Analysis of relationships between the concentrations of total testosterone and dehydroepiandrosterone sulfate and the occurrence of selected metabolic disorders in aging men

Abstract

Objective: The evaluation of relationships between the concentrations of dehydroepiandrosterone sulfate (DHEAS) and total testosterone (TT) and the occurrence of metabolic disorders, including metabolic syndrome (MetS).

Method: The participants were subjected to anthropometric measurements and were tested for DHEAS, TT, lipid parameters and carbohydrate parameters.

Result: We observed a lower concentration of DHEAS in the men with hypertension (HT) compared to those without HT. In the men with MetS, HT, overweight and obesity, the concentration of TT was lower than in the men without these problems. We found statistically significant positive correlations (DHEAS– total cholesterol [TCh], DHEAS– low-density lipoprotein [LDL], TT– high-density lipoprotein [HDL], TT–waist-to-hip ratio [WHR]) and negative correlations (DHEAS–age, TT–body weight, TT– body mass index [BMI], TT–abdominal circumference [AC], TT–hip circumference [HC], TT– triglyceride [TG], TT– fasting plasma glucose [FPG], TT– serum insulin levels [I], TT– Homeostasis Model Assessment–Insulin Resistance [HOMA–IR]). Using logistic regression it was ascertained that lower TT levels increase the risk of HT, and were also associated with obesity.

Conclusion: Our research indicates relationships between TT and the occurrence of MetS and its individual components. Excess body weight in men is a factor associated with lower TT levels. It seems necessary to determine TT in men with MetS and overweight or obesity. DHEAS did not show any significant relations with MetS and its parameters. Age was the most crucial factor responsible for the decrease in DHEAS.

• Aging men
• dehydroepiandrosterone
• metabolic disorders
• testosterone

View correction statement:

Corrigendum

Introduction

In men, the concentrations of the androgens dehydroepiandrosterone (DHEA), dehydroepiandrosterone sulfate (DHEAS) and testosterone (T) decrease with age. A decrease in DHEA is known as adrenopause, while a decrease in T may lead to the late-onset hypogonadism (LOH) [1].

DHEA in men is the precursor of about 50% of androgens [2]. Its highest concentrations are observed in men between the 20th and 30th year of age, while in elderly men (between the 70th and 80th year of age) DHEA levels do not exceed 10%–20% of those found in young men [3]. The cause of this age-related decrease in DHEA – also observed for DHEAS – has not been unequivocally established. It is supposed that a significant role may be played by the gradual disappearance of the zona reticularis [4].

The age-related decrease in DHEA is conducive to the occurrence of sarcopenia, ostheopenia, atherosclerosis and the development of cardiovascular diseases (CVD), immunological disorders and cognitive impairment [5–8]. In addition, low DHEA levels may lead to erectile dysfunction, even in men below 60 years of age, as well as decreased libido and depressed mood [9–11]. Low DHEAS in men, especially those older than 70 and in tobacco smokers, is associated with a high relative risk of death [12,13]. Accordingly, DHEA has been marketed as an “anti-aging” supplement claimed to protect against CVD and other chronic illnesses [14].

The production of T and its circadian secretion are subject to age-related impairment, which is associated with a dysregulation of the hypothalamic pulse generator. These changes are related to the natural course of aging, genetic factors and regressive changes in gonadotropic cells and in hypothalamic and pituitary circulation [15,16]. The testicles of aging men have a 50% lower number of Leydig and Sertoli cells than in young men. In addition, aging men experience deterioration in the functioning and blood supply to the testicles and regressive changes in the stroma of these organs [17–19]. These changes may be genetic, but are also associated with lifestyle, including addictions, intake of drugs and toxins, as well as atherosclerosis [16,18,19].

The results of a Massachusetts Male Aging Study (MMAS) indicate that from the 30th year of age the concentration of total testosterone (TT) reduces by 0.8–1% per year [20], while according to a multi-center European Male Aging Study (EMAS) TT concentration decreases by 0.4% annually, while for free testosterone (FT) this decrease is 1.3% per annum [21]. The age-dependent reference range for T has not been defined and criteria on the cutoff levels for hypogonadism remain somewhat controversial [22]. The EMAS research group proposes that LOH should be defined as the presence of at least three symptoms (less frequent sexual desire, erectile dysfunction and a decrease in the frequency of early morning erections), and a TT level lower than 11 nmol/l (3.2 ng/ml) or FT below 220 pmol/l (64 pg/ml) [23,24], but this is still not a uniformly accepted standard.

