World Journal of Urology

, Volume 29, Issue 4, pp 541–546

The influence of age on bioavailable and free testosterone is independent of body mass index and glucose levels

Authors

    • Curso de Pós-Graduação em Ciências da SaúdeUniversidade Federal de Ciências da Saúde de Porto Alegre (UFCSPA)
  • Ernani Luis Rhoden
    • Universidade Federal de Ciências da Saúde de Porto Alegre (UFCSPA)
    • CNPq-Conselho Nacional de Pesquisa, Researcher Level 2
    • Médico Hospital Moinhos de Vento
  • Charles Edison Riedner
    • Urology Unit of Military Hospital of Porto Alegre
Original Article

DOI: 10.1007/s00345-011-0724-x

Cite this article as:
Halmenschlager, G., Rhoden, E.L. & Riedner, C.E. World J Urol (2011) 29: 541. doi:10.1007/s00345-011-0724-x

Abstract

Purpose

To evaluate the influence of age on serum levels of total testosterone (TT), bioavailable testosterone (BT), free testosterone (FT), and sex-hormone binding globulin (SHBG), considering the presence of fasting blood glucose levels and body mass index (BMI) in a selected male population.

Methods

A total of 428 men were analyzed. Anthropometry was taken from all, considering BMI as general obesity indicative variable. Fasting blood samples were drawn for determination of plasma glucose levels and serum levels of albumin, TT, and SHBG. The values of BT and FT were calculated from TT, SHBG, and albumin, by Vermeulen’s equation. Statistical significance was set at P ≤ 0.05.

Results

Age was negatively correlated to BT (r = −0.301; P < 0.001) and FT (r = −0.273; P < 0.001), but not to TT levels (r = 0.002, P = 0.974). Age was positively correlated to SHBG (r = 0.376; P < 0.001). Age was independently associated with the occurrence of high SHBG levels (OR = 1.07, 95%CI = 1.05–1.10, P < 0.001) and of low BT (OR = 1.04, 95%CI = 1.02–1.07, P < 0.001) and FT levels (OR = 1.05, 95%CI = 1.03–1.08, P < 0.001), but not with low levels of TT (P = 0.08).

Conclusions

Age was significantly associated to high levels of SHBG and to low levels of BT and FT, without significant association to TT. This pattern was independent of BMI and glucose levels.

Keywords

AgingAndrogensBlood glucoseBody mass index

Introduction

The ageing process in men is accompanied by many health-related problems, such as cardiovascular diseases, diabetes, osteoporosis, sexual dysfunction, depressed mood, endocrine deficiencies, decreased muscle and bone mass, and decreased muscle strength as well as increased body fat [1, 2]. Many of these age-related clinical symptoms are closely associated with androgen deficiency (AD), which has been defined as a clinical syndrome characterized by subnormal levels of serum testosterone [3].

The age-related decline in androgen levels is well documented by both cross-sectional [410] and longitudinal [1114] studies, but the data have not been conclusive, especially, when different populations are considered as well as when potential confounding factors are included in the analyses. The Massachusetts Male Aging Study (MMAS) demonstrated that total testosterone (TT) and bioavailable testosterone (BT) levels decline longitudinally by 1.6 and 2–3% per year, respectively [12]. However, in some studies, lower levels of TT were not clearly associated to increasing age [710], raising questions regarding the role of chronic conditions related to the ageing process versus the ageing process per se in changing androgen levels, especially diabetes and obesity, two common conditions in aged men [15, 16].

Although it is well known that obesity and the presence of high fasting serum glucose levels and diabetes significantly exacerbate the age-related decrease of serum TT levels in men [16], little is known about the influence of overweight on androgen levels associated with increasing age. Furthermore, few studies have adjusted their results for obesity and glucose levels concurrently [8, 11, 14].

Therefore, the aim of the present study was to evaluate the influence of age on serum androgen levels considering the presence of fasting blood glucose levels and body mass index (BMI) in a selected male population presenting to a primary medical care practice for routine evaluation.

Materials and methods

Study design and population

This cross-sectional study was carried out in a large public outpatient clinic of primary health care, since 2009. A total of 428 men, aged 30–83 years, were eligible and were recruited for this study, during their regular visits to a primary care physician. Exclusion criteria included the presence of self-reported diabetes mellitus (DM) or use of hypoglycemic drugs, cirrhosis or any other liver disease or neoplasic condition, pelvic surgery due to prostate cancer [17], psychiatric and cardiovascular diseases, use of mood stabilizers, psychotropic and anxiolytic agents as well as medications that affect the endocrine system, use of any kind of drug abuse (but cigarette and alcohol intake), illiteracy, and individuals with obesity class II and III according to the National Institutes of Health guidelines [18].

