Article Text

Protocol
Effectiveness and acceptability of different lifestyle interventions for women with polycystic ovary syndrome: protocol for a systematic review and network meta-analysis
  1. Fuliang Shangguan1,
  2. Hui Liu1,
  3. Yu Guo1,
  4. Juping Yu2,
  5. Yinni Liang1,
  6. Huixi Yu1,
  7. Yinhua Su1,
  8. Zhongyu Li1
  1. 1 University of South China School of Nursing, Hengyang, Hunan, China
  2. 2 Faculty of Life Sciences and Education, University of South Wales, Pontypridd, UK
  1. Correspondence to Professor Zhongyu Li; lzhy1023{at}hotmail.com; Dr Yinhua Su; 382373646{at}qq.com

Abstract

Introduction Women with polycystic ovary syndrome (PCOS) experience various metabolic, endocrine, reproductive and psychosocial manifestations. Lifestyle modification is crucial for the management of PCOS to reduce long-term complications. Nonetheless, the efficacy and acceptability of lifestyle interventions differs, and there are no uniform methods of clinical application. Hence, a systematic review and network meta-analysis (NMA) are needed to explore the efficacy and acceptability of lifestyle interventions to inform clinical practice.

Methods and analysis Ten databases (Cochrane Gynaecology and Fertility Specialised Register, Cochrane Register of Studies Online, PubMed, EMBASE (Excerpta Medica Database), PsycINFO, CINAHL (Cumulative Index to Nursing and Allied Health Literature), Chinese National Knowledge Infrastructure, WanFang, VIP, and Sinomed) and four clinical trial registry platforms will be searched to identify literature published in English or Chinese reporting results of randomised clinical trials conducted to evaluate the effects of lifestyle interventions for women with PCOS. The reference lists of the included studies will be manually searched. Primary outcomes will include biochemical and clinical hyperandrogenism, recruitment and retention rates. Secondary outcomes will encompass menstrual regularity, ovulation, anthropometry and quality of life. Literature selection and extraction of data will be performed independently by at least two researchers. An NMA random-effects model will be implemented for amalgamating evidence. All treatments will be ranked based on the value of p. OpenBUGS will be used for Bayesian modelling, with output verifications generated in Stata and R. The quality of evidence supporting network estimates of major outcomes will also be appraised using the Grading of Recommendations Assessment, Development, and Evaluation framework.

Ethics and dissemination Ethical approval is not required for this review as no data will be collected from human participants. Results will be presented in a peer-reviewed publication.

PROSPERO registration number CRD42024499819

  • Meta-Analysis
  • Gynaecology
  • Nurses
http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Strengths and limitations of this study

  • Random-effects network meta-analysis (NMA) will facilitate a simultaneous comparison of multiple lifestyle interventions in women with polycystic ovary syndrome (PCOS) in a single model.

  • A systematic search strategy will be developed to systematically identify all available evidence on lifestyle interventions in women with PCOS.

  • The study encompasses women with all types of PCOS, without a narrow focus on a specific target population.

  • Only articles published in English or Chinese will be included in the review.

  • The results may be affected by publication bias, as is the case with all systematic reviews and NMA.

Introduction

Polycystic ovary syndrome (PCOS) is a complex endocrine and metabolic disorder, featuring anovulation, hyperandrogenism and polycystic ovarian morphology.1 2 It is estimated that 5%–18% of reproductive age women worldwide are affected by the condition.3 Its treatment varies and needs to be tailored according to symptoms, age and metabolic risks.4 Although its underlying mechanism is unclear, it is believed that hyperandrogenaemia, hyperinsulinaemia and family history (heritability of around 70%) play a key role.5–8 Studies show that up to 80% of patients with PCOS are overweight or obese, contributing to insulin resistance and hyperinsulinaemia which could lead to diabetes mellitus, cardiometabolic diseases and reproductive sequelae.9–11 There are stigmas attached to hyperandrogenism as hirsutism, acne and alopecia are its common clinical expressions.12 This can significantly affect the quality of life and mental health of women suffering from hyperandrogenism.13–15 Hence, effective weight control plays a crucial role in managing overall health in management of PCOS.9 16

Lifestyle modification, which involves dietary, exercise and behavioural changes, is recommended to women with PCOS as first-line management.2 17 Studies have shown significant improvements in fertility and metabolic indicators among women with PCOS following structural dietary and exercise interventions.18 19 Behavioural interventions, such as cognitive behavioural therapy, play an crucial role in increasing self-efficacy and improving metabolism.20 Hormonal imbalance significantly exacerbates the clinical manifestation of PCOS, while lifestyle change has showed positive effects on PCOS symptoms.21