Data on the frequency of occurrence of LOH are varied and depend on the accepted criteria, both regarding the clinical symptoms and the accepted cut-off for T [23,25,26]. In an observational cohort study by MMAS conducted on healthy men aged 40–70 years, the prevalence of androgen deficiency (TT < 400 ng/dl) ranged from 25.3% to 39.3%, but when considering the presence of at least three signs or symptoms of low T, the prevalence dropped to 6%–12% [25]. According to the results of a multicenter EMAS study, LOH concerns 2.1% of men between 40 and 79 years of age and its incidence increases with age. When using more flexible criteria, the incidence of LOH increases in the abovementioned age range to 5.6%–12% [22].

The age-related decrease in T is associated with a higher incidence of depression, fatigue, osteoporosis and reduced muscle mass [23,27,28]. Low concentrations of T also lead to an increased risk of cardiovascular disease and premature mortality [29]. The presence of testosterone receptors in the myocardium indicates that T might have a direct impact on cardiac remodeling and the renin angiotensin system, contributing to congestive heart failure [30].

Studies suggest an association between age-related hypogonadism and the prevalence of metabolic syndrome (MetS), characterized by the presence of central obesity, insulin resistance, dyslipidemia and hypertension (HT) [31]. It is believed that central adiposity is the primary factor in the development of MetS, which contributes to the development of high blood pressure, increased concentration of low-density lipoprotein (LDL), reduced high-density lipoprotein (HDL) and increased glycemia [32].

The aim of this study was to evaluate the associations between serum DHEAS and TT and the presence of selected metabolic disorders in aging men. The metabolic disorders included MetS and its various parameters, as well as lipid profile (total cholesterol [TCh] and LDL) and carbohydrate metabolism (serum insulin levels [I] and Homeostasis Model Assessment–Insulin Resistance [HOMA–IR]), which are not included in the diagnostic criteria for MetS. We also analyzed the relation between DHEAS and TT and the indicators of nutritional status, including body weight, body mass index (BMI) and waist-to-hip ratio (WHR) in aging men.

Methods

The study involved 313 men aged 50–75 years (mean age 61.3 ± 6.3 years) who had volunteered after receiving information about the course and the aim of the study from their family doctors at the primary health care centers operating in the city of Szczecin (Poland). Criteria for exclusion from participation in the study included cancer treatments, receiving steroids (including DHEA and T), thyroid disease, receiving neuroleptics and antidepressants. The studies lasted from March 2013 to February 2014. The study was approved by the Bioethics Committee of the Pomeranian Medical University in Szczecin (KB-0012/159/12). The men participating in the study were informed about the course of the research project and completed a written consent to participate in the study.

The respondents were surveyed concerning demographics and the presence of chronic diseases. Blood pressure, weight, height, and abdominal and hip circumference (HC) were measured. BMI and WHR were calculated. One hundred twenty-one respondents (38.7%) declared professional activity, 37 patients (11.8%) received disability pension, 136 (43.5%) were retired and 19 men (6.1%) were unemployed.

Blood samples were taken in the morning on an empty stomach from an ulnar vein between the hours of 7.30 am and 9 am. For the biochemical and hormonal tests, the blood was collected into a tube with a coagulation activator and gel separator, and then centrifuged. The resulting serum was stored at −70 °C.

In the surveyed men, ELISA was used to determine the serum concentrations of DHEAS, TT and I, with the use of reagent kits (DRG Medtek, Warsaw, Poland). In addition, the concentrations of glucose, TCh, LDL, HDL and triglyceride (TG) were determined by spectrophotometric method using specialised reagent kits (BIOLABO, Aqua-Med, Łódź, Poland). In the non-diabetic subjects, HOMA–IR was calculated according to the following formula: HOMA–IR = fasting plasma glucose (mmol/L) × fasting insulin level (µU/ml)/22.5. MetS was diagnosed according to the criteria of the International Diabetes Federation (IDF) from 2005 [33].

