Published in the Bulletin of Experimental Treatments for AIDS Summer 1999 issue, by the San Francisco AIDS Foundation.

Related BETA Article:
Safer Sex Guidelines

Summer 1999 Table of Contents

Main Page

beta@sfaf.org

Sexual Transmission in the Era of New Treatments

By Liz Highleyman

The advent of protease inhibitors in 1996 and the widespread use of highly active antiretroviral therapy (HAART) in the ensuing years has had a dramatic effect on HIV disease progression and rates of death due to AIDS. But many people also wonder whether HAART affects HIV transmission.

There are several ways in which antiretroviral therapy could impact HIV transmission. Some are biological, such as the possibility that HIV may not be as easily transmitted by people who have a low viral load. Others are psychological and social, and are related to the relaxation of safer sex behavior by people who believe that HIV/AIDS is no longer the threat it once was. Recent media reports of increases in unprotected sex among gay men have brought renewed attention to the issue of sexual transmission of HIV, and underscore the need to understand better how the virus is transmitted in the HAART era.

The Biology of Sexual Transmission

HIV may be transmitted sexually when HIV in one partner's semen, cervical/vaginal fluid, or blood comes into contact with the bloodstream or mucous membranes of another partner.

Much of what is known about sexual transmission of HIV comes from studies in animals. In a monkey model of vaginal infection with simian immunodeficiency virus (SIV), a relative of HIV, the virus binds to and is engulfed by immune system cells in the vaginal lining that carry the CD4 surface receptor (macrophages and dendritic cells). Infected dendritic cells known as Langerhans cells then transport the virus to local lymph nodes, where it is presented to CD4 helper T-cells. Within two days, SIV can be found in lymph nodes in the groin area; within five days, SIV-infected T-cells and free virus can be detected in the blood. Target cells susceptible to HIV infection have also been found in the cervix (especially the columnar zone, so named for the predominant type of cell), foreskin, urethra, and rectum.

At the 6th Conference on Retroviruses and Opportunistic Infections (CROI) in February 1999, R. Otten and colleagues reported that HIV-2 RNA (genetic material) could be detected in the cervical/vaginal fluid of infected macaque monkeys as early as one week after exposure, even before the virus was detectable in the blood. Vaginal and blood viral loads peaked at two to three weeks after exposure, and HIV-2 antibodies were detectable in the blood by three to four weeks following exposure. In humans, HIV antibodies can usually be detected in the blood by 12 weeks after exposure, and in almost all cases by six months, the widely quoted "window period" between initial exposure and the point at which an HIV antibody test will typically show a positive result.

Factors Affecting Sexual Transmission

It is estimated that three-quarters or more of all HIV infections worldwide are transmitted sexually. However, it is clear that HIV transmission does not occur during every unprotected sexual contact. Sexual transmission is a product of the interaction of a wide range of host factors, viral factors, and environmental factors.

Host Factors

Host factors are characteristics of the person transmitting or receiving the virus. HIV transmissibility is influenced both by the infectiousness of the source partner and the susceptibility of the recipient partner. These factors vary greatly among individuals. There are cases in which HIV has been transmitted by a single sexual exposure, and other cases in which people have been exposed multiple times without becoming infected.

Infectiousness of the source partner involves several factors. The likelihood of transmission is known to be associated with the stage of HIV disease. In the September 1994 issue of AIDS, Pietro Vernazza and colleagues reported that people with lower CD4 cell counts and late-stage, symptomatic HIV disease were more likely to transmit the virus during sex than people with higher CD4 cell counts and asymptomatic disease. This may be related to the fact that people with later-stage disease typically have higher viral loads and more virulent strains of HIV. Several studies have shown that higher levels of HIV in the blood are associated with a higher likelihood of sexual transmission. Women with later-stage disease and higher blood viral loads are also more likely to transmit HIV vertically to their infants.

Studies have shown that HIV viral load levels in the semen and cervical/vaginal fluid also affect transmission. For example, M. Dalod and colleagues reported at the 37th Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC) in September 1997 that in a study of 15 heterosexual couples in France, men who transmitted HIV to their female partners had significantly higher semen viral loads than men who did not transmit the virus, despite having similar blood viral loads. Genital fluid and blood viral loads may not necessarily correlate, and they may be affected differently by anti-HIV treatment.

Viral load levels in the blood and genital fluids are often very high immediately following exposure, during a stage known as primary HIV infection. Evidence indicates that newly infected persons are more likely to transmit HIV than those at subsequent stages of the disease, especially during the "window period" before seroconversion. Some researchers believe that transmission from people in the earliest stages of HIV disease may play a major role in the progression of the epidemic.

Other factors that may influence infectiousness include pregnancy and phase of the menstrual cycle. The European Study Group on Heterosexual Transmission (ESGHT) reported in the March 28, 1992 issue of the British Medical Journal that women are more likely to transmit HIV to their male partners during menstruation. Another study showed that women are three times more likely to have HIV in their cervical/vaginal fluids during pregnancy. In addition, cervical/vaginal viral load increases in the period immediately following a hysterectomy (removal of the uterus).

