The association between the combined oral contraceptive pill and insulin resistance, dysglycemia and dyslipidemia in women with polycystic ovary syndrome: a systematic review and meta-analysis of observational studies

Ilana J. Halperin 1 Shoba Sujana Kumar 1 Donna F. Stroup 0 Sheila E. Laredo 1 0 Data for Solutions, Inc., PO Box 894, Decatur, GA 30031-0894, USA 1 Department of Medicine, Women's College Hospital,University of Toronto , 1650-111 Elizabeth St. Toronto, Toronto , Ontario,Canada M5S 1B6 background: Polycystic ovary syndrome (PCOS) is the most common endocrine disorder of young women. First-line treatment is often the oral contraceptive pill (OC), but evidence suggests that OC may worsen metabolic outcomes in this population. We undertook this meta-analysis of observational studies and cohorts from within randomized controlled studies to investigate the association between OC use and dysglycemia, dyslipidemia and insulin resistance (IR) in women with PCOS. methods: We searched MEDLINE (1966 - April 2010), EMBASE (1980 - April 2010) and All EBM Reviews. We included prospective cohorts and RCTs that treated women, aged 13 - 44, with PCOS with OC for at least 3 months. Blinded quality assessment and data extraction were conducted on 35 included studies by two independent reviewers. We used random effects methods to calculate weighted mean differences as the effect size. We investigated heterogeneity using sequential removal of studies, subgroup analysis and meta-regression. results: OC use was significantly associated with an increase in high-density lipoprotein cholesterol (HDL-C) (P 0.004) and triglycerides (P 0.004). Significant heterogeneity was found in glucose, cholesterol, HDL-C, low-density lipoprotein cholesterol triglycerides, fasting glucose to insulin ratios and homeostatic model assessments-IR. Study characteristics such as mean BMI, mean age and duration of study could explain some of the heterogeneity. conclusions: Use of OC was not associated with clinically significant adverse metabolic consequences. Because of limitations of the underlying studies, further research including rigorously designed randomized trials would more definitively confirm our findings. Introduction Polycystic ovary syndrome (PCOS) is the leading cause of anovulatory infertility in developed countries (ESHRE/ASRM, 2004). The clinical features of the syndrome include oligomenorrhea, acne, hirsutism, obesity and insulin resistance (IR). IR is present in the majority of women with PCOS, regardless of BMI (Dunaif et al., 1989). Women with PCOS have an increased risk for impaired glucose tolerance (IGT) (Legro et al., 1999). IGT or type two diabetes mellitus (T2DM) develops in .40% of obese women with PCOS by the age of 30 (Ehrmann et al., 1999). Decreased high-density lipoprotein cholesterol (HDL-C) levels, and increased total cholesterol, lowdensity lipoprotein cholesterol (LDL-C) and triglyceride levels are also associated with PCOS (Talbott et al., 1995). & The Author 2010. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. For Permissions, please email: The oral contraceptive pill (OC) has been recommended for .30 years to treat women with PCOS (Ehrmann, 2005). Its beneficial effects include reducing acne, improving hirsutism and correcting oligomenorrhea (Ehrmann, 2005). Less well understood is the effect of OCs on the metabolic risk profile in PCOS. Studies in the general population have linked OCs, composed of 30 50 mg of ethinyl estradiol (EE), to reduced glucose tolerance and IR (Rimm et al., 1992; Watanabe et al., 1994). The link between OC and dysglycemia is unclear with low-dose OCs (ChasenTaber et al., 1997), but the risks reported in the general population may underestimate the risk for women with PCOS. Dyslipidemia has been linked to the progestogen component of OCs (Ball et al., 1990; Van Rooijen et al., 2002). If OC use can induce IR, dysglycemia and dyslipidemia in the general population, then women with PCOS may have a greater risk for adverse metabolic outcomes when exposed to OCs. Until now, published studies examining the effect of the OC in women with PCOS have been small and yielded conflicting results. Thus, clinicians lack definitive data on the metabolic changes associated with OC treatment in women with PCOS. To investigate the association between metabolic changes and OC use in women with PCOS, we undertook a systematic review and meta-analysis. We hypothesized that OC use is associated with worsening of IR, lipid profiles and glucose metabolism. We expected that differences in effect size may be explained by differences in patient age, duration of OC use, BMI, EE dose and progestogen type. Materials and Methods Search strategy We conducted searches with the assistance of a professional librarian. We used the following MeSH terms: polycystic ovary syndrome, hyperandrogen*, contraceptives, oral, estrogen, lipids, dysglycem* and insulin resistance (Supplementary data, Appendix I). We searched All EBM Reviews (a compilation database through the University of Toronto that includes: Cochrane Database of Systematic Reviews, ACP Journal Club, Database of abstracts of Reviews of Effects, Cochrane Central Register of Controlled Trials, Cochrane Methodology Register, Health Technology Assessment and National Health Service Economic Evaluation), Medline (1966 April 2010) and EMBASE (1980 April 2010) We searched abstracts from the conference proceedings of the European Society of Human Reproduction and Embryology, the American Society of Reproductive Medicine and the Endocrine Society (2005 2009) for relevant unpublished literature and contacted the authors for their data. Experts in the field were contacted for possible unpublished research on the topic. Eligibility of relevant studies We included studies that used the NIH (Zawadzki and Dunaif, 1992) or Rotterdam (ASRM/ESHRE, 2004) criteria for diagnosis of PCOS. We also included studies that did not specify the terms NIH or Rotterdam criteria but described the presence of at least two of the following three clinical criteria in their study cohort, consistent with NIH or Rotterdam criteria: oligoovulation (nine or fewer menses per year), hyperandrogenism and polycystic ovaries on ultrasound. We excluded studies of women with pre-existing diabetes. The intervention of interest was contraceptive dose EE, combined with any type of progestogen. Studies involving women aged 13 44 years were included. The minimum follow-up period required for inclusion was 3 months to allow the menstrual rate, sex hormone binding globulin and androgens time to normalize. In studies with more than one end-point, the follow-up closest to 12 months was used. Initial inclusion criteria were restricted to RCTs that compared either OC use to non-use or two different types of OCs. The first search using these criteria resulted in six possible studies (Creatsas et al., 2000; Mastorakos et al., 2002; Cagnacci et al., 2003; Vrbikova et al., 2004; Mastorakos et al., 2006; Charitidou et al., 2008). None of the RCTs compared OC with placebo and thus the restric (...truncated)