Correlation of Hyperandrogenemia and Metabolic Syndrome in Thai Women With Polycystic Ovary Syndrome (PCOS).

Objective: To investigate the prevalence of metabolic syndrome compared between hyperandrogenemia and non-hyperandrogenemia in Thai women with PCOS, and to identify factors signicantly associated with metabolic syndrome in this PCOS population. Methods: Thai PCOS women were conducted during 2010-2018. Patients were categorized into the non-hyperandrogenemia group or the hyperandrogenemia group dened by total testosterone >0.8 ng/mL or free testosterone >0.006 ng/mL or dehydroepiandrosterone sulfate (DHEA-S) >350 mcg/dL. Metabolic syndrome was diagnosed according to National Cholesterol Education Program-Adult Treatment Panel III criteria. Demographic, anthropometric, clinical, and biochemical blood data were collected and analyzed. Results: 520 PCOS women were included. 22.6% had metabolic syndrome and 75.0% had hyperandrogenemia. Free testosterone cut-off to dene hyperandrogenemia for determining metabolic syndrome in PCOS yielded the highest sensitivity (88.9%) and the highest negative predictive value (90.8%). The prevalence of metabolic syndrome was 27.1% in hyperandrogenemia and 9.2% in non-hyperandrogenemia. Factors signicantly associating with metabolic syndrome in Thai PCOS women were age, BMI, free testosterone, and DHEA-S. Conclusion: The prevalence of metabolic syndrome was to be signicantly higher in hyperandrogenemia women than in their non-hyperandrogenemia counterparts. Older age, higher BMI, higher free testosterone, and lower DHEA-S were all identied as factors signicantly associated with metabolic syndrome.


Introduction
Polycystic ovary syndrome (PCOS) is a common endocrine disorder that most often develops in reproductive-aged women [1]. The incidence of PCOS was reported to be 5-10% in women of childbearing age [2]. PCOS is diagnosed when a woman has at least two of the following: irregular menstruation, hyperandrogenism, and/or polycystic ovary that causes disruption of normal ovarian function. PCOS is also associated with the metabolic disorders that characterize metabolic syndrome, including insulin resistance [3], dyslipidemia, elevated blood pressure, and prothrombotic and proin ammatory factors that contribute to cardiovascular disease and diabetes mellitus. These factors were also found to be associated with increased risk of ischemic heart disease in patients with metabolic syndrome [4]. Therefore, early detection and treatment of metabolic syndrome is essential for preventing or minimizing adverse cardiovascular consequences.
Androgen excess is associated with detrimental changes in lipid metabolism by increased hepatic lipase activity, which results in increased degradation and decreased serum level of high-density lipoprotein (HDL) cholesterol. Moreover, androgens increase plasma low-density lipoprotein (LDL) cholesterol by increasing hepatic lipase activity, which enhances very low-density lipoprotein (VLDL) cholesterol catabolism and downregulates LDL receptors [5]. Previous study reported hyperandrogenism to be associated with metabolic syndrome in PCOS women [6]. A 2014 study from Korea found association between obesity and metabolic syndrome in women with PCOS [4]. In contrast, another study found no signi cant association between androgen excess and metabolic syndrome in obese women with PCOS [7].
Our review of the literature revealed no study that has investigated association between hyperandrogenism and metabolic syndrome in Thai PCOS women. Accordingly, the aim of this study was to investigate the prevalence of metabolic syndrome compared between hyperandrogenemia and nonhyperandrogenemia in Thai women with PCOS, and to identify factors signi cantly associated with metabolic syndrome in this PCOS population. Data from this study will improve screening, which will result in early diagnosis and treatment, improved patient outcomes, and reduced healthcare costs related to the adverse cardiovascular consequences of this condition. Demographic (age, gender), anthropometric (height, weight, waist circumference), clinical (blood pressure), and biochemical blood data were collected and analyzed. The details of measurement at our clinic for each parameter were reported elsewhere [9]. The biochemical blood tests evaluated to diagnose metabolic syndrome included fasting blood sugar (FBS), total cholesterol (TC), triglycerides (TG), highdensity lipoprotein cholesterol (HDL-C), and low-density lipoprotein (LDL-C). Other blood tests included total testosterone, dehydroepiandrosterone sulphate (DHEA-S), sex hormone binding globulin (SHBG), and albumin. Free testosterone was classi ed by computerized calculator using SHBG and albumin [10]. The biochemical assays were performed at the biochemical laboratory unit of the Department of Clinical Pathology, Faculty of Medicine Siriraj Hospital, Mahidol University, which is an ISO 15189 certi ed laboratory. Patients having incomplete metabolic syndrome and/or hyperandrogenemia investigation data were excluded.