Susceptibility to HIV is also affected by several different factors. Transmission is more likely when there is damage or injury to the mucous membranes that come into contact with infected semen or cervical/vaginal fluid. In men, greater susceptibility to HIV infection has been associated with non-circumcision (perhaps because Langerhans cells in the foreskin act as targets for the virus) and possibly with the use of vasodilating drugs such as poppers. In women, greater susceptibility has been associated with higher vaginal pH (a more acidic vaginal environment is protective against infection), cervical ectopy (the presence of a certain type of cell on the exterior of the cervix, a condition particularly common in young women), and contraceptive use (IUDs are associated with greater susceptibility, while there is contradictory evidence regarding the influence of oral contraceptives).

Researchers have also determined that susceptibility to HIV infection involves genetic co-factors. People who have two copies of a specific mutation of the gene that encodes the cell surface co-receptor CCR5 appear less likely to become infected with certain strains of HIV. People with one mutant copy of the gene tend to experience slower HIV disease progression. The frequency of the mutation differs among population groups, being higher among people of European descent and lower among people of Asian or African descent. The type of HIV that is most likely to be transmitted sexually (M-tropic strains) uses the CCR5 co-receptor to enter human host cells. However, the protection afforded by this mutation is not absolute; soon after the CCR5 hypothesis was proposed, researchers began to uncover cases in which HIV infection occurred in persons with two mutant copies of the gene (see "Research Notes," BETA, June 1997). A different, newly discovered CCR5 mutation that is more common in African-Americans appears to increase susceptibility to HIV infection (see "News Briefs," BETA, April 1999).

The presence of sexually transmitted diseases (STDs) can also impact sexual transmission of HIV, increasing both infectiousness and susceptibility. At the 12th World AIDS Conference in June 1998, researchers from the Centers for Disease Control and Prevention (CDC) presented information showing that in people with certain STDs, epithelial cells lining the genital and urinary tracts do not function properly; gaps can form between the cells, making it easier for HIV to enter. In addition, dendritic cells collect in inflamed genital tissues, where they can become targets for HIV infection. In the June 28, 1997 issue of the Lancet, Myron Cohen, MD, and colleagues reported that in 135 HIV positive Malawian men, HIV viral loads in the semen were eight times higher in men with urethritis (inflammation of the urethra often associated with STDs) compared to men without the condition, despite similar blood viral loads. Semen HIV viral loads decreased when men were given antibiotic treatment for urethritis. D. Cameron and colleagues reported in the Lancet in 1989 that women with genital ulcers associated with STDs were more likely to transmit HIV to their male partners than women without such ulcers. Genital lesions may be a symptom of syphilis, herpes simplex, or chancroid. Several research groups have presented evidence suggesting that prevention and treatment of STDs can help reduce the rate of HIV transmission, especially in areas at an early stage of the HIV epidemic. For example, in a Tanzanian study, villages in which a STD treatment program was implemented witnessed a 40% reduction in HIV incidence (new cases) over two years compared to villages without such a program.

Viral Factors

Specific viral factors also influence HIV transmissibility. There are two major types of HIV: M-tropic strains, which preferentially infect macrophages and dendritic cells, and T-tropic strains, which preferentially infect T-cells. T-tropic strains are syncytium-inducing (SI), that is, they cause infected cells to clump together, and tend to lead to more rapid disease progression. M-tropic strains are non-syncytium-inducing (NSI) and tend to lead to less aggressive disease, at least initially. M-tropic strains are most likely to be transmitted sexually.

Viral subtype may also contribute to transmission risk. A study in Thailand indicated that subtype E strains of HIV-1 may be more likely to infect Langerhans cells in the vagina and penis than subtype B strains. Some researchers have suggested that this may be why HIV tends to be more often transmitted by heterosexual sex in this part of the world.

Environmental Factors

Environmental factors that influence HIV transmission are many and varied. They include societal norms concerning sexual behavior and relationships; specific sexual activities and practices; patterns of partnering, including the rate of partner change; use of protection such as condoms; HIV prevalence in a population, which determines the number of infected people in a pool of potential sexual partners; and the amount of time the epidemic has been present in a given geographic area.

Anal, Vaginal, Oral-How Risky Are They?

Different sexual acts are associated with varying likelihoods of transmitting HIV. It is clear that unprotected anal intercourse has the highest potential for transmitting the virus, followed by unprotected vaginal intercourse. Anal sex is more likely than vaginal sex to transmit HIV because the rectal lining is thinner and more prone to minor damage that can facilitate viral access. In both anal and vaginal unprotected intercourse, the receptive partner (the "bottom") is at a significantly higher risk than the insertive partner (the "top"), because the receptive partner accepts semen into the rectum or vagina, allowing prolonged contact with mucous membranes. The ESGHT found that male-to-female transmission is 1.9 times more likely than female-to-male transmission during vaginal intercourse. Although the risk of infection is significantly less for the insertive male partner, it has been documented.

There has been considerable debate concerning the risk of HIV transmission through oral sex. Epidemiological evidence indicates that oral sex carries a much lower risk of HIV transmission than anal or vaginal intercourse. There have been only about 30 reports in the medical literature of documented HIV transmission via receptive fellatio (oral sex on a man), and fewer documented cases attributable to giving cunnilingus (oral sex on a woman); there have been only three reports of transmission to the insertive partner in fellatio. Most population-based epidemiological studies have not found significant evidence of transmission via fellatio; however, research in this area is difficult because few people have only oral sex to the exclusion of other sexual activities.

