Heterosexual transmission accounts for the fastest growing risk group in the United States. AIDS cases attributed to heterosexual transmission increased from 3% of all cumulative AIDS cases among women in 1983 and 1984 to 39% as of December 31, 2002.(6) Regions with the highest rates in women correspond to those in which seroprevalence is highest in male IDUs.(7)

The true incidence of heterosexually transmitted cases is probably even higher, but cases are sometimes unrecognized owing to the U.S. Centers for Disease Control and Prevention (CDC) reporting hierarchy. For example, if a woman had heterosexual contact multiple times and injected intravenous drugs once, she is categorized as having injection drug use as the source of her infection. Of the 126,880 cases ever classified as "risk not reported or identified" through June of 2000, (the last year for reporting on this type of information), 59,692 have been reclassified. Of this group, 15,300 are women, and 10,404 (68%) of them have been reclassified as transmission through heterosexual contact.(8) As recently as 2002, 7,532 (66%) of the total 11,279 adult women diagnosed with AIDS were classified as either "risk not reported or identified" (4,029 [36%]) or as "transmission secondary to heterosexual contact with a HIV-infected person, risk not specified" (3,502 [31%]).(6)

It is estimated that if information were to be obtained on this latter group of 3,502 that at least 68% would be reclassified as heterosexual transmission. These statistics have demonstrated similar trends in the last decade of the epidemic and identify a fundamental problem: women often have limited knowledge about their partners' risk factors for HIV/AIDS and often have the misperception that they are at low risk of contracting AIDS. It remains disturbingly common at San Francisco General Hospital for a woman to present with a new diagnosis of Pneumocystis pneumonia, AIDS, <100 CD4 cells, and no knowledge that she had ever been exposed to HIV.

In industrialized countries, the IDU epidemic in men is reflected in the heterosexual epidemic in women, which is in turn reflected in the epidemic among children. The most alarming trends of heterosexual transmission are now occurring in young women.

Cumulative data of AIDS cases in the United States as of December 2001 illustrate the preponderance of heterosexual sex as the mode of transmission for young women ages 13-25. In this age group, heterosexual contact dominates as a means of transmission for women, whereas it is a relatively uncommon mode of transmission in young men of this age in the United States.(9) Most of these young women have not reported intravenous drug use and are likely to have partners an average of 5 years older than themselves. It is well documented that women frequently have trouble negotiating with their male sexual partners for condom use, and one could speculate that the age gap between younger women and their sexual partners is likely to worsen this trend.(10-12) Finally, normal physiology in young women includes cervical ectopy, where the columnar cells of the endocervix are more exposed and allow more efficient transmission of HIV. Younger populations of women, particularly women of color, require urgent outreach, and intervention measures.(13-15)

Among HIV-discordant couples (those with only a single HIV-positive partner), heterosexual transmission from the male to the female partner is approximately eight times more efficient than female-to-male transmission. Most infections have occurred by the vaginal route, although participation in anal sex increases the risk.(16-18) The cumulative incidence of transmission within discordant couples suggests an approximate 20% risk of transmission from male to female from unprotected sex over a sustained period in a fixed partnership. A study from northern California, which began prospectively following patients in 1985, found that only 2 (2.4%) of the 82 male partners of HIV-infected females became infected, but 68 (19%) of the 360 female partners of infected men became infected.(16,17) Male-to-female transmission risk per sexual contact in this study was estimated to be 0.0009.(16)

In a more recent study from Rakai, Uganda, monogamous serodiscordant couples were followed prospectively. In this cohort of 97 seropositive men, 17 (17.5%) of their wives became infected with HIV. Among the 77 seropositive women, 21 (27.3%) of their husbands became infected.(19) Many factors may explain this difference in female-to-male transmission risk compared to U.S. studies, including HIV viral load,(20) concurrent genital ulcer disease, and HIV subtype. (See section below: "Risk Factors Associated with Heterosexual Transmission.")

Factors associated with increased transmission include other sexually transmitted diseases (STDs), lack of condom use, advanced disease state (measured by CD4, viral load, or AIDS diagnosis), anal intercourse, number of sexual contacts, genital ulcerative disease, and use of an intrauterine contraceptive device (IUD) (Table 1).(16,21,22,23) Cervical ectopy is emerging as a risk in women for acquisition of HIV from infected men. It is possible that the increased levels of heterosexual transmission in adolescent women may be partially explained by the natural occurrence of cervical ectopy in this age group.(24-26)

Increased HIV viral load in the infected partner is associated with increased risk of transmission in heterosexual couples. In the Rakai study, each 10-fold increase in viral load was associated with an increase by a factor of 2.45 (95% confidence interval [CI]: 1.85, 3.26) in the risk of transmission, and no cases of HIV transmission were reported within serodiscordant couples in which the seropositive partner had an HIV viral load of <1,500 copies/mL.(20) It should be kept in mind that this study was conducted in a population without access to antiretroviral drugs. Infectious virus in genital secretions may persist despite undetectable serum viral load in individuals taking antiretroviral therapy,(27) and resistance mutations not seen in plasma virus may occur in the genital compartment.(28) Accordingly, for a given plasma viral load, the risk of transmission of HIV from an individual on antiretroviral treatment may be different from the risk observed in the Rakai study, and the resulting infection may be more difficult to treat because of preexisting drug resistance.