All the volunteers signed the written informed consent. The study protocol was approved by our local Institutional Ethical Committee for Research.

Considering the objective of the study men were divided into five different groups according to age ranges, as follows: 30–40 years (N = 74), 41–50 (N = 121), 51–60 (N = 138), 61–70 (N = 79), and >70 (N = 16).

Clinical examination

All the clinical examination was performed by a single trained medical professional (ELR). Weight (kg) was measured using a scale (Filizola, São Paulo, Brazil) and height (cm) was obtained using a stadiometer. BMI was calculated dividing the weight (kg) by the height squared (m2). The anthropometry was performed in standing subjects wearing light clothes and without shoes.

Blood pressure was measured twice with a standard sphygmomanometer at the right arm after a 10-min rest in supine position, and the average of the measurements was used in the analyses.

Potential confounding factors

Some common potential confounding factors associated with alteration in hormone levels in aged men, such as fasting blood glucose [16, 19] and BMI, were investigated and included in the analyses.

Laboratory measurements

Blood samples were drawn between 8:00 and 12:00 AM after an overnight fast for determination of plasmatic glucose and serum levels of albumin, TT, and sex-hormone binding globulin (SHBG).

Glucose and albumin were measured by colorimetric assay (Labquest–Labtest Diagnostic, MG, Brazil). The levels of TT were determined by electrochemiluminescence immunoassay–ECLIA method (Elecsys 2010; Roche Diagnostics, Indianapolis, IN, USA), and the levels of SHBG were measured by chemiluminescent immunometric assay (Immulite®, DPC, Los Angeles, CA, USA). The values of free testosterone (FT) and of BT were calculated from TT, SHBG, and albumin according to a valid method proposed by Vermeulen et al. [20].

The normal ranges for the exams considered in the study were as follows: glucose (70–99 mg/dL), albumin (3.5–5.5 g/dL), TT (280–800 ng/dL), FT (2.62–16.7 ng/dL), BT (131–682 ng/dL), and SHBG (13–71 nmol/L). All assays were performed in the same laboratory.

Statistical analyses

Data were expressed as mean ± standard deviation (SD), unless otherwise stated. All variables were tested for normality, and the non-parametric data were log transformed. The correlations between age and serum hormone levels were tested by Pearson’s test. Analyses of covariance (ANCOVA) were used to estimate adjusted means of TT, BT, FT, and SHBG in the different age groups, controlled by fasting glucose levels and BMI. These two control variables were taken also in logistic regression models of age to predict the risk of low hormone levels, defined as the lowest quartile of values for FT (≤6.5 ng/dL), BT (≤156.0 ng/dL), and TT (≤333.0 ng/dL), and of high SHBG levels, defined as the highest quartile of values (≥43.8 nmol/L). Odds ratio (OR) and 95% confidence intervals (95%CI) were obtained.

All the statistical analyses were performed using SPSS statistical software package version 17 for Windows (SPSS Inc., Chicago, IL, USA). Statistical significance was set at P ≤ 0.05.

Results

The demographic and clinical characteristics of the study population are summarized in Table 1. The volunteers were 30–83 years old with mean ± SD age of 51.5 ± 10.7. Besides that, 24.1% were current smokers, and 94.4% of the studied sample was considered Caucasian.
Table 1

Demographic and clinical characteristics of the study population

Characteristics

General population (N = 428)

Age (years)

51.5 ± 10.7

Current smoker (%)

24.1

Caucasian (%)

94.4

Weight (kg)

80.6 ± 12.0

Height (m)

1.72 ± 0.06

Body mass index (kg/m2)

26.9 ± 3.5

Systolic blood pressure (mmHg)

126.3 ± 13.7

Diastolic blood pressure (mmHg)

82.5 ± 8.8

Total testosterone (ng/dL)

440.7 ± 155.1

Bioavailable testosterone (ng/dL)

214.3 ± 82.7

Free testosterone (ng/dL)

8.9 ± 3.6

Sex-hormone binding globulin (nmol/L)

31.1 [21.4–43.8]

Glycaemia (mg/dL)a

94.7 ± 30.3

Albumin (g/dL)