The risk of transmission via fellatio is likely to be higher if the receptive partner has gum disease, other oral infections, or injuries in his or her mouth, or if blood is present, for example due to bleeding gums. The risk of transmission via cunnilingus is greater if the receptive partner is menstruating. In their review of oral transmission of HIV in the November 12, 1998 issue of AIDS, Richard Rothenberg and colleagues concluded that "unprotected receptive oral intercourse is of low but non-trivial infectivity and widespread occurrence." However, there has been a trend toward increased numbers of case reports of HIV transmission via oral sex in recent years, possibly due to a higher incidence of oral sex in lieu of more risky anal sex. There has been one reported case of HIV infection through receptive oral/anal sex.

Female-to-female transmission of HIV is theoretically possible, but epidemiological evidence indicates that it is rare. There have been only two documented reports of probable transmission via lesbian sex in the medical literature, one of which involved exposure to blood.

Several researchers have attempted to quantify the HIV transmission risk associated with specific sexual activities, with widely varying results. At the 5th CROI in February 1998, E. Vittinghoff presented estimates of the per-contact risk of various male-to-male sexual activities. He found a 0.0024 probability of transmission per contact for unprotected receptive anal sex, and a 0.0003 probability for both unprotected insertive anal intercourse and unprotected receptive oral sex. Mitchell Katz, MD, of the San Francisco Department of Public Health, and Julie Gerberding, MD, now with the CDC, reviewed the medical literature on the probability of HIV transmission via several specific routes in the April 10, 1997 issue of the New England Journal of Medicine (see table, below).

Probability of Transmission Associated with Specific Types of Exposure to HIV *
Unprotected receptive anal intercourse:	0.008 to 0.032 (80 in 10,000 to 320 in 10,000)
Use of HIV-contaminated drug injection equipment:	0.0067 (67 in 10,000)
Contaminated needlestick exposure by health-care worker:	0.0032 (32 in 10,000)
Unprotected receptive vaginal intercourse:	0.0005 to 0.0015 (5 in 10,000 to 15 in 10,000)
Unprotected insertive vaginal intercourse:	0.0003 to 0.0009 (3 in 10,000 to 9 in 10,000)
Unprotected insertive anal intercourse:	no data, but believed to be similar to unprotected insertive vaginal intercourse
Unprotected oral sex:	no per-exposure estimates, but cases have been documented in the medical literature
*Rates are for a single exposure of the type indicated; ranges indicate the rates found in different studies.
Adapted from Katz, M.H. and Gerberding, J.L. Postexposure treatment of people exposed to the human immunodeficiency virus through sexual contact or injection drug use. New England Journal of Medicine 336(15): 1097-1099. April 10, 1997.

Genital Viral Load

Semen

HIV has been detected in T-cells and macrophages in the seminal fluid, but is probably not present in sperm cells themselves. Several studies have examined semen viral load and its relation to blood viral load; unfortunately, different studies have yielded divergent results, making interpretation difficult.

Vernazza and colleagues reported in the July 1997 issue of AIDS that semen viral load was correlated with blood viral load and immune status as measured by CD4 cell counts; people with lower CD4 cell counts had higher blood and semen viral loads. At the February 1999 CROI, J. Evans and colleagues reported that blood and semen viral load levels are correlated, but that semen viral load is on average one log (ten-fold) lower than blood viral load. Other studies suggest that there is little correlation between blood and semen viral load levels, and that a higher semen viral load is not necessarily associated with the degree of immunosuppression. For example, J. Krieger and colleagues reported in the February 1991 Journal of Infectious Diseases that shedding of HIV in the semen was not correlated with disease stage or CD4 cell count. K. Hamed's group reported similar results in the April 1993 issue of the same journal.

Phalguni Gupta of the University of Pittsburgh and colleagues reported at the February 1999 CROI that semen viral load may vary widely over time in a given individual. The researchers found that in a group of 18 men, 28% did not shed HIV in their semen, 28% shed HIV consistently, and 44% shed HIV intermittently. Gupta concluded that many of the men "had very high levels of virus present in their semen at certain time points and then at other times had a lower amount.... In some HIV-infected men, the levels of virus in semen can vary from one day to the next.... This indicates that some men may be more likely to transmit the virus during risky sexual behavior at certain times." Gupta added that people "can't rely on blood tests" to assess the risk of sexual transmission of HIV.

Although semen viral load generally increases as HIV disease progresses, virus may be shed in the semen early in the course of HIV disease; in fact, shedding may "spike" soon after infection. In the September 1992 issue of AIDS, B. Tindall reported that he was able to detect HIV in the semen of three men using a polymerase chain reaction (PCR) assay as early as 17 days after the onset of primary HIV infection. He suggested that men may be "potentially infectious through sexual transmission during the first few weeks after infection."

Pre-Ejaculatory Fluid

HIV has been detected in pre-ejaculatory fluid ("pre-cum"), but in smaller amounts than in semen. In the December 12, 1992 issue of the Lancet, Jeffrey Pudney and colleagues reported that in laboratory tests they found HIV-infected cells in pre-ejaculate samples from six of nine HIV positive men. HIV was detected in the pre-ejaculatory fluid of both symptomatic and asymptomatic men, and in men both taking and not taking AZT (Retrovir). Epidemiological evidence indicates that anal, vaginal, or oral exposure to pre-ejaculate is a significantly less likely route of HIV transmission than exposure to semen. However, the possibility of transmission via pre-ejaculatory fluid cannot be ruled out, and there have been several anecdotal reports of transmission by this route. Pudney and colleagues concluded that further studies are needed, but "until then, it should be considered that pre-ejaculatory fluid is potentially infectious, and mucosal exposure to the fluid should be avoided through protected sex."

