual cycle of > 35 days) or amenorrhea (i.e. no menstrual bleeding during the last 3 months). Hyperandrogenism was clinically defined as the presence of clinical signs of hirsutism (modified Ferriman and Gallwey score > = 3), obvious acne (the number of comedones (>10) and/or inflammatory papules/pustules as well as nodules/cysts (>10) spread on the face, back and chest), and/or levels of total testosterone (TT) higher than 37.14 ng/dl (1.21 nmol/l), and/or levels of bioavailable testosterone (Bio-T) higher than 206 pg/ml, and/or levels of dehydroepiandrosterone sulfate (DHEAS) higher than 2500 ng/ml.

Assessment of sexual function and quality of life

The Chinese version of FSFI was used to assess female sexual function [29]. Individual subdomain scores were obtained by weighting as original score times severity score. A total score of less than 26.55 was defined as FSD [30]. According to the Diagnostic and Statistical Manual of mental disorders and to the main scientific societies, the concept of a clinically significant ‘distress’ related to sexual symptoms is mandatory for the definition of FSD. The present study failed to evaluate this parameter and therefore we can only talk about a pathological FSFI score and a risk for sexual dysfunction. The Chinese SF-12 includes 12 questions, which is divided into two subscales of mental (mental component summary, MCS) and physical health (physical component summary, PCS). The greatest score obtained for each section or subscale is 100 and the least score is zero with a higher score indicating a better health status [31].

Assay methods

Baseline blood samples (in early follicular phase and/or if amenorrhoeic) were collected after overnight fasting for at least 12 h. Serum follicle-stimulating hormone (FSH) and luteinizing hormone (LH) were determined by the ADVIA Centaur XP Chemiluminescence Analyzer (Siemens Healthcare Diagnostics Inc.). Serum TT, DHEAS and 17-hydroxyprogesterone (17-OHP) were quantified with an in-house developed LC-MS/MS method that was established and validated previously [32]. Bio-T using the parameters of total testosterone, sex-hormone-binding globulin (SHBG) and albumin (online calculator website: http://www.issam.ch/freetesto.htm) [33]. Biochemical parameters including fasting blood glucose and fasting insulin were measured using a chemistry analyzer (Beckman Coulter LX20, USA) in the clinical chemistry laboratory. The homeostasis model assessment of insulin resistance (HOMA-IR) was calculated according to the suggested formula: [glucose (nmol/L) × insulin (μU/mL)/22.5].

Patient sample calculation

This is an observational cross-sectional study. Based on a single population proportion formula, the minimum sample formula was calculated: N = u_α² × prevalence (1 − prevalence)/δ². The prevalence of FSD was estimated as 60.2% according to our earlier internet-based study in young Chinese women [8]; the evaluated error δ was 5%; the minimum sample content was n = (1.96² × 0.602 × 0.398)/0.05² = 368. For our study, we were primarily able to recruit almost three times the number of PCOS patients during our defined recruitment time of 12 months (intent to treat analysis), and more than twice the number of PCOS patients who could finally be analyzed according to the study protocol (per protocol analysis). Enough samples were analyzed in our study for controls.

Statistics

Statistical analyses were performed using SPSS 23.0 software. Continuous variables were verified for normality by the Kolmogorov-Smirnov test. Values were described as mean ± standard deviation (SD) and percent for categorical variables. The independent samples t-test and Mann-Whitney U test were used for normally and non-normally distributed quantitative variables, respectively. Chi-square (χ2) and Fisher’s exact tests were used to compare categorical variables between the groups. One-way ANOVA was used for the normal variables and Kruskal-Wallis test was used for the nonnormal variables among four PCOS clinical phenotype groups and control group. Post-hoc analysis with the Mann-Whitney U test and Dunn Bonferroni correction was applied to variables whose Kruskal-Wallis test results were found to be significant. Pearson and Spearman’s correlation coefficients were obtained to find the association between total and domain FSFI scores and clinical variables and SF-12 scores. Multivariate linear regression analysis was used to estimate risk factors for FSD and to control for confounding factors. Two-sided p values <0.05 were considered to be significant.

Results

Clinical characteristics

A total of 910 PCOS women and 468 control women participated in the study, and 809 PCOS women and 385 control women with complete parameters were suitable to evaluate the sexual function status and their correlations with clinical characteristics and SF-12 scores after screening. Phenotype distributions in PCOS women were phenotype A in 408 (50.43%), B in 114 (14.09%), C in 73 (9.02%) and D in 214 (26.45%). The mean age of PCOS women and control women was 27–28 years, and there was no significant difference among different PCOS phenotype groups and control group. The details of clinical characteristics are provided in Table 1.

