Should women be treated with testosterone?
H. Burger,S. Davis
DOI: https://doi.org/10.1046/j.1365-2265.1998.00524.x
1998-08-01
Clinical Endocrinology
Abstract:The use of testosterone therapeutically in women is a controversial issue in contemporary endocrinology and gynaecology. Perhaps least debated is its administration to postmenopausal women, particularly those who have undergone bilateral oophorectomy, and who complain of loss of libido and impaired sexual function, despite adequate oestrogen replacement (Burgeret al., 1984, 1987; Sherwinet al., 1985; Davis et al., 1995; Davis & Burger, 1996). In these circumstances, combined oestrogentestosterone replacement results in substantial benefit, without significant adverse effects. The rationale for such treatment is based on one of the known consequences of bilateral oophorectomy, a 50% fall in circulating testosterone concentrations. Testosterone supplementation, particularly using a 50 mg implant of crystalline testosterone approximately every 6 months in combination with oestradiol 50 mg causes no adverse changes in circulating lipids. In fact the favourable effects of postmenopausal oestrogen replacement on total and LDL-cholesterol are not diminished by the addition of either oral or parenteral testosterone replacement (Watts et al., 1995; Daviset al., 1995). Even for this indication, however, testosterone therapy for women is not in widespread use: it is not officially approved of by most regulatory authorities. The dose of testosterone required for therapeutic efficacy in postmenopausal women has not been established rigorously, though published experience indicates that 50 mg every 6 months provides satisfactory benefit (Burgeret al., 1987). At this dose virilizing side-effects have not been encountered and it may be that 100 mg or more is in excess of therapeutic requirements. Administration of androgens to women in their reproductive years, with intact ovarian function, has been much less studied. Hence the paper by Bucklert al. (1998) in this issue of Clinical Endocrinology addressing the role of androgen replacement in the management of PMS tackles a very common and important women’s health issue. However, the design of the study and the conclusions reached merit further discussion. The study was small, 22 women in each group, and no data are given concerning the definition of premenstrual syndrome (PMS) nor how severity was ascertained. However, the major methodological flaw is that no baseline data are provided for any parameter. Therefore, not only do we not know whether the two groups were comparable pretherapy, we also do not know the effect of treatment from baseline within the group who received testosterone. There are no placebo treated controls. No indication is given regarding the number of women who discontinued testosterone therapy which had been initiated for PMS, and it is for the reader to assume that those who persisted with treatment for an average of 3·3 years experienced some benefit. Since we therefore have no evidence that the two groups of women are actually comparable it is impossible to be confident of the conclusions drawn. This is further underlined by the probability of selection bias and the possibility of a Type II error. The authors describe their testosterone treatment as ‘low dose’ though this description can be argued. In fact, reported mean testosterone levels were more than twice the upper limit of normal, at 4·5 nmol/l compared with 1·0 nmol in the control group. Clinically, it is somewhat surprising that none of the women complained of masculinizing side-effects but we are given no indication that such information was formally sought or possible changes evaluated. The only adverse metabolic effects were a 16% decrease in apolipoprotein A1 levels and a 13% decrease in high density lipoprotein cholesterol levels in comparison with controls. There was a doubling of very low lipoprotein cholesterol levels in the treated patients. There were, however, no changes of significance in other lipid markers nor markers of activation of the coagulation cascade. The significance of the lipid changes is difficult to assess critically. This group of women received treatment over an average three year period but it is unlikely that they would be treated for more than a total of 5–10 years given that their average age at study was already 39·6 years. Hence, whether these adverse lipid changes would really be of overall cardiovascular significance is uncertain but perhaps not great. A lower dose of testosterone could be expected to have a less adverse effect and might be therapeutically equivalent. The therapeutic use of testosterone in PMS has not been advocated widely and is to some degree a novel concept. It would be of great interest to establish whether it has a scientific rationale. Relevant information to take into account in this context is the study by Zumoff et al. (1995) which showed that 24 h mean plasma testosterone concentrations declined by 50% when women in their early 40 s were compared with 20 year olds. It is widespread clinical experience that the prevalence Commentary