HIV Epidemiology and Breakthroughs in Prevention 30 Years Into the AIDS Epidemic Susan P

IAS–USA Topics in Antiviral MedicineConference Highlights—Epidemiology and Prevention Volume 19 Issue 2 May/June 2011

HIV Epidemiology and Breakthroughs in Prevention 30 Years Into the AIDS Epidemic Susan P. Buchbinder, MD

Thirty years after the first AIDS cases were reported in the United States, the at particularly high risk. In examin- HIV epidemic continues to be heavily concentrated among men who have sex ing drivers of the epidemic in MSM, with men (MSM) in the United States. MSM are heavily impacted throughout numerous studies have shown lower most of the world and are the predominant risk group throughout the levels of reported sexual risk and drug Americas and Western Europe; heterosexuals are the predominant risk group use among black and Latino MSM in sub-Saharan Africa; and injection drug users predominate throughout than among white MSM. These racial Eastern Europe and Southeast Asia. In the United States, blacks and Latinos and ethnic disparities may arise as a continue to be disproportionately affected, despite overall advances in HIV result of differences in background testing and care. The 2011 Conference on Retroviruses and Opportunistic prevalence, patterns of intraracial mix- Infections focused on populations heavily impacted throughout the world: ing, prevalence of sexually transmit- adolescents, women, MSM, and serodiscordant couples. Several presentations ted infections, access to antiretroviral focused on the unique relationship between herpes simplex virus type 2 therapy, and rates of undiagnosed HIV (HSV-2) and HIV-1; although many opportunistic infections increase HIV infection, all of which may drive in- acquisition risk, HSV-2 is likely the only one whose effective prevention or creased rates in black and Latino MSM. treatment could substantially influence HIV infection rates, because of the high Millett and colleagues also present- prevalence and persistence of HSV-2. The 2011 conference also celebrated the ed data from 1214 black and Latino substantial advances made in the use of antiretroviral drugs for prevention of MSM enrolled in the Brothers y Her- HIV acquisition (eg, oral preexposure prophylaxis, topical microbicides) and manos study in New York City, Phila- transmission (eg, antiretroviral therapy). Further progress is also being made delphia, and Los Angeles (Abstract in evaluating other prevention strategies and their rollout, including male 131LB). Overall, 12% were HIV-sero- condoms, male circumcision, and HIV testing and linkage to care. positive and unaware of their infection, with a higher rate among black The US HIV Epidemic the United States who have ever been than Latino men (18% vs 5%, respec- tested for HIV has risen to 45%, and tively; P < .001). Among both groups, Mermin provided an overview of the the proportion of persons with AIDS having a low perceived risk of testing US epidemic and strategies for imple- diagnosed within 12 months of their HIV-seropositive and endorsing the menting high-impact prevention (Ab- first HIV-seropositive test has dropped belief that having sex with men of the stract 19). He reminded the audience to 32%. However, given limited re- same race or ethnicity reduces the risk of the substantial disparities in HIV sources, Mermin called for targeted of HIV acquisition were independently infection, with new infections more prevention strategies based on knowl- associated with being HIV-seroposi- than 40 times more likely to be in men edge of effectiveness, cost, scalability, tive and unaware. Among black MSM, who have sex with men (MSM) than and coverage of affected populations. having disclosed sexual identity to a in other men and women and more Millett further explored the US epi- health care practitioner, having health than 8 times more likely in blacks and demic in MSM, the only risk group insurance, and having fewer than 3 times more likely in Latinos than in in the United States in whom new 3 lifetime HIV tests were also indepen- whites. Tremendous strides have been infections continue to rise (Abstract dently associated with HIV-seroposi- made in prevention, with community- 69). Modeling suggests that even if tive–unaware status. initiated behavior change leading to an MSM and heterosexuals had similar Millett pointed to the need to ad- 89% reduction in the transmission rate numbers of sexual partners and rates dress the misperceived risk of HIV ac- per 100 HIV-infected persons, thereby of unprotected intercourse, incidence quisition, including the risk associated averting an estimated 350,000 new rates in MSM would be higher because with intraracial partnerships. He also HIV infections since the beginning of of higher background prevalence rates, highlighted the responsibility of prac- the epidemic. Through expanded HIV increased risk of anal versus vaginal titioners to offer frequent HIV tests to testing, the proportion of people in sex, and role versatility in which many MSM, as even black MSM who disclose MSM serve both insertive and recep- their risk to health care practitioners tive roles, thereby accelerating trans- and who have health insurance appear Dr Buchbinder is director of the HIV Re- search Section at the San Francisco Depart- mission through partner networks. to be at elevated risk of HIV seroposi- ment of Public Health and associate clinical MSM are disproportionately affect- tivity. Millett also called for multilevel professor of medicine and epidemiology at ed within all racial and ethnic groups, approaches to prevention and treat- the University of California San Francisco. and young black and Latino MSM are ment for MSM, including those using

