Prace oryginalne/Original papers

The reference values of sex hormones and SHBG serum levels in subjects over 65 years old — The PolSenior Study

Wartości referencyjne stężenia hormonów płciowych oraz SHBG w surowicy u osób powyżej 65. roku życia — badanie PolSenior

Andrzej Milewicz1, Barbara Krzyżanowska-Świniarska2, Tomasz Miazgowski 2, Diana Jędrzejuk1, Anna Arkowska1, Urszula Mieszczanowicz1, Ewa Bar-Andziak3

1Medical University, Department of Endocrinology, Diabetology and Isotope Therapy, Wrocław, Poland
2Pomeranian Medical University, Department of Hypertension and Internal Medicine, Szczecin, Poland
3Medical University, Department of Internal Medicine and Endocrinology, Warsaw, Poland

Barbara Krzyżanowska-Świniarska M.D. Ph.D., Pomeranian Medical University, Department of Hypertension and Internal Medicine, Unii Lubelskiej 1, 71–252 Szczecin, Poland, tel.: +48 91 425 35 30, e-mail: bksam@pum.edu.pl

Abstract

Introduction: Over the last decade, average life expectancy has continuously increased. There has been no data on normal sex hormone (SH) levels in a Polish elderly population. In this study, we assessed SH in the PolSenior cohort to determine normal reference ranges in relation to gender, age, and cardiovascular disease risk factors (CVDRFs).

Material and methods: The study was performed with 4,352 participants (2,168 men and 2,088 women), aged from 55 to over 90 years, stratified in five-year age groups. Pre-elderly subjects (55–59 years of age) served as the reference group. We assessed total testosterone (TT), estradiol (TE2) and DHEA-S (by RIA) SHBG and FSH (by IRMA) and calculated free androgen and free estrogen indices (FAI and FEI). Percentage body fat (%BF) was measured by bioelectric impedance analysis.

The CVDRFs assessment included blood pressure and biochemical (blood glucose, high-density lipoproteins, triglycerides) and anthropometric (waist circumference) components of the metabolic syndrome.

Results: TT was low in 19.9%, normal in 78.2%, and high in 1.8% of men. TE2 was low in 94.6% of women. Age and CVDRFs significantly influenced values of SHBG, FSH, TT, FAI, FEI, and DHEA-S in men, while in women values of FSH, TT and TE2 did not change. BMI and %BF affected SH regardless of the age groups and CVDRFs.

Conclusions: Our findings suggest that the reference ranges stratified by the five-year age bands seem more accurate than those given for the overall population over 60 years of age. The clinical relevance of these reference ranges increases when they are considered in relation to CVDRFs, BMI and %BF. (Endokrynol Pol 2013; 64 (2): 82–93)

Key words: testosterone, estradiol, indices of free sex hormones, SHBG, DHEA-S, elderly people, reference values 

Streszczenie

Wstęp: W ostatniej dekadzie wzrasta przewidywana długość życia społeczeństw. W Polsce brak danych o stężeniach hormonów płciowych (SH) u starzejących się osób. Celem naszej pracy było przedstawienie wartości referencyjnych SH w grupie osób uczestniczących w badaniu PolSenior w zależności od płci, wieku i czynników ryzyka chorób układu sercowo-naczyniowego (CVDRFs).

Materiał i metody: Badania wykonano u 4352 osób (2168 mężczyzn i 2088 kobiet) w wieku od 55. do powyżej 90. roku życia, podzielonych na 5-letnie przedziały wieku. Osoby na przedpolu starości (wiek 55–59 lat) stanowiły grupę referencyjną. Oznaczano testosteron całkowity (TT), estradiol całkowity (TE2) i DHEAS metodami RIA, FSH i SHBG metodą IRMA wyliczono wskaźniki wolnych androgenów (FAI) i estrogenów (FEI). Odsetek tłuszczu ciała (%BF) oznaczano metodą bioimpedancji. CVDRFs analizowano w odniesieniu do wartości ciśnienia tętniczego oraz biochemicznych (stężenie glukozy, HDL i triglicerydów) i antropometrycznych (obwód talii) składowych zespołu metabolicznego.

Wyniki: Prawie 20% mężczyzn miało stężenia TT poniżej normy, 78,2% w zakresie norm a 1,8% powyżej górnego zakresu normy. Stężenia TE2 poniżej normy dla kobiet po menopauzie miało 94,6% kobiet. Wiek i liczba CVDRFs istotnie zmieniały stężenia SH za wyjątkiem TE2u mężczyzn i FSH i TT u kobiet. BMI i %BF istotnie wpływało na większość oznaczanych hormonów niezależnie od wieku.

Wnioski: Przedstawiono wartości referencyjne dla stężeń SH u mężczyzn i kobiet od 65. do powyżej 90. roku życia. Nasze badania sugerują, że wartości referencyjne przedstawione w 5. letnich przedziałach wieku wydają się bardziej precyzyjne niż dla całej populacji powyżej 60. roku życia. Ich przydatność kliniczna zwiększa się, jeśli uwzględni się liczbę CVDRFs oraz BMI i %BF. (Endokrynol Pol 2013; 64 (2): 82–93)

Słowa kluczowe: testosteron, estradiol, wskaźniki wolnych hormonów, SHBG, DHEA-S, ludzie w podeszłym wieku, wartości referencyjne

Introduction

Over the last decade, average life expectancy has continuously increased. In Poland, compared to the middle of the last century, men live longer by approximately 16 years and women by almost 19 years, including individuals over 60 years old. In 2010, the predicted life expectancy at age 75 was 9.47 years for men and 11.85 years for women [1]. Therefore, there have been increasing numbers of multidisciplinary studies aiming to assess the age-dependent health risk factors that affect life expectancy in the elderly. The PolSenior Study is the first cross-sectional, national, and multicentre health survey designed to evaluate the medical, psychological, and socio-economic aspects of ageing in Poland [2].

The ageing process, which is still poorly understood, seems to be associated, among a variety of mechanisms, with a decline in the activity of the hypothalamic-pituitary-gonadal axis in postmenopausal women, and, to a lesser extent, in men over 65 years old. Many symptoms observed in elderly people such as the decrease in physical activity associated with decline in muscle mass and the weakening of muscle contraction, mood changes, impaired cognitive processes, impaired sexual performance, increased rate of obesity and bone mass deterioration that lead to disability, may be associated, at least partially, with impaired gonad function. The main sex hormones, including testosterone in men and estrogen in women, are transported in the blood to target cells by carrier proteins, mainly by sex hormone-binding globulin (SHBG) and albumin. The sex hormones’ biological activity is associated with albumin-bound and albumin-free fractions that decide on their bioavailability. In men over 60 years of age, the main sources of estrogens are the liver, skin, and adipose tissue, where biosynthesis of estrogen is based on aromatisation of their adrenal precursors [3]. In women, estrogens are synthesised also in ovarian theca cells [4].