Sexually transmitted infections, particularly those associated with symptoms or gential ulcerations, appear to increase the risk of HIV transmission. A study of heterosexual couples in Haiti found an increased risk of HIV acquisition when genital ulcers were reported in the initially seronegative partner.(29) The Rakai study also found an increased rate of HIV transmission within couples with a seropositive partner who experienced genital discharge or dysuria. In a study in Kenya, higher levels of cervical herpes simplex virus (HSV) type 2 were associated with higher levels of HIV and with more frequent detection of HIV-infected cells in cervical secretions.(30)

HIV subtype C and circulating recombinant form (CRF) 01 (A/E) (formerly called subtype E) have been found to replicate more efficiently than subtype B in Langerhans cells, the antigen-presenting cells hypothesized to be responsible for vaginal transmission. In Thailand, where CRF01(A/E ) predominates, and in India and southern Africa, where subtype C is common, heterosexual transmission is the major route of HIV infection, whereas in North America, where subtype B predominates, heterosexual transmission is less common. It has been speculated that the introduction of CRF01(A/E) or subtype C into North America might result in increased heterosexual transmission.(31,32)

Lack of circumcision has been associated with increased risk of HIV acquisition by men in Africa,(20) and Kreiss et al found that uncircumcised homosexual men in the United States had a twofold increased risk of HIV infection.(33)

Transmission of HIV between women was reported as early as 1984,(34) but documented cases are very rare. A number of small studies have been conducted but tend to include only small numbers of women having sex with women (WSW) who have no other HIV risk factors, such as injection drug use. Through September 1989, 79 women with AIDS had reported sex with a female only, but 95% were IDUs.(35) Sporadic cases of direct transmission have been reported, but analysis is hampered by lack of data regarding baseline serology or blood exposure during sex.(36)

When the CDC surveillance hierarchy of risk factors is used, an infected WSW who has injected drugs or had heterosexual contact will be classified into the highest category: IDU or heterosexual transmission (HT). Currently, WSW who have HIV infection but have no other risk factors are classified as "no identified risk," that is, not having reported behaviors known to efficiently transmit HIV. It is therefore possible that rare cases of female-to-female sexual transmission of HIV have been systematically misclassified. Although HIV is present in the cervical and vaginal secretions of HIV-infected women, there is still a paucity of data on HIV in female genital fluids and the risks of transmission from various WSW sexual practices.(37)

There has been no evidence of female-to-female transmission in a study of 960,000 female blood donors in the United States. Of the total, there were 96 seropositive women, of whom, 3 reported sex with men and women and none with women only.(38) Lesbians have thus been considered at low risk. If providers assume that lesbians are not engaging in sex with men, however, they may be overlooking risk behaviors that may be more common than most providers suspect. In 1993, the San Francisco Department of Health studied 498 women, recruited at public social events aimed at lesbian and bisexual women in San Francisco and Berkeley, California. In the cohort, 68% self-identified as lesbians and 32% as bisexuals, 81% reported having sex with men, including unprotected oral (56.3%), vaginal (39%), and anal sex (10.9%). High rates of unprotected oral, vaginal, or anal sex with MSM (14.6%, 9.6%, and 3.2%, respectively) were also noted.(39) Multiple other studies conducted in New York City, Milwaukee, and Seattle corroborate these results, although they are not ideal for deriving precise prevalence data about WSW and probably involve only segments of the WSW community. These studies reemphasize the important point that sexual identity is a poor marker for sexual behaviors. Because they have been considered "safe," lesbians have not been targeted for HIV/AIDS education and may be engaging in multiple unsafe sexual practices.(40)

WSW may engage in sexual activity with men for pleasure, as commercial sex workers, or may be victims of sexual assault. Some women IDUs who engage in sex with other women are associated with riskier injection drug use and sexual behaviors and therefore are at a higher risk of acquiring HIV than other women IDUs.(37) WSW who are also IDUs were more likely than other IDUs to share needles, to exchange sex for drugs or money, to be homeless, and to seroconvert. This more marginalized group and their sexual partners would likely benefit from specially targeted prevention programs.

Although the risk of sexual transmission of HIV between WSW appears to be low, women who are partners of women infected with HIV should observe principles of safer sexual activity and avoid mucous membrane contact with all potentially infectious secretions (Table 2).(41,42)

World health leaders have generally recognized the need for female-controlled contraceptive anti-infectives and microbicides. The realities of nonconsensual sex, domestic violence, and economic abandonment limit the usefulness of the male condom in the lives of many women throughout the world. The ideal female-controlled method would be easy to use, would prevent other STDs, and could be used without a partner's consent or knowledge. The female condom, first marketed in 1993, is a polyurethane sheath that partially covers the external genitalia and may offer protection against all STDs. The polyurethane membrane is 40% stronger than latex and inhibits HIV, cytomegalovirus (CMV), and hepatitis B. As a contraceptive, it is comparable to other barrier methods including the male condom, diaphragm, cervical cap, and contraceptive sponge. The female condom has yet to attain full acceptability but has the potential to emerge as an important female-controlled contraceptive.

