Diet and exercise as a management strategy for symptoms of polycystic ovary syndrome (PCOS): the views of women in the community with PCOS. CURRENT

Background The international clinical practice guidelines for PCOS emphasize and as first-line management of clinical signs and symptoms. This study aimed to describe the patterns, perceptions and experiences of lifestyle interventions for women in the community with PCOS. Method An electronic survey of 493 members of two PCOS consumer support groups described women’s types and patterns of diet and exercise, experiences and perceptions of effectiveness. Women were recruited from the Polycystic Ovary Association of Australia (POSAA) and from the Facebook group, PCOS University Research Group. Associations between participants perceptions of effectiveness, and diet types and exercise patterns were assessed using logistic regression. Response bias for the POSAA group was assessed with a continuum of resistance model. Results 91% of POSAA members aged 16-50 years responded to the survey. Nearly all women reported adjusting their dietary and exercise practices with the aim to improve their health and/or PCOS (82% and 73% respectively), however less than 13% reported achievement of health goals (12.2% and 8.1% respectively). Low carbohydrate, high protein diets, and vigorous activity were associated with self-perceived effectiveness (r.0.16, p<0.01; r.0.15 p<0.01 and r.0.2 p<0.01 respectively). Barriers for lifestyle interventions included psychosocial factors. Response bias was not assessed for the Facebook group, however self-reported PCOS aligned with prevalence of clinical phenotypes and suggests results are generalizable to clinical populations of women with PCOS, who are responsible for self-directing and administering lifestyle interventions to manage their PCOS. Conclusions Perceptions of effectiveness for lifestyle interventions by women with PCOS may be complicated by a lack of rigorous evidence. The strength of recommendations in clinical practice guidelines may be enhanced by clinical trials investigating flexible and feasible lifestyle interventions for women in the community with PCOS.


Background
Polycystic ovary syndrome (PCOS) is the most common reproductive endocrinopathy affecting up to one in five women of reproductive age. 1,2 The cause of PCOS remains unknown however increased prevalence coincides with a global increased incidence of metabolic syndrome. Metabolic disturbance is associated with increased adiposity and clinically worsens menstrual cyclicity, hyperandrogenism, ovulation rates, fertility, maternal and neonatal outcomes and increases women's risks for diabetes, cardiovascular disease and cancer. 2,3 The international evidence based guidelines (EBG) for the management of PCOS highlights personal lifestyle behaviours to reduce and prevent risk factors significantly associated with increased body weight in women with PCOS. 4 Diet and exercise interventions aimed at weight reduction and prevention of weight gain are first-line management strategies due to their significant impact on clinical outcomes, including reproductive endocrinology and clinical signs and symptoms. [3][4][5] However, current evidence for lifestyle intervention for women with PCOS is not well established. Many studies informing the evidence base are underpowered and report high risk of bias, particularly from high attrition. [5][6][7][8] The effectiveness of specific dietary and exercise practices for women centred outcomes has not been established. It has not been shown that if, and to what degree diet and exercise performed outside the confines of study intervention environments correlate with improved health status for women with PCOS in the community. 9 Dietary and exercise practices of women may be influenced by a wide range of ecological factors including inter and intrapersonal, environmental and psychosocial characteristics of individuals plus levels of health literacy and potential for self-efficacy. [10][11][12][13][14] Health promoting behaviours by women with PCOS may be complicated by a lack of evidence informing women of the benefits of lifestyle intervention, 4 and may exacerbate psychological morbidity, which has a high prevalence in women with PCOS, 15 and frustration with current medical treatments. 16,17 To date there are few data describing the views of women with PCOS for engagement and optimal effectiveness of dietary and exercise strategies in community-based environments.
This study has two aims; to describe the symptoms of concern to women with PCOS living in the community, and to explore perceptions about effectiveness for dietary and exercise interventions for PCOS.

