icon-    folder.gif   Conference Reports for NATAP  
  Conference on Retroviruses
and Opportunistic Infections (CROI)
February 22-25, 2016, Boston MA
Back grey_arrow_rt.gif
HIV Prevention at CROI 2016 - Jared Baten MD, Connie Celum MD
  Conference on Retroviruses and Opportunistic Infections
Seattle, WA, USA
February 22-25, 2016
Jared Baeten, MD PhD
Connie Celum, MD MPH
University of Washington
CROI 2015, held in relatively warm Boston this year, again brought together a broad group of researchers, clinicians, policymakers, advocates, funders and others to talk about the pathogenesis, prevention, treatment, and cure of HIV and associated infections. The past CROI meetings have seen high enthusiasm regarding new HIV prevention and treatment strategies, which, if implemented at scale, could result in substantial effects on the global epidemic. 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.
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. Since 2011, PrEP has been a major focal point of CROI, with major clinical trials, translational science, safety studies, and implementation projects reporting findings. This year was no different, with important abstracts presented in all of these areas.
Large-scale efficacy clinical trials of tenofovir-based PrEP - as both pills and topical vaginal gels, many reported at CROI - have demonstrated that PrEP is an effective and safe intervention for HIV prevention. Since the last CROI, the World Health Organization has issued guidance recommending PrEP using tenofovir-based pills as part of combination prevention initiatives for persons and groups at substantial HIV risk, and several countries, including Canada, France, Kenya, and South Africa, have made regulatory decisions that open the door for PrEP to be a major part of their prevention efforts. However, oral PrEP is not a workable strategy for all individuals and clinical trials of PrEP have illustrated important challenges: in 2012, 2013, and 2015 the results of three major trials of PrEP among African women (FEM-PrEP, VOICE, and FACTS 001) reported their findings, showing no protection against HIV, as a result of very low adherence. While other studies have demonstrated that PrEP can be efficacious in preventing HIV in women, the results of those three studies raised important warning flags about developing and implementing prevention options for African women, who face tremendous HIV risk.
A great Tuesday afternoon symposium discussed PrEP innovations - in delivery (http://www.croiwebcasts.org/console/player/29560?mediaType=podiumVideo&), STI control (http://www.croiwebcasts.org/console/player/29561?mediaType=podiumVideo&), making PrEP appealing (http://www.croiwebcasts.org/console/player/29562?mediaType=podiumVideo&), and the promise and challenges of sustained delivery PrEP (http://www.croiwebcasts.org/console/player/29563?mediaType=podiumVideo&).
New clinical trials of PrEP
At CROI 2016, two phase III clinical trials of a new PrEP modality - a vaginal ring, worn continuously for a month at a time, and containing the non-nucleoside reverse transcriptase inhibitor dapivirine - reported their results. (Full disclosure - one of us [JMB] led one of these two trials.) The first (Baeten, abstract 109LB, http://www.croiwebcasts.org/console/player/29584?mediaType=podiumVideo&), called MTN-020/ASPIRE, enrolled and followed 2629 women at 15 research sites in 4 countries (Malawi, South Africa, Uganda, Zimbabwe) between 2012 and 2015. Their median age was 26 years, median follow-up was 1.6 years, and women attended 91% of expected monthly visits. Among women assigned to the active dapivirine vaginal ring arm, approximately 80% had dapivirine detected in plasma; however, detection of dapivirine in plasma was reported by the authors to potentially overestimate adherence (for example, if women only used the ring on the day of the clinic visit). A total of 168 post-randomization HIV-1 infections occurred: 71 among those assigned the dapivirine vaginal ring (incidence 3.3 per 100 person-years) and 97 among those assigned the placebo ring (incidence 4.5 per 100 person-years).  