38 Conference Highlights—Epidemiology and Prevention Volume 19 Issue 2 May/June 2011 individual-, interpersonal-, and struc- study of seroadaptive behavior among sented from a study conducted in Can- tural-level interventions. 1207 men recruited from December ada. Brenner and colleagues reported Heffelfinger and colleagues report- 2007 to October 2008 in San Fran- on the spread of HIV among MSM ed on recent HIV infections among cisco (Abstract 133). Behavioral prac- in Montreal, Canada, from December MSM in 21 high-prevalence US cities tices were evaluated at the individual, 2005 to September 2009 (Abstract enrolled in the 2008 National HIV Be- dyad, and episode levels, and were cat- 1046). HIV sequence data were col- havioral Surveillance System (NHBS) egorized into mutually exclusive prac- lected from surveillance of primary (Abstract 130). Of 6864 evaluable tices based on highest to lowest HIV HIV-1 infections (PHIs) and divided MSM, 4% had new infections, defined transmission risk. Seroadaptation was into unique transmissions, small clus- as having an HIV-seropositive test re- reported consistently by 39% of men, ters (2−4 PHIs), and large clusters sult with a reported last HIV-serone- whereas only 25% reported 100% con- (5−31 PHIs). Large clusters of infec- gative test result within the past 12 dom use, 14% no oral or anal sex, and tion accounted for the fastest grow- months. Independent risk factors for 12% oral sex only. When the unit of ing subepidemic, accounting for 25% recent infection (compared with unin- evaluation was partnerships, 100% of all transmissions in 2005 to 39% fected men) were younger than age 30 condom use was the most common in 2009 (P < .001). The 34% of infec- years; black, non-Hispanic race; His- practice (33%) compared with sero- tions occurring from MSM born out- panic ethnicity; other nonwhite race or adaptation (26%). When the unit of side of Canada were predominantly ethnicity; completing less than a high evaluation was sexual episode, oral sex unique transmission events. Given school education; having no insurance was the most common practice (65% the unique sociodemographic and be- or public insurance; and having had of acts), and anal sex with a condom havioral characteristics of these 3 dif- 2 or more HIV tests in the prior 24 was next most common (16%). Over- ferent types of transmission groups, months. Risk practices were not sta- all, more than 90% of all individuals, prevention strategies may need to be tistically significantly associated with dyads, and episodes used some form targeted differently to reach all 3 sub- recent infection. of safer sex or seroadaptation, suggest- populations of MSM contributing to this This report extends the data reporting that MSM use several strategies to epidemic. ed by Millett and colleagues about the manage their HIV risk. independent association of sociode- Golden and colleagues presented Populations at High Risk of mographic variables with HIV acqui- data on the differential impact of sero- HIV-1 Acquisition sition risk, in the absence of reported sorting by race among MSM in Seattle, differences in sexual practices or drug Washington (Abstract 1037). In their Youth use. Heffelfinger suggested 3 possible study of 7620 white and black MSM explanations: (1) increased prevalence who received HIV testing at a sexually Pettifor reminded the audience that and sexual mixing patterns within transmitted diseases clinic in Seattle approximately half of all new HIV subgroups; (2) differences in access to from 2006 to 2010, 266 participants infections globally occur in persons care among subgroups; and (3) differ- received a new diagnosis of HIV infec- younger than 25 years, with 35% oc- ential underreporting of risk practices. tion. White and black MSM reported curring in 15- to 24-year-olds (Abstract Regardless of reason, successful, cul- serosorting (unprotected anal sex only 66). There are also marked sex dispari- turally appropriate interventions need with partners of the same serostatus) ties: young women have HIV infection to be developed and tested in high-in- at 30% and 28% of their clinic visits, rates 2 to 3 times those of men in sub- cidence populations. respectively. Although serosorting was Saharan Africa, but HIV-infected men Oster and colleagues reported on a associated with lower HIV infection substantially outnumber HIV-infected network analysis of 23 black men aged risk among white MSM (odds ratio women in the Americas and Europe, 17 years to 25 years newly diagnosed [OR], 0.48; 95% confidence interval where the epidemic occurs predomi- with HIV infection from 2006 to 2008 [CI], 0.35−0.66), there was no such nantly in MSM. Many potential factors in Jackson, Mississippi (Abstract 1044). protection among black MSM (OR, drive this epidemic in younger per- They found that all men were linked 1.04; 95% CI, 0.47−2.30; P value for sons, including biological susceptibili- by few venues, suggesting that these interaction, .02). The reasons for these ty, increased individual behavioral risk, venues should be targeted for testing differences are not clear, as the mean as well as societal forces such as age and prevention intervention. time since the last HIV test was not dif- and power inequities within relation- ferent between newly diagnosed white ships and poverty driving the need for Seroadaptation and black men. A possible explanation transactional sex. is a higher rate of undiagnosed or un- Pettifor focused on recent prom- Several presentations focused on disclosed HIV infection among part- ising results from cash-transfer pro- seroadaptation, the practice of alter- ners of black men. grams whereby families receive cash ing sexual behavior based on self- Additional data on the possible con- incentives either unconditionally or and partner HIV serostatus. Truong tribution of sexual mixing patterns in conditional upon some requirements and colleagues presented data from a transmission among MSM were pre- (eg, girls must attend school). The