Cervical/Vaginal Secretions

HIV can be detected in the cervical and vaginal secretions of HIV positive women. At the February 1999 CROI, Susan Cu-Uvin, MD, and colleagues reported that HIV in cervicovaginal lavage (CVL) fluid was correlated with blood viral load levels. In a study of 95 HIV positive women, of those who had a blood viral load below 400 copies/mL, none had a detectable CVL viral load; of those who had a blood viral load of 401-9,999 copies/mL, 4% had a detectable CVL viral load; and of those who had a blood viral load greater than 10,000 copies/mL, 53% had a detectable CVL viral load. The women were also more likely to have a detectable CVL viral load if their CD4 cell count was lower. In another report by Cu-Uvin, presented at the 1998 World AIDS Conference, 45 of 172 HIV positive women had a blood viral load higher than their CVL viral load, while 16 women had a higher CVL viral load. In this study, 96% of women who had an undetectable blood viral load also had an undetectable CVL viral load. The researchers did not find a blood viral load threshold below which genital shedding of HIV did not occur.

Levels of HIV in cervical/vaginal fluid may impact sexual transmission as well as vertical transmission from mother to infant. In the Otten monkey study described previously, female macaques had detectable HIV genetic material in their cervical/vaginal fluid as soon as one week after exposure, before HIV was detectable in the blood, suggesting that HIV transmission is possible very soon after initial infection. P. Reichelderfer and colleagues reported at the February 1999 CROI that shedding of HIV in human cervical/vaginal fluid appears to vary over the course of the menstrual cycle, being lowest during the follicular phase and highest during menses.

Rectal Viral Load

Secretions in the rectum may also harbor HIV. C. Celum and colleagues reported at the 1997 annual meeting of the Infectious Diseases Society of America (IDSA) that in a study of ten HIV positive men, nine of whom were taking antiretroviral therapy, six had detectable rectal viral loads. Out of 33 total rectal samples (four from each man, minus seven invalid samples), four samples indicated a rectal viral load greater than the corresponding blood viral load. The presence of HIV in the rectum has implications for the transmission of HIV to an insertive partner during anal intercourse.

Compartmentalization of HIV

Recent work by Joseph Eron, MD, Vernazza, and colleagues contributes to a better understanding of how HIV behaves in different parts of the body. The researchers reported in the October 22, 1998 issue of AIDS that HIV can evolve into genetically distinct strains in different body compartments such as the blood and the genital tract. In a study of 11 HIV positive men, the researchers found that certain anti-HIV drugs do not appear to make their way to the male genital tract in sufficient concentrations to suppress the virus completely. Some of the previously treated men in this study had drug-resistant mutant strains of HIV in their semen, and in some cases these strains were different than the drug-resistant strains in the men's blood. According to Eron, this research indicates that, "if men-and presumably women-do not have adequate suppression of their virus...they are very likely to shed drug-resistant strains of HIV in their genital secretions. And that is virus that's available for transmission."

The researchers suggested that in the future, physicians may need to target the blood and the genital tract separately when making decisions about antiretroviral treatment, noting that "therapeutic strategies that fully suppress HIV-1 in the genital tract should be a public health priority." It is possible that regimens of the future will include specific drugs that penetrate different body compartments.

Gupta reported that the HIV strains in the semen and blood of the persistent shedders in his 1999 CROI study were very similar, while the semen and blood strains of the intermittent shedders were distinct. He suggested that HIV in the semen of the latter group of men may originate from a different site (possibly the prostate) than HIV in the blood. Echoing Eron, Gupta said that physicians "may be able to target treatments to this site and suppress replication of viral strains which may be different than the ones found circulating throughout the body. This is absolutely crucial to stemming transmission of the virus."

Other research lends weight to the compartmentalization hypothesis. At the February 1999 CROI, L. Ping and colleagues reported that they had found genetically distinct strains of HIV in the blood and semen of 50% of Malawian men studied. Timothy Schacker, MD, and colleagues found similar discordance in U.S. men examined near the time of primary infection.

At the September 1997 ICAAC, S. Phipott and colleagues reported on a case in which an HIV positive woman had a CVL viral load of 2,500,000 copies/mL compared to a blood viral load of 180,000 copies/mL. Genetic analysis showed that the HIV strains in her blood and genital fluid were distinct, with the HIV in the latter being more genetically varied. G. Fang reported similar findings at the same conference, adding that patterns of drug resistance were different in the blood and cervical/vaginal fluid, with fewer drug-resistant mutations in the CVL than in the blood.

Does Anti-HIV Therapy Reduce Infectiousness?

Studies have not yet definitively shown whether HAART reduces the likelihood of transmitting HIV, but there have been indications that treatment that reduces viral load in the semen and cervical/vaginal fluid may contribute to a decrease in infectiousness. In the September 12, 1994 issue of the Annals of Internal Medicine, M. Musicco and colleagues reported that men treated with AZT were 50% less likely to transmit HIV to their female partners than untreated men. AZT has also been shown to reduce vertical transmission from mother to infant by as much as two-thirds (see "Strategies for Preventing Late-Term Vertical HIV Transmission," BETA, January 1999). As previously noted, it is likely that more effective combination regimens that penetrate into the genital compartment will have an even greater impact.

