Sexual dysfunction in polycystic ovary syndrome: a systematic review and meta-analysis

Polycystic ovarian syndrome (PCOS) is a common disorder characterized by clinical or biochemical hyperandrogenism and ovulary dysfunction. Female sexual dysfunction (FSD) adversely affects quality of life and interpersonal relationships. We aimed to compare the prevalence of FSD in women with and without PCOS. We pooled data from 28 observational studies involving 6256 women. Apart from the total prevalence of FSD, subgroup analyses based on different PCOS diagnostic criteria and obesity status (body mass index [BMI] ≥ 25 kg/m2) were performed. The differences in total and subscale scores of the Female Sexual Function Index (FSFI) among women with and without PCOS were also compared. Women with PCOS were younger (mean ± SD 28.56 ± 3.0 vs 31.5 ± 3.2 years, p < 0.001) with higher BMI (28.5 ± 4.2 vs 27.0 ± 6.1 kg/m2, p < 0.001), Ferriman-Gallwey score (10.0 ± 3.2 vs 4.0 ± 2.1, p < 0.001), and serum total testosterone level (2.34 ± 0.58 nmol/L vs 1.57 ± 0.60 nmol/L, p < 0.001) compared with women without PCOS. The prevalence of FSD among women with and without PCOS was 35% and 29.6%, respectively. There was no significant difference in total FSFI score (24.59 ± 3.97 vs 26.04 ± 3.05, p = 0.237) between the two groups. Women with PCOS, however, had significantly lower scores in the pain (p < 0.001) and satisfaction subscales (p = 0.010) compared with women without PCOS. Women with PCOS had 1.32 higher odds (95% CI 1.07, 1.61) of having FSD than women without PCOS. Women with PCOS have a higher risk of FSD than those without PCOS. Although total FSFI scores were not significantly different, women with PCOS tended to report dyspareunia and lack of sexual satisfaction.


Introduction
Polycystic ovary syndrome (PCOS) is a common disorder affecting women of reproductive age. Its prevalence differs according to ethnicity and diagnostic criteria, which ranges from 5.5% among Caucasians, using the 1990 National Institute of Health (NIH) criteria, to 16% among women in the Middle East, using the 2003 Rotterdam criteria [1]. PCOS is characterized by clinical or biochemical hyperandrogenism, such as hirsutism, acne, alopecia and seborrhea, and ovulary dysfunction with or without polycystic ovaries [2]. Notably, PCOS has been increasingly recognized as a complex illness which not only has long-term health consequences, such as insulin resistance, cardiovascular disease, and endometrial cancer, but also has a psychosocial impact related to changes in physical appearance [3,4]. Apart from emotional disturbances, other important aspects of psychological well-being, namely, sexual function and satisfaction, in women with PCOS are often overlooked [5,6].
Female sexual dysfunction (FSD) is very common, with a prevalence of 21-28% among premenopausal women [7]. The International Classification of Diseases (ICD)-11 defines FSD as a frequent and persistent disorder involving difficulty in experiencing personally satisfying, non-coercive sexual activities, leading to clinically significant distress [8]. The Sexual Function Health Council of the American Foundation for Urologic Disease (AFUD) has classified FSD into five main categories, namely, hypoactive sexual desire disorder, sexual aversion disorder, sexual arousal disorder, orgasmic disorder, and sexual pain disorders [9].
In women with PCOS, androgen excess and psychosocial changes may affect their sexual functions, although the existing data are as yet too few to draw definitive conclusions [10]. Women with total testosterone levels more than one standard deviation (SD) above the mean had significantly better sexual functioning compared with those within one SD and those with more than one SD below the mean [11]. Similarly, serum total testosterone levels correlated positively with sexual function in two other studies [12,13], but negatively in another [14]. Women with PCOS and hirsutism were found to have low self-esteem, and their self-confidence was greatly affected by the presence of facial hair [15], although this association was not found in another study [16]. The higher body mass index (BMI) seen in women with PCOS also unfavorably affects self-esteem and body satisfaction while leading to fear of negative appearance [15,17,18].
In the present meta-analysis, we aimed to compare the prevalence of FSD in women with and without PCOS. We also did separate analyses for subgroups diagnosed with PCOS using the Rotterdam criteria and for those with a BMI ≥ 25 kg/m 2 [19].

