Combined contraceptives, including oral contraceptive pills, are commonly prescribed for adults and adolescents with PCOS to ameliorate the clinical symptoms and associated hormonal disturbances. The effects of COCPs on menstrual cycle, hirsutism, weight loss, waist/hip ratio, testosterone concentrations, lipid profile and blood sugar levels are

variably reported and depend on type of COCP used, duration of use, severity of presentation/phenotype, adherence to the regimen, among other factors. Different combinations of COCPs are available with heterogeneous estrogen and progestin preparations with varying pharmacological and clinical properties. Thus, the efficacy and consequences of COCPs in PCOS may vary. Some preparations also comprise natural estrogen instead of synthetic ethinyloestradiol (EE) with benefits and contraindications considered similar.

International Guideline group on PCOS (2018)

Clinical question

Is the combined oral contraceptive pill (COCP) alone or in combination effective for management of hormonal and clinical PCOS features in adolescents and adults with PCOS?

Summary of literature

Summary of systematic review evidence – COCP alone

Research evidence - ADOLESCENTS

COCP versus placebo

One randomised controlled trial (RCT) was identified to address this comparison in adolescents (Hoeger, 2008). There was a statistically significant improvement with COCP (compared to placebo) for high-density lipoprotein (HDL) in this very low quality study with low certainty. No statistically significant differences were found for outcomes: body mass index (BMI) (kg/m2); Waist (cm); Total testosterone (ng/dl); sex hormone-binding globulin (SHBG) (nmol/liter); free androgen index (FAI); Hirsutism (FG score); Total cholesterol (mg/dl); LDL (mg/dl); Triglycerides (mg/dl); Fasting insulin (IU/ml); Fasting blood sugar (mg/dl); CRP (mg/l); PAI-1. Side effects were not reported.

COCP versus lifestyle

One RCT was identified to address this comparison in adolescents (Hoeger, 2008). There was a statistically significant improvement with lifestyle (compared to COCP) for low-density lipoprotein (LDL) in this very low quality study with very low certainty. No statistically significant differences were found for: BMI (kg/m2); Total testosterone (ng/dl); SHBG (nmol/liter); FAI; Hirsutism (FG score); Total cholesterol (mg/dl); HDL (mg/dl); Triglycerides (mg/dl); Fasting insulin (IU/ml); Fasting blood sugar (mg/dl); CRP (mg/l); PAI-1. Side effects were not reported.

COCP versus metformin

A systematic review including four RCTs that address this comparison in adolescents was identified (Al Khalifah, 2016). The evidence team conducted additional analysis of outcomes not addressed in the systematic review. While a statistically significant improvement was found in BMI and LDL with use of metformin over COCP; and a statistically significant improvement was found in menstrual regulation with use of COCP over metformin, we remain cautious due to very low certainty in effect estimates and the quality of evidence. A statistically significant improvement in dysglycemia (OGTT) was found with the use of metformin over COCP, however it should be noted that there is low certainty in the effect estimates and the quality of evidence. No statistically significant differences were found for: Hirsutism; Total Testosterone (nmol/L); Triglyceride (mg/dL); Total Cholesterol (mg/dL); HDL (mg/dL); Weight (kg); Fasting insulin; SHBG; FAI; Fasting blood sugar (mg/dL); CRP (mg/L); PAI-1. Side effects included weight gain with COCP; and side effects were not specified with metformin.

COCP versus metformin + anti-androgen

One RCT was identified to address this comparison is adolescents (Ibanez, 2004). Due to the lack of direct comparisons between groups (no p values reported for between groups) it is uncertain whether there were any differences in this very low quality study with very low certainty for outcomes: BMI (kg/m2); Hirsutism (FG score); Glucose/insulin ratio; SHBG (ug/dl); Testosterone (ng/dl); Triglycerides (mg/dl); HDL (mg/dl); LDL (mg/dl); Cycle regularity; Weight. Side effects were not reported.