Compared to those assigned placebo, women assigned the dapivirine ring had a 27% (95% CI 1-46, p=0.046) relative reduction in HIV-1 incidence overall. Several analyses were done to assess HIV protection in populations with evidence of better adherence, most notably finding a 37% (95% CI 12-56, p=0.007) reduction in an analysis excluding data from two sites with lower retention and adherence and a 56% (95% CI 31-71, p=0.0003) reduction in an analysis among women older than 21 years of age. Adherence was lower in women aged 18-21 compared to women older than 21. The second (Nel, abstract 110LB, http://www.croiwebcasts.org/console/player/29585?mediaType=podiumVideo&), called The Ring Study, enrolled 1959 women at 7 sites in South Africa and Uganda. Like MTN-020/ASPIRE, The Ring Study also had high retention (82% of visits) and adherence (83% by objective markers, again with a caveat about overestimating adherence). Overall, there were 133 HIV seroconversions: 77 among 1300 women assigned the dapivirine vaginal ring (incidence 4.1 per 100 person-years) and 56 among 650 women assigned placebo (incidence 6.1 per 100 person-years), a 31% reduction (95% CI 1-52, p=0.04). Women older than 21 years of age had somewhat greater HIV protection (37%, 95% CI 3-59%). A post-hoc analysis found a strong trend between the amount of dapivirine eluted from the ring during the month (another marker of adherence) and HIV protection. Thus, overall, these two studies had similar results - partial HIV protection overall, higher in subgroups with better use of the product. In both studies the ring was safe, with no increase in adverse events or in drug resistance among women who acquired HIV. The fact that adherence predicted HIV protection should come as no surprise to anyone following this field for the last several years, and younger age being related to less use and less HIV protection is consistent with other PrEP findings as well. Notably, HIV incidence in the placebo groups for this study was >4% per year, and >8% in some research sites, emphasizing the incredible need for HIV prevention strategies that women can use. Next steps for the dapivirine vaginal ring will include regulatory review and open label studies. Thinking back, one remembers 2011 and the presentation of the iPrEx trial results at CROI - a 44% reduction in HIV risk (and lower among younger participants) in that study of tenofovir-based pills among men who have sex with men and transgender women - as well as the substantially greater adherence and HIV protection seen in follow-on open-label projects among men who have sex with men subsequently.
PrEP clinical safety and implementation
Understanding the scope and risks of clinical complications of tenofovir-based PrEP received a fair amount of attention at CROI 2016. A follow-on analysis from the iPrEx clinical trial of PrEP assessed bone mineral density (BMD) during PrEP use and after PrEP discontinuation (Grant, abstract 48LB, http://www.croiwebcasts.org/console/player/29449?mediaType=slideVideo&). A total of 498 people were enrolled. DXA scans of the hip and spine were done at baseline and every 24 weeks, plus 24 weeks after stopping PrEP use. Importantly, data were analyzed by level of PrEP adherence, as measured by levels of tenofovir in blood samples. As previously reported by this group, BMD declined on PrEP. However, when PrEP was stopped, BMD recovered, quickly, returning to baseline levels 6 months after stopping PrEP. When stratified by age, bone recovery was also seen for young participants. Thus, bone density seems to return to baseline levels in those who complete PrEP use, which is very reassuring, as PrEP is generally a time-limited strategy.
An analysis from the PROMISE study provided reassuring data that tenofovir use during pregnancy may not result in substantial bone mineral density problems for infants (Siberry, abstract 36, http://www.croiwebcasts.org/console/player/29477?mediaType=slideVideo&). While this was a study of HIV infected pregnant women, its carryover potential to PrEP is notable.
