Impact of FAI on Metabolism and Adrenal Hyperandrogenism in Women with Polycystic Ovary Syndrome: a cross-sectional study

doi:10.21203/rs.3.rs-4894808/v1

Background

Although PCOS is a heterogeneous endocrinopathy in reproductive-aged women characterized by reproductive, endocrine, metabolic and psychological abnormalities, hyperandrogenism seems to be the most consistent feature. Therefore, the aim of the present study was to assess three hyperandrogenism parameters, FAI, TT and DHEAS, and their relationships with diverse metabolic indices, metabolic derangements and adrenal hyperandrogenism in women diagnosed with PCOS.

Methods

In this single-center cross-sectional study, 217 women aged 18–45 years with PCOS were enrolled. Baseline phenotypic, endocrine, and metabolic parameters, as well as multiple endocrine hormone levels, including insulin levels, thyroid function, sex hormones, ACTH, and cortisone levels, were collected. Dynamic function tests, such as OGTT, 1-mg dexamethasone suppression test, and ACTH stimulation test, were performed. Liver and vaginal ultrasound scans were also conducted to fully assess the metabolic and endocrine status of the participants.

Results

FAI was positively associated with BMI, waist circumference, hip circumference, waist-to-hip ratio, SBP, DBP, FPG, PPG, HbA1c, HOMA-IR, AUCGLU, AUCINS, TG, TC, LDL-C and UA (p < 0.05). Multivariate logistic regression revealed that FAI was an independent risk factor for multiple metabolic disorders, including overweight/obesity, fatty liver, dyslipidemia, insulin resistance, hyperglycemia, metabolic syndrome, and hyperuricemia. Moreover, the FAI was positively correlated with both TT and DHEAS (p < 0.05). In addition, the FAI was negatively related to basal, stimulated and suppressed cortisol and positively associated with ACTH, DHEAS and stimulated 17OHP in PCOS patients (p < 0.05). According to age range-based cutoff values of circulating DHEAS, 85 of 217 PCOS women with adrenal hyperandrogenism and higher levels of FAI likely experienced more severe adrenal hyperandrogenism. A ROC analysis was performed with the best cutoff point (FAI = 5.29, AUC = 0.724, 95% CI: 0.654–0.793, p = 0.000, sensitivity = 71.8%, specificity = 64.2%).

Conclusions

The FAI is superior to TT and DHEAS in reflecting comprehensive features in terms of their correlation with phenotypic and metabolic parameters in patients with PCOS. Moreover, the FAI is also a promising biomarker for predicting adrenal hyperandrogenism, with the best cutoff point (value = 5.29).

Trial Registration

Chinese Clinical Trials Registry (registration number ChiCTR2000040904, date of registration 20201215).

PCOS

FAI

Metabolism

Arenal hyperandrogenism

Polycystic ovary syndrome (PCOS) is a complex endocrinopathy in reproductive-aged women that is characterized by reproductive, endocrine, metabolic and psychological features.[1] Initially, reproductive hormone abnormalities involving luteinizing hormone (LH) hypersecretion and hyperandrogenism, resulting in acne and hirsutism, as well as subfertility, are the consequences of aberrant follicular maturation, ovulatory disturbance, and miscarriage. Subsequently, associated metabolic dysregulations, such as hyperinsulinemia, insulin resistance, metabolic syndrome, obesity, dyslipidemia, and hepatic steatosis, which predispose affected women to an increased risk of type 2 diabetes mellitus and cardiovascular disease, have gradually been recognized.[2]

Due to the heterogeneity of the clinical phenotype, the diagnostic criteria for PCOS have varied in recent years until international evidence-based guidelines for the assessment and management of PCOS endorsed the use of the Rotterdam criteria declared in 2018.[1] However, currently, the pathogenesis of PCOS remains undecipherable, and mechanism-based treatment is still unavailable. Hyperandrogenism, as the most consistent feature of PCOS, has received abundant attention. Hyperandrogenic PCOS women exhibit an increase in various androgens, including serum levels of total testosterone (TT), free testosterone (FT), androstenedione (A4), dehydroepiandrosterone (DHEA) and dehydroepiandrosterone sulfate (DHEAS), as well as the calculated free testosterone and free androgen index (FAI).[3] Moreover, detectable hyperandrogenism may contribute to adverse metabolic phenotypes, such as hyperglycemia, insulin resistance, and hepatic steatosis, in individuals with PCOS.[4, 5] Animal models of PCOS induced by androgen also display metabolic profiles, including obesity, adipose dysfunction, insulin resistance and non-alcoholic fatty liver disease(NAFLD), closely mimicking clinical PCOS traits.[6, 7] All of these findings lead to the hypothesis that hyperandrogenism might be the cause of metabolic disorders rather than the consequence.

Moreover, various studies have reported the dysregulation of steroid metabolism through the hypothalamus–pituitary–adrenal (HPA) axis in PCOS patients.[8, 9] In particular, studies have reported an increasing association between psychological distress and PCOS.[10] The role of cortisol, a major stress marker, has drawn increasing attention. However, alterations in cortisol metabolism have not been consistent across studies, and few studies have investigated its correlation with hyperandrogenism.

Therefore, the aim of the present study was to assess three hyperandrogenism parameters, FAI, TT and DHEAS, and their relationships with diverse metabolic indices, metabolic derangements and adrenal hyperandrogenism in women diagnosed with PCOS.

