Oral abstracts of the 21st International AIDS Conference

No abstract available. (Published: 27 July 2017) Abstracts of the 9th IAS Conference on HIV Science (IAS 2017)  Journal of the International AIDS Society  2017,  20 :22253 http://www.jiasociety.org/index.php/jias/article/view/22253  |  http://dx.doi.org/10.7448/IAS.20.6.22253

Background: HIV persistence during antiretroviral therapy (ART) is a substantial obstacle to HIV cure. HIV-infected cells can undergo clonal expansion, and specific clones may be highly expanded. We previously reported one provirus integrated in the HORMAD2 gene that accounted for 20-40% of all proviruses in one patient. Mechanisms by which clones emerge and persist over time are uncertain. To investigate the dynamics of HIV clonal expansion, we developed multiplexed droplet digital approaches (ddPCR) to quantify HIV proviruses, including specific integrants, prior to and following prolonged ART. Methods: HIV-infected ART-naive individuals (N = 11) underwent ART and were followed for a median of 13.7 years (range 4.3-16.4). Cell-associated DNA (CA-DNA) from peripheral blood lymphocytes pre-ART, during first-and second-phase viral decay and after prolonged ART was quantified using multiplexed ddPCR assays targeting HIV gag, LTR and tat/rev regions, as well as a host gene (CCR5). We designed a specific ddPCR primer set overlapping the host-HIV junction to quantify the HIV integrant in HORMAD2. Results: All patients had successful suppression of HIV RNA on ART to <50 copies/ml plasma within five months; HIV DNA copies/million CD4 + cells decreased for gag (average 15-fold), LTR (average 9.3-fold) and tat/rev (average 20-fold) regions. In 10/11 patients, the LTR:gag and LTR:tat/rev ratios increased progressively after second-phase decay (average 6-fold and 6.4-fold, respectively, p < 0. 01, paired T-test), demonstrating loss of full-length proviruses; in 1/11 patients, LTR:gag and LTR:tat/rev ratios remained stable. The HORMAD2 integrant was undetectable at pre-ART and one month and two months on ART (<1 copy in 500,000 infected cells). After one year on suppressive ART, however, the HORMAD2 integrant was present at a frequency of 30% of all infected cells and persisted for six years on ART. Conclusions: Progressive appearance of deleted proviruses is detectable in most but not all patients undergoing ART. Substantial deletion did not appear during first-or second-phase viral decay, but only after one to four years during suppressive therapy. Clonal expansion of HIV-infected cells can be rapid and sustained at stable levels during prolonged ART, suggesting that both antigen-induced clonal expansion and homeostatic proliferation maintain HIV populations.

MOAA0105
The impact of treatment duration on defective and expanded identical HIV genomes in T-cell subsets from peripheral blood and tissues Institute of Allergy and Infectious Diseases, National Institutes of Health, Human Immunology Section, Vaccine Research Center, Bethesda, USA. 9 Centre de recherche du CHUM and Department of Microbiology, Infectiology and Immunology, Université de Montréal, Montreal, Canada Presenting author email: eunok.lee@sydney.edu.au Background: Understanding the impact of antiretroviral therapy (ART) duration on HIV-1 reservoirs is critical for implementing effective curative strategies. We studied the distribution of defective viral genomes in T-cell subsets within blood and tissues over 3-18 years of effective ART. We also examined expansions of identical HIV-1 DNA sequences (EIS) to assess the contribution of cellular proliferation to viral persistence during ART. Methods: Using single-genome/proviral sequencing, we performed inter-patient analysis of 479 HIV-1 p6-RT RNA sequences from preand early-ART plasma and 2329 HIV-1 DNA sequences from naive (N), central (CM), transitional (TM), effector (EM), gut homing and lymph homing memory CD4+ T cells sorted from peripheral blood (PB), lymph node (LN) and gut tissues from 14 participants who initiated ART during chronic infection (3-18 years on ART). Defective viral sequences had hypermutation, premature stops, frameshifts and large deletions. EIS were determined as ≥2 identical intact or defective HIV-1 DNA sequences across all cell types from all anatomic sites. Results: Defective HIV-1 DNA sequences did not appear to accumulate substantially over 3-18 years of ART in any anatomic site (odds ratio = 0.90-1.02/year, p = 0.14-0.98). The viral sequences derived from pre-and early-ART plasma samples were more often genetically intact