HIV Prevention at CROI 2018 and PrEP
Updates on PrEP - Disparities; Current & Future; Delivery/Use/Barriers; Women. Transmission Clusters
Conference on Retroviruses and Opportunistic Infections |
Boston, MA, USA
March 4-7, 2018
Jared Baeten, MD PhD
Connie Celum, MD MPH
University of Washington
CROI 2018 brought together researchers, clinicians, policymakers, advocates, funders and others to talk about the pathogenesis, prevention, treatment, and cure of HIV and associated infections. Prevention has been a major topic at CROI for many years now, with game-changing new options in prevention frequently being reported for the first time there. Importantly, the last few years have shown an increasing number of presentations on how best to delivery successful prevention strategies and achieve impact at scale - thus, prevention that can change the global epidemic. This year's CROI included a number of major advances in the prevention and implementation sphere. As in previous years, oral sessions were recorded and are available online (http://www.croiwebcasts.org/) along with full copies of abstracts (http://www.croiconference.org/abstracts/search-abstracts/). We have cited specific links here for certain oral sessions - these contain both slides and video, but alternative links with slides with audio only are available.
HIV prevention in plenary sessions
CROI frequently has fantastic plenary sessions, and this year was no exception. Particularly notable for HIV prevention were the two Sunday evening opening named lectureships: the Bernard Fields Lecture, which focused on the interface of virology, immunology, and clinical and epidemiologic science (Julie Overbaugh, abstract 10, http://www.croiwebcasts.org/console/player/37038?mediaType=slideVideo&) and the N'Galy-Mann Lecture, which also addressed international collaboration in HIV treatment and prevention in Kenya (Elizabeth Bukusi, abstract 11, http://www.croiwebcasts.org/console/player/37039?mediaType=slideVideo&). Most notably, both speakers specifically focused on how multidisciplinary collaboration in global health amplified and made maximally relevant the work they were able to do.
Morning plenary sessions that focused on HIV prevention included Monday's first plenary on the early days of HIV (Jaffe, abstract 12, http://www.croiwebcasts.org/console/player/37043?mediaType=slideVideo&), Tuesday's session on the interface of mental health and HIV (Remien, abstract 65, http://www.croiwebcasts.org/console/player/37158?mediaType=slideVideo&), and a Wednesday session that tied together prevention goals to address what is need to bend the curve (and then to end) the global epidemic (Weiss, abstract 110, http://www.croiwebcasts.org/console/player/37270?mediaType=slideVideo&). All are very much worth watching online.
Pre-exposure prophylaxis for HIV prevention
In PrEP, an HIV uninfected person uses an antiretroviral medication ahead of an HIV exposure in order to prevent infection. At each CROI since 2011, PrEP has been a major focal point, with major clinical trials, translational science, safety studies, and, most recently, implementation projects reporting findings. Large-scale efficacy clinical trials of PrEP - first as pills and then last year as an antiretroviral-containing vaginal ring - have demonstrated that PrEP is an effective and safe intervention for HIV prevention. Efficacy evaluations of injectable delivery of PrEP is ongoing, and additional delivery approaches (e.g., implants) are in very early phase testing, all potentially heralding a future where multiple options exist for individuals to choose among for prevention. A Tuesday oral abstract session presented new data on PrEP for HIV prevention, with studies spanning from preclinical animal models to population-level implementation.
Oral PrEP using TDF/FTC (Truvada®) is licensed proven and recommended, and the integrase inhibitor cabotegravir, formulated as a long-acting injectable nanosuspension given every 2 months, is being tested in phase III studies. Non-human primate models have shown protection in intravaginal intrarectal and intravenous challenge. An abstract tested a novel atraumatic SHIV penile challenge model, thus mimicking HIV exposure in heterosexual men and insertive men who have sex with men (Dobard, abstract 83, http://www.croiwebcasts.org/console/player/37185?mediaType=slideVideo&). First, TDF/FTC was tested, to validate the model - 6 male macaques were challenged weekly and PrEP efficacy was 93% (p=0.0008). Then, cabotegravir was tested, with a 3-monthly intramuscular injection at 50 mg/kg, which maintains plasma levels equivalent to human dosing); HIV prevention efficacy was 93%, p=0.0001 (1 of 6 animals infected at 12 weeks). Plasma cabotegravir levels were predictive of infection - levels remained high in 5 animals but one animal had lower levels. In sum, the new penile transmission model appears to be valid and shows protection for the investigation product cabotegravir long-acting.
