A CLOSER LOOK AT THE BIOLOGICAL BASIS OF WOMEN’S SEXUAL DESIRE AND AROUSABILITY INCLUDING THE ROLE OF ANDROGENS
The neuroendocrine basis of sexual desire/interest is poorly understood. The effects on sexuality of medications with known or partially known mechanisms of action suggest that more than 30 neurotransmitters, peptides, and hormones are involved in the sexual response. Currently, the most clinically important include noradrenaline, dopamine, oxytocin, and serotonin via 5HT1A and 5HT2C receptors—all considered to be prosexual. Serotonin acting via most 5HT receptor sites, prolactin, and GABA, are considered sexually negative. The role of dopamine has been investigated particularly in rodents. Dopamin- ergic input from the ventral tegmental area, particularly to the nucleus accumbens and forebrain is important for cognitive and reward processes. Dopamine admin- istration into the nuclear accumbens has been found to stimulate the anticipatory phase (or appetitive phase) of a sexual activity (54). The paraventricular nucleus and the medial preoptic area of the hypothalamus regulate the anticipatory/ motivational phases of rat copulation as well as the physiological changes of genital engorgement. Introducing a male hamster increases the dopamine in the nucleus accumbens in the female hamster along with her increased sexual activity. Even in animals, the effects of experience can be seen—there is more dopamine accumulation and for a longer time period in female hamsters that are sexually experienced than in those who are sexually na¨ıve (55). In oophorized female hamsters, progesterone administration after estrogen priming leads to increased numbers of sex hormone receptors in the medial preoptic area. Interest- ingly, dopamine administration has the same effect as does environmental change—namely the presence of a male hamster.
Brain imaging of women during sexual arousal shows activation of areas involved in cognitive appraisal of the stimuli, namely the orbital frontal and anterior cingulate areas, and other areas involved in the emotional response to arousal including the rostral anterior cingulate (56). The latter and the posterior hypothalamus also imaged, are involved in the organization and perception of genital reflexes. Of interest, areas in the basal ganglia and temporal lobes that had shown activity in the nonsexually aroused state are no longer imaged during arousal, suggesting that they are involved in tonic inhibition.
Hormones can be measured during the sexual response, but these findings may reflect the consequence of sexual response rather than cause (e.g., oxytocin increases with arousal and prolactin increases after orgasm).
Estrogen is known to affect mood and sleep and so its central action may indirectly influence sexual response. That postmenopausal estrogen therapy causes improvement in well being, sleep, and vasomotor symptoms, is evidence based (5), but there are few scientific data to suggest that sexual benefit is afforded by relief of these particular symptoms. The role of androgen in women’s sexual desire and arousability is currently under investigation. Although there is consensus that androgens are needed for sexual response, scientific study of androgen therapy with physiological amounts of androgen is only just beginning. It is also unclear whether the aromatization of testosterone to estradiol within the cell is essential, or whether instead or in addition, activation of the androgen receptor is essential. Areas of high density androgen receptors in women’s brains also have high aromatase activity. Thus, the whole question of whether any benefit of testosterone administration to women is actually due to making estrogen more available (by decreasing SHBG) remains unsolved. The major androgens include the proandrogens, dehydroepiandrosterone sulfate (DHEAS), dehydroepiandrosterone (DHEA), androstenedione, plus testosterone (T), and dehydrotestosterone (DHT). Output of adrenal androgen decreases from the early 30s onwards. Ovarian androstenedione is consistently reduced in mid- and later life. Studies are less conclusive regarding ovarian T production after natural menopause, with evidence of both reduced and increased production (58,59). Two recent small studies have shown a gradual decrease of T in women through their 40s with loss of mid-cycle peaks of T and androstenedione (60,61). Studies across the menopause transition show either a minimal decrease or even an increase (62 – 64). Despite further reduction in adrenal androgen, in some women there may be increased production of ovarian T through the next two decades (59,62). On the other hand, some women show very low levels of ovarian production given the T levels in a large group of older women, after natural or surgical menopause were similar. Both of these groups of women were receiving estrogen therapy (58).
