Hormonal homeostasis in a group of 216 aging Czech males and correlation with responses to a questionnaire of the University of St Louis

Abstract

The male aging process is accompanied by changes in the levels of several types of hormones. Falling levels of androgenic-anabolic steroids (total testosterone, free testosterone, biologically accessible testosterone, dehydroepiandrosterone) correspond to a group of symptoms referred to as PADAM syndrome (Partial Androgen Deficiency in the Aging Male). In the case of those carefully examined patients with symptoms of PADAM and proven hypogonadism, administering androgen supplements can alleviate some of the undesirable manifestations. In its literature, the University of St Louis repeatedly refers to its questionnaire as a verbal tool for the detection of possible hypogonadism.

The aim of this study was to ascertain to what extent the aging process is evident in hormonal homeostasis detected in laboratory testing, and the extent to which this data is in accord with the evaluation of responses to questions in the University of St Louis questionnaire. Method: 216 men aged over 50 years were examined. Measurements were taken of: testosterone; the index of free testosterone; androstenedione; dihydrotestosterone; dehydroepiandrosterone and its sulfate; isomers 7α- and 7β-hydroxydehydroepiandrosterone; epitestosterone; luteinizing hormone (LH); follicle-stimulating hormone (FSH); prolactin; and sexual hormone-binding globulin (SHBG). Evaluations of the patients' responses to the University of St Louis questionnaire were compared with the results of the laboratory tests. Results: The study confirms that the most prominent phenomenon is that of an age-related decrease in the index of free testosterone, which is indicated in particular by an increase in the level of SHBG, and by a decrease in dehydroepiandrosterone and its derivatives. No significant correlation was found between levels of hormones and single items on the questionnaire, nor with the overall score arrived at by studying the patients' data.

Keywords:

• Aging
• PADAM
• andropause
• androgens
• dehydroepiandrosterone
• St Louis questionnaire

Introduction

In recent decades, countries with developed economies and a corresponding standard of healthcare have seen a significant increase in average life expectancy. The increase is not the same for both sexes: for men the average age is 5–7 years shorter. Sources of the WHO [1] estimate that in 2025 the average life expectancy will be 30 years longer than that of the 1950s.

The study of the changes which accompany the male aging process has gradually led to the recognition of a group of symptoms considered to share a common basis – a hormone deficiency, of androgens in particular. At a conference of the Austrian Urological Society in 1994, the acronym PADAM (Partial Androgen Deficiency in the Aging Male) [2] was suggested for the condition created by this symptomatology, with psychological, vasomotoric, metabolic, somatic, sexual and other symptoms. Sometimes the less precise term ‘andropause’ is used for the PADAM syndrome, as an analogue of the menopause. The PADAM syndrome, which can be considered a clinical unity, is not a precise equivalent of the female menopause, since the decrease in sexual hormones is a long-term phenomenon, clinical symptoms and laboratory values are highly variable, and fertility is maintained.

Hitherto, the normal values of older people for the evaluation of laboratory findings have not been well charted. In pathological findings relating to the elderly, the standard deviation is taken as double that of the physiological values of thirty-year-old males [3]. Hormone deficiency, in particular of androgens, led logically to the idea of hormone supplements, just as in the case of women, where the positive results of such treatment have been proven, albeit with known risks accompanying long-term treatment. The decision to commence hormone treatment is not an easy one: it demands an evaluation of the patient's current state of health, including the relevant laboratory values. The patient must be informed about possible undesirable side effects and carefully monitored during treatment.

Aim of the study

The primary, stated aim of the study was to establish the serum levels of a select group of androgens and other hormones in older Czech males, and then to evaluate responses to the University of St Louis questionnaire – which is often cited as a verbal guide for the detection of hypogonadism – and compare these responses to the laboratory findings for the individual patients [4].