The few studies performed on subjects aged 60–65 years have demonstrated that, in men, SHBG levels increase and free testosterone (FT) decreases with age. The reports on total testosterone (TT) in men over 65 have yielded conflicting results. Some studies have shown a progressive decrease in TT levels [5, 6], while others have not found significant age-dependent changes in TT [7, 8]. Similarly, there have been only a few studies that have provided inconsistent results regarding total estradiol (TE2) levels and free estradiol index (FEI) in ageing men [5].

In women over 60 years of age, as in men, studies have shown increases in SHBG levels with age [9, 10]. In women over 70 years of age, some studies have demonstrated decreases in TT, FT, and dehydroepiandrosterone sulfate (DHEA-S) levels [11]. On the other hand, van Geel et al. [12] found that, in women aged 55–84 years, SHBG, TT, FAI, and TE2 did not change, while FEI significantly decreased with age.

SHBG and sex hormone levels in the elderly are influenced by many factors. Aside from chronic kidney and liver diseases, obesity, thyroid diseases, hypopituitarism, dyslipidaemia, insulin resistance and type 2 diabetes, metabolic cardiovascular risk factors (CVDRFs), heart failure, and coronary heart disease may influence serum concentrations [13–15]. Increasing evidence suggests that the normal reference values for sex hormones in elderly subjects should be determined because they could help both in the assessment of hypogonadism rates [16] and in monitoring the safety and efficacy of hormone replacement therapy in elderly patients.

In this study, we assessed sex hormone levels in the PolSenior cohort, stratified by five-year age groups, to determine reference ranges in relation to gender, age, and cardiovascular disease risk factors (CVDRFs).

Material and methods

Study population

This study was performed with 5,695 subjects (2,899 men and 2,796 women) participating in the PolSenior Study, who were randomly selected from 16 administrative centres of Poland by the three-stage, proportional, and stratified-by-age group selection process, as described elsewhere [2]. In brief, stratified random sampling was used with the aim of recruiting elderly men and women in six five-year age groups: 65–69, 70–74, 75–79, 80–84, 85–89, and 90 years and over. Additionally, from the same cohort, pre-elderly subjects aged 55–59 years served as the reference group. Subjects were invited by letter to attend a session for an interviewer-administered questionnaire; assessment of height, weight, and waist and hip circumference, and blood sampling. The structured questionnaire (available at http://polsenior.iimcb.gov.pl/ankiety/) included details about medical history and socio-economic status of the subjects. Blood sampling, blood pressure and measurements of height, weight, waist circumference, and hip circumference were performed according to the standard protocol, as described elsewhere [17]. We included only those responders who were able to complete study procedures, and excluded immobilised patients.

The study complied fully with all applicable institutional and governmental regulations concerning the ethical use of human volunteers and with the terms of the Helsinki Declaration. The institutional review board approved the study protocol, and all the recruited subjects gave their written informed consent.

Anthropometric measurements

Height, waist circumference (WC) and hip circumference were measured to the nearest 0.5 cm. Percentage body fat (%BF) was measured by bioelectric impedance (Tanita BC-536, Tanita Corporation, Japan). %BF was calculated from the measurements of resistance made at 50 kHz using the formula provided by the manufacturer.

Assays

We collected blood samples from 2,168 men and 2,088 women. Biochemical assessments included SHBG, FSH, TT, TE2, DHEA-S, lipid profiles, fasting glucose and insulin were measured in all subjects in one batch of serum or plasma that had been thawed for the first time. Blood samples were divided in aliquots and frozen at –20°C until further analyses. Sex hormones, DHEA-S and insulin were assessed by radioimmunological methods using Wallac 1470 Wizard gamma counter. SHBG and FSH were assessed by IRMA using commercially available assays (Immunotech, Prague). TT, TE2, DHEA-S, and insulin were assessed using the Siemens assays (Los Angeles, USA). The free androgen index (FAI) and FEI were calculated from SHBG, TT, and TE2 using the following formulas: FAI = TT(nmol/L) × 100/SHBG(nmol/L); FEI = 100 × E2 (pmol/L)/SHBG (nmol/L) [18, 19]. Blood glucose and lipid profiles, including total cholesterol, high-density lipoproteins (HDL), low-density lipoproteins (LDL) and triglycerides (TG) were assessed by automatic methods (Modular PPE, Roche Diagnostics). Insulin resistance was calculated from fasting glucose and insulin using the following formula: HOMA-IR (Homeostasis Model Assessment– Insulin Resistance) = glucose (mmol/L) × insulin (mIU/L)/22.5 [20].

The CVDRFs assessment included waist circumference, blood pressure, blood glucose, HDL, and TG, according to current guidelines and diagnostic criteria for metabolic syndrome [21].

Statistical analysis

The distribution of continuous variables was tested for normality by the Kolmogorov-Smirnov test. Because the majority of the studied variables showed markedly skewed distribution, logarithmic transformations of these measurements were performed before calculations, and the results were presented as median (range), median (25–75 quartiles) or mean (95% CI). Differences among the groups were evaluated by Mann-Whitney U test or Wilcoxon signed rank test as appropriate. Intra-gender differences (between age bands, separately for men and women) were evaluated using ANOVA/MANOVA followed by Tukey’s post-hoc test. Linear Spearman’s rank correlation coefficients and multiple regression analyses were used to determine relationships between studied variables. We used the 5% significance levels for all tests. All calculations were performed with the Statistica 9.0 software package (StatSoft, Tulsa, OK, USA).

Results

Baseline characteristics are shown in Table I. As expected, men had higher TE2, FEI and DHEA-S levels than women. They had also more favourable lipid profile. SHBG concentrations were comparable in both sexes. Medical history revealed that overall prevalence of known diabetes was 14.5% (308 subjects) in men and 18% (396 subjects) in women. However, in the studied cohort, as many as 33.8% of individuals (1,202 men and women) had increased fasting glucose concentration, of whom nearly two thirds had glucose levels above 7.0 mmol/L (126 mg/dL). TT was low in 19.9%, normal in 78.2% and high in 1.8% of men compared to the normal reference range for men aged 60 years. Moreover, almost all men (99.6%) had TE2 levels below the normal reference range. On the other hand, only 3.6% of females had TE2 levels above reference values for postmenopausal women. Compared to the normal reference ranges, TT was lowered in 46.5%, normal in 50.1%, and increased in 3.4% of women. Sex hormones, SHBG, and DHEA-S significantly correlated with the majority of anthropometric and biochemical CVDRFs; the results of these analyses have been published elsewhere [22].

