Introduction
Polycystic ovary syndrome (PCOS) is the most frequent endocrine disorder in women of reproductive age with a reported prevalence of 9–18% (March et al., Reference March, Moore, Willson, Phillips, Norman and Davies2009). Its main clinical and diagnostic features include menstrual irregularity, biochemical and/or clinical hyperandrogenism and presence of polycystic ovaries on ultrasound (Aziz et al., Reference Aziz, Carmina, Dewailly, Diamanti-Kandarakis, Escobar-Morreale and Futterweit2006). Women with PCOS can have severe metabolic and reproductive manifestations (Azziz et al., Reference Azziz, Woods, Reyna, Key, Knochenhauer and Yildiz2004). PCOS is often also associated with excess body weight which can worsen the clinical presentation (Lim et al., Reference Lim, Norman, Davies and Moran2013). Compared with women without PCOS, women with PCOS show an increased prevalence of clinical depression (28–64% v. 7.1–8%), anxiety (34–57% v. 18%) (Deeks et al., Reference Deeks, Gibson-Helm, Paul and Teede2011) and other manifestations of psychological distress (Himelein and Thatcher, Reference Himelein and Thatcher2006; Kerchner et al., Reference Kerchner, Lester, Stuart and Dokras2009; Açmaz et al., Reference Açmaz, Albayrak, Acmaz, Başer, Soyak, Zararsız and İpekMüderris2013; Rowlands et al., Reference Rowlands, Teede, Lucke, Dobson and Mishra2016).
The aetiology of this observed increased prevalence of depression and anxiety in PCOS is still unclear. Possible explanatory factors that have been investigated as a source of distress include visible features such as excess weight, clinical hyperandrogenism (hirsutism, acne or androgenic alopecia) (Barry et al., Reference Barry, Kuczmierczyk and Hardiman2011; Veltman-Verhulst et al., Reference Veltman-Verhulst, Boivin, Eijkemans and Fauser2012), medical consequences such as infertility (Himelein and Thatcher, Reference Himelein and Thatcher2006; Tan et al., Reference Tan, Hahn, Benson, Janssen, Dietz, Kimmig, Hesse-Hussain, Mann, Schedlowski and Arck2008; Deeks et al., Reference Deeks, Gibson-Helm and Teede2010), concerns relating to diagnosis and fear regarding long-term health complications (Trent et al., Reference Trent, Rich, Austin and Gordon2003; Moran et al., Reference Moran, Gibson-Helm, Teede and Deeks2010; Deeks et al., Reference Deeks, Gibson-Helm, Paul and Teede2011). However, the contributing factors to depression and anxiety are currently unclear and results from studies are often conflicting. Only a few of the existing studies (Himelein and Thatcher, Reference Himelein and Thatcher2006; Benson et al., Reference Benson, Hahn, Tan, Mann, Janssen, Schedlowski and Elsenbruch2009b; Moran et al., Reference Moran, Gibson-Helm, Teede and Deeks2010; Deeks et al., Reference Deeks, Gibson-Helm, Paul and Teede2011) have adjusted for factors known to have an association with depression and anxiety in the general population such as body mass index (BMI) (Simon et al., Reference Simon, Von Korff, Saunders, Miglioretti, Crane, Van Belle and Kessler2006), infertility (Cousineau and Domar, Reference Cousineau and Domar2007) and socio-demographic factors such as ethnicity, household income and marital status (Moreno-Peral et al., Reference Moreno-Peral, de Dios Luna, Marston, King, Nazareth, Motrico, GildeGómez-Barragán, Torres-González, Montón-Franco and Sánchez-Celaya2014). It is also well documented that a chronic illness is a stressful condition and chronic stress is associated with depression and anxiety in the general population (De Ridder et al., Reference De Ridder, Geenen, Kuijer and van Middendorp2008; Hammen et al., Reference Hammen, Kim, Eberhart and Brennan2009; Lahey, Reference Lahey2009). While women with PCOS reported a significantly increased physiological reaction to stress compared with controls (Benson et al., Reference Benson, Arck, Tan, Hahn, Mann, Rifaie, Janssen, Schedlowski and Elsenbruch2009a; Farrell and Antoni, Reference Farrell and Antoni2010), there is a paucity of studies investigating the relationship of stress with psychological health in PCOS (Guidi et al., Reference Guidi, Gambineri, Zanotti, Fanelli, Fava and Pasquali2015). The prevalence of stress and its potential contribution to depression and anxiety in PCOS therefore warrants further investigation.