Patients and methods

Patients

A total of 216 male volunteers aged over 50 years, who met the requirement of a three-year monitoring period, were examined. The composition of the group tested was the closest possible match for the overall composition of the clientele of the authors' urological outpatient clinic (Urocentrum Prague).

The algorithm of the examination was as follows: the patient was given printed information offering general guidance about the PADAM syndrome, and was offered the opportunity of a comprehensive examination focusing on andropause. Those who agreed (out of 225 questioned, 216 consented) were invited to a focused initial examination and received the University of St Louis questionnaire (see Table I). Two professional translators translated the questionnaire into and then from Czech, so that errors relating to the language barrier were minimalized.

Table I.  The University of St Louis questionnaire as a general guide to the discovery of androgen deficiency

1. Do you have a decrease in libido (sex drive)?

2. Do you have a lack of energy?

3. Do you have a decrease in strength and/or endurance?

4. Have you lost height?

5. Have you noticed a decreased ‘enjoyment of life’?

6. Are you sad and/or grumpy?

7. Are your erections less strong?

8. Have you noticed a recent deterioration in your ability to play sports?

9. Do you fall asleep after dinner?

10. Has there been a recent deterioration in your work performance?

Table II.  The association between hormonal indices and values reflecting responses to the University of St Louis questionnaire

– No correlations with serum hormone levels were found for questions 1, 4, 5 and 7.

2. Do you have a lack of energy?

– There was a lower level of epitestosterone in men who did not profess a loss of energy (Figure 4A).

3. Do you have a decrease in strength and/or endurance?

– There was a higher level of DHEA in men not professing a loss of strength (Figure 4B).

6. Are you sad and/or grumpy?

– Individuals not professing feelings of unhappiness had higher levels of androstendione, DHEA and 7β-OH-DHEA, and lower levels of LH (Figures 4C–F).

8. Have you noticed a recent deterioration in your ability to play sports?

– Individuals not professing a drop in their satisfaction with sport had lower levels of DHEAS (Figure 4G).

9. Do you fall asleep after dinner?

– Individuals not professing drowsiness after meals had higher levels of 7α-hydroxyled-DHEA (Figure 4H).

10. Has there been a recent deterioration in your work performance?

– Individuals not professing deterioration in work performance had higher levels of T and SHBG and a lower index of FT and LH (with more profound differences in subjects who were originally manual workers) (Figure 5).

The examination included recording the patient's history (work, family, social and sexual), evaluating his physical condition and psychological activity, and measuring his blood pressure, height and weight. Comorbidity and existing pharmacotherapy were established, and ultrasonography of the urogenital tract performed.

Laboratory methods

Laboratory blood testing and urine examination followed, and always took place between 8 a.m. and 9 a.m. Blood samples were centrifuged, and plasma was placed in freezing boxes at a temperature of −20°C. Minerals, urea, creatinine, aspartate transaminase (AST), alanine transaminase (ALT), glycemia, testosterone (T), the index of free testosterone (IFT), dihydrotestosterone (DHT), androstenedione, prolactin, sexual hormone-binding globulin (SHBG), luteinizing hormone (LH), follicle-stimulating hormone (FSH), dehydroepiandrosterone (DHEA), dehydroepiandrosterone sulfate (DHEAS), 7-hydroxyled metabolites of DHEA (7α- and 7β-OH-DHEA) and epitestosterone (epiT), as well as full blood count and urine sedimentation, were evaluated. In addition, a microbiological examination of the urine was performed. For the purposes of communication the laboratory hormonal indices, the patient's profession, possible obesity, blood tension and current pharmacotherapy were also monitored. Possible relationships between hormone levels and answers to the University of St Louis questionnaire were also evaluated.

Commercial RIA and IRMA kits from Immunotech and Orion-Diagnostica were used for measurement of the hormones, with the exceptions of epiT and 7-hydroxyled-DHEA; these latter steroids were measured using methods developed in the steroid laboratory of the Institute of Endocrinology in Prague [5-7].