A Wednesday afternoon poster discussion session was devoted to renal and STI issues related to PrEP. In a secondary analysis from the iPrEx OLE cohort (Gandhi, abstract 866, http://www.croiwebcasts.org/console/player/29667?mediaType=slideVideo&) greater cumulative exposure to TDF-FTC PrEP, as measured by higher levels in hair samples, was associated with a greater decline in renal function, although on average the decline was only 5.6% from baseline even in those with daily dosing. For those with a baseline creatinine clearance of <90 mL/min, there was a 27% change of falling to <70 mL/min over the first year, and this risk was greater in those over 40 years of age. In a demonstration project of PrEP in the US, another analysis (Liu, abstract 867, http://www.croiwebcasts.org/console/player/29668?mediaType=slideVideo&) assessed the risk of a significant renal function decline on PrEP. Overall, the mean change in estimated glomerular filtration rate was -2.8% over the first year, and the decline seemed to occur in the first three months and was stable thereafter. Only 2% of patients had a ≥10% decline, none declined to <60 mL/min, and PrEP was held due to an elevated creatinine in only three patients and then only temporarily; however, decline to <70 mL/min was more common in those with baseline <90 mL/min. Finally, a secondary analysis of the Partners PrEP Study (Mugwanya, abstract 868, http://www.croiwebcasts.org/console/player/29669?mediaType=slideVideo&) assessed risk of renal tubulopathy using several urine toxicity measures. Tubulopathy was very rare (<2% of subjects, and no different than placebo in this clinical trial) and was not associated with declines in glomerular filtration. Taken together, these studies remind us, of course, that tenofovir-based pills for PrEP have, on average, a small and not clinically significant effect, on renal function, that is reversible after PrEP discontinuation. There is much question about how much renal monitoring is necessary to ensure safety with PrEP - these findings, and others reported in the past couple of years, suggest that toxicity is very rare. Current CDC guidance is to monitor renal function every 6 months while on PrEP, with consideration for more careful attention for those aged >40 and with lower baseline renal function, an approach that is supported by these data.
Two abstracts (Golub, abstract 869, http://www.croiwebcasts.org/console/player/29670?mediaType=slideVideo& and Cohen, abstract 870, http://www.croiwebcasts.org/console/player/29671?mediaType=slideVideo&) assessed STI frequency among men who have sex with men taking PrEP. STI rates were high, perhaps not surprisingly for high-risk persons eligible for PrEP, and more frequent screening detected more STIs. Future work will need to assess the optimal timing and cost-benefit of frequent STI screening.
Finally, a much-discussed case of a PrEP breakthrough infection was presented (Knox, abstract 169aLB, http://www.croiwebcasts.org/console/player/29737?mediaType=audio&). The individual, a 43 year-old gay man acquired HIV in spite of good evidence of TDF-FTC PrEP use; the virus acquired had multiclass resistance, including to TDF and FTC, suggesting transmission of a strain that was unable to be blocked by PrEP. The occurrence of this case is not surprising - PrEP may offer less than absolutely 100% protection, particularly if faced with resistant virus and it should not diminish efforts to provide PrEP to individuals at risk.
Translational work on PrEP and new PrEP agents
While oral tenofovir disoproxil fumarate (TDF), often combined with emtricitabine (FTC) (the combination branded as Truvada), is the only licensed PrEP medication on the market, new PrEP agents - pills, topical delivery approaches, injections, and other modalities - are still needed as choice will be essential to respond to the prevention needs of different individuals and populations. Several studies of new PrEP approaches were presented at CROI, including tenofovir-based alternatives to TDF/FTC, other oral agents for PrEP, and new work with injectable agents.
The new oral antiretroviral tenofovir alafenamide (TAF) is, like TDF, a prodrug of tenofovir, but TAF has a pharmacokinetic profile that results in lower plasma but higher intracellular concentrations of tenofovir than TDF resulting in less renal and bone toxicity. In an established macaque rectal SHIV challenge model that was previously used to demonstrate the protective efficacy of TDF/FTC for HIV prevention, researchers studied TAF/FTC (Garcia-Lerma, abstract 107, http://www.croiwebcasts.org/console/player/29582?mediaType=podiumVideo&). Overall, TAF/FTC, when provided in a dosage comparable to standard 25 mg of TAF daily human dosing (which is the dose approved, with other medications, for treatment of HIV infection), provided 100% protection against rectal challenge. As expected, plasma concentrations were lower of tenofovir for TAF, and, interestingly rectal concentrations were also lower, than for TDF, but protection was preserved. These data suggest that TAF-based PrEP may be effective - next steps must include evaluations in people.
A phase I, single-dose pharmacokinetics study of TAF among women (Garrett, abstract LB102, http://www.croiwebcasts.org/console/player/29628?mediaType=slideVideo&) assessed plasma and mucosal concentrations of tenofovir and phosphorylated tenofovir. Not surprisingly, plasma concentrations were considerably lower (19-fold) and intracellular PBMC concentrations of phosphorylated tenofovir were considerably higher (9-fold) than after TDF dosing (as observed in a prior single-dose study). Notably, tissue concentrations of phosphorylated tenofovir were also lower: 2-fold for cervicovaginal tissue and 13-fold for rectal tissue, with a majority of samples undetectable. More work is needed to assess whether TAF provides sufficient be a potential PrEP agent.