Furthermore, the majority of studies in PCOS have utilised clinic-based samples, convenience control groups, internet surveys (Benson et al., Reference Benson, Arck, Tan, Hahn, Mann, Rifaie, Janssen, Schedlowski and Elsenbruch2009a) or health service databases (Altinok et al., Reference Altinok, Glintborg, Depont Christensen, Hallas and Andersen2014; Hung et al., Reference Hung, Hu, Tsai, Yang, Huang, Chen, Wang, Lu and Shen2014). This limits the generalisability of their findings as referral bias may lead to findings of higher rates of overweight and obesity in such studies (Ezeh et al., Reference Ezeh, Yildiz and Azziz2013) and consequently higher rates of depression and anxiety as excess adiposity is related to worsened psychological health both in women with PCOS and in the general population (Simon et al., Reference Simon, Von Korff, Saunders, Miglioretti, Crane, Van Belle and Kessler2006; Álvarez-Blasco et al., Reference Álvarez-Blasco, Luque-Ramírez and Escobar-Morreale2010). The limited research in community-based studies in PCOS reports greater psychological distress in adolescents and young adults (Guidi et al., Reference Guidi, Gambineri, Zanotti, Fanelli, Fava and Pasquali2015; Rowlands et al., Reference Rowlands, Teede, Lucke, Dobson and Mishra2016) but none in adult women.
The aim of the present study was to assess depression, anxiety, perceived daily stress, psychotropic medication use and health-seeking behaviours with regards to managing psychological health in a large community-based, representative sample of Australian women, comparing women with a self-reported medical diagnosis of PCOS to those who did not report PCOS. We additionally aimed to evaluate the role of stress in contributing to the relationship between PCOS, depression and anxiety.
Methods
Study population and protocol
This study is based on data from the Australian Longitudinal Study on Women's Health (ALSWH), a longitudinal population-based study of four age cohorts of Australian women. The study aims are to examine the physical and mental health of women across the life span assessing biological, psychological, social and lifestyle aspects as well as use and satisfaction with health care services. The ALSWH first collected mailed survey data from three age cohorts studied in 1996. Women were randomly selected from the national health insurance database Medicare (which includes all permanent residents of Australia) and invited to participate in the project. Women were recruited nationally with intentional oversampling from rural and remote areas (Lee et al., Reference Lee, Dobson, Brown, Bryson, Byles, Warner-Smith and Young2005). Further information about the methods used and sample characteristics has been reported elsewhere (Brown et al., Reference Brown, Bryson, Byles, Dobson, Lee, Mishra and Schofield1999; Powers and Loxton, Reference Powers and Loxton2010) and are available on the ALSWH website (http://www.alswh.org.au). The Human Research Ethics Committees of the University of Newcastle and the University of Queensland approved the study methods and informed written consent was obtained from each participant. The current study is based on data from the cohort born between 1973 and 1978 who first completed the survey at age 18–23 years. We analysed data from survey 4 (2006) and included women who responded to the question on PCOS (online Supplementary Fig. S1).
Demographic measures and participant characteristics
(1) PCOS: Women were asked ‘In the last three years, have you been diagnosed or treated for polycystic ovary syndrome?’. Women who responded ‘yes’ were classified as PCOS, whereas all other women who responded to the question were classified as not having PCOS.
(2) BMI: Weight and height were self-reported. BMI (weight in kg/height in metres2) was calculated from self-reported height and weight. BMI was then classified into categories, with overweight and obesity defined by the World Health Organization criteria (BMI 25.0–29.9 kg/m2 for overweight and ⩾30 kg/m2 for obese classification) (World Health Organization, 1999). Socio-demographic variables included age, education, occupation, marital status, number of children, ethnicity and personal income.
(3) Fertility was assessed by the question ‘Have you and your partner (current or previous) ever had problems with fertility – that is, tried unsuccessfully for 12 months or more to get pregnant?’ with the answers: (1) no never tried to get pregnant; (2) no, had no problem with fertility; (3) yes, but have not sought help/treatment and (4) yes, and have sought help/treatment available. Women who responded ‘no never tried to get pregnant’ or ‘no, had no problem with fertility’ were classified as not knowing/not experiencing infertility. Those who responded ‘yes, but have not sought help/treatment’ or ‘yes, and have sought help/treatment’ were classified as having experienced infertility.