Statistics

A single person made the evaluations and a single specialist undertook the computer analyses. The age dependence of the hormones was evaluated using backward stepwise polynomial regression; because of the non-Gaussian distribution in the concentrations of the hormones, the original data were subjected to power transformation to attain the minimum skewness of the studentized residuals with an absolute value of less than 2 (in order to eliminate the effect of outliers on the shape of distribution). The retransformed 95% confidence intervals of prediction for the occurrence of individual experimental points over an area covering 95% of the subjects may be considered the age-dependent reference limits for the given hormone. The degree of the polynomial was determined using the correlation coefficient of multiple regression, adjusted to degrees of freedom, standard error of estimation (the square root of the standard quadratic error of estimation), mean absolute error and Akaike information criterion. The degree of the polynomial did not exceed a value of 1 and, therefore, either straight linear age dependence or independence was observed. The eventual monotonic curvilinear relationships resulting from the non-Gaussian data distribution were linearized using power transformations. The transformations also improved the data distribution and stabilized the variance. The statistical significance of the model was determined using Fisher's test of the ratio of the variance explained by the model to unexplained variance. Regression diagnostic plots (Williams graph, rankit graph, variance graph and graph of studentized residual symmetry) were used for the detection of outliers and high leverage points. The Williams graph was compiled using Statgraphics Plus v3.0 statistical software. The remaining diagnostic graphs were an integral part of the software. All outliers and high leverage points excluded from the computation of predictions and confidence intervals were, however, retained in the figures for completeness.

For the evaluation of the relationship between the questions of the University of St Louis questionnaire and serum hormone levels, a one- or two-way analysis of covariance adjusted to constant age was used. The first factor was the type of response (yes/no) and, in the cases of the evaluation of work performance and physical activity, the factor of original employment (intellectual/manual) was also taken into account. As in the case of the regression analysis, in order to stabilize the group variance and to approximate a Gaussian data distribution, the original data were transformed by power transformation for the minimum skewness of the normalized residuals with an absolute value of less than 2 (in order to eliminate the effect of outliers on the shape of distribution). The residuals with an absolute value exceeding a value of 2 were considered outliers, and the corresponding values were not included in the analysis. Analysis of the residuals was carried out using diagnostic graphs for one-dimensional data with the use of Statgraphics Plus. The group means with 95% confidence intervals were retransformed to the original scale.

Results

The pattern of age-related hormonal values is illustrated graphically in Figures 1A–D, 2A–D and 3A–D. The marked regression curve with 95% reliability intervals and 95% confidence intervals for the incidence of the individual experimental points delimits a band which should include 95% of the subjects from the population studied. Of the main androgenic indicators, T and DHT did not show a significant reduction with age, while the index of FT decreased significantly, in line with the increase of SHBG. A drop in androstenedione was also evident. In the case of T (Figure 1A), however, a significant number of values lay below the levels for which replacement therapy is recommended (under 11 nmol/l), and 9 out of 212, i.e. 4.25%, were in the severe deficiency range. As with T, the level of DHT (Figure 1B) did not show a descending trend, but the number of probands with below-normal values (under 0.9 nmol/l) was significant (8.9%). By contrast, FT decreased in a linear manner depending on age (Figure 1C), as did androstenedione (Figure 1D).

Figure 1. Correlation between age and serum levels of testosterone, dihydrotestosterone, the index of free testosterone and androstenedione. The solid line shows the retransformed regression curve, the dashed line closer to the regression curve the 95% confidence interval. The area surrounded by dashed lines further from the regression curve defines a confidence interval of prediction for an individual subject, should include 95% of subjects, and is considered to be an age-dependent interval of reference values.

The testosterone precursors DHEA (Figure 2A) and DHEAS (Figure 2B) both fell significantly with increasing age. An age-related decrease was also found in both 7-hydroxy isomers of DHEA (Figures 2C, D).

