Transdermal testosterone administration in women with acquired immunodeficiency syndrome wasting: a pilot study

Although human immunodeficiency virus (HIV) disease is increasing rapidly among women, no prior studies have investigated gender-based therapeutic strategies for the treatment of acquired immunodeficiency syndrome (AIDS) and its complications in this population. Markedly decreased serum androgen levels have been demonstrated in women with AIDS and may be a contributing factor to the wasting syndrome in this population. To assess the effects of androgen replacement therapy in women with AIDS wasting, we conducted a randomized, placebo-controlled, pilot study of transdermal testosterone administration. The primary aim of the study was to determine efficacy in terms of the change in serum testosterone levels, safety parameters and tolerability. A secondary aim of the study was to investigate testosterone effects on weight, body composition, quality of life, and functional indexes. Fifty-three ambulatory women with the AIDS wasting syndrome defined as weight less than 90% of ideal body weight or weight loss of more than 10% of the preillness maximum, free of new opportunistic infection within 6 weeks of study initiation, and with screening serum levels of free testosterone less than the mean of the normal reference range (< 3 pg/mL) were enrolled in the study. Subjects were age 37 +/- 1 yr old (mean +/- SEM), weighed 92 +/- 2% of ideal body weight, and had lost 17 +/- 1% of their maximum weight. CD4 count was 324 +/- 36 cells/mm3, and viral burden was 102,382 +/- 28,580 copies. Subjects were randomized into three treatment groups, in which two placebo patches (PP), one active/one placebo patch (AP group), or two active patches (AA group) were applied twice weekly to the abdomen for 12 weeks. The expected nominal delivery rates of testosterone were 150 and 300 microg/day, respectively, for the AP and AA groups. Forty-five subjects completed the study (PP group, n = 13; AP group, n = 14; AA group, n = 18). Two additional subjects from the PP group and two from the AP group were included in the intent to treat analysis. Serum free testosterone levels increased significantly from 1.2 +/- 0.2 to 5.9 +/- 0.8 pg/mL (AP) and from 1.9 +/- 0.4 to 12.4 +/- 1.6 pg/mL (AA) in response to testosterone administration (P < 0.0001 for comparison of AA vs. PP and AP vs. PP; normal range, 1.3-6.8 pg/mL). Testosterone administration was generally well tolerated locally and systemically, with no adverse trends in hirsutism scores, lipid profiles, or liver function tests. Weight increased significantly in the AP group (1.9 +/- 0.7 kg) vs. the PP group (0.6 +/- 0.8 kg; P = 0.043), but did not increase significantly in the AA group (0.9 +/- 0.4 kg; P = 0.263 vs. PP, by mixed effects model assessing the interaction of time and treatment on all available data, one-tailed test). Improved social functioning (P = 0.024, by one-tailed test) and a trend toward improved pain score (P = 0.059) were observed in the AP vs. the PP-treated patients (RAND 36-Item Health Survey questionnaire). Five of six previously amenorrheic patients in the AP group had spontaneous resumption of menses compared to only one of four amenorrheic patients in the AA group (P = 0.045 for comparison of actual number of periods during the study). This study is the first investigation of testosterone administration in women with AIDS wasting. We demonstrate a novel method to augment testosterone levels in such patients that is safe and well tolerated during short term administration. At the lower of the two doses administered in this study, testosterone therapy was associated with positive trends in weight gain and quality of life. Higher, more supraphysiological, dosing was not associated with positive trends in weight or overall well-being. These data suggest that testosterone administration may improve the status of women with AIDS wasting. Further studies are needed to assess the effects of testosterone on weight in HIV-infected women and to define the optimal therapeutic window for test