Statistical analysis was performed using Stat View software v5.0 (SAS Institute Inc., Cary, NC). The group characteristics for the studied traits included arithmetic mean (AM) and standard deviation (SD), median (Med) and range. The normality of distributions was tested with the use of the Shapiro–Wilk tests. The analysis of correlations was performed using ANOVA tests. The correlations between the analyzed variables were calculated using Spearman’s rank correlation coefficients (r_s). Logistic regression was used for the calculation of the odds ratios (OR) and 95% confidence intervals (CI). The level of significance was p ≤ 0.005.

Results

In the participants, average body weight was 87 ± 15 kg, BMI 28.3 ± 7.7, waist circumference (WC) 102 ± 12 cm, WHR 1.0 ± 0.1, arterial blood pressure (ABP) systolic 135 ± 20 mmHg, and ABP diastolic 84 ± 20 mmHg (Table 1). TT concentration was 3.9 ± 1.6 ng/ml, while DHEAS concentrations ranged 0.01–4.61 μg/ml.

Table 1. Characteristics of the study group.

Parameter	n	AM ± SD	Med	Range
Height (m)	313	175.2 ± 0.6	175.0	158.0–194.0
Body weight (kg)	313	87.0 ± 15.0	85.0	52.0–135.0
BMI (mg/m^2)	313	28.3 ± 7.7	27.7	19.5–46.7
WC (cm)	313	102.0 ± 12.0	100.0	75.0–146.0
HC (cm)	313	103.0 ± 8.0	103.0	82.0–129.0
WHR	313	1.0 ± 0.1	1.0	0.8–1.2
ABP systolic (mmHg)	313	135.0 ± 20.0	135.0	80.0–220.0
ABP diastolic (mmHg)	313	84.0 ± 20.0	80.0	50.0–140.0
TT (ng/ml)	313	3.9 ± 1.6	3.5	0.6–9.8
FT (pg/ml)	313	105.2 ± 58.1	95.9	6.7–423.7
SHBG (nmol/l)	313	45.3 ± 24.3	42.3	3.9–192.0
E2 (pg/ml)	313	39.1 ± 21.5	35.4	5.6–133.7
DHEAS (μg/ml)	313	1.49 ± 0.84	1.3	0.01–4.61
TG (mg/dl)	313	144.4 ± 83.6	121.0	37.0–581.0
TCh (mg/dl)	313	219.0 ± 57.1	217.0	96.0–416.0
LDL (mg/dl)	313	142.0 ± 54.0	138.0	20.0–342.0
HDL (mg/dl)	313	48.6 ± 12.4	47.0	19.0–95.0
FPG (mg/dl)	313	107.0 ± 21.5	104.0	53.0–237.0
I (μIU/ml)	257	14.8 ± 7.7	13.4	1.2–66.4
HOMA–IR	257	4.0 ± 2.2	3.4	0.3–15.8

AM, arithmetic mean; SD, standard deviation; Med, median; BMI, Body Mass index; WC, Waist circumference; HC, Hip circumference; WHR, waist-to-hip ratio; ABP, arterial blood pressure; TT, total testosterone; FT, free testosterone; SHBG, sex hormone binding globulin; E₂, estradiol; DHEAS, dehydroepiandrosterone sulfate; TG, triglyceride; TCh, total cholesterol; LDL, low-density lipoprotein; HDL, high-density lipoprotein; FPG, fasting plasma glucose; I, serum insulin levels; HOMA–IR, Homeostasis Model Assessment–insulin resistance.

Type 2 diabetes mellitus (T2DM), HT and MetS was diagnosed in 17.6%, 54.6% and 51.4% of participants, respectively (Table 2). 16.3% of men smoked cigarettes while 15.7% were being treated with statins. No statistically significant differences were found in the concentrations of DHEAS between patients with and without MetS, and those with and without T2DM. Participants with HT had significantly lower DHEAS levels compared to those without HT. Patients treated with statins had lower DHEAS levels. There was no relationship between the level of DHEAS and nutritional status as measured by BMI (Table 2).

Table 2. The concentration of dehydroepiandrosterone sulfate (DHEAS) and total testosterone (TT) in relation to diseases and health problems in a study group.

