the chapter is then organized by health outcome. For each health outcome section there is a brief introduction on epidemiology, risk factors, and biological plausibility, followed by an overview of studies that have been conducted on the correlations between the outcome and changes in endogenous testosterone levels during aging. A description of the randomized placebo-controlled trials in older men is provided in each section, with detailed tables on the results specific to that outcome.
Early studies of testosterone levels and aging found conflicting evidence regarding changes in endogenous testosterone levels, but recent studies have consistently reported declining levels with aging. Some of the earlier discrepancies have been attributed to various health conditions and inconsistent timing of sera drawn for testosterone measures (Tenover, 1994). Normal values of testosterone vary widely in older men, and the particular level that is considered to be abnormally low is not consistent in the literature. Additionally, whether total testosterone, free testosterone, bioavailable testosterone, or some combination is the most appropriate measure has been debated. This section highlights the results of several large cohort studies that have compared endogenous testosterone levels among various age groups (Box 2-1). Many of the studies are cross-sectional in design, with serum hormone level and age considered at the same point in time. Blood specimens for these studies (Table 2-1) were collected from participants in the morning.
Harman and colleagues (2001) examined changes in testosterone and sex hormone binding globulin (SHBG) levels over time among participants in the Baltimore Longitudinal Study of Aging (BLSA) (Table 2-1). During a 6-month period in 1995, sera from 890 participants’ most recent and several previous visits (up to 10 samples per man) were retrieved. Cross-sectional plots of earliest total testosterone, SHBG, and free testosterone indices [(FTI) = total T/SHBG] versus age show a negative association with age for the two testosterone measures. An increase in SHBG with age was more apparent at older ages (>50 years) than among the younger decades of age. Longitudinal analysis based on all men with sera for at least two visits (N = 702) showed similar downward trends of testosterone for each decade of age from the 30s to the 80s; downward trends for FTI were found for each decade except the 80s (Figure 2-1). Multivariable analysis found age associated with a decrease in testosterone and FTI at a relatively constant rate, independent of obesity, illness, medications, cigarette smoking, or alcohol intake. Total testosterone decreased an aver-