Data sources and extraction
We performed a systematic search of all English-language medical literature published from inception until January 2020 in the PubMed, CINAHL, and Medline databases using the following Medical Subject Headings: ((sexual dysfunction) OR (sexual dysfunctions) OR (sexual disorder) OR (sexual function)) AND ((polycystic ovary) OR (polycystic ovarian syndrome) OR (polycystic ovary syndrome)). We also looked through the references of the selected articles. When the articles were not available or the cohort study data were inadequate, we attempted to contact the respective authors via email to obtain the full articles and detailed data. To be included in the analysis, studies had to fulfill the following criteria: (i) diagnosis of PCOS based on the Rotterdam criteria, NIH criteria, or the Androgen Excess and PCOS Society criteria, (ii) use of a validated questionnaire or visual analog scales, and (iii) reported either proportion of FSD among PCOS or scores of the validated questionnaire. Studies were excluded if the subjects recruited had other concurrent medical illnesses or diseases or reported only on health-related quality of life, mental health, or idiopathic hyperandrogenism. Review articles, posters, abstracts, and theses were also excluded. Two independent reviewers (HHL and HSL) screened the titles and abstracts obtained through the electronic search and analyzed the full-text articles. All duplicates were removed. Data such as study sites, study types, sample size, PCOS diagnostic criteria, type of questionnaire, baseline characteristics, Ferriman-Gallwey score, total testosterone, Female Sexual Function Index (FSFI) scores, and proportion of FSD were extracted using a standardized data extraction form. Any discrepancies were resolved by the third reviewer (LLL). Wherever data were not provided numerically, they would be read off graphs. Data from eligible studies were extracted by HHL and all extracted data were reviewed by AY and SK.
Structured interviews and self-reported questionnaires, such as the FSFI, have been widely used to assess FSD. FSFI is a validated 19-item self-report scale that evaluates sexual function of women within a 4-week window, based on six subscales namely sexual desire, arousal, lubrication, orgasm, satisfaction, and degree of pain. The maximum score of each subscale is 6 and the maximum total score is 36. A score of < 26 defines FSD [20]. In this meta-analysis, total and subscale scores of FSFI among women with and without PCOS were compared.

Quality assessment
HHL and AY independently assessed the quality of the methodology and reporting of all studies using the Newcastle-Ottawa Scoring (NOS) Scale for Case-Control Studies or the NOS Scale adapted for Cross-Sectional Studies, as appropriate. Any discrepancies were discussed with the third reviewer (LLL) until agreement was reached. The NOS scale was developed to assess the quality of non-randomized case-control studies for the interpretation of meta-analysis results. Based on a "star-graded" system, each study is assessed in three broad categories, namely, group selection, group comparability, and ascertainment of the outcome of interest (exposure). In the original NOS Scale for Case-Control Studies, each study can be awarded a maximum of one star for every numbered item (four in the "selection" category and three in the "exposure" category) and a maximum of two stars in the "comparability" category. In the NOS Scale adapted for crosssectional studies, each study can be awarded a maximum of five stars in the "selection" category, two stars for the "comparability" category, and three stars for the "outcome" category. Both scales have a maximum score of 10.

Statistical analysis
Qualitative All abstracted information was tabulated. A qualitative metaanalysis was conducted to summarize, compare, and contrast the abstracted data.

Quantitative
All data analyses were performed using Stats Direct (version 2.7.9). The presence of heterogeneity between the trials was tested using the I-squared (I 2 ) statistic. An I 2 of more than 40% indicated significant heterogeneity. If the I 2 was significant, we pooled the data by using random-effects (DerSimonian-Laird). Conversely, we pooled the data by using fixed-effects (Mantel-Haenszel, Rothman Boice). We also assessed publication bias with the Begg-Mazumdar and Egger tests. For dichotomous outcomes, we estimated the odds ratio (OR) with 95% confidence intervals (CI) using the fixeed-effects model.For continuous outcomes, we estimated the weighted mean difference in FSFI score with 95% confidence interval (CI) if the mean and standard deviation (SD) of the outcomes were presented in the original articles.