Psychological outcomes
Both validated scales and more self-reported questions were used to investigate psychological features. Primary outcomes were depression and anxiety using standardised and validated tools.
(4) Depression: Depression symptoms were assessed using the 10-item Centre for Epidemiologic Studies Depression Scale (CESD-10) with a score of ⩾10 used as a categorical cut-off for clinically significant symptoms (Andresen et al., Reference Andresen, Malmgren, Carter and Patrick1994).
(5) Anxiety: Anxiety symptoms were assessed using the anxiety subscale of the Goldberg Depression and Anxiety Scale (GADS) with a score of ⩾5 used as a categorical cut-off for risk of clinical anxiety as previously described (Goldberg et al., Reference Goldberg, Bridges, Duncan–Jones and Grayson1987, Reference Goldberg, Bridges, Duncan-Jones and Grayson1988).
(6) Perceived stress: Perceived stress was analysed as a continuous variable assessed by the Perceived Stress Questionnaire (PPQ) (Bell and Lee, Reference Bell and Lee2002, Reference Bell and Lee2003) which is a scale that has been developed for and validated in the ALSWH. The total score ranges from 0 to 4 with a score >2 indicating moderate levels of stress and a score >3 or >4 indicating, respectively, a very and extremely stressful condition.
Secondary measures included self-reported information concerning medical diagnosis of mental illness, psychological symptoms and health-seeking behaviours relating to mental health management.
(7) Mental illness diagnosis: A self-reported medical diagnosis of depression, anxiety or other major mental illnesses was assessed with the question ‘In the last three years have you been diagnosed or treated for’. Women could respond to these questions with either ‘yes’ or ‘no’ responses (yes/no).
(8) Psychological symptoms and seeking behaviours: Self-reported symptoms of depression, episodes of intense anxiety (e.g. panic attacks) or other mental problems and the proportion of women seeking help for those symptoms were investigated asking ‘in the last 12 months have you had one of the following? (no, rarely, sometimes, often)’. Only women who responded ‘often’ were categorised as reporting symptoms. The question was then asked ‘If yes, did you seek help for this problem? (yes/no)’ and these data were also included in the analysis.
(9) Other health-seeking behaviours relating to mental health management: psychotropic medication use was investigated using the following two questions: ‘During the past 4 weeks have you used medications that were: prescription medication for depression (e.g. Zoloft, Aropax, Lexapro, Cipramil, etc.), prescription medication for your nerves/anxiety (e.g. Valium, Serapax, Kalma, Ducene, etc.)’. Women could respond to these questions with either ‘yes’ or ‘no’ responses (yes/no). We also included the question ‘Have you consulted a counsellor or other mental health worker for your own health in the last 12 months?’ with ‘(yes/no)’ responses in the analysis.
Statistical analysis
Data were analysed using Stata software version 11.0 (StataCorp, Texas, Lakeway, USA).
Categorical data were expressed as proportion and continuous data as mean ± standard deviation. Comparison between women with and without PCOS was performed by χ2 for categorical variables and independent Student's t test for continuous variables. Binary logistic regression analyses were used to assess the relationship between depression or anxiety and PCOS, and linear regression models were used to assess the association between perceived stress and PCOS. Potential confounding variables in these relationships were assessed and incorporated in the model including socio-demographic factors, BMI (categorical variable) and infertility. The selection of variables was based on identifying all measured variables of known or suspected relevant effect on depression, anxiety and perceived stress, and/or exhibiting p < 0.10 on univariable analysis. In addition to examination as a dependent variable, perceived stress was also considered as a potential confounding variable in the association between PCOS, depression and anxiety. Since stress was also considered to be a mediator in the relationship between PCOS, depression and anxiety, binary mediation analysis was performed to partition the indirect effect of PCOS on both depression and anxiety. The significance level was set at a two-tailed 5%. Analysis was also weighted by area of residence to adjust for the deliberate oversampling in rural and remote areas.