39 IAS–USA Topics in Antiviral Medicine

Schooling Income and HIV Risk (SIHR) HIV incidence in pregnant women in could substantially alter the overall Trial conducted in Malawi and re- sub-Saharan Africa is high (4.3/100 HIV epidemic (Abstract 68). In a com- ported at the 2010 International AIDS women-years; 95% CI, 3.9−4.6). Data prehensive review of global prevalence Society meeting in Vienna found that comparing the peripartum risk with data in MSM, Beyrer split countries into HIV prevalence was 60% lower in com- risk in nonpregnant women suggest a 4 scenarios: (1) countries where the munities randomly assigned to the modest increase in HIV acquisition risk epidemic predominantly affects MSM conditional and unconditional cash- (OR, 1.3; 95% CI, 0.96−1.6). Several (much of the Americas, Ghana); (2) transfer groups than in control com- factors may lead to this increase in HIV epidemics driven by injection drug us- munities.1 It appears that changes in acquisition, including behavior change ers (IDUs) (Eastern Europe and Central individual risk behavior accounted for in male partners, or the biological fea- Asia); (3) countries with a generalized less than half of the beneficial effect; a tures (hormonal, immunologic, and lo- heterosexual epidemic (much of sub- possible mechanism was that girls in cal genital tract changes) that occur in Saharan Africa); and (4) countries with the intervention groups were less likely pregnancy. This is of concern for the a mixed epidemic in MSM, IDU, and to have older male partners and less women and their infants, as the risk heterosexuals (South and Southeast likely to receive cash from their male of HIV transmission through breast Asia, Senegal, Egypt). Unfortunately, partners. milk is substantially elevated during there are 94 countries for which there Ott and colleagues also presented the acute infection period. Moreover, are no data for MSM available, includ- data on age mixing in sexual relation- if women are not aware of their HIV- ing three-fourths of African countries. ships from a population-based surveil- seropositive status, they may also be Of note, HIV prevalence is high in MSM lance study in rural KwaZulu-Natal, less likely to receive antiretroviral ther- in all 4 scenarios. South Africa (Abstract 1030). In this apy for prevention of mother-to-child In sub-Saharan Africa, prevalence community, casual relationships with transmission (PMTCT). in MSM exceeds that in heterosexu- “sugar daddies” (ie, men at least 10 The authors stated that in the period al men in all countries and exceeds years younger than their casual part- before initiation of PMTCT programs prevalence in women in all countries ners) are much less common than in Zimbabwe, acute HIV infection ac- except South Africa, Botswana, and marriages in which the woman is sub- counted for only an estimated 6% of Namibia. In modeling the impact of in- stantially younger than the man, lead- infant infections, whereas after PMTCT creasing prevention and treatment for ing to an increased risk of HIV acquisi- programs began, acute infection could MSM (condom and lubricant availabil- tion in young women. account for 44% of new infections in ity, community-based prevention pro- Santelli and colleagues found de- infants. John-Stewart outlined poten- grams, and antiretroviral therapy avail- mographic factors, risk practices, and tial behavioral and biomedical preven- ability for HIV-infected MSM), Beyrer sexually transmitted infections increased tion strategies to prevent peripartum showed that such programs would the rate of HIV acquisition among infections that include recognizing have a substantial impact in countries youth 15 years to 24 years of age en- the desire for pregnancy among many in all 4 types of scenarios, enhanced rolled in the Rakai Community Cohort women and studying the safety and by drug treatment for IDUs in scenar- Study (Abstract 690). Incidence in efficacy of prevention strategies in the ios 2 and 4. He ended with a tribute this group remained at 1% to 2% per peripartum period. to David Kato, the Ugandan activist year from 1999 to 2008. In multivari- Meditz and colleagues explored im- recently murdered for his work on hu- ate analysis, independent risk factors munologic reasons for an increasing man rights for MSM. Beyrer reminded for women and men were lower lev- rate of infections in women aged 40 the audience that improving access to els of education, increased numbers years of age or older (Abstract 33). prevention and treatment and address- of sexual partners, being separated or They reported that 24 postmeno- ing human rights issues are central to divorced, and having sexually trans- pausal women had higher CC chemo- the HIV practitioner community’s abil- mitted disease symptoms. Alcohol kine receptor 5 (CCR5) expression on ity to impact the global HIV epidemic. use was an independent risk factor for CD4+ cells and a higher proportion New data were presented on men, whereas for women, alcohol use of activated CD4+ cells in the periph- MSM in Kenya and Thailand as well. by the last partner was an important eral blood and the cervix than did 21 Sanders and colleagues reported that risk factor. premenopausal women. This disparity HIV-1 incidence was 6.5 per 100 per- may provide some explanation for in- son-years among 666 men with vari- Women creased susceptibility in these women. ous sex partners or recent anal sex (within 3 months of screening) at a John-Stewart and colleagues reported Men Who Have Sex With Men clinic in coastal Kenya (Abstract 1042). on the peripartum risk of HIV acqui- Incidence was highest among men re- sition and factors that drive increased Beyrer described the current state of porting sex with men only (21.7/100 HIV acquisition risk among pregnant knowledge of the global epidemic in person-years; 95% CI, 15.9−29.5), inter- women (Abstract 67). In combining a MSM and provided modeling for how mediate in men who reported sex with number of studies, they observed that increased prevention and treatment men and women (4.9/100 person-years;

