Relationship between testosterone serum levels and lifestyle in aging men

Abstract

Objective. The aim of this study was to evaluate the association between serum levels of testosterone and free testosterone to lifestyle in aging males.

Methods. Men between 45 and 85 years were assessed regarding body mass index (BMI), nicotine and alcohol consumption, stress level, physical and social activity, and sleeping quality by a self-administered questionnaire. In parallel, serum levels of testosterone (T), free testosterone (fT), LH, FSH, DHEA-S, E2 and SHBG were obtained.

Results. In total, 375 men with a mean age of 59.9 years (9.2 ± SD) entered this study; 25.4% and 27.4% had hypogonadal testosterone or free testosterone serum levels, respectively. Nicotine consumption (smokers had higher levels of T and fT; p < 0.01), BMI (negative correlation to T; p < 0.01) and age (negative correlation to fT; p < 0.001) correlated with serum levels of testosterone or free testosterone. Physical and social activity, nicotine and alcohol consumption, stress level and sleep quality did not show a significant association with serum androgen levels.

Conclusion. This prospective study of 375 men aged 45 to 85 years confirms the correlation between age, BMI and smoking with serum levels of testosterone and free testosterone, whereas the investigated variety of lifestyle factors did not show a significant association to serum androgen levels.

Keywords:

• Androgen status
• aging
• BMI
• smoking

Introduction

The continuous decline of serum levels of testosterone (T) and free testosterone (f T) serum levels in aging men has been repeatedly demonstrated [1-4]. In view of its clinical relevance, there is an ongoing discussion regarding the impact of this decrease on a variety of symptoms such as a reduction of muscle mass, bone mineral density, well-being and cognitive function, and on the increased incidence of mood disorders, abdominal obesity and reduction of sexual activity and libido [5].

Furthermore, serum levels of T and fT are potentially associated with various lifestyle factors. Although results have often been contradictory, body mass index (BMI), nicotine and alcohol consumption, mental and physical stress, physical and social activity and sleeping habits may impact androgen levels in aging men [6-8].

It is of substantial interest to have detailed knowledge on factors others than aging influencing serum levels of T and fT, for two main reasons. Firstly, because of the possible confounding effects in clinical studies and secondly, because of the growing number of men seeking professional help to increase quality of life in the aging process. Therefore, it is of interest to investigate a possible correlation of T and fT serum levels with lifestyle parameters in aging men.

Methods

Participants

Men aged 45 years or older, seen at 11 different participating urological practices or outpatient clinics for routine urological check-up, were recruited for this study, which took place over a 6 month period. Participation was voluntary: exclusion criteria were liver or renal failure as assessed by medical history and laboratory tests; medication with a known effect on the endocrine system (e.g., finasteride) or other forms of hormonal therapy; or any malignant disease.

Lifestyle assessment

Lifestyle factors (social activity, physical activity, sleep quality and smoking/drinking habits), sociodemographic parameters (age, height, weight, matrimony/partnership, educational status) and medical history (previous surgical procedures, ongoing medication and previous cardiovascular, endocrine, gastrointestinal or urogenital tract diseases) were determined by a self-administered, standardized questionnaire.

Social activity was assessed by 8 questions proposed by Pincus et al. [9]. Sleep quality was quantified by 5 questions, as described by Buysse et al. [10]. Physical activity and drinking habits were quantified by one question with four possible answers (less than weekly, once per week, more than once per week, daily). Smoking status was determined by two questions, including the average number of cigarettes smoked daily. Stress level was evaluated by three questions. Those men who suffered from stress had to answer questions on frequency of stress episodes (less than weekly, once per week, more than once per week, daily) and on the degree of stress (mild, moderate, strong).

Laboratory tests

Blood samples were obtained from the study participants between 8:00 and 10:00. Eleven different laboratories provided T and fT determination, all using RIA for T and fT quantification. To avoid a type 2 statistical error due to possible differences of testosterone values when measured in different laboratories, each patient was classified according to his individual laboratory results to hypogonadal level (group I, below the lower cut-off value of his laboratory), lower half (group II, within the lower half of reference values of his laboratory) and higher half (group III, within the upper half of reference values of his laboratory) of T and fT.

This form of multicenter (and multilaboratory) design was chosen as it reflects daily practice; each urologist has to make a decision based on the locally available laboratory methods and results, each laboratory providing slightly different cut-off values for T and fT. In addition, a routine laboratory study (white and red blood count, liver and kidney function tests, lipid profile) and PSA was obtained.