Results
On 9145 responders to the survey, 8612 answered to the question of PCOS status and were included in the analysis (online Supplementary Fig. S1). Four hundred and seventy-eight were classified as reporting PCOS and 8134 as not reporting PCOS (abbreviated as with and without PCOS, respectively), with an estimated prevalence of self-reported PCOS of 5.8% (95% CI 5.3–6.3). Participant characteristics for the entire group and the PCOS and non-PCOS groups are reported in Table 1. The mean age was 30.6 ± 1.5 years and the mean BMI was 25.2 ± 5.7 kg/m2. As previously reported (Teede et al., Reference Teede, Joham, Paul, Moran, Loxton, Jolley and Lombard2013), women with PCOS reported a greater BMI compared with women without PCOS (PCOS 28.1 ± 7.2 kg/m2 v. non-PCOS 25.1 ± 5.6 kg/m2) and a significantly higher proportion were in the overweight (PCOS 26.1% v. non-PCOS 23.2%) and obese range (PCOS 33.2% v. non-PCOS 16.2%) (p < 0.001). Women with PCOS were more likely to report experiencing infertility (PCOS 47.4% v. non-PCOS 9.2%) (p < 0.001) (Table 1). Women with and without PCOS did not differ significantly in socio-demographic variables.
Characteristics for women with and without PCOS
Values are reported as mean (±s.d.) or number (%).
Data were analysed by t test for continuous variables and χ2 test for categorical variables.
PCOS, polycystic ovary syndrome; BMI, body mass index; WHO, World Health Organization.
On unadjusted analysis, a significantly higher proportion of women with PCOS reported clinically significant depression (PCOS 27.3% v. non-PCOS 18.8%) and anxiety symptoms (PCOS 50.0% v. non-PCOS 39.2%) (p < 0.001). Women with PCOS also had higher levels of perceived stress compared with women without PCOS (PCOS 1.06 ± 0.6 v. non-PCOS 0.88 ± 0.5) (p < 0.001) (Table 2). On univariable regression analysis, the odds of reporting depression symptoms or a score in the clinical risk range of depression was 1.61-fold higher in women with PCOS (95% CI 1.29–2.03, p < 0.001). The odds of reporting anxiety symptoms risk score was 1.46-fold higher in women with PCOS (95% CI 1.20–1.79, p < 0.001). The odds of reporting perceived stress was 0.16 in women with PCOS (95% CI 0.11–0.22, p < 0.001) (Table 2). Women with PCOS reported a significantly higher score than those without PCOS in all secondary psychological outcomes with the exception of seeking help for anxiety symptoms and other mental problems in the last 12 months. Women with PCOS were more likely to report psychotropic medication use both for depression (PCOS 12.6% v. non-PCOS 7.03%, p = 0.001) and anxiety (PCOS 5.4% v. non-PCOS 2.8%, p = 0.001) (Table 1, online Supplementary Data).
Depression, anxiety and perceived stress in women with and without polycystic ovary syndrome
Values are reported as mean (±s.d.), number (%) or mean, 95% CI.
Data were analysed by t test for continuous variables and χ2 test for categorical variables and survey-weighted univariable and multivariable logistic analysis.
PCOS, polycystic ovary syndrome; CESD-10, Centre for Epidemiologic Studies Depression Scale; GADS, Goldberg Anxiety Depression Scale; PSQ, Perceived Stress Questionnaire.
a Adjusted for BMI, infertility and socio-demographic factors (age, education, occupation, marital status, number of children, ethnicity, personal income).
b Adjusted for perceived stress, BMI, infertility, socio-demographic factors (age, education, occupation, marital status, number of children, ethnicity, personal income).
In multivariable regression analysis, after adjusting for age, BMI, infertility, education, occupation, marital status, number of children, ethnicity and personal income, the effect size was slightly reduced for the outcomes of depression (OR 1.39, 95% CI 1.04–1.86, p = 0.027), anxiety (OR 1.37, 95% CI 1.07–1.76, p = 0.012) and perceived stress (coefficient β 0.15, 95% CI 0.09−0.23, p < 0.001) (Tables 3–5). A second multivariable analysis was performed both for depression and anxiety, adding perceived stress to the other potential confounding variables described above. On these analyses, the association between PCOS with both depression and anxiety symptoms was no longer significant (Tables 4 and 5). On adjusted analyses, for a one-unit increase in stress, the odds of being clinically depressed and anxious increased by a factor of 9.93 and 11.36 (p < 0.001), respectively.
Univariable and multivariable regression analysis – depression
Data were analysed by survey-weighted univariable and multivariable logistic analysis.
PCOS, polycystic ovary syndrome; BMI, body mass index; WHO, World Health Organization.
a Adjusted for BMI, infertility and socio-demographic factors (age, ethnicity, education, marital status, number of children, personal income).
b Adjusted for perceived stress, BMI, infertility, socio-demographic factors (age, ethnicity, education, marital status, number of children, personal income).