Another study evaluated use, safety, and acceptability of TDF-based pills and tenofovir gel (a reduced glycerin formulation appropriate for rectal use) as prevention agents among men who have sex with men and transgender women (Cranston, abstract 108LB, http://www.croiwebcasts.org/console/player/29583?mediaType=podiumVideo&). This study, MTN-017, was a large, open-label, phase II, cross-over study among 195 subjects - each completed 8 weeks of pills (FTC/TDF daily), daily gel, and gel used before and after sex; it was the first phase II study of a rectal microbicide ever completed. All three approaches were safe, and the pills and the with-sex gel were more acceptable than daily gel. Adherence was high and further work from this study will report out on the innovative approach to adherence counseling done by the team. The sum of this work is that a with-sex approach to gel use may be feasible, useable, and safe as a rectal microbicide - spurring the need to evaluate in larger-scale studies. At the same time, another abstract (Shieh, abstract 169bLB, http://www.croiwebcasts.org/console/player/29738?mediaType=audio&) assessed distribution of a gel product in the colon using SPECT/CT imaging, finding that manual application, as opposed to using a formal applicator device, resulted in low and variable delivery. Gel applicators are cumbersome and easier methods to deliver a rectal microbicide would be desirable, but these findings offer some caution that all approaches may not be offer a strong chance of delivering sufficient product for HIV prevention. Two linked abstracts evaluated the safety of daily dosing of the oral pill maraviroc (MVC) as a PrEP agent (Gulick, abstract 103, http://www.croiwebcasts.org/console/player/29578?mediaType=podiumVideo& & McGowan, abstract 104, http://www.croiwebcasts.org/console/player/29579?mediaType=podiumVideo&). This study (HPTN 069/ACTG A5305) was a randomized phase II evaluation of daily MVC alone, MVC + FTC, MVC + TDF, or TDF + FTC, for 48 weeks, among 406 men who have sex with men and transgender women in the US. All four medication approaches were safe, and side effects (including gastrointestinal and renal side effects) were comparable among them all. Adherence was ∼80% and there was no indication that MVC interacted adversely with FTC or TDF. Of note, 5 HIV infections occurred: 1 had been assigned TDF + FTC but had no evidence of use and 4 had been assigned MVC, of whom 2 had some evidence of use. In a rectal explant substudy, no increased CCR5 expression or T cell activation was seen, addressing prior concerns about use of MVC for prevention. In explant viral challenge experiments, MVC did not protect against HIV; however, the explant model was assessed as potentially not working well for MVC. All in all, this study demonstrates that MVC may be an alternative to TDF-based PrEP, at least in terms of comparable safety; the breakthrough infections and the explant challenge data (recognizing that the assay may be non-ideal) do raise serious questions about whether MVC would effectively prevent HIV.
Two oral abstracts addressed a new injectable PrEP agent: cabotegravir long-acting (CAB-LA), an analogue of the oral treatment integrase inhibitor dolutegravir. Prior studies have shown the CAB-LA provides very high protection from rectal and vaginal SHIV challenge in macaques, and at CROI 2016, a macaque IV challenge model was presented (Andrews, abstract 105, http://www.croiwebcasts.org/console/player/29580?mediaType=podiumVideo&). At a dosing comparable to that used in humans, CAB-LA protected 7/8 animals from infection; a lower pre-exposure dose was perhaps slightly less effective (6/8 animals protected). Overall, the results suggest high, though potentially imperfect, protection by CAB-LA to an IV viral challenge, supporting the idea that CAB-LA could protect against injection exposures. A second study (Markowitz, abstract 106, http://www.croiwebcasts.org/console/player/29581?mediaType=podiumVideo&) assessed CAB-LA in humans - ECLAIR was a phase II evaluation among low-risk men who received a 4 week oral run-in of CAB followed by intramuscular CAB-LA every 12 weeks for 36 weeks, or placebo of 106 men who started on the CAB arm, 94 completed the oral run-in and 87 finished the injection phase (for a total of 82% completion). Safety was generally supported, although injection site pain was common (92%), although was described as of mild duration and lasting ∼5 days. Pharmacokinetic data found absorption of the medication was greater than expected, with higher peak and lower trough levels: 70% of subjects had trough levels of CAB <4x the IC90 (a level that provided protection against infection in macaque challenge models) and 15-30% <1x IC90, suggesting potentially that every 12 weeks may be too infrequent to maximize HIV protection for this dosing of CAB-LA. Overall, these important phase II data advance the pathway for CAB-LA as a potential new PrEP medication.