Figure 2. Correlation between age and serum levels of dehydroepiandrosterone (DHEA), its sulfate (DHEAS), 7α-hydroxydehydroepiandrosterone (7α-OH-DHEA) and 7β-hydroxydehydroepiandrosterone (7β-OH-DHEA). The solid line shows the retransformed regression curve, the dashed line closer to the regression curve the 95% confidence interval. The area surrounded by dashed lines further from the regression curve defines a confidence interval of prediction for an individual subject, should include 95% of subjects, and is considered to be an age-dependent interval of reference values.

A significant rise was confirmed in both gonadotropins (LH, FSH) and in SHBG with increasing age (Figures 3A, B, D). Epitestosterone (data not shown) and prolactin, by contrast, were independent of age (Figure 3C); in the latter hormone, however, almost 8% of the values slightly exceeded the upper reference limit (15–20 mg/l).

Figure 3. Correlation between age and serum levels of LH, FSH, prolactin and SHBG. The solid line shows the retransformed regression curve, the dashed line closer to the regression curve the 95% confidence interval. The area surrounded by dashed lines further from the regression curve defines a confidence interval of prediction for an individual subject, should include 95% of subjects, and is considered to be an age-dependent interval of reference values.

Accordingly, the possible relationships were assessed between responses to individual questions on the University of St Louis questionnaire. Of the values found, only those where the association reached significance were addressed (Figures 4, 5, and see also Table II).

Figure 4. Correlation between the questions in the University of St Louis questionnaire and serum hormone levels. Columns with error bars represent the retransformed group median values and their 95% confidence intervals. The first value in the frame defines the significance level of the differences between the subjects answering positively and the subjects answering negatively. The particular question always figures in the graph headline. The description on the second line shows the significance level of the age dependency of a particular hormone (****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05), with its positive (+) or negative (−) relationship noted.

Figure 5. Correlation between question 10 of the University of St Louis questionnaire concerning the deterioration of work performance and selected androgen parameters. Columns with error bars represent retransformed group median values and their 95% confidence intervals. The subjects were subdivided into those who worked intellectually and those worked manually during their active life. The first value in the frame defines the significance level of the differences between the subjects answering positively and the subjects answering negatively. The particular question always figures in the graph headline. The second value in the frame defines the significance level of the differences between the types of occupation (manual or intellectual). The third value shows the interaction between the type of occupation and the question in the graph headline, and represents the extent of the differences between the positive/negative answers and the original occupation type. The description behind the word age shows the significance level of the age dependency for a particular hormone (****p < 0.0001, ***p < 0.001, **p < 0.01, *p < 0.05), with its positive (+) or negative (−) relationship noted.

Discussion

The group of psychosomatic and internal changes which accompany the aging process of men in the last third of their lives extends over a long period, during which there is great symptomatological variability and a visible tendency for some hormonal indicators to fall or, in some cases, rise.

The usual comparison made is that between the hormone levels of aging males and the concentrations of hormones at their optimal levels, i.e. at around the age of thirty. As a rule, the decrease in androgens and the simultaneous slight increase in gonadotropins and SHBG are seen as characteristic [8-10]. The decrease in FT from the age of 40 at a rate of around 1% per year is stated, along with the fact that around 20% of men aged between 60 and 80 have values of FT below the lower norm [11]. The authors were able to confirm a decrease in the index of FT in the study group of men aged over 60; this drop is due to the continuous rise of SHGB, which binds free testosterone, rather than to a decrease in total testosterone. The level of total testosterone in the study group of men over 50 years (with a higher proportion of probands in the 60–80 age range) did not fall with increased age; this is in accord with the observations of Barrett-Connor et al. [12] in their monitoring of population groups in the US Rancho Bernardo study. In the present group, the levels of DHT and prolactin did not change.