Univariable and multivariable regression analysis – anxiety
Data were analysed by survey-weighted univariable and multivariable logistic analysis.
PCOS, polycystic ovary syndrome; BMI, body mass index; WHO, World Health Organization.
a Adjusted for BMI, infertility and socio-demographic factors (age, ethnicity, education, marital status, number of children, personal income).
b Adjusted for perceived stress, BMI, infertility, socio-demographic factors (age, ethnicity, education, marital status, number of children, personal income).
Univariable and multivariable regression analysis – stress
Data were analysed by survey-weighted univariable and multivariable regression analysis.
PCOS, polycystic ovary syndrome; BMI, body mass index; WHO, World Health Organization.
a Adjusted for BMI, infertility and socio-demographic factors (age, ethnicity, education, marital status, number of children, personal income).
On mediation analysis, we found that stress showed a high-level mediation effect between the relationship between PCOS and depression, with a proportion of total effect moderated = 0.71. The indirect effect of PCOS on depression was given by OR 1.18 (95% CI 0.92–1.51). A high-level mediation effect of stress between PCOS and anxiety was also found, with a proportion of total effect moderated = 0.71. The indirect effect of PCOS on anxiety was given by OR 1.18 (95% CI 0.95–1.47).
Discussion
The present study assessed for the first time depression and anxiety symptoms and perceived stress in adult women with PCOS in a large community-based cohort study. Women reporting PCOS reported an increased prevalence of depression and anxiety symptoms, perceived stress, self-reported medical diagnoses of depression, anxiety or other major mental illnesses and treatment for psychological conditions or mental illness. We also report here for the first time the confounding and the mediating effect of stress in the association between both depression and anxiety and PCOS.
We report here an increased prevalence of clinical depression of 27.3% and 18.8% and anxiety of 50.0% and 39.2% in women with and without PCOS, respectively, compared with 7.1–8% and 18% in the general population (Deeks et al., Reference Deeks, Gibson-Helm, Paul and Teede2011). In agreement with previous research (Barry et al., Reference Barry, Kuczmierczyk and Hardiman2011; Dokras et al., Reference Dokras, Clifton, Futterweit and Wild2011; Veltman-Verhulst et al., Reference Veltman-Verhulst, Boivin, Eijkemans and Fauser2012), women with PCOS had increased risk of clinically significant depression and anxiety symptoms (1.39- and 1.37-fold respectively). These increased odds were lower than reported in a recent meta-analysis (4.03- and 6.88-fold) (Dokras et al., Reference Dokras, Clifton, Futterweit and Wild2011, Reference Dokras, Clifton, Futterweit and Wild2012), which may be because the meta-analysis included clinical-based studies while our study used a community-based sample. Many factors in women may contribute to depression including hormonal and biological (e.g. infertility, child birth and premenstrual syndrome) and psychosocial (e.g. stress, socio-economic advantage and violence) factors (National Institute of Mental Health, 1995). We considered in our study BMI, infertility and socio-demographic variables as specific risk factors for depression for the general population (Simon et al., Reference Simon, Von Korff, Saunders, Miglioretti, Crane, Van Belle and Kessler2006). While some of these factors have been previously considered in investigating the relationship between PCOS and psychological variables (Himelein and Thatcher, Reference Himelein and Thatcher2006; Benson et al., Reference Benson, Hahn, Tan, Mann, Janssen, Schedlowski and Elsenbruch2009b; Moran et al., Reference Moran, Gibson-Helm, Teede and Deeks2010; Deeks et al., Reference Deeks, Gibson-Helm, Paul and Teede2011; Rowlands et al., Reference Rowlands, Teede, Lucke, Dobson and Mishra2016), to our knowledge, this is the first community-based study considering all of them together in one analysis.
In the present study, the association between PCOS, depression and anxiety was attenuated but maintained on adjustment for BMI, infertility and socio-demographic variables. This indicates that although the presence of overweight or obesity and infertility may worsen depression and anxiety as reported in the general population (Simon et al., Reference Simon, Von Korff, Saunders, Miglioretti, Crane, Van Belle and Kessler2006), PCOS status per se is likely to have an independent effect on psychological function (Deeks et al., Reference Deeks, Gibson-Helm and Teede2010). This may be related to visible features, the frustration of having a chronic condition (Kozica et al., Reference Kozica, Gibson-Helm, Teede and Moran2013) or the perceived risk of future health complications (Moran et al., Reference Moran, Gibson-Helm, Teede and Deeks2010).