Adherence to lifestyle changes can be challenging, making lifestyle interventions unattainable for many women with PCOS, as evidenced by drop rates.4 22 For example, studies show an average up to 50% drop rates in weight loss trials in PCOS,5 compared with 31% in the general population (aggregated data n=26 455 from 80 studies.23 Acceptability for lifestyle interventions among individuals with PCOS is a major concern.24 25 Hence, an understanding of issues around compliance as reported in such interventions is crucial to the improvement of compliance.26–28

There is a lack of evidence to guide the selection of optimal lifestyle interventions due to the lack of conclusive conclusions.20 Network meta-analysis (NMA) is a method for comparing a network of interventions simultaneously by combining both direct and indirect evidence from randomised controlled trials (RCTs) to allow estimation of the ranking and hierarchy of interventions, which helps practitioners to choose the most appropriate strategies.29 A comprehensive summary of evidence from lifestyle interventions, adhering to the broadest inclusive criteria, would greatly aid in addressing clinically relevant questions about enhancing metabolic health in women with PCOS. Therefore, the aim of this systematic review and NMA is to systematically review all the available evidence and to evaluate the study acceptability and the comparative effectiveness of lifestyle interventions on high androgen-related indicators, menstrual regularity, anthropometric measures and quality of life.30

Methods and analysis

Study registration

The protocol was registered on PROSPERO (No. CRD42024499819), adhering to the Preferred Reporting Items for Systematic Review and Meta‐Analysis Protocols (PRISMA-P) guidelines (Research Checklist).31 The results will be reported following the PRISMA-NMA guidelines for reporting a systematic review involving an NMA.

Inclusion criteria

Types of studies

Studies to be retrieved will include RCTs using a parallel group design or a cross-over design to evaluate the effect of lifestyle interventions and meet the ‘PICOT’ criteria described below. For RCTs using a cross-over design, only the data from the first phase will be included in the meta-analyses due to the anticipated lasting effects of the intervention under investigation. Studies can be carried out in any country.

Participants (P)

Participants of the included studies must include women of reproductive age (15–49 years) with PCOS. We will include studies using any definition of PCOS in this review, with the trialist’s definition of PCOS described. Women with PCOS who have any comorbidities such as infertility or obesity will be included in our study. There will be no restrictions on participants’ race, nationality and education levels.

Interventions (I)

The review will focus on lifestyle interventions including dietary interventions, exercise interventions and behavioural interventions used as the main treatment or main adjuvant treatment.2 Studies that involve multiple lifestyle interventions will also be included. Dietary interventions often involve calorie reduction or diet structure change (carbohydrate counting, fat counting and protein counting).21 Exercise interventions often include resistance or aerobic exercise.21 Behavioural interventions apply modern theories of learning to address behavioural disorders by goal setting, self-monitoring, problem-solving, assertiveness training, reinforcing changes and relapse prevention.32 Figure 1 presents the included studies based on our previous literature searches.

Figure 1

Network of conceivable pairwise contrasts between viable interventions. RESMENA (The MEtabolic Syndrome REduction in NAvarra) diet is characterised by increased meal frequency (7 meals/day), low glycemic load, high antioxidant capacity, high n-3 FAs, high protein, and healthy FA content.

Comparison (C)

All studies included in the review will have an intervention group and a control group. Participants in the control group may receive a different intervention, routine treatment or no treatment.

Outcome measures (O)

Primary outcomes
  1. Hyperandrogenaemia: clinical symptoms (eg, hirsutism, acne and alopecia) and biochemical markers, including total testosterone and bioavailable testosterone.31

  2. Dropout rates: proportion of patients who withdrew for any reason.

Secondary outcomes
  1. Menstrual regularity: an initiation of menses or significant shortening of cycle length where possible.33

  2. Ovulation: the number of ovulatory menstrual cycles where possible.33

  3. Anthropometric: including weight, body mass index and adiposity distribution (by measures including waist circumference and waist-to-hip ratio).34 35

  4. Quality of life: a multidimensional concept that encompasses the physical, emotional and social aspects associated with a disease.34 36

Time frame of outcome evaluation (T)

Only studies with a follow-up period of more than 6 months will be included to ensure sufficient time to evaluate any change of clinical expressions of hyperandrogenism and biochemical makers.5 For RCTs where the outcome was measured more than once, results from the last data collection point will be used.

Exclusion criteria

Studies will be excluded if:

  1. A non-RCT design was used.

  2. Participants included those with conditions akin to PCOS (eg, congenital adrenal hyperplasia, Cushing’s syndrome, hyperprolactinaemia, thyroid disorders and androgen-secreting tumours).

