Circadian Variation in Testosterone, Sex Hormone-Binding Globulin, and Calculated Non-Sex Hormone-Binding Globulin Bound Testosterone in Healthy Young and Elderly Men

Previous studies concerning the relationship of serum testosterone levels to aging in normal men have yielded apparently inconsistent results. Studies performing blood sampling in the morning have often shown an age-related decrease in testosterone levels, while those using afternoon samples have failed to show such a decrease. These results suggested to us the possibility that the circadian rhythm in serum testosterone levels might be altered with normal aging in men. Hourly blood samples were obtained for 24 h from 1 young (mean age, 52.2 yr) and 12 old (mean age, 17 yr) healthy men. Total testosterone levels were measured by RIA. The circadian rhythm in serum testosterone levels found in normal young men was markedly attenuated or absent in healthy elderly men; the early morning rise in testosterone levels characteristic of young men was not present in old age. Mean testosterone levels for the entire 24-h day were lower in healthy old men than in young men. These results demonstrate a clear decrease in serum testosterone levels in healthy old men compared to those in young men and provide an explanation for the inability to demonstrate an age-related decline in testosterone levels in earlier studies using serum samples obtained in the afternoon.

Serum testosterone (T) levels in men decline with age while serum LH levels, as measured by RIA, increase. To assess if the decline in serum T levels in healthy aging men is paralleled by an age-related decline in the bioavailable non-sex hormone-binding globulin (SHBG)-bound fraction of T and to determine whether there are age-related changes in LH secretion or LH control of T production, we studied 29 young (aged 22-35 yr) and 26 elderly (aged 65-84 yr) healthy men. All men had single random blood samples drawn, and 14 men in each age group underwent frequent blood sampling for 24 h, both before and after 7 days of clomiphene citrate (CC) administration. Both mean 24-h serum total T levels and non-SHBG-bound T were reduced in elderly men compared to those in young men (P less than 0.05), while estradiol and SHBG levels were similar in the 2 age groups. Serum FSH determined by RIA and LH by RIA and bioassay were higher in the elderly men compared to those in young men (P less than 0.05), but the ratios of LH bioactivity to immunoreactivity and the LH pulse frequency and amplitude were similar. After CC administration, mean serum total T and non-SHBG-bound levels in young men increased by 100% and 304%, respectively, while in older men these values increased by only 32% and 8%, respectively. However, CC-stimulated LH pulse characteristics and serum levels of estradiol, SHBG, FSH, and bioactive and immunoreactive LH were similar in the 2 groups. Thus, both at baseline and after CC stimulation, elderly men had significantly lower serum total T and non-SHBG-bound (bioavailable) T levels than did young men, despite similar or increased levels of bioactive LH and similar bioactive to immunoreactive LH ratios and LH pulse characteristics. These results suggest that major age-related changes in the hypothalamic-pituitary-testicular axis occur at the level of the testes and are manifested by decreased responsiveness to bioactive LH. Administration of CC to young and elderly men resulted in similar changes in LH pulse characteristics and LH bioactivity and immunoreactivity, suggesting preserved hypothalamic-pituitary responsiveness in the elderly.

The effect of advancing age on the chronobiology of testosterone (T) and luteinizing hormone (LH) secretion in healthy men was investigated. Twenty young (average age 30.4 yrs) and 14 elderly (average age 70.4 yrs) men underwent 10 min blood sampling for 25 hrs to evaluate the circadian periodicity of LH, LH pulse frequency, and T. Using cosinor regression analysis, young men were found to have a significant (p less than .05) circadian variation in LH pulse frequency, with slowing of LH pulses during the night (maximum slowing at 2230 hr). There was also a tendency for LH pulse amplitude to increase at night (p = .06) in young men. However, no significant circadian pattern in LH pulse frequency or amplitude was detected in the elderly men. Mean LH by radioimmunoassay (RIA) and bioassay did not vary over the 24-hr period in either age group. Both young and elderly men had significant circadian rhythms in serum T, although the rhythm in elderly men was considerably blunted and was shifted in time compared to the young. These data provide evidence for age-related changes in the circadian rhythms of LH pulse frequency and T secretion and suggest that the LHRH pulse generator loses its circadian rhythmicity with normal aging in men.