Abstract
Introduction
Testosterone is an androgenic hormone found in men and, at lower concentrations, in women. 1 Testosterone is synthesized from cholesterol by Leydig cells of the testis, and is transported in serum with approximately 98% bound to protein. Only 1% to 2% of testosterone is not bound to protein, commonly referred to as “free testosterone.” Testosterone can act directly on target tissues or via conversion to the more potent androgen dihydrotestosterone (DHT) by 5-α reductase. Testosterone and DHT bind to the androgen receptor and regulate genes that influence sexual function and many other processes, such as insulin sensitivity and direct vasodilation.2,3
It is estimated that testosterone levels decline by approximately 1% per year among men starting at the age of 40 years, primarily because of testicular functional decline. Testosterone decline is also associated with reduced stimulation via the hypothalamic–pituitary–testis axis. 4 The diagnosis of treatable hypogonadism generally requires both symptoms (including low libido, erectile dysfunction, or reduced testicle size) and low serum testosterone. 5 There is no consensus on a lower limit of normal for serum testosterone levels; however, it has been suggested that serum testosterone greater than 350 ng/dL usually does not require treatment while men with levels less than 230 ng/dL may benefit from treatment. 6 Many men have low levels of testosterone but do not exhibit symptoms of hypogonadism; for example, one study found that 39% of men 45 years and older presenting to a primary care practice had testosterone levels less than 300 ng/dL. 7 However, the prevalence of symptomatic men with androgen deficiency has been estimated at 6%. 8 Unfortunately, there are known risks associated with testosterone treatment and limited data to ensure safety in long-term use. 9
The effects of testosterone impact multiple organ systems. Short-term intracoronary administration of testosterone was found to induce coronary artery dilatation and increased blood flow in men with heart disease. 10 Although cardiovascular disease is more common in men than women, men with lower levels of testosterone have an increased risk of cardiovascular disease,11-16 as well as an increased risk of obesity, diabetes, and all-cause mortality.17-20 Proposed mechanisms include modulation of arterial smooth muscle tone and blood pressure,21-23 lipid and lipoprotein metabolism24-28 and insulin or inflammatory signaling.29-31 Treatment with exogenous testosterone has been shown to improve several of these parameters, although the long-term consequences of such treatment are unknown.32-41
However, conflicting research has been reported as to whether these associations are mutually independent or are caused by a multifactorial pathway involving obesity, inflammatory cascades, or direct physiologic changes. Even less understood is the association between testosterone levels and lifestyle factors such as smoking and diet. A study of military males found those who were physically active had significantly higher testosterone levels compared with their matched controls (ie, low physical activity), 42 although an older study did not show a conclusive relationship. 43 Several studies have also found that smoking was associated with either an increase or a decrease in testosterone levels that may depend on the age that smoking was initiated and the history of nonsmokers (never smoked vs quit smoking).44,45 The purpose of the current study was to assess the relationship between free testosterone, chronic disease states, and lifestyle factors such as diet, smoking, and exercise status.
Methods
Participant Selection
The study was a cross-sectional study involving a convenience sample of 571 non-Hispanic whites, non-Hispanic blacks, and Hispanics/Latinos recruited from 12 participating primary care sites. Participants were eligible for the study if they were older than 44 years, self-identified as non-Hispanic white, non-Hispanic black, or Hispanic/Latino, and had no history of self-reported cardiovascular disease (coronary artery disease, peripheral arterial disease, history of myocardial infarction or stroke, or congestive heart failure), renal failure, or cirrhosis. All consented participants underwent a 1-hour face-to-face interview that assessed demographic, psychosocial, and medical history measures and a 25-mL blood draw. Three milliliters were sequestered in 352 samples and manipulated for blood spots and specimen storage in a monitored −80°F freezer. The present substudy was a post hoc evaluation using 159 frozen male serum samples for testosterone analyses. All study procedures were approved by the University of North Texas Health Science Center and John Peter Smith Health System Institutional Review Boards.
