In the past decade, there has been increasing emphasis on addressing the HIV transmission behaviors of HIV-positive persons. As people with HIV live longer and healthier lives, their prevention needs are increasingly prominent in studies and discussions of HIV prevention. Despite increased prevalence of antiretroviral therapy use, HIV incidence has remained high in many communities, suggesting that prevention programs targeting only HIV-negative persons may be inadequate to curb the epidemic, and emphasizing the need to address HIV prevention among persons with HIV infection.

Perceived responsibility for prevention of HIV transmission is a key area of concern for HIV-positive individuals, who often report that they would like to have prevention discussions with their medical providers. Nevertheless, many providers do not discuss prevention: Recent studies find that between 25% and 30% of HIV-positive patients report never discussing safer sex with their health care providers,(1-3) and that persons at higher transmission risk are no more likely than those at low risk to report having received counseling.(4) In a study of prevention practices in a sample of 16 U.S. government-funded HIV clinics, 68% of the participants reported being sexually active in the preceding 6 months; however, only 25% of the participants reported having had a discussion about safer sex with their provider that day, of whom only 6% reported discussing specific sexual acts with their provider.(5)

The focus on integrating prevention into care has become a priority in the United States. In 2003, the Centers for Disease Control and Prevention (CDC) along with the Health Resources and Services Administration, the National Institutes of Health (NIH), and the HIV Medicine Association disseminated a set of guidelines titled "Incorporating HIV Prevention into the Medical Care of Persons Living with HIV".(6)

The goal of this chapter is to review factors related to risk of HIV transmission, to assist clinicians and other health care providers in assessing their patients' potential for transmitting HIV, and to provide information to help facilitate sensitive, nonjudgmental discussions with HIV-positive persons about their sexual behavior.

HIV infections rates among men who have sex with men (MSM) remain high, and MSM continue to account for the largest proportion of new HIV infections in the United States. In San Francisco, estimates of infection rates among MSM increased from 1.1/100 person-years in 1997 to 2.2/100 person-years in 2000. HIV incidence was 5.3% among MSM seeking sexually transmitted disease (STD) services.(7) These high rates are reported despite other evidence that HIV infectivity declined by as much as 60% among MSM in San Francisco following the introduction of highly active antiretroviral therapy (ART).(8) The CDC Behavioral Surveillance Survey, a probability-based survey of MSM in urban venues, reported HIV incidence rates ranging from 1.2% to 8% in urban settings in the United States.(9) The EXPLORE study of high-risk HIV-uninfected MSM in 6 U.S. urban centers enrolled between 1999 and 2001 and followed for up to 4 years reported an HIV incidence of 2.1%.(10) Young men of color may be at especially high risk for infection: a large urban study of MSM <22 years of age found that HIV infection rates among Latino and black participants were 2.3 times and 6.3 times higher, respectively, than those of white MSM.(11,12) The study by Valleroy et al also found that 82% of the HIV-positive young men were unaware that they were HIV infected, suggesting a need to emphasize frequent testing among this population.(11)

Among women, blacks and Latinas are at continued high risk for HIV infection.(13) It is estimated that 3 out of 10 new HIV cases in the United States will occur in women. Among new HIV cases reported in 2001, women accounted for 56% of cases among those 13-19 years of age, 36% of the cases in those 20-24 years of age, and 37% of all cases.(14) It appears that young women are at least as likely as young men to become HIV infected. Most striking in these statistics is the fact that two thirds of women becoming infected with HIV in the United States are black, and nearly 20% are Latina. Among 33 states with HIV reporting, 50% of newly diagnosed HIV infections occurred in blacks.(15)

Among populations of injection drug users (IDUs), HIV incidence appears to be in decline in some U.S. cities.(16,17) The availability of needle exchange programs, clean injection equipment, and community outreach to IDUs has contributed to this decline.(18) Recent analyses of HIV transmission behavior among IDUs suggest sexual risk behavior as the main contributor driving new infections, and that HIV seroconversion among IDUs is now being fueled by men having sex with men, and by women trading sex for money or drugs.(19,20) Demographic trends of HIV incidence among IDUs are similar to those among non-IDU MSM and women, with HIV becoming overrepresented among black and Latino IDU populations.(21) Among women, 70% of those who reported their risk as heterosexual contact identified their male partners as IDUs, and among male IDUs, 54% were black.(15)