De nitions
Metabolic syndrome was diagnosed by National Cholesterol Education program-Adult Treatment Panel III (NCEP-ATP III) criteria [11]. According to NCEP ATP III criteria, metabolic syndrome is diagnosed when three or more of the following risk factors are present: waist circumference ≥ 80 cm, fasting plasma glucose ≥ 100 mg/dL, serum triglyceride ≥ 150 mg/dL, HDL-cholesterol < 50 mg/dL, and/or blood pressure ≥ 130/85 mmHg.
Hyperandrogenemia was diagnosed if a patient had at least one abnormal serum androgen value (total testosterone > 0.8 ng/mL or free testosterone > 0.006 ng/mL or DHEA-S > 350 microgram/dL).

Sample size calculation and statistical analysis
Using an estimated prevalence of 10% in the non-hyperandrogenemia group, an estimated prevalence of 20% in the hyperandrogenemia group, a probability of a type 1 error (alpha, α) of 0.05, and a 20% increase to compensate for lost or missing data, a minimum sample size of 130 in the non-hyperandrogenemia group, and of 390 in the hyperandrogenemia group was calculated.
Data were analyzed using SPSS Statistics version 18 (SPSS, Inc., Chicago, IL, USA). Data are presented as mean ± standard deviation, number and percentage, or percentage and 95% con dence interval.
Student's t-test or Mann-Whitney U test was used to compare continuous data, and chi-square test or Fisher's exact test was used to compare categorical data. A p-value less than 0.05 was regarded as being statistically signi cant.

Results
Among 520 Thai women with PCOS, 118 (22.6%) had metabolic syndrome and 390 (75.0%) had hyperandrogenemia. The mean age of all participants was 25.5 ± 5.59 years, and the mean body mass index (BMI) was 26.40 ± 7.47 kg/m 2 . The proportion of PCOS women that presented with elevated systolic blood pressure, elevated diastolic blood pressure, central obesity, impaired fasting glucose, and dyslipidemia was 16.9%, 10.3%, 52.2%, 8.5%, and 39.7% respectively. The prevalence of women who had hyperandrogenemia de ned by total testosterone, free testosterone, or DHEA-S was 25.4%, 73.0%, and 12.7%, respectively (Table 1).  Table 2. The prevalence of metabolic syndrome was found to be signi cantly higher in the hyperandrogenemia group than in the non-hyperandrogenemia group (27.1% vs. 9.2%, respectively; p < 0.001). Mean BMI was also signi cantly higher among hyperandrogenemia women than among their non-hyperandrogenemia counterparts (p < 0.001). All components of metabolic syndrome, including blood pressure, waist circumference, fasting blood glucose, and triglycerides were signi cantly higher in the hyperandrogenemia group than in the non-hyperandrogenemia group (all p < 0.05). Factors identi ed as being signi cantly associated with metabolic syndrome in Thai women with PCOS were older age, high body mass index (BMI), high free testosterone level, and low DHEA-S level (Table 3).
Free testosterone cut-off value to de ne hyperandrogenemia for determining metabolic syndrome in PCOS yielded the highest sensitivity (88.9%) and the highest negative predictive value (90.8%) (data not shown).