Table 1. Clinical characteristics in patients with four PCOS phenotypes and control women.

	Control	Phenotype A	Phenotype B	Phenotype C	Phenotype D	P value
N (%)	385	408 (50.43)	114 (14.09)	73 (9.02)	214 (26.45)
Age (years)	27.56 ± 3.83	26.89 ± 3.95	27.47 ± 3.88	28.17 ± 4.38	27.87 ± 3.91	0.181
BMI (kg/m^2)	23.35 ± 3.53^a^bc	26.78 ± 3.89^d	25.77 ± 4.42^d	25.06 ± 3.53	23.24 ± 3.70	<0.01
WC (cm)	79.10 ± 10.47^a^b	87.45 ± 12.44^c^d	86.09 ± 10.81^d	83.86 ± 10.62	83.39 ± 11.23	<0.001
F-G score	1.23 ± 0.51^a^bcd	3.15 ± 1.42 ^c^d	3.15 ± 1.38^d	2.75 ± 1.46	1.77 ± 0.77	<0.05
LH/FSH	1.11 ± 0.16^a^bcd	2.24 ± 0.48^d	2.18 ± 0.88^d	2.19 ± 0.61^d	1.95 ± 0.44	<0.001
TT (nmol/L)	0.92 ± 0.14^a^bcd	2.62 ± 0.49^b^cd	2.47 ± 0.59^c^d	2.09 ± 0.59^d	1.39 ± 0.41	<0.01
SHBG (nmol/L)	57.89 ± 42.67^a^bcd	30.21 ± 26.22^d	30.34 ± 25.79^d	34.22 ± 29.03^d	47.65 ± 34.02	<0.01
Bio-T (pg/ml)	73.45 ± 52.12^a^bcd	186.49 ± 111.19^c^d	179.42 ± 102.64^c^d	143.66 ± 107.43	142.92 ± 78.11	<0.001
DHEAS (ng/ml)	1201.33 ± 573.19^a^bcd	2792.57 ± 899.65^c^d	2682.11 ± 935.32^c^d	2200.12 ± 877.68	2039.71 ± 745.41	<0.05
17-OHP (ng/ml)	0.74 ± 0.55	0.92 ± 0.62	0.89 ± 0.47	0.72 ± 0.49	0.81 ± 0.51	0.271
HOMA-IR	1.46 ± 0.64^a^bc	2.61 ± 0.78^c^d	2.52 ± 0.92^c^d	2.17 ± 0.61	1.93 ± 0.55	<0.001
Infertility [n (%)]^#	0 (0)	346 (84.80)^c^d	94 (82.46)^c^d	43 (58.90)	134 (62.62)	<0.001
Infertility treatment  [n (%)]^#	0 (0)	179 (43.87)^c^d	48 (42.11)^c^d	11 (15.07)	31 (14.49)	<0.001

PCOS: polycystic ovary syndrome; BMI, body mass index; WC, waist circumference; F-G score, Ferriman and Gallwey score; LH/FSH, luteinizing hormone/follicle-stimulating hormone; TT, total testosterone; SHBG, sex hormone-binding globulin; Bio-T, bioavailable testosterone; DHEAS, dehydroepiandrosterone sulfate; 17-OHP, 17-hydroxyprogesterone; HOMA-IR, homeostasis model assessment for insulin resistance index;.

P values refer to the Kruskal-Wallis test.

^# Values are given as number (%) by using Chi-square test.

^a P < 0.05 compared with phenotype A;

^b P < 0.05 compared with phenotype B;

^c P < 0.05 compared with phenotype C;

^d P < 0.05 compared with phenotype D.

SF-12 scores

In terms of SF-12, the mean scores in MCS domain were significantly lower in phenotypes A and B compared with phenotypes C and control group (p < 0.05), but no significant difference was observed in PCS domain (p = 0.057). The details of MCS and PCS scores are provided in Table 2.

Table 2. SF-12 scores in patients with four PCOS phenotypes and control women.

	Control	Phenotype A	Phenotype B	Phenotype C	Phenotype D	P value
N (%)	385	408 (50.43)	114 (14.09)	73 (9.02)	214 (26.45)
MCS	54.32 ± 19.82^a^b	47.67 ± 16.54^c	49.33 ± 17.21^c	53.07 ± 16.34	52.15 ± 20.18	<0.05
PCS	51.17 ± 21.38	46.49 ± 17.66	46.51 ± 15.49	49.22 ± 16.72	49.53 ± 18.81	0.057

SF-12, the 12-item short form health survey; PCOS: polycystic ovary syndrome; MCS: mental component summary; PCS: physical component summary.