Study design and subjects

In this observational, cross-sectional study, women of reproductive age (18–45 years) who were newly or previously diagnosed with PCOS at The Third Affiliated Hospital of Guangzhou Medical University without any medication therapy for PCOS or metabolic syndrome or other steroid therapy within the past 3 months were eligible for inclusion. According to the revised Rotterdam criteria, PCOS is diagnosed when either two or three of the key features are present: ovulatory dysfunction, polycystic ovarian morphology and clinical and/or biochemical hyperandrogenism.[11] All patients underwent a standardized examination involving anthropometric measurements, metabolic and hormonal evaluation to rule out other causes of hyperandrogenism, such as congenital adrenal hyperplasia (CAH), Cushing’s syndrome, subclinical Cushing syndrome or androgen-secreting neoplasm. Women who were currently pregnant, postabortion, postpartum, or breast feeding in the past 3 months or who refused to complete the above test were excluded. In total, approximately 217 women were diagnosed and phenotyped by this standardized screening method. The survey was approved by the Medical Ethics Committee of The Third Affiliated Hospital of Guangzhou Medical University, Guangzhou, Guangdong, China. All patients provided written informed consent before entering the study, which was conducted in accordance with the Declaration of Helsinki and registered at the Chinese Clinical Trials Registry (ChiCTR2000040904).

Clinical parameters

All the participants underwent demographic and anthropometric assessments after screening. Menstrual irregularity (MI) was defined as an interval > 35 days or < 8 menstrual cycles in the past year. Hirsutism was assessed according to a modified Ferriman-Gallwey (FG) score ≥ 5.[12]

Metabolism Measurements

Fasting blood samples were collected in the morning after fasting for at least 8 hours to analyze fasting plasma glucose (FPG), fasting insulin (Fins), HbA1c, triglyceride (TG), total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and uric acid (UA) levels. Each participant underwent a 75-g oral glucose tolerance test (OGTT) with blood sampling at 30 minutes, 1 hour, 2 hours and 3 hours for plasma glucose and insulin. The areas under the curve (AUCs) for glucose (AUC_GLU) and insulin (AUC_INS) were calculated via the trapezoid method. The homoeostasis model assessment of insulin resistance (HOMA-IR) was calculated using the following formula: [HOMA-IR = Fins ×FPG/22.5]. The insulin sensitivity index (ISI) was calculated using the following formula:

Metabolic syndrome was diagnosed according to any three out of the following five criteria: waist circumference > 88 cm, blood pressure ≥ 130/85 mmHg, HDL < 1.3 mmol/L (50 mg/dL), TG ≥ 1.7 mmol/L (150 mg/dL), and FPG ≥ 6.1 mmol/L (110 mg/dL).[14] Adrenal hyperandrogenism was diagnosed according to age range-based cutoff values of circulating DHEAS concentrations for the definition of adrenal hyperandrogenism (18–24 years old: 3,402 ng/ml; 25–34 years old: 2,982 ng/ml; 35–49 years old: 2,582 ng/ml).[15]

Hormone Measurements

Fasting blood samples before 8 am for sex hormone, including serum TT, DHEAS, SHBG, FSH, LH, estradiol (E2), prolactin (PRL), 17-hydroxyprogesterone (17OHP) and Anti-Müllerian Hormone (AMH), were collected on the third day of menstruation or the same day when samples of metabolic biomarkers were collected for participants with amenorrhea.

Serum TT was analyzed by solid-phase competitive chemiluminescence (analytical sensitivity 0.5 nmol/l). SHBG and DHEAS were measured by an immunometric chemiluminescent enzyme immunoassay (with an analytical sensitivity of 0.1 nmol/l and 0.1 µmol/L, respectively). The FAI was calculated by using the following formula: [TT (nmol/L) × 100/SHBG (nmol/L)]. FSH, LH, E2, PRL and 17OHP were measured by radioimmunoassay. AMH was assessed via batch analysis using a DS2 ELISA robot with a single lot reagent (AMH Gen II ELISA, Beckman Coulter; Inc.).

Other endocrine hormone measurements, including thyroid function, adrenocorticotropic hormone (ACTH), and the circadian rhythm of serum cortisone levels, were detected by radioimmunoassay. The ACTH stimulation test was conducted the next morning between 8:00 AM and 9:00 AM. All the recruited participants underwent serum assessment for cortisol and 17OHP at baseline and 30 minutes and 60 minutes after 0.25 mg of cosyntropin.

Afterward, a 1-mg dexamethasone suppression test (DST) was performed on a different day to exclude any Cushing’s syndrome or subclinical Cushing’s syndrome.

Vaginal ultrasound examination

All the participants underwent vaginal ultrasound examination to determine the antral follicle count (AFC) and bilateral ovarian volume. Polycystic ovary morphology (PCOM) was considered the presence of ≥ 12 antral follicles (with a diameter of 2–9 mm) and/or an ovarian volume > 10 million or both.[11]

Liver ultrasound scan

A liver ultrasound scan of the liver was carried out to assess fatty liver.

Statistical analysis

The measurement variables are presented as the means ± SEMs and percentiles. Categorical variables are expressed as proportions (percentages). For correlation analysis, Pearson correlation was utilized with the correlation coefficient. Multivariate logistic regression analysis was performed to calculate the odds ratios (OR) and 95% confidence intervals (CIs) for the independent predictors of crucial PCOS-related clinical metabolic phenotypes. The results are expressed as age- and BMI-adjusted OR with 95% CI and two-sided p values. Multivariate logistic regression was applied with forward stepwise selection between metabolic disturbances and several clinical or laboratory variables only if they had qualified associations through univariate logistic regression analysis. More than two group means were compared using mixed-model repeated-measures analysis of covariance (ANCOVA) to estimate the change in variables, with age and BMI of the corresponding dependent variable as covariates. P value of 0.05 was considered to indicate statistical significance. Statistical analysis was performed using IBM SPSS Statistics, version 26 (IBM Corp, Armonk, NY).