Although a desire to protect sexual partners is a major reason for condom use among all HIV-infected risk groups, research has consistently demonstrated that many HIV-infected persons continue to engage in high-risk sexual behaviors that may transmit HIV, despite knowledge of their HIV infection. Across studies of HIV-positive MSM, women, and IDUs, between 17% and 38% report unprotected vaginal or anal intercourse (many as recently as their last sexual encounter) with partners who are HIV negative or of unknown HIV status.(22-27)

High-risk sexual behavior among HIV-infected persons is not limited to interactions with casual or anonymous partners. Multiple studies have found that safer sex precautions are less likely to be adopted in relationships characterized by affection and in ongoing sexual relationships than in casual or transient partnerships.(28-31) This pattern has been found not only in the case of monogamous serodiscordant male couples, but also among affectionate relationships that are not mutually exclusive and in which partners do not know each other's serostatus.(32,33) In one analysis of couples in serodiscordant relationships, 31% reported unprotected anal sex with their primary partner at least once in the past 12 months. This behavior may put heterosexual women especially at risk, as they are more likely than other groups to have a single high-risk partner rather than multiple partners.(34)

Patterns of risk behavior also may vary by the stage of HIV infection and knowledge of infection status. A metaanalysis of high-risk sexual behavior among persons infected with HIV found that prevalence of high-risk behavior was 68% lower in HIV-positive persons aware of their status than in persons unaware of their status.(35) Several studies have suggested that HIV-positive persons go through a period of sexual abstinence as they adjust to their infection status, but later resume their sexual activity.(36,37) However, one study of newly infected persons found that 11% reported unprotected insertive anal sex and 26% reported unprotected receptive anal sex with unknown-serostatus or HIV-negative partners within a 6-month period after infection, suggesting a need to address behavior change early following the diagnosis of HIV infection.(38)

Counseling interventions have been shown effective in reducing self-reported risk behavior of HIV-infected persons, with studies indicating that interventions lead to significant reductions in both unprotected anal intercourse and numbers of sexual partners.(27,39,40) Clinician-based counseling to encourage behavior change has been highly effective in reducing tobacco and alcohol use.(41) Studies examining the efficacy of medical provider counseling have shown reductions in sexual risk behavior among specific groups of HIV-infected patients, although limitations in study design prevent definitive conclusions about the effect of such interventions.(42-45) Richardson et al found that brief prevention counseling delivered by a physician was effective in reducing unprotected anal or vaginal sex among HIV-infected individuals who reported 2 or more partners.(45) In addition, the degree of behavior change necessary to reduce HIV infection rates remains to be determined. In the EXPLORE study, for instance, despite a 20% reduction in serodiscordant unprotected intercourse in the intervention arm, HIV seroincidence was not significantly different among persons in the intensive behavioral intervention arm compared with persons in the standard counseling and testing arm.(10) Research in this area is ongoing and will provide further evidence and direction to clinicians on the most effective approaches to use with patients engaging in high-risk behavior. In particular, the acceptability and feasibility of conducting prevention intervention among HIV-positive patients in a nonresearch context has yet to undergo careful evaluation.

When engaging patients in discussions about reducing behaviors with the potential to transmit HIV, clinicians should consider adapting the conceptual framework from the general theory of behavior change used in the above trials, asking patients:

• How important is it to you to reduce your risk behavior?

• How confident are you that you can change your behavior?

• How ready are you to change your behavior?

In contrast to counseling that promotes behaviors to improve a patient's own health, counseling HIV-positive patients about the risk of sexual transmission emphasizes changing behavior to protect others from infection. This may present a conflict for the clinician, yet HIV-positive patients themselves may derive health benefits from changes in transmission risk behaviors associated with sex or drug use. These potential health benefits include reduced likelihood of contracting other STDs and, possibly, reduced potential for superinfection with drug-resistant HIV. By engaging patients in frank discussions about the risks of HIV transmission, their current risk behavior, and factors associated with increased risk, providers can begin to tailor risk-reduction strategies based on patients' needs and desires. The following sections address issues that should be discussed with patients, over a series of visits, to assess risk fully and to help patients develop strategies for behavioral change. After the initial assessment, the provider should reassess frequently to determine if behaviors have changed or if there are new factors influencing risk.