	DHEAS		TT
Qualitative variables	AM ± SD (μg/ml)	p	AM ± SD (ng/ml)	p
MetS (n = 161)	1.34 ± 0.78	0.51	3.40 ± 1.33	<0.0001
Without MetS (n = 152)	1.41 ± 0.88		4.36 ± 1.76
T2DM (n = 55)	1.35 ± 0.80	0.17	3.48 ± 1.53	0.06
Without T2DM (n = 258)	1.52 ± 0.98		3.94 ± 1.63
HT (n = 171)	1.26 ± 0.78	0.008	3.43 ± 1.42	<0.0001
Without HT (n = 142)	1.52 ± 0.88		4.38 ± 1.71
Statin treatment (n = 49)	1.12 ± 0.59	0.019	3.70 ± 1.58	0.48
No statin treatment (n = 264)	1.43 ± 0.87		3.91 ± 1.64
Smoking (n = 51)	1.51 ± 0.84	0.13	3.89 ± 1.85	0.79
No smoking (n = 262)	1.45 ± 0.88		3.81 ± 1.58
Normal body weight (n = 62)	1.35 ± 0.80	p (0–1) = 0.68	4.67 ± 0.68	p (0–1) = 0.004
Overweight (n = 164)	1.30 ± 0.79	p (0–2) = 0.18	3.97 ± 1.53	p (0–2) < 0.0001
Obesity (n = 87)	1.46 ± 0.93	p (1–2) = 0.09	3.11 ± 0.86	p (1–2) < 0.0001

DHEAS, dehydroepiandrosterone sulfate; TT, total testosterone; AM, arithmetic mean; SD, standard deviation; p, probability for trend; MetS, metabolic syndrome; T2DM, type 2 diabetes mellitus; HT, hypertension.

In assessing the correlation between DHEAS and the evaluated parameters in the surveyed men, we found a statistically significant negative correlation between DHEAS and age, and positive relations between DHEAS and TCh and between DHEAS and LDL (Table 3). Logistic regression showed that a higher level of DHEAS decreased the risk of HT (p = 0.009, OR = 0.69, 95%CI: 0.529–0.915).

Table 3. Spearman correlation coefficients between age, anthropometric and biochemical parameters and the serum dehydroepiandrosterone sulfate (DHEAS) and total testosterone (TT).

	DHEAS		TT
Parameter	rs	p	rs	p
Age	−0.310	<0.0001	0.07	0.22
Body weight	0.077	0.17	−0.365	<0.0001
BMI	0.022	0.69	−0.406	<0.0001
WC	0.139	0.80	−0.377	<0.0001
HC	−0.034	0.80	−0.297	<0.0001
WHR	0.060	0.28	−0.352	<0.0001
ABP systolic	−0.040	0.47	−0.075	0.18
ABP diastolic	0.035	0.53	−0.089	0.11
TG	0.084	0.14	−0.314	<0.0001
TCh	0.209	0.0002	0.059	0.30
HDL	−0.069	0.22	0.197	0.0004
LDL	0.209	0.0002	0.081	0.15
FPG	−0.020	0.74	−0.145	0.02
I	0.075	0.53	−0.263	<0.0001
HOMA–IR	0.068	0.27	−0.291	<0.0001

DHEAS, dehydroepiandrosterone sulfate; TT, total testosterone; BMI, Body Mass index; WC, waist circumference; HC, hip circumference; WHR, waist-to-hip ratio; ABP, arterial blood pressure; TG, triglyceride; TCh, total cholesterol; HDL, high-density lipoprotein; LDL, low-density lipoprotein; FPG, fasting plasma glucose; I, serum insulin levels; HOMA–IR, Homeostasis Model Assessment–insulin resistance; p, probability for trend.

In the participants with MetS (51.4%), TT concentration was significantly lower than in those without MetS (Table 2). Between the diabetic and non-diabetic men the difference in TT concentration was close to statistical significance. In the participants with HT we found a highly significant difference in the concentration of TT compared to those without HT. Smokers had a higher TT concentration than non-smokers, but the difference was not statistically significant. In the overweight men (BMI between 25 and 29.99), TT was significantly lower than those with normal body weight. The obese men (BMI ≥ 30) were characterized by a highly significantly lower concentration of TT than the overweight men (Table 2).