Results
The initial search identified a total of 2135 articles: 302 from PubMed, 299 from CINAHL, and 2097 from MEDLINE. After the screening of titles and abstracts and the removal of duplicate publications and screening of full-texts, we included 28 full-text articles in the present systematic review and metaanalysis ( Fig. 1).

FSFI scores
Fourteen cohort studies reported total FSFI scores for women with PCOS [5,6,14,24,27,30,34,[36][37][38][39][40], while eight cohort studies were compared with the control group [5,14,24,27,30,34]. Only two studies showed significant differences in the total FSFI scores between women with PCOS and the control group consisting of healthy women [27,30]. Six studies reported the scores of each FSFI subscale [5,14,27,30,33,34]. For the orgasm subscale, none of the studies reported a significant difference between women with and without PCOS. For the desire subscale, one study showed a significantly lower score among women with PCOS [30], whereas another study showed a significantly higher score compared with women without PCOS [5]. For the pain subscale, one study showed higher scores among women with PCOS [27], whereas the other two studies showed a lower score when compared with women without PCOS [5,30]. Women with PCOS scored significantly lower for lubrication, arousal, orgasm, and satisfaction subscales compared with women without PCOS [27,30] .

Meta-analysis
In the pooled cohort, there was no significant difference in total the FSFI score (24.59 ± 3.97 vs 26.04 ± 3.05, p = 0.237) in women with and without PCOS (Fig. 2 When we sub-analyzed studies using the Rotterdam criteria for diagnosis of PCOS and reporting total FSFI scores, there was also no significant difference between the two groups (p = 0.237) (Fig. 3). Similarly, there was no significant difference when we included only studies of PCOS women with BMI ≥ 25 kg/m 2 (p = 0.483) (Fig. 4).

Other sexual questionnaire scoring
Eleven studies utilized other forms of validated sexual questionnaire to ascertain presence of FSD among women with PCOS compared with healthy women [11-13, 21-23, 25, 26, 28, 29, 31, 32]. Five studies found no significant difference between the two groups [12,13,23,25,31]. The detailed results of these studies are shown in the Supplementary Table.

Proportion of FSD
Sixteen studies reported the proportion of sexual dysfunction among women with PCOS (Table 1), which ranged between 4 and 69.7%. In the fixed-effects model, women with PCOS had 1.32 higher odds (95% CI 1.07, 1.61; p = 0.010) of having FSD than women without PCOS (Fig. 5). These significant results were not altered when we included only studies using the Rotterdam criteria for diagnosis of PCOS (Fig. 6) as well as the cohort with BMI ≥ 25 kg/m 2 (Fig. 7).