Research evidence – ADULTS

COCP versus metformin

Nine RCTs were identified to address this comparison (Dardzinska, 2014; Harborne, 2003; Kilic, 2011; Meyer, 2007; Morin-Papunen, 2003; Morin-Papunen, 2000; Moro, 2013; Rautio, 2005; Wu, 2008). There were statistically significant improvements with metformin (compared with COCP) for fasting insulin, including for both BMI subgroups. Metformin improved HDL in the BMI > 25 subgroup but not in the BMI < 25 subgroup or when all participants were combined; and improved triglycerides when all participants were combined, in the BMI > 25 subgroup and in the subgroup where BMI was not defined, but not in the BMI < 25 subgroup. There were statistically significant improvements with COCP (compared with metformin) for SHBG, FAI, total testosterone and irregular cycles, including for all BMI subgroups. COCP improved LDL in the BMI > 25 subgroup but not in the BMI < 25 subgroup or when all participants were combined. No statistically significant differences were found for: Weight; Clamp (M value); homeostatic model assessment (HOMA) (change from baseline); BMI (kg/m2); waist-hip-ratio (WHR); Hirsutism [FG score]; Fasting glucose [mmol/l], Total cholesterol [mmol/l]. Metformin use increased GI-related events, whereas the COCP group had none. While one of the included studies was of moderate quality and certainty, the majority of studies in these meta-analyses were of low to very low certainty in effect estimates and the quality of evidence and therefore all findings should be interpreted with caution.

COCP versus COCP + metformin

Six RCTs were identified to address this comparison (Cibula, 2005; Elter, 2002; Essah, 2011; Feng, 2016, Moro, 2013; Wu, 2008; ). There was a statistically significant improvement with COCP alone (compared with COCP plus metformin) for triglycerides. COCP alone improved SHBG in the BMI < 25 subgroup but not in the BMI > 25 subgroup or when all participants were combined. There were statistically significant improvements with COCP plus metformin (compared with COCP alone) for FAI. COCP plus metformin improved testosterone when all participants were combined but not in BMI subgroups; improved hirsutism and fasting glucose when all participants were combined and in the BMI > 25 subgroup but not in the BMI < 25 subgroup; improved SHBG and fasting insulin in the BMI < 25 subgroup but not when all participants were combined or in the BMI > 25 subgroup; and improved total cholesterol in the BMI > 25 subgroup but not when all participants were combined or in the BMI < 25 subgroup. No statistically significant differences were found for: HOMA; Weight (kg); BMI (kg/m2); WHR; HDL [mmol/l]; LDL [mmol/l]. The addition of metformin to COCP increased GI-related events, whereas the COCP alone group had none. While one of the included studies was of moderate quality and certainty, the majority of studies in these meta-analyses were of low to very low certainty in effect estimates and the quality of evidence and therefore all findings should be interpreted with caution.

COCP versus COCP + metformin + anti-androgen

One RCT was identified to address this comparison in adults (Ibanez, 2004). Due to the lack of direct comparisons between groups (no p values reported for between groups) it is uncertain whether there were any differences in this very low quality study with very low certainty for outcomes: BMI (kg/m2); Hirsutism (FG score); Glucose/insulin ratio; SHBG (ug/dl); Testosterone (ng/dl); Triglycerides (mg/dl); HDL (mg/dl); LDL (mg/dl); Cycle regularity; Weight. Side effects were not reported.

COCP versus anti-androgen

One RCT was identified to address this comparison is adults (Spritzer, 2000). Due to the lack of direct comparisons between groups (no p values reported for between groups) it is uncertain whether there were any differences in this very low quality study with very low certainty for the outcome hirsutism (FG score).

Summary of systematic review evidence – COCP combined with other agents

Research evidence - ADOLESCENTS

COCP + metformin+ lifestyle versus COCP + lifestyle + placebo

One RCT was identified to address this comparison is adolescents (Hoeger, 2008). There was a statistically significant improvement with the addition of metformin to COCP and lifestyle (compared to COCP and lifestyle plus placebo) for testosterone and HDL in this very low quality study with very low certainty. No statistically significant differences were

found for: BMI (kg/m2); Waist (cm); SHBG (nmol/l); FAI; Hirsutism (FG score); Total cholesterol (mg/dl); LDL (mg/dl); Triglycerides (mg/dl); Fasting insulin (IU/ml); Fasting blood sugar (mg/dl); CRP (mg/liter). One in each group stopped metformin or placebo due to GI effects.

COCP + anti-androgen versus COCP + anti-androgen + metformin

One RCT was identified to address this comparison in adolescents (Ibanez, 2005). Due to the lack of direct comparisons between groups (no p values reported for between groups) it is uncertain whether there were any differences in this very low quality study with very low certainty for outcomes: BMI (kg/m2); Fasting glucose/insulin ratio SHBG (μg/dl); Testosterone (ng/dl); LDL (mg/dl); HDL (mg/dl); Triglycerides (mg/dl). Side effects were not reported.