Table I. Baseline characteristics. Values are expressed as median and Q25–Q75

Tabela I. Charakterystyka badanych. Wartości przedstawiono jako mediana i Q25–Q75

	Men (n = 1,649)	Women (n = 1,547)	Z test
Age (years)	76.0 [69.0–85.0]	74.0 [67.0–83.0]	ns
Height [cm]	170.0 [164.0–173.6]	156.0 [152.0–161.0]	p < 0.000
Body mass [kg]	77.7 [69.0–87.2]	69.5 [60.4–80.0]	p < 0.000
Waist circumference [cm]	100.0 [93.0-109.0]	96.0 [87 0–105.0]	p < 0.000
Hip circumference [cm]	104.0 [100.0–110.0]	108.0 [101.0–116.0]	p < 0.02
BMI [kg/m2]	27.2 [24.6–30.1]	28.3 [ 25.1–32.4]	p < 0.000
%BF	27.7 [22.6–32.9]	37.4 [31.8–41.9]	p < 0.000
SHBG [nmol/L]	51.7 [38.6–70.2]	58.2 [41.8–80.8]	ns
FSH [IU/L]	10.1 [6.4–19.5]	59.7 [44.0–75.0]	p < 0.000
TT [ nmol/L]	15.3 [11.3–19.4]	0.8 [0.4–1.1]	p < 0.000
FAI	29.8 [21.7–39.6]	1.3 [0.7–2.2]	p < 0.000
TE2 [pmol/L]	48.4 [36.0–63.4]	23.8 [18.3–31.2]	p < 0.000
FEI	91.6 [59.3–141.2]	41.3 [26.1-66.2]	p < 0.000
DHEA-S [mg/dL]	766.0 [463.0–1,169.0]	543.0 [335.0–886.0]	p < 0.001
Total cholesterol (TC) [mmol/L]	5.1 [4.3–5.8]	5.5 [4.7–6.3]	p < 0.004
LDL [mmol/L]	2.9 [2.3–3,7]	3.2 [2.5–4.0]	ns
HDL [mmol/L]	1.2 [0.9–1,5]	1.4 [1.1–1.6]	p < 0.000
Triglycerides (TG) [mmol/L]	1.2 [0.9–1.6]	1.4 [1.1–1.82]	p < 0.000
Glucose [mmol/L]	5.3 [4.8–6.1]	5.2 [4.8–5.9]	ns
Insulin [mIU/L]	5.1 [2.8–9.4]	6.3 [3.6–10.6]	ns
HOMA IR	1.2 [0.6–2.4]	1.5 [0.8–2.7]	ns
Diabetes (%)	14.5	18.0
Hypertension (%)	57.0	41.3
Z — Wald-Wolfowitz test

The results of multiple regression analyses that assessed the associations of sex hormones and DHEA-S with selected anthropometric (Model 1) and biochemical (Model 2) parameters as independent variables are shown in Table II. Sex hormones (except for FSH in women) and DHEA-S were strongly associated with age.

Table II. Most relevant associations between SHBG, DHEA-S, FSH, TT, TE2, FAI, FEI, age and anthropometric (β1) and metabolic (β2) measurements

Tabela II. Najważniejsze zależności między SHBG, DHEA-S, TT, FSH, TT, TE2, FAI, wiekiem i pomiarami antropometrycznymi (β1) oraz metabolicznymi (β2)

Men

Women

SHBG [nmol/L]

FSH [IU/L]

TT [nmol/L]

FAI

E2 [pmol/L]

FEI

DHEA-S [mg/dL]

SHBG [nmol/L]

FSH [IU/L]

TT [nmol/L]

FAI

E2 [pmol/L]

FEI

DHEA-S [mg/dL]

β1

β1

β1

β1

β1

β1

β1

β1

β1

β1

β1

β1

β1

Age (years)

0.25c

0.37c

–0.23c

0.36c

–0.12c

–0.27c

–0.49c

0.23c

ns

–0.10c

–0.07b

–0.21c

–0.27b

BF%

–0.08b

ns

–0.08c

ns

ns

ns

ns

–0.06a

ns

0.08a

ns

ns

ns

WC [cm]

–0.13b

–0.08b

–0.10b

ns

ns

ns

ns

–0.17c

–0.14c

ns

ns

0.09a

ns

BMI [kg/m2]

–0.18c

ns

ns

0.08a

ns

0.17c

ns

–0.24c

–0.22c

0.21c

0.26c

0.33c

ns

β2

β2

β2

β2

β2

β2

β2

β2

β2

β2

β2

β2

β2

TC [mmol/L]

ns

0.14b

ns

ns

0.06a

ns

0.24c

0.21c

ns

ns

ns

–0.18b

ns

LDL [mmol/L]

ns

ns

ns

ns

ns

ns

ns

–0.15b

ns

ns

ns

0.16b

ns

HDL [mmol/L]

0.07b

ns

ns

ns

ns

–0.07 a

0.08b

ns

0.18c

ns

ns

ns

ns

TG [mmol/L]

–0.22c

–0.07a

–0.14c

ns

ns

0.15c

0.10c

–0.25c

ns

0.09b

ns

0.17c

ns

HOMA–IR

–0.07c

ns

ns

ns

ns

ns

ns

–0.25c

–0.12c

0.22c

0.17c

0.28c

0.12c

ap < 0.05; bp < 01; cp < 0.001; Model 1 (β1) includes age, BF%, waist circumference, and BMI; Model 2 (β2) includes TC, LDL-C, HDL-C, TG, and HOMA IR; TC — total cholesterol

Changes in SHBG, TT, FAI, FEI, FSH, and DHEA-S levels within age groups analysed by one-way ANOVA are shown in Table III and Figure 1. In men, compared to the pre-elderly and across some of the elderly age groups, mean values of FSH, TT, FAI, FEI, and DHEA-S decreased and SHBG increased. In women, as in men, SHBG increased and DHEA-S and free sex hormone indices decreased, while FSH and TT did not change with ageing.