Nonoxynol-9 is a spermicide with potent viricidal effect in vitro and effectiveness at killing other STD-causing agents.(43) It may, however, produce vaginal irritation when used frequently.(43,44,45-47) At the Research Clinic for Prostitutes in Nairobi, Kenya, 138 seronegative prostitutes were evaluated, half routinely using a nonoxynol-9-treated sponge and half using a placebo. There were more genital ulcers and vulvitis in the nonoxynol-9 group, and the rate of HIV acquisition was similar; thus, nonoxynol-9 was not protective.(44) Nonoxynol-9, in a potentially less irritating bioadhesive compound (COL-1492), was tested in a randomized, blinded, placebo-controlled trial in 892 female sex workers in Benin, Cote d'Ivoire, South Africa, and Thailand. Unfortunately, this study found a higher rate of HIV infection among women in the nonoxynol-9 arm than among placebo users, with nonoxynol-9 nearly doubling the risk of HIV infection in women using the microbicide >3 times per day.(48) Nonoxynol-9 was also associated with an increased number of vaginal lesions. Therefore, nonoxynol-9 should not be used as a vaginal microbicide.

Other microbicides in planned or ongoing clinical trials include: carrageenan, which interferes with viral adsorption to the mucosal surface; BufferGel, which maintains vaginal pH at a level of acidity that inhibits HIV (49); PRO 2000 gel (a synthetic naphthalene sulfonate polymer shown to block HIV by binding to the viral envelope and interfering with CD4 binding) (49); and dextrin sulfate (a sulfated polysaccharide that acts at the cell surface to inhibit entry of HIV).(50) No efficacy data are available for these products.

Inhibitors of viral replication are also being considered, and a small study in an SIV-macaque model using a vaginal gel of the antiretroviral drug tenofovir was encouraging, as was a trial using the antiretroviral loviride and the disinfectant chlorhexidine in a suppository.(51,52) The lack of in vitro systems and small animal models for testing, as well as deficient commercial interest, has slowed development of all of these products.(51) More knowledge about Langerhans cells and other sites for HIV and STD acquisition in the female genital tract is urgently needed so that effective and affordable methods can be developed.

AIDS-defining diagnoses seen frequently in data combined from a number of small cohorts of women include PCP, esophageal candidiasis, disseminated Mycobacterium avium, and mucocutaneous HSV.(58) Women are not protected from any of the other AIDS-defining opportunistic infections.(59) Bacterial infections, especially respiratory infections with such encapsulated organisms as Streptococcus pneumoniae and Haemophilus influenzae, occur more frequently in IDUs than in homosexual men and occur with equal frequency in heterosexual men and women.(22,60) One of the earliest and largest prospectively followed cohorts of women with HIV was initiated in the early 1980s in Rhode Island, where 200 HIV-infected women (most of whom were white) have been followed at regular intervals. The initial clinical manifestations of HIV infection in 117 symptomatic women were candida vaginitis (n = 43), lymphadenopathy (n = 17), bacterial pneumonia (n = 15), acute retroviral syndrome (n = 8), and constitutional symptoms, such as unexplained weight loss of 10 pounds or more or diarrhea for 4 weeks or more (n = 8). The rest had syndromes suggestive of HIV infection (including thrush, tuberculosis, hairy leukoplakia, herpes zoster, PCP, HIV encephalopathy, or CMV retinitis).(53) Nonspecific conditions of vaginitis, pneumonia, and constitutional symptoms were more frequent than in HIV-uninfected women. In this cohort, PCP was the AIDS-defining manifestation in only 20%, and 34% presented with esophageal candidiasis as their AIDS-defining diagnosis. Another study from France also found that esophageal candidiasis was more common in a cohort of women than was PCP. These studies are in sharp contrast to U.S. statistics and other cohorts in which, prior to 1993, PCP was the leading AIDS diagnosis.(61-64) The presence of gynecologic conditions suggesting immunosuppression (Table 3) should prompt referral for counseling and HIV testing.

Appropriate opportunistic infection (OI) prophylaxis is necessary for both women and men, and no gender-specific recommendations can be made. In the United States, immunoglobulin G (IgG) antibodies to CMV are almost always positive in homosexual men, but the prevalence of CMV antibodies in women are more like the rates in the general adult population (approximately 50%). The CMV-negative patient in need of a transfusion should always receive CMV-negative blood to avoid future risk of CMV end-organ disease.(65,66)

Results of initial early studies analyzing progression and survival in HIV disease suggested a difference based on gender. Many early studies indicated that the prognosis for women was worse than for men regardless of risk group or race. A study undertaken in New York City early in the epidemic looked at 544 women compared to 2,526 men with first-episode PCP. Fewer women were white; more women were admitted through emergency departments; more women received care at hospitals with less experience in caring for PCP; and fewer women underwent bronchoscopy. In addition, more women were admitted to the ICU. This study reflected late access to limited care--a common scenario for women in the AIDS epidemic.(67) All of the more recent work disputes that women have increased levels of morbidity and mortality as a result of gender and supports the contention that earlier results reflected poor access to care.(68,69)

