Context Despite gut microbiome being widely studied in metabolic diseases, its role in polycystic ovary syndrome (PCOS) has been scarcely investigated. Objective Compare the gut microbiome in late fertile age women with and without PCOS and investigate whether changes in the gut microbiome correlate with PCOS-related metabolic parameters. Design Prospective, case-control study using the Northern Finland Birth Cohort 1966. Setting General community. Participants 102 PCOS women and 201 age- and body mass index (BMI)-matched non-PCOS control women. Clinical and biochemical characteristics of the participants were assessed at ages 31 and 46 and analyzed in the context of gut microbiome data at the age of 46. Intervention(s) None Main outcome measure(s) Bacterial diversity, relative abundance, and correlations with PCOS-related metabolic measures. Results Bacterial diversity indices did not differ significantly between PCOS and controls (Shannon diversity p = 0.979, unweighted UniFrac p = 0.175). Four genera whose balance helps to differentiate between PCOS and non-PCOS were identified. In the whole cohort, the abundance of two genera from Clostridiales, Ruminococcaceae UCG-002 and Clostridiales Family XIII AD3011 group, were correlated with several PCOS-related markers. Prediabetic PCOS women had significantly lower alpha diversity (Shannon diversity p = 0.018) and markedly increased abundance of genus Dorea (FDR = 0.03) compared to women with normal glucose tolerance. Conclusion PCOS and non-PCOS women at late fertile age with similar BMI do not significantly differ in their gut microbial profiles. However, there are significant microbial changes in PCOS individuals depending on their metabolic health.

ObjectiveTo date, little is known about differences in the knowledge, diagnosis making and treatment strategies of health care providers regarding polycystic ovary syndrome (PCOS) across different disciplines in countries with similar health care systems. To inform guideline translation, we aimed to study physician reported awareness, diagnosis and management of PCOS and to explore differences between medical disciplines in the Nordic countries and Estonia.MethodsThis cross-sectional survey was conducted among 382 endocrinologists and obstetrician-gynaecologists in the Nordic countries and Estonia in 2015–2016. Of the participating physicians, 43% resided in Finland, 18% in Denmark, 16% in Norway, 13% in Estonia, and 10% in Sweden or Iceland, and 75% were obstetrician-gynaecologists. Multivariable logistic regression models were run to identify health care provider characteristics for awareness, diagnosis and treatment of PCOS.ResultsClinical features, lifestyle management and comorbidity were commonly recognized in women with PCOS, while impairment in psychosocial wellbeing was not well acknowledged. Over two-thirds of the physicians used the Rotterdam diagnostic criteria for PCOS. Medical endocrinologists more often recommended lifestyle management (OR = 3.6, CI 1.6–8.1) or metformin (OR = 5.0, CI 2.5–10.2), but less frequently OCP (OR = 0.5, CI 0.2–0.9) for non-fertility concerns than general obstetrician-gynaecologists. The physicians aged <35 years were 2.2 times (95% CI 1.1–4.3) more likely than older physicians to recommend lifestyle management for patients with PCOS for fertility concerns. Physicians aged 46–55 years were less likely to recommend oral contraceptive pills (OCP) for patients with PCOS than physicians aged >56 (adjusted odds ratio (OR) = 0.4, 95% CI 0.2–0.8).ConclusionDespite well-organized healthcare, awareness, diagnosis and management of PCOS is suboptimal, especially in relation to psychosocial comorbidities, among physicians in the Nordic countries and Estonia. Physicians need more education on PCOS and evidence-based information on Rotterdam diagnostic criteria, psychosocial features and treatment of PCOS, with the recently published international PCOS guideline well needed and welcomed.