At the February 1999 CROI, Jonathan Kaplan, MD, of the CDC reviewed what is known about the effect of antiretroviral therapy on transmission. Kaplan suggested that people taking potent antiretroviral drugs are likely to be less infectious. However, the availability of HAART has so far not had an effect on HIV incidence (frequency of new infections). Kaplan attributed this in part to the fact that newly infected people with high viral loads often do not know that they are infected, and thus are not receiving treatment-and often are not practicing safer sex. In other words, those most likely to transmit HIV may not be receiving antiretroviral therapy.

It is not completely clear what effect antiretroviral therapy has on semen and cervical/vaginal fluid viral loads. In the August 1997 issue of AIDS, Eron and colleagues showed that two- or three-drug treatment regimens reduced HIV levels in the semen by 90%, and that reductions in semen viral load parallel drops in blood plasma viral load. Several other research teams have also reported that HAART reduces semen viral load. Kenneth Mayer, MD, and colleagues reported at the September 1997 IDSA meeting that 60% of the 22 men in their study had detectable semen viral load levels before the initiation of HAART, compared to 20% after starting therapy. J. Evans and colleagues reported at the February 1998 CROI that a "distinctly higher proportion" of men on antiretroviral therapy had an undetectable semen viral load compared to men not on therapy. Cu-Uvin's group found that 88%-94% of women taking anti-HIV therapy (depending on the type of treatment) had an undetectable CVL viral load, in contrast to 61% of women not on treatment

A. Pereira and colleagues presented findings at this year's CROI that suggest that it may take longer after initiating antiretroviral therapy for viral load to be reduced in the semen compared to the blood. Blood viral load became undetectable in all of the nine men in this study by day 14 after starting treatment, but semen viral load did not become undetectable in all nine men until day 56.

Recent research suggests that even the most effective HAART does not completely eliminate HIV from the semen and cervical/vaginal fluid. Roger Pomerantz, MD, and colleagues reported in the December 17, 1998 issue of the New England Journal of Medicine that they were able to detect proviral HIV DNA in the semen of men with undetectable blood viral loads. Proviral DNA is HIV genetic material integrated into a human host cell chromosome. Using a PCR assay, the researchers found this type of genetic material in latently infected semen cells of four of seven men who were taking HAART and had undetectable blood plasma viral loads. The researchers were able to trigger HIV from the semen of two of these men to produce virus capable of infecting other cells. The semen HIV strains detected by Pomerantz's group were M-tropic-the type most likely to be transmitted sexually-and the virus could potentially be reactivated in the body of the recipient. Pomerantz concluded that "HIV-1-infected men on HAART without detectable levels of blood plasma HIV-1 RNA could still be a source of sexual transmission.... Even people who are doing well on the new treatment regimens should use condoms and practice safe sex."

Ashley Haase, MD, and Schacker underscored this point in an editorial in the same journal. They called Pomerantz's research "one of the most iron-clad pieces of evidence that people on [antiretroviral] drugs can still transmit the virus." Haase and Schacker advocate an education campaign to counter the "widespread myth" that HAART eliminates the risk of HIV transmission.

Cu-Uvin and colleagues similarly found that HIV genetic material could still be detected in the genital secretions of 3% of HIV positive women who were taking HAART and had undetectable blood viral loads.

While it is increasingly clear that blood viral load measurements are not a reliable indicator of viral load in genital fluids, it is not known how low a person's semen or cervical/vaginal viral load must be to render him or her unable to transmit HIV. In the case of mother-to-child HIV transmission, while treatment with AZT and lower blood viral load levels are associated with decreased risk of transmission, there is so far no blood viral load level below which transmission is not known to occur. The same may be true for sexual transmission.

Transmission of Drug-Resistant HIV Strains

The widespread use of antiretroviral drug regimens-which are complicated and often difficult to adhere to-encourages the development of drug-resistant strains of HIV. Becoming infected with a drug-resistant strain means that the newly infected person has fewer effective treatment options.

Frederick Hecht, MD, and colleagues made headlines around the world last summer (1998) when they reported at the 12th World AIDS Conference that HIV that is resistant to multiple antiretroviral drugs could be transmitted sexually. Previous studies had shown that HIV strains that were resistant to nucleoside analogs and non-nucleoside reverse transcriptase inhibitors (NNRTIs) could be transmitted via sex; Hecht's report, and a similar report from Switzerland, added protease inhibitors to the list. Hecht's group described an HIV positive man in San Francisco who had been treated with multiple antiretroviral drugs and who transmitted multidrug-resistant HIV to his male partner through unprotected anal intercourse. The transmitted strain was resistant to AZT, 3TC (Epivir), indinavir (Crixivan), nelfinavir (Viracept), ritonavir (Norvir), and saquinavir (Fortovase). The report appeared in the July 30, 1998 issue of the New England Journal of Medicine (see "Selected Highlights from the 12th World AIDS Conference," BETA, October 1998).