Another macaque challenge study (Massud, abstract 85, http://www.croiwebcasts.org/console/player/37187?mediaType=slideVideo&) explored the effect of tenofovir alafenamide (TAF), the relatively new tenofovir prodrug that is licensed for HIV treatment and is being tested for PrEP in men who have sex with men currently. Previous work has shown that TAF (in combination with FTC) protects against rectal challenge in macaques. This abstract tested TAF for vaginal SHIV challenge. First, a pharmacokinetics validation was done, to find the dose of TAF that would achieve an equivalent level found from oral pill dosing in humans; then, viral challenge was done to test the efficacy of the drug, with drug given before and after each exposure. FTC/TAF was highly efficacious against vaginal SHIV challenges - hazard ratio 5.6 p=0.004, 1 of 6 animals infected during 16 exposures; the one animal infected interestingly had mostly TFV-DP levels below the limits of quantification, although FTC-TP concentrations were high. In sum, FTC/TAF given 24 h before and 2 h after vaginal virus exposure prevented SHIV infection with a degree similar to FTC/TDF. The finding support that FTC/TAF might be a PrEP agent for women - given differences between TAF and TDF in distribution of the drug in human tissues, human studies are needed before TAF is used in practice.
A third macaque challenge study explored the new drug MK-8591 (Markowitz, abstract 89LB, http://www.croiwebcasts.org/console/player/37191?mediaType=slideVideo&). MK-8591 is a potent and long-acting nucleoside reverse transcriptase translocation inhibitor; weekly dosing of results in robust activity. Dosing studies and viral challenges were done. In the challenge studies, 8/8 animals remained uninfected at 1.3 and 0.43 mg/kg dosed weekly. At 0.1 mg/kg 2/8 animals became infected (still 92% reduction compared to controls, p=0.0004). In sum, these results help dissect the necessary level of protection for MK-8591 and reemphasize the potential for this against for extended duration prophylaxis against HIV. Human studies are potentially in the future for this molecule.
From the CDC, estimates of the numbers of people with indications for PrEP use were done (Smith, abstract 86, http://www.croiwebcasts.org/console/player/37188?mediaType=slideVideo&). CDC has previously published the total number of persons in the US with a potential PrEP indication, broken down by transmission group (men who have sex with men, people who inject drugs, heterosexuals). This abstract broke this need down by state and race, combining CDC data sources to better define these populations. The updated estimate is that 1.1 million in US have indications - of these, 44% are black, 25% Hispanic, 27% white. Men who have sex with men account for 71% overall, and women account for 15%. The highest concentrations of black individuals with PrEP indication are in the South and Midwest. CDC also analyzed PrEP coverage currently through national pharmacy data, covering ∼85% of prescriptions in the US. PrEP coverage is highest among white Americans - although even for that group only 14% in 2015-2016 were estimated to have been prescribed PrEP - for black Americans the coverage was estimated to be on the order of 1%. These results emphasize strongly that there remains a tremendous disparity in PrEP use among black populations in the US, where the need is highest - the size of the black population with PrEP indication is sobering.
Two abstracts addressed PrEP roll-out in major cities. From San Francisco (Buchbinder, abstract 87, http://www.croiwebcasts.org/console/player/37189?mediaType=slideVideo&), the experience with PrEP, as part of that city's "Getting to Zero" campaign, was discussed. Between 2006 and 2013, San Francisco implemented easier HIV testing, ART at diagnosis, and other strategies to optimize HIV care; the Getting to Zero campaign was launched in 2014, linking multilevel stakeholders and city-wide programs in PrEP, rapid ART, and linkage and retention in care. As of the end of 2016, SF has seen a 51% reduction in new diagnoses (since 2012); diagnoses have fallen in black, Latino, and white men who have sex with men, although disparities still exist, with black men having the highest rate. PrEP use in the city was estimated from multiple data sources - in 2014, ∼10% to ∼35-50% in 2017 were on PrEP (up to 20,000 men on PrEP in 2017). The percentage of men on PrEP has increased in all racial/ethnic groups, but the increase has been slowest in black men and in youngest men, with older (age 45+) "catching up". In sum, new infections are continuing to decrease in San Francisco, PrEP use is increasing over time. San Francisco is actively working to increase PrEP messaging and access for those subgroups for whom the gap is greatest.