Significant negative correlation coefficients were found for the relationships: TT–body weight, TT–BMI, TT–AC and TT–HC and TT–WHR (Table 3). The correlations TT–APB systolic and TT–APB diastolic were negative but not statistically significant. With respect to lipid parameters, there was a highly significant negative correlation TT–TG and a positive correlation TT–HDL (Table 3). The correlations TT–LDL and TT–TCh were positive, but not statistically significant. Among the parameters of glucose metabolism, significant correlations were found for the following pairs: TT–fasting plasma glucose (FPG), TT–I and TT–HOMA–IR (Table 3).

Using logistic regression it was found that a higher concentration of TT decreased the risk of both HT (p < 0.0001, OR = 0.67, 95%CI: 0.572–0.788) and MetS (p < 0.0001, OR = 0.66, 95%CI: 0.56–0.779). A lower TT level was also associated with the occurrence of overweight (p = 0.0087, OR = 1.798, 95%CI: 1.675–1.945) and obesity (p < 0.0001, OR = 1.48, 95%CI: 1.375–1.631).

Discussion

The results confirm that age is an important factor in the reduction in DHEAS. However, we found no relationship between DHEAS and the occurrence of MetS. This is consistent with the results of a study on non-smoking men (n = 130) by Blouin et al. [34] who also show no association between the occurrence of MetS and the concentration of DHEAS when adjusted for age. They suggest that age per se is a factor significantly associated with reduced levels of DHEA and with a higher incidence of metabolic disorders. Similarly, Maggio et al. [35] also do not find any association between MetS and the concentration of DHEAS in men from Italia over 65, just like Akishita et al. [36] in middle-aged men from Japan. In contrast, Rabijewski et al. [37] show that DHEA deficiency is a significant independent risk factor for MetS in non-obese men aged 60–70 years.

Our present study showed that men with MetS had significantly lower levels of TT, which is consistent with the studies by Laaksonen et al. [38] and Muller et al. [39]. Also Blouin et al. [34] demonstrate that a reduction in T levels significantly correlate to an increased risk of developing MetS, which is supported by Maggio et al. [35] who report higher frequency of MetS in men from Italia with lower TT levels. The results of Rodriguez et al. [40] also show that the prevalence of MetS increases with age and is associated with lower androgen levels.

The relationship between reduced levels of T and MetS is presented in cross-sectional and longitudinal studies of Muller et al. [39] and Laaksonen et al. [38]. Laaksonen et al. [38] stress that the relationship between low T and MetS levels and its components is independent of BMI. MetS can result in the development of CVD, the most frequent cause of death. Corona et al. [41] report a lower T concentration in men with CVD.

In Caucasian men with diabetes, the prevalence of hypogonadotropic hypogonadism ranges 25–40% [42–45]. A study on a representative sample of 1226 Americans by Li et al. [46] shows low TT levels being significantly associated with high levels of HOMA–IR, fasting serum insulin and T2DM. In this study, similar relations were found for TT and FPG, I, and HOMA–IR, and in subjects with T2DM – the difference in the TT concentration compared to non-diabetic men was close to statistically significant levels. Corona et al. [41] report lower T concentrations in obese men with T2DM, and so adipose tissue may modify the relationship between T and diabetes. In 2010 the Endocrine Society recommended the determination of T levels in men with DM who had reported sexual dysfunction, weakness, weight loss and mobility limitation [47]. According to the latest recommendations of The International Society for the Study of the Aging Male, it is advisable to determine T levels in men with obesity, MetS, and insulin resistance [48].

In our study we found no statistically significant differences in the DHEAS concentrations between the men with and without T2DM. There were statistically significant correlations for DHEA–FPG, DHEA–HOMA–IR and DHEA–I. The 5-year follow-up examinations by Kameda et al. [49] show a reduction in DHEAS levels in men from Japan being associated with an increased risk of T2DM, which is confirmed by Oh et al. [50]. In addition, Tchernof et al. [51] indicate that a decreased DHEAS concentration may contribute to the development of insulin resistance and T2DM, mainly due to the relations between T2DM and obesity.

Low T levels have also been reported to correlate with increased blood pressure [52]. Heufelder et al. [53] show that testosterone treatment of men with hypogonadism, MetS and newly diagnosed diabetes reduces blood pressure more than diet and exercise. A reduced TT concentration promotes arterial stiffness, which in turn is associated with HT [54].

