Anti‐obesity pharmacological agents for polycystic ovary syndrome: A systematic review and meta‐analysis to inform the 2023 international evidence‐based guideline

This systematic review and meta‐analysis evaluated the efficacy of anti‐obesity agents for hormonal, reproductive, metabolic, and psychological outcomes in polycystic ovary syndrome (PCOS) to inform the 2023 update of the International Evidence‐based Guideline on PCOS. We searched Medline, EMBASE, PsycInfo, and CINAHL until July 2022 with a 10‐year limit to focus on newer agents. Eleven trials (545 and 451 participants in intervention and control arms respectively, 12 comparisons) were included. On descriptive analyses, most agents improved anthropometric outcomes; liraglutide, semaglutide and orlistat appeared superior to placebo for anthropometric outcomes. Meta‐analyses were possible for two comparisons (exenatide vs. metformin and orlistat + combined oral contraceptive pill [COCP] vs. COCP alone). On meta‐analysis, no differences were identified between exenatide versus metformin for anthropometric, biochemical hyperandrogenism, and metabolic outcomes, other than lower fasting blood glucose more with metformin than exenatide (MD: 0.10 mmol/L, CI 0.02–0.17, I2 = 18%, 2 trials). Orlistat + COCP did not improve metabolic outcomes compared with COCP alone (fasting insulin MD: −8.65 pmol/L, −33.55 to 16.26, I2 = 67%, 2 trials). Published data examining the effects of anti‐obesity agents in women with PCOS are very limited. The role of these agents in PCOS should be a high priority for future research.

interaction of altered hypothalamic-pituitary-ovarian function and concomitant hyperinsulinemia/insulin resistance with promotion of androgen excess underlie the pathophysiology of PCOS. 2 Characterized by heterogeneous features including reproductive, psychological, and metabolic sequelae, PCOS has been diagnosed applying the Rotterdam Criteria over the past two decades requiring two of three features (hyperandrogenism, ovulatory dysfunction, and/or polycystic ovarian morphology), after excluding other mimicking conditions. 3nce this time in both 2018 and 2023, diagnosis has been upgraded from consensus-based Rotterdam criteria to International evidencebased Guideline PCOS criteria, endorsed by 40 Societies internationally.This differentiated adolescent and adult criteria defined each component of the criteria and included Anti-Mullerian Hormone as an alternative to ultrasonography for determining polycystic ovarian morphology. 4Due to the increased risk for obesity, diabetes, metabolic pregnancy complications, cardiovascular disease, 5 and sleep apnea among individuals with PCOS, prevention of adverse metabolic consequences is crucial. 6e association between obesity and PCOS is complex and bidirectional. 7Obesity genes are noted on genetic studies in PCOS, and cluster analyses, alongside epidemiological and longitudinal studies, show that obesity is increased in PCOS, is causal of PCOS, and exacerbates PCOS clinical features and as such is common in women with PCOS presenting to the clinic. 8,91][12] However, lifestyle modifications, including diet and exercise, are challenging to maintain and often insufficient to lead to meaningful weight loss. 13armacotherapy is recommended in the general population, as an adjunct to lifestyle approaches to optimize weight loss success and efficacy in obesity.Glucagon-like peptide 1 receptor agonists (GLP-1 RAs) including liraglutide, exenatide, or semaglutide are indicated for both weight loss and type 2 diabetes treatment.Predominantly studied in individuals with BMI ≥ 30 kg/m 2 , these agents have been shown to promote weight loss by a variety of mechanisms, including suppression of post prandial glucagon, inhibition of glucose production, slowed gastric emptying, and increased satiety to reduce food intake. 14Orlistat is a long acting reversible pancreatic lipase inhibitor that is designed to reduce absorption of digestive fat, increasing fecal fat excretion. 15Phentermine and topiramate are central acting medications that have been approved in some countries for management of obesity based on their appetite reducing properties, which occur through their effects on γ-Aminobutyric Acid (GABA) receptors and increasing norepinephrine in the hypothalamus. 16Centrally acting anti-obesity agents, including naltrexone/ bupropion and locarserin, target hypothalamic brain signaling as well as dopamine 17 and serotonin receptors, 18 respectively, to diminish food intake.0][21] The need for alternative pharmacotherapies as adjunctive treatments to promote weight loss in PCOS is a key priority with weight gain a primary concern expressed by those with the condition.Although none of the antiobesity agents outlined here have been approved for PCOS alone, their effects on weight loss and insulin resistance make them potentially important future PCOS therapies.
To clarify their potential utility in PCOS, and in the context of informing the 2023 update of the International evidence-based PCOS Guideline, the aim of this systematic review and meta-analysis was to evaluate the efficacy of anti-obesity pharmacological agents alone, or in combination, for the management of hormonal, reproductive, metabolic, and psychological outcomes in adolescents and adults with PCOS, with a focus on newer anti-obesity agents.