P values refer to the Kruskal-Wallis test.

^a P < 0.05 compared with phenotype A;

^b P < 0.05 compared with phenotype B;

^c P < 0.05 compared with phenotype C.

FSFI scores and percentage at risk of sexual dysfunction

The FSFI scores and percentage at risk of sexual dysfunction in patients with four PCOS phenotypes and control women are given in Table 3. Phenotype A had lower total FSFI mean score (23.14 ± 3.22) compared with phenotype D and control group (p < 0.05). The control group had the highest total FSFI mean score (24.98 ± 3.78). For the percentage at risk of sexual dysfunction, phenotype A (87.5%) and phenotype B (82.46%) had a higher risk of FSD than that in phenotype C (75.34%), phenotype D (70.56%) and the control group (61.30%) (p < 0.05). For all PCOS women, the mean total FSFI score was (24.15 ± 2.44) and percentage at risk of sexual dysfunction was 81.11%. For the FSFI subscores in six domains, there were significant differences between the PCOS subgroups and the control group in terms of desire, lubrication, orgasm and satisfaction, but no significant difference was observed in arousal score and pain score (all p > 0.05).

Table 3. FSFI scores and proportion of participants with FSFI score < 26.55 in patients with four PCOS phenotypes and control women (n = 809).

	Control	Phenotype A	Phenotype B	Phenotype C	Phenotype D	P value
N (%)	385	408 (50.43)	114 (14.09)	73 (9.02)	214 (26.45)
Total FSFI Score	24.98 ± 3.78^a^b	23.14 ± 3.22^d	24.15 ± 1.97	24.33 ± 2.33	24.47 ± 2.57	<0.01
Desire	3.81 ± 0.85^a^b	3.21 ± 0.91^c^d	3.56 ± 0.79^d	3.68 ± 0.74	3.72 ± 0.72	<0.05
Arousal	3.99 ± 1.02	3.80 ± 0.93	3.84 ± 0.86	3.88 ± 0.81	3.79 ± 0.92	0.292
Lubrication	4.58 ± 1.17^a^c	4.38 ± 1.21^b^cd	4.57 ± 0.93	4.51 ± 1.08	4.56 ± 1.15	<0.05
Orgasm	4.03 ± 0.93^a	3.70 ± 0.74^d	3.85 ± 0.89^d	3.92 ± 0.92	3.96 ± 0.86	<0.05
Satisfaction	4.54 ± 1.27^a^bcd	4.00 ± 1.02^d	4.35 ± 1.15	4.40 ± 1.12	4.47 ± 1.08	<0.001
Pain	4.01 ± 0.89	3.98 ± 1.13	3.98 ± 0.87	3.96 ± 0.93	3.97 ± 1.04	0.147
Number at risk of sexual dysfunction	236	357	94	55	151
Percentage at risk of sexual dysfunction ^#	61.30%^a^bc	87.5%^c^d	82.46%^c^d	75.34%	70.56%	<0.05

FSFI: Female Sexual Function Index. PCOS: polycystic ovary syndrome.

P values refer to ANOVA.

^# Values are given as number (%) by using Chi-square test..

^a P < 0.05 compared with phenotype A;

^b P < 0.05 compared with phenotype B;

^c P < 0.05 compared with phenotype C;

^d P < 0.05 compared with phenotype D.

Multivariate linear regression analysis

Spearman’s correlation analysis was conducted to assess the correlation of total FSFI scores and its domains with clinical characteristics and SF-12 scores. FSFI scores were negatively correlated with age, BMI, WC, Ferriman and Gallwey scores (F-G scores), LH/FSH, TT, Bio-T, DHEAS, HOMA-IR, infertility and infertility treatment, and positively correlated with SHBG, MCS and PCS. There were no significant associations between FSFI scores and 17-OHP. For total FSFI scores, multivariate linear regression analysis demonstrated that infertility treatment, Bio-T, age and WC were risk factors for FSD. Meanwhile, high MCS scores seemed to be a protective factor for FSD, positively correlated with total FSFI scores (Table 4).

Table 4. Risk factors for FSD analyzed by multivariate linear regression analysis.