Discussion
In the present meta-analysis of 28 observational studies, we aimed to evaluate the prevalence of FSD among women with PCOS. Our results showed that women with PCOS have a more than 30% higher risk of having FSD than women without PCOS. Although total FSFI scores were not significantly different, women with PCOS had lower scores in pain and satistifaction subscales than those without PCOS. This suggests that although FSFI is one of the widely used measures to assess various domains of sexual function, it has marked  There seemed to be overlap between sexual arousal and desire which could be attributed to the inappropriate definition of desire used in FSFI [43]. The results could also be biased by the limited number of studies reporting these scores.
One-third of women with PCOS have FSD, suggesting that this coexistence is more common than expected and should not be overlooked in routine clinical practice. The present meta-analysis is the largest to date, and our results add to two previous meta-analyses which included only 18 and 10 observational studies, respectively [44,45]. In contrast to the present meta-analysis, Zhao et al. reported no significant association between FSD and PCOS, although with significantly lower scores in the FSFI subscales, namely arousal and lubrication [45]. Pastoor et al. also reported lower scores in the arousal, lubrication, orgasm, and satisfaction subscales [44]. Given the adverse impact of PCOS on sexual satisfaction and quality of life, our results have indicated the need for a greater understanding of FSD in women with this disorder in order to identify new treatment strategies for improved care.
To date, the exact mechanisms of FSD in PCOS are not entirely clear, although changes in sex hormones and psychosocial well-being have been hypothesized [10,46]. Menstrual irregularities and subfertility could lead to low self-esteem and emotional distress, such as depression and anxiety, which might impair sexual function and interpersonal relationships with partners [12,21,[46][47][48]. In addition, women with PCOS might find themselves less attractive due to body dissatisfaction and potential loss of feminine identity as a consequence of obesity and androgen excess [49]. In a case-control study involving 200 Italian premenopausal women, 45% of women with metabolic syndrome had FSD, compared with only 23% in age-and body weight-matched controls [50]. On the other  PCOS is characterized by high levels of androgens (dehydroepiandrosterone, androstenedione, and testosterone) and luteinizing hormone (LH), and increased LH/follicle stimulating hormone (FSH) ratio [52]. While androgen deficiency is linked to reduced sexual drive [53,54], data on women with PCOS are limited and with conflicting results [10,46,55]. Nevertheless, our findings were similar to those of a Brazilian study including 88 women with PCOS (mean age 27 years), which showed a positive association of either androgen excess or an elevated LH level with FSD [13]. Notably, PCOS arises from a vicious cycle of androgen excess, which promotes insulin resistance and compensatory hyperinsulinemia, as well as an amplification of LH-stimulated androgen secretion by the ovarian theca cells and adrenal glands, which can worsen sexual function [2]. In addition, due to the androgen-induced insensitivity of hypothalamic gonadotrophin (GnRH) pulse generator to suppression by estrogen and progesterone [56], there is persistent and rapid GnRH pulsatility with preferential synthesis and release of LH over FSH, in which FSH is physiologically regulated by slow GnRH pulse frequency [57]. The relatively low FSH level prevents ovarian follicular growth, resulting in estrogen deficiency with possible vaginal atrophy. This may explain the increased risk of FSD, particularly dyspareunia, as reported in the present analysis, although no significant differences were found in lubrication subscale scores [2,46].
There is a paucity of data on the treatment effects of PCOS on FSD [10]. In an open-label, observational study of 64 women with PCOS (mean age 29.3 years), 6 months of metformin treatment reduced dyspareunia and improved sexual Fig. 6 Prevalence of female sexual dysfunction in women with and without polycystic ovary syndrome diagnosed via the Rotterdam criteria Fig. 7 Prevalence of female sexual dysfunction in women with and without polycystic ovary syndrome with BMI ≥ 25 kg/m 2 satisfaction and frequency of sexual intercourse, presumably due to improved insulin resistance [58]. Another observational study involving 72 women with androgen excess (mean age 24.3 years) treated with anti-androgenic oral contraceptive pills (OCP), reported significant improvement in hirsutism, sexual pain, orgasm, and satisfaction as early as the 6th cycle and sustained until the 9th cycle of OCP administration [59]. Further research with larger sample sizes is required to validate these interesting findings.
Our results can be generalized to a number of populations, given that the pooled cohort included women from North and South America, Europe, the Middle East, and Asia. Despite this diversity of study populations, the pooled analysis was of low heterogeneity, as evidenced by the I 2 statistic. Our study has several limitations. First, all included studies used self-reported questionnaires, which could be subject to information bias. Second, given that all studies were observational in design, we were unable to provide causal inference. Last, we were unable to evaluate the intervention effect on FSD due to the limited number of studies.

Conclusion
FSD, particularly dyspareunia and a lack of sexual satisfaction, is a prevalent and disabling condition in young women with PCOS. Sensitive probing into the intimate aspects of their sex lives is needed to further understand the struggles that afflict women with PCOS. While more resources should be directed toward the screening of FSD in these women, parallel efforts should be undertaken to investigate the impact of new treatment strategies.

Code availability Not applicable
Authors' contributions HHL, HSL, LLL, AY, and SK performed the systematic search and data extraction. HHL, AY, and LLL assessed the quality of the methodology. AY and SK analyzed the data. HHL, SK, BF, and LLL drafted the manuscript. All authors read and approved the final manuscript. Data availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflicts of interest.
Ethics approval Not applicable.

Consent to participate Not applicable.
Consent for publication Not applicable.