Three recent findings announced at the February 1999 CROI shed light on the extent of drug resistance in people newly infected with HIV. Susan Little, MD, and colleagues reported on genotypic and phenotypic analysis of HIV from 69 people with primary HIV infection in Boston, Dallas, Denver, Los Angeles, and San Diego; none had received antiretroviral drugs for more than seven days. The researchers found that 28% of the HIV strains had at least some degree of resistance to at least one drug (3% to nucleoside analogs, 17% to NNRTIs, and 13% to protease inhibitors). S. Wegner and colleagues reported an overall resistance mutation rate of 21% in 114 newly infected, treatment-naive persons; up to 3.2% (depending on the test used) showed resistance to drugs from all three classes. M. DePasquale and colleagues reported on the first discovery of a protease inhibitor-resistant mutant strain of HIV in semen; Eron and others had previously shown that semen contained HIV strains resistant to nucleoside analogs and NNRTIs.

In an editorial accompanying Hecht's report in the New England Journal of Medicine, Oren Cohen, MD, and Anthony Fauci, MD, stated that the transmission of multidrug-resistant HIV "should serve as a wake-up call on multiple fronts in the battle against HIV," and that "the availability of HAART cannot be relied on as primary prevention." They concluded that "since HIV may be compartmentalized in genital tract secretions, even persons with undetectable levels of [blood] plasma viremia may be capable of transmitting the virus if they do not take safe sex precautions."

What About Reinfection?

Many HIV prevention educators recommend that HIV positive persons should practice protected anal and vaginal intercourse even with others who are HIV positive because of the possibility of reinfection with new, possibly more virulent or more drug-resistant strains of HIV.

However, there is little research to indicate that reinfection actually occurs, although this would be difficult to document due to the existence of multiple strains ("quasispecies") of HIV in some infected people. There is no mention of reinfection in three years' worth of ICAAC or CROI conference abstracts or in AIDSLINE, a compendium of HIV/AIDS-related medical articles. Jay Levy, MD, of the University of California at San Francisco (UCSF) writes in his book HIV and the Pathogenesis of AIDS that "the prevalence of infection by more than one strain of HIV-1 in vivo is unknown, but is most probably rare." Likewise, Bernard Branson, MD, of the CDC, who has studied the phenomenon, suggests that "reinfection, if it occurs, is rare." However, Branson notes that "the tools necessary to document [reinfection] have only recently become available."

Some research does exist concerning superinfection of host cells by more than one strain of HIV. It appears that initial infection with the first strain tends to interfere with infection by subsequent strains by blocking host cell receptors.

There are also a few reports of dual infection with two strains of HIV in a single person. This phenomenon was once thought to be rare, but recent research suggests that it may be more common than previously realized. Patricia Fultz, MD, and colleagues reported in the April 1998 issue of the Journal of Virology and at the September 1998 ICAAC that PCR tests of HIV genetic material from the blood and lymph nodes of chimpanzees showed that a second strain of HIV could be detected in a superinfected individual, but only within the first six weeks after exposure to the new strain. The researchers were able to detect the second strain in all the chimpanzees that were superinfected. Because the first strain was predominant, this determination required a strain-specific PCR test; standard PCR tests tend not to detect minority strains, and thus may fail to reveal dual infection. Fultz suggested that the first infection somehow kept the subsequent strain under control.

At the February 1999 CROI, J. Mellquist and colleagues presented two abstracts on dual infection with HIV subtypes A and D in a Ugandan mother and child. Standard PCR tests tend to detect only a single, majority strain, but the researchers were able to detect the second strain using multiple PCR tests with different primers. Based on the similarity of strains in the mother and infant, the researchers believe that both strains were transmitted perinatally.

In summary, it is not clear whether reinfection is a normal occurrence or whether infection with one strain of HIV protects against infection with new strains. Although reinfection with new strains of HIV has not been definitively shown to occur, people with HIV remain at risk for other STDs through unprotected sex. STDs can be especially dangerous for immunocompromised people, since individuals whose CD4 counts have fallen-even if they have subsequently risen again due to HAART-may not have enough of the immune system cells necessary to combat sexually transmitted organisms. In addition, STDs may promote an increase in HIV replication, leading to higher viral loads.

Do Better Drugs Mean Less Caution?

In addition to the impact HAART may have on the biology of HIV transmission, the availability of effective treatments may also affect risk behavior by influencing people's perception of the gravity of HIV disease. If people believe that HIV is a chronic, manageable illness and no longer a death sentence, will they be less likely to take precautions to prevent infection?

According to a study by Stephen Kravcik, MD, and colleagues published in the October 1, 1998 issue of the Journal of Acquired Immune Deficiency Syndromes and Human Retrovirology, "a small but significant proportion of people with HIV believes that safe sex is less important for those receiving antiretroviral therapy." The study included 147 HIV positive persons. Nineteen percent of those taking HAART believed there was less need for safer sex practices, and 20.4% believed that the risk of transmission was reduced. Kravcik suggested that transmission of HIV may increase "if changes in knowledge and attitudes toward HIV risk lead to a change in sexual and drug use behaviors and practices." He cautioned that the use of terms such as "undetectable" to describe viral load may lead people to underestimate the risk of HIV transmission by people on HAART. The same caution applies to speaking about the eradication or elimination of HIV-a prospect that does not appear possible with current therapies.

