Division of Endocrinology, San Francisco General Hospital, and Department of Medicine University of CaliforniaSan Francisco, San Francisco, California 94110
Address correspondence to: Joan C. Lo, M.D., Division of Endocrinology, San Francisco General Hospital, Building 30, Room 3501-K, San Francisco, California 94110.
Since the advent of highly active antiretroviral therapy (HAART), the resultant decline in mortality rate associated with human immunodeficiency virus (HIV) infection (1) has allowed patients and clinicians to consider other important health issues beyond short-term survival. Included among these issues is the maintenance of normal gonadal function, as well as ethical and practical considerations for those individuals who wish to conceive and have children. Here, we review the pathophysiology and treatment of abnormalities in gonadal and sexual function and consider some of the challenges surrounding reproductive health among HIV-infected men and women.
Reproductive health in HIV-infected men
Testicular function and sex hormones. Autopsy studies in men with the acquired immune deficiency syndrome (AIDS) reveal a number of histopathologic changes in the testes, including hypospermatogenesis, spermatogenic arrest, and "Sertoli-only" findings, as well as tubular basement membrane thickening and atrophy and peritubular fibrosis (2, 3, 4, 5). On average, approximately 30% of testicular germ cells are HIV infected (4). Within the last decade, a more pronounced loss of germ cells has been evident on histologic examination and is likely due to several factors, such as duration of HIV infection, direct HIV cytopathic effects, the use of gonadotoxic or antiandrogenic drugs, malnutrition, wasting, fever, and other systemic illnesses or chronic infections (5). Direct testicular involvement by opportunistic infections has also been reported in men with AIDS, with cytomegalovirus, Toxoplasma gondii, and Mycobacterium avium intracellulare, the most common pathogens (2, 3, 5). In addition, neoplastic infiltration of the testes by Kaposis sarcoma and lymphoma has been noted in patients with evidence of disseminated disease at autopsy (5, 6).
The extent to which specific effects on gonadal endocrine function can be attributed to HIV per se, associated opportunistic infections, or neoplasms is not completely understood, but certain general patterns have been observed. Early in the course of HIV infection, total and free testosterone levels have been reported as normal or, in some cases, elevated (7), accompanied by an exaggerated response of LH to infusion of GnRH (8). Thereafter, testosterone levels tend to fall with progression of HIV disease, as a consequence of both gonadal and extragonadal factors that contribute to testicular dysfunction (7, 9). At that point, measurement of free or bioavailable testosterone may be the most sensitive indicator of hypogonadism, because increased sex hormone-binding globulin levels have been observed in HIV infection (9, 10, 11). Although primary hypogonadism and testicular atrophy have been reported in men with AIDS (12), hypogonadotropic (or secondary) hypogonadism occurs more commonly (6, 13). For the majority of cases, the etiology is multifactorial and may be attributable to systemic illness, opportunistic infections, malnutrition, cachexia and weight loss, and HIV-associated cytokines (6, 13, 14, 15). A significant number of patients with the AIDS wasting syndrome have some degree of androgen deficiency, and it has been hypothesized that the decline in circulating testosterone levels may contribute to the critical loss of lean body tissue and muscle mass in these patients (11, 16). Furthermore, pharmacologic doses of systemic glucocorticoids or megestrol acetate, the latter used as an appetite stimulant in the treatment of AIDS wasting, are known to suppress the hypothalamic-pituitary-gonadal axis (17, 18). Ketoconazole, particularly at higher doses, inhibits testosterone biosynthesis (19). Chronic use of alcohol, opiates, and marijuana has also been associated with diminished testosterone production (20). Direct pituitary or hypothalamic destruction by secondary opportunistic infection (e.g. cytomegalovirus and toxoplasmosis) is rare (21, 22).