Variable	Standardized coefficients	Standard error	P value
Infertility treatment	−0.387	0.18	0.007
Bio-T	−0.27	0.023	<0.001
MCS	0.057	0.01	<0.001
Age	−0.046	0.013	0.000
WC	−0.001	0	<0.001

FSD, female sexual dysfunction; Bio-T, bioavailable testosterone; MCS: mental component summary; WC, waist circumference.

Discussion

In this study, we analyzed the current sexual function of women with PCOS according to the four clinical phenotypes. In accordance with previous reports [34–36], phenotype A was the most common (50.39% of the PCOS diagnoses) in our study population, whereas phenotype B resulted in less frequent diagnoses in 14.12%. The normo-androgenic phenotype D accounted for 26.47%, which is different from results observed in some other studies [37, 38]. There is no dispute that phenotype A as the severe group is the most prevalent phenotype, although the relative prevalence of each phenotype varies in different ethnic groups. The disagreement is mainly concentrated on the normo-androgenic phenotype D, considered the ‘mild’ form of PCOS, due to the subjectivity involved in evaluating clinical hyperandrogenism with great interobserver variability.

In the present study, 61.3% of control women and 80% of PCOS patients were at high risk of FSD based on the total FSFI score, which is comparable with results observed in our previous studies in China [7–10] and higher than reported in some other studies [11–13]. Loh et al. examined 28 observational studies comprising a total of 6256 women with PCOS [14]. The prevalence of FSD was 45% among women with PCOS and 29.6% among those without PCOS. The probability of FSD was 1.32-fold higher in women with PCOS [14]. The healthy control group showed a mean FSFI score in the pathological range (24.98) and the high prevalence of FSD in Chinese women may be due in part to different demographic characteristics of the women and their social and cultural background, such as the conservative beliefs and neglect of FSD in China. For PCOS women, it may also be related to the internal characteristics of PCOS. In this study, phenotype A and B had higher risk of FSD than that in phenotypes C, D and the control group. For subscores, desire, lubrication, orgasm and satisfaction scores were lower in phenotype A than in phenotype D and control group. Desire and satisfaction scores were lower in phenotype B than in the control group. The status of FSD in PCOS women seems to be severe and associated with PCOS phenotypes.

There may be several reasons for these differences in risk of FSD in different phenotypes. Here, we focus on phenotype A and B group, with lower total scores and higher risks of FSD, occurring in the context of the numerous clinical features in patients with PCOS. Firstly, phenotype A and phenotype B are also called the ‘classical phenotype’ of PCOS with HA and OA, often with high severity. Severe obesity and metabolic syndrome [20–22], frequent menstrual irregularities [23, 24] and severe negative clinical manifestations of hyperandrogenism [19, 39] are all associated with higher risks of FSD. Our correlation analysis observed that FSFI scores were negatively correlated with BMI, WC, F-G scores, TT, Bio-T, DHEAS and HOMA-IR. In this study, most of these negative parameters were more serious in the phenotype A, B group compared with phenotype C, D and the control group.

Multivariate linear regression analysis demonstrated that Bio-T and WC were risk factors for FSD, and we also observed significantly higher levels of Bio-T and WC in the ‘classical phenotype’ of PCOS. The association between levels androgen and sexual function remain inconsistent. Ercan et al. [40] reported a significant negative correlation between the scores of total FSFI and total testosterone and free testosterone. Veras et al. [41] reported a negative collection between sexual function and total testosterone levels and DHEAS. However, Mansson et al. and Stovall et al. [11, 42] showed an inverse relationship. Maseroli et al. [43] reported a moderate association between total T and a better sexual desire/global sexual function, although weak. Similar results on desire were obtained for free testosterone and FAI. Rellini et al. [44] reported clinical signs of hyperandrogenism are associated with sexual clues driving desire for sex more than biochemical parameters of hyperandrogenism. The real mechanism of androgen on sexuality is complicated. Paick et al. [45] reported ‘Mirodenafil’, in doses of 50 or 100 mg, significantly improved erectile function, inferring androgens may play a role in genital vascularization. Whilst some patients with hypoactive sexual desire disorder (HSDD) can be treated with testosterone, an explanation for this obvious paradox may be that PCOS women with higher bio-T tend to have a negative self-image, obesity and psychosocial and emotional disorders [40]. A large WC puts women at a higher risk of cardiometabolic diseases that in turn are linked with impaired sexual activity and type 2 diabetes, as frequently seen in ‘classical phenotype’ PCOS women [20–22, 46], and also a higher risk of endocrine hormone disorders associated with FSD [22].