Studies of disease progression in women are limited and, when available, are derived from three principal sources: large national databases, which usually include the diagnosis of AIDS and the date of death with little additional interval information; small- to moderate-sized cohort studies of 50-200 women in industrialized countries, often followed for <3 years; and larger cohort studies conducted in developing countries.(70) One report summarized a dozen studies that overall reveal that the survival and course of disease in men and women is not substantially different when patients are matched for socioeconomic status, risk group, and access to care.(14) Reports from the CDC suggest that survival time of women and heterosexual men is similar. In the Bronx, access to or adherence to follow-up care strongly influenced survival time, with survival substantially shortened in women who had no prior care before a severe HIV-related infection. A trial of 1,372 patients in Baltimore found that gender had no relationship to disease progression.(71) A population-based rural cohort study in Uganda also found no difference in gender-specific mortality rates.(72)

A U.S. study of 3,779 men and 768 women, with median T-cell counts higher in the group of women (240 cells/µL, versus 137 cells/µL in the men), found that women were at increased risk of death (relative risk: 1.3) during a 15-month period. It is important to note that there was no increased risk of HIV disease progression in women, that increased risk of death was found primarily among IDUs, and that the deaths were secondary to bacterial pneumonia and endocarditis, both likely related to injection drug use. The study authors noted that these findings may represent differential access to care, treatments, or social support.(68) The most important predictors of survival or progression appear to be CD4 count, viral load, and the specific AIDS-defining diagnosis--not gender.(11,62,73-76)

In a review of AIDS in women reported to the CDC by December 1990, 73% of women with AIDS in the United States were residents of large metropolitan areas of >1 million people, mostly on the East Coast. The other 26% were from smaller cities, making it harder to target a specific population.(62) New CDC data mapping the estimated prevalence of adults and adolescents living with HIV or with AIDS in the United States as of 2002 suggest that this distribution has not changed.(6)

Because women frequently access health care through emergency departments, family planning clinics, STD clinics, youth guidance centers, jail clinic facilities, and drug treatment units, these sites are ideal for targeting efforts at early diagnosis and use of early intervention with antiviral and prophylactic therapies.(77)

Lack of access to care, minimal self-motivation, attention to the health care of their children over that of themselves, and disenfranchisement among a large proportion of women all contribute to decreased rates of early detection and intervention.(67) Medical providers need to maintain a low threshold to counsel and test for HIV, remembering that women infected with HIV through heterosexual transmission may be unaware of their partners' HIV status and therefore may not perceive themselves at risk of infection. HIV and AIDS for women is an issue of access to health care, and the care system is not always well suited to their needs.(58,78,79)

Several studies have found that women have lower viral loads than a men at a similar CD4 cell count.(80)

Investigation of HIV viral load levels in 527 IDUs at baseline (using branched DNA [bDNA]) and from 285 of the 527 at follow-up 3 years later (using polymerase chain reaction [PCR]) found that women IDUs had lower median viral load measurements than men with similar CD4 counts. Using bDNA, initial viral load levels were 3,365 copies/mL for women (median CD4 count 518/µL) and 8,907 copies/mL for men (median CD4 count 518/µL). Using PCR at follow-up, viral load levels were 45,416 copies/mL for women (median CD4 count 417/µL) and 93,130 copies/mL for men (median CD4 count 390 cells/µL). These results suggested that women had median viral load levels that were only about half the viral loads of men with a similar CD4 cell count, and that women with the same viral load as men had a 1.6-fold higher risk of AIDS (95% CI: 1.10, 2.32).(81)

A separate study found that HIV RNA was not different in women and men with CD4 counts below 200 cells/µL, but that at CD4 counts of 200-500 cells/µL, women demonstrated 40% lower HIV RNA levels than men, and that at CD4 counts above 500/µL, levels were 24% lower in women.(82)

These studies were further corroborated in a large CDC cohort of 3,776 participants (2,467 men and 1,309 women). Women with CD4 counts of 0-199 cells/µL had viral loads that were 40% lower than men; women with CD4 counts 200-499 cells/µL had viral loads that were 48% lower; and those women with CD4 of >500 cells/µL had viral loads that were 57% lower. All of these results were statistically significant, indicating that women have lower viral loads on average than men with comparable CD4 counts.(83,84)

How can one best use this information? There is a paucity of data supporting the use of a specific viral load threshold as an independent factor for initiating antiretroviral therapy. Moreover, multiple studies over the last decade highlight that women and men who have equal access to care have similar rates of progression to AIDS and death. The precise timing for initiating antiretrovirals in patients with >200 CD4 cells/µL thus remains controversial.