Table IIIA. Mean and 95% confidence intervals (antilogarithms) values of SHBG, FSH, TT, FAI, TE2, FEI and DHEA-S levels in five-year age bands in men

Tabela IIIA. Średnie i 95% przedziały ufności (antylogarytmy) stężenia SHBG, FSH, TT, FAI, TE2, FEI i DHEA-S w 5-letnich przedziałach wiekowych u mężczyzn

Age group (years)	n	SHBG [nmol/L]	FSH [UI/L]	TT [nmol/L]	FAI	E2 [pmol/L]	FEI	DHEA-S [ng/mL]
55–59	233	41.6	6.9	16.1	38.6	50.8	122.0	1,397.8
		39.3–44.1	6.3–1.1	14.6–17.1	35.0–42.6	47.8–53.9	112.5–132.4	1,288.2–1,516.8
65–69	273	45.3	8.5	15.7	34.6	49.0	108.0	973.1***
		42.0–47.9	7.7–9.3	14.3–17.2	31.5–38.1	446.2–51.9	99.8–116.8	899.1–1,053.2
70–74	317	50.3***	10.1	14.2	28.2**	46.6	92.6***	802.8***
		47.7–53.0	9.2–11.1	13.0–15.5	25.8–30.9	44.1–49.2	85.9–99.8	744.4–865.8
75–79	312	53.2***	12.1***	15.0	28.2**	49.3	92.6***	737.9***
		50.6–56.0	11.0–13.2	13.8–16.4	25.8–30.8	46.7–52.0	86.1–99.5	685.9–793.8
80–84	286	57.0***	13.1***	12.5***	22.0***	45.3	79.4***	590.1***
		53.9–60.3	11.9–14.5	11.4–13.8	19.9–24.2	42.7–48.1	73.3–86.1	544.1–640.0
85–89	315	61.7***	17.1***	10.8***	17.6***	44.4**	72.0***	510.6***
		65.0–65.0	15.6–18.8	9.9–11.9	16.0–19.2	42.0–47.0	66.8–77.6	473.5–550.6
> 90	219	68.5***	22.8***	9.5***	13.8***	41.4**	60.5	408.1
		64.5–72.6	20.6–25.3	8.5–10.5	12.5–15.3	38.9–44.1	55.6–65.8	374.8–444.3
p (ANOVA)		p < 0.000	p < 0.000	p < 0.000	p < 0.000	p < 0.000	p < 0.000	p < 0.000
Differences in serum SHBG, FSH, TT, FAI, TE2, FEI, and DHEA-S levels between age groups were calculated using Tukey’s post hoc test ; **p > 0.00 v. group of 55–59 year; ***p > 0.000 v. group of 55–59 year

 

Figure 1A. Distribution of values of SHBG, FSH, TT, FAI, FEI and DHEA-S concentrations (mean and 95% CI; logarithms) by age groups in men. TE2 decreased significantly only in subjects aged > 85 (not shown); B. Distribution of values of SHBG, FSH, FAI, TE2, FEI and DHEAS concentrations (mean and 95% CI; logarithms) by age groups in women. FSH and TT (not shown) did not change. TE2 decreased significantly only in subjects aged > 90

Rycina 1A. Rozkład stężeń SHBG, FSH, TT, FAI, FEI i DHEA-s ( średnia i 95% CI; logarytmy) w grupach wieku u mężczyzn. Stężenia TE2 zmniejszały się tylko u osób > 85. roku życia (nie przedstawione); B. Rozkład stężeń SHBG, FSH, FAI, TE2 , FEI i DHEA-S (średnia i 95% CI; logarytmy) w grupach wieku u kobiet. Stężenia FSH i TT (nie przedstawione) nie zmieniały się. Stężenia TE2 zmniejszały się istotnie dopiero > 90. roku życia

Among all the factors that influenced sex hormone levels, the majority were CVDRFs. Overall, more than three CVDRFs were present in as many as 41.4% of men and 58.6% of women. Although the majority of CVDRFs — adjusted for gender, age, and BMI — were strong single predictors of sex hormone levels in multiple regression analysis, further analyses using MANOVA revealed that, in women, only age (stratified by age groups) and number of CVDRFs (< 3 or ≥ 3) were significant determinants of sex hormone levels (Table IV, Figure 2). In contrast, in men, the number of CVDRFs and age were associated with sex hormone levels only in a MANOVA model that also included %BF (p = 0.04). In women, %BF (< 30% or ≥ 30%) was a third factor that significantly modified sex hormone levels.

Table IIIB. Mean and 95% confidence intervals (antilogarithms) values of SHBG, FSH, TT, FAI, TE2, FEI, and DHEA-S in five-year age bands in women

Tabela IIIB. Średnie i 95% przedziały ufności (antylogarytmy) stężenia SHBG, FSH, TT, FAI, TE2, FEI i DHEA-S w 5-letnich przedziałach wiekowych u kobiet

Age group (years)	n	SHBG [nmol/L]	FSH [IU/L]	TT [nmol/L]	FAI	E2 [pmol/L]	FEI	DHEA-S [ng/mL]
55–59	288	49.7	56.3	0.72	1.45	26.5	53.2	779.6
		47.1–52.4	52.8–60.1	0.66–0.79	1.31–1.61	25.1–27.9	49.1–57.6	719.4–844.8
65–69	285	51.9	55.5	0.71	1.37	24.6	47.4	647.2***
		49.0–55.0	51.7–59.5	0.65–0.78	1.23–1.53	23.2–26.0	43.4–51.6	593.4–706.0
70–74	319	55.8	56.3	0.75	1.34	23.8	42.7	534.0***
		52.7–59.1	52.5–60.3	0.68–0.82	1.21–1.50	22.5–25.2	39.2–46.6	489.9–582.2
75–79	262	59.3***	49.4	0.72	1.21	24.4	41.2***	451.3***
		55.8–63.1	45.8–53.2	0.65–0.80	1.08–1.36	23.0–26.0	37.6–45.1	411.4–494.9
80–84	223	65.6***	53.7	0.70	1.06*	23.1	35.2***	395.9***
		61.6–70.0	49.7–58.0	0.63–0.77	0.94–1.19	21.7–24.6	32.0–38.7	359.7–435.8
85–89	237	71.8***	53.9	0.71	0.99***	23.3	32.4***	444.0***
		67.6–76.3	50.0–58.0	0.64–0.78	0.88–1.11	21.9–24.8	29.6–35.5	405.3–486.4
> 90	201	80.0***	59.4	0.74	0.93***	22.2*	27.8***	378.4***
		75.0–85.3	55.0–64.2	0.67–0.83	0.82–1.05	20.8–23.7	25.2–30.5	343.7–416.6
p (ANOVA)		p = 0.000	p = 0.0799	p = 0.8903	p = 0.002	p = 0.000	p = 0.000	p = 0.000
Differences in serum SHBG, FSH, TT, FAI, TE2, FEI, and DHEA-S levels between age groups were calculated using Tukey’s post hoc test ; *p < 0.05 v. group of 55–59 yr; ***p < 0.000 v. group of 55–59 yr

Table IVA. Mean and 95% confidence intervals (antilogarithms) values of SHBG, FSH, TT, FAI, TE2, FEI and DHEA-S in five-year age bands in men with CVDRF < 3 and CVDRF ≥ 3. Differences between age groups were calculated using Tukey’s post hoc test. MANOVA; the Wilks test = 0.97; p = 0.04

