| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Editorial |
Division of Endocrinology, Diabetes, and Metabolism, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104-6149
Address all correspondence and requests for reprints to: Peter J. Snyder, 700 Clinical Research Building, 415 Curie Boulevard, Philadelphia, Pennsylvania 19104-6149. E-mail: pjs{at}pobox.upenn.edu.
The role of gender in mortality has long fascinated medical investigators. Mortality due to cardiovascular disease in men exceeds that in women until the age of menopause, when mortality in women gradually catches up. One obvious explanation for this discrepancy is that estradiol production during premenopausal years partially protects women from cardiovascular disease, but for many years there was a lingering suspicion that testosterone itself is harmful.
Recent studies do not support the hypothesis that testosterone adversely affects overall mortality. Two recent prospective studies in community-dwelling men, for example, have shown no association between serum testosterone concentration and overall mortality. In one study, in Wales, 2323 men ages 45–59 were followed for an average of 16.5 yr, and during this time 482 men died (1). Hazard ratios for all deaths did not vary by quintile of testosterone and showed a weak inverse correlation with ischemic heart disease deaths. In another study, in Massachusetts, 1686 men ages 40–70 were followed for an average of 15.3 yr, and during this time 395 deaths occurred (2). The relative risk of death from all causes was not associated with either total or free testosterone.
A new study, published in this issue of the Journal (3), suggests the possibility that it is low levels—rather than normal or high levels—of testosterone that might be harmful. A community-dwelling group of 794 men, ages 50–91 (mean, 73.6) yr in Rancho Bernardo, California, had serum drawn and then were followed for an average of 11.8 yr. During this time 538 men died. The results showed that "men whose total testosterone levels were in the lowest quartile (<241 ng/dl) were 40% more likely to die than those with higher levels.... " Low testosterone was also associated with a similarly high (38%) risk of cardiovascular mortality. Low bioavailable testosterone was also associated with increased mortality. These associations were independent of age, adiposity, lifestyle, and metabolic disease.
At least two differences between the populations in Wales and Massachusetts and the population in Rancho Bernardo could explain why the former two showed no association between testosterone and overall mortality and the last did show a relationship. The first is that the populations in Wales and Massachusetts were younger by approximately two decades than that in Rancho Bernardo—mean or median ages in the early 50s vs. early 70s—raising the possibility that factors that contribute to mortality, including those related to testosterone, differ at different ages. Another difference between populations, which might be related to the age difference, is the serum testosterone concentration itself. The lowest quintile of serum testosterone concentration in the Massachusetts study was less than 370 ng/dl, whereas the lowest quartile in the Rancho Bernardo study was less than 241 ng/dl. Whether this difference was due to differences in methodology, age, or other factors is not known. If the difference is not the result of methodology and testosterone does contribute to, or is at least associated with, greater mortality, one would expect a greater effect on mortality with a more abnormal serum testosterone concentration.
Did low testosterone cause increased mortality in Rancho Bernardo men? To evaluate the possibility that the association of low testosterone with increased mortality was, conversely, the result of the condition(s) that caused the mortality, the investigators excluded the first 5 yr of observation and still found that in 5–20 yr of observation, men in the lowest quartile of total testosterone had a greater chance of mortality from all causes than those in highest quartile.
If low testosterone does cause increased mortality, what might be the mechanism? Low testosterone might be expected to lead to decreased muscle mass and bone strength, and thereby to more fractures, complications of which could lead to deaths that might be recorded as cardiovascular or respiratory. The data collected, however, do not allow evaluation of this possibility.
Because half of the deaths were recorded as due to cardiovascular disease, a link between low testosterone and cardiovascular disease should be explored. An effect of testosterone on serum lipids seems an unlikely link. A meta-analysis showed a weak effect of im testosterone on both HDL and LDL cholesterol (4), and a randomized, placebo-controlled trial of testosterone in elderly men with low-normal serum testosterone concentrations showed no effect on serum lipids, apolipoproteins or lipoprotein (a) (5). The possibility that testosterone could decrease blood clotting has been examined in small studies, but these are inconclusive.
Another possible link between low testosterone and cardiovascular mortality is metabolic risk factors, because in epidemiological studies low testosterone precedes the development of central obesity, the metabolic syndrome, and diabetes mellitus (6, 7, 8). In the Rancho Bernardo men reported herein, low testosterone was associated with central obesity and higher glucose and insulin concentrations. However, excluding or adjusting for metabolic syndrome, insulin resistance or diabetes did not diminish the association between low testosterone and increased mortality. Furthermore, studies of the effect of testosterone treatment do not support an effect of testosterone on the location of adipose tissue usually associated with cardiovascular risk. For example, in a study of 13 unequivocally hypogonadal men who were treated with testosterone enanthate for 18 months, sc adipose tissue, assessed by computerized tomography, decreased statistically significantly by 12%, but the 6% decrease in visceral adipose tissue was of borderline significance (9). Another example is a randomized, placebo-controlled study of elderly men with low-normal testosterone concentrations, in which testosterone-treated men experienced an overall decrease of 3 kg of fat mass, as assessed by dual-energy x-ray absorptiometry, in 3 yr, but the decrease was mainly appendicular, not central (10). All in all, the mechanism that links low testosterone to increased mortality remains elusive.
The significance of these new data from Rancho Bernardo is that they tend to swing the pendulum further. Whereas testosterone was once thought to be associated with increased mortality and more recent data suggested it had no effect, these new data suggest that a sufficiency of testosterone might even be beneficial to survival.
What studies should be done next? Additional epidemiological studies, especially in large populations of elderly men with a wide range of testosterone concentrations in which testosterone is measured by mass spectroscopy, would confirm or refute the suggestion of these new Rancho Bernardo data that a low testosterone concentration is associated with increased mortality. A randomized, placebo-controlled study of testosterone in elderly men with unequivocally low serum testosterone concentrations would provide information about possible effects of testosterone on known cardiovascular and metabolic risk factors, as well as on other areas that could indirectly affect mortality, such as mobility, cognition, and bone quality. If such effects are clearly demonstrated, a long-term clinical trial of the effect of testosterone on clinical cardiovascular disease and mortality of all causes would be warranted.
Received November 9, 2007.
Accepted November 13, 2007.
References
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