  3. Participants had active eating disorders, or using drugs for dyslipidaemia, hypertension, diabetes/impaired glucose tolerance, or hormonal treatment.

  4. Studies have high potential bias (eg, lack of randomisation or blinding).

  5. Articles were published neither in English nor in Chinese.

Search strategy

Ten databases will be used to identify relevant literature, including Cochrane Gynaecology and Fertility Specialised Register, CENTRAL via Cochrane Register of Studies Online, PubMed, EMBASE, PsycINFO, CINAHL, Chinese National Knowledge Infrastructure, WanFang Data, the Chongqing VIP Database and China Biology Medicine (Sinomed). In addition, four clinical trials registry platforms will be searched, including ClinicalTrials.gov and the WHO International Clinical Trials Registry Platform, grey literature in OpenSIGLE, and Latin American and Caribbean trials in the Latin American and Caribbean Health Sciences Literature database.

The following search terms will be used: ‘lifestyle intervention’, ‘PCOS’ and ‘RCT’ based on the principle of mesh terms combined with entry terms. Appropriate adjustments will be made accordingly. A comprehensive search strategy for each database is available (see online supplemental appendices). The additional four clinical trials registry platforms will be searched using ‘polycystic ovary syndrome’, ‘PCOS’ as the keywords.

Supplemental material

The selected studies will include all published, unpublished and ongoing studies that are conducted to evaluate the effects of lifestyle interventions in women with PCOS. The language will be limited to only Chinese and English. It is important to include unpublished studies, since publication bias can lead to exaggerated effect sizes, affecting the results.37 Data restrictions will not be applied to any of the searches. The reference lists of included studies will be searched for additional studies. Where eligible studies are found, unpublished data will be requested from the investigators.

Data collection and analysis

Study selection

Covidence (www.covidence.org) will be used to manage the retrieved studies and remove duplicates. First, two researchers (FS and HL) will independently screen the titles and abstracts of retrieved studies against the eligible criteria. Second, the same researchers will conduct the full text screening independently. Any disagreement will be solved by discussion and a third researcher will be consulted if needed. If both reviewers agree that a study does not meet eligibility criteria, it will be excluded and the reasons for exclusion will be recorded. The included studies will be crosschecked. The selection procedure is detailed in a PRISMA flow chart (figure 2).

Figure 2

Preferred Reporting Items for Systematic Review and Meta‐Analysis (PRISMA) flow diagram.

Data extraction and management

FS and HL will extract data independently. The data to be extracted will include five broad areas.

  1. Basic information (authors, year of publication, country where the study was conducted, study design and funding source).

  2. Participants (demographics, diagnostic criteria, sample size, types of symptoms and duration of illness).

  3. Interventions (intervention type, content, frequency, duration and dosage of intervention; allocation; blinding).

  4. Controls (control, content, frequency, duration and dosage of treatment).

  5. Outcomes (measurement data of primary and secondary outcomes; how adherence was measured; follow-up time, missing data handling, and conclusions).

Corresponding authors will be contacted for further details if needed. FS and HL will verify postextracted date. Any disagreement will be solved by discussion and a third researcher will be consulted if needed.

Assessment of risk of bias

FS and HL will evaluate the risk of bias of the included studies independently using the Cochrane Collaboration’s tool for assessing risk of bias V.2.0.38 Five domains will be scrutinised: (1) Randomisation process bias, (2) Deviations from intended intervention bias, (3) Missing outcome data bias, (4) Outcome measurement bias, and (5) Selective reporting bias.39 In cases where allocation concealment and trial characteristics were insufficiently detailed in an article, supplemental information will be requested from the corresponding author of the article. The overall risk of bias will be classified as low if all domains are categorised as low risk. The risk of bias will be classified as high if one domain is labelled high risk or multiple domains exhibit some concerns, potentially undermining the study’s integrity. Corresponding authors will be contacted for critical information. FS and HL will conduct cross-verification of the postassessment data. Disputes will be addressed by discussions until a consensus is reached. A third reviewer will be consulted if needed.

Evaluation of certainty of evidence

FS and HL will evaluate the evidential strength of each study endpoint independently using the Grading of Recommendations Assessment, Development, and Evaluation framework.40 The certainty of evidence will be classified as high, moderate, low or extremely low. Any disagreement will be solved by discussion and a third researcher (YG) will be consulted if needed.