40 Conference Highlights—Epidemiology and Prevention Volume 19 Issue 2 May/June 2011

95% CI, 3.3−7.4), and lowest among 0.49; 95% CI, 0.32−0.77), reminding 3.9; 95% CI, 1.3−12.4). Among 164 men who reported sex only with women practitioners of the challenges of con- women, HSV-2 incidence was 22.1 (1.1/100 person-years; 95% CI, 0.4−2.8). dom acceptability among men. per 100 person-years (P < .001 com- Edwards-Jackson and colleagues re- Hughes and colleagues studied a pared with men) and was associated ported on 200 HIV-seropositive MSM cohort of 3297 serodiscordant couples with incident HIV-1 infection (aIRR, recruited from an anonymous clinic in to probe factors affecting the per-act 8.9; 95% CI, 3.6−21.8). Interestingly, Bangkok, Thailand (Abstract 1039). At infectivity of HIV-1 (Abstract 135). genital washing with soap was protec- their most recent sexual encounter, 17% They observed a 2.9-fold increase in tive against HSV-2 acquisition in men of men reported engaging in unpro- infectivity per 1 log10 increase in plas- (aIRR, 0.3; 95% CI, 0.1−0.8), but vagi- tected anal sex, and only 26% disclosed ma viral load in the infected partner. nal washing with soap increased the their HIV-seropositive status. Despite Differences in male-to-female and fe- risk of HSV-2 acquisition in women the lack of disclosure, men were more male-to-male transmission rate were (aIRR, 1.9; 95% CI, 1.0−3.4). likely to report condom use during anal driven by the plasma viral load in the Tenofovir gel halved the risk of sex with partners of unknown serosta- HIV-infected partner, herpes simplex HSV-2 acquisition in the CAPRISA tus (91%) and HIV-seronegative part- virus type 2 (HSV-2) serostatus, and (Center for the AIDS Programme of ners (85%) than with HIV-seropositive the age of the HIV-uninfected partner. Research in South Africa) 004 study.2 partners (61%; P = .001). Baeten and colleagues evaluated Lama and colleagues presented data the association of genital tract HIV-1 from the iPrEx (Chemoprophylaxis for Serodiscordant Couples RNA levels with the risk of HIV trans- HIV Prevention in Men) trial of 2499 mission in the same cohort of serodis- MSM, half of whom were assigned to Transmission between partners in sta- cordant couples (Abstract 154). They receive oral tenofovir/emtricitabine ble serodiscordant relationships may reported that genital tract HIV-1 RNA and half placebo (Abstract 1002). In account for a substantial proportion of level was independently associated this study, HSV-2 incidence was simi- new HIV infections globally. Ndase and with HIV-1 transmission risk after ad- lar among those assigned to the active colleagues point out that interventions justing for plasma HIV-1 RNA levels. A study drug and those assigned to pla- for serodiscordant couples also need to total of 11 transmissions occurred in cebo (6.2 vs 5.8 per 100 person-years, take into account the possibility of out- couples with very low or undetectable respectively). There was no difference side partnerships (Abstract 1040). In genital HIV-1 RNA, but all had detect- in the proportion of genital ulcers in their study of 3380 HIV-1 serodiscor- able plasma HIV-1 RNA. the 2 groups, although there was a re- dant couples observed over a minimum duction in participants with anal ulcers of 24 months, there was a statistically Herpes Simplex Virus Type 2 (relative risk [RR], 0.4 vs placebo; 95% significant decline in the proportion of CI, 0.22−0.85) and a trend toward couples who engaged in sexual activity HSV-2 causes genital ulcerations and a reduction in herpes genital ulcer– (from 94% at enrollment to 73% at 24 has previously been associated with defined adverse events of grade 2 or months; P < .001) and an increase in an increased rate of HIV acquisition higher (13 in the treatment group vs the proportion of HIV-uninfected persons among heterosexual men and women 24 in the placebo group; P = .06). having an outside partner (from 3% to as well as MSM. McClellan and col- Tan and colleagues reported no de- 14%, respectively; P < .001). The rate leagues reported that HSV-2 preva- crease in HSV-1 or HSV-2 shedding of outside partnerships is likely higher lence and incidence were substantially from oral, genital, or anal mucosa than reported, as 22% of seroconvert- higher among women than men en- among 40 HIV-infected patients taking participants who reported no outside rolled in a cohort study of 5543 Zim- ing oral tenofovir compared with those partners were infected with genetically babweans (Abstract 1028). Prevalent not receiving tenofovir (Abstract 979). distinct viruses from their seropositive HSV-2 infection more than doubled the These data suggest that higher levels main partner. This proportion was sub- odds of acquiring HIV-1 infection in of tenofovir at mucosal surfaces may stantially higher among seroconverting men and women in the study. Hughes be required to reduce the risk of HSV participants who did report outside part- and colleagues also found a dou- acquisition and viral shedding among nerships (86%; P < .001). bling of HIV-1 acquisition risk among those already HSV-infected. Ngolobe reported on 444 serodiscor- HSV-2-seropositive men and women in dant couples in Uganda (Abstract 1041). 3297 heterosexual HIV-1–serodiscor- HIV Testing Strategies On multivariate analysis, condom use dant couples in Africa (Abstract 135). was not associated with antiretroviral Okuku and colleagues evaluated HIV testing among at-risk persons re- drug use (adjusted odds ratio [AOR], risk factors for HSV-2 acquisition in a mains suboptimal globally, as does 1.26; 95% CI, 0.81−1.96), suggesting cohort study in Kenya (Abstract 29). knowledge of HIV serostatus. Oster no risk compensation associated with Among 443 men, HSV-2 incidence was and colleagues reported on adherence treatment. However, condom use at last 9.0 per 100 person-years and was as- to HIV testing guidelines among 7271 sex was inversely associated with male- sociated with incident HIV-1 infection MSM participating in the 2008 NHBS controlled sexual decision making (AOR, (adjusted incidence rate ratio [aIRR], (Abstract 1048). Older men were less