Tabela IVA. Średnie i 95% przedziały ufności (antylogarytmy) stężenia SHBG, FSH, TT, FAI, TE2, FEI i DHEA-S w 5-letnich przedziałach wiekowych u mężczyzn z CVDRF < 3 i CVDRF ≥ 3. Różnice między przedziałami wieku oceniano testem post hoc Tukey’a. MANOVA; test Wilksa = 0.97; p = 0.04

	Age group (years)	n	SHBG [nmol/L]	FSH [IU/L]	TT [mmol/L]	FAI	TE2 [pmol/L]	FEI	DHEAS[ng/mL]
CVDRF < 3
	55–59	122	52.7	7.3	19.7	37.4	52.9	100.3	1,401.8
			48.4–57.5	6.2–8.6	16.9–23.0	32.0–43.8	48.1–58.3	88.3–114.0	1,231.2–1,596.1
	65–69	113	53.6	8.6	16.7	31.1	46.6	87.0	969.6
			48.6–59.1	7.2–10.3	14.0–19.9	26.0–37.2	41.8–52.0	75.2–100.5	836.5–1,123.7
	70–74	144	63.2	9.6	17.6	27.9	44.7	70.8	767.6
			58.0–68.8	8.3–11.3	15.2–20.5	23.9–32.6	40.7–49.2	62.4–80.3	675.1–872.8
	75–79	146	59.8	11.0	18.2	30.3	55.8	93.2	774.2
			54.1–66.2	9.2–13.3	15.2–21.7	25.3–36.4	49.9–62.3	80.4–108.0	666.3–899.6
	80–84	134	62.2	12.6	12.9	20.7	44.0	70.7	582.6
			57.4–67.4	10.9–14.7	11.1–14.8	17.8–23.9	40.2–48.1	62.7–79.6	516.1–657.6
	85–89	173	68.2	15.7	13.1	19.1	44.7	65.4	500.4
			63.5–79.6	13.8–17.9	11.5–14.8	16.8–21.8	41.2–48.4	58.8–72.8	449.1–557.6
	> 90	122	70.9	23.0	10.2	14.3	45.0	63.5	392.5
			65.3–76.9	19.8–26.7	10.2–11.7	12.4–16.6	41.1–49.3	56.3–71.6	347.4–443.5
CVDR ≥ 3
	55–59	92	31.1	5.8	12.2	39.2	56.4	181.8	1,229.7
			25.8–37.5	4.1–8.2	8.8–17.0	27.9–55.1	45.9–69.5	138.0–239.5	928.8–1,628.0
	65–69	126	48.4	8.0	16.3	33.7	55.6	114.8	1,286.0
			39.4–59.5	5.5–11.6	11.3–23.5	23.2–49.1	44.2–69.9	84.7–155.7	943.2–1,753.5
	70–74	145	47.0	9.2	14.4	30.6	48.6	103.5	671.4
			41.4–53.4	7.3–11.6	11.5–18.0	24.3–38.6	42.2–56.1	85.7–124.9	554.5–813.0
	75–79	130	49.1	11.6	15.6	31.7	50.9	103.7	666.0
			41.4–58.2	8.5–15.8	11.5–21.0	23.3–43.2	42.2–61.5	80.7–133.2	516.0–859.4
	80–84	111	54.6	12.8	12.6	23.1	50.7	92.8	447.2
			46.0–64.8	9.4–17.5	9.3–17.1	17.0–31.5	41.9–61.3	72.1–119.3	346.2–577.8
	85–89	97	50.2	14.5	8.6	17.1	45.5	90.7	497.3
			42.3–59.6	10.6–19.8	6.4–11.7	12.6–23.4	37.6–55.1	70.4–116.7	384.5–643.1
	> 90	43	61.9	13.2	5.1	8.3	33.2	53.7	453.0
			48.5–79.0	8.5–20.7	3.3–7.9	8.3–12.9	25.4–43.6	37.5–76.8	314.4–652.7

Table IVB. Mean and 95% confidence intervals (antilogarithms) values of SHBG, FSH, TT, FAI, TE2, FEI, and DHEA-S in five–year age groups in women with CVDRF < 3 and CVDRF ≥ 3. Differences between age groups were calculated using Tukey’s post hoc test. MANOVA; the Wilks test = 0.97; p = 0.004

Tabela IVB. Średnie i 95% przedziały ufności (antylogarytmy) stężenia SHBG, FSH, TT, FAI, TE2, FEI i DHEA-S w 5–letnich przedziałach wiekowych u kobiet z CVDRF < 3 i CVDRF ≥ 3. Różnice między przedziałami wieku oceniani testem post hoc Tukey’a. MANOVA: test Wilksa = 0.97; p = 0,004

	Age group ) [years]	n	log SHBG [nmol/L]	log FSH [IU/L]	log TT [nmol/L]	log FAI [nmol/L]	log E2 [pmol/L]	log FEI	log DHEAS [ng/mL]
CVDRFs < 3
	55–69	148	56.1	58.7	0.67	1.19	25.9	46.1	770.7
			52.1–60.5	53.6–64.2	0.6–0.8	1.03–1.36	24.0–27.9	41.2–51.5	689.0–862.1
	65–69	90	56.7	58.7	0.76	1.34	24.1	42.5	686.1
			51.4–62.4	52.3–65.9	0.6–0.9	1.12–1.60	21.9–26.5	36.8–49.0	594.3–792.1
	70–74	85	63.3	59.0	0.77	1.22	22.5	35.6	539.7
			57.4–69.8	52.4–66.5	0.7–0.9	1.02–1.47	20.4–24.8	30.7–41.2	465.5–625.7
	75–79	77	65.5	47.0	0.70	1.08	24.9	38.1	404.2
			59.1–72.6	41.5–53.3	0.6–0.8	0.89–1.31	22.5–27.6	32.6–44.4	346.0–472.1
	80–84	77	68.1	51.3	0.64	0.93	22.6	33.2	353.0
			61.5–75.5	45.3–58.2	0.5–0.8	0.77–1.13	20.4–25.1	28.4–38.7	302.3–412.4
	85–89	90	74.0	56.1	0.68	0.91	23.3	31.4	410.4
			67.3–81.4	50.0–63.0	0.6–0.8	0.76–1.09	21.1–25.6	27.3–36.2	355.5–473.8
	> 90	97	83.7	63.0	0.70	0.83	19.9	23.8	336.3
			76.4–91.8	56.3–70.4	0.6–0.8	0.70–0.99	18.2–21.8	20.7–27.3	292.9–386.3
CVDRFs ≥ 3
	55–59	140	44.0	54.1	0.78	1.78	27.1	61.5	788.6
			40.8–47.5	49.3–59.4	0.7–0.9	1.54–2.05	25.1–29.2	54.8–68.9	702.8–884.0
	65–69	195	47.5	52.5	0.67	1.40	25.1	52.8	610.5
			44.5–50.7	48.5–56.8	0.6–0.7	1.24–1.59	23.5–26.8	47.9–58.2	553.8–673.1
	70–74	234	49.1	53.6	0.72	1.47	25.2	51.4	528.5
			46.3–52.1	49.9–57.6	0.7–0.8	1.32–1.65	23.8–26.8	47.0–56.1	483.4–577.7
	75–79	185	53.7	51.8	0.73	1.37	23.9	44.5	503.8
			50.3–57.4	47.8–56.2	0.7–0.8	1.21–1.55	22.4–25.5	40.3–49.1	455.7–556.9
	80–84	146	63.2	56.2	0.76	1.20	23.7	37.4	444.0
			58.7–68.2	51.3–61.5	0.7–0.9	1.04–1.38	22.0–25.5	33.5–41.8	296.7–497.1
	85–89	147	69.7	51.7	0.74	1.07	23.3	33.5	480.3
			64.6–75.1	47.2–56.6	0.7–0.8	0.93–1.23	21.7–25.1	30.4–37.5	429.2–537.4
	> 90	104	76.4	56.1	0.79	1.04	24.8	32.4	425.7
			70.0–83.5	50.4–62.5	0.7–0.9	0.88–1.23	22.7–27.1	28.4–37.0	372.4–486.5