Finally, the first plenary on Tuesday morning discussed the use of monoclonal antibodies for HIV prevention and treatment (http://www.croiwebcasts.org/console/player/29438?mediaType=slideVideo&). A phase I safety and pharmacokinetics study of one of these antibodies (VRC01) was presented in an oral abstract session (Mayer, abstract 90, http://www.croiwebcasts.org/console/player/29613?mediaType=slideVideo&); the results showed that the antibody (administered as an intravenous infusion or a more-frequent subcutaneous injection) was well-tolerated. Larger-scale phase IIB studies of this antibody are planned but have not yet initiated.
Prevention of mother-to-child transmission of HIV
Prevention of mother-to-child transmission (PMTCT) of HIV is a tremendous public health achievement. Understanding how best to deliver PMTCT services is the current research priority. Tuesday morning featured a powerful oral abstract session on PMTCT-related topics.
In Swaziland, an implementation science study using a stepped-wedge trial design evaluated roll-out of Option B+ across 10 sites which were randomly assigned to transition from Option A to Option B+ (Abrams, abstract 34, http://www.croiwebcasts.org/console/player/29475?mediaType=slideVideo&). 2315 women (1272 A, 1043 B+) were followed, of whom 53% were newly diagnosed. ART uptake was 35% under A and 94% under Option B+ (p<0.001), including when limited to those with CD4 counts <350 (66% vs. 94%, p<0.001). Antenatal retention was also increased under Option B+ (54% vs. 68%, p<0.001), although interestingly this was lower among women with CD4 >350 who were on ART. Postnatal retention was also higher on B+ (50% vs. 26%, P<0.001), although still low overall. Thus, transitioning clinics to Option B+ greatly improved ART initiation and coverage, but retention was overall low, particularly postpartum. Efforts are needed to improve continued engagement in care for women in Option B+ programs.
In Malawi, a national survey of the Option B+ program was performed (Barr, abstract 35LB, http://www.croiwebcasts.org/console/player/29476?mediaType=slideVideo&). Overall transmission was 4.1% (3.2-5.2), substantially lower for those on ART vs. not on ART (2.9% vs. 20.3%). Transmission was lower with earlier ART start (1st trimester < later in pregnancy < postpartum). Overall, four years after implementing Option B+ in Malawi, ART coverage among HIV infected pregnant women is high, HIV transmission is relatively low and very low among those on ART. Another abstract from Malawi (Gupta, abstract 789, http://www.croiwebcasts.org/console/player/29759?mediaType=slideVideo&) detailed the rapid and successful decentralization of Option B+ into smaller clinics in Malawi.
Prevention of HIV transmission to women during pregnancy and optimizing the ability of HIV infected women to take treatment both could be supported by having better HIV testing of male partners. A randomized trial in Kenya compared home-based HIV testing of partners of women attending their first antenatal clinic visit to standard-of-care invitation of male partners to come to the clinic for testing (Farquhar, abstract 49, http://www.croiwebcasts.org/console/player/29464?mediaType=slideVideo&). The rationale is that although there are many benefits to male partner involvement in antenatal care, only 5-33% of men attend antenatal clinics. A total of 601 women enrolled: 87% of male partners assigned to the home-based testing offer tested vs. 39% of male partners invited to the clinic for testing, and similar percentages of the women knew their partners' status. Testing as a couple was higher as well (77% vs. 24%) and a tripling of identification of serodiscordant status occurred (13% vs. 4%). Intimate partner violence was infrequent - observed in 3% of visits, of which 50% occurred in concordant negative couples, and none were considered by participants or staff as study related. Thus, this program of home visits, partner education and HIV testing for couples in Kenya was able to double the proportion of men who tested during their partner's pregnancy. This is substantially better than other programs that have been evaluated to increase male partner testing during pregnancy, such as letters to invite male partners to attend HIV testing at antenatal clinics and making antenatal clinics more 'male friendly'.