In the present study men with HT had lower DHEAS levels. This is consistent with the results of Carroll et al. [55]. This relation does weaken when adjusted for age, BMI, health behaviors and socio-demographic factors, but remains on the borderline of statistical significance (p = 0.06). In contrast, Schunkert et al. [56] report a positive correlation between serum DHEA and ABP systolic. However, an unambiguous evaluation of the relationship between the DHEA concentrations and ABP requires further research; the difficulty lies in a large number of confounding factors that affect ABP.

Results of a multicenter population-based study of EMAS indicate that men with excess body weight have lower TT and FT levels [21]. This is consistent with studies by Cao et al. [57] conducted on a group of 314 men aged 65 years. Fui et al. [58] based on studies of other authors, confirm the existence of a strong link between obesity and low T concentrations.

We found that with an increase in body weight the TT concentration decreases, but no such relationship was found for DHEAS. The findings of other authors concerning the relationship between reduced DHEA of and obesity are ambiguous. Abbasi et al. [59] show no association between DHEAS and BMI in men aged 60–80 years, while in women they demonstrate a positive correlation between these two parameters. Rabijewski et al. [60] report that a shortage of DHEAS and TT are independent factors associated with obesity and insulin resistance. Tchernof et al. [51,61,62] in a meta-analysis present two studies that indicate a statistically significant negative correlation between serum DHEA in men and measure of abdominal fat distribution. Those studies suggest that lower concentrations of DHEA may contribute to increased accumulation of visceral fat. However, there are two other studies that present different results [63,64]. Tchernof et al. [63] report that in many studies the relationship between serum DHEAS and BMI becomes statistically insignificant when adjusted for age.

The results of this present study suggest that lower TT concentrations are conducive to higher TG and lower HDL. Stanworth et al. [65] demonstrate a statistically significant positive correlation of TT–HDL in men with T2DM. Agledahl et al. [66] in a cross-sectional study show that men with an unfavorable lipid profile (HDL < 0.90 and TG > 1.8) have significantly lower levels of TT when adjusted for age and BMI.

The relationship between DHEA and lipid parameters has been confirmed in many studies that suggest a beneficial effect of DHEA and its sulfate on the concentrations of HDL and TG [67–69]. Studies conducted among men from the Czech Republic also demonstrate a beneficial correlation between DHEA and its sulfate and lipid parameters, but following adjustment for age or age and BMI these correlations disappear, which suggests that the most crucial factor is the relationship between DHEA, age and BMI [70].

This theory is supported by Srinivasan et al. [14] according to whom long-term intake of DHEA does not affect the concentration of lipoproteins in aging men. Those authors are of the opinion that worse lipid profile parameters in aging men are mainly associated with age. However, many authors do not demonstrate associations between DHEA and DHEAS with lipid metabolism. Also in the present study we found no association between serum TG and HDL, or parameters considered as criteria for MetS and the concentration of DHEAS, while a positive correlation was found between Ch and LDL and the concentration of DHEAS.

Recently published research indicates that five-year testosterone replacement therapy in men with hypogonadism has a beneficial effect on body composition, metabolic parameters, body weight (weight loss), WC and BMI [71,72]. Traish et al. [73] point out that long-term treatment with T significantly reduces the concentration of TCh, LDL and TG, increases HDL, and has a positive effect on carbohydrate parameters and blood pressure, and consequently reduces the risk of CVD.

Tchernof and Labrie [51] in a meta-analysis emphasized the inconsistency of results in research on the relationships between DHEA or DHEAS and the parameters of lipid metabolism. Those researchers believe that many factors, such as age or the accumulation of visceral fat, have an impact on the relationship between DHEA and lipid profile.

Conclusions

Our research indicates a substantial connection between the concentration of TT and the occurrence of MetS and its various components. Excessive weight is a factor associated with a lower TT concentration. It therefore seems reasonable to determine TT concentrations in men with MetS and excessive BMI. DHEAS does not show significant associations with MetS and its parameters. Age is a key factor responsible for lower DHEAS levels.

Declaration of interest

The authors report no conflicts of interest in this work. The study was financed as research project no. FSN (stimulation fund research) WNoZ 321-10/13 by the Pomeranian Medical University in Szczecin.