41 IAS–USA Topics in Antiviral Medicine

likely than younger men to have been this clinical testing within 90 days dom use reduced risk by 78% (95% CI, tested in the previous year, and there (P < .001). Both Los Angeles and San 58%−89%). Bachanas and colleagues were no differences in testing rates Francisco Public Health Departments reported on condom use by 3538 HIV- by race or ethnicity. However, among plan to use an RTA at all of their test seropositive patients attending clin- men reporting an HIV test within the sites by July 2011. ics in Kenya, Namibia, and Tanzania prior year, the proportion testing new- Choko and colleagues reported on (Abstract 136). Overall, 54% of partici- ly positive was 14% for blacks, 7% for the first evaluation of self-testing in Af- pants had an HIV-seronegative partner Hispanics, and 3% for whites. Of the rica using an oral HIV test kit (Abstract or a partner of unknown serostatus. 5864 (81%) of the sample reporting 42). Four geographic areas in Blantyre, Inconsistent condom use was statisti- 1 or more high-risk characteristics Malawi, were selected for participa- cally significantly associated with be- for whom testing is recommended tion, and 92% of participants opted ing female, desiring pregnancy, being at least every 6 months, only 44% for self-testing over clinic-based or no a spouse (compared with casual and reported receiving an HIV test in the HIV testing. Sensitivity and specificity steady, nonmarital partners), and not past 6 months. This suggests the need were excellent (97.9% and 100%, re- taking antiretroviral therapy. for both better adherence to testing spectively), and overall, 99.2% of self- guidelines and guidelines targeted test users read an accurate result on Male Circumcision for highest incidence populations. their first try. However, 10% of partici- Calderon and colleagues reported on pants needed help beyond the initial Kong and colleagues reported on long- a novel community pharmacy–based instructions, 10% made some type of term effects of male circumcision in HIV testing and counseling program in error in preparing the test, and more Rakai, Uganda (Abstract 36). Overall, New York City (Abstract 1052). Nearly than half stated that they thought that 80% of control subjects returning for three-fourths of eligible patients of- some type of additional counseling a visit chose male circumcision. Over- fered HIV testing accepted. was needed with HIV testing. Overall, all efficacy remains high through 4.8 Delaney and colleagues presented 99% stated that they would be likely to years of follow-up (adjusted effective- data from a randomized trial of a rapid self-test again, and self-testing was the ness, 73%; 95% CI, 55%−84%). Risk HIV testing algorithm (RTA), in which most common choice for the next test behavior increased comparably in both a second rapid test was used to con- that participants would like to take. circumcised and uncircumcised men, firm an initial positive rapid test result The same test was evaluated for suggesting that risk compensation (Abstract 132LB). This approach was 987 participants in Singapore (Abstract has not been observed in this setting. compared with standard HIV testing: 1075). Among HIV–seronegative, at-risk Tobian and colleagues reported that one rapid test followed by off-site con- participants, sensitivity and specificity male circumcision does not decrease firmatory laboratory testing. An RTA of self-testing was 100%. As reported human papillomavirus (HPV) trans- was implemented at 9 testing sites in in the previous abstract, participants mission from HIV-seropositive men to Los Angeles and San Francisco to eval- responded favorably to the testing (eg, their female partners (Abstract 1008). uate the impact on referral to health 89% liked the privacy of testing, and Several other abstracts focused on care practitioners and on CD4+ cell 96% found the instructions easy to fol- strategies for scale-up of male circum- counts or viral load within 90 days low), but nearly three-fourths felt the cision services (Abstracts 1005−1007). of the initial positive rapid test, as need for confidential pre- and posttest compared with standard testing at 23 counseling. This suggests that self-test- Preexposure Prophylaxis control sites. ing may increase the uptake of testing The positive predictive value of the in various populations but that accom- Efficacy, adherence, safety and resis- RTA was 100% and of the initial rapid modations should be made to provide tance. This year, 1 plenary (Session 35) test using a standard algorithm was additional counseling, as needed. and 2 oral abstract sessions (Sessions 86%. All persons receiving the RTA 8 and 25) focused on the strategy of received referrals to medical care, HIV Prevention Strategies using topical or oral antiretroviral whereas only 47% of those with a drugs to prevent HIV acquisition, that positive rapid test result who received Conference presentations on HIV vac- is, preexposure prophylaxis (PrEP). Ce- standard testing (requiring that they cine development are reviewed by lum provided a framework for think- return for their confirmatory test re- Watkins elsewhere in this issue (see ing about how and when antiretroviral sults) actually returned for care. Over- pages 36–37). drugs are used for prevention: drugs all, two-thirds of participants referred initiated before exposure (PrEP), drugs to care received a CD4+ cell count or Condoms used for postexposure prophylaxis viral load test within 90 days of their (PEP), and those used as therapy after initial test (regardless of whether they Hughes and colleagues reported data infection (Abstract 120). were referred from the intervention or on determinants of per-contact HIV-1 Grant and colleagues presented up- control groups), whereas only half of infectivity among serodiscordant cou- dated data from the iPrEx trial, a ran- those not receiving a referral received ples (Abstract 135). Self-reported con- domized, placebo-controlled trial of