Discussion
Among the physiologic hormonal changes that occur in PCOS, reproductive endocrine function might especially in uence the risk of metabolic derangement. Larger waist circumference, higher blood pressure, lower insulin sensitivity, and a greater degree of dyslipidemia are known to develop in PCOS women. As a result, the prevalence of metabolic syndrome was found to be higher in PCOS women, especially those with older age, obesity, and hyperandrogenemia status. In our study, metabolic syndrome was more prevalent in the hyperandrogenemia PCOS group than in non-hyperandrogenemia PCOS group. The prevalence of metabolic syndrome was 27.1% and 9.2% in hyperandrogenemia and nonhyperandrogenemia groups, respectively. Prevalence of metabolic syndrome was statistically signi cantly different between hyperandrogenemic PCOS and non-hyperandrogenemic PCOS, which is comparable to the results of two previous studies [4,5]. Similar to previous studies, PCOS women with hyperandrogenemia in the present study had more obesity [4], and a higher level of abnormality among the components of metabolic syndrome [12], including blood pressure, waist circumference, fasting blood glucose, triglycerides, and HDL-C. In contrast, The present study had the difference in prevalence of metabolic syndrome according to difference of mean age, mean BMI, race and criteria for diagnosis of metabolic syndrome when compared with previous studies [4,12]. Obesity and androgen excess in PCOS may involve the aggravation of metabolic abnormalities, such as increased visceral fat, decreased lipolysis in subcutaneous fat, insulin resistance in adipose tissue and skeletal muscle, and lipid metabolism[6]. These factors in uence an increased risk of metabolic syndrome.
Many studies reported obesity and hyperandrogenism to be factors that in uence the development of metabolic syndrome in PCOS women [4,5]. However, hyperandrogenism and its associated parameters, including free androgen index and SHBG, were reported to be signi cantly associated with metabolic syndrome in non-obese PCOS women [13]. Another study found a signi cant difference in abnormal insulin resistance and lipid metabolism between PCOS with and without hyperandrogenism [5]. As such, several factors that relate to hyperandrogenism in uence metabolic syndrome. Similar to previous study [13], we found age, BMI, free testosterone, and DHEA-S to be factors signi cantly associated with metabolic syndrome in Thai women with PCOS.
Although we found higher free testosterone and lower DHEA-S to be associated with metabolic syndrome, no association was observed for total testosterone. A recent study reported that high levels of circulating androgens derived from the adrenal gland, and that lower levels of ovarian androgens exert a bene cial effect on metabolic disturbances in PCOS women [14]. There are several mechanisms that might explain the protective effect of DHEA-S in PCOS women, including its effect on insulin function [14]. DHEA decreases gluconeogenesis by suppressing the activity and expression of glucose-6-phosphatase and phosphoenolpyruvate carboxykinase [15]. DHEA-S increases glucose uptake in hepatocytes and increases insulin binding to its receptor [15]. High circulating levels of glucose or insulin might impact DHEA-S synthesis in the adrenal gland, which might be mediated via insulin action [16]. Therefore, a high DHEA-S might be a defensive effect to prevent a metabolic syndrome event. Our ndings revealed free testosterone cut-off for diagnosing hyperandrogenemia to be the only parameter signi cantly associated with metabolic syndrome. Therefore, hyperandrogenemia using free testosterone cut-off should be considered a potential indicator for predicting metabolic syndrome in PCOS women.

Limitations
This study has some mentionable limitations. First and consistent with the retrospective nature of this study, some patient data may have been missing or incomplete, and no causal relationships could be conclusively established. Second, this was a single-center study. Third, metabolic syndrome is a multifactorial disorder, and our study was not able to adjust for other confounding factors, including nutritional status, exercise, lifestyle and smoking. Further prospective study is needed to support the ndings of this study and to include and assess the aforementioned unevaluated factors.

Conclusion
The results of this study revealed the prevalence of metabolic syndrome to be signi cantly higher in hyperandrogenemia women than in their non-hyperandrogenemia counterparts. Older age, higher BMI, higher free testosterone, and lower DHEA-S were all identi ed as factors signi cantly associated with metabolic syndrome.