Study Procedures
Participants completed weight, height, body mass index, and blood pressure (millimeters of mercury [mm Hg]) measurements. Automated sphygmomanometers were used to measure systolic and diastolic blood pressures in each arm using a size-appropriate cuff; an average systolic and diastolic blood pressure was calculated for each subject using these measurements. Elevated blood pressure was defined as an average systolic blood pressure ≥140 mm Hg or a diastolic blood pressure ≥90 mm Hg.
Demographic and Health Behavior Measures
The study used standardized questions from the Behavioral Risk Factor Surveillance System to collect a selected number of demographic and health behavior information. Age was registered as years and gender was recorded. Race/ethnicity was assessed using the federal Office of Management and Budget standards and then categorized as non-Hispanic white, non-Hispanic black, and Hispanic/Latino. Smoking status was assessed by asking, “Have you smoked at least 100 cigarettes in your entire lifetime?” Diet was categorized as healthy (high amounts of fiber, fruits, vegetables, poultry, and fish) or unhealthy (high amount of red meats, fried foods, and fast-foods). Exercise was measured by assessing the extent of physical activity outside of work and during normal daily activities for exercise within the past month.
Medical Histories and Classifications
Participants were asked if they were currently taking any medications or had been diagnosed with a chronic medical condition by a medical professional. Participants were classified as having hypertension if they reported either a history of hypertension, were actively taking antihypertensive medications, or had a systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg. Subjects were categorized as having dyslipidemia if they reported a history of dyslipidemia, were on a lipid lowering medication, or had a fasting low-density lipoprotein level ≥160 mg/dL. Subjects were considered diabetic if they reported a history of diabetes, were taking antiglycemic medications, or had fasting glucose level of ≥126 mg/dL.
Testosterone Analyses
Free testosterone levels were measured from archived plasma samples using a 96-well, enzyme-linked immunosorbent assay–based method, according to the manufacturer’s protocol (BioVendor Research & Diagnostic Products, Candler, NC). Quantitation is enabled by including a standard curve along with the experimental samples in each 96-well plate. Final free testosterone levels are reported in ng/mL and function as this study’s dependent variable.
Statistical Methods
Statistical analyses were performed using Statistical Package for the Social Sciences (SPSS) version 19.0. Descriptive statistics are provided for demographic characteristics. Counts and frequencies are listed for categorical variables; means and standard deviations are provided for continuous variables. Stem and leaf plots and kurtosis/skewedness statistics were performed to assess for data normality. This was performed to not violate assumptions related to the regression analyses. Twelve outliers were excluded in the analyses resulting in a final N of 147. Simple and multiple linear regression models were used to assess for associations between free testosterone levels and age, race/ethnicity, smoking, diet, exercise, obesity, diabetes, hypertension, and dyslipidemia status. Multicollinearity was assessed by tolerance and variation inflation factor in the adjusted models. No interaction terms or collinear relationships were identified between potential confounding variables.
Results
Study population demographics are shown in Table 1. The participants had a mean free testosterone level of 3.1 ng/mL (SD = 1.5) and average age of 56.8 years (SD = 7.9). There was a fairly even distribution of whites, blacks, and Hispanics/Latinos. Most participants smoked (59%), rated their diet as healthy (61%), engaged in regular exercise (76%), and were overweight or obese (86%). There was an equal distribution of participants with and without dyslipidemia. However, a majority had a history of hypertension (61%) and a history without diabetes (77%).
Study Population Demographics (N = 147). a
Total may not add up to 147 because of missing data.
The simple and multiple linear regression results are shown in Table 2. Simple linear regression analyses did not find significant associations between free testosterone levels and race/ethnicity, smoking, exercise, obesity, and dyslipidemia status. Increasing age, unhealthy diet, and presence of diabetes and hypertension were significantly associated with lower free testosterone levels.
Linear Regression Analyses of Free Testosterone, Lifestyle Factors, and Chronic Disease Status (N = 147).
In the multiple regression models, increasing age continued to be significantly associated with lower free testosterone levels. Compared with participants who reported a healthy diet, those reporting unhealthy diets were significantly more likely to have lower free testosterone levels after controlling for all potential covariates. In addition, diabetics, regardless of all other covariates, showed a trend toward significance for an association with lower free testosterone levels. However, hypertension was no longer significantly associated with free testosterone levels in the multiple regression model.
