Somayeh Hashemi1, Fahimeh Ramezani Tehrani2, Mahsa Noroozzadeh1, Fereidoun Azizi1. 1. Reproductive Endocrinology Research Center, Research Institute for Endocrine Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran. 2. Reproductive Endocrinology Research Center, Research Institute for Endocrine Sciences, Shahid Beheshti University of Medical Sciences, Tehran, Iran. Electronic address: ramezani@endocrine.ac.ir.
Abstract
OBJECTIVE: The normal cut-off values for hyperandrogenaemia differ between populations. Although there are various ways to determine the normal range of hormones in a population, use of a gold standard and the 95th percentile approach are not appropriate to diagnose hyperandrogenism. This study aimed to determine normative cut-off androgen levels for Iranian women of reproductive age using k-means cluster analysis. STUDY DESIGN: The subjects were 1126 women, aged 18-45 years, initially selected at random from women of reproductive age from different geographic regions of Iran. Serum levels of hormones were measured, and complete data were available for 923 women following the implementation of exclusion criteria. The reference group comprised 423 eumenorrhoeic non-hirsute women selected from the total population. Normative cut-off levels were computed using k-means cluster analysis (k=3). Analyses were also conducted based on age and body mass index (BMI) groups for the 923 subjects. RESULTS: Cluster cut-off values for the total study population were similar to the 95th percentile values, whereas the percentiles for the free androgen index, total testosterone, dehydroepiandrostenedione sulphate and androstenedione were 79.3%, 74.1%, 96.7% and 71%, respectively, in the reference group. CONCLUSIONS: Determination of cut-off values for hyperandrogenism based on age and BMI categories provides the most objective way of determining normal upper limits of androgen levels among Iranian women of reproductive age.
OBJECTIVE: The normal cut-off values for hyperandrogenaemia differ between populations. Although there are various ways to determine the normal range of hormones in a population, use of a gold standard and the 95th percentile approach are not appropriate to diagnose hyperandrogenism. This study aimed to determine normative cut-off androgen levels for Iranian women of reproductive age using k-means cluster analysis. STUDY DESIGN: The subjects were 1126 women, aged 18-45 years, initially selected at random from women of reproductive age from different geographic regions of Iran. Serum levels of hormones were measured, and complete data were available for 923 women following the implementation of exclusion criteria. The reference group comprised 423 eumenorrhoeic non-hirsute women selected from the total population. Normative cut-off levels were computed using k-means cluster analysis (k=3). Analyses were also conducted based on age and body mass index (BMI) groups for the 923 subjects. RESULTS: Cluster cut-off values for the total study population were similar to the 95th percentile values, whereas the percentiles for the free androgen index, total testosterone, dehydroepiandrostenedione sulphate and androstenedione were 79.3%, 74.1%, 96.7% and 71%, respectively, in the reference group. CONCLUSIONS: Determination of cut-off values for hyperandrogenism based on age and BMI categories provides the most objective way of determining normal upper limits of androgen levels among Iranian women of reproductive age.
Authors: Hasina Maredia; Geralyn M Lambert-Messerlian; Glenn E Palomaki; Satupaitea Viali; Nicola L Hawley; Stephen T McGarvey Journal: Clin Biochem Date: 2016-02-18 Impact factor: 3.281