In agreement with prior results (Guidi et al., Reference Guidi, Gambineri, Zanotti, Fanelli, Fava and Pasquali2015), being stressed was more prevalent in PCOS even after consideration of BMI, infertility and socio-demographic variables. The questionnaire used here assesses perceived sources of stress in several domains of daily life (Bell and Lee, Reference Bell and Lee2002, Reference Bell and Lee2003) rather than being designed for the assessment of health-related stress. However, it also assesses perceived stress in life domains potentially impacted by PCOS such as health, motherhood or social relationships (Farkas et al., Reference Farkas, Rigó and Demetrovics2014). Future work should investigate if women with PCOS display higher stress related to their condition rather than other daily life domains or if stress precedes PCOS. We also report here for the first time the role of stress both as confounder and mediator variable between the relationship between PCOS and psychological outcomes. The significant association between PCOS and depression and anxiety was lost on adjustment for stress, and stress showed a strong association both with depression and anxiety. This suggests an independent confounding relationship of stress both with PCOS status and with depression or anxiety and that higher depression and anxiety in PCOS may be related to higher stress levels in PCOS. Mediation analysis gave further clarification as stress mediated for a large proportion of the relationship between PCOS and both depression and anxiety. This could suggest that stress can have a relevant direct effect in depression and anxiety symptoms in women with PCOS, rather than other factors considered up to now. This is consistent with prior research on the relationship between stress, chronic illness and psychological morbidity in the general population (Hammen et al., Reference Hammen, Kim, Eberhart and Brennan2009; Lahey, Reference Lahey2009). This may explain why previous studies have not found any direct causal relationship between PCOS and depression or anxiety.
As endocrine systems may be more vulnerable to the physiological effects of stress due to the pathophysiological features of PCOS such as hypothalamic–pituitary–adrenal axis (HPA) and sympathetic nervous system (SNS) hyperactivity or low-grade immune system inflammation (Benson et al., Reference Benson, Janssen, Hahn, Tan, Dietz, Mann, Pleger, Schedlowski, Arck and Elsenbruch2008), even low levels of perceived stress may have a clinically significant impact in PCOS (Farrell and Antoni, Reference Farrell and Antoni2010; Barry et al., Reference Barry, Kuczmierczyk and Hardiman2011). In light of this, the evaluation of coping and adjustment strategies in PCOS is relevant as key factors relating to psychological outcomes (Benson et al., Reference Benson, Hahn, Tan, Janssen, Schedlowski and Elsenbruch2010) and preliminary trials of stress management interventions have reported encouraging results on the amelioration of psychological outcomes in PCOS (Raja-Khan et al., Reference Raja-Khan, Agito, Shah, Stetter, Gustafson, Socolow, Kunselman, Reibel and Legro2015; Stefanaki et al., Reference Stefanaki, Bacopoulou, Livadas, Kandaraki, Karachalios, Chrousos and Diamanti-Kandarakis2015).
The findings of the present study therefore confirm the importance of consideration of assessment and management of stress, even at non-clinical levels, in PCOS in addition to depression and anxiety. Furthermore, the independent association of BMI and infertility with both depression and anxiety was also lost on inclusion of stress into the multivariable models. Stress may therefore have a stronger impact on psychological distress than other factors previously highlighted as risk factors for depression and anxiety both in PCOS and in the general population.
