What Are "Normal" Testosterone Levels for Women?

Susan Davis and Jane Tran

The Jean Hailes Foundation Clayton, Victoria 3168, Australia

To the editor:

The cross-sectional study by Laughlin et al. (1), comparing endogenous sex hormones levels with hysterectomy and oophorectomy status, chronological age, and years since menopause in older women, reinforces the possible adverse endocrine sequelae of oophorectomy in older women. However, we question some aspects of the data analysis and, hence, the clinical significance and interpretation of some of the said findings.

The inference from the manuscript is that, in women with intact ovaries, there is an increase in total testosterone levels after the time of menopause and with increasing age reaching premenopausal levels concomitant with falling androstenedione levels. That testosterone should rise while androstenedione falls is incongruous because adrenal androgen precursor levels fall linearly with age (2) and, following the menopause, peripheral conversion of androstenedione becomes a major source of circulating testosterone (3). We would appreciate the authors proposing a hypothesis for this incongruity.

The reason for adjusting total testosterone and bioavailable testosterone for body mass index (BMI) in postmenopausal women is not justified. Because weight may vary significantly with increasing age and years since menopause, and the authors suggest that sex hormone-binding globulin concentrations were inversely related to BMI, one is left to wonder if without adjustment for BMI whether any variation occurred. The authors do not report a relationship between either testosterone or androstenedione and BMI, making the need for the adjustment most curious. Adjustment for BMI is not clinically informative, because the absolute bioavailable circulating levels are the values of physiological significance in terms of direct androgen action and as precursors for extragonadal estrogen biosynthesis in tissues such as bone. Indeed, was a difference seen for bioavailable testosterone not adjusted for BMI? Non sex hormone-binding globulin-bound testosterone includes albumin-bound testosterone, which may account for up to 20% of total testosterone (4). In view of the age of the subjects, if one is to adjust for BMI, surely one should also adjust for variations in serum albumin. Caution in adjusting for covariates in reporting clinical findings has recently been highlighted (5). Presentation of the unadjusted data from this study would be very informative.

The authors report an assay sensitivity of 0.07 nmol/L for testosterone, however, standard RIAs for testosterone are notoriously inaccurate at the lower end of the female range. We are surprised that such distinct differences were observable in values below 0.5 nmol/L. Furthermore, the actual RIA used for measuring total testosterone is not stated. The most pronounced difference in total testosterone increase was reported to occur between the 6th and 7th decades, yet for the 6th decade there were only 29 women with intact ovaries and seven after oophorectomy. Hence, we caution a 35% increase on a mean value of less than 0.5 nmol/L in so few women becoming a generalizable fact. Furthermore, reporting of values adjusted for BMI gives us little feel for what the actual normal values were, and this is exacerbated in Table 2 in which values were adjusted for both age and BMI.

Although ovarian stromal hypertrophy and hyperplasia may sometimes persist or develop after menopause, probably secondary to elevated LH levels and individual sensitivity, resulting in increased testosterone production (6), other researchers have not found this to be the general rule (7). That "... levels in women more than 70 yr of age or 20 yr post menopause were comparable to premenopausal levels" is also questionable. In Fig. 3 in the article, the dotted lines indicate the mean testosterone level for premenopausal women as being between 0.6–0.7 nmol/L. However, according to Sinha-Hikim et al. (8), who defined the range of total and free testosterone levels during the normal menstrual cycle in healthy women, the mean testosterone level across the cycle was 1.20 ± 0.69 nmol/L and the mean free testosterone was 12.8 ± 5.59 pmol/L. These levels were measured using an equilibrium dialysis method, and the sensitivity of the total testosterone assay was increased to 0.008 nmol/L. The results from that study are sound and have been used as references in many other studies. Furthermore, these testosterone levels are in agreement with those reported in other literature (9, 10).

If we accept the data of Sinha-Hikim et al. (8) as reference levels, then the testosterone levels in the "intact postmenopausal women" in the study by Laughlin et al. (1) are, in fact, very low and do not actually increase back to premenopausal levels (1.20 ± 0.69 nmol/L). Although in Results the authors state (referencing a 1997 paper) these premenopausal values as being from the same laboratory as used for this study, they are inconsistent with the widely accepted range. Were these also BMI adjusted? Either way, the comparison is not meaningful. Of note, the actual citation for the testosterone levels for premenopausal women in Fig. 3 is that of an article by S. E. Bulun and E. R. Simpson, a study of levels of aromatase cytochrome P450 transcripts in adipose tissue of women. The bioavailable values reported are ~10-fold greater than the normal range for free testosterone. Reporting of free testosterone would have been far more meaningful.

The role of androgens in women is becoming increasingly more recognized and established. Certainly, the use of androgens, particularly testosterone, has been shown to influence life aspects, such as mood, women’s general well being and restoration of sexual desire. However, there is limited data establishing normal androgen values for women of differing ages, to enable us to define those with "androgen deficiency." It is, therefore, necessary to highlight the incongruencies and short-comings of the paper by Laughlin et al. (1), and the need for larger prospective studies to establish the variations in testosterone levels in women with age.

References

  1. Laughlin G, Barrett-Connor E, Kritz-Silverstein D, Von Muhlen D. 2000 Hysterectomy, oophorectomy, and endogenous sex hormone levels in older women: the Rancho Bernado Study. J Clin Endocrinol Metab. 85:645–651.[Abstract/Free Full Text]
  2. Zumoff B, Rosenfeld RS, Strain GW. 1980 Sex differences in the 24 hour mean plasma concentrations of dehydroisoandrosterone (DHA) and dehydroisoandrosterone sulfate (DHAS) and the DHA to DHAS ratio in normal adults. J Clin Endocrinol Metab. 51:330–334.[Abstract]
  3. Judd HL, Lucas WE, Yen SSC. 1994 Effect of oophorectomy on circulating testosterone and androstenedione levels in patients with endometrial cancer. Am J Obstet Gynecol. 118:793–798.
  4. Luthold WW, et al. 1993 Serum testosterone fractions in women: normal and abnormal clinical states. Metabolism. 42:638–643.[Medline]
  5. Assmann S, Pocock S, Enos L, Kasten L. 2000 Subgroup analysis and other(mis)uses of baseline data in clinical trials. Lancet. 355:1064–1069.[CrossRef][Medline]
  6. Procope B. 1968 Studies on the urinary excretion, biological effects and origin of estrogens in postmenopausal women. Acta Endocrinol (Copenh). 135:1–86.
  7. Labrie F, Belanger A, Cusan L, Gomez J-L, Candas B. 1997 Marked decline in serum concentrations of adrenal C19 sex steroid precursors and conjugated andrgoen metaboliltes during aging. J Clin Endocrinol Metab. 82:2396–2402.[Abstract/Free Full Text]
  8. Sinha-Hikim I, et al. 1998 The use of a sensitive equilibrium dialysis method for the measurement of free testosterone levels in healthy, cycling women and in human immunodeficiency virus-infected women. J Clin Endocrinol Metab. 83:1312–1318.[Abstract/Free Full Text]
  9. Judd HL, Yen SSC. 1973 Serum and androstenedione and testosterone levels during the menstrual cycle. J Clin Endocrinol Metab. 36:475–481.[Medline]
  10. Abraham GE. 1974 Ovarian and adrenal contribution to peripheral androgens during the menstrual cycle, J Clin Endocrinol Metab. 39:340–346.[Medline]