Population characteristics

The average age of the 217 women with PCOS in the studied population was 28.2 ± 0.3 years, the BMI was 25.0 ± 0.3 kg/m2, the waist circumference was 83.68 ± 0.81 cm, the FPG was 5.01 ± 0.06 mmol/L, the 2-hour postprandial glucose (2h PPG) level was 7.76 ± 0.18 mmol/L, the HbA1c level was 5.57 ± 0.06%, the HOMA-IR was 3.40 ± 0.20, the TG level was 1.55 ± 0.08 mmol/L, the TC level was 4.86 ± 0.07 mmol/L, the LDL-C level was 3.10 ± 0.07 mmol/L, the HDL-C level was 1.46 ± 0.03 mmol/L, the UA level was 353.41 ± 6.18 µmol/L, and the Thyroid Stimulating Hormone (TSH) level was 2.43 ± 0.40 mU/L. The mean TT was 2.50 ± 0.08 nmol/L, the mean DHEAS was 289.67 ± 8.14 µg/dL, and the mean FAI was 7.53 ± 0.46. Overall, the prevalence of various metabolic and endocrine disorders is presented in Table 1.

Table 1

Characteristics of recruited 217 PCOS women
Characteristics	Case (n)	Percentage (%) (total n = 217)
Menstrual irregularity	213	98.2
PCOM	197	90.8
Bilateral	180	82.9
Unilateral	17	7.8
mFG ≥ 5	5	2.3
TT ≥ 1.97 nmol/L	162	74.7
Overweight/Obese	127	58.5
28kg/m2>BMI ≥ 24 kg/m²	49	22.6
BMI ≥ 28 kg/m²	73	33.6
Waist I ≥ 85 cm	87	40.1
Hypertension	31	14.3
SBP ≥ 140 mmHg	18	8.3
DBP ≥ 90 mmHg	27	12.4
Hyperglycaemia	86	39.6
FPG ≥ 6.1 mmol/L	19	8.8
FPG ≥ 7.0 mmol/L	8	3.7
2h PPG ≥ 7.8 mmol/L	86	39.6
2h PPG ≥ 11.1 mmol/L	22	10.1
HbA1c ≥ 6.5%	16	7.4
HOMA-IR ≥ 2.69	102	47.5
Hyperlipidaemia	122	56.2
TG ≥ 1.7 mmol/L	59	27.2
TC ≥ 5.2 mmol/L	69	31.8
LDL_C ≥ 3.4 mmol/L	72	33.2
HDL_C < 1.0 mmol/L	25	11.5
Hyperuricaemia	92	42.4
UA ≥ 360 umol/L	92	42.4
Metabolic Syndrome	50	23.0
Fatty Liver	74	34.1
Thyroid Disfunction	9	4.1
TSH ≥ 5 mU/L	6	2.8
TSH < 0.35 mU/L	3	1.4
Adrenal Hyperandrogenism	85	39.2
BMI body mass index, AG abdominal girth, WHR waist-hip rate, TG total triglyceride, TC total cholesterol, HDL- high density lipopretein cholesterol, LDL-c low density lipopretein cholesterol, FPG fasting plasma glucose, PPG 2-h postprandial plasma glucose, HOMA-IR homeostasis model assessment of insulin resistance, HOMA-B homeostasis model assessment of b-cell function

Correlations of metabolic parameters with FAI, TT and DHEAS

The present correlation analysis demonstrated that FAI was positively associated with BMI, waist circumference, hip circumference, waist-to-hip ratio, SBP, DBP, FPG, PPG, HbA1c, HOMA-IR, AUC_GLU, AUC_INS, TG, TC, LDL-C and UA (p < 0.05). In addition, the FAI was negatively associated with the ISI and HDL-C, whereas the TT was negatively correlated with the ISI, as indicated by a lower correlation coefficient (p < 0.05) (Table 2).

Table 2

Correlation analysis of metabolic biomarkers and FAI, TT and DHEAS.
Characteristics	FAI		TT		DHEAS		SHBG
Characteristics	r	p	r	p	r	p	r	p
Age	-0.031	0.654	-0.108	0.112	-0.156	0.022	-0.109	0.111
BMI	0.521	0.000	0.047	0.490	0.039	0.569	-0.416	0.000
Waist circumference	0.471	0.000	0.076	0.284	0.043	0.546	-0.396	0.000
Hip circumference	0.401	0.000	0.033	0.656	0.023	0.753	-0.313	0.000
Waist-to-hip ratio	0.279	0.000	0.106	0.144	0.088	0.223	-0.290	0.000
SBP	0.166	0.014	0.049	0.470	0.041	0.548	-0.077	0.261
DBP	0.179	0.008	-0.014	0.837	0.012	0.855	-0.154	0.023
FPG	0.318	0.000	-0.083	0.221	0.033	0.634	-0.205	0.002
PPG	0.313	0.000	0.003	0.960	0.052	0.450	-0.246	0.000
HbA1c	0.327	0.000	0.044	0.594	0.068	0.411	-0.271	0.001
HOMA-IR	0.285	0.000	0.127	0.062	0.005	0.937	-0.162	0.017
AUC_GLU	0.373	0.000	-0.084	0.240	-0.008	0.914	-0.171	0.016
AUC_INS	0.281	0.000	0.081	0.260	-0.021	0.776	-0.232	0.001
ISI	-0.406	0.000	-0.147	0.031	-0.063	0.361	0.298	0.000
TG	0.167	0.014	-0.048	0.486	-0.060	0.377	-0.080	0.238
TC	0.151	0.026	0.119	0.080	0.055	0.420	-0.020	0.768
LDL-C	0.313	0.000	0.078	0.262	0.093	0.182	-0.294	0.000
HDL-C	-0.419	0.000	0.087	0.218	-0.054	0.446	-0.529	0.000
UA	0.534	0.000	0.043	0.533	0.101	0.140	-0.410	0.000

Independent influential factors of PCOS-related metabolic disturbances

According to multivariate logistic regression, the independent risk factors for PCOS-related metabolic clinical phenotypes are displayed in Table 3. In women with PCOS, the FAI was an independent risk factor for overweight/obesity. Moreover, both overweight/obesity and FAI were risk factors for dyslipidemia, fatty liver, hyperglycemia, metabolic syndrome, and hyperuricemia. For insulin resistance, overweight/obesity and FAI were risk factors, whereas age was a protective factor. Moreover, overweight/obesity was the only risk factor for hypertension. (Table 3).