Clinicians have the responsibility to discuss patients' medication regimens and viral load in the context of transmission risk. The availability of effective HIV treatments in the United States has been correlated with increasing rates of HIV infection.(46) ART also has been linked to biologic markers of risk behavior: Among persons diagnosed with AIDS, receipt of ART was independently associated with a 4-fold increased likelihood of being diagnosed with an STD.(47) On the other hand, metaanalyses have shown that persons receiving ART are no more likely than HIV-infected individuals not on ART to engage in high-risk behavior.(48)

Compelling and widely cited evidence indicates that, as plasma viral load increases, the risk of HIV transmission increases, but the correlation between plasma viral load and risk of sexual transmission may vary from one situation to another. Data from a study of treatment-naive heterosexual couples in Uganda indicate that, for every 2.5 log increase in plasma viral load, the risk of transmission increases approximately 10-fold.(49) In the setting of ART, however, correlation between viral load and risk of HIV transmission, and in particular the potential for transmission of resistant strains, remains to be determined. Furthermore, although some studies have shown a moderate association between plasma viral load and the viral load in genital secretions, others have found no significant association.(50,51) Nevertheless, several studies have found that, compared with persons who do not believe low viral load reduces transmission risk, those who do hold this belief engage in higher-risk behaviors.(39,52-55) A metaanalysis of research on the relationship between viral load and risk among HIV-positive persons found that neither taking medications nor having an undetectable viral load was related to risk; however, believing that taking ART or having an undetectable viral load protected against transmission was related to risk regardless of serostatus.(48) Another analysis found that many men with serodiscordant partners discuss viral load specifically to guide sexual risk behavior.(56)

These observations illustrate the importance of eliciting patients' understanding of what being on medication means with regard to transmission risk, the impact of viral load on transmission risk, and their sexual practices in relationship to treatment and viral load results. There is unlikely to be any plasma viral load level below which it would be absolutely "safe" for patients to engage in high-risk behaviors, and clinicians should reinforce this message with their patients.

Although the absolute risk of infection from any given exposure is low, providers may find it most useful to discuss the relative risk of infection of specific behaviors with their patients. Patients often are surprised to learn that unprotected insertive anal intercourse is not orders of magnitudes less risky than receptive anal intercourse, but "only" about 4 times less risky. Although oral sex is much less risky than unprotected anal sex, there is strong evidence that some men have become infected through receptive oral sex.(64)

Framing of relative risks may be particularly appropriate for HIV-positive persons who perceive their sexual activity as low risk. For example, in reviewing such information, an HIV-infected male who has used condoms only as the insertive partner (in order to protect his partners) may be convinced to use condoms also when having receptive anal sex. Ranges of absolute per-contact risk estimates from a variety of studies are shown in Table 1.(65,66)

Note that the absolute per-contact estimates should be used with caution, if at all, from a prevention perspective on an individual basis, given the multitude of variables that can influence the infectiousness of a host and susceptibility of the partner. Higher viral loads, the presence of an STD in either partner, the presence of trauma, and a host of complicated genetic and immune factors influence the risk of transmission for any of the above acts. For a given high-risk exposure, usually it is not possible to predict the probability of HIV transmission with any accuracy.

Although such data may relieve anxiety around recent exposures among HIV-negative persons and help them decide whether to receive interventions such as postexposure prophylaxis, HIV-infected persons who continue to engage in high-risk behaviors should be asked to consider the cumulative risk of their behaviors: With repeated exposures over time, it is likely that they will infect a partner.

Lack of knowledge of partner serostatus may lead some HIV-infected individuals to engage in high-risk behaviors due to incorrect assumptions that their partner also is HIV infected. One study reported that up to 30% of HIV-infected adults had not informed any partners of their HIV status.(67) Another study of HIV-positive persons in clinical care found that 42% of MSM, 19% of heterosexual women, and 17% of women reported at least 1 sexual contact without disclosure.(68) HIV status may be revealed only after sexual activity, when patients take their medications or allude to some HIV-related complication. Many patients use nonverbal cues to assess HIV status, and may assume that if status is not discussed, their partner is of the same serostatus. Disclosure of serostatus is high between primary partners,(69-75) but lower with nonprimary partners.(76,77)

Whether disclosure alone is sufficient to change risk behavior is debatable. One study of an intervention emphasizing disclosure as a way to reduce risk found this strategy to reduce risk behavior.(39) Yet, for some persons, disclosure alone as a risk-reduction strategy may not be effective. Data demonstrate that unsafe sex is high among known serodiscordant couples.(62,78) In addition, due to gaps between HIV testing and the antibody window period, in which recently infected individuals may be highly infectious but appear negative on HIV testing, some persons may incorrectly identify themselves as HIV negative. The risk of unprotected sex during this period significantly increases. Finally, HIV-positive persons may feel uncomfortable revealing their status to new or casual partners, for fear of rejection, other discrimination, or even violence.(79,80)

Recently, the concept of "serosorting," defined as deliberately limiting some or all sexual risk activities to encounters with partners of one's own serostatus, has received increased attention among researchers, and has been hypothesized to account for the fact that increases in STD rates in some communities have been accompanied by stable or even decreasing HIV infection rates.(26,81,82) In theory, if persons were 100% adherent to serosorting for all sexual activities, sexual transmission of HIV would stop. However, serosorting depends on the assumption that persons know their current HIV status, as well as that of their partners. Persons recently infected--and who therefore may be the most highly infectious due to high viral loads--may inadvertently put others at risk by serosorting with other persons who test HIV negative. The efficacy of serosorting as an HIV prevention strategy among either HIV-negative or HIV-positive persons has not been evaluated rigorously.