 

Figure 2A (men), B (women) — Distribution of sex hormones SHBG and DHEA-S concentrations (mean and 95% CI; logarithms) by age groups in persons with CVDRFs < 3 and CVDRFs ≥ 3

Rycina 2A (mężczyźni), B (kobiety) — Rozkład stężeń hormonów płciowych, SHBG i DHEA-S w grupach wieku (średnia i 95% CI; logarytmy) u osób mających < 3 CHDRFs oraz ≥ 3 CHDRFs

Median values of SHBG, TT, FAI, TE2, FEI and DHEA-S in subjects with more and less than three CVDRFs are given in Table V. In both genders, all studied parameters (except for estradiol in men) significantly differed between subjects with ≥ 3 and < 3 CVDRFs. Sex hormones and SHBG concentrations significantly changed with age in men; however, changes in SHBG and FEI were more pronounced in individuals with more than three CVDRFs. FSH and TT in women did not change with age, and CVDRFs and individual significant changes in other parameters were not observed in all age groups. Overall, we found significant differences in sex hormone levels (except FSH and TT) between women with less and more than three CVDRFs, but these differences were less pronounced than in men (Fig. 2).

Table V. Median values (range) of SHBG, TT, FAI, TE2, FEI, FSH, and DHEA-S concentrations in men and women with CVDRF < 3 and CVDRF ≥ 3

Tabela V. Mediany (rozstęp) stężeń SHBG, TT, FAI, TE2 ,FEI, FSH i DHEA-S u mężczyzn i kobiet z liczbą czynników ryzyka sercowo-naczyniowego poniżej 3 i powyżej 3 (CVDRF < 3 i CVDRF ≥ 3)

	Men		p	Women		p
	CVDRFs < 3	CVDRFs ≥ 3		CVDRFs < 3	CVDRFs ≥ 3
SHBG [nmol/L]	60.1 (321.7)	45.4 (215.9)	p = 0.000	65.5 (226.2)	54.9 (228.8)	p = 0.000
FSH [IU/L]	10.7 (166.6)	9.7 (106.9)	p = 0.002	63.2 (210.7)	57.6 (275.7)	p = 0.000
TT [ nmol/L]	16.2 (62.9)	14.0 (36.9)	p = 0.000	0.7 (27.2)	0.8 (65.8)	p = 0.04
FAI	28.3 (124.9)	31.4 (119.3)	p = 0.000	1.1 (38.8)	1.3 (89.4)	p = 0.000
E2 [pmol/L]	47.7 (231.2)	48.8 (261.0)	ns	22.7 (367.0)	24.2 (472.3)	p = 0.002
FEI	79.1 (595.1)	105.9 (715.6)	p = 0.000	34.2 (398.2)	43.9 (624.2)	p = 0.000
DHEA–S [mg/dL]	723.0 (6,473.0)	772.0 (4,942.0)	p = 0.007	503.0 (4,100.0)	554.0 (4,786.0)	p = 0.02

In univariate analysis, the associations of BMI and %BF with SHBG, FSH, FEI, and DHEA-S, although statistically significant, explained only a small percentage of the variation in sex hormones, except for SHBG level. In women, as in men, BMI markedly influenced the variation in SHBG. However, the impact of BMI on variations in other sex hormones (including FAI) was relatively low (Table VI).

Table VI. Associations between BMI, %BF, SHBG, TT, FAI, TE2, FEI, FSH, and DHEA-S in men and between BMI, SHBG, TT, FAI, TE2, FEI, FSH, and DHEAS in women

Tabela VI. Zależności między BMI, BF%, SHBG, TT, FAI, TE2, FEI, FSH i DHEA-S u mężczyzn oraz BMI, SHBG, TT, FAI, TE2, FEI, FSH i DHEA-S u kobiet

log BMI
Men	β	SE	p	R2	The regression equation
SHBG [nmol/L]	–1.00	0.06	0.000	0.12	log SHBG = 3.15–1.00 * log BMI
FSH [IU/mL]	–0.44	–0.44	0.000	0.01	log FSH = 1.65–0.44 * log BMI
TT [nmol/L]	–0.51	0.11	0.000	0.01	log TT = 1.84–0.51* log BMI
FAI	0.51	0.11	0.000	0.01	log FAI = 0.67 + 0.51* log BMI
TE2 [pmol/L]	0.09	0.07	ns
FEI	1.11	0.09	0.000	0.07	log FEI = 0.365 + 1.10* log BMI
DHEAS [ng/mL]	0.39	0.10	0.000	0.01	log DHEAS = 2.28 + 0.39 * log BMI
log %BF
Men	β	SE	p	R2	The regression equation
SHBG [nmol/L]	–0.40	0.04	0.000	0.07	log SHBG = 2.29–0.40 * log %BF
FSH	–0.17	0.06	0.008	0.00	log FSH = 1.30–0.17 * log %BF
TT [nmol/L]	–0.31	0.06	0.000	0.01	log TT = 1.57–0.31* log %BF
FAI	0.09	0.06	ns
TE2 [pmol/L]	–0.03	0.04	ns
FEI	0.38	0.05	0.000	0.03	log FEI = 1.40 + 0.38 * log %BF
DHEAS [ng/mL]	0.14	0.06	0.011	0.00	log DHEAS = 2.66 + 0.14 * %BF
log BMI
Women	β	SE	p	R2	The regression equation
SHBG [nmol/L]	–0.97	0.06	0.000	0.13	log SHBG = 3.19–0.97 * log BMI
FSH [IU/L]	–0.58	0.07	0.000	0.04	log FSH = 2.57–0.58 * log BMI
TT [nmol/L]	0.28	0.09	0.002	0.00	log TT = –0.51 + 0.28 * log BMI
FAI	1.25	0.10	0.000	0.07	log FAI = –1.701 + 1.2295 * log BMI
TE2 [pmol/L]	0.50	0.06	0.000	0.04	log E2 = 0.61 + 0.50 * log BMI
FEI	1.48	0.08	0.000	0.14	log FEI = –0.57 + 1.48 * log BMI
DHEAS [ng/mL]	0.27	0.09	0.004	0.00	log DHEAS = 2.37 + 0.27 * log BMI