Table 3

Independent predictors of PCOS metabolic related phenotype.
	Predictors	OR	95% CI	p
Overweight/Obese	FAI	1.241	1.145–1.344	0.000
Fatty liver	BMI ≥ 24 Kg/m²	25.757	7.553–87.831	0.000
	FAI	1.128	1.060–1.200	0.000
Dyslipidaemia	BMI ≥ 24 Kg/m²	2.132	1.159–3.924	0.015
	FAI	1.077	1.020–1.137	0.007
Insulin resistance	BMI ≥ 24 Kg/m²	7.468	3.606–15.465	0.000
	FAI	1.139	1.065–1.218	0.000
	Age	0.910	0.839–0.987	0.024
Hyperglycaemia	BMI ≥ 24 Kg/m²	3.055	1.600-5.831	0.001
	FAI	1.049	1.001–1.099	0.045
Metabolic syndrome	BMI ≥ 24 Kg/m²	7.450	2.437–22.770	0.000
	FAI	1.085	1.029–1.144	0.002
Hyperuricemia	BMI ≥ 24 Kg/m²	2.847	1.474–5.501	0.002
	FAI	1.050	1.001–1.102	0.049
Hypertension	BMI ≥ 24 Kg/m²	6.253	2.104–18.585	0.001
Note: Presented as age, BMI, TT and DHEAS adjusted OR with 95% CI and two-sided p value.

Correlations of sex hormones with FAI, TT and DHEAS

The FAI was positively correlated with both TT and DHEAS (p < 0.05). There was a significant negative correlation between FAI and SHBG, with the highest correlation coefficient compared to that between TT and SHGB or between DHEAS and SHBG (p < 0.05). AMH was only positively related to TT (p < 0.05). TT was also positively correlated with E2, LH and LH/FSH (p < 0.05) (Table 4).

Table 4

Correlation analysis of hormone and FAI, TT and DHEAS.
Characteristics	FAI		TT		DHEAS		SHBG
Characteristics	r	p	r	p	r	p	r	p
FAI	/	/	0.204	0.003	0.176	0.010	-0.632	0.000
TT	0.204	0.003	/	/	0.037	0.591	0.075	0.270
DHEAS	0.176	0.010	0.037	0.591	/	/	-0.218	0.001
FSH	-0.019	0.781	0.033	0.628	0.120	0.080	-0.001	0.992
LH	-0.098	0.150	0.244	0.000	-0.033	0.633	0.131	0.054
LH/FSH	-0.082	0.231	0.222	0.001	-0.035	0.608	0.107	0.115
E2	0.005	0.940	0.195	0.004	-0.017	0.800	-0.044	0.517
P	0.064	0.364	-0.004	0.957	-0.020	0.778	-0.059	0.396
SHBG	-0.632	0.000	-0.149	0.028	-0.218	0.001	/	/
AMH	-0.075	0.343	0.278	0.000	-0.114	0.154	0.145	0.066
Prolactin	-0.123	0.072	-0.015	0.826	-0.020	0.766	0.158	0.020
TSH	-0.045	0.508	-0.050	0.465	-0.058	0.395	-0.014	0.840
FT3	-0.024	0.722	0.186	0.006	-0.027	0.692	0.163	0.017
FT4	-0.244	0.000	0.029	0.671	0.059	0.395	0.294	0.000
ACTH	0.207	0.003	0.051	0.461	0.040	0.569	-0.065	0.347
Cortisol circadian rhythm
8am cortisol	-0.268	0.000	0.068	0.322	0.048	0.488	0.491	0.000
4pm cortisol	-0.231	0.001	-0.086	0.209	-0.087	0.205	0.283	0.000
0am cortisol	-0.185	0.007	-0.065	0.343	0.091	0.186	0.259	0.000
ACTH stimulation test
0’ cortisol	-0.268	0.000	0.068	0.322	0.048	0.488	0.491	0.000
30’ cortisol	-0.248	0.005	0.021	0.818	0.008	0.929	0.316	0.000
60’ cortisol	-0.284	0.001	0.112	0.207	-0.027	0.762	0.340	0.000
0’ 17OHP	0.106	0.120	0.052	0.450	0.171	0.012	-0.157	0.021
30’ 17OHP	0.259	0.003	-0.091	0.312	0.139	0.118	-0.306	0.000
60’ 17OHP	0.264	0.003	-0.122	0.170	0.183	0.040	-0.287	0.001
1-mg dexamethasone suppression test
8am cortisol	-0.369	0.000	-0.142	0.081	-0.126	0.122	0.486	0.000

Correlations of other hormones with FAI, TT and DHEAS

For other hormones, FAI had a positive correlation with ACTH and a negative correlation with serum cortisol levels at 8 am, 4 pm and 0 am, which was consistent with the circadian rhythm of cortisol (p < 0.05). Moreover, the FAI was also negatively correlated with the cortisol concentration at baseline and after 30 and 60 minutes of ACTH stimulation and with the cortisol concentration following DST (p < 0.05). Although the FAI had no relationship with baseline 17OHP, it was positively associated with 17OHP at 30 minutes and 60 minutes after ACTH stimulation (p < 0.05). Furthermore, DHEAS was positively related to 17OHP at baseline and after 60 minutes of ACTH stimulation (p < 0.05) (Table 4). FAI, TT and DHEAS are not related to TSH or prolactin.