Populations / Epidemiology and HIV Transmission
From the US, a study from the CDC estimated the lifetime risk of a diagnosis of HIV (Hess, abstract 52, http://www.croiwebcasts.org/console/player/29467?mediaType=slideVideo&). Lifetime risk (=cumulative probability of being diagnosed with HIV from birth) may be a good way to communicate risk to public. The investigators based the new analysis on (1) HIV diagnosis data from the National HIV Surveillance System, (2) mortality data from the National Center for Health Statistics, and (3) population data from the US Census for the years 2009-2013. The results found that one in 64 in US will get HIV in their life time. However the risks in subgroups are much higher, including 1 in 6 men who have sex with men (MSM), 1 in 2 of all black MSM, and 1 in 4 of all Hispanic MSM. In comparison 1 in 23 of women who inject drugs (1 in 23) will get HIV in their lifetime. This analysis highlights the need for prevention targeted to specific populations in the US. From South Africa (Oldenfeld, abstract 167, http://www.croiwebcasts.org/console/player/29734?mediaType=slideVideo&) a population-based study assessed ART use and HIV acquisition. Data were from the Africa Centre demographic surveillance project. Using data from the population survey, a cohort of cohabiting couples was created and ART use was linked from public health records (thus, couples were not specifically enrolled but identified through surveillance records). HIV incidence by partner status was: 0.3% per year for those with an HIV negative partner, 1.4% for those with an HIV positive partner on ART, and 5.6% for those with an HIV positive partner not on ART, translating into a 77% reduction in transmission risk with ART. These real-world data assess HIV risk reduction from ART.
Finally, An excellent plenary on HIV in transgender populations is absolutely worth watching (http://www.croiwebcasts.org/console/player/29575?mediaType=podiumVideo&).
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.
From Kenya, a cluster randomized trial assessed the effectiveness of immediate versus delayed provider-assessed partner notification services for HIV (Cherutich, abstract 50, http://www.croiwebcasts.org/console/player/29465?mediaType=slideVideo&). A total of 18 HIV testing sites were assigned to use immediate versus delayed (to 6 weeks) of partner tracing of persons newly identified as HIV infected. The outcomes were HIV testing, case finding, and enrollment in HIV care of sexual partners; a secondary objective was to compare regional and urban/rural differences in the number of index cases needed to intervention to identify new HIV infected partners. A total of 550 were randomized to the immediate arm (57% women) and 569 to the delayed notification (65% women). Detailed sexual history information was collected and 1872 sexual partners were named. The intervention included community tracing and testing of sexual partners of newly identified HIV+ partners with phone calls, home or workplace visits. Outcomes were assessed at 6 weeks. There were significant increases across all outcomes with 4.8 fold increase in the number of partners tested; a 14.8 fold increase in the number of newly tested partners; and a 4.4 fold increase in partners who were newly enrolled in HIV care. The number needed to test to identify an HIV infected case was 4.2 index cases to find one partner tested for first time; there was a higher yield of 3 NNT in rural areas, perhaps due to lower penetration of HIV testing there. Partner tracing was not associated with an increased risk of intimate partner violence. This study demonstrates the feasibility and effectiveness of partner services in Africa.