| Study design
This systematic review protocol was prospectively registered with PROSPERO (CRD42022347314) and followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. 22The review was intended to inform recommendations for the following clinical question: "Are anti-obesity pharmacological agents alone or in combination, effective for management of hormonal and clinical PCOS features and weight in adolescents and adults with PCOS?"This clinical question was prioritized by consumers, content experts within the guideline development group, and the expert evidence synthesis team, who devised eligibility criteria using the patient, intervention, comparison, and outcome (PICO) framework outlined below.

| Search strategy and selection criteria
The search strategy and selection criteria were developed by an international team of evidence synthesis experts and clinical leads (key contacts) including endocrinologists and a general practitioner (family physician).We searched Medline (OVID), EMBASE (OVID), PsycInfo (EBSCO), and CINAHL (EBSCO) on July 22, 2022 with a limit set for 10 years in order to focus the review on newer agents.We searched reference lists of relevant systematic reviews, and the key contact team reviewed the final list of included randomized controlled trials (RCTs) to ensure no important trials were missing.The search strategy included terms for PCOS, anti-obesity medications, and RCTs (see Appendix A).Citations were imported into Covidence, 23  (2) intervention: anti-obesity pharmacological agents (including, but not limited to, orlistat, GLP-1 RAs, phentermine/topiramate, lorcarserin, or naltrexone/bupropion), provided for a minimum of 3 months, alone or in combination with lifestyle, metformin, the combined oral contraceptive pill (COCP) or anti-androgens; (3) comparison: placebo or any other intervention listed in the intervention or combinations of those listed in the intervention; (4) outcomes: hormonal, metabolic, lipids, psychological, or anthropometric outcomes, and adverse effects (see Appendix B for full list of eligible outcomes).Only RCT designs were eligible for inclusion, and crossover trials were included only for the phase before the crossover.Quasi-randomized trials, conference abstracts, and any trials not published in English were excluded.

| Integrity assessment
Trial integrity was assessed by the Research Integrity Team following the "Research Integrity in Guideline Development (RIGID)" framework developed by Mousa et al. (2023; unpublished), as detailed in Section 6.7 of the guideline technical report. 24Here, studies were assessed using the Trustworthiness in Randomised Controlled Trials (TRACT) checklist, 25 an integrity assessment tool based on the Cochrane Research Integrity Assessment tool, 26 which classifies studies on multiple domains related to integrity.Following this process, studies were classified as low, moderate, or high risk for integrity concerns.Low-risk studies were included, and authors for moderate-and high-risk studies were contacted to clarify integrity concerns.Where a satisfactory response was received, those studies were subsequently "included."Studies with no response were "not included," whereas studies requiring additional time to provide the necessary information (e.g., raw data and ethics protocols) are "awaiting classification" and have not been included in the review or analysis at this stage.

| Quality appraisal
Risk of bias was assessed at the study-level (i.e., for each trial) independently by two reviewers (AG, CE, SG, VR, JL), with disagreements resolved by discussion.We used the Cochrane Risk of Bias 1 tool to assess random sequence generation (selection bias), allocation concealment (selection bias), blinding of participants and personnel (performance bias), blinding of outcome assessment (detection bias), incomplete outcome data (attrition bias), selective reporting (reporting bias), and other biases.
The quality of the evidence at the outcome-level was assessed using the Grading of Recommendations, Assessment, Development and Evaluations (GRADE) approach by CE, as outlined in the GRADE handbook. 27The evidence can be downgraded from "high certainty" by one or two levels for serious or very serious limitations, respectively, for each of four main domains: risk of bias, indirectness of evidence, inconsistency and imprecision of effect estimates, and for other biases including potential publication bias.