The U.S. Department of Health and Human Services ARV treatment guidelines summarize the situation as follows: "Taken together, these data suggest that gender-based differences in viral load occur predominantly during a window of time when the CD4⁺ T-cell count is relatively preserved, and treatment is recommended only in the setting of high levels of plasma HIV RNA. Clinicians may wish to consider lower plasma HIV RNA thresholds for initiating therapy in women with CD4⁺ T-cell counts >350 cells/mm³, although there are insufficient data to determine an appropriate threshold. In patients with CD4⁺ T-cell counts <350 cells/mm³, very small sex-based differences in viral load are apparent; therefore, no changes in treatment guidelines for women are recommended for this group."(85)

It is estimated that 90% of HIV infections that occur among children are a result of vertical transmission from a mother to her infant. Without the use of ART, vertical transmission rates of HIV are estimated to be 15-25% in industrialized nations and 25-45% in developing countries.(75) Unequivocally, one of the most successful uses of antiretroviral medications has been the prevention of vertical transmission. Numerous studies show that brief courses of antiretroviral treatment given to mother and infant can dramatically reduce the risk vertical transmission, with little or no significant long-term sequelae to the infant. (For more information, health care providers should contact the Antiretroviral Pregnancy Registry, 800-258-4263.)

Perhaps the most effective means of reducing mother-to-child transmission is to provide fully suppressive ART to the mother in the long term, thereby not only reducing the risk of vertical transmission, but also sustaining the life and health of the mother while the child is growing up. Obstetrical risk factors that increase risk of vertical transmission include invasive fetal procedures, chorioamnionitis, active genital HSV or other STD while in labor, and premature rupture of membranes coupled with a low CD4 count. It is important to provide ongoing prenatal care to address and modify as many of these factors as early as possible to minimize vertical transmission.

In some resource-constrained countries, breast-feeding is often the safest or only option a mother has to feed her child, despite the attendant risks of HIV transmission and increased maternal mortality. It is estimated that vertical transmission in the postpartum period occurs in 15% of breast-fed children with a chronically HIV-infected mother and in 29% of infants with a breast-feeding mother who becomes infected with HIV during the postpartum period.(86) To minimize vertical transmission and provide sustained benefit to mother and baby, it is incumbent that we work to ensure that safe alternative feeding options and long-term effective antiretroviral therapy are available for all HIV-infected mothers.

Kaposi sarcoma (KS) has occurred much less frequently in women than in men in the United States, although the rates in MSM are much higher than in heterosexual men with HIV.(87) In Africa, the AIDS epidemic has resulted in a decrease in the male-to-female ratio of KS (from 19:1 to 1.7:1), suggesting that route of transmission or other factors may affect the risk in women of developing KS.(88)

Non-Hodgkin lymphoma, an AIDS-defining diagnosis, is encountered too infrequently in women to compare gender-specific incidence rates.(89,90) It does, however, constitute an AIDS-defining diagnosis in both sexes. Breast cancer, lung cancer, and other malignancies (with the exception of cervical carcinoma, discussed below) have not been found to occur with higher frequency in HIV-infected women than in the general population.(91)

Much attention has been given to disorders that may be more frequent, more severe, and less responsive to therapy in HIV-infected women (particularly those with advanced immunosuppression) than in women without HIV infection: human papillomavirus (HPV)-associated cervical disorders such as cervical intraepithelial neoplasia (CIN); Candida vaginitis; and pelvic inflammatory disease (PID).(92,93-97) These disorders are recognized as HIV-associated conditions in the expanded CDC case definition.(98,99)