COCP versus COCP + anti-androgen

Four RCTs were identified to address this comparison in adults (Hagag, 2014; Leelaphiwat, 2015; Tartagni, 2000; Vieira, 2012). There was a statistically significant improvement with COCP alone (compared with COCP plus anti-androgen) for BMI and LDL. No statistically significant differences were found for: Weight (kg); WHR; Hirsutism (FG score); FAI (%); Testosterone (nmol/L); SHBG [nmol/l]; Fasting insulin (uIU/ml); Fasting glucose [mmol/l]; Total cholesterol [mmol/l]; HDL [mmol/l]; Triglycerides (mg/dL); HOMA; CRP (mg/l); Headache; Breast-related side effects; Vomit/Nausea; Minor depressive state; Liver function. The majority of studies in these meta-analyses were of low to very low certainty in effect estimates and the quality of evidence and therefore all findings should be interpreted with caution.

COCP (DRP-containing COCs) + metformin + lifestyle versus COCP (CPA-containing COCs) + metformin + lifestyle

One study was identified to address this comparison in adults (Wang, 2016). There were no statistically significant differences reported between the two interventions (differing by the combination in the COCP) in this very low quality study of very low certainty for outcomes: WHR; Fasting plasma glucose (mmol/L); HbA1c (%); Total cholesterol (mmol/L); LDL (mmol/L). Side effects were not reported.

Summary of narrative review evidence

Evidence on COCP use from the general population also informed recommendations. Consideration of adverse effects is needed before prescribing COCPs. Absolute contraindications for COCP use according to world health organisation (WHO) include a history of migraine with aura, deep vein thrombosis (DVT)/pulmonary emboli (PE), known

thrombogenic mutations, multiple risk factors for arterial cardiovascular disease, history of ischemic heart disease or stroke, complicated valvular heart disease, breast cancer, neuropathy, severe cirrhosis and malignant liver tumours (World Health Organization, 2015). Other risk factors for DVT need consideration including postpartum immobility, transfusion at delivery, BMI > 30 kg/m2, postpartum haemorrhage, immediately post-caesarean delivery, preeclampsia or smoking. Current evidence suggests that COCPs containing levonorgestrel, norethisterone and norgestimate are associated with the lowest relative risk of DVT. Also, WHO recommends that COCPs with 35 micrograms of EE and cyproterone acetate should only be used when treating moderate to severe hirsutism or acne due to higher DVT risk. For contraception, irregular menstrual cycles and mild to moderate hirsutism, other lower risk preparations are recommended first line (World Health Organization, 2018).

Recommendations

Recommendations – COCP alone

4.2.1

4.2.2

4.2.3

4.2.4

4.2.5

4.2.6

Recommendations – COCP in combination with other agents

4.3.1

4.3.2

4.3.3

4.3.4

4.3.5

4.3.6

Justifications

Although relatively safe, COCPs have absolute and relative contraindications and risks and benefits in the general population that need consideration by health professionals and women.

Although combined metformin and COCP offers additional benefits, these did not surpass the impact of COCP plus lifestyle intervention. Hence the combination is indicated where COCP and lifestyle have failed to meet goals. A combination regime may also lead to increased mild gastrointestinal side effects, which can impact on adherence. Strategies to reduce side effects are available (see metformin recommendations below). With metformin therapy in addition to COCP, women with PCOS and obesity may yield the greatest benefit. The PCOS phenotype, BMI, ethnicity and the informed preference of the individual with PCOS need to be considered when recommending pharmacological agents for the treatment of PCOS.

COCPs, metformin and anti-androgens are off label treatments specifically for treatment of PCOS. However, use is evidence-based for the treatment of clinical features of PCOS and is generally not restricted for use in PCOS. Women should be informed of the benefits and risks and the regulation status of relevant medications. The combination of off label treatments with the COCP is not routine practice and will require education and integration into algorithms. This is anticipated to significantly change practice. Due to subgroup differences in recommendations, the personal characteristics of all women need to be considered.

The results of the summary of literature were adopted from the international guideline PCOS (2018). More information about the methodology of the international guideline PCOS can be found in Chapter 6 Guideline Development Methods (p.124-136) of the guideline, which can be consulted at the ESHRE website (www.eshre.eu/guidelines). The Evidence tables, Table of excluded studies and Literature search strategy can be found under Technical report (supporting documents).