Assessment of similarity and consistency

Similarity and consistency are two main factors that should be evaluated to obtain valid and credible results. Similarity will be assessed according to clinical and methodological characteristics owing to the challenges in clarifying similarity by statistical analysis. Study designs, participant characteristics and interventions will be assessed. Local inconsistency will be evaluated using the node-splitting analysis. The value of p>0.05 is not considered statistically significant indicating that it is consistent with the direct and indirect comparison. The value of p<0.05 is considered statistically significant indicating inconsistency. A consistency model or inconsistency model will be chosen based on the results. A potential scale reduced factor (PSRF) will be used to determine convergence. PSRF close to 1 indicates convergence.

Selective outcome reporting will be rated with regard to the two primary outcomes (hyperandrogenism, and recruitment and retention rates). The risk of bias will be considered low if the number of respondents and the number of total dropouts are reported. The risk of bias will be considered high if neither the number of respondents nor the number of total dropouts was reported. All other cases will be classified as unclear risk of bias.

Loss to follow-up is typically associated with the outcome and the intervention received.41 Lifestyle interventions often require long-term adherence, and so some dropouts are typically unavoidable. Inappropriate methods to impute data such as the LOCF (Last Observation Carried Forward) approach are often applied and are known to produce biased results.42 However, even application of appropriate methods such as multiple imputations in practice often results in a random assumption. Consequently, the studies concerning attrition bias will be classified as (1) Low risk if an appropriate imputation method was employed that accounts for the different reasons for dropouts between arms (especially in placebo-controlled trials, where the lack of active comparator can affect the dropout), or if the missing data were 20% or less and were balanced between arms; (2) High risk if dropout was unbalanced between the arms, and an inappropriate imputation method such as LOCF was used to impute dropouts; (3) All other cases will be classified as unclear risk.

Statistical synthesis

Characteristics of included studies and information flow in the network

Descriptive statistics will be conducted to describe author, year of publication, sponsorship, study population and clinical settings.

The evidence inventory will be exhibited by a network diagram. Information presented in the diagram will be distinguished by the size of the nodes (presenting the total number of participants) and the thickness of the links or edges (showing the number of studies comparing two interventions). The contribution matrix will be used to discern the network’s most impactful comparisons and how direct and indirect evidence influences the final summary data by detailing the percentage contribution of each direct comparison to the aggregate evidence.43

Pairwise meta-analysis

Stata software V.15.0 (Stata Corp) will be used for the data analysis. Heterogeneity between studies will be measured with I2. I250% indicates that heterogeneity is acceptable, and a fixed-effects model will be chosen. I 2>50% indicates that heterogeneity is evident, and a random-effects model will be selected.

Network meta-analysis

Bayesian network analysis will be conducted to synthesise data.44 Stata software V.15.0 will be used to compare the outcomes of included study and forest plots will be generated to present the NMA results.45 The rank of various lifestyle interventions will then be generated. Comparisons between interventions will be presented as a network plot. Rank plots will be used to present the contribution of different designs to the final effect size of the NMA. Lifestyle interventions will be ranked based on the value of p, which determines the extent of certainty when the intervention group is superior compared with the control group. A p value of 100% indicates that the treatment is more effective than the control whereas a p value of 0% indicates that the treatment is less effective compared with the control.

Subgroup analysis, meta-regression analysis and sensitivity analysis

Subgroup and meta-regression analyses will be carried out to identify the potential sources of variability and inconsistencies. If compatible data exist, the data will be separated according to the type of lifestyle interventions. Meta-regression will be conducted to examine PCOS duration, intervention dosage, obesity levels, age and origin of participants. Furthermore, sensitivity analysis will be executed by gradually excluding studies to verify the robustness of the outcomes.

Publication bias assessment

A comparison-adjusted funnel plot will be generated to detect the reporting bias if more than 10 studies are included.46

Patients and public involvement

None.

Ethics and dissemination

Ethical approval will not be required because no data will be collected from human participants. The results will be published in a peer-reviewed journal and disseminated at relevant conferences.

Ethics statements

Patient consent for publication

References

Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

Footnotes

  • FS and HL contributed equally.

  • Contributors FS and HL designed the study and drafted the manuscript. FS, HL, YG, YL, HY revised the protocol and assisted with the study design. FS and YL contributed to the initial search strategy. FS and YG provided input on the protocol and designed the analysis plan. ZL, YS and JY critically revised the manuscript for important intellectual content and assisted with the study design. All authors have read and approved the final manuscript. ZL and YS are the study guarantors.

  • Funding This work was supported by the National Natural Science Foundation of China (No. 82272383 and 32070189) and Hunan Natural Science Foundation (2024JJ5349).

  • Competing interests None declared.

  • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.