Method
The study was approved by the University of Western Sydney Human Research and Ethics Committee

Questionnaire design
A 37-item anonymous questionnaire was designed to describe the signs and symptoms of women with PCOS and women's use of medical treatment and diet and exercise to manage PCOS. (Supplementary File 1). Fifteen items sought information regarding dietary and exercise practices and included types of diets and exercise, types and patterns of exercise practice in terms of frequency, duration and intensity. Moderate intensity exercise was described as causing an increased breath and heart rate but still able to talk. Vigorous exercise was described as causing markedly increased breath and heart rate and not being able to talk during activity. Additional questions examined barriers to engagement including reasons for not partaking or disengaging with therapeutic diet and exercise practices and questions about self-perceived achievement of goals and satisfaction with diets and exercise. Sociodemographic characteristics were also collected. Multiple response options were available for participants to indicate which signs and symptoms of PCOS they were experiencing, and the types of dietary and exercise practices used. A cover letter informed participants' that consent was implied on completion of the questionnaire. The questionnaire took 15-20 minutes to complete. The stability of the questionnaire was tested on a pilot sample of eight women including four with PCOS to review the applicability and accuracy of the questionnaire.
The electronic survey was loaded onto Survey Monkey 18 and made available to both members of POSAA and members of the Facebook group. A single response function was applied to each member and to devices.

Data analyses
Responses were analysed using the statistical package SPSS version 21.0. 19 Categorical responses were reported as percentages and proportions with 95% confidence intervals. Relationships between variables were explored using Pearson's correlation (two tail) for parametric data and Spearman's Rho correlations or Gamma co-efficient (collapsed less than six variables) for non-parametric measurements. Associations between perceptions of effectiveness (dependent variable) and type and pattern of diet and exercise (covariates) were assessed using logistic regression. A p value of 0.05 was considered statistically significant.
Non-response bias was assessed using a continuum of resistance model in which late responders were viewed as proxy non-responders. 20 Participants who responded to the survey four months or more after the final reminder, were classified as late responders. This method was chosen due to the likelihood of early response by women who found the topic interesting. 21 The impact of non-response bias was assessed by comparing early and late responders.

Results
Four hundred and ninety-three participants aged between 16-50 years responded to the survey; 235 of the 258 invited members of the POSAA group responded (91.1%) including 53 late responders, and 311 women replied to snowballing using the University Research Facebook group page. No differences in demographics or PCOS signs and symptoms were found between early and late responders or the two consumer support groups (POSAA and Facebook) (Supplementary file 2.