Gynecomastia in HIV-infected men is typically associated with low testosterone and/or elevated estrogen levels, liver disease, and the use of alcohol, marijuana, and other recreational drugs. Medications such as ketoconazole, cimetidine, steroid hormones, and hormone antagonists have also been implicated. More recently, there have been reports of gynecomastia occurring in men treated with HAART, including a protease inhibitor, in the absence of any other apparent cause (23, 24, 25). Although HIV protease inhibitors seem to inhibit cytochrome P450 3A4-mediated metabolism of testosterone in human liver microsomes in vitro (26), effects on testosterone metabolism in vivo have not been observed. Typically, gynecomastia does not resolve after cessation of protease inhibitor therapy, and the role of these drugs, as well as other antiretroviral agents (27), in promoting breast enlargement has not been elucidated. Gynecomastia has also been reported in the setting of peripheral fat loss and central fat accumulation and may, indeed, be another facet of the HIV-related fat redistribution syndrome (28).
Treatment considerations. Treatment of hypogonadism associated with HIV infection depends on an understanding of the underlying etiology, the severity of symptoms, and the specific purpose of therapy (29). For patients who are symptomatic due to primary or secondary hypogonadism, treatment with replacement doses of testosterone is warranted, after excluding reversible etiologies. There are three main routes of testosterone delivery. The testosterone esters (e.g. enanthate and cypionate) are administered by im injection every 23 weeks and provide a relatively safe, effective, and inexpensive approach to androgen replacement therapy. However, because intermittent injections can lead to large fluctuations in circulating testosterone levels (30), the newer transdermal delivery systems have been developed to provide a more stable, continuous mode of testosterone replacement. Many patients prefer the testosterone patch to im injections, although they are more costly, require daily application, and may be associated with skin irritation. Up to 30% of men develop a local rash with the Androderm patch; in these cases, pretreatment with triamcinolone cream under the patch is generally effective in preventing the rash and does not interfere with drug absorption. A new gel formulation (Androgel) has also been shown to maintain serum testosterone levels within the normal range without significant adverse effects (31, 32), but patients should be aware that vigorous skin contact (e.g. with a female partner) may lead to significant drug transfer, as a large percentage of the dose remains on the skin after drying (31). In addition, the testosterone gel and patch may not deliver an adequate testosterone level at the recommended doses in every patient, so that measurement to determine whether therapeutic testosterone concentrations have been achieved should be considered.
Testosterone replacement in hypogonadal men without HIV infection has been associated with improvement in body composition, bone density, sexual functioning, quality of life, cognitive function, and moodpotential benefits that are likely to be evident in the setting of HIV infection as well (29). In hypogonadal men with HIV-associated weight loss, physiologic testosterone administration alone or in combination with resistance training has been shown to increase lean body mass, muscle strength, and quality of life (33, 34, 35). Despite an increase in muscle mass, the role of higher dose testosterone therapy in eugonadal men with HIV-associated weight loss is less clear, given the potential adverse effects on high-density lipoprotein cholesterol and lack of long-term safety data (36). Substantial increases in lean tissue accrual and strength gains from resistance exercise training have also been reported with supraphysiologic androgen therapy including an anabolic steroid, but with significant reduction in high-density lipoprotein cholesterol levels (37). The impact of these anabolic therapies on future reproductive function and fertility has not been systematically evaluated.
Whereas diminished libido and impotence in HIV-infected men have generally been associated with low testosterone levels, sexual dysfunction may also be related to other factors, including neurological disorders, systemic disease, medication effects, weakness, fatigue, and psychosexual issues (38, 39, 40, 41). The overall prevalence of erectile and ejaculatory dysfunction has been estimated to be in the range of 60% among men with advanced HIV disease (40). Sexual dysfunction has also been reported in men receiving protease inhibitor therapy, in the absence of any clear identifying cause; these preliminary reports suggest a possible association between protease inhibitors and the development of sexual dysfunction, the mechanism of which remains unclear (42, 43). Many patients have benefited from the use of sildenafil citrate (Viagra) to improve erectile function, although clinicians should be aware that sildenafil is also metabolized by cytochrome P450 3A4, the microsomal enzyme involved in the hepatic metabolism of protease inhibitors such as indinavir, ritonavir, saquinavir, and nelfinavir. Protease inhibitor pharmacokinetics are not significantly altered during sildenafil coadministration, but sildenafil concentrations can be substantially elevated (44, 45), increasing the risk of sildenafil-related adverse events. Hence, only the lowest starting dose of sildenafil (25 mg) should be used, and with particular caution, in patients receiving these protease inhibitors concomitantly (45).
