New Horizons in Testosterone and the Ageing Male

Tomás Ahern; Frederick C. W. Wu


Age Ageing. 2015;44(2):188-195. 

In This Article

Multiple Mechanisms Contribute to the Fall in Testosterone Levels With Ageing

Testicular and hypothalamic function decline with age. Study of post-mortem specimens shows that Leydig cell number is ~44% lower in men aged 50–76 than in men aged 20–48.[10] Congruently, the secretory capacity of the testes is substantially lower in older men than in younger men.[11] Although declining testicular function appears to be the main cause of low testosterone levels due to ageing, hypothalamic GnRH secretion, but not pituitary LH reserve, is lower in older men.[12]

Obesity contributes to the decline in testosterone levels with ageing. Total and visceral fat mass increase with ageing peaking normally at 65 years.[13] Obese men (BMI > 30 kg/m2) have lower total and free testosterone concentrations than lean men (BMI 20–25 kg/m2), and their testosterone concentrations decline more quickly.[2] Despite lower testosterone concentrations than lean men, LH concentrations are not elevated in those with obesity suggesting a hypothalamic-pituitary defect.[2] Possible explanations for hypothalamic-pituitary disruption in obese men include elevated cytokine concentrations[14] and insulin resistance.[15]

Chronic illness, which accompanies ageing, plays also a role in the fall in testosterone levels with ageing. Men with chronic illness have lower testosterone levels than healthy men.[2] Like men with obesity, LH concentrations are not elevated in those with chronic illness, suggesting a hypothalamic-pituitary defect.[2] Chronic illnesses, such as cardiovascular disease (CVD) and type 2 diabetes (T2DM), are associated with increased concentrations of pro-inflammatory cytokines,[16] which may disrupt the hypothalamus resulting in lower testosterone levels.[14]

Other possible, although less likely, causes for lower testosterone levels in older men include statin use[17] and vitamin D deficiency.[18]

Symptoms of Adult Onset Male Hypogonadism are Non-specific and Overlap With Many Symptoms That Develop With Normal Ageing

MMAS found that the prevalence of loss of libido increased, over the course of 9 years, from 30.6 to 41.1% and that the prevalence of erectile dysfunction increased from 37.4 to 42.3%.[19] Counterintuitively, symptoms of hypogonadism do not correspond to low testosterone concentrations (poor positive predictive value) and are not sensitive.[7,8] This is exemplified in BACH where among men aged over 50 only 20.2% of those with symptoms of hypogonadism had a low total testosterone level (≤10.5 nmol/l), and of men with a low testosterone level, only 20.1% reported low libido and only 29.0% reported erectile dysfunction.[8]

To overcome these difficulties, EMAS investigators defined LOH as the presence of three sexual symptoms (decreased frequency of morning erection, erectile dysfunction and decreased frequency of sexual thoughts) together with a total testosterone concentration <11 nmol/l and a free testosterone concentration <220 pmol/l[1] (Figure 1). Such a syndrome appears to affect ~3% of men aged 60–69 and ~0.1% of men aged 40–49 (Figure 2).[1] Over 9 years, the incidence rate is 10% and the remission rate is 55%.[19]

Figure 1.

Multiple correspondence analysis (MCA) showing associations between symptoms and levels of total testosterone and free testosterone in the training and validation sets [1]. In this MCA plot, variables (including low or normal testosterone levels and the presence or absence of symptoms) are considered to be highly associated if they are at the same distance and in the same direction from the origin where the horizontal axis (Axis 1) and the vertical axis (Axis 2) cross in the training set and the validation set. Thus, the clustering of the categories of the variables in close proximity to one another is indicative of a syndromic association, which is highlighted by broken-line circle. The values along the axes are indexes of the strength of the association between variables. On Axis 1, the presence of symptoms has positive coordinates (to the right of the origin), compared with the absence of symptoms, with negative coordinates (to the left of the origin). Axis 2 helps identify symptoms that are related to a low testosterone level. The red clusters indicate the presence of the three sexual symptoms, with coordinates similar to those of a low testosterone level. The blue clusters indicate the absence of symptoms, with coordinates similar to those of a normal testosterone level. In contrast, the three psychological symptoms and, to a lesser extent, the three physical symptoms are located far from coordinates for normal and low testosterone, indicating that these symptoms are unrelated or weakly related to the testosterone level. The cluster patterns of symptoms in relation to total or free testosterone levels in the training set are virtually identical to those in the validation set. To convert the values for total testosterone to nanograms per millilitre, divide by 3.467. To convert the values for free testosterone to picograms per millilitre, divide by 3.467. Full colour figures are available online.

Figure 2.

The prevalence of the syndrome in EMAS, overall and stratified by age, BMI and co-morbidity [1]. The syndrome of LOH as defined by at least three sexual symptoms associated with total testosterone levels of <11 nmol/l and free testosterone of <220 pmol/l. The overall prevalence of LOH is 2.1% (a), the prevalence of LOH increased with age from 0.1, 0.6, 3.2 to 5.1% at 40–49, 50–59, 60–69 and >70 years, respectively (a). The prevalence of LOH also increased with BMI (b) and co-morbidity (number of coexisting illnesses) (c).