Tables 1 and 2).
Four hundred and thirty-four (88.0%) subjects completed all parts of the survey, with the remainder using the skip patterns inserted into the electronic survey or completing only part of the survey.
The demographics of participants are presented in Table 1   The majority of women reported they did not fully achieve their health goals using dietary changes.
Over thirty three percent (± 4.5 n = 140) reported they did not achieve their weight loss goals or any Pilates, dance, personal trainer, tennis and team sports.
Sixty percent of responders (± 5.45) were exercising specifically to manage their PCOS.
One hundred and fifteen participants did not exercise ( Nearly forty percent (39.4%, ± 5.35, n = 127) reported non achievement of health goals with exercise.
The majority reported part achievement using exercise (51.8%, ± 5.46, n = 166), and only eight percent (± 2.93, n = 25) of respondents felt they had achieved their health goals with exercise.
We found that increased frequency and duration of exercise was significantly associated with the achievement of health goals and an inverse association between the duration of moderate exercise and the achievement of health goals. (Table 4).  Women with PCOS in the community have been found to lack support and information for selfmanaging PCOS symptoms using lifestyle interventions. 9,17 In a survey of 1385 women with PCOS, only 11.9% reported satisfaction with information provided about beneficial diet and exercise. 17 In the same study, less than 4% of participants were satisfied with information provided about the emotional features of PCOS, with no information or support being offered in most cases. 17 Lack of information and support for community-based women with PCOS may have influenced the low perceptions of efficacy for lifestyle interventions found in the present study. Women's information needs include details about types and patterns of safe and effective lifestyle interventions, and mediators for success and goal achievement including the impact of social support and influence of expectations and experiences of lifestyle interventions. 15,22 Views of community-based women with PCOS presented here contrast with findings of a randomized control trial (RCT) evaluating diet and exercise on the motivators and barriers of exercise for women with PCOS. 8 This clinical investigation found women's perceptions of exercise were significantly improved over 20 weeks in all three study arms (diet only, diet plus aerobic exercise and diet plus aerobic and progressive resistance exercises), and correlated with improved anthropometry, health related quality of life and less depression. 8 Greatest perceived benefits were on the psychological outlook and social interaction subscales of the Exercise Benefits/Barriers Scale (EBBS) (p ≤ 0.001).
Other significant findings were reduced barriers on subscales of exercise milieu (atmosphere), time expenditure and physical exertion (p ≤ 0.003). The contrasting findings found in the present study may be explained by the recruitment strategy of women with PCOS from different populations and by differences in the nature and structure of lifestyle interventions.
The present study recruited participants from consumer support groups in order to assess the views of women in the community, whereas recruitment to the clinical investigation was conducted via public advertisements via medical speciality clinics. Women with PCOS seeking medical support have been shown to display different characteristics of PCOS compared to community-based women 2 which may explain variable perceptions of exercise interventions found in the present study. In addition, the nature of the interventions provided in the clinical trial were structured with weekly contacts, compared to self-informed, self-initiated, self-administered and self-accounted lifestyle interventions typically used by women in the community. Although nearly two thirds of participants reported engagement with supervised/structured exercise, a lack of specific information about evidence-based lifestyle recommendations for PCOS, may have limited guidance of exercise supervisors, with respect to the types and patterns of exercise provided, and a lack of information may have mediated women's expectations and experiences of efficacy. 23 It is unclear which dietary and exercise strategies are optimal for women with PCOS. Current recommendations based on reduced caloric intake, combined with moderate energy expenditure, are described as being most likely to produce sustained weight loss and favourable endocrine and reproductive outcomes. 5,24−27 Women in this study reported similar optimal dietary and exercise practices on symptoms of PCOS, with greatest effectiveness reported for low carbohydrate and high protein diets compared to low calorie diets. Women's perceptions of effective exercise practices highlighted the importance of including vigorous activity, however many respondents reported no engagement in any vigorous activity at all. Various reasons were cited, including time limitations, feelings of embarrassment and personal injury. Additionally, exercise preferences of women with PCOS may favour non-vigorous activity. A study into the comparative exercise practices of women with and without PCOS, controlled for BMI, (n = 163) showed that women with PCOS were less likely to engage in vigorous activity 28 and was strongly determined by personal circumstances and characteristics.
Despite the present study finding most women's exercise activity aligned with recommendations in clinical practice guidelines, 29 the respondents perceptions of low efficacy elucidates a gap in information 17 and a need for guidance despite the absence of high quality evidence. 5

Limitations and strengths
There are some methodological limitations of this study including the non-response and selection bias of recruitment from Facebook support groups. We were not able to assess unsolicited invitations from group administrators and response bias analyses was not possible. Women volunteered to participate, which may have introduced response bias with responders potentially reporting more positive lifestyle behaviours compared to non-responders. 32 Generalizations are limited to computer literate women with English language skills, other groups of women with PCOS may be underrepresented. The questionnaire was not validated against medical records for PCOS lifestyle interventions; however, it was designed to investigate women with PCOS in the community and their perceptions, usage and experiences of lifestyle interventions, which may not coincide with the views of clinical groups. PCOS was self-reported however a similar symptom profile to clinical populations 1,15,33 was described, and these results may be generalisable to women with medically diagnosed PCOS, living in the community. Genetic analyses of PCOS susceptibility has demonstrated consistency between woman's self-reports of the condition and rigorous diagnoses. 34,35 Conclusion Women in the community with PCOS report frequent engagement in a variety of dietary and exercise practices according to the clinical practice guidelines, and they perceive their health goals are partially met. Self-perceived lack of efficacy for dietary and exercise interventions may be due to the lack of high-quality evidence of effectiveness for specific lifestyle interventions. The quality of evidence might be improved by pragmatic RCTs of community-based samples, comparing lifestyle interventions on outcomes of value to women with PCOS. An amendment to the recruitment protocol was approved on 19th August 2013. The reference number was H9341.

Consent to publish
The Participant Information Form informed of potential risks from participation in the survey, and that consent to publish the results was implied through participation and submission of responses. It also explained that that responses could not be withdrawn due to anonymity of the survey.