Results presented at the American Public Health Association (APHA) conference in November 1998 suggest that the use of combination anti-HIV therapy is associated with increases in unprotected sexual activity. Ralph DiClemente, PhD, and colleagues reported that there was a "trend towards reduced condom usage" among sexually active HIV positive persons on HAART. Participants in this study were recruited from six public clinics in Alabama. The researchers found that people taking regimens containing protease inhibitors were approximately 2.5 times more likely to engage in unprotected sex than those not taking the drugs. People on HAART were also more likely to state that they never used condoms. The difference was more pronounced in certain subpopulations. Among men who have sex with men, those on HAART were 3.8 times more likely to report that they used condoms less than half the time and 5.4 times more likely to say that they never used condoms. Such differences were not seen in heterosexual men and women, leading the researchers to speculate that perhaps gay and bisexual men had had more exposure to overly optimistic media reports about the benefits of HAART.

Researchers are also studying whether the availability of postexposure prophylaxis (PEP) might influence people to engage in more risky sexual behavior. Currently there are no clinical trials showing that PEP is effective for sexual exposures to HIV; however, a case-control study of health-care workers in the U.S., U.K., and France (published in the December 22, 1995 issue of Morbidity and Mortality Weekly Report) showed that AZT reduced the risk of HIV infection following a needlestick injury by 79%. PEP may reduce viral load during primary HIV infection, but this has not yet been demonstrated. Joshua Bamberger, MD, of the San Francisco PEP Project is adamant in emphasizing that "there is no morning-after pill" to prevent HIV infection (see "Postexposure Prophylaxis: Still More Questions than Answers," BETA, October 1998).

Increasing Sexual Risk Behavior

Research trends appear to confirm that unprotected anal sex is increasing among men who have sex with men. In the October 1, 1998 issue of the Journal of Acquired Immune Deficiency Syndromes and Human Retrovirology, Katz and colleagues stated that self-reported rates of unprotected sex among gay and bisexual men was no lower in 1994-1995 than in 1992-1993. Katz concluded that "our data clearly show that the problem of high-risk behavior among young [men who have sex with men] continues unabated."

Researchers from the CDC recently reported that rates of unprotected anal sex appear to be on the rise among gay and bisexual men. The January 29, 1998 issue of Morbidity and Mortality Weekly Report published results of a survey conducted by the Stop AIDS Project in San Francisco. Surveys of 21,857 gay and bisexual men revealed increased rates of unprotected anal intercourse. The percentage of men who reported unprotected anal sex increased from 30.4% in 1994 to 39.2% in 1997. The percentage of men reporting unprotected anal sex with multiple partners increased from 23.6% in 1994 to 33.3% in 1997. Younger men in particular were more likely to have multiple partners and use condoms less frequently. Richard Steketee of the CDC's Division of AIDS Prevention suggested that news about rapid advances in anti-HIV treatment may have led some men to perceive that it is no longer necessary to take precautions to prevent infection.

The same report revealed that rates of rectal gonorrhea in men seen at San Francisco STD clinics nearly doubled between 1994 and 1997. According to the San Francisco Department of Public Health, rectal gonorrhea rates increased from 21 cases to 38 cases per 100,000 men in that period. The increase represents a reversal of a three-year decline; however, the current rate remains far below that of the early 1980s, when rectal gonorrhea cases reached 5,000 or more per year. Because rectal gonorrhea is transmitted through unprotected anal sex, it is viewed as a marker for sexual behavior that can also transmit HIV.

Research from the East Coast also reveals some disturbing numbers. A study conducted by Mary Ann Chiasson, MD, and colleagues from the New York City Health Department presented at the February 1999 CROI showed that HIV prevalence rates among young men who have sex with men in that city are high. The results of the Young Men's Survey indicate that 12% of the 545 surveyed gay and bisexual men aged 15-22 are HIV positive. Forty-six percent of the young men surveyed self-reported unprotected anal sex in the past six months.

Prevention in the HAART Era

As the HIV/AIDS epidemic nears its third decade, many people have begun to reexamine the effectiveness of HIV prevention messages that focus on "use a condom every time." It has become now clear that HIV is a long-term problem, and that prevention measures must be timely and relevant, as well as sustainable over the course of years-and possibly lifetimes.

Many people find it difficult to maintain safer sex behaviors for long periods of time. Several gay men have written eloquently about the profound loss they feel at having to give up semen exchange and sexual contact without barriers. Recent media attention has focused on the phenomenon of "barebacking"-intentional, eroticized anal sex without condoms-among gay men, sometimes in lurid and sensationalistic detail. While barebacking as a culture and identity involves only a small minority of gay men, surveys show that many more men occasionally "slip up" by engaging in unprotected sex despite the best intentions.

News stories about effective new treatments for HIV, as well as the possibility of PEP, have combined with a host of other psychological and social factors to influence sexual behavior in the era of HAART, and AIDS prevention education has evolved to meet these new challenges.

HIV/AIDS prevention has become more sophisticated, eschewing simplistic messages and focusing on the many factors involved in maintaining sexual behavior change. There has been a more realistic assessment of the relative risks of various sexual activities (for example, oral sex), and an attempt by some people to introduce "harm reduction" measures to maximize the safety of whatever sexual behaviors one practices. Safer sex education is now often tailored differently for HIV positive and HIV negative persons, and "negotiated safety" (a mutual decision to engage in unprotected sex) has become a recognized option for HIV negative people in long-term, committed relationships. Perhaps most notable, the focus has shifted from "one size fits all" recommendations to individualized risk assessment and decision-making.