Cervical Disorders

Human Papilloma Virus and Cervical Neoplasia Limited evidence suggests an increased occurrence and aggressiveness of cervical cancer in women with HIV infection.(61,100) As of 1993, cervical cancer constitutes an AIDS diagnosis.(98) In the U.S. population, invasive cervical cancer is more common among black women living in the South.(101-105) HPV is believed to be an etiologic factor in human cervical cancer and has been investigated since the 1970s. Approximately 95% of cervical condyloma, all grades of CIN, and invasive cervical cancer contain HPV DNA. HPV types 16 and 18 have been found most commonly in cervical cancer, and types 6 and 11 are most frequently associated with benign condylomas or low-grade squamous intraepithelial lesions (SILs).(106,107) Immune suppression appears to make one particularly susceptible to infection by oncogenic strains of HPV. Reports indicate that autograft transplant recipients receiving immunosuppressive treatments have a ninefold increase in incidence of HPV anogenital infection and a 16-fold increase in incidence of CIN compared with the general population. HPV prevalence, acquisition, and retention are higher in HIV-positive women than in matched controls. By DNA analysis, types 16, 18, and 33 are associated most frequently with CIN in HIV-positive women and HIV-negative controls,(95,108) especially in women with CD4 counts of <200 cells/µL.(109-117) An increased frequency of abnormal Papanicolaou (Pap) smear results was first noted in women attending clinics for HIV, methadone maintenance, and cervical dysplasia.(93,94) In an early study, 40% of 35 HIV-positive women had SILs on cervical cytologic examinations, compared with 9% of 32 HIV-negative women.(94) In another early study of 32 HIV-infected women who underwent colposcopy, 78% had a normal Pap smear result, and only 3% had cytologic Pap smear findings suggesting CIN. Colposcopy associated with biopsy, however, disclosed that 41% had cervical abnormalities.(96) This finding had profound implications, because colposcopy is more costly and more time consuming and requires specially trained personnel for proper interpretation. Subsequent studies have not confirmed these findings and, in fact, suggest that a correctly obtained and expertly read Pap smear (at least under research conditions) has sufficient sensitivity and specificity to detect cervical abnormalities.(118) A large, prospective, multicenter study conducted by the CDC in conjunction with investigators in Miami and New York compared HIV-positive and -negative women, all of whom had Pap smears, colposcopy, and evaluation for HPV by PCR. This controlled, prospective trial confirmed numerous prior uncontrolled studies of HIV-positive cohorts in finding that CIN is more frequent in HIV-positive women (20% in HIV-positive compared with 4% in HIV-negative women) and that more advanced CIN is more likely with a more advanced stage of immunosuppression. By multivariate analysis, HPV (odds ratio (OR) = 9.8), HIV infection (OR = 3.5) and CD4 count (OR = 2.7) were independently associated with CIN.(105,106) The Women's Interagency HIV Study (WIHS), a large natural history study of HIV-infected women, reported on a cohort of 2,015 HIV-infected women and 577 HIV-seronegative controls. Those women who were HIV infected had the highest rates of HPV infection (58%), compared with seronegative controls (26%). HIV-infected women also demonstrated a 42% prevalence of infection with multiple types of HPV versus a 16% prevalence in uninfected women. Risk factors for CIN/SIL were CD4 lymphocytes <200 cells/µL, higher risk HPV types, and infection with multiple types of HPV. In a meta-analysis examining 19 studies incorporating 728 HIV-infected women, dysplasia was found to occur in 64% of those with AIDS and in 36% of HIV-positive women.(119,120) Are Pap Smears Sensitive in HIV-Infected Women? Because CIN is more frequent and more aggressive in women with severe immunosuppression, some practitioners routinely recommend a Pap smear every 6 months, for women with CD4 counts of <200 cells/µL.(121) The WIHS study also suggested that Pap smears were highly sensitive in HIV-positive women. Subsequent studies confirm that Pap smears were equally sensitive for detecting cervical disorders in HIV-positive and -negative women.(120,122) One study also demonstrated, however, that 15% of the dysplastic lesions in HIV-positive women seen in a large dysplasia referral clinic consisted of vulvar, vaginal, or perianal lesions detected only by colposcopy. These lesions would have been missed if only a cervical Pap smear had been performed.(120) Although colposcopy would be the most sensitive and specific diagnostic tool, the general lack of availability and standardization necessitates assiduous attention to appropriate Pap testing and follow-up. At this time, performing a Pap screening every 6 months along with careful vulvar, vaginal, and anal inspection appears prudent, especially in patients with CD4 counts <200 cells/µL. Colposcopic evaluation should be performed if a Pap smear reveals any atypical squamous cells of unknown significance (ASCUS), atypical glandular cells of unknown significance (AGCUS), low-grade or high-grade SIL, or any persistent inflammation that is unresolved after treatment for infections such as gonorrhea, trichomonas, or chlamydia. If the initial inspection reveals suspicious vulvar, vaginal, or anal lesions, colposcopy should be performed in those areas to evaluate for dysplastic lesions. Colposcopy should be considered as the initial test for women with poor likelihood of follow-up, or suspicion of extracervical disease (Table 4).(120,123) Pregnancy did not affect the incidence of abnormal smears in women enrolled in the Women and Infants Transmission Study.(124) Treatment of CIN As in HIV-uninfected women, CIN 1 appears to have a low rate of progression to CIN 2 or 3. It should also be noted that treatment of CIN 1 has a high failure rate in HIV-infected women.(125) An approach used at San Francisco General Hospital is to observe a woman with CIN 1 if she has a CD4 count of 200 cells/µL or higher and has demonstrated the willingness and ability to commit to follow-up appointments. Careful surveillance allows the clinician to treat any progression of CIN. Cryotherapy is appropriate for CIN 1 if a woman has a CD4 count <200 cells/µL or is likely to be lost to follow-up evaluations. Appropriate treatment of CIN 2 or 3 requires use of an ablative or an excisional procedure. Cryotherapy is appropriate for CIN 2 or 3 if there is a satisfactory colposcopy and no prior cervical treatment, and if the lesion is completely visible, <2 cm in diameter, and is affecting no more than 2 quadrants. Laser ablation can be useful for women who have lesions that are >2 cm or involve 3 or more quadrants. Laser ablation can also be used to treat vaginal lesions. Loop electrical excisional procedure (LEEP) is helpful when cryotherapy is not appropriate due to size of the lesion or if the location of the lesion is high in the endocervix. LEEP is not possible when cervical architecture is disrupted secondary to a prior LEEP or cone biopsy. Cold knife cone biopsy cannot be done in the clinic and requires an operating room. It is best reserved for a high-grade lesion where malignancy is detected on Pap smear and microinvasive disease or a glandular lesion may be present. Cone biopsy may also be used for a diagnostic test if LEEP cannot be done for the reasons described above. Topical application of the antineoplastic agent 5-fluorouracil (5-FU) appears to be a useful adjunctive modality after standard excisional or ablative treatment of CIN 2 or 3.(126) The role of smoking and CIN is explained to all patients, and smoking cessation is recommended and supported in all patients.(125) In HIV-negative women, there is a small risk (5-10%) of CIN 2 or 3 recurring after therapy. All of the treatments for CIN appear less effective in HIV-positive women, especially those women with more advanced immunosuppression, and the risk of recurrence is much higher.(127) Careful surveillance with treatment of recurrences is necessary to prevent the development of invasive cervical cancer. Two small studies suggest that effective ART plays a beneficial role in decreasing recurrence and progression of CIN.(125,128)