In one of the most exciting presentations of the entire meeting, data from Australia described roll-out of PrEP and population-level impact (Grulich, abstract 88, http://www.croiwebcasts.org/console/player/37190?mediaType=slideVideo&) . The state of New South Wales (where Sydney is located) has been implementing a population-wide study of rapid introduction of PrEP, called EPIC-NSW, since 2016. The setting has had a stable annual HIV diagnosis rate for a decade, has high ART coverage (reached UNAIDS 90/90/90), and a get to zero campaign. EPIC-NSW is implementing PrEP in individuals with higher-risk (STI, meth use, etc.) funded by the government. The goals were pragmatic - looking for impact but also balancing against what could be achieved through completely real-world delivery. A target of 3700 individuals to start PrEP was set, to be started on PrEP in a year. Surprisingly, uptake exceeded all expectations: as of the end of December 2017 there are over 7500 who have started PrEP - and the 3700 target had been met in just the first 8 months. For the 3700, nearly all were male. Nearly all had some follow-up HIV testing. The medication possession ratio (a measure of adherence, which is the fraction of days when someone had PrEP pills, taking into account medication pickups at pharmacies) median was 98%, but 3% never came back and another 27% had an medication possession ratio <80%. In the cohort of people on PrEP, only 2 new infections occurred - one in a man who never started and a second who stopped - the incidence 0.05% per year (compared to an expected 2+% per year). More impressively, HIV incidence was measured in the entire state of New South Wales, measured using surveillance data and comparing to historical HIV rates of recent HIV infection. During the 12 months after the project started there has been a 32% (confidence interval 24-40) reduction in new HIV infections. The benefit was greatest in Australia-born and in central Sydney. This is amazing work - and a clear demonstration that impact at scale can be achieved with PrEP.
A themed discussion on Tuesday on PrEP in women included brief presentations about PrEP eligibility and willingness in the US and reaching high risk young women in southern and East Africa.
From the US, predictors of PrEP eligibility among at-risk women in the southern US was evaluated (Patel, abstract 1048, http://www.croiwebcasts.org/console/player/37233?mediaType=slideVideo&), applying CDC criteria for PrEP use to at-risk HIV negative women enrolled in the Southern sites (Atlanta, Chapel Hill, Birmingham/Jackson, Miami) of the Women's Interagency HIV Study (WIHS) from 2014-15. PrEP eligibility was determined using number of male sex partners, partner HIV status, condom use, and injection drug use in the past 6 months. Of 225 women, 187 (83%) identified as African-American, median age was 45 years, and 120 (53%) had health insurance. One third of women met CDC criteria for PrEP and the most common PrEP indicator was condomless sex. Self-perception of HIV risk was highly predictive of PrEP eligibility. At baseline, 24 (11%) women previously heard of PrEP, and only 1 reported prior use, but 84% were willing to take PrEP, including the majority of PrEP-eligible women (86%).
Predictors of willingness to take PrEP among black and Latina transgender women was evaluated (Poteat, abstract 1045, http://www.croiwebcasts.org/console/player/37234?mediaType=slideVideo&). Data were from Baltimore, MD and Washington, DC from April 2016 - May 2017, via community health centers, outreach, and network referrals. Among 201 interviewed, 86% had heard of PrEP, and of those, 80% knew where to get it. Among self-reported HIV-negative or HIV-unknown participants who had not taken PrEP, 78% were willing to take it, yet only 39% had ever done so. Only 30 participants were on PrEP which limited the power to detect significant predictors of PrEP uptake. Greater transgender pride, history of exchange sex, and higher HIV risk perception were positively associated with PrEP willingness, while legal gender affirmation was negatively associated. History of exchange sex and legal gender affirmation remained significant in the multivariable model. More than three-fourths of HIV-negative black and Latina transgender women reported awareness and willingness to take PrEP, although uptake remains quite low.
Data on PrEP and early ART for female sex workers in South Africa (the TAPS project) were presented (Venter, abstract 1046, http://www.croiwebcasts.org/console/player/37235?mediaType=slideVideo&). TAPS was a prospective, observational cohort study with two arms delivered in existing service settings: 1) PrEP for HIV-negative FSWs, and 2) early ART for HIV-positive FSWs. Of the 947 FSWs seen in clinic, 692 were HIV tested and HIV prevalence was 49%. Among those returning to clinic after testing and confirmed clinical eligibility, 98% (219/224) and 94% (139/148) took up PrEP and early ART, respectively. Of those enrolled, 22% on PrEP and 60% on early ART were seen at 12 months. Little change was seen over time in consistent condom use or the number of sexual partners, with high levels of consistent condom use with clients and low use with main partners in both study arms. There were no seroconversions on PrEP and seven virological failures on early ART. Total cost of service delivery was approximately $126 for PrEP and $406 for early ART per person-year. The TAPS project demonstrated feasibility of PrEP and early ART services within FSWs routine services in high prevalence, urban settings, with good uptake but challenging retention rates. The TAPS demonstration project results provided the basis for the first government PrEP and early ART guidelines and the roll out of a national sex worker plan in South Africa.