42 Conference Highlights—Epidemiology and Prevention Volume 19 Issue 2 May/June 2011

daily tenofovir/emtricitabine in 2499 respectively; P < .001). Drug was most design (Abstract 995). Tenofovir levels MSM and transgender women in North commonly detected in men report- in scalp hair were strongly correlated and South America, Africa, and Asia ing unprotected receptive anal sex, with dose, suggesting that this may be (Abstract 92). Extending data from less common in sexually active men a useful strategy for monitoring adher- the published interim analysis cen- without this risk, and least common in ence in clinical trials. sored on May 1, 2010,3 Grant and col- men reporting no sexual activity in the Concerns have been raised about leagues presented safety and efficacy prior 12 weeks (in 76%, 59%, and 35%, whether tenofovir-based regimens data through August 2010, including respectively; P=.003), suggesting that would be associated with an unaccept- HIV seroconversions occurring by 8 men may have used sexual risk to de- able rate of adverse events (AEs) in an weeks after the study drug was discon- termine whether or not to take the otherwise healthy population. Grant tinued. There were 37 additional HIV study drug. and colleagues reported that partici- infections since the interim analysis, Celum reviewed data from the IAVI pants randomly assigned to take te- but overall efficacy remained largely (International AIDS Vaccine Initiative) nofovir/emtricitabine in the iPrEx trial unchanged (modified intention-to- E001 and E002 studies and reminded had no higher rate of serious, grade treat analysis; efficacy, 42%; 95% CI, attendees that adherence by electronic 3, or grade 4 AEs than did placebo 18%−60%). These efficacy analyses measurement of opening pill bottles participants (Abstract 92). The only did not vary by age, race, education, or was highest with daily dosing; inter- symptoms reported more common- geographic location. mediate with fixed-dose, twice-per- ly in participants in the active study In exploratory analyses, efficacy week dosing; and lowest for fixed-dose group than in the placebo recipients was somewhat lower in uncircumcised plus postcoital dosing (Abstract 120). were nausea (2% vs <1%, respectively; than in circumcised men (efficacy, 36% There also appears to be substantial P=.04) and weight loss (2% vs 1%, re- vs 83%, respectively; P = .10) and lower heterogeneity between populations, spectively; P = .04). Participants given among men not reporting unprotected with adherence higher among 34 US tenofovir/emtricitabine were some- receptive anal sex at baseline than iPrEx study participants than among what more likely to have elevated cre- among men who did report this risk 145 non-US iPrEx study participants atinine levels (2% vs 1%, respectively; (efficacy, −25% vs 52%, respectively; (P < .0001) (Abstract 96LB). In an ad- P = .08), although only 5 participants P = .03). These estimates are not adherence substudy from the Partners in the tenofovir/emtricitabine group justed for other potentially confound- PrEP efficacy trial in Uganda, median had elevated creatinine levels lasting ing variables but raise the possibility adherence as measured by pill count for 2 or more visits. All creatinine el- of differential PrEP efficacy by route and unannounced home visits was evations resolved after drug discontin- of exposure. Efficacy data from het- greater than 99% (Abstract 488). uation; 4 of the participants restarted erosexual men and women and IDUs Amico and colleagues reported on the study drug and exhibited no recur- should be available in the next few years the correlation of self-report, pill count, rence of creatinine elevation. (see Global Advocacy for HIV Prevention drug dispensation records, and blood Mulligan and colleagues reported Web site, http://www.avac.org). detection of study drug among the on bone mineral density (BMD) among Grant and colleagues also presented same 179 iPrEx study participants in- a subset of 503 iPrEx trial participants data demonstrating that efficacy was cluded in Anderson’s presentation (Ab- on 4 continents (Abstract 94LB). At highest among those reporting taking stract 95LB). Men had a median self- baseline, before initiation of the study their study drug more than 90% of the reported adherence level of 100% by drug, BMD was low (ie, z score, <−1.0) time, intermediate in those reporting each of 4 measures, despite no detect- in 36% of participants in the spine and 50% to 90% adherence, and lowest in able study drug in half of the samples in 18% in the hip. There were small those reporting taking less than half (ef- tested. Even among men self-reporting but statistically significant decreases ficacy estimates, 68%, 34%, and 16%, never missing a pill in the prior month, in BMD in participants receiving te- respectively). Anderson and colleagues study drug was detectable in only 68%. nofovir/emtricitabine compared with reported on drug levels among a sub- On the other hand, reports of low lev- those receiving placebo for the total sample of 179 iPrEx trial participants els of adherence (less than half of pills hip (−0.65 at 24 weeks; −0.95 at 48 in the active study drug group at the taken) were uncommon (2%), but in weeks) and spine (−0.95 at 24 weeks), 24-week study visit (Abstract 96LB). this group, study drug was substantially although no difference was observed Overall, 50% had detectable metabo- less likely to be detected (22%). between the groups in bone fractures lites of tenofovir and emtricitabine in Clearly, better measures of ad- or international standards for low BMD their peripheral blood mononuclear herence than self-reporting and pill (z score,<−2.0). cells (PMBCs), indicating no study counts are needed. Liu and colleagues Liu and colleagues also reported on drug had been taken for at least 1 reported on drug levels present in hair BMD changes in a group of 200 HIV- week to 2 weeks before the visit. Par- samples among 15 HIV-uninfected seronegative men in San Francisco ticipants 25 years or older were more persons taking tenofovir for 2, 4, or 7 enrolled in a phase II tenofovir PrEP likely to have detectable drug than days per week under modified directly study (Abstract 93). They reported that younger participants (73% vs 44%, observed dosing, in a cross-over study 10% of men had low BMD at baseline