In this study, we observed a significantly higher risk of FSD in our largest group of patients undergoing infertility treatment. More infertility and infertility treatments were seen in phenotype A and B group compared with phenotype C and D group. This may be a key factor contributing to the higher risk of FSD in the ‘classical phenotype’ group. Several studies confirmed that infertile women had problems with sexual function [47–49]. For women diagnosed with infertility, sexual intercourse may lose its spontaneity and erotic value because the main aim becomes conception. More seriously, the treatment of infertility dictates the frequency and timing of sexual intercourse. The usual intimate event becomes regulated and controlled and medical factors may present during the personal act [50]. This may affect the ability for intimate sexuality and can provoke sexual dysfunction [51, 52]. Many women describe the period of diagnosis and treatment of infertility as the most stressful period of their life [51]. This is in line with several other studies, where psychosexual implications were found to cause profound emotional distress which could further aggravate sexual dysfunction [23–25].

In the present study, high mental health scores seemed to be a protective factor for FSD analyzed by multiple regression analysis and positively correlated with total FSFI scores. However, significantly lower SF-12 MCS scores were determined in the phenotype A, B groups than in the healthy controls. According to our results, women in phenotype A and B groups may experience emotional conditions such as depression, anxiety and low self-esteem that are known causative factors of sexual dysfunction. Recently, Seher et al. found that depression and body image scores were higher in the PCOS phenotype A and B group compared with the control group [53]. Using the SF-36 scale, Bazarganipour et al. [26] reported that the psychological dimension was more affected than the physical domain, and psychological impairment was higher in patients with HA and menstrual irregularities than in the two other phenotypes. This finding is consistent with the results of our study. Furthermore, although the age range of our participants was narrow, increasing age was still found to be associated with a high risk of FSD, similar to results in other studies [8, 54].

Strengths and limitations of the study

The strengths of our study are that we report, to our knowledge, the first large-scale face to face study of sexual function status stratified according to the four main phenotypes to minimize variability and confounding factors in Chinese women with PCOS. It is also a strength that all women were patients from one center where data was collected, and all clinical examinations were performed. Our center is the first officially acknowledged ‘Department of Gynaecological Endocrinology’ in China, and the center with the longest and greatest experience in treating PCOS. Because the women came from all over the country, our results should be representative for China. The main limitation is selection bias. Despite the statistical significance of the differences among the different phenotypes and controls, the mean FSFI is quite similar, with a difference of approximately 2 points. The clinical implication on the findings is limited. It would be desirable to repeat our analysis within an even larger community population, including also data of information on education level and income level, to obtain sufficient patient samples in subgroups of PCOS. In addition, the status of mental health of PCOS women, distress, confidence and self-esteem, also are important factors of sexual function, which needs specialized research. The present study failed to evaluate this parameter of ‘distress’ related to sexual symptoms for the definition of FSD. we can only talk about a risk for sexual dysfunction. However, we think that our study is a large-scale study, presently is unique for China and certainly will stimulate us for further research.

Conclusion

The risk of FSD in PCOS women is associated with PCOS clinical phenotypes. The ‘classical phenotype’ of PCOS with HA and OA had significantly higher risk of sexual dysfunction than in the healthy control. Infertility treatment, Bio-T, psychological factors, age and WC presented negative correlations with female sexual function. Hence, clinicians should consider sexual dysfunction as well as clinical and psychological factors with regard to possible effects on sex life in the individual management of PCOS.

Author contributions

XT: conceptualization: equal, data curation: lead, writing original draft: lead; XR: conceptualization: lead, data curation: lead, funding acquisition: lead, project administration: lead, supervision: lead, writing-review and editing: lead; JD: data curation: equal; JC: data curation: equal; RJ: data curation: equal; AOM: conceptualization: equal, methodology: supporting, supervision: equal, writing-review and editing: lead.

Consent to participate

Informed consent was obtained from all individual participants included in the study.

Consent to publish

Not applicable.

Ethics approval

This study was performed in line with the principles of the Declaration of Helsinki. Approval was granted by the Ethics Committee and Institutional Review Board of Beijing Obstetrics and Gynecology Hospital, Capital Medical University, People’s Republic of China (protocol no. 2019-KY-039-01).

Acknowledgments

The authors especially thank Prof. Xingming Li of the Capital Medical University (Beijing, China) for his assistance in statistical analysis of the data.

Data availability statement

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

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

Supported by the Beijing Municipal Administration of Hospitals’ Ascent Plan (No. DFL20181401).