| Data analysis
The following outcomes were rated as critical: modified Ferriman-Gallwey (mFG) score, free androgen index (FAI), homeostatic model assessment of insulin resistance (HOMA-IR), 2-h glucose after 75-g oral glucose tolerance test (OGTT), and BMI.The remaining outcomes were judged as important but not critical.Outcome data were extracted from original intention-to-treat results wherever possible or from per-protocol results if these were the only outcomes available.
For trials that used the same assessment method and provided continuous data, we reported mean difference (MD) and 95% confidence intervals (CIs), converting units of measurement to standardized units where required.Where trials did not use the same assessment methods, we reported standardized mean differences (SMD) and 95% confidence intervals.Heterogeneity was assessed using the I 2 statistic.Outcomes from individual studies were pooled using random-effects models.All statistical analyses were performed using Review Manager. 28We had planned subgroup analyses to separate those who were in the post-menopausal stage, adolescents versus adults, and by BMI category but were unable to conduct these due to the small number of included trials.

| Study characteristics
The PRISMA flowchart of study selection is presented in Figure 1.A total of 782 citations were identified, with 675 remaining after duplicates were removed.After title and abstract screening, 647 were excluded, and 28 full-text manuscripts were screened for eligibility.

| Participants
Eligibility criteria were adults with PCOS with a BMI in the overweight range or above with the exception of Nylander et al., 43 who enrolled individuals who either had a BMI ≥ 25 kg/m 2 and/or insulin resistance (defined as fasting plasma C-peptide >0.6 nmol/L at screening).
Overweight was generally classified as BMI ≥ 25 kg/m 46 or BMI ≥ 24 kg/m 2 , 31,35,40 and one study did not define the BMI cut-off for overweight. 30 One stdy 31 required a concurrent diagnosis of insulin resistance, defined as fasting insulin >10 mIU/L, and another 46 required a diagnosis of prediabetes (defined as fasting plasma glucose 5.6-6.9 mol/L and/or 2 h post glucose 7.8-11.0 mmol/L o OGTT). 46One study 39,40 enrolled individuals who had also been diagnosed with infertility due to PCOS. Wefound no studies on adolescents.

| Outcomes
1][42]46,47 Four studies collected glucose 2 h post 75-g oral glucose. 30,40,41,47 With he exception of a single study, 47 all studies collected data on biochemical hyperandrogenism.Only one study 30 collected data on clinical hyperandrogenism (hirsutism).The most frequently collected outcomes were total testosterone (TT), [29][30][31][32]35,[40][41][42][43]46 FAI, [29][30][31][32]40,43,46 sex hormone binding globulin (SHBG), 30,31,35,40,43,46 and dehydroepiandrosterone sulfate (DHEAS).[29][30][31][32]41,46 All studies collected data on lipid profile, three studies collected data on highly sensitive C-reactive protein (hsCRP), 30,40,41 and one study collected C-reactive protein (CRP) values.35 No studies collected data on quality of life or psychological outcomes. Five studies collcted reproductive outcomes, specifically menstrual regularity 30,32,40,43 and clinical pregnancy rate. 31,40 Liu et al. als collected data on pregnancy complications.31,40 All but one study 35 reported on gastrointestinal (GI) adverse events (AEs), and seven reported on other AEs.[29][30][31][41][42][43]46 Duration of follow-up ranged from 12 to 32 weeks, and Liu et al. 40 followed up pregnancy outcomes for up to 64 weeks.The majority of trials were at unclear risk of selection bias, mainly due to failure to specify if or how allocation was concealed. More than h of the trials were at high risk of performance bias due to lack of blinding of participants and personnel.Three quarters of the trials were at unclear risk of detection bias, and more than half were at high or unclear risk for reporting bias.Greater than a quarter of trials were at high risk of other biases, mainly due to conflicts of interest.

| Effects of interventions
We report on a total of 12 comparisons.Meta-analyses were conducted for the following two comparisons: (1) exenatide v metformin (2 RCTs) 30,46 and (2) orlistat + lifestyle + COCP v lifestyle + COCP (2 RCTs). 31,35For the remaining comparisons, a narrative synthesis was provided as a meta-analyses were not possible on any outcomes either due to the comparison only having one representative RCT or RCTs reporting non-parametric data (median and interquartile range/ IQR) or change scores without any information on standard deviation or standard error.