Fertility and reproductive issues. In early HIV disease, semen parameters (i.e. sperm count and morphology) are generally normal and consistent with fertility, whereas untreated men with advanced stages of HIV disease have been reported to have both a reduced sperm count and an increased percentage of abnormal sperm forms (46, 47). HIV-infected men tend to produce more viscous semen containing fewer motile sperm and round cells, compared with healthy seronegative controls (48). These early studies also showed that administration of zidovudine had no deleterious effect on sperm production or other semen parameters (46, 47, 48).
The main ethical considerations surrounding reproduction for men with HIV infection have focused on risks to the future child, as well as to the female partner if she is HIV negative. Even in patients with undetectable levels of plasma HIV RNA, the virus remains detectable in semen (49). For those HIV-discordant couples who desire children, semen washing, followed by artificial insemination, has been investigated as a potentially safer method of reproduction, based on evidence that the primary reservoir for HIV in semen is the seminal plasma and nongerminal cells rather than the spermatozoa (50). Preliminary studies suggest that sperm washing can yield undetectable levels of HIV RNA in semen and substantially reduce HIV transmission rates compared with natural conception (51, 52, 53, 54). However, because the procedure is costly and the risk of infection is not completely eliminated, this approach has not been widely accepted. In vitro fertilization using intracytoplasmic sperm injection has been proposed as an alternative method in cases in which semen quantity is insufficient or there is a lack of fallopian tube patency (55); this method might further reduce the risk of HIV transmission, although larger trials are needed.
Reproductive health in HIV-infected women
Ovarian function and sex hormones. Because the ovaries have not been systematically examined during autopsy surveys in women with AIDS, remarkably little is known regarding ovarian pathology in HIV infection. Presumably, the ovaries are susceptible to HIV and secondary opportunistic infections, and, indeed, there is a single case of cytomegalovirus oophoritis reported in the literature, occurring in a woman with disseminated cytomegalovirus infection (56). It has also been shown that HIV can infect cells and tissues from the female reproductive tract, including the fallopian tubes, uterus, and cervix (57). Shedding of HIV-infected cells occurs in both the endocervix and vagina throughout the menstrual cycle (58) and seems to be increased by hormone contraceptive use, vitamin A deficiency, and gonorrheal or candidal infection (59).
Unlike the direct or associated pathologic effects of HIV infection on ovaries, the abnormalities in menstrual cycle and reproductive function have been better characterized in the last decade. Increased rates of oligoamenorrhea and amenorrhea have been reported in HIV-infected women without AIDS-defining illnesses (60), as well as in women with the AIDS wasting syndrome compared with those women with AIDS who have stable weight or manifest only mild weight loss (61). These latter findings are not unexpected for women with a severe catabolic illness complicating their HIV infection. In some women, changes in menstrual pattern have been associated with a past history of substance abuse, suggesting that socioeconomic factors may be involved (60). In addition, narcotics, marijuana, and chronic alcohol consumption are known to affect menstrual function and ovulation (20). Other surveys indicate that infection with HIV does not seem to have clinically significant effects on menstruation or vaginal bleeding rates (62, 63). Recent prospective data from two large cohorts demonstrate that HIV serostatus has little overall effect on amenorrhea, menstrual cycle length, or variability; however, among HIV-infected women, higher viral loads and lower CD4 cell counts were associated with increased cycle variability and polymenorrhea (64). In women with self-reported regular menstrual cycles, normal levels of estradiol and progesterone have been observed during the follicular and luteal phases (65).