Baseline data were presented on risk behavior, perception, and reasons for PrEP among young African women in a prospective study of PrEP delivery HPTN 082 (Celum, abstract 1049, http://www.croiwebcasts.org/console/player/37236?mediaType=slideVideo&). HPTN 082 is an open label PrEP study in Cape Town and Johannesburg, South Africa and Harare, Zimbabwe. Sexually active HIV-negative women ages 16-25 were enrolled regardless of initial decision to initiate PrEP. Of 434 enrolled, 396 initiated PrEP at and 13 after enrollment (94%). Median age was 21 years. 84% of acceptors reported a primary sex partner--60% were thought to be HIV negative, 19% of unknown status, and 1% HIV positive -- and most women thought their partners had other partners (24%) or were not sure (60%). Most women reported risk behaviors, with 66% reporting inconsistent or infrequent condom use, 22% transactional sex in the past 3 months, and 50% intimate partner violence in the past year. STI prevalence was very high: 30% C. trachomatis, 8% N. gonorrhoeae, and 7% T. vaginalis. 41% had depression based on a CES-D-10 score ≥11. Only 16% reported a moderate or high chance of acquiring HIV in the next year. Uptake of PrEP is very high among young African women participating in HPTN 082, who had a high prevalence of risk behaviors and STIs.
Patterns of oral PrEP adherence and HIV risk among eastern African women were analyzed from another prospective study (Pyra, abstract 1043, http://www.croiwebcasts.org/console/player/37237?mediaType=slideVideo&). 233 HIV-uninfected women were enrolled as members of serodiscordant couples in a PrEP demonstration project in Kenya and Uganda and were assessed by daily electronic monitoring via MEMS over the first six months after PrEP initiation. She identified four unique adherence patterns identified: women with high steady adherence (55% of population); moderate steady adherence (29%); late declining adherence (8%); and early declining adherence (9%). Adherence patterns differed by average weekly doses (6.7 vs 5.4 vs 4.1 vs 1.5, respectively). No baseline characteristics were significantly different across adherence patterns. Women with steady HIV risk were more likely to have high steady adherence compared to women with declining HIV risk (61% vs 35%); women with steady HIV risk were also less likely to have early (6% vs 17%) or late (4% vs 19%) declining adherence compared to those with declining HIV risk. Thus, patterns of adherence to oral PrEP among women were associated with their concurrent HIV risk, and women with steady HIV risk were more likely to have high steady adherence and less likely to have declining adherence over the first six months of PrEP use compared to women with declining HIV risk.
Finally, PrEP uptake was high among Kenyan pregnant women offered PrEP during antenatal care (Kinuthia, abstract 1047, http://www.croiwebcasts.org/console/player/37238?mediaType=slideVideo&). HIV-uninfected pregnant women seeking routine antenatal (ANC) services at 10 maternal and child health clinics in Kisumu County, Kenya from June to August 2017 were approached about PrEP following Kenyan PrEP guidelines. A total of 1,008 pregnant women were screened for willingness to be counseled about PrEP. The median age was 23 years, and one-third of women accepted PrEP counseling. The most significant difference in women who accepted PrEP counseling compared to those who declined it was having a partner of unknown HIV status (81% vs 19%, p<0.001), STI diagnosis (6% vs 1%, p<0.001), engaging in transactional sex (3% vs 1%, p=0.02), and forced to have sex (2% vs 1%, p=0.02) in the last 6 months. There were no differences in gestational age between women who accepted and declined PrEP counseling (median 28 [IQR 24-32] vs 28 [IQR 23-34] weeks, p=0.26). Of the 347 women counseled for PrEP, 73% wanted to initiate PrEP and were prescribed PrEP the same day. Compared to women who did not choose to initiate PrEP, initiators more frequently had a known HIV-infected partner (9% vs 2%, p<0.001) and >1 sex partner (6% vs 1%, p=0.04). Among women who did not initiate PrEP, the most frequently reported fears were pill burden (28%) and stigma (13%). This demonstration project indicates that it was feasible to implement PrEP during ANC in a high HIV prevalence region, and one quarter of pregnant women chose to initiate PrEP who more frequently had risk factors for HIV than those who did not initiate PrEP.