In a poster discussion, one presentation (Chamie, abstract 979, http://www.croiwebcasts.org/console/player/29528?mediaType=slideVideo&) described a highly innovative and effective hybrid approach to achieve high testing coverage in the SEARCH trial taking place in Uganda and Kenya. SEARCH is a multisite cluster randomized trial of treatment as a prevention in 16 communities of 10,000 persons each. In the first year of the SEARCH, the team achieved a 89% coverage of adult HIV testing through two week multi-disease community health campaigns. For those who didn't come to the campaigns, home-based HIV testing was conducted. A year later (2014-15), repeat mobilization was conducted and new services of men's health and urgent care were added based on community input, again followed by home-based HIV testing and counseling. They were able to achieve a remarkable 94% HIV testing coverage. Coverage was 92% among men and 96% among women, including high coverage in key populations (95% of bar workers, 85% of transport workers, and 89% of fishing workers). Notably, 50% of adults who did not test in year 1 tested in year 2, and they had a similar prevalence (9.7%) to those who had tested in year 1. Thus, this mobile model bringing together in East Africa was highly effective in achieving almost universal testing, including among men.
In a Wednesday symposium on challenges in getting to 90-90-90, a presentation focused on gaps in testing and treatment of men (http://www.croiwebcasts.org/console/player/29676?mediaType=slideVideo&). Barriers include men's perceptions that they don't get sick and find clinics unfriendly, as primarily women's spaces. As a result, in the PopART community randomized trial of treatment as prevention in Zambia and the western Cape of South Africa, a larger proportion of HIV+ men didn't know their status (posters Shanaube abstract 981 & Floyd abstract 982). It was also more difficult to reach men through home-based HIV testing as they were not at home when the community health workers visited. The PopART team designed strategies to reach men based on what men indicated they wanted. Community campaigns called Man Out were conducted which were multidisease campaigns, including urgent care, male sexuality, diabetes, hypertension, and eye screening. The campaigns were conducted in fishing sites, gardens, and work places to reach men, and were able to achieve high (90%) uptake of HIV testing in men.
In the PopART trial (Fidler, abstract 114, http://www.croiwebcasts.org/console/player/29603?mediaType=slideVideo&), preliminary results from home-based testing and community health worker efforts to link to care increased ART coverage from 49% to 71%, with a reduction of 43% in the proportion of those who were HIV infected by untreated. While the results are encouraging, they emphasize how difficult it is to achieve the second "90" of the UNAIDS target of 90-90-90 (90% tested, linked to care, and virally suppressed).
Linkage to and Engagement in Care
A clinic-based cluster randomized trial from Uganda assessed a strategy for earlier linkage to care (Amanyire, abstract 112, http://www.croiwebcasts.org/console/player/29601?mediaType=slideVideo&). The rationale for the trial stemmed from the observation that delays in ART initiation among patients who are ART eligible included sluggish diffusion of contemporary evidence about risks of delay (especially among symptomatic patients), time to do CD4 processing and therefore inefficient determination of treatment eligibility, and multiple pre-treatment adherence counseling norms. The intervention, which was rolled out in a stepped wedge fashion to 20 Uganda Ministry of Health clinics, was comprised of opinion-leader led teaching and coaching about risks of delayed initiation; introduction of a point-of-care CD4 testing platform (Alere PIMA); a revised counseling approach without mandatory multiple pre-initiation sessions; and quantitative clinic feedback on ART initiation rates to the clinic staff. Outcomes were ART initiation in all treatment eligible adults and HIV RNA suppression in a random sample one year after ART eligibility. Among 12,024 treatment eligible patients with a median CD4 level of 310/µl (IQR: 179-424), 79.6% met the primary outcome of ART initiation two weeks after eligibility in the intervention condition vs. 37.7% in the control condition (risk difference (RD): 41.9%, 95% CI: 40.1%-43.8%, p<0.0001).  Same-day ART initiation rose from 18.3% to 70.8% (RD: 52.5%, 95% CI: 50.7%-54.3%, p<0.0001).  Among the 335 patients (76% of those randomly selected) in whom HIV RNA was successfully obtained one year later, suppression was 86.2% in intervention and 70.6% in control condition (RD=15.6%, 95% CI: 4.4%-26.7%, p=0.0078); however, suppression was not significantly different if those who did not have an RNA obtained were counted as not suppressed. Thus, this multi-component intervention targeting health care worker behavior doubled the probability of ART initiation 14 days after eligibility and suggested improved HIV RNA suppression.