Since the advent of HAART, specific body composition changes have been reported in women, primarily, but not exclusively, in the presence of protease inhibitor therapy. These include breast enlargement, abdominal obesity, and wasting of fat in the lower extremities and gluteal region, but without overt endocrine perturbations (66, 67, 68). Interestingly, in one patient, this redistribution of body fat was also associated with coexisting evidence of the polycystic ovary syndrome, including hirsutism, polycystic ovaries, and an increased LH to FSH ratio (69). However, the majority of women experiencing HIV-associated fat redistribution do not have typical features of hyperandrogenism. A recent pilot study in nine HIV-infected women with the fat redistribution syndrome, five of whom were eumenorrheic, found that androgen levels were increased compared with HIV-infected women without fat redistribution and healthy seronegative controls (70). These findings were accompanied by higher fasting insulin levels, dyslipidemia, and an increased LH to FSH ratio. Although the biochemical features in this small cohort are reminiscent of the polycystic ovary syndrome in seronegative women, additional studies are needed in a larger population to confirm these results and to explore potential pathophysiologic mechanisms.
Treatment with protease inhibitors has also been associated with the development of hypermenorrhea and anemia in four women, which was severe enough in two cases to require blood transfusions (71). No abnormalities were found on pelvic examination, platelet counts were normal, and the women did not report intermenstrual bleeding. One patient also experienced hemoptysis, and the possibility of decreased hemostatic function was raised. In a separate series of four patients, initiation of protease inhibitors (for HAART or post-HIV exposure prophylaxis) was associated with galactorrhea (72). In three of these cases, marked hyperprolactinemia was reported; of these, two patients had received metoclopramide and a third patient, fluoxetine, both of which may increase serum PRL levels. However, symptom resolution was only evident after the protease inhibitors were discontinued. These cases may reflect a direct effect of protease inhibitors, or rather, an indirect effect through potentiation of the dopamine antagonist effect of other drugs (72).
Treatment considerations. Because there have been few studies specific to women with HIV infection, the treatment considerations for affected women with ovarian dysfunction or failure are generally similar to that for immunocompetent individuals. Remarkably, little is known about therapies to restore hormonal balance in HIV-infected women for the purpose of improving reproductive potential. Recommendations for contraception include a combination of a barrier method and another form of contraception suitable to the woman, for maximum safety (73). Treatment with the combined oral contraceptive pill may increase the risk of HIV shedding in the lower genital tract (59), and there remains the possibility that contraceptive efficacy is affected by antiretroviral drugs and other medications (73). For postmenopausal women with symptoms of estrogen deficiency, hormone replacement therapy should be considered (74). It may also protect against the bone mineral loss associated with estrogen deficiency. The cardioprotective effects of hormone replacement are more controversial, based on current data showing no benefit of estrogen plus progestin in seronegative postmenopausal women with established coronary heart disease (75).
Given recent observations that circulating testosterone levels are lower in HIV-infected women compared with healthy seronegative women (76), increasing attention has been directed toward the role of androgen replacement therapy in these individuals, including androgenic anabolic steroids. Estratest provides combined estrogen-androgen replacement but contains methyl testosterone, which may contribute to hepatic dysfunction; it should also be noted that concomitant treatment with a progestin is necessary for women with an intact uterus to prevent endometrial hyperplasia. Preliminary studies indicate that low-dose twice-weekly transdermal testosterone administration in HIV-infected women is well tolerated (77, 78); furthermore, in those with the AIDS wasting syndrome, physiologic testosterone treatment was associated with positive trends in weight gain and quality of life (77). Additional studies are needed to determine the long-term consequences of physiologic androgen replacement on ovarian physiology, sexual function, body composition, functional status, quality of life, and other health-related outcomes.