For many women, including women in developing countries where the epidemic is spreading most rapidly, the issue is often one of control. Women may not be able to insist that their male partners use condoms. The recently developed internal "female" condom is an important step toward female-controlled HIV prevention, and is also an additional tool for men who practice receptive anal sex. The development of safe and effective microbicides that can help prevent HIV infection in women and men who practice receptive intercourse is a critically important goal. Ultimately, a safe and effective vaccine is the best hope for HIV/AIDS prevention.

Programs to prevent and treat STDs are likely to have a profound impact in reducing the rate of HIV transmission in many populations. Encouragement of HIV testing-and expanded use of tests that can detect HIV before antibodies are present in the blood-combined with early treatment and safer sex education may help to reduce transmission during the earliest stages of infection when people may be most likely to transmit the virus. In a review article on sexual transmission of HIV in the April 10, 1997 issue of the New England Journal of Medicine, Rachel Royce, PhD, and colleagues concluded that "breaking the chain of transmission during the period of primary and early infection is potentially the most effective intervention." It remains to be seen what contribution HAART will make toward this goal.

Still No Cure

Despite promising news about antiretroviral therapy, it is clear that there is still no cure for HIV disease. Current drug regimens are inconvenient, expensive, and entail a plethora of side effects and drug-drug interactions. Although PEP may be possible for some people who have been sexually exposed to HIV, there is no "morning-after pill." As has been the case since the early 1980s, prevention of HIV transmission remains the best hope for controlling the AIDS epidemic.

Liz Highleyman is a freelance writer based in San Francisco and a former member of the BETA editorial staff.

Thanks to Joe Headlee, Eric Rofes, and Michael Discepola for their helpful comments on this article.

Centers for Disease Control and Prevention. Increases in unsafe sex and rectal gonorrhea among men who have sex with men, San Francisco, 1994-1997. Morbidity and Mortality Weekly Report 48(3): 45-48. January 29, 1999.

Cohen, O. and Fauci, A. Transmission of multidrug-resistant human immunodeficiency virus-the wake-up call. New England Journal of Medicine 339(5): 341. July 30, 1998.

Cu-Uvin, S. and others. Genital tract HIV detection in women: association with plasma viral load, CD4 cell count, antiretroviral therapy, mode of transmission. 6th Conference on Retroviruses and Opportunistic Infections. Chicago. January 31-February 4, 1999. Abstract 264.

Eron, J. and others. Resistance of HIV-1 to antiretroviral agents in blood and seminal plasma: implications for transmission. AIDS 12(15): 181-189. October 22, 1998.

Fultz, P. Are infections with two (or more) HIV-1 strains really rare events? 38th Interscience Conference on Antimicrobial Agents and Chemotherapy. San Diego. September 24-27, 1998. Abstract I-33.

Gupta, P. and others. Mechanism of shedding of HIV-1 in semen. 6th CROI. Abstract 302.

Haase, A. and Schacker, T. Potential for the transmission of HIV-1 despite highly active antiretroviral therapy. New England Journal of Medicine 339(25): 1846-1847. December 17, 1998.

Hecht, F. and others. Sexual transmission of an HIV-1 variant resistant to multiple reverse transcriptase inhibitors and protease inhibitors. New England Journal of Medicine 339(5): 307. July 30, 1998.

Howard, D. Do you need to use condoms for positive-positive sex? Bay Area Reporter 29(10). March 3, 1999.

Kaplan, J. Highly active antiretroviral therapy: is it impacting transmission of HIV? 6th CROI. Abstract S30.

Katz, M.H. and Gerberding, J.L. Postexposure treatment of people exposed to the human immunodeficiency virus through sexual contact or injection drug use. New England Journal of Medicine 336(15): 1097-1099. April 10, 1997.

Kravcik, S. and others. Effect of antiretroviral therapy and viral load on the perceived risk of HIV transmission and the need for safer sexual practices. Journal of AIDS and Human Retrovirology 19: 124-129. October 1, 1998.

Osmond, D. Sexual transmission of HIV. The AIDS Knowledge Base, Third Edition, HIV InSite version (http://hivinsite.ucsf.edu/akb). P.T. Cohen and others (eds). 1998.

Otten, R. and others. Evidence for cell-free virus in cervicovaginal lavage (CVL) specimens as an early indicator of productive infection following vaginal exposure in a pigtail macaque model for human retroviral (HIV-2) infection. 6th CROI. Abstract 303.

Pudney, J. and others. Pre-ejaculatory fluid as potential vector for sexual transmission of HIV-1. The Lancet 340(8833): 1470. December 12, 1992.

Rothenberg, R.B. and others. Oral transmission of HIV. AIDS 12(16): 2095-2105. November 12, 1998.

Royce, R.A. and others. Current concepts: Sexual transmission of HIV. New England Journal of Medicine 336(15): 1072-1078. April 10, 1997.

Vernazza, P.L. and others. Sexual transmission of HIV: infectiousness and prevention. AIDS 13(2): 155-166. January 1999.

Zhang, H. and others. Human immunodeficiency virus type 1 in the semen of men receiving highly active antiretroviral therapy. New England Journal of Medicine 339(25): 1803-1847. December 17, 1998.

Page last updated 5 October 1999