Given the above considerations, providers should evaluate the risk-taking of their HIV-infected patients and determine whether disclosure or serosorting is likely to lead to behavioral change, taking into account type(s) of partners, relationship status, and interpersonal dynamics of partnerships.(80,83,84)

Recreational drug and alcohol use is associated with increased risk behavior.(85,86) Drugs repeatedly linked to increased risk behavior across multiple risk groups include amphetamines (both injected and noninjected), amyl nitrite ("poppers"), and cocaine. Most studies also demonstrate a relationship between high levels of alcohol use and high levels of sexual risk.(87-90)

Drugs may have a disinhibitory effect on individuals who would otherwise have protected sex. By contributing to miscommunication of one's own or one's partner's serostatus, drug use also may lead to unknowing engagement in riskier behavior. Drugs also may be used intentionally prior to engaging in high-risk behavior in order to reduce anxiety or inhibitions about potential disease transmission, termed "cognitive escape" by some.(91) Health care providers should inquire about their patients' recreational drug and alcohol use, discourage use during sexual activity, and make appropriate referrals for drug and alcohol treatment and counseling. Whereas optimal formats for substance-use counseling for HIV-infected persons merit further investigation, most studies indicate that persons who enter and remain in drug treatment reduce their HIV risk behavior.(92,93)

In addition to use of illicit recreational drugs, use of drugs prescribed for erectile dysfunction has been associated independently with unprotected anal sex with partners of unknown or opposite HIV serostatus.(85,94) Providers often prescribe sildenafil (Viagra), tadalafil (Cialis), or vardenafil (Levitra) to treat medically diagnosed sexual dysfunction associated with HIV.(95) Among some recreational drug users, it is likely that these erectile dysfunction treatments are being used to offset sexual dysfunction due to concomitant use of "club" drugs such as ecstasy, amphetamine, and ketamine, and there are reports of sildenafil being distributed at parties and clubs. The combination of sildenafil with methamphetamine use may lead to especially high-risk sexual behavior.(96) These findings suggest that health care providers should reinforce safer-sex messages prior to prescribing treatments for erectile dysfunction, and that HIV prevention programs should address potential risks associated with recreational use of these drugs.

Environmental factors influence high-risk sexual behavior. MSM and other risk groups are more likely to engage in high-risk sexual encounters among partners met in parks, sex clubs, and bathhouses.(97,98) Several studies have shown that the Internet is linked to high-risk sexual encounters.(99,100) In San Francisco, an outbreak of syphilis was traced to contacts established using the Internet.(101,102) In Denver, among persons presenting to public counseling and testing sites, clients seeking sexual encounters through the Internet reported greater risk behavior than non-Internet sex-seekers.(103,104) As part of risk assessments, providers should inquire about whether patients meet partners over the Internet or engage in sexual activity in public venues. Patients who report engaging in such activities may be considered at high risk of transmitting HIV and warrant further risk-reduction counseling and regular STD screening.

As might be expected, HIV-positive men who have high-risk sex with multiple partners show greater sexual preoccupation and greater behavioral disinhibition than those who do not engage in unprotected sex.(105-108) The relationship between depression and sexual behaviors of HIV-positive persons is inconclusive, with the primary work conducted prior to the availability of effective ART. Some researchers have reported an increase in unprotected intercourse among depressed HIV-positive patients, although others have found a reduction in sexual behavior among depressed patients.(109) A recent review concluded that there may be a relationship between risk and depression for HIV-positive IDUs.(110) Further studies have not defined a clear relationship between depression and risk behavior.(111,112) However, some studies are indicating that treating depression may decrease sexual risk behavior. Clearly, this question merits further investigation, given the range of treatments available for depression. For now, it appears best to screen patients for depression or other mood disorders, treat the underlying condition, and continue to assess potential high-risk behavior.