43 IAS–USA Topics in Antiviral Medicine

(z score, < −2.0). In addition, low BMD rectal challenges with SHIV162p3M184V. became infected compared with 3 of 4 was statistically significantly higher All 5 control animals became infected control animals (14% vs 75%, respec- in men reporting amphetamine (OR, (P = .0008). This suggests that oral tively; P = .04). Silicone vaginal rings 5.9; P <.001) or amyl nitrite (OR, 4.6; PrEP may protect against some resis- tested without study drug were report- P =.002) use and statistically signifi- tant viral strains. ed to be safe and well tolerated among cantly lower in men reporting use of 169 women in South Africa and Tanza- multivitamins, calcium, or vitamin D Drug penetration into genital tissue. nia (Abstract 1004). (OR, 0.3; P <.001). Men receiving teno- Celum reminded attendees that the Hendrix and colleagues reported fovir also had small but statistically sig- effectiveness of PrEP will depend on on the relative safety, adherence, and nificant decreases in BMD in the femo- “getting the right drug to the right acceptability of oral versus vaginal te- ral neck (−0.4%; P =.004) and total hip place at the right time” (Abstract 120). nofovir (Abstract 35LB). In total, 144 (−0.8%; P =.003) but not in the spine Several presentations focused on the women were enrolled at 4 US and 3 (−0.7%; P =.13). There was no statisti- timing and levels of drug penetration African sites and assigned to receive cally significant decrease in BMD after into genital tissues. Dobard and col- sequential 6-week periods of vaginal, 12 months on the study drug, and no leagues presented in vitro data sug- oral, or both formulations. Although differences were observed in fractures gesting that tenofovir/emtricitabine relatively infrequent, women reported between study groups. The clinical provides protection only 2 hours after substantially more nausea, diarrhea, importance of both the higher-than- challenge, whereas raltegravir provides and headache during periods when expected proportion of men with low protection up to 8 hours to 10 hours they were taking tenofovir orally. Self- BMD at baseline and the small reduc- after challenge (Abstract 30). In a low- reported adherence was high, but drug tions in BMD in MSM taking tenofovir- dose, twice-weekly, vaginal-challenge was detected in only 35% to 65% of based PrEP regimens is not yet known. model, 1% raltegravir gel administered samples, with no difference between Drug resistance is another concern 3 hours after challenge protected 5 of 6 study groups. Tissue levels observed that has been raised about the use of macaques through 20 challenges. All 4 with vaginal dosing were more than PrEP. Grant and colleagues reported control animals became infected. This 100 times higher than with oral dos- no additional cases of drug resistance suggests that integrase inhibitors may ing; addition of oral dosing did not in the iPrEx study; the only cases pre- be particularly useful as PEP drugs. further raise tissue levels. US women viously reported consisted of 3 partici- Brown and colleagues reported on preferred the oral formulation, where- pants who were already HIV-infected concentrations of darunavir, ritonavir, as African women were evenly divid- at enrollment (Abstract 92). Liegler and and etravirine in seminal plasma and ed between preference for oral and colleagues searched for minor variant rectal tissue of 12 HIV-seropositive vaginal use. drug resistance among iPrEx trial par- men, to consider these drugs’ utility ticipants and found 1 K65R minor vari- for PrEP in HIV-uninfected persons Rectal delivery of pre- and postex- ant and 1 M184V minor variant in 2 (Abstract 992). Seminal plasma levels posure prophylaxis. Anton and col- placebo recipients (Abstract 97LB). No were 80% to 93% lower than blood leagues reported on the safety and ac- minor variants were found among par- plasma levels for the 3 drugs, and ceptability of this 1% tenofovir gel used ticipants given active drug, including rectal tissue levels were 3 to 13 times rectally among 18 men and women. the 3 participants with breakthrough higher, perhaps reflecting fecal elimi- This hyperosmolar gel was neither well infections, in whom low levels of study nation of these drugs. tolerated (ie, many reported lower-gas- drug were detected at the first-infec- Nel and colleagues presented phar- trointestinal AEs) nor highly accept- tion time point. Levels of drug-resis- macokinetic and safety data on the able (ie, only 25% of participants liked tant virus in the 2 participants HIV-in- investigational dapivirine vaginal ring the gel) (Abstract 34LB). However, a fected at baseline who received study in 48 women (Abstract 1001). The ring single rectal dose resulted in 100 times drug declined to below the lower limit was well tolerated and was not associ- the rectal tissue concentration of teno- of detection (< 0.5%) through week 40 ated with serious AEs. Drug concentra- fovir compared with a single oral dose. of follow-up, suggesting that drug-resis- tions remained high through 35 days Dezzutti and colleagues presented tant virus, when it emerges, may revert of use. Celum reported on plans to data on a reformulated tenofovir gel quickly to wild type. conduct an efficacy trial of the dapiv- prepared for rectal use (Abstract 983). Garcia-Lerma presented data on irine ring (Abstract 120). The reformulated gel had lower os- the efficacy of oral tenofovir/emtric- Singer and colleagues reported on molality, increased spreadability, en- itabine against an emtricitabine-resis- the ethylene-vinyl acetate (EVA) ring hanced transepithelial resistance, and tant simian-HIV (SHIV) variant (Cong containing 100 mg of the investiga- elimination of the epithelial stripping et al, Abstract 31). In this study, all 5 tional nonnucleoside analogue reverse of the colorectal implants that was ob- animals treated with twice-weekly oral transcriptase inhibitor (NNRTI) MIV- served with tenofovir gel. The anti-HIV tenofovir/emtricitabine (3 days be- 150 in a macaque high-dose, vaginal- activity was similar between gels, sug- fore and 2 hours after challenge) were challenge model (Abstract 1003). Only gesting that the reformulated gel may protected against 14 weekly, low-dose, 2 of 14 animals with the MIV-150 ring be a useful rectal microbicide.