| Exenatide versus metformin
Exenatide 20 μg/day versus metformin 1.5-2 g/day Two trials were included in this comparison. 30,46Sample sizes were 100 46 and 63 30 per study for the two arms.Both trials enrolled individuals who were overweight/obese.Participants in the study by Tao et al. 46 also had a concurrent diagnosis of prediabetes.Both trials provided exenatide up to 20 μg daily for 12 weeks, together with diet and exercise guidance without active lifestyle interventions.The metformin dose was 1.5-2 g daily in one study 46 and 2 g daily in the other. 30See Table 2 for a summary of GRADE assessments for this comparison.
The oral contraceptive pill (cyproterone acetate and ethinyl estradioldose not specified) was provided in both groups in the first 12 weeks.
In descriptive analysis of this trial, 40   In relation to reproductive outcomes, the ratio of actual menses to expected menses was higher in the intervention group compared with control (0.90 ± 0.13 vs. 0.68 ± 0.03, p < 0.001) at 12 weeks post-treatment.After 12 weeks, the natural pregnancy rate was higher in the exenatide + COCP group (43% vs. 18.7% in metformin + COCP, p < 0.05), but there were no differences for total pregnancy or live birth rates at 64 weeks follow-up.There was no difference between groups for risks of miscarriage, preterm delivery, gestational diabetes mellitus, gestational hypertension, or fetal macrosomia.GI AEs were more frequent in the exenatide + COCP group compared with metformin + COCP.

| Exenatide + metformin versus metformin alone
Exenatide (10-20 μg/day) + metformin (1.5-2 g/day) versus metformin alone Tao et al. 46 also reported on the comparison of a combination of exenatide (10-20 μg daily) and metformin (1.5-2 g daily) versus metformin alone (1.5-2 g daily).AEs were only reported in the context of reason for withdrawal.In this study, there were no between-group differences for weight, BMI, or any of the androgen levels collected (FAI, TT, SHBG, DHEAS, androstenedione).There were no differences in metabolic markers including FBG, FINS, HOMA-IR, or lipids (TC, HDL, TG), with the exception of metformin being superior to   p < 0.001). 33There was no difference between groups for percentage body fat.Liraglutide was superior to placebo for FBG (mean reduction of 0.24 mM, 95% CI 0.05-0.43,p < 0.05 in the liraglutide group compared with placebo) but not for the Matsuda index.There were no between-group differences for lipids including TC, TG, HDL, and LDL or hsCRP.
There were no differences between groups for fat free mass.

| Semaglutide versus placebo
Semaglutide (1 mg weekly) versus placebo One trial (n = 25) 47 reported on the comparison of semaglutide 1 mg weekly versus placebo for 16 weeks.There was very serious imprecision (single trial, very small sample size), and the study was rated at moderate risk of bias due to potential conflict of interest and selective outcome reporting.There was more nausea in the semaglutide group compared with placebo.

| Orlistat versus placebo
Orlistat (120 mg three times/day) + lifestyle versus placebo + lifestyle One trial (n = 86) reported on the comparison of orlistat 120 mg three times per day versus placebo for 3 months. 42Both groups were prescribed a hypocaloric mono-unsaturated fatty-acid (MUFA) diet of 1200-1800 kcal/day and were encouraged to walk for 30 min daily.
More than half of participants reported urgency to go to the bathroom, and 30% reported oily spotting in undergarments.About one in five participants reported oily or fatty stool.
Orlistat + lifestyle was reported to be superior to placebo  31 Both trials provided all participants with lifestyle interventions (dietician-prescribed personalized balanced nutrition diet 31 or low-fat diet 35 ) and the COCP (drospirenone/EE 35 and cyproterone acetate/EE 31 ).The intervention groups received orlistat 120 mg three times/day in both trials.A small proportion of participants reported GI AEs with orlistat (flatulence and oily spotting).
Meta-analysis was conducted on six outcomes.Orlistat + lifestyle and the COCP were superior to lifestyle and the COCP alone for SHBG (MD 14.30 nmol/L, 2.94-25.66,I 2 = 0%, 2 trials, low certainty evidence, Figure 4A).No between-group differences were evident for metabolic outcomes including FBG (MD À0.00 nmol/L, À0.17 to 0.17, I 2 = 0%, 2 trials, low certainty evidence, Figure 4B), fasting insu-  35 There was no difference between groups for waist circumference. 35Note that numerical data on endpoint weight and BMI were not available from the Song trial as only a figure lacking data labels was displayed.However, Song et al. reported   between-group differences for weight and BMI favoring the intervention group. 31See Table 3 for a summary of GRADE assessments for this comparison.