4.5 ± 0.6

Data are expressed as mean ± SD, as median [quartiles], and as percentages

aN = 426

Using Pearson’s correlation test, age was negatively correlated to plasma concentration of BT (r = −0.301; P < 0.001) and FT (r = −0.273; P < 0.001), but not to TT (r = 0.002; P = 0.974). Additionally, age was significantly correlated to SHBG (r = 0.376; P < 0.001) (Fig. 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs00345-011-0724-x/MediaObjects/345_2011_724_Fig1_HTML.gif
Fig. 1

Correlation between a age and total testosterone, b age and bioavailable testosterone, c age and free testosterone, and d age and sex-hormone binding globulin (SHBG). SHBG was log transformed

Table 2 depicts the mean and 95% confidence interval (95%CI) of the serum hormones considered in this study, adjusted for fasting blood glucose and BMI. There was a stepwise linear decrease in the mean values of BT and FT with increasing age. In contrast, TT remained relatively stable among age groups. In addition, a stepwise linear increase in the values of SHBG was observed with increasing age.
Table 2

Adjusted means (95% confidence intervals) of hormones and protein levels according to age range in men

Variable

Age range

 

30–40 years (n = 74)

41–50 years (n = 121)

51–60 years (n = 136)

61–70 years (n = 79)

>70 years (n = 16)

P trend*

TT (ng/dL)

439.0 (404.8–473.1)

437.6 (411.0–464.1)

434.2 (409.2–459.2)

475.5 (442.6–508.3)

379.1 (305.9–452.3)

0.632

BT (ng/dL)

255.6 (237.4–273.9)

227.0 (212.9–241.2)

200.0 (186.7–213.3)

192.2 (174.7–209.7)

167.3 (128.3–206.4)

<0.001

FT (ng/dL)

10.4 (9.6–11.2)

9.6 (8.9–10.2)

8.2 (7.6–8.7)

8.1 (7.3–8.9)

6.8 (5.1–8.5)

<0.001

SHBG (nmol/L)

1.37 (1.32–1.41)

1.42 (1.39–1.46)

1.53 (1.50–1.57)

1.64 (1.60–1.69)

1.58 (1.48–1.68)

<0.001

TT total testosterone, BT bioavailable testosterone, FT free testosterone, SHBG sex-hormone binding globulin

Adjusted for fasting blood glucose and BMI

P for linear trend obtained by linear regression models

log transformed

A multivariate analysis was conducted in order to identify the OR and 95%CI for occurrence of low hormones and high SHBG levels according to age, considering glucose levels and BMI as the adjustment factors. Low hormone levels were defined as the lowest quartile of values for FT (≤6.5 ng/dL), BT (≤156 ng/dL), and TT (≤333.0 ng/dL). High SHBG levels were defined as the highest quartile of values (≥43.8 nmol/L). Results showed that age was independently associated with the occurrence of high SHBG levels (OR = 1.07, 95%CI = 1.05–1.10, P < 0.001) and of low BT (OR = 1.04, 95%CI = 1.02–1.07, P < 0.001) and FT levels (OR = 1.05, 95%CI = 1.03–1.08, P < 0.001), but not with low levels of TT (P = 0.08).

Discussion

The present study demonstrated an age-related decrease of both BT and FT, paralleling an age-related increase of SHBG. Conversely, TT remained stable among age groups. Additionally, the age factor was independent of BMI and glucose levels in the association with higher levels of SHBG and with lower levels of BT and FT, keeping absence of importance in respect of TT.

Our results are consistent with some large cross-sectional studies published so far [710]. Yeap et al. [7] analyzed 3645 men, aged 70–89 years, and found a 0.8% annual decline in FT, offset by an annual rise of SHBG in the order of approximately 1.6%, without significant changes in TT levels with increasing age. Although this sample was composed basically by older men and may not represent the general population as a whole, the volunteers of the abovementioned study were relatively healthy, which allows us to extend the findings to a typical population, aged 40–80 years. Similarly, a study conducted by Wu et al. [8] with 3220 men, aged 40–79 years, found that increasing age was significantly associated with a decrease in TT and FT levels and with an increase in SHBG levels. However, their data showed that the age trend in TT became statistically non-significant after adjusting for covariates (BMI, comorbidity, smoking, and alcohol intake), suggesting that the risk factors mentioned above may confound any effects of age [8]. Atlantis et al. [9] analyzed 1141 men, aged 35–81 years, and also found that increasing age was significantly associated with a decrease in BT and with an increase in SHBG levels, without significant association with TT, after adjustment for physical and lifestyle factors. Likewise, Fukai et al. [10] demonstrated that age was negatively associated with FT, but not with TT.