In a Wednesday themed discussion, one interesting presentation from South Africa described an assessment of the acceptability of integrating contraception and PrEP discussions into hair salons in South Africa (Bassett, abstract 1070, http://www.croiwebcasts.org/console/player/37358?mediaType=slideVideo&). Most owners (11/17, 65%) were female and comfortable (94%) with a nurse offering health services in their salon. Most stylists (75/92, 82%) were women and 98% reported that if trained they would be willing to talk to clients about health topics and refer to a nurse for services. Among 326 female clients, 73% currently use contraception; the majority (97%) visit the salon at least every 2 months, attend the same salon for most visits (80%) and spend >1 hour (83%). 91% reported willingness to receive injectable contraception, 93% oral contraceptive pills, 74% HIV testing and 77% PrEP at the salon. Access to a private salon room was also significantly related to client willingness to receive health services (p=0.005). Women already on contraception were more likely to agree to contraception in the salon (p=0.028); contraceptive use was not related to willingness to receive PrEP (p=0.82).
On Wednesday morning, two abstracts presented new data on the dapivirine vaginal ring for HIV prevention (Baeten, abstract 143LB, http://www.croiwebcasts.org/console/player/37318?mediaType=slideVideo&, and Nel [presented by Rosenberg], abstract 144LB, http://www.croiwebcasts.org/console/player/37319?mediaType=slideVideo&). At CROI 2016, two trials (ASPIRE and The Ring Study), found that a vaginal ring containing dapivirine, worn for a month at a time, was well-tolerated and reduced HIV incidence by ∼30% overall. This year, the two studies presented exciting data from interim analyses of their ongoing open-label extension studies, called HOPE and DREAM, respectively. The open-label extensions are a bridge from the phase III trials to potential implementation, and this path was walked about five years ago for oral PrEP. HOPE and DREAM enrolled participants who had previously been in ASPIRE and The Ring Study; the results of HOPE and DREAM are remarkably similar. Both studies found high uptake of the ring (>90%), adherence (∼90% of returned rings have residual levels of dapivirine consistent with at least some use), and retention. HIV incidence in both studies to date is much lower than expected - 1.9% per year for HOPE and 1.8% per year for DREAM; these are much lower than the placebo incidences in ASPIRE and The Ring Study (both >4% per year). However, the populations for the phase III trials and the open-label extensions were slightly different, if for no other reason than the open-label studies were older (since participants had aged). To account for these differences between phase III and open-label, both studies did rigorous statistical modeling to compare the HIV incidence in the open-label extensions to what might have been expected if there was no access to the ring, using bootstrap sampling to create a counterfactual comparison group from the placebo arms of the previous phase III trials (ASPIRE/The Ring Study) and controlling for factors that relate to HIV risk such as age, site, and whether women had a curable sexually transmitted infection. The results of the bootstrapping analysis were remarkably similar - each study showed that the HIV incidence observed in the open-label extension studies was 54% lower than would have been expected, and both studies found not a single bootstrap sample to have an HIV incidence as low as 1.8-1.9% per year, in 10,000 samples for each trial. In summary, HOPE and DREAM are the first open-label data for the dapivirine vaginal ring, and they are showing high uptake, objective evidence of adherence, and HIV incidence half of what was expected. These results compare very favorably to open-label extension data from the oral PrEP field - for example, the iPrEx trial had a placebo arm incidence about 4% per year and an open-label extension incidence of 2% per year. More analyses from the dapivirine vaginal ring studies will be coming in the next couple of years.
An additional promising strategy for HIV prevention is the passive administration of broadly neutralizing antibodies (bNAbs) (see Alexandra Trkola's excellent summary on Advances in Immunology in the Sunday workshop for new investigators, http://www.croiwebcasts.org/console/player/37004?mediaType=slideVideo&). The bNAbs 3BNC117 and 10-1074 target the CD4 binding site and V3 glycan supersite on HIV env, respectively, and have been shown to protect macaques against rectal SHIV challenge. An oral abstract tested these two bNAbs in a vaginal challenge model (Garber, abstract 82, http://www.croiwebcasts.org/console/player/37184?mediaType=slideVideo&). After a single subcutaneous injection of the bNAbs, repeated vaginal SHIV challenges in macaques were done. Groups of six female rhesus macaques were injected subcutaneously once with either a single bNAb (3BNC117, 10mg/kg) or two bNAbs (3BNC117 and 10-1074, 10mg each/kg) or no bNAbs and repeatedly challenged once weekly intravaginally until systemic SHIV infection was confirmed. Maximum bNAb concentrations were observed at 1 week post-administration, and the estimated plasma half-lives of 3BNC117 and 10-1074 were between 8-10 days. Macaques administered 3BNC117 alone exhibited significantly delayed SHIV acquisition (median of 5 challenges to infection vs 2 in untreated controls, p=0.002). Animals that received 10-1074 in combination with 3BNC117 exhibited even greater protection against SHIV acquisition (median 11.5 challenges) compared to 3BNC117-alone (p=0.0005) or to untreated controls (p=0.002). Thus, one subcutaneous administration of 3BNC117 singly, or in combination with 10-1074, conferred significant protection in macaques against repeated vaginal challenges with SHIV. These bNAbs are now in human clinical trials. Together these results support continued development of these (and other) bNAbs as potential HIV prevention interventions.