| DISCUSSION
This meta-analysis considered the effects of anti-obesity agents on hormonal, metabolic, anthropometric, and reproductive outcomes in individuals with PCOS.We found that GLP-1 RAs (exenatide, liraglutide, and semaglutide) have variable weight reduction efficacy within 12 weeks that appears to mostly correspond with metabolic and reproductive benefits in PCOS.Orlistat was superior to the COCP alone for some anthropometric outcomes but not for metabolic outcomes.No differences were observed between exenatide and phentermine/topiramate for anthropometric, biochemical hyperandrogenism, metabolic, and lipid outcomes.Evidence on fertility outcomes was limited to one trial that suggested increased pregnancy rates with pregestational exenatide compared with metformin; however, there was no difference between groups for live birth rate.Of note, all agents consistently resulted in greater AEs than controls, including metformin.
The interest in using GLP-1 RAs in women with PCOS has increased, due to general efficacy in weight reduction.Previous narrative reviews supporting the use of GLP-1 RAs in PCOS focused on single-center RCT and observational studies. 49Two meta-analyses of GLP-1 RAs versus metformin in PCOS included six and eight RCT studies, respectively, 50,51 which concluded GLP-1 RAs were beneficial with or without metformin for metabolic, reproductive, and anthropometric parameters.However, these analyses combined exenatide and  network meta-analysis on anti-obesity agents for PCOS and found that liraglutide was most effective for lowering BMI.Our results differed from these findings as we evaluated the effect of individual anti-obesity medications against a comparator, whereas Wang et al.
combined medication classes and varying doses. 52Furthermore, some studies in our review are awaiting classification after an integrity check aligned to Cochrane processes. 25Overall, these agents show promise, based on general population data and emerging data in PCOS; however, inadequate quality trials have culminated in only low to very low certainty evidence with research now a major priority.
Our review found that exenatide used twice daily did not confer significant benefits over metformin and was associated with more AEs.Clinical outcomes in PCOS have been postulated to be correlated with degree of weight loss.Because our analysis failed to find clear differences in anthropometric outcomes, no differences in metabolic and biochemical parameters were anticipated or identified.This observation is consistent with studies in non-PCOS populations, where twice daily exenatide was compared with longer acting GLP1 RA.Exenatide twice daily induced modest changes in body weight ranging from +0.3 to À2.96 kg. 53In contrast, liraglutide led to larger weight reduction, ranging from 0.3 to 3.38 kg, and semaglutide 1 mg led to greater weight loss again ranging from À3.47 to À6.5 kg.Concordantly, we have shown that clinical impacts were more likely to be observed with liraglutide and semaglutide, in PCOS.
We found that although liraglutide was superior to placebo for anthropometric outcomes, liraglutide alone (without lifestyle co-interventions) resulted in more lean body mass loss than placebo.When liraglutide was delivered with lifestyle co-interventions, there was no difference between groups for lean body mass.This finding supports that physical activity (specifically resistance training) should be part of a recommended approach to preserve lean body mass and promote weight maintenance alongside the use of GLP-1 RAs. 54though metformin was superior to exenatide for lowering fasting glucose concentration, mean fasting glucose concentrations were less than 5.6 mmol/L for both interventions.Whether the minimal difference in fasting glucose between metformin and exenatide contributes to future metabolic co-morbidities is unknown.The associated glycated hemoglobin reduction noted with twice daily exenatide was less than that seen with long-acting agents.Longer acting GLP-1 RA medications have advantages with improved adherence, glycaemic effects, and tolerance. 53,55mitations of these studies include the rapid evolution of the GLP-1 drug class with limited studies with each agent and with variable doses.As with many areas of PCOS, funding is limited and quality trials woefully inadequate given the prevalence and impact of the condition. 56Many of the studies included suboptimal liraglutide doses; the 3-mg dose has been shown to optimize weight loss. 52Many of the studies used a 12-week protocol that limits the ability to demonstrate changes in important clinical outcomes such as hirsutism and fertility.Longer studies allowing for effect of full dose GLP-1 RA medication may also lead to more substantial benefits in this population.
Additionally, semaglutide is the most potent long acting GLP-1 RA 57 but has only been studied in one small pilot study in patients with PCOS showing benefits compared with placebo for anthropometric measures (including visceral body fat), with no reproductive outcomes and some but not all metabolic and lipid parameters showing only modest benefits that are unlikely to be clinically significant.More high quality, multicenter studies of semaglutide in PCOS are urgently needed, incorporating reproductive, metabolic, and psychological outcomes, in addition to anthropometric outcomes.Further, with the FDA approval of a new dual acting GLP1-RA along with gastric inhibitory polypeptide receptor activator (tirzepatide), and the prospect of additional newer agents on the horizon, medical weight management in those with PCOS will be continue to be an area of interest.
For other agents, a previous meta-analysis comparing orlistat and metformin in patients with PCOS, Graff et al. 58 reported benefits of orlistat for weight reduction and reduction in HOMA-IR, insulin, and testosterone (2 RCTs).Our review included two RCTs with orlistat, with or without lifestyle or the COCP, being superior to lifestyle or the COCP alone for some outcomes.Orlistat had high AEs as similarly reported by Graff et al. 58 Data regarding the efficacy of phenterminetopiramate in those with PCOS were limited with no superiority over exenatide for any reported outcomes and more AEs.No other evidence was found on other anti-obesity agents.However, their use is of interest, based on evidence from the general population with obesity, where naltrexone/bupropion and lorcaserin have each led to weight loss of up to 5% at 12 weeks. 59,60Overall, those with PCOS and their healthcare professionals need to consider both evidence of potential benefits and AEs in shared decision making on the use of anti-obesity for weight loss in PCOS.
The strengths of this study include the rigorous design and conduct.Wherever possible, we have conducted meta-analyses on individual agents such as individual GLP-1 RAs rather than pooled agents.
We excluded trials of uncertain integrity to reduce potential erroneous conclusions, and as such, the validity and trustworthiness of our results is strengthened.An extensive global prioritization exercise identified the need for this review, which was conducted by a multidisciplinary team.Limitations are that only published studies, available in English, were included, and due to resource and time limitations, grey literature was not searched.Despite little to no integrity concerns identified in the included studies, the quality of these studies (in terms of risk of bias) and small sample sizes decreased the level of certainty of the evidence presented.Data on adolescents were not available, and due to time limitations, we did not include searches of grey literature or clinical trial registries.