Fertility and reproductive issues. Retrospective studies demonstrate that pregnancy and birth rates among women with HIV infection are lower than the general population and the rates of therapeutic abortion are higher, outcomes which, in part, are related to knowledge of HIV antibody status (79, 80, 81). A history of iv drug use and other sociodemographic factors may contribute to increased rates of pregnancy termination (80). Large cohort studies indicate that sexual risk behavior and unplanned pregnancies remain prevalent among HIV-infected women, emphasizing the need for further implementation of safer sex practices and contraceptive counseling (79, 81, 82). Because discovery of HIV diagnosis is generally followed by a decrease in sexual activity (81), it is difficult to predict the direct biologic impact of HIV infection on fertility in women. Certainly, other mechanisms might also impact reproductive potential, including weight loss, systemic illness, drug abuse, and sexually transmitted diseases of the reproductive tract (20, 83). It has been suggested that rates of spontaneous fetal loss are increased in HIV-infected women and related to HIV transmission and fetal thymic dysfunction (84, 85), but these observations were made in women who did not receive antiretroviral therapy during pregnancy and may not reflect trends in the current era. Overall, among HIV-infected women with a low prevalence of AIDS, maternal seropositivity seems to have little demonstrable impact on the health status at birth of live newborns (86).
Increased routine screening of all pregnant women has identified a substantial number of HIV-infected women early in pregnancy, enabling them to make timely reproductive decisions and seek appropriate health care. The use of zidovudine during pregnancy reduces the risk of vertical transmission of HIV from mother to infant by nearly 70% (87). In addition, effective maternal viral load suppression with aggressive antiretroviral therapy seems to further minimize this risk, as may other adjunctive measures such as elective cesarean section and avoidance of breast feeding (88, 89, 90, 91, 92). While the issues concerning conception in women with HIV infection are becoming increasingly relevant in the current era of effective antiretroviral therapy, they remain complex and involve a range of ethical and sociocultural issues (93, 94, 95). For example, in a recent survey of selected fertility centers in the United Kingdom, only a few centers offered intrauterine insemination or in vitro fertilization in cases in which the female partner was found to be seropositive (96).
For women with HIV infection who do become pregnant, several large cohort studies conducted in industrialized countries have shown that there are no apparent deleterious effects of pregnancy on HIV disease progression or associated immunologic parameters (97, 98, 99). In general, antiretroviral treatment recommendations for pregnant women have been based on the premise that therapies of known benefit to women should not be withheld during pregnancy unless there are known risks that outweigh the benefit (91). Important drug considerations unique to pregnancy include the pharmacokinetic changes that occur during gestation, the effectiveness in reducing perinatal HIV transmission risk, and the potential for adverse effects on the fetus and newborn (91, 100). Careful evaluation of patients during gestation and close coordination of care between the HIV clinician and obstetrician are recommended. Currently, experience with combination antiretroviral drug treatment in pregnancy remains limited and further observational studies are needed to determine overall safety (101). Very preliminary data suggest that adverse effects related to protease inhibitors are relatively uncommon during gestation (102), although women may be at higher risk for glucose intolerance. In addition, recent cases of fatal lactic acidosis in pregnant women receiving didanosine and stavudine in combination have led to heightened concern regarding the risk for lactic acidosis in patients treated with this drug combination, as well as nucleoside reverse transcriptase inhibitors in general. Additional studies are necessary to determine whether there are long-term adverse effects of perinatal antiretroviral drug exposure on the offspring, particularly in light of the possible association between perinatal nucleoside analog therapy and mitochondrial dysfunction in children exposed in utero and after birth (103, 104).
Summary
HIV infection is associated with unique and varied abnormalities in gonadal function in affected men and women. There remains a need for greater insight into the direct and indirect effects of HIV infection and its therapies on reproductive health. Future research should identify effective therapeutic strategies to prevent and treat gonadal dysfunction, as well as to minimize the risk of HIV transmission associated with reproduction. Finally, the existing socioethical challenges surrounding reproductive considerations in patients with HIV infection have yet to be resolved.
Footnotes
1 Supported in part by the NIH (Grants DK45833 and RR-00083). J.C.L.
is a recipient of a Clinical Associate Physician Award from the
National Center for Research Resources.
Received February 6, 2001.
Accepted March 18, 2001.
References