Genital Warts

Genital warts (GW) may be large, multifocal, and more prone to recurrence in HIV-positive individuals than in those without HIV infection. Traditional topical treatments such as trichloroacetic acid (TCA), cryotherapy with liquid nitrogen, or cryoprobe are usually effective. Liquid nitrogen may be less painful at the time of application than TCA. In applying TCA, it may be difficult to prevent the compound from affecting adjacent healthy tissue. Liquid nitrogen may allow the clinician to apply the therapy to a more circumscribed area. Laser or surgical excision may be required in cases recalcitrant to these therapies or in lesions involving large surface areas. Topical application of imiquimod analogue 2% cream has been tested in women without HIV infection, in men in a number of small trials, and in one trial involving HIV-infected patients. In a trial enrolling 60 HIV-uninfected women, 25 of 30 women in the 2% imiquimod group had achieved cure, whereas only 1 of 30 in the placebo group achieved cure. Five of 26 relapsed at 11 months.(129) In a similar trial of 60 men, at the completion of 4 weeks of imiquimod treatment, 70% (21 of 30) were cured of all warts versus 3 of 30 in the placebo group. At 18 months, 1 patient in the imiquimod group and 2 in the placebo group had relapsed.(129) A systematic review in 2001 demonstrated that complete clearance of warts occurred in only 51% of HIV-negative patients treated with imiquimod 2% or 5%; one trial in HIV-positive patients did not find imiquimod to be effective.(130) Treatment success of imiquimod in HIV uninfected patients is reasonable but not outstanding, and initial evidence suggests treatment in HIV-infected patients is not effective. Although imiquimod offers the HIV-infected patient the greater privacy of home administration, it has not replaced the traditional treatments of TCA and liquid nitrogen.

Anal Squamous Intraepithelial Neoplasia

Anal squamous intraepithelial neoplasia (ASIL) is also more prevalent and increasingly recognized in HIV-infected individuals.(131) More aggressive growth of HPV in the setting of immunosuppression may result in ASIL and invasive anal cancer. Receptive anal intercourse undoubtedly plays a role in some of the transmission of anal HPV in women and men,(114,132,133) but ASIL is often present in women with no history of ever having anal intercourse. In most HIV-infected women, detection of HPV in either cervical or anal sampling correlated with presence of HPV in the other site as well.(134) It is uncertain whether fingers or sexual devices with HPV contamination are capable of spreading infection in either men or women. In one study of 102 HIV-infected women and 96 HIV-uninfected women, anal cytologic abnormalities were significantly more common in HIV-infected women (26%) than in HIV-uninfected women (7%). In the HIV-infected group, 95% of the abnormalities consisted of mild squamous cytologic atypia, and only 5% were low-grade ASIL. Notably, the likelihood of cytologic abnormalities was greater in women with lower CD4 lymphocyte counts. In a group of women IDUs in San Francisco, ASIL was as prevalent as CIN.(114,135-137) At this time, it is difficult to make specific recommendations for screening or for treatment for ASIL in women because the natural history of ASIL and the likelihood of progression to anal carcinoma in women is unknown. Studies are ongoing in the United States to determine optimal screening and treatment regimens for both men and women. Women with other sites of HPV anogenital disease and lower CD4 lymphocyte counts are more likely to have ASIL. If a demonstrated decrease in morbidity and mortality resulted from screening and treatment, screening would be justified for women with low CD4 lymphocyte counts, and even for women without known HPV infection but with clinical prognosticators suggesting long survival. In all cases, morbidity and mortality considerations for the individual patient should weigh heavily into a decision of whether to screen or treat for ASIL.

Vaginal Candidiasis

Vaginal candidiasis may cause more severe morbidity in HIV-infected women than in the general population.(53,64,138-140) The revised CDC case definition of 1993 designated severe vaginal candidiasis as an HIV-associated symptomatic disorder.(26) Vaginal candidiasis may occur in early or late HIV disease, but, surprisingly, many women with severe immunosuppression do not have Candida vaginitis. Physicians who see women with recurrent or refractory vaginal candidiasis should offer HIV counseling and testing.(141,142) (See chapter "Candidiasis and HIV.")