Variation in steroid regulation with distinct FAI levels

The FAI was also evaluated as percentiles. According to the 25th, 50th and 75th percentiles, FAI was divided into the following four subgroups: FAI ≤ 2.26, 2.26 < FAI ≤ 5.32, 5.32 < FAI ≤ 10.65 and FAI > 10.65. With augmented FAI levels, although ACTH, TSH and prolactin did not significantly change (p > 0.05), serum cortisol levels significantly decreased (p < 0.05). Cortisol levels at 8 am, 4 pm and 0 am, as measured by the circadian rhythm, were significantly greater in individuals with lower FAI values than in those with higher FAI and cortisol levels after DST (p = 0.000). Similarly, the cortisol levels in the ACTH stimulation test also revealed the same patterns (p = 0.001). In contrast, 17OHP levels in the ACTH stimulation test were significantly lower in patients with lower FAI values than in those with higher FAI values, especially at 30 and 60 minutes (p = 0.001). At the same time, DHEAS manifested similar trends with variations in FAI levels (p = 0.003) (Fig. 1).

Eighty-five of 217 PCOS patients who were positive for adrenal hyperandrogenism were included. To assess the predictive value of the FAI for adrenal hyperandrogenism, ROC analysis was performed (Fig. 2). The AUC was 0.724 (95% CI: 0.654–0.793, p = 0.000), and the optimal cutoff value of the FAI was 5.29, with a sensitivity of 71.8%, specificity of 64.2%, positive predictive value of 56.5% and negative predictive value of 77.9%. In total, 109 patients in this cohort of 217 PCOS patients had an elevated FAI of more than 5.29. According to the sensitivity of 71.8% of this index, it is predicted that 78 patients will have adrenal-origin androgen dominance.

The evaluation of detectable hyperandrogenism and its effect on metabolic profiles and hormone alterations provides promising insight into the underlying mechanism of PCOS. Nonetheless, some studies have illustrated a correlation between hyperandrogenism and metabolic disorders, and few studies have reported an association between hyperandrogenism and cortisol dysregulation. To this end, our assessment revealed the close impact of FAI on a variety of metabolic biomarkers, metabolic conditions, and steroid dysregulation.

In terms of three androgen indicators, FAI, TT and DHEAS, we found that FAI was positively related to metabolic parameters in PCOS women, including BMI, waist circumference, hip circumference, waist-to-hip ratio, blood pressure (BP), FPG, PPG, HbA1c, AUC_GLU, AUC_INS, HOMA-IR, TG, TC, LDL-C and UA. It also had a negative correlation with the ISI score and HDL-C level. This finding is in line with a previous cross-sectional study demonstrating that FAI is superior to FT and TT because of its association with FPG, HOMA-IR and various lipid indices and it is a more sensitive indicator of weight, BMI, waist circumference, hip circumference, and waist-to-hip ratio. [16] Although recent international guidelines suggest that the calculated FAI, FT or calculated bioavailable testosterone should be preferable for distinguishing hyperandrogenism in patients with PCOS,(1)¹ further research indicates that the FAI, as it considers both the influences of testosterone and SHBG, appears to be a better parameter.[16] One study revealed that HbA1c, plasma glucose concentrations, and insulin levels increased with increasing FAI.[17] Another study reported that TT was positively associated with UA levels and the prevalence of hyperuricemia in females with PCOS.[18] However, we are not aware of any data on FAI and its relationship with UA. In addition, DHEAS, which reflects only adrenal hyperandrogenism, was not related to metabolic abnormalities. The present study suggested that the FAI is a preferred hyperandrogenism index for the prediction of hyperuricemia compared to TT and DHEAS. Thus, FAI is a more potent predictor of most abnormal metabolic parameters in women with PCOS.

Additionally, FAI was independently associated with the incidence of multiple metabolic disorders, such as overweight/obesity, fatty liver, dyslipidemia, insulin resistance, hyperglycemia, metabolic syndrome, and hyperuricemia. Previously, in a study, independently associated with NAFLD in women with PCOS.[5] Other researchers have suggested that the FAI might serve as an indicator of insulin resistance and glucose intolerance in patients with PCOS.[17] There is a review revealed that 38–88% of PCOS patients are overweight or obese.[19] In contrast to nonobese PCOS subjects, obese subjects are characterized by higher TT levels.[20] However, to our knowledge, no study has revealed any association between FAI and the incidence of overweight/obesity and hyperuricemia. Our research is the first to show that the FAI is a distinctive predictor of overweight/obesity and hyperuricemia in patients with PCOS. Overall, we demonstrated a greater FAI in patients with most abnormal metabolic indices, and moreover, we found the FAI to be an independent risk factor for the incidence of various metabolic disturbances in women with PCOS.

The findings regarding reproductive hormones showed that TT was correlated with LH, LH/FSH, E2, SHBG and AMH in concert with the secretion of TT via the HPA axis. In contrast, FAI and DHEAS were negatively associated with SHBG but not with other reproductive hormones. Testosterone and estrogen are mainly synthesized in both the ovaries and adrenal cortex.[20] DHEAS is the only androgen secreted from the adrenal cortex in response to adrenocorticotropic hormone (ACTH) to induce adrenal hyperandrogenism without regulating GnRH-LH.[21] Thus, DHEAS is not related to gonadotropin but is positively correlated with FAI.