  In another study (Hoffman, abstract 113LB, http://www.croiwebcasts.org/console/player/29602?mediaType=slideVideo&) another trial of linkage to care and ART initiation was presented. The trial individually randomized 2398 persons testing positive for HIV to standard of care linkage, immediate point-of-care (POC) CD4 count, combined POC CD4 plus care facilitation, or POC CD4 plus transport reimbursement. There was no difference in entry into care within 90 days across the arms, and only 32% overall had documented entry to care within 90 days and only 15% initiated ART. The data emphasize that knowing one's status is not enough and true linkage to care is very challenging.
A third study (Rosen, abstract 28, http://www.croiwebcasts.org/console/player/29456?mediaType=slideVideo&) was a randomized trial of initiating ART at a patient's first clinic visit, similar to the Uganda study above. This trial was done in South Africa and involved randomizing ART eligible patients to rapid or standard ART initiation arms, at an individual level. Immediate initiation was facilitated by point of care tests, including CD4 testing to determine treatment eligibility. The study was conducted in two public sector clinics in Johannesburg, South Africa from 2013-2015 with all study procedures performed at the enrollment visit. Outcomes were assessed through passive medical record review without further interactions with study participants; primary outcomes were initiation of ART by 90 days, initiated, retained in care, and virally suppressed by 10 months of study enrollment. A total of 229 were enrolled in the standard arm and 234 in rapid arm. The primary outcome of ART initiation by 90 days was achieved by 72% in the standard arm and 97% in the rapid arm, for a risk difference of 25% and a RR of 1.36 (95% CI 1.24-1.49). All five patients who didn't initiate in the rapid arm were lost in the course of workup for a potential TB diagnosis. The median time in clinic between study enrollment and ARV dispensing in rapid ART arm was 2.4 hours for all the steps. The outcome of ART initiation by ≤90 days, retained and virally suppressed by 10 months was achieved by 51% in the standard arm and 64% in the rapid arm, for a risk difference of 13% and a RR of 1.26 (95% CI 1.05-1.50). Most of the loss to follow up in the standard ART initiation arm was before ART initiation and was after ART initiation in the rapid ART initiation arm. The effect was greater in the primary health clinic than in hospital based HIV clinic due to bigger room for improvement in the primary health clinic, which is important given that most South Africans receive their HIV care through primary care. This study demonstrated that it is possible to initiate nearly all eligible patients on ART with three-quarters starting ART on the same day. Same day ART initiation improved ART initiation by 36% and viral suppression by 26%. Screening for TB is one of the most important aspects of rapid ART initiation to avoid delays, but 90% do not have TB in South Africa in spite of high TB/HIV co-infection rates.
From the prevention of mother-to-child transmission sector, an assessment of reaching 90-90-90 in Malawi was reported (Jahn, abstract 168LB, http://www.croiwebcasts.org/console/player/29735?mediaType=slideVideo&). Malawi was one of the first countries to implement Option B+ and thus it is an early 'Treat All' model. Program data from 717 PMTCT and ART sites in Malawi were analyzed. Between 2011 and 2015, the proportion of pregnant women with HIV who knew their status was estimated to increase from 49% to 80%, and 78% of pregnant HIV infected women nationally were on ART. Overall, these data estimate that Malawi's 90-90-90 cascade is at 80-56-48 in 2015 compared to 49-3-2 in 2011.
From the US, an analysis from the CDC assessed viral suppression from 2009-2013 (Bradley, abstract 53, http://www.croiwebcasts.org/console/player/29468?mediaType=audio&). Increased HIV viral suppression among US adults receiving medical care was seen between 2009 and 13. Data were from the Medical Monitoring Project with cross-sectional estimates from 23 jurisdictions and weighted percentages by year, overall and subgroups (age, ethnicity, and gender) were calculated. Overall a 2% increase in viral suppression per year seen. In the subgroups of women, ages <40 years, and African Americans, who had lower suppression rates in 2009, greater improvements were seen (>10% more suppression over the five years) than for overall (8% more suppression over the five years). The data are encouraging and may reflect increasing ART prescribing, guideline changes, reduced ADAP wait lists, and improvements in ART regimens with reduced pill burden.