Genital Ulcerative Disease

Genital ulcerative disease is a well-described risk factor for transmission of HIV.(143) A wide array of conditions may present with genital ulcerations, including HSV, CMV, syphilis, chancroid, gonorrhea, acid-fast bacterial infections, other bacterial or fungal pathogens, malignancies, and drug reactions. Genital ulcers may be painful, disabling, difficult to treat, and merit a full diagnostic evaluation. Herpes Simplex Virus Severe ulcerative genital herpes is a common AIDS-defining diagnosis in HIV-infected women.(53) Genital HSV infections are prevalent in the population at large and may be particularly refractory in HIV-infected men and women. In HIV-infected women, HSV may present and recur frequently on the labia majora, labia minora, and the less common sites of the sacrum and buttocks.(144,145) HSV-2 shedding is increased in HIV-positive women and is usually asymptomatic. As immunosuppression progresses, as measured by declining CD4 count, shedding of HSV-2 becomes more common.(146) HSV-2 infection is associated with increased transmission of HIV in men.(147) and in heterosexual couples.(148) It is reasonable to assume that HIV occurs in HSV-2 lesions of coinfected women just as it has been shown to occur in men.(149) (See chapter "Herpes Simplex Virus and HIV.") Giant Idiopathic Aphthous Genital Ulcers Giant idiopathic aphthous genital ulcers occur infrequently but are painful, disabling, and difficult to treat. Such ulcers should be evaluated for the presence of HSV, CMV, syphilis, chancroid, gonorrhea, bacterial or fungal pathogens, and malignancies. The pathogenesis may be similar to giant esophageal aphthous ulcers. The mainstay of treatment of oral or genital aphthous ulcers is a 1- to 2-week course of topical steroids and, if unsuccessful, systemic steroids. Thalidomide may be effective for esophageal apthous ulcers associated with HIV,(150) but there is little published experience on the treatment of aphthous genital ulcers in women.(151) In the acute retroviral syndrome of early HIV infection, genital ulcerations generally present only in those who have acquired HIV through sexual transmission. Because sexual exposure is the most common route of transmission in women, assessment of possible acute retroviral syndrome should include evaluation for genital aphthous ulcerations.(152)

Pelvic Inflammatory Disease

PID is not infrequent in women with HIV, and it may be more severe with advanced HIV disease.(153,154) Early findings of the Multicenter HIV and PID Study Group did not confirm a substantially more severe clinical presentation. PID responded to therapy equally in HIV-infected and -uninfected groups.(95,155) HIV-positive patients may present with lower white blood cell counts than their HIV-uninfected counterparts. Tubo-ovarian abscess formation has been reported to occur in as many as 25% of HIV-infected women with PID versus 12% of HIV-uninfected controls with PID. Therefore, more surgical intervention is required, especially in patients with more advanced HIV disease. Standard antibiotic regimens that include anaerobic coverage (such as metronidazole plus a fluoroquinolone) can be used initially in the less complicated patient. Most clinicians maintain a low threshold to hospitalize, and treat with IV antibiotics if the HIV disease is advanced.(156)

Menstrual Disorders

Currently, no data are available to suggest that the diagnosis or treatment of menstrual disorders should be modified in HIV-positive women, but documentation of specific types of menstrual irregularity in this group is lacking. Menstrual disorders are relevant not only because of personal discomfort but because unsuspected pregnancy may masquerade as a delayed period. Menstrual irregularities make prediction of time of fertility difficult for those wishing to conceive or avoid pregnancy, and may potentially increase partner exposure to menstrual blood.(157) Sporadic reports suggest that HIV-infected women experience higher rates of menstrual irregularities, including amenorrhea, menorrhagia, intermenstrual bleeding, and worse premenstrual syndrome (PMS). A New York study of IV drug using identified more menstrual abnormalities the HIV-positive women than in HIV-negative women (41% vs 24%).(158) Amenorrhea and between-period bleeding were noted particularly. Other small studies support the finding that approximately one third of HIV-positive women have either excessive bleeding or amenorrhea.(24,159,160) Menorrhagia, in particular, may be more common at higher CD4 counts.(161) A small controlled trial of HIV-negative and -positive women with midstage disease found no menstrual differences between groups.(140) The fluctuations of progesterone, estradiol, and cortisol throughout the menstrual cycle of HIV-infected women with normal cycles was identical to well-established norms in the population at large.(162) Data from a WIHS cohort of 802 HIV-infected women and 273 HIV-uninfected women revealed that seropositivity may slightly increase the probability of very short cycles, but that HIV seropositivity had minimal impact on amenorrhea, menstrual cycle length or cycle variability after adjustment for body mass index, substance use, and demographic characteristics. As in other studies, higher viral loads and lower CD4 counts were correlated with increased cycle variability and polymenorrhea.(163) A small study assessed anovulation and early menopause (defined using follicle-stimulating hormone [FSH] and progesterone levels) in 33 women aged 20-42 years enrolled in studies of the AIDS Clinical Trials Group (ACTG). Anovulation was noted in 16 of 33 women (48%), and early menopause was noted in 2 of 24 (8%) of women in whom FSH levels were measured. It is important to note that the 8% found in this study is congruent with rates of early menopause in the general population. Finally, women in the study who ovulated had higher CD4 cells and were less likely to report a recent change in menstrual periods.(164) In contrast, another small study from WIHS that evaluated ovulation among 14 women found that anovulatory women tended to be older (mean age of 38 vs 31 years) and to have higher CD4 cell counts (mean 632 vs 392 cells/µL) than women who ovulated.(165) The interaction between HIV infection, immune status, and sex hormones requires further elucidation. These relationships are likely to differ depending on immune status, and may a change dramatically as the individual receiving effective antiretroviral therapy moves from advanced HIV infection to improved immune function.