The FAI is a characterization index of testosterone levels after correcting for SHBG, which is produced by the liver to bind to circulating sex steroids with high affinity, regulate the concentration of bioactive sex hormones in the blood and affect their bioavailability as a transporter of sex hormones. Furthermore, insulin can affect the expression of androgens by suppressing the concentration of SHBG, and decreased SHBG secretion is considered a marker of both IR and hyperandrogenaemia. Therefore, without overlap between PCOS patients and non-PCOS patients, the FAI is a comprehensive indicator of the level of bioactive testosterone after adjusting for abnormal SHBG and reflects not only the level of FT but also the level of IR, another crucial indicator of PCOS.[22] Similarly, another Chinese study also demonstrated that women with PCOS and a high FAI (FAI ≥ 5) had significantly greater TT, DHEAS, and HOMA-IR than women with a low FAI (FAI < 5) or controls. Similarly, there was no difference in LH/FSH, prolactin or progesterone between different levels of FAI.[23] According to a descriptive study, compared with the FAI, DHEAS was a weaker marker for diagnosis according to receiver operating characteristic (ROC) curve analysis.[24] Another study reported that the optimal cutoff value of the FAI measured by LC‒MS/MS was 2.5, with a sensitivity of 87.0% and specificity of 92.68% for representing hyperandrogenism in patients with PCOS.[25] In addition, a recent meta-analysis revealed that a moderate diagnostic value of the FAI for PCOS could be improved by a better cutoff point.[26] Thus, the FAI is a more comprehensive marker for PCOS than TT and DHEAS are.

In the present study, we first showed that the FAI was negatively related to adrenal steroids in women diagnosed with PCOS. Women with PCOS tend to have clinical and/or biochemical hyperandrogenism.[27] Current or other studies have confirmed that hyperandrogenism is closely related to obesity and insulin resistance.[28] Multiple studies have also reported that, compared to that in the general population, the incidence of depression in the PCOS population is increasing, but the underlying mechanism between hyperandrogenism and depression has not been identified.[29] Although cortisol is acknowledged as a highly sensitive biomarker for stress-related changes in metabolic homeostasis, high levels of salivary cortisol have been reported in PCOS patients.[30] Nevertheless, a recent case‒control study revealed a nonsignificant difference between PCOS patients and healthy controls.[31] However, there is still a lack of research on the relationship between hyperandrogenism and the level of cortisol in the PCOS population. However, the present study revealed a negative association between the FAI and the amount of cortisol filling in the gaps.

The present study revealed that the FAI was negatively related to basal, stimulated and suppressed cortisol levels and was positively associated with ACTH, DHEAS and stimulated 17OHP in patients with PCOS. The findings established that patients with a greater FAI were more likely to have adrenal hyperandrogenism due to a decrease in circulating cortisol with overactivation of ACTH and hyperstimulation of the adrenal androgen pathway, which suggested that twenty-one-hydroxylase–deficient nonclassic congenital adrenal hyperplasia (NC-CAH) has a familial predisposition under a complex genetic mechanism.[32] However, the current study examined 17OHP levels both at baseline and after ACTH stimulation to exclude NC-CAH.[32, 33] A meta-analysis revealed that the prevalence of adrenal hyperandrogenism in women with the classic PCOS phenotype was 28% (95% CI: 15–43%), with high heterogeneity.[34] A recent Chinese study reported that 17% of PCOS patients had adrenal-origin androgen dominance, with TT significantly decreasing after 2 days of dexamethasone administration.[25] In the present study, 40% of PCOS patients had adrenal hyperandrogenism, and PCOS patients with higher levels of FAI likely experienced more severe adrenal hyperandrogenism. On the other hand, some basic investigations have shown that a reduction in extra adrenal cortisol regeneration from cortisone might be caused by decreased hepatic type 1 11β-hydroxysteroid dehydrogenase activity resulting from ovarian hyperandrogenaemia.[15] In addition, hyperinsulinism, compared with insulin resistance, could also increase liver cortisol clearance via 5β-reductase. Subsequently, a decrease in circulating cortisol was driven by both scenarios with compensatory activation of the hypothalamic‒pituitary‒adrenal axis.[15] Additionally, hyperinsulinism itself and dietary factors favor P450c17 Δ⁵-17,20-lyase activity by means of Ser/Thr residue phosphorylation and thus 17OHP to DHEA in both the adrenal gland and ovary.[35]³⁷ Furthermore, although age range-based cutoff values of circulating DHEAS were utilized to define adrenal hyperandrogenism, the easily acquired FAI could be a promising biomarker for predicting adrenal hyperandrogenism. We first performed ROC analysis, and the optimal cutoff value for the FAI was 5.29, with a sensitivity of 71.8% and a specificity of 64.2%. Larger-scale clinical studies should be performed to further confirm these findings.

Limitations of the current study will also be addressed. This was a single-center cross-sectional study with a limited number of patients included, although a multicenter study is preferred. In addition, the present study did not recruit healthy controls for comparison, as there were logistical difficulties for healthy participants in accomplishing the 1-mg dexamethasone suppression test and ACTH stimulation test. Accordingly, other basic studies are needed to determine the exact mechanism of adrenal hyperandrogenism and its effect.

The present study showed that the FAI is a more comprehensive biomarker for PCOS than TT and DHEAS because it is an independent risk factor for multiple metabolic disorders, including overweight/obesity, fatty liver disease, dyslipidemia, insulin resistance, hyperglycemia, metabolic syndrome, and hyperuricemia. Moreover, the FAI is a promising biomarker for predicting adrenal hyperandrogenism, with the best cutoff point (value = 5.29). These novel results suggest that the FAI should be a better diagnostic and follow-up measurement for Chinese women with PCOS and that it should also be an alternative marker of adrenal hyperandrogenism. Our observations encourage further exploration, which is warranted to examine the mechanism of adrenal hyperandrogenism and its further effect, as well as to perform larger-scale clinical trials to confirm the current findings.