Discussions of safer sex need not be limited to emphasizing the need to protect others from infection; providers also should encourage patients to adopt safer behaviors to protect themselves from other STDs and from possible superinfection (reinfection by a second strain of HIV) with drug-resistant virus.(109)

Superinfection

Although several reports have confirmed that superinfection with HIV does occur,(113,114) the clinical significance remains to be determined. Several studies are in progress to determine whether superinfection is associated with poor health outcomes. Clinicians should ask their patients about their understanding and concerns regarding superinfection, and patients should be counseled regarding the potential risks to themselves and their partners of unprotected sexual acts with the potential to transmit antiretroviral-resistant strains. Despite ongoing debate within the scientific community about the clinical and epidemiologic impact of superinfection, many HIV-infected persons believe it occurs: In one sample of HIV-positive MSM in San Francisco, 74% reported that concern about superinfection led them to practice safer sex.(115) No studies have examined beliefs about superinfection among HIV-positive IDUs or women.

STDs

Outbreaks of syphilis and increases in gonorrhea have been reported among MSM in San Francisco, Seattle, Los Angeles, and Boston.(42,101) Federal guidelines recommend that sexually active MSM be screened for syphilis annually, although more frequent STD screening (every 3 to 6 months) may be indicated for MSM at highest risk (eg, those with multiple anonymous partners or those having sex in conjunction with illicit drug use).(116-118) The San Francisco Department of Public Health routinely suggests syphilis screening every 6 months for sexually active MSM. Providers should strongly consider annual STD screening among all sexually active HIV-positive patients; one study showed that 7.5% of those screened in an urban HIV primary care clinic reported infection with gonorrhea or chlamydia in the prior 12 months.(23,42) A discussion about STD screening can also lead into a broader discussion about placing others at risk, given the overwhelming evidence that STDs enhance transmission of HIV.(116,119-121)

As with many other health-related behaviors, risk-taking is determined not only by individual-level or partner-specific variables, but also by a multitude of factors at the environmental, community, and societal level. Although the ability to influence these variables directly may be beyond the scope of most patient-provider visits, providers should acknowledge to patients the complexity of factors contributing to risk, and should be sensitive to the discrimination that many HIV-positive persons feel on a daily basis. Discrimination and marginalization involving groups most affected by HIV are associated with high rates of risk behavior. Homophobia, histories of childhood abuse, and lack of connection to the gay community have been associated with increased risk of HIV infection and risk behavior among MSM.(122,123) Poverty, discrimination, and violence are being recognized in the literature as risk factors for HIV infection and for lack of clinical care, particularly among communities of color.(124,125) These circumstances, which may exist for many individuals who are at risk for HIV, do not disappear once an individual is diagnosed as seropositive.

Several ongoing behavioral studies are examining the impact of individual risk-reduction counseling, as well as group counseling, including peer-led groups, on reducing risk behavior of HIV-positive individuals. Within the multisite NIH-sponsored HIV Prevention Trials Network, a variety of microbicides, including PRO 2000 and BufferGel, currently are being tested in clinical studies. As well, studies to determine the efficacy of the cervical diaphragm in decreasing HIV rates among women currently are under way. A trial of providing ART to the positive member of a serodiscordant couple, with the goal of reducing HIV transmission to the HIV-negative partner, is ongoing. The CDC-funded study Project MIX is a behavioral intervention targeting MSM substance users, designed to determine whether a risk-reduction approach to substance use and sexual risk behavior is effective in this population. The results of these trials may provide a variety of behavioral and biological methods for reducing the risk of HIV transmission.

Given the multitude of health issues facing HIV-positive patients, many clinicians may not feel that they have time to engage patients in discussions about high-risk behavior. Nevertheless, patient and community health depend on patient risk behaviors, which therefore must be assessed over the duration of the patient-clinician relationship. Providers should commit to regular discussion of sexual risk behavior with HIV-positive patients in the setting of HIV care. A comprehensive "prevention checklist" should include:

• Assessing the patient's current level of sexual behavior and its potential relationship to current mental and physical health.

• Emphasizing that undetectable plasma viral loads should not be equated with noninfectiousness.

• Assessing current use of barrier methods (eg, condoms), and discussing proper use and alternatives.

• Explaining the relative risks of various sexual acts, emphasizing that even "low-risk" acts may result in transmission.

• Screening annually for STDs, and more frequently if patients are at high risk for STD acquisition.

• Discussing frankly how the patient assesses partner serostatus, and how the patient meets partners.

• Assessing alcohol and recreational drug use, including erection-sustaining drugs, and the relation of drug use to sexual behavior.