Statistical analysis

All statistical analyses were conducted using the Statistical Package for Social Sciences, version 10.0.7 (SPSS Inc., Chicago, IL). Correlation between serum levels of testosterone and free testosterone, BMI, age and lifestyle factors were calculated by Pearson's Chi-square test. Results were corrected for age and BMI whenever possible. Multiple regression models were not performed. Statistical significance was set at p < 0.05 for all analyses.

Results

Cohort characteristics

A total of 375 men were recruited. Age of the study population was 59.9 ± 9.2 years (mean ± SD) and distributed as following: 45–55 years, n = 99; 56–65 years, n = 109; 66–75 years, n = 92; 76–85 years, n = 75. BMI averaged at 27.6 ± 3.9. In total, 91.7% of men were in a stable relationship, and cohabited with either a permanent partner or wife.

Alcohol and nicotine consumption, smoking status, stress, physical activity and subjective sleep quality were distributed as shown in Table I. A subnormal testosterone fT serum level was seen in 25.4% and 27.4% of men, respectively.

Table I.  Distribution of lifestyle-factors to fT and T-levels.

			Free Testosterone				Testosterone
		Total	Group I	Group II	Group III		Group I	Group II	Group III
		n = 375	n = 103	n = 120	n = 152	p	n = 95	n = 122	n = 158	p
Age (mean)		59.9	62.9	62.2	55.6	0.001	60.6	59.3	59.8	NS
BMI (mean)		27.6	28.2	27.6	26.8	NS	28.6	27.5	26.6	0.01
Smoking status (%)	Non-smoker	61.3	73.8	66.7	43.0		74.8	58.9	53.2
	Ex-smoker	17.1	14.8	11.1	17.2		14.4	16.7	18.8
	Smoker	21.6	9.8	22.2	39.7	0.01	9.9	23.3	27.9	0.01
Alcohol consumption (%)	<1/week	39.5	42.2	37.7	40.0		40.0	37.8	38.6
	1/week	19.0	26.7	9.8	20.0		15.6	18.9	15.0
	>1/week	28.5	17.8	32.8	20.0		33.3	27.0	22.2
	Daily	13.0	11.1	9.8	20.0	NS	6.7	13.5	15.0	NS
Stress level (%)	None	53.9	51.1	65.6	53.0		47.8	61.3	55.2
	Mild	20.0	20.0	18.0	14.0		22.2	16.2	20.8
	Strong	26.1	28.9	16.4	33.0	NS	30.0	22.5	24.0	NS
Social activity (%)	Low	22.2	24.0	19.0	19.0		25.0	29.0	20.0
	Moderate	48.8	46.0	58.0	52.0		50.0	51.0	56.0
	High	29.0	30.0	23.0	29.0	NS	25.0	30.0	24.0	NS
Physical activity (%)	<1/week	46.9	55.6	42.6	43.1		51.1	46.8	42.5
	1/week	22.3	13.3	19.7	26.0		17.8	20.7	17.0
	>1/week	24.1	20.0	19.7	22.0		20.0	22.5	20.3
	Daily	6.7	0.0	6.6	1.7	NS	0.0	3.6	6.5	NS
Sleep quality (%)	Low	26.4	27.0	20.0	40.0		20.0	31.0	34.0
	Moderate	44.5	30.0	46.0	36.0		50.0	40.0	42.0
	Good	29.1	43.0	34.0	24.0	NS	30.0	29.0	24.0	NS

Mean value (age, BMI, LH, FSH) and percentage (smoking, alcohol, stress, social activity, physical activity, sleep); Group I: hypogonadal levels; Group II: low levels; Group III: high levels. NS, not significant. p-value <0.001.

Lifestyle and total testosterone

As described above, men were divided into three groups according to their individual total T serum levels: hypogonadal levels (Group I; n = 95), lower half (Group II; n = 122) and upper half (Group III; n = 158) of normal range. Nicotine consumption (smokers had higher levels of T; p < 0.01) and BMI (negative correlation to T; p < 0.01), but not age, correlated with serum levels of testosterone. Reported physical and social activity, alcohol consumption, stress level and sleep quality were homogeneously distributed within the three groups of total testosterone status, and there was no statistically significant difference between men with hypogonadal (Group I), low (Group II) or high (Group III) testosterone levels (Table I).

Lifestyle and fT

The corresponding three groups for fT serum levels included 103 men in group I (hypogonadal level), 120 men in the lower half of normal range (Group II) and 152 men in the upper half of normal range. Of the hypogonadal men according to the total T values (Group I, total testosterone), 87.5% were also hypogonadal as classified by their fT values (Group I, fT); the corresponding figures for Groups II and III were 83.4% and 88.2%, respectively.