Discussion
Literature on the physiologic effects of endogenous testosterone and testosterone replacement therapy has grown substantially over recent years. The explosion of testosterone deficiency awareness has resulted in a large increase in men inquiring about testing and treatment. 46 As a result, the medical profession has a need to better understand the needs, benefits, and risks associated with testosterone testing and treatment. The current study provides insight about the importance of diet and its relationship to free testosterone levels independent of chronic disease status, age, or obesity status.
In our study, free testosterone levels were inversely correlated with age, consistent with previous reports. 4 Self-reported “unhealthy diet” status was also significantly associated with lower free testosterone levels. Diabetes and hypertension were significantly associated with low testosterone in the simple regression but not in the multiple regression analysis, although diabetes approached significance (P = .07). Research on the relationship between free testosterone and diet is limited, and suggests an important area of further investigation. Several possible explanations may be considered. First, high calorie intake is a major contributor to insulin resistance, and dysregulation of insulin signaling may in turn reduce the production of testosterone by Leydig cells of the testis.20,47 Second, poor diet may induce dyslipidemia, including low levels of high-density lipoprotein cholesterol (HDL-C). HDL is an important source of cholesterol substrate for steroidogenesis,48,49 and individuals with low HDL-C due to genetic conditions have reduced steroid production evidenced by lower urinary steroids. 50 Indeed, in the current study, low HDL-C levels were associated with reduced free testosterone in a separate analysis, lending some support to the possibility that poor diet may decrease testosterone via effects on HDL.
Finally, it has been postulated that obesity may be a significant contributor to testosterone deficiency because of the aromatization of testosterone to estrogen in adipose tissue. This pathway is especially relevant in elderly men because it is known that aromatase-specific activity increases with age.51,52 However, as stated above, the current study did not find obesity status to be associated with free testosterone levels in the regression models. This is in contrast to several studies, including the European Male Ageing Study, which found that high body mass index was associated with reduced testosterone 44 and that weight loss was associated with increased testosterone levels during aging. 45 These differences may be because of the study population, the method of analysis, or the testosterone outcome measure (total or free testosterone). It is important to note that poor diet was associated with lower free testosterone levels even when the model was controlled for age, obesity, dyslipidemia, hypertension, and diabetes, suggesting that diet may contribute to low testosterone levels in ways that are not completely understood. For example, it is likely that patients reporting a poor diet in this study consumed excess calories, but these diets may also be relatively deficient in one or more nutrients that promote normal gonadal function. Further research is needed to assess the impact of lifestyle interventions, including dietary modification, on testosterone levels and the associated comorbidities of testosterone deficiency.
There are several strengths and limitations to this study. The cross-sectional nature of the study precludes determination of causation. Testosterone is a protein-bound element; however, the current study used free testosterone to overcome this limitation. Since the study population originated from north Texas, generalizability to other populations remains undetermined. Texas has one of the highest smoking and obesity rates in the United States, hence, not necessarily representing the general population. 53 In addition, the original study in which the data are derived from was focused on cardiovascular health disparities and the study intentionally oversampled African American and Hispanics, which may have affected physiologic measures and lifestyle rates. Nonetheless, the multivariate regression analyses controlled for these potentially confounding factors. Diet, smoking, and exercise measures may underestimate true behaviors of participants. However, previously published outcomes by the authors found consistent relationships between expected associations of smoking and diet and cardiovascular measures such as coronary calcium score and visceral adipose tissue measures.54,55 There may be other unmeasured factors that may impact the study’s results such as psychosocial factors including stress. Although it is beyond the scope of the presented analyses, stress, especially among under-represented populations, may lead to unhealthy lifestyle choices or have direct physiologic consequences related to testosterone production. 56
In conclusion, this study implicates diet, in addition to advanced age, as a possible risk factor in the development of low testosterone levels. Further research is needed to understand the mechanism by which diet and other lifestyle factors affect testosterone levels.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the National Institute on Minority Health (P20MD001633) and Health Disparities (NIMHD) and NIH Loan Repayment Program.