Finally, a fantastic Tuesday symposium focused on big questions in PrEP delivery, including the intersection with STIs (Schillinger, abstract 105, http://www.croiwebcasts.org/console/player/37262?mediaType=slideVideo&&crd_fl=0&ssmsrq=1522159415746&ctms=5000&csmsrq=5006), measuring HIV incidence (Coutinho, abstract 106, http://www.croiwebcasts.org/console/player/37263?mediaType=slideVideo&), confronting big controversies in PrEP - a talk that ties up all lingering questions for PrEP delivery (Bekker, abstract 107, http://www.croiwebcasts.org/console/player/37264?mediaType=slideVideo&), and roll out of PrEP in Kenya as an example of country-level decision-making (Mugo, abstract 108, http://www.croiwebcasts.org/console/player/37265?mediaType=slideVideo&).
Populations / Epidemiology and HIV Transmission
A Monday oral abstract session was devoted to epidemiology and transmission dynamics.
There is increasing interest in the use of cutting-edge molecular sequencing data to study HIV transmission dynamics and potentially to intervene on transmission clusters. One abstract described the use of routine genotyping sequence data to identify growing molecular clusters, transmission rates, and characteristics of those in the clusters, from the US (France, abstract 40, http://www.croiwebcasts.org/console/player/37083?mediaType=slideVideo). Clusters represent transmissions within 3 years - defined by a 0.5% genetic distance threshold; priority clusters had at least 5 transmissions. Total data based on 51,750 persons with a new diagnosis 2013-2016. 60 priority clusters (5-42 persons) were identified, from all regions of the US, representing ∼900 transmissions. Overall transmission rate was estimated to be 4/100 person-years but was 44/100 person-years in the priority clusters (range 21-132). Priority clusters were predominantly MSM (83% vs. 59%) and young (<30). This use of sequencing data is potentially quite valuable as these priority clusters represent a leading edge of transmission. With these data, CDC already took some public health action, providing technical support to local public health jurisdictions to guide prevention actions in those settings.
A second sequencing study used molecular methods to understand the shifting drivers of the epidemic in Switzerland - specifically focusing on the push-pull of treatment as prevention and changing mores related to sexual risk behavior (Bachmann, abstract 41, http://www.croiwebcasts.org/console/player/37084?mediaType=slideVideo&). Data were from the Swiss HIV Cohort Study - which has enrolled 75% of MSM with HIV in Switzerland. Phylogenetic data, incorporating sequences from outside of Switzerland, were used to identify clusters (defined as >80% of sequences grouped in a cluster coming from Switzerland). The results suggested declining infectivity (with greater ART usage) and increasing risk, a shift from diagnosed to undiagnosed individuals as drivers of the epidemic, with clusters pointing to groups where undiagnosed driving individuals may be.
A third abstract (Ragonnet-Cronin, abstract 42, http://www.croiwebcasts.org/console/player/37085?mediaType=slideVideo&) used phylogenetics to understand HIV transmission among transgender women in Los Angeles. Transgender women had the highest odds of clustering (2.3-fold higher than cisgender women, slightly higher than MSM), and transgender women sequences were often linked strongly to each other. These data might be used to identify access points for prevention and treatment services among transgender women and their partners.
From South Africa, data from a large population-based surveillance program in a rural area, with ∼90,000 people under demographic surveillance at any time, were studied from 2004 to present (Vandormael, abstract 46, http://www.croiwebcasts.org/console/player/37089?mediaType=slideVideo&). HIV prevalence is very high in this area (∼29%) and ART has had increasing availability over the past decade. Since 2004, HVI incidence has consistently been greater in women than men. But since 2010, HIV incidence has been declining in men, but not in women. The change in HIV incidence tracks with increasing ART coverage in women (and decreasing viremia in women) and increases in male circumcision coverage in men. In contrast, for new HIV transmissions to women, male ART coverage has not increased and viremia not decreased in men. Similar results were found in a population-based study in Uganda (which was published last year = Grabowski, N Engl J Med 2017). These results emphasize the need to engage men in HIV care.