As described for total testosterone, smokers had significantly higher levels of fT (p < 0.01) than non-smokers. Furthermore, fT was significantly correlated to age (p < 0.001). None of the investigated lifestyle factors (alcohol consumption, stress level, social and physical activity, sleep quality) revealed a significant association to fT serum levels (Table I).

Discussion

The male aging process is accompanied by a number of endocrine changes, as, for example, was convincingly demonstrated by the Massachusetts Male Aging Study in 1991 [1]. Consequently, a number of symptoms such as reduced muscle mass, bone mineral density, well-being, cognitive function, libido, and increased incidence of mood disorders and abdominal obesity – frequently seen in aging men (also designated as partial androgen deficiency of the aging male (PADAM)) – have been associated with the described endocrine changes [5].

In the ongoing discussion concerning the clinical impact of decreasing androgen levels, different aspects of lifestyle turned out to potentially influencing both the endocrine system and the aging process in general. We therefore set up a prospective study to elucidate the association between lifestyle, testosterone and fT in aging urological patients.

Age, BMI

In contrast to previous reports [1-3], in our cohort of men undergoing a routine urological check-up, age correlated only with free (p < 0.001) but not with total testosterone. Only a few studies, for example the Rancho Bernardo study, also found that total testosterone did not decrease significantly with age [4].

BMI, defined as kg/m², is a useful tool used to measure obesity, reflecting the ratio of two important “lifestyle parameters”: diet and physical activity. In general, a negative correlation to serum testosterone is well established [11-13]. In concordance with previous reports by Field et al. [7], total testosterone, but not fT, was correlated significantly with BMI in our population (p < 0.0001). The lack of correlation with fT may be explained by the decrease of SHBG with an increasing BMI [8].

Smoking, drinking

After adjustment for age and BMI, nicotine consumption showed a significant positive correlation to total and fT levels (p < 0.01), and this confirms previous results [6],[7],[11]. This relationship is still elusive and requires further investigation.

Alcohol consumption did not show a T or fT, dependent distribution. In fact, alcohol consumption is not precisely estimated from self-report: honesty of participants on this point could have obscured the association. Furthermore, severe forms of alcohol abuse were potentially excluded in our collective, as men with impaired liver function test results were excluded, as mentioned above.

In contrast to two large population-based studies [6],[7], caffeine consumption was not assessed in our questionnaire.

Sleep quality and stress level

To our knowledge, only one study has investigated the possibility of a correlation between sleep and endogenous androgen levels so far, indicating a link between the circadian testosterone rhythm and the timing of rapid eye movement (REM) phases [14]. In our study, standardized questionnaires were used to assess sleep quality, sleep latency, sleep duration, sleep disturbances and daytime dysfunction [10]; no correlation of these parameters with T or fT was found.

Stress level was also determined by three questions: “Did you suffer from stress the last 4 weeks?”; if the response was positive, the participant was then asked to select from the following: “How often?” (less than weekly, once per week, more than once per week, daily) and “How strong?” (mild, moderate, strong). The working hypothesis was that mental stress would lower testosterone levels by raising cortisol levels, as described by Francis [15]. In fact, the subjective reported stress level did not show relevant association to testosterone levels in our cohort.

Physical and social activity

There are numerous reports on the effects of physical exercise on testosterone levels, reporting decreased T and fT in endurance trained men [16],[17], whereas persons involved in continuous strength training do not show changes in testosterone levels [18-20]. The only population-based study investigating physical activity in aging men found no significant correlation to total or fT level [6], as was also demonstrated in our 375 urological patients. While this setting is not adequate for proving whether physical training does have a significant influence on endogenous androgen levels, it would be of interest for daily routine to develop an easy tool to get fast information about this topic. Social activity was not found to be significantly influenced by T or fT serum levels, or vice versa.

Conclusions

In general, a correlation of serum testosterone and lifestyle factors was not found. Our prospective study of 375 men at urological routine check-up, aged 45 to 85 years, elucidates no significant relationship between the investigated lifestyle factors and serum levels of T and fT in aging men, but confirms the known correlation with age (for fT), BMI (for total testosterone) and nicotine consumption (for both fT and total testosterone).

Nevertheless, we believe that an important future field of research for urologists will be the aging process of the endocrine system, as well as the impact of lifestyle factors on health, well-being and aging, and their respective relationships.