ACTH	Adrenocorticotropic Hormone
A4	Androstenedione
AFC	Antral Follicle Count
AMH	Anti-Müllerian Hormone
ANCOVA	Analysis of Covariance
AUC_GLU	Areas Under the Curve for Glucose
AUC_INS	Areas Under the Curve for Insulin
BMI	Body Mass Index
BP	Blood Pressure
CAH	Congenital Adrenal Hyperplasia
CI	Confidence Interval
DBP	Diastolic Blood Pressure
DHEA	Dehydroepiandrosterone
DHEAS	Dehydroepiandrosterone Sulfate
DST	Dexamethasone Suppression Test
E2	Estradiol
FAI	Free Androgen Index
FG	Ferriman-Gallwey
Fins	fasting insulin
FPG	Fasting Plasma Glucose
FSH	Follicle-Stimulating Hormone
FT	Free Testosterone
GnRH	Gonadotropin-Releasing Hormone
HbA1c	Hemoglobin A1c
HDL-C	High-Density Lipoprotein Cholesterol
HOMA-IR	Homeostatic Model Assessment of Insulin Resistance
HPA	Hypothalamus-Pituitary-Adrenal
ISI	Insulin Sensitivity Index
LDL-C	Low-Density Lipoprotein Cholesterol
LH	Luteinizing Hormone
MI NAFLD	Menstrual Irregularity Non-alcoholic fatty liver disease
OGTT	Oral Glucose Tolerance Test
OR	Odds Ratio
PCOM	Polycystic ovary morphology
PCOS	Polycystic Ovary Syndrome
PRL	Prolactin
ROC	Receiver Operating Characteristic
SBP	Systolic Blood Pressure
SEM	Standard Error of the Mean
SHBG	Sex Hormone-Binding Globulin
17OHP	17-Hydroxyprogesterone
TC	Total Cholesterol
TG	Triglycerides
TSH	Thyroid Stimulating Hormone
TT	Total Testosterone
2h PPG	2-hour postprandial glucose
UA	Uric Acid

Ethics approval and consent to participate

The study received approval from the Ethics Committee of The Third Affiliated Hospital of Guangzhou Medical University. All participants were informed about the purpose of the study and provided written informed consent before participation.

Consent for publication

Not applicable.

Availability of data and materials

The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests

The authors declare that they have no competing interests.

Funding

This study was supported by Teaching Reform Research Project of Clinical Teaching Base in Guangdong Province (No. 2021JD102), and Basic and Applied Basic Research Program of Guangzhou Municipal Science and Technology Bureau (No. 202201020122).

Author contributions

Renyuan Li, Yirui Zhu and Yijuan Xie contributed to the preliminary study design, data collection and statistical analysis. Renyuan Li was involved in the integrity of the data and manuscript writing. Siyuan Zheng contributed to the revision and submission. Ying Zhang is the guarantor of this work and, as such, has full access to all the data in the study and takes responsibility for the final study design and quality control of the clinical trial. All the authors approved the final version of the manuscript.

Acknowledgments

The authors thank the staff at the Endocrinology Department, the Third Affiliated Hospital of Guangzhou Medical University, for their assistance with study execution and data collection. In particular, we want to thank the patients for their participation in this trial.