| CONCLUSION
Our findings support the need for further investigations of antiobesity agents in PCOS.On the basis of our analyses, we cannot provide definitive recommendations at this time due to the small number of trials, short follow-up periods, and overall high or unclear risk of bias in the majority of trials.Given the association of metabolic and reproductive benefits that appear to have a dose response with degree of weight loss, anti-obesity medications including liraglutide, semaglutide, GLP-1 RAs, and orlistat could be considered, in addition to active lifestyle intervention, for the management of higher weight in adult women with PCOS as per general population guidelines.
Weight management is an important outcome for those with PCOS, and further studies in this area need to be prioritized.In particular, the need for placebo-controlled trials is urgent.With increasing popularity but limited initial data, more trials to assess the efficacy of these wherein duplicates were removed.Two reviewers (AG, CE) independently screened titles and abstracts and then full-text manuscripts using Covidence web-based software, and disagreements were resolved by discussion, with a third reviewer to adjudicate if needed.Data were extracted on study characteristics, participant characteristics at baseline, intervention, and outcomes, using a data extraction template created by the guideline evidence team.Data were extracted by independent reviewers (AG, CE, SG, VR, JL) in duplicate, with any disagreements resolved by discussion.Inclusion criteria using the Participants-Interventions-Comparators-Outcomes (PICO) framework were as follows: (1) participants: individuals with PCOS diagnosed by Rotterdam, original National Institutes of Health (NIH) or Androgen Excess and Polycystic Ovary Syndrome society (AE-PCOS) criteria of any age, ethnicity or weight;

Figure
Figure 2A,B summarizes the assessed risk of bias of the included trials.
was no difference between groups for FBG or lipids (TC, low-density lipoprotein cholesterol/LDL-C, high-density lipoprotein cholesterol/ HDL-C, TG) or for any biochemical hyperandrogenism outcomes including TT, FAI, or SHBG.

F I G U R E 3
Forest plots of exenatide v metformin.(A) Fasting glucose, (B) fasting insulin, (C) homeostatic model assessment of insulin resistance (HOMA-IR), and (D) 2-h insulin.
agents are needed, particularly for GLP-1 RAs given their promising benefits and minor AEs.Future research should also examine weight regain in PCOS following cessation of anti-obesity agents and evaluate the impact of anti-obesity agents on quality of life and clinical hyperandrogenism.Longer follow-up periods are also required to demonstrate meaningful clinical benefits.With PCOS currently impacting approximately 10% of reproductive-aged women, this research should be designated as high priority.

Table 1
Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flowchart of study selection.
Risk of bias.(B) Risk of bias summary.T A B L E 2 GRADE assessment-Exenatide v metformin.