Discussion

In this study, we determined, for the first time, reference ranges for SHBG, TT, TE2, FAI, FEI, FSH, and DHEA-S in Polish men and women over 65 years of age, stratified by five-year age groups and including subjects over 90 years of age. In the literature, there have been only a few reports on the reference values of sex hormones in people over 65 years of age [5, 11, 23–25]. In our study, sex hormone levels showed high individual variations, which has also been reported in earlier studies [5, 23]. These variations may result, at least partially, from the relatively less strict inclusion criteria applied in our study. However, they may also result from well-known environmental and metabolic factors as well as less-known genetic factors, which account for approximately 30-60% of the variation in SHBG [26] — a globulin that determines both binding and bioavailability of sex hormones. We have demonstrated that, compared to the pre-elderly and across the elderly age groups, mean values of TT, DHEA-S, FAI, and FEI decreased, SHBG increased, and TE2 decreased only in men over 85 years of age. In women, as in men, SHBG increased, free sex hormones indices and DHEA-S decreased, but TT, FSH and TE2 (except for women 90 years of age) did not change with age. Overall, these findings are consistent with the results of earlier studies [5, 7], although, unlike our study, they could not find any associations between age and TT in men. We also confirmed previous reports that have found increases in SHBG levels with age in women [9, 10]. On the other hand, in line with some studies [11, 12], but in contrast to other studies [24, 23], we could not demonstrate the impact of ageing on TT levels in women. As in men, we observed age-dependent decreases in DHEA-S, FAI, and FEI in women. These discrepancies may result from differences in methods of hormone assessment and their free fraction calculation, ethnicity, age of studied populations, and sample size.

The cross-sectional nature of our study does not allow exploration of a cause-effect relationship; thus, our results do not explain age-dependent increases in SHBG levels in the ageing population. However, our results seem to confirm that some of these changes might be associated with changes in insulin sensitivity and BMI, which progressively decrease with age (data not shown). Free sex hormones indices seem to reflect changes in SHBG levels. Moreover, the progressive FSH increase and TT decrease observed in our study may suggest that the age-related decline of gonadal function in elderly men is rather a slow process, while in women, the pituitary-gonadal axis seems to remain relatively stable throughout the entire postmenopausal period up to 85 years of age.

Our results suggest that both in men and women, CVDRFs may influence sex hormone levels. Men with more than three CVDRFs had decreased values of SHBG and TT, higher FAI and FEI, but similar TE2 compared to those with fewer CVDRFs. These findings are consistent with the majority of previous studies [8, 9, 27] and Maggio et al. [28], who found lowered TE2 concentration in older men in the InCHIANTI Study. In our study, women with three or more CVDRFs, like men, had lowered SHBG levels but higher TT, TE2, and FAI, which has also been reported by previous studies [9, 29] and confirmed in a recent meta-analysis of 52 observational studies [27]. In addition to these observations, we also found in this population higher FEI and DHEA-S concentrations.

Based on our results and the results of previously cited reports, we present, for the first time in the literature, reference ranges for SHBG, TT, TE2, DHEA-S, FSH, FAI, and FEI in elderly Polish men and women, stratified by five-year age groups and number of CVDRFs. Our results suggest that the reference ranges for sex hormones are not only gender- and age-specific, but also should be evaluated in relation to the number of CVDRFs. In particular, this might facilitate decisions regarding the application of hormone replacement therapy in elderly cases with hypogonadism.

The strength of this study is that we analysed a large, representative, randomly selected sample of elderly Polish people recruited from different administrative centres of Poland using the three-stage, proportional and stratified-by-age group selection process. However, some potential limitations should be considered in interpreting our results. First, the cross-sectional nature of the study is problematic for differentiating cause and effect from simple association. Second, we did not use strict exclusion criteria, which might have potentially influenced the variability of hormonal assessments. Like Orwoll et al. [5] and Yeap et al. [7], we did not exclude cases with benign prostatic hyperplasia (2.5% of men in the PolSenior cohort), prostate cancer (0.85%) that might be treated with antiandrogens [30], or hypothyroidism (10.3% of women and 5.6% of men, including 60% with asymptomatic disease) [31]. We also did not exclude patients with diabetes because in the multiple regression analysis, history of diabetes, along with BMI and age bands, did not significantly influence sex hormone levels (data not shown). Third, in this study we evaluated TT with the RIA method. It has been suggested that liquid chromatography-mass spectrometry has the highest sensitivity and selectivity in the determination of total and free testosterone. However, recent studies have not shown significant advantages for either method [32].

In conclusion, in this study, we have presented normal reference values for sex hormones in Polish elderly men and women. Our data suggests that the reference ranges stratified by five-year age bands seem more accurate than those given for the overall population over 60 years of age. The clinical relevance of these reference ranges increases when they are considered in relation to CVDRFs, BMI, and %BF.