Teede HJ, Misso ML, Costello MF, Dokras A, Laven J, Moran L, et al. Recommendations from the international evidence-based guideline for the assessment and management of polycystic ovary syndrome†‡. Hum Reprod. 2018;33(9):1602–18.
Emanuel RHK, Roberts J, Docherty PD, Lunt H, Campbell RE, Möller K. A review of the hormones involved in the endocrine dysfunctions of polycystic ovary syndrome and their interactions. Front Endocrinol. 2022;13:1017468.
Elhassan YS, Idkowiak J, Smith K, Asia M, Gleeson H, Webster R, et al. Causes, Patterns, and Severity of Androgen Excess in 1205 Consecutively Recruited Women. J Clin Endocrinol Metab. 2018;103(3):1214–23.
Borzan V, Lerchbaum E, Missbrenner C, Heijboer AC, Goschnik M, Trummer C, et al. Risk of Insulin Resistance and Metabolic Syndrome in Women with Hyperandrogenemia: A Comparison between PCOS Phenotypes and Beyond. J Clin Med. 2021;10(4):829.
Hong S hyeon, Sung YA, Hong YS, Song DK, Jung H, Jeong K, et al. Non-alcoholic fatty liver disease is associated with hyperandrogenism in women with polycystic ovary syndrome. Sci Rep. 2023;13(1):13397.
Aflatounian A, Edwards MC, Paris VR, Bertoldo MJ, Desai R, Gilchrist RB, et al. Androgen signaling pathways driving reproductive and metabolic phenotypes in a PCOS mouse model. 2020;245(3):381–95.
Cui P, Hu W, Ma T, Hu M, Tong X, Zhang F, et al. Long-term androgen excess induces insulin resistance and non-alcoholic fatty liver disease in PCOS-like rats. J Steroid Biochem Mol Biol. 2021;208:105829.
Kumari S, Agarwal R. Comparative study of cortisol level between PCOS women and normal women. Int J Clin Obstet Gynaecol. 2020;4(6):341–4.
Benjamin JJ, Kuppusamy M, Koshy T, Kalburgi Narayana M, Ramaswamy P. Cortisol and polycystic ovarian syndrome – a systematic search and meta-analysis of case–control studies. Gynecol Endocrinol. 2021;37(11):961–7.
Damone AL, Joham AE, Loxton D, Earnest A, Teede HJ, Moran LJ. Depression, anxiety and perceived stress in women with and without PCOS: a community-based study. Psychol Med. 2019;49(9):1510–20.
The Rotterdam ESHRE/ASRM‐sponsored PCOS consensus workshop group. Revised 2003 consensus on diagnostic criteria and long‐term health risks related to polycystic ovary syndrome (PCOS). Hum Reprod. 2004;19(1):41–7.
Zhao X, Ni R, Li L, Mo Y, Huang J, Huang M, et al. Defining hirsutism in Chinese women: a cross-sectional study. Fertil Steril. 2011;96(3):792–6.
Matsuda M, DeFronzo RA. Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp. Diabetes Care. 1999;22(9):1462–70.
Grundy SM, Brewer HB, Cleeman JI, Smith SC, Lenfant C. Definition of Metabolic Syndrome. Circulation. 2004;109(3):433–8.
Luque-Ramírez M, Escobar-Morreale HF. Adrenal Hyperandrogenism and Polycystic Ovary Syndrome. Curr Pharm Des. 2016;22(36):5588–602.
Zhang D, Gao J, Liu X, Qin H, Wu X. Effect of Three Androgen Indexes (FAI, FT, and TT) on Clinical, Biochemical, and Fertility Outcomes in Women with Polycystic Ovary Syndrome. Reprod Sci. 2021;28(3):775–84.
Zhang B, Wang J, Shen S, Liu J, Sun J, Gu T, et al. Association of Androgen Excess with Glucose Intolerance in Women with Polycystic Ovary Syndrome. BioMed Res Int. 2018;2018(1):6869705.
Mu L, Pan J, Yang L, Chen Q, Chen Y, Teng Y, et al. Association between the prevalence of hyperuricemia and reproductive hormones in polycystic ovary syndrome. Reprod Biol Endocrinol. 2018;16(1):104.
Barber TM, Franks S. Obesity and polycystic ovary syndrome. Clin Endocrinol (Oxf). 2021;95(4):531–41.
Moran C, Arriaga M, Rodriguez G, Moran S. Obesity differentially affects phenotypes of polycystic ovary syndrome. Int J Endocrinol. 2012;2012(1):317241.
Miller WL, Auchus RJ. The molecular biology, biochemistry, and physiology of human steroidogenesis and its disorders. Endocr Rev. 2011;32(1):81–151.
Rosenfield RL, Ehrmann DA. The Pathogenesis of Polycystic Ovary Syndrome (PCOS): The Hypothesis of PCOS as Functional Ovarian Hyperandrogenism Revisited. Endocr Rev. 2016;37(5):467–520.
Li H, Xu X, Wang X, Liao X, Li L, Yang G, et al. Free androgen index and Irisin in polycystic ovary syndrome. J Endocrinol Invest. 2016;39(5):549–56.
Khan SH, Rizvi SA, Shahid R, Manzoor R. Dehydroepiandrosterone Sulfate (DHEAS) Levels in Polycystic Ovarian Syndrome (PCOS). J Coll Physicians Surg Pak. 2021;31(3):253–7.
Chen F, Chen M, Zhang W, Yin H, Chen G, Huang Q, et al. Comparison of the efficacy of different androgens measured by LC-MS/MS in representing hyperandrogenemia and an evaluation of adrenal-origin androgens with a dexamethasone suppression test in patients with PCOS. J Ovarian Res. 2021;14(1):32.
Wang L, Li J. The value of serum-free androgen index in the diagnosis of polycystic ovary syndrome: A systematic review and meta-analysis. J Obstet Gynaecol Res. 2021;47(4):1221–31.
Naamneh Elzenaty R, du Toit T, Flück CE. Basics of androgen synthesis and action. Best Pract Res Clin Endocrinol Metab. 2022;36(4):101665.
Zeng X, Xie Y jie, Liu Y ting, Long S lian, Mo Z cheng. Polycystic ovarian syndrome: Correlation between hyperandrogenism, insulin resistance and obesity. Clin Chim Acta. 2020;502:214–21.
Kolhe JV, Chhipa AS, Butani S, Chavda V, Patel SS. PCOS and Depression: Common Links and Potential Targets. Reprod Sci Thousand Oaks Calif. 2022;29(11):3106–23.
Basu BR, Chowdhury O, Saha SK. Possible Link Between Stress-related Factors and Altered Body Composition in Women with Polycystic Ovarian Syndrome. J Hum Reprod Sci. 2018;11(1):10–8.
Marschalek ML, Marculescu R, Schneeberger C, Marschalek J, Dewailly D, Ott J. A case-control study about markers of stress in normal-/overweight women with polycystic ovary syndrome and in controls. Front Endocrinol. 2023;14:1173422.
Carmina E, Dewailly D, Escobar-Morreale HF, Kelestimur F, Moran C, Oberfield S, et al. Non-classic congenital adrenal hyperplasia due to 21-hydroxylase deficiency revisited: an update with a special focus on adolescent and adult women. Hum Reprod Update. 2017;23(5):580–99.
Maffazioli GDN, Bachega TASS, Hayashida SAY, Gomes LG, Valassi HPL, Marcondes JAM, et al. Steroid Screening Tools Differentiating Nonclassical Congenital Adrenal Hyperplasia and Polycystic Ovary Syndrome. J Clin Endocrinol Metab. 2020;105(8):dgaa369.
Barendregt JJ, Doi SA, Lee YY, Norman RE, Vos T. Meta-analysis of prevalence. J Epidemiol Community Health. 2013;67(11):974–8.
Li X, Hu S, Zhu Q, Yao G, Yao J, Li J, et al. Addressing the role of 11β-hydroxysteroid dehydrogenase type 1 in the development of polycystic ovary syndrome and the putative therapeutic effects of its selective inhibition in a preclinical model. Metabolism. 2021;119:154749.

No competing interests reported.

Impact of FAI on Metabolism and Adrenal Hyperandrogenism in Women with Polycystic Ovary Syndrome: a cross-sectional study

Status:

Version 1

Abstract

Background

Methods

Results

Conclusions

Trial Registration

Figures

Introduction

Methods

Statistical analysis

Results

Discussion

Conclusion

Abbreviations

Declarations

References

Additional Declarations

Status:

Version 1