Cross-sectional and cohort studies of sexual response and T values are inconclusive. Either there is no correlation between T levels and sexual variables (65) correlation with estradiol levels but not T (63), or a correlation of free-T with levels of sexual desire (66). There have been several short-term randomized con- trolled studies of T administration to women complaining of diminished sexual interest and satisfaction. An improved outcome has been found by most but not all of these trials, but the T levels produced were not clearly within the phys- iological range. The study with levels closest to the physiological (25) was of oophorized women, and showed benefit only in older women receiving 300 mg/day of transdermal T, with corresponding blood levels at or slightly
above the normal range for premenopausal women. Of note, the correct range for postmenopausal women is unclear. A very recent study of T administration to premenopausal women did show benefit over placebo, but the free androgen index was above the upper limit for normal premenopausal women (67). Of major importance is the fact that these studies have been only of short duration, and, therefore, safety data are very limited. Moreover, only estrogen replete women have been studied.
Despite documented progressive loss of DHEA and DHEAS in women from late 30s onwards, the results of DHEA supplementation to improve sexual health have been conflicting (68 – 72).
The term androgen deficiency syndrome has been used recently (73). However, the usual criteria used in endocrinology for establishment of a deficiency state have not been met. These include:
1. Symptoms regularly associated with low levels of the hormone;
2. Relationship of symptoms to the established biological actions of the hormone;
3. Reversal of symptoms on administration of the hormone in doses which are physiological and not pharmacological.
None of these criteria is fully met in the case of “androgen deficiency syndrome” (74). In addition, a specific level of testosterone in women, which can be con- sidered diagnostic of androgen deficiency, has not been established.
Some of this confusion may be in part owing to problems in measuring T, including a lack of assay specificity. Free-T is preferably measured by equili- brium dialysis, but this is rarely available in clinical practice. Free-T correlates more closely with the biological effects of the hormone than does the total because most of the circulating T is bound to SHBG which prevents diffusion into tissues. Unfortunately, the analogue assays for free-T are inaccurate. Free-T can be calculated if the total T, albumin, and SHBG are known. However, at the low levels of T found in women, few assays of total T are reliable. Whichever assay is used, thorough validation is necessary. Another major complicating factor is that much T activity within the cell is derived intracellularly from ovarian adrenal precursors (75). This intracellular T cannot be measured. Estimating T activity from measuring testosterone metabolites is not yet standardized.
There is clearly a clinical dilemma. Clinicians repeatedly see previously responsive women markedly distressed from their lost arousability—none of their formerly useful stimuli are effective. Typically, this is of gradual onset in the late 40s or early 50s. Loss of innate sexual thoughts and fantasies is not the issue. The context of their sexual lives has not changed—they speak of a sexual “deadness”. Accurate measurements of T activity and long-term random- ized controlled trials of physiological T therapy are very much needed. Clearly, this loss of arousability appertains to just a subgroup of mid-life women— perhaps partially explaining the inconsistencies amongst reports of T levels of women in mid-life and older in the general population.
The free-T can be reduced by 50% by many oral contraceptive pills and by administration of glucocorticoids (76). There has been little research in these areas that is helpful to clinicians.
The risks of T administration include those that are familiar, for example, greater sebum production, acne, loss of scalp hair, stimulation of facial and other body hair, as well as other potential risks including metabolic dysfunction in some women. This is based on the fact that although in the condition of polycys- tic ovarian syndrome, it appears that hyperinsulinemia is usually the cause of the hyperandrogenism, there are some reports of situations in which hyperandrogen- ism causes insulin resistance (77). There is also a risk that other concerns will come to light if women are given testosterone when estrogen deficient, in view of the recent withdrawal of large numbers of women from estrogen therapy owing to the results of the women’s health initiative study (78).