STUDY QUESTION Can we identify novel variants associated with polycystic ovary syndrome (PCOS) by leveraging the unique population history of Northern Europe? SUMMARY ANSWER We identified three novel genome-wide significant associations with PCOS, with two putative independent causal variants in the checkpoint kinase 2 (CHEK2) gene and a third in myosin X (MYO10). WHAT IS KNOWN ALREADY PCOS is a common, complex disorder with unknown aetiology. While previous genome-wide association studies (GWAS) have mapped several loci associated with PCOS, the analysis of populations with unique population history and genetic makeup has the potential to uncover new low-frequency variants with larger effects. STUDY DESIGN, SIZE, DURATION A population-based case–control GWAS was carried out. PARTICIPANTS/MATERIALS, SETTING, METHODS We identified PCOS cases from national registers by ICD codes (ICD-10 E28.2, ICD-9 256.4, or ICD-8 256.90), and all remaining women were considered controls. We then conducted a three-stage case–control GWAS: in the discovery phase, we had a total of 797 cases and 140 558 controls from the FinnGen study. For validation, we used an independent dataset from the Estonian Biobank, including 2812 cases and 89 230 controls. Finally, we performed a joint meta-analysis of 3609 cases and 229 788 controls from both cohorts. Additionally, we reran the association analyses including BMI as a covariate, with 2169 cases and 160 321 controls from both cohorts. MAIN RESULTS AND THE ROLE OF CHANCE Two out of the three novel genome-wide significant variants associating with PCOS, rs145598156 (P = 3.6×10−8, odds ratio (OR) = 3.01 [2.02–4.50] minor allele frequency (MAF) = 0.005) and rs182075939 (P = 1.9×10−16, OR = 1.69 [1.49–1.91], MAF = 0.04), were found to be enriched in the Finnish and Estonian populations and are tightly linked to a deletion c.1100delC (r2 = 0.95) and a missense I157T (r2 = 0.83) in CHEK2. The third novel association is a common variant near MYO10 (rs9312937, P = 1.7 × 10−8, OR = 1.16 [1.10–1.23], MAF = 0.44). We also replicated four previous reported associations near the genes Erb-B2 Receptor Tyrosine Kinase 4 (ERBB4), DENN Domain Containing 1A (DENND1A), FSH Subunit Beta (FSHB) and Zinc Finger And BTB Domain Containing 16 (ZBTB16). When adding BMI as a covariate only one of the novel variants remained genome-wide significant in the meta-analysis (the EstBB lead signal in CHEK2 rs182075939, P = 1.9×10−16, OR = 1.74 [1.5–2.01]) possibly owing to reduced sample size. LARGE SCALE DATA The age- and BMI-adjusted GWAS meta-analysis summary statistics are available for download from the GWAS Catalog with accession numbers GCST90044902 and GCST90044903. LIMITATIONS, REASONS FOR CAUTION The main limitation was the low prevalence of PCOS in registers; however, the ones with the diagnosis most likely represent the most severe cases. Also, BMI data were not available for all (63% for FinnGen, 76% for EstBB), and the biobank setting limited the accessibility of PCOS phenotypes and laboratory values. WIDER IMPLICATIONS OF THE FINDINGS This study encourages the use of isolated populations to perform genetic association studies for the identification of rare variants contributing to the genetic landscape of complex diseases such as PCOS. STUDY FUNDING/COMPETING INTEREST(S) This work has received funding from the European Union’s Horizon 2020 research and innovation programme under the MATER Marie Skłodowska-Curie grant agreement No. 813707 (N.P.-G., T.L., T.P.), the Estonian Research Council grant (PRG687, T.L.), the Academy of Finland grants 315921 (T.P.), 321763 (T.P.), 297338 (J.K.), 307247 (J.K.), 344695 (H.L.), Novo Nordisk Foundation grant NNF17OC0026062 (J.K.), the Sigrid Juselius Foundation project grants (T.L., J.K., T.P.), Finska Läkaresällskapet (H.L.) and Jane and Aatos Erkko Foundation (H.L.). The funders had no role in study design, data collection and analysis, publishing or preparation of the manuscript. The authors declare no conflicts of interest.