The recent paper by Yeap et al. addresses an important and emerging issue: the association between testosterone and cardiovascular mortality in elderly men. Designed as a cohort study, the population was derived from community-dwelling elderly men in Perth, Western Australia. The men were given an initial work-up between the years 1996 and 1999, followed by another assessment in the years 2001–2004.
In total, 3690 men were followed for a mean of 6.7 yr. During this time, there were 947 deaths, with 325 directly attributable to ischemic heart disease (IHD). Overall, men who died had lower mean baseline levels of serum testosterone (12.8 vs 13.2 nmol/l; p = 0.013), dihydrotestosterone (DHT) (1.4 vs 1.5 nmol/l;p = 0.002), and estradiol (71.6 vs 74.0 pmol/l; p = 0.022). Following adjustment for several risk factors, both testosterone and DHT were found to be associated with all-cause mortality. Furthermore, men with baseline serum testosterone levels in the middle two quartiles (between 283 ng/dl and 453 ng/dl) had lower rates of mortality from IHD compared to men in the lowest or highest quartiles.
The findings of Yeap et al. illustrate an interesting U-shaped association between serum testosterone levels and cardiovascular mortality in which moderate levels appear to be protective. Furthermore, these findings appear consistent with recent studies in which elderly men with cardiovascular disease were examined , as well as data from the Testosterone in Older Men (TOM) trial .
In the first of these studies, Vigen et al.  concluded that an association existed between testosterone supplementation therapy (TST) and cardiovascular morbidity in men >60 yr of age. However, on more detailed analysis, the men on TST were found to have mean serum testosterone levels of only approximately 320 ng/dl, with the majority of men with . Thus, if data from the Vigen et al. study are transposed to the quartiles identified in the current study by Yeap et al., a significant proportion of men on TST actually corresponded to the lowest quartile (, as well as to large studies on male veterans that show hypogonadism is an important risk factor for increased mortality  and . Taken together, it can be concluded that hypogonadism alone, independent of TST, remains an independent risk factor for cardiovascular morbidity and mortality in elderly men.
In the TOM trial , a population of older men with limitations in mobility and several comorbidities was assessed. Application of testosterone gel (100 mg/d) was associated with an increased risk for adverse cardiovascular events. In this trial, the mean testosterone level in men on TST was 574 ng/dl . If the post-treatment testosterone level was optimal, or within the middle quartiles, one can only speculate whether mortality would have actually been different.
The question now becomes (while acknowledging the inherent variability in serum testosterone assays), how does one apply these findings to clinical practice? Should clinicians aim for certain serum testosterone values in elderly men on TST? For example, while injectable TST (ie, testosterone cypionate) is a less expensive modality, it yields higher and more variable levels of serum testosterone, often reaching supraphysiologic levels. Would lower, more physiologically stable levels obtained from subcutaneous implantable pellets or gels be more desirable for lower long-term cardiovascular protection? While it is known that testosterone has a dose-dependent stimulatory effect on erythropoiesis that is more pronounced in older men , it is thus theoretically possible that increased blood viscosity could aggravate vascular disease in coronary, cerebrovascular, or peripheral vascular circulation in this elderly population . Given the preponderance of the current evidence, an association between cardiovascular disease, TST, and serum testosterone levels in elderly men with multiple comorbidities exists, but what remains to be determined are the optimal treatment modalities and target testosterone levels in elderly men requiring TST.
Conflicts of interest
L. Lipshultz has been a clinical trials participant, consultant, and speaker for Auxilium and a clinical trials consultant and speaker for Endo. The other authors have nothing to disclose.