References

1. 1. Polish Central Statistical Office 2010: http://www.stat.gov.pl/cps/rde/xbcr/gus/lud_trwanie_zycia_2010.pdf.
2. 2. Blędowski P, Mossakowska M, Chudek J et al. Medical, psychological and socioeconomic aspects of aging in Poland. Assumptions and objectives of the PolSenior project. Exp Gerontol 2011; 46: 1003–1009.
3. 3. Czajka-Oraniec I, Simpson ER. Aromatase research and its clinical significance. Endokrynol Pol 2010; 61: 126–134.
4. 4. Fogle RH, Stanczyk FZ, Zhang × et al. Ovarian androgen production in postmenopausal women. J Clin Endocrinol Metab 2007; 92: 3040–343.
5. 5. Orwoll E, Lambert LC, Marshall LM et al. Testosterone and estradiol in older men. J Clin Endocrinol Metab 2006; 91: 1336–1344.
6. 6. Li JY, Li XY, Li M et al. Decline of serum levels of free testosterone in aging healthy Chinese men. Aging Male 2005; 8: 203–206.
7. 7. Yeap BB, Almeida OP, Hyde Z et al. In men older than 70 years, total testosterone remains stable while free testosterone declines with age. The Health In Men Study. Eur J Endocrinol 2007; 156: 585–594.
8. 8. Chubb SA, Hyde Z, Almeida OP et al. Lower sex hormone-binding globulin is more strongly associated with metabolic syndrome than lower total testosterone in older men: the Health in Men Study. Eur J Endocrinol 2008; 158: 785–792.
9. 9. Maggio M, Lauretani F, Basaria S et al. Sex hormone-binding globulin levels across the adult lifespan in women — the role of body mass index and fasting insulin. J Endocrinol Invest 2008; 31: 597–601.
10. 10. Davison SL, Bell R, Donath S et al. Androgen levels in adult females: changes with age, menopause, and oophorectomy. J Clin Endocrinol Metab 2005; 90: 3847–3853.
11. 11. Elmlinger MW, Kühnel W, Wormstall H, Döller. Reference intervals for testosterone, androstenedione and SHBG levels in healthy females and males from birth until old age. Clin Lab 2005; 51: 625–632.
12. 12. van Geel TA, Geusens PP, Winkens B et al. Measures of bioavailable serum testosterone and estradiol and their relationships with muscle mass, muscle strength and bone mineral density in postmenopausal women: a cross-sectional study. Eur J Endocrinol 2009; 160: 681–687 .
13. 13. Soriguer F, Rubio-Martín E, Fernández D et al. Testosterone, SHBG and risk of type 2 diabetes in the second evaluation of the Pizarra cohort study. Eur J Clin Invest 2012; 42: 79–85.
14. 14. Rexrode KM, Manson JE, Lee IM et al. Sex hormone levels and risk of cardiovascular events in postmenopausal women. Circulation 2003; 108: 1688–1693.
15. 15. Onat A, Hergenç G, Karabulut A et al. Serum sex hormone-binding globulin, a determinant of cardiometabolic disorders independent of abdominal obesity and insulin resistance in elderly men and women. Metabolism Clinical and Experimental 2007; 56: 1356–1362.
16. 16. Rosner W, Auchus RJ, Azziz R et al. Position statement: Utility, limitations, and pitfalls in measuring testosterone: an Endocrine Society position statement. J Clin Endocrinol Metab 2007; 92: 405–413.
17. 17. Mossakowka M, Szybalska A, Błędowski P. Geneza, cele i zakres badania. In: Mossakowska M, Więcek A, Błędowski P (ed.). Aspekty medyczne, psychologiczne, socjologiczne i ekonomiczne starzenia się ludzi w Polsce. Termedia Wydawnictwa Medyczne, Poznań 2012: 27–44.
18. 18. Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation of simple methods for the estimation of free testosterone in serum. J Clin Endocrinol Metab 1999; 84: 3666–3672.
19. 19. Moreno SA, Shyam A, Morgentaler AJ. Comparison of free testosterone results by analog radioimmunoassay and calculated free testosterone in an ambulatory clinical population. J Sex Med 2010; 7: 1948–1953.
20. 20. Matthews DR, Hosker JP, Rudenski AS el al. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985; 28: 412–419.
21. 21. Alberti KG, Eckel RH, Grundy SM et al. Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation 2009; 120: 1640–1645.
22. 22. Milewicz A, Arkowska A, Mieszczanowicz U et al. Indeks wolnych androgenów i wolnego estradiolu jako czynnik ryzyka chorób układu krążenia u Polaków po 55. roku życia. In: Mossakowska M, Więcek A, Błędowski P (ed.). Aspekty medyczne, psychologiczne, socjologiczne i ekonomiczne starzenia się ludzi w Polsce. Termedia Wydawnictwa Medyczne, Poznań 2012: 307–315.
23. 23. Haring R, Hannemann A, John U el al. Age-specific reference ranges for serum testosterone and androstenedione concentrations in women measured by liquid chromatography-tandem mass spectrometry. J Clin Endocrinol Metab 2012; 97: 408–415.
24. 24. Cappola AR, Ratcliffe SJ, Bhasin S et al. Determinants of serum total and free testosterone levels in women over the age of 65 years. J Clin Endocrinol Metab 2007; 92: 509–516.
25. 25. Bjerner J, Biernat D, Fosså SD, Bjøro T. Reference intervals for serum testosterone, SHBG, LH and FSH in males from the NORIP project. Scand J Clin Lab Invest 2009; 69: 873–879.e1–11.
26. 26. Vanbillemont G, Bogaert V, De Bacquer D et al. Polymorphisms of the SHBG gene contribute to the interindividual variation of sex steroid hormone blood levels in young, middle-aged and elderly men. Clin Endocrinol (Oxf) 2009; 70: 303–310.
27. 27. Brand JS, van der Tweel I, Grobbee DE et al. Testosterone, sex hormone-binding globulin and the metabolic syndrome: a systematic review and meta-analysis of observational studies. Int J Epidemiol 2011; 40: 189–207.
28. 28. Maggio M, Lauretani F, Ceda GP et al. Estradiol and metabolic syndrome in older italian men: The InCHIANTI Study. J Androl 2010; 31: 155–162.
29. 29. Weinberg ME, Manson JE, Buring JE et al. Low sex hormone-binding globulin is associated with the metabolic syndrome in postmenopausal women. Metabolism. 2006; 55: 1473-1480. Erratum in: Metabolism. 2007; 56: 296.
30. 30. Prajsner A, Chudek J, Dąbrowski P et al. Choroby gruczołu krokowego u osób w wieku podeszłym w Polsce. In: Mossakowska M, Więcek A, Błędowski P (ed.). Aspekty medyczne, psychologiczne, socjologiczne i ekonomiczne starzenia się ludzi w Polsce. Termedia Wydawnictwa Medyczne, Poznań 2012; 317–334.
31. 31. Bar-Andziak E, Milewicz A, Jędrzejuk D et al. Thyroid dysfunction and thyroid autoimmunity in a large unselected population of elderly subjects in Poland — the ‚PolSenior’ multicentre crossover study. Endokrynol Pol 2012; 63: 346–355
32. 32. Janse F, Eijkemans MJ, Goverde AJ et al. Assessment of androgen concentration in women: liquid chromatography-tandem mass spectrometry and extraction RIA show comparable results. Eur J Endocrinol 2011;165: 925–933.