As in previous literature, significant decreases in DHEA, DHEAS, both 7-hydroxyled isomers of DHEA metabolites and in androstenedione were observed. The varied effects of DHEA and its sulfate have been documented for several years in the literature, and these are customarily recommended as hormones which restore youth. Even though a relatively significant or even dramatic fall in dehydroepiandrosterone and its derivatives with advancing age has been proven, the effect is sometimes subject to doubt [13-18]. For the moment, the extent to which these adrenal prehormones are in a causal relationship with some aspects of PADAM, e.g. depression, erectile dysfunction, lower immunity or glycide metabolism disorders, cannot be determined. However, as a treatment administered in a dosage, which raises the circulating level to that of a younger age, DHEA has, in some cases, been used in the therapy of erectile dysfunction, certain autoimmune disorders, and depression [19-22]; changes in levels of DHEA and DHEAS could therefore be added to the analysis of androgens as a reliable marker of adrenopause. According to certain data [23], since the 7-hydroxyled metabolites of DHEA are even greater bearers of the positive effects of DHEA, particularly in the area of the modification of immune response, their monitoring might also contribute information on the state of hormone regulation. This is also the case with epitestosterone, which is sometimes adjudged to have an antiandrogenic effect [24].

As regards the questions used in the University of St Louis questionnaire – often cited as a useful tool in detecting hypogonadism [4],[25]– a possible correlation was sought in the values of the variables monitored. A positive response to questions 1 or 7 should signal an androgen deficiency, as should positive answers to any other three questions. However, the comparison of positive answers to drops in the androgen levels in this group revealed no significant relationship.

The authors are therefore of the opinion that the answers to the questions posed by the questionnaire do not necessarily indicate possible hypogonadism, and recommend that further information on the patient be sought before reaching a decision on the approach to further diagnosis. In this respect, a questionnaire targeting further essential information (e.g. the existence of a sexual partner, habits and medication), such as that used by Nicopoulou and Adamopoulos [26], provides more comprehensive data.

Some of the patients studied were sexually abstinent after the loss of a partner, while in other cases pressure of work or serious comorbidity had had a marked negative impact. Financial problems, too, can sometimes play an important role; lifestyle and personal values had a significant influence. It is essential to recognize and distinguish these personal preferences properly, since they document the patient's concept of quality of life. No evidence was found of hormone deficiency, which was significantly linked to particular questions [1],[7], nor was it proven that a drop in androgens related to positive answers to the first three questions.

A wide variance in the relationship between symptoms and hormone levels is evidenced by the results of this study, and this is supported by the literature [27,28]. Only a few links were significant in statistical terms. Decisions regarding hormone supplements demand a very thorough and faithful appraisal of all of the information [29,30]; the regular monitoring of patients is essential [25],[31]. The active interest of the patient himself in the improvement of his own quality of life has a fundamental effect on the success of the treatment. Preparations which work selectively and contain modified molecules possessing an affinity with specific receptors undoubtedly represent a promising type of therapy in the future [32]; such treatment would reduce the undesirable side effects of the androgens administered hitherto.

Conclusions

The aim of this project was to provide the authors with their own information on the age-related changes in the serum levels of androgens in a group of 216 men aged over 50. The results did not reveal statistically significant drops in the most important androgens, testosterone and dihydrotestosterone, often cited in the literature. After reaching the age of 60, however, the decrease was not very pronounced. Furthermore, evaluation of the responses to the University of St Louis questionnaire did not provide sufficiently reliable information about possible androgen deficiency.

Even so, the values ascertained and the corresponding changes in serum levels can provide important guidance when the administration of hormone supplements is considered. Treatment programs should include dietary recommendations, suitable forms of exercise, stimulation of psychological interests, modification of the daily regime and, of course, the treatment of comorbidity. This thorough approach should be coordinated by an urologist, in close collaboration with other specialists. Such systematic care may then represent a qualitatively new medical service for the aging male.

Acknowledgements

This study was supported by research project MZ 9A8255-3 of the Internal Grant Agency (IGA) of the Ministry of Health of the Czech Republic.