A second abstract from that same rural South African cohort used a multidisciplinary assessment to understand an expanding cluster of infections (Coltart, abstract 47LB, http://www.croiwebcasts.org/console/player/37090?mediaType=slideVideo&). Phylogenetic analysis revealed a cluster of HIV transmissions, geospatial analyses identified two microregions where 40% each of transmissions were clustered one of which was near a new economic zone (a coal mine), ethnographic work explored how that economic development might have been linked to HIV risk, through increased expendable wealth in a more transient population. The work underlines the importance of HIV prevention to be flexible, adaptive, and data-driven.
From the US, data from the National HIV Behavioral Surveillance were used to demonstrate the potential of serial cross-sectional surveys to estimate HIV incidence (Burnett, abstract 44, http://www.croiwebcasts.org/console/player/37087?mediaType=slideVideo&&crd_fl=0&ssmsrq=1522673782114&ctms=5000&csmsrq=5010) The NHBS does anonymous, standardized surveys in major US cities with high HIV burden. The data from the surveys were combined and estimated HIV incidence in men who have sex with men to be 2.5 per 100 person-years (95% CI 2.1-3.0), in persons who inject drugs to be 0.6 (95% CI 0.5-0.7), and in heterosexuals with low socioeconomic status to be 0.4 (95% CI 0.2-0.6). These results are consistent with data from prospective cohort studies in these populations done in the last several years. In conclusion, successive cross-sectional surveys can be used to estimate HIV incidence in high risk populations.
A multinational prospective study was done to assess the effect of pregnancy on HIV acquisition risk (Heffron, abstract 45, http://www.croiwebcasts.org/console/player/37088?mediaType=slideVideo&&crd_fl=0&ssmsrq=1522674103021&ctms=5000&csmsrq=5005). Prior epidemiologic studies have suggested that women have higher risk of HIV acquisition during pregnancy but the potential for confounding in such analyses prompted this analysis (for example, sexual behavior may change with potentially decreased condom use in pregnancy and decreased sexual activity in later pregnancy). Data were from prospectively followed cohorts of HIV serodiscordant couples from several countries in Africa. Data were collected on sexual behavior and were used to calculate per-contact HIV acquisition risk. Data were compared to non-pregnant periods and were assessed as early pregnancy, late pregnancy, and the post-partum period. Overall HIV incidence was 1.62 per 100 person-years. Infectivity per sex act increased in a step-wise fashion over the comparison periods - lowest in non-pregnant periods and increasing thereafter: risk ratios were 2.07 (p=0.14) in early pregnancy, 2.82 (p=0.01) in late pregnancy, and 3.97 (p=0.01) in the post-partum period; results were consistent across multiple sensitivity analyses. These findings emphasize that pregnancy is a period of increased HIV risk and offers great opportunity for prevention services, like PrEP.
One of the presentations in a Tuesday themed discussion about the role of the microbiome in HIV infection (and vice versa) was relevant to the intensifying interest in the vaginal microbiome and HIV acquisition (Burgener, abstract 271, http://www.croiwebcasts.org/console/player/37205?mediaType=slideVideo&). The study utilized a metaproteomics approach to identify cervicovaginal host and bacterial factors underlying HIV acquisition risk in women from the CAPRISA 004 trial, and examined functional relationships using rhesus macaques and in vitro models. Cervicovaginal lavage samples from last HIV negative time point in 63 women who acquired HIV within the trial (cases), and 638 women who remained uninfected (controls), were analyzed by mass spectrometry. A total of 5,335 host and bacterial proteins were identified. Cox PH models and hierarchical clustering were used to determine relationships to risk of HIV infection. Bacterial co-cultures were used to study bacteria-proteome interactions. In vivo immunofluorescence and RNAseq data of early infectious foci was collected from vaginal mucosal tissues of dual-reporter SIV/SIV challenged animals. Twenty proteins were differentially abundant in cases (P<0.05, HR>1.6, FD>1). Eleven downregulated proteins best classified cases from controls by cluster analysis, localized to the upper vaginal epithelium signifying epithelial barrier disruption. Epithelial barrier disruption was most 4.6 fold more frequent in women with vaginal community state type IV bacteria (Prevotella, Mobiluncus, others). HIV risk with epithelial barrier disruption was exacerbated by CST-IV bacteria; women with both EBD and CST-IV bacteria were the highest risk group identified, with a 8.9-fold increase compared to women without EBD and CST-I bacteria (L. crispatus) (HR=9.9). This study identifies a high-risk phenotype and provides evidence that epithelial disruption exacerbated by vaginal bacteria increases HIV acquisition risk in young women, which may represent vulnerable tissue sites favorable for virus.