We report here a more comprehensive assessment of psychological distress in women with and without PCOS than previously reported. A significantly higher proportion of women with PCOS reported having been diagnosed or treated for depression, anxiety or other mental illness, used psychotropic medications and suffered from or sought help for psychological distress symptoms or consulted a mental health professional. This is in keeping with the higher prevalence of depression and anxiety in PCOS. However, the self-reported prevalence of symptoms or episodes of depression or anxiety and the percentage of women engaged in psychological health-seeking behaviours was also lower than the prevalence of depression and anxiety symptoms using validated scales. This suggests that further research is warranted investigating if women with PCOS are aware of treatment options and strategies to encourage them to seek help for psychological distress. While there are numerous studies assessing psychological distress in women with PCOS, there is relatively limited literature either examining health-seeking behaviours relating to mental health management (Benson et al., Reference Benson, Hahn, Tan, Mann, Janssen, Schedlowski and Elsenbruch2009b; Moran et al., Reference Moran, Gibson-Helm, Teede and Deeks2010; Altinok et al., Reference Altinok, Glintborg, Depont Christensen, Hallas and Andersen2014) or psychological treatments (Rofey et al., Reference Rofey, Szigethy, Noll, Dahl, Lobst and Arslanian2008; Raja-Khan et al., Reference Raja-Khan, Agito, Shah, Stetter, Gustafson, Socolow, Kunselman, Reibel and Legro2015) either in clinical or community samples. As far as we know that is only the second study (Altinok et al., Reference Altinok, Glintborg, Depont Christensen, Hallas and Andersen2014) investigating the use of psychotropic medication among women with PCOS. In accordance with these previous results, we found that women with PCOS were more likely to use psychotropic medication than women without PCOS. This is an important aspect of health-seeking behaviours that need to be further explored.
There is therefore a need to report community-based data on current uptake of management strategies for psychological health. This may encourage clinicians to discuss management options with patients for psychological features.
Strengths of this study include the use of a community-based cohort which reduces the bias of clinical-based studies in overestimating overweight and obesity (Ezeh et al., Reference Ezeh, Yildiz and Azziz2013), reproductive (Khan et al., Reference Khan, Daya, Collins and Walter1996) and potentially psychological issues. Further strengths include the consideration of a broad range of potential confounding variables, the use of valid and reliable psychological measures as primary outcomes, the assessment of multiple less commonly investigated psychological outcomes and the examination of the effect of stress in psychological distress in women with PCOS. This is a cross-sectional analysis and neither causality nor a longitudinal evaluation between PCOS and psychological distress can be established. The major limitations of this study are the use of self-reported information such as PCOS status, infertility, BMI and psychological outcomes. However, this is reasonable in large epidemiologic studies due to feasibility and economy. Furthermore, validation studies have previously reported the validity of self-report measures for anthropometric measures as well as medical and psychological records in the general population or primary care setting (Von Korff et al., Reference Von Korff, Ustun, Ormel, Kaplan and Simon1996; Haapanen et al., Reference Haapanen, Miilunpalo, Pasanen, Oja and Vuori1997; Spencer et al., Reference Spencer, Appleby, Davey and Key2002). Another limitation is the lack of information about treatments such as antiandrogen therapy and contraception which may have potential effect on mood (Davis and Tran, Reference Davis and Tran2001; Oinonen and Mazmanian, Reference Oinonen and Mazmanian2002). The women in this study were also all aged between 28 and 33 years, and these results may not be applicable to women in other age groups. This is the first study assessing depression, anxiety symptoms and perceived stress in adult women with and without PCOS in a large, unselected community cohort. This study confirms findings of worsened psychological distress and highlights the role of even non-clinical stress levels in potentially playing a role in the association among PCOS, depression and anxiety. It emphasises the need for a multidisciplinary approach to this increasing and debilitating condition. Future work should consider more specific assessments of health-related stress in women with PCOS and the potential role for stress management in the treatment of psychological distress in PCOS.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/S0033291718002076.
Acknowledgements
The research on which this paper is based was conducted as part of the Australian Longitudinal Study on Women's Health undertaken by The University of Newcastle and The University of Queensland. The authors are grateful to the Australian Government Department of Health for funding and to the women who provided the survey data.
Author contributions
AJ, DL, HT and LM were responsible for substantial contributions to the study conception and design. DL and HT were responsible for data acquisition. ALD, AJ, AE and LM were responsible for the analysis and interpretation of data. ALD, AJ, DL, AE, HT and LM were responsible for drafting the article or revising it critically for important intellectual content, and final approval of the version to be published.
Financial support
This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors. The ALSWH is funded by the Australian Government Department of Health. ALD received a scholarship from the University of Bologna (Italy) for her secondment at Monash University. HJT holds a National Health and Medical Research Council of Australia (NHMRC) Practitioner fellowship. AJ holds a NHMRC early career fellowship. HJT holds a National Health and Medical Research Council of Australia (NHMRC) Practitioner fellowship. LM is supported by a SACVRDP fellowship; a program collaboratively funded by the NHF, the South Australian Department of Health and the South Australian Health and Medical Research Institute and a NHF Future Leader fellowship.
Conflict of interest
None.
