Abstract
Background: Previous studies demonstrate that aggregation of Asian ethnic subgroups masks variation in the prevalence of diabetes (DM) across these groups. Population data also suggest that Asian ethnic differences exist in the prevalence of diagnosed polycystic ovary syndrome (PCOS), but few studies have compared the risk of PCOS and DM in the same Asian subgroups, accounting for differences in BMI. We previously found that the prevalence of PCOS based on ≥1 diagnosis ranged from 1.6–1.9% in White, Hispanic, and Black women and 1.2%, 1.7%, and 3.5% for Chinese, Filipina, and South Asian women age 21–44 years. In this study, we further examined and compared the risk of clinical PCOS and clinical DM by weight status in these Asian subgroups. METHODS: Using data from women receiving care in a large integrated healthcare system, we conducted a cross-sectional, retrospective study of 19,848 Chinese, 23,890 Filipina, and 19,905 South Asian women aged 21–44 years who had ≥1 clinic visit in 2016. Ethnicity was based on data from the electronic health record (EHR), and those identified in the EHR as Asian with unspecified ethnicity were assigned as Chinese, Filipina, and South Asian based on algorithms that included primary language and surname. BMI was calculated from non-gestational height and weight and classified as healthy, overweight (23 to <27.5), or obese (≥27.5 kg/m2) using Asian BMI thresholds. Clinical PCOS was defined by ≥2 ambulatory PCOS diagnoses (ICD-9 256.4, ICD-10 E28.2) in 2015–2016 and clinical DM by ≥2 ambulatory DM diagnoses (ICD-9 250, ICD-10 E10-11,13) in 2015–2016 and history of DM pharmacotherapy. Multivariable logistic regression was used to produce adjusted odds ratios (aOR) for PCOS and DM comparing Filipina and South Asian to Chinese women after adjusting for age and weight status. Results: Among Chinese, Filipina, and South Asian women, respectively, prevalence of PCOS was 1.0%, 1.5%, and 3.2%, prevalence of DM was 1.2%, 4.2%, and 2.5%, and prevalence of obesity was 15.6%, 38.5%, and 30.1%. The prevalence of obesity was high in women with PCOS (58.9%) and women with diabetes (70.0%). South Asian women had a 2.5-fold higher adjusted odds of PCOS than Chinese women (95% CI 2.1–3.0), with no difference for Filipina versus Chinese women (aOR 1.0, 95% CI 0.8–1.2). In contrast, South Asian (aOR 1.5, 95% CI 1.3–1.8) and Filipina women (aOR 2.2, 95% CI 1.9–2.6) had higher adjusted odds of DM than Chinese women. Conclusion: Despite the known metabolic association of PCOS and insulin resistance, we found that the risk profiles of PCOS and DM differed by Asian ethnicity, with risk of PCOS highest in South Asian women and risk of DM highest in Filipina women. While our analyses were limited to women with BMI assessed at a clinical visit and not a screened population, the differential risk of PCOS and DM in Asian subgroups and higher burden of PCOS in South Asian women, independent of BMI, warrant further study.
Referral Bias in Defining the Phenotype and Prevalence of Obesity in Polycystic Ovary Syndrome