There is general agreement that serum testosterone levels decrease with increasing age. However, the exact mechanism by which age affects sex-hormone levels is still debated. It is believed that the declining gonadal function in aging men occurs due to a decrease in testicular Leydig cell numbers and in their capacity to secrete testosterone, as well as to a central defect in down-regulation phenomenon of the hypothalamic-pituitary feedback system [21, 22].

The causes of the increase in SHBG levels in aging men remain unclear; however, there are theories suggesting that the age-related decrease in growth hormone (GH) and in insulin-like growth factor I (IGF-I) might contribute to the increased SHBG levels [23]. It is well known that approximately 44% of circulating testosterone is inactive and tightly bound to SHBG, 54% is loosely bound to albumin, and 2% is a free hormone [7, 24]. Consequently, serum SHBG levels are the main factor that influences the biologically active fractions of testosterone, as albumin generally does not vary significantly in healthy men [4]. It is interesting to observe in the present study that TT serum levels remained relatively stable in the different age groups, with progressive increase of SHBG levels, which increases the binding of testosterone, reducing the proportion of BT and FT as age increased.

It is well recognized that obesity and type 2 DM are both conditions associated with low TT levels [8, 1619, 2527]. Besides that, the prevalence of both conditions increases with age, leading to a high degree of overlap between obesity, type 2 DM, and hypogonadism [15]. Nevertheless, the mechanisms and the cause-and-effect relationship between the abovementioned conditions remain to be elucidated [15, 28]. Recent studies have shown that TT, FT, and BT decrease with increasing BMI [29]. In this regard, Wu et al. [8] observed that obesity was the most important factor that influenced TT levels, independent of age. Kaplan et al. [16] clearly demonstrated that obesity and the presence of high fasting serum glucose levels are the key components of the metabolic syndrome that exacerbates the age-related decrease in TT levels. Furthermore, it has been documented that men with impaired glucose tolerance (IGT) have low TT levels and that TT levels are inversely associated to fasting plasma glucose [25]. Additionally, low TT levels and symptomatic hypogonadism are common conditions in men with type 2 DM and also with increased insulin resistance [25, 26].

Although there is a considerable body of evidence in the literature documenting the age-related decline in androgen levels [414], few studies have adjusted their results to BMI and glucose levels concurrently [8, 11, 14]. The Baltimore Longitudinal Study of Aging (BLSA) [11] demonstrated a longitudinal effect of age on TT and FT index, and such effect was independent of obesity, illness (including DM), medications, cigarette smoking, and alcohol intake. Correspondingly, the Massachusetts Male Aging Study (MMAS) estimated a longitudinal decline in TT by −1.1% per year, independent of chronic illness (DM included), general health, medications, smoking, BMI, employment, and marital status [12]. Therefore, cross-sectional and longitudinal studies that include obesity, overweight, and high fasting serum glucose/diabetes as confounders are strongly encouraged.

Regarding these aspects, the present study has the merit of evaluating the association between age and TT, FT, BT, and SHBG levels, independently of glycemia and BMI, in individuals who did not have DM diagnosis. This analysis enhances the evidence assembled from BLSA [11] and MMAS [12] , which focused on the general population, including individuals with long-term medical illness.

Some limitations should be recognized in the current study. First, both BT and FT were calculated from TT, SHBG, and albumin using the Vermeulen’s equation [20]. Although calculated FT is considered a reliable index of FT by equilibrium dialysis [3], calculated BT is criticized for its great uncertainty [30]. Instead of calculating BT, it is recommended to measure BT using the ammonium sulfate precipitation method [3, 30]. Unfortunately, this method was not available and is also expensive and not practical. Second, we have used a single blood sample to determine TT and SHBG levels. Greater precision could have been achieved by taking 2 or more samples. Third, we have assessed general obesity using crude BMI, thus residual confounding can not be excluded. Finally, we have not demonstrated the cross-sectional percentage of androgen decline, because longitudinal analyses are more reliable and may show up to fourfold-higher changes than cross-sectional studies [13].

In conclusion, the current study demonstrated that increasing age was significantly associated with higher levels of SHBG and with lower levels of BT and FT, without significant association with TT. This pattern is independent of BMI and glucose levels.

Acknowledgments

Dr. Rhoden is a researcher of ConselhoNacionaldeDesenvolvimentoCientíficoeTecnológico (CNPq)–“National Counsel of Technological and Scientific Development”. M.Sc. Halmenschlager is supported by CoordenaçãodeAperfeiçoamento de PessoaldeNívelSuperiorCAPES. We thank Federal University of Health Sciences of Porto Alegre (UFCSPA) for financial support.

Conflict of interest

None.

Copyright information

© Springer-Verlag 2011