HIV Testing and Linkage to Care and Treatment as Prevention
HIV testing, linkage to care, initiation of ART, continued engagement in care, and viral suppression define the continuum that is key to maximizing the benefits of treatment as prevention. A number of studies were presented that suggest ways to improve the steps in this continuum. A Wednesday oral abstract session had several excellent presentations on studies of testing and linkage, particularly with a focus on HIV self-testing (HIVST) , which is increasingly finding its place in the spectrum of tools to combat the epidemic.
From Zambia, a 3 arm cluster randomized trial of three different HIV testing modalities to encourage HIV testing among female sex workers was done: peer educator FSWs randomized to direct delivery of oral HIVST to peers, coupon distribution for an oral HIVST to pick up at a local drug store, or standard of care (= standard availability of HIV testing at a facility) (Oldenburg, abstract 148, http://www.croiwebcasts.org/console/player/37323?mediaType=slideVideo&). The primary results did not show a major effect on HIV testing - but HIV testing was high. There was a statistically significant reduction in number of partners (about 20%) in those who received HIV self-tests. Condom use was not altered by access to HIVST.
From South Africa, another HIV self-testing study provided tests to men who have sex with men in South Africa (Lippman, abstract 149, http://www.croiwebcasts.org/console/player/37324?mediaType=slideVideo&). 65% of participants were 18-24%, 83% had regular male partners. Only 38% had been testing every 6 months and 15% had never tested. Overall 91% tested during the 6 month follow-up. 55% chose blood testing, 20% oral fluid, 25% both at different times. 6 HIV seroconversions were observed. 96% gave some tests to a friend, 66% to partner, 84% to family member. 83% preferred HIV ST for their next test and 65% preferred blood testing (which may reflect the standard of blood testing in clinics). In summary, there was a preference for HIVST over clinic-based testing and would potentially increase the frequency for testing. There was no evidence of social harms. Having both blood and oral fluid options would have more choices.
From Zimbabwe (Sibanda, abstract 150LB, http://www.croiwebcasts.org/console/player/37325?mediaType=slideVideo&), a cluster randomized trial tested 1) whether conditional incentives given to lay HIVST distributors would increase linkage to care, 2) non-randomized assessment compared against communities without HIVST in the same area. Outcome was done through a population-based survey done 6 weeks after the HIVST campaign. 38 communities were included, ∼80,000 HIVST distributed, outcome assessed in ∼7000 community members. 89% had ever tested for HIV - higher than expected - and ∼60% had recently been tested with the HISVT, and about 1/3 had tested for the first time with the HIVST. However, incentives to lay workers did not increase the linkage to services in general, but for HIV+ persons not on ART, there was an increase in linkages to services (adjusted PR 1.59, 95% CI 1.05-2.39, 76% vs. 50%). Compared to the communities not in the HIVST area, there was higher ART initiation during the HIVST campaign (RR 1.27, 95% CI 1.14-1.43) but not before (as expected) or after. In sum, a small incentive (0.20 per client linked) did not result in an overall increase in uptake of post-test services compared to just paying lay workers, although there was small improvement in HIV+ retesting for HIV+ not on ART.
Pre-conference workshop on clinical trial design and analysis
An excellent pre-conference workshop on clinical trials is very much worth watching. There were three presentations. First, an extremely clear explanation of the use of phylogenetics for understanding transmission clusters when combined with HIV incidence assays and the potential utility of targeting highly effective prevention interventions to highly connected and recent transmission networks (Little, abstract 6, http://www.croiwebcasts.org/console/player/37024?mediaType=slideVideo&). The use of phylogenetic data to guide public health interventions (as detailed above) is an emerging opportunity. Included in the talk is a compelling rationale for the need to decriminalize non-disclosure of HIV status, given that similar viruses do not verify transmission and since most laws do not consider ART (U=U), PrEP or condom use. Second, a run-through of the use of population-level measures for impact evaluation (Justman, abstract 7, http://www.croiwebcasts.org/console/player/37025?mediaType=slideVideo&). Included in this talk was a summary of the population-based HIV impact (PHIA) project, funded by PEPFAR to measure national HIV incidence and national and subnational rates of viral suppression in 14 countries. The PHIA project uses cross-sectional estimates of HIV incidence based on limiting antigen avidity index and viral load testing and has collected biomarker and HIV incidence data from >250,000 adults, providing important data on geographic and gender patterns of HIV prevalence and viral suppression. Third, pragmatic trials (Grant, abstract 8, http://www.croiwebcasts.org/console/player/37026?mediaType=slideVideo&), including the pros and cons, and a framework for understanding trial results.