Cut-off levels for hyperandrogenemia among Samoan women: An improved methodology for deriving normative data in an obese population

Abstract

Objective To define biochemical hyperandrogenemia (HA) among a population-based sample of reproductive-aged Samoan women, taking into consideration their high BMI levels. Design and Methods A secondary analysis was performed among a cross-sectional sample of Samoan women aged 25–39 (n=494) who were part of a larger genome-wide association study (GWAS) of adiposity. Women who were pregnant/lactating, had a hysterectomy, oophorectomy, other pelvic surgery, and/or indicated use of contraceptive injections were excluded from the study. We analyzed the distribution of free androgen index (FAI) values to establish normative androgen data among Samoan women of reproductive age. Using the lowest tertile of body mass index (BMI), we defined HA as free androgen index (FAI) values > 95th percentile in that subsample. We compared the anthropometric and metabolic characteristics of those with HA to those with normal androgen levels. Results HA was defined as FAI > 8.5. Using this definition, 14% of women were classified as hyperandrogenemic. Women with HA had significantly higher BMI, abdominal circumferences, fasting lipids, and insulin and significantly lower adiponectin levels. Conclusion This study is the first to define normative androgen values among Samoan women with a quantitative assessment of the relationship between adiposity and androgen levels. The uniquely high BMI levels in the population not only provide important clinical insight into normative androgen values among Samoan women, but also serve as a reference for the clinical assessment of HA, taking into consideration BMI, in other populations.

Reproductive health, obesity, and cardiometabolic risk factors among Samoan women

“…The Kar et al (2013) findings are consistent with our descriptive findings of more extreme abnormalities among women with OM/AM + HA and women with OM/AM + HA + PCO. In our earlier study of HA in this Samoan sample, HA was associated with greater total, central, and peripheral adiposity, and higher triglycerides, insulin, and HOMA-IR (Maredia et al 2016). Our findings suggest an underlying pathophysiology that influences reproductive and metabolic health together, perhaps driven by HA secondary to adiposity.…”

“…In our prior report in a different sample of Samoan women, with a less robust definition of HA, without control for injectable contraceptive use, and without an estimate of PCO, the frequency of having both OM/AM and HA was 8.4% (Lambert-Messerlian et al 2011). Considering the high prevalence of obesity, type 2 diabetes (Hawley et al 2014), and high androgen levels (Maredia et al 2016) in Samoan women 25–39 years of age, one might expect a higher prevalence of PCOS. Recent reviews emphasize the complex etiology and presentation of PCOS and the potential for ethnic diversity in prevalence and risk factors (Lizneva et al 2016).…”

“…Anthropometry, cardiometabolic factors, and androgen measurements for the sample have previously been described (Hawley et al 2014; Maredia et al 2016). Material lifestyle scores (MLS), an inventory of household assets, was calculated as a measure of socioeconomic position as in prior studies (Choy et al 2016).…”

“…SHBG levels above the 95 th percentile of the population were checked to exclude heterophilic interference. In our recent paper on androgen levels in Samoan women, we defined HA using the 95 th FAI percentile of the lowest BMI tertile, FAI > 8.5, and use that cutoff here (Maredia et al 2016). …”

“…Specifically, we improve the earlier study of menstrual irregularity by excluding women who report use of contraceptive injections, using a revised estimate of biochemical hyperandrogenism (Maredia et al 2016), and applying AMH levels as a marker of polycystic ovary morphology.…”