44 Conference Highlights—Epidemiology and Prevention Volume 19 Issue 2 May/June 2011

Leyva and colleagues evaluated 3 drop below 350/µL, PrEP again be- among a cohort of IDUs in Baltimore enemas of varying osmolality in 9 men comes more cost-effective at lower (Abstract 484). Castel and colleagues re- (Abstract 993). Hyperosmolar enemas thresholds of effectiveness because of ported on CVL as a population-based bio- (sodium phosphate) caused the most the lower possibility of transmission marker of HIV transmission in Washing- epithelial disruption in the colorectum. from partners with CD4+ cell counts ton, DC, from 2004 to 2008 (Abstract Hypoosmolar enemas (distilled water) above 350/µL. 1023). Although only half of the more had the greatest colonic permeability. Park presented an evaluation of the than 15,000 HIV cases diagnosed dur- Isoosmolar enemas had the best co- cost-effectiveness of PrEP in South Af- ing that time had a viral load measure- lonic distribution and retention and rican women (Walensky et al, Abstract ment available, mean CVL decreased were the most preferred by partici- 37LB). At the efficacy ranges observed substantially over that time. CVL was pants, suggesting that if larger quanti- in the CAPRISA 004 and iPrEx trials, highest in geographic areas with the ties of rectal microbicides are required, they report that PrEP would be cost- highest levels of poverty and unem- isoosmolar enemas may be explored effective (≤ $4600/year of life saved). ployment and the lowest proportion as a vehicle for microbicide delivery. If PrEP could be targeted to women at of high school graduates. The mean of very high risk (ie, incidence >9%/year), the most recent viral load was highest Models and surveys on preexposure be very effective (ie, >70%), and cost among women, blacks, and those in- prophlaxis. Abbas and colleagues (Ab- less than $40 per year, PrEP could be fected heterosexually, through IDU or stract 98LB) and Hallett and colleagues cost-saving for South African women. “other” modes of transmission. (Abstract 99LB) modeled the relative Mayer and colleagues examined Laraque and colleagues also showed benefits and risks of using only anti- practitioner preferences for oral ver- disparities in CVL in New York City, retroviral therapy for HIV-infected per- sus topical PrEP (Abstract 1000). More with higher viral loads observed in sons versus combining antiretroviral than two-thirds of the 121 physicians men, young and middle-aged adults, therapy with PrEP for HIV-uninfected in Massachusetts completing the sur- MSM, persons with AIDS or low CD4+ persons; the models examined effects vey preferred topical PrEP, because cell counts, persons with more recently on the HIV epidemic in South Africa of perceptions of fewer AEs (93%), in- diagnosed cases, and persons in spe- and among serodiscordant couples, creased ease of use (66%), and com- cific neighborhoods (Abstract 1024). respectively. In Abbas and colleagues’ mon use of lubricants for sex (54%). Terzian and colleagues also reported model, the use of both antiretroviral Nearly all (97%) stated that the major on CVL in New York City to monitor therapy and PrEP in a community led factor influencing their prescribing the effectiveness of care (Abstract to a larger prevention impact on the PrEP would be formal guidelines from 1025). Most HIV-seropositive persons epidemic than either alone. Antiretro- the US Centers for Disease Control and had repeated viral load testing, sug- viral therapy is predicted to contribute Prevention (CDC). gesting they were receiving ongoing substantially more HIV resistance at a clinical care. Although nearly half had community level than PrEP, although Treatment as Prevention fully suppressed viral load over the inadvertent PrEP use among HIV-in- prior year, a small proportion had sus- fected persons would also contribute One themed discussion session (Ses- tained high viral load, and these per- to cases of resistance. sion 42) and several additional posters sons were more likely to be younger, Hallett and colleagues posed the addressed how summary measures of black, or female. question of whether it would be more viral load within communities (com- Several models suggest that although cost-effective in preventing HIV trans- munity viral load, CVL) are related to treatment is likely to have a beneficial ef- mission to provide PrEP to the HIV- HIV infection rates and provision of fect on HIV transmission rates, preven- uninfected partner or antiretroviral care in different US cities. Das and tion interventions must also be used to therapy earlier to the HIV-seropositive colleagues reported that progress change the course of the current HIV partner. Assuming all HIV-seropositive has been made in San Francisco in epidemic. Prabhu and colleagues used partners are treated when CD4+ cell mean CD4+ cell count at diagnosis, data from South Africa, Kenya, Malawi, count falls below 200/µL, PrEP would rates of antiretroviral therapy initia- and Mozambique to project the probe more cost-effective only if the cost tion, and time to virologic suppres- portion of new infections attributable of PrEP is less than 40% of the cost sion (Abstract 1022). In particular, to different stages of HIV infections of antiretroviral therapy and if PrEP time from diagnosis of HIV infection (Abstract 482). Less than 10% of new is more than 60% as effective. PrEP to virologic suppression decreased infections are attributable to untreat- would be more cost-effective at lower substantially from 32 months in 2004 to ed HIV infection, whereas two-thirds levels of effectiveness when used by 8 months in 2008 (P<.001). Decreases in to three-fourths of new infections are higher-risk couples (eg, in couples for CVL also correlated with decreases associated with chronic, undiagnosed whom the HIV-uninfected partner may in newly diagnosed and reported HIV infection, and one-fifth to one-fourth be at risk from outside partners). Addi- cases (P <.001). are attributable to acute infection. This tionally, if all HIV-seropositive persons Similarly, HIV incidence decline corre- would suggest that substantial effort are treated when CD4+ cell counts lated temporally with a reduction in CVL should be focused on increasing HIV

45 IAS–USA Topics in Antiviral Medicine

testing uptake, particularly for those over time. Only by adding prevention 2. Abdool Karim Q, Abdool Karim SS, with established infection. to treatment would new infections Frohlich JA, et al. Effectiveness and safety Van Sighem and colleagues used a and AIDS deaths decline and costs of tenofovir gel, an antiretroviral microbicide, for the prevention of HIV infection mathematic model to evaluate the im- decline over time. This led to the con- in women. Science. 2010;329:1168-1174. pact of various interventions on the an- clusions from this report that “treat- 3. Grant RM, Lama JR, Anderson PL, et al. nual number of new infections in MSM ment costs…are unsustainable” and Preexposure chemoprophylaxis for HIV in the Netherlands (Abstract 483). Al- “greater emphasis must be placed on prevention in men who have sex with though immediate treatment for all prevention of new infections.” men. N Engl J Med. 2010;363:2587-2599. HIV-infected persons would lead to a 4. Institute of Medicine Board on Global Health. Preparing for the future of HIV/ rapid decrease in HIV infection rates, Financial Disclosure: Dr Buchbinder has no AIDS in Africa: a shared responsibil- this decline would not be sustained. relevant financial affiliations to disclose. ity. November 29, 2010. http://www. Decreasing risk practices and reducing iom.edu/Reports/2010/Preparing-for-the- the time from infection to diagnosis Future-of-HIVAIDS-in-Africa-A-Shared- (leading to a decrease in risk behav- A list of all cited abstracts Responsibility.aspx. Accessed April 17, 2011. iors) are needed to fundamentally alter appears on pages 99–106. the trajectory of new HIV infections. Bongaarts presented models from a 2010 Institute of Medicine report (Bon- References gaarts and Pelletier, Abstract 173).4 In 1. World Bank. Malawi and Tanzania re- projecting the future of the African epi- search shows promise in preventing demic, increasing rates of treatment HIV and sexually transmitted infections. would slow the rates of new infections July 18, 2010. http://go.worldbank.org/ Top Antivir Med. 2011;19(2):38–46 and AIDS deaths but increase costs YVMPZBKC00. Accessed April 17, 2011. ©2011, IAS–USA

Dermatologic Manifestations of HIV Infection in Africa Resource Card

Based onCards the Topics Available in HIV Medicine article from February/March 2010, this folding card is available on request by visiting www.iasusa.org. Included are brief descriptions of selected dermatologic manifestations, along with their differential diagnoses and treatment options.