Medicare Drug Beneficiaries and Cost-Lowering Strategies

In Reply: Dr Daniell raises possible confounders of the results of our study that demonstrated an inverse relationship between brain levels of testosterone and development of AD. Although we were unable to directly control for obesity, we believe that the contribution of an association with low testosterone to our observations is minimal. Obesity is also associated with elevated estradiol levels and we observed no significant relationship between brain levels of estradiolandneuropathologicalstatus.Wealsodonotthink that expectant care and opioid use as features of the agonal state explain our results. While it is likely that agonal state may significantly differ between individuals with and without AD, our design indirectly controlled for this possibility with the use of a third group, men with mild neuropathological changes who also showed low levels of testosterone. We reason that because death in this group was unrelatedtoneurologicaldisease,agonalstateinthecontrolgroup and the mild neuropathological changes group should be similar. Moreover, our findings are consistent with the recent results of Moffat et al, 1 who reported that serum levels of testosterone are negatively associated with clinical measures of AD. In their longitudinal study, agonal state was not a relevant factor, and the relationship between testosterone and AD was unaffected by variables such as obesity. Dr Bowen and colleagues discuss the possibility that development of AD may be associated with elevated gonadotropins, particularly luteinizing hormone, rather than low testosterone. However, as argued by Hogervorst et al, 2 any relationshipbetweenelevatedserumgonadotropinsandAD would appear to be most relevant in patients older than 80 years, the age when luteinizing hormone normally shows modest but significant elevations in men. 3 In our study, we comparedADneuropathologicalstatusinmenyoungerthan 80 years. We do not think that the actions of elevated luteinizing hormone rather than reduced androgens as an explanation of the finding of significantly increased brain levels of -amyloid following castration of male rodents 4 is a likely possibility in our paradigm. -Amyloid levels in castratedrodentsarereducedbyandrogenbutnotestradioltreatment 4 ; however, estradiol treatment reverses elevations in luteinizing hormone caused by castration of male rodents. 5 Thus, our human data, as well as rodent data, implicate androgens in regulation of AD-related neuropathology. However, we agree with the need to understand the broader role of normal, age-related disruption of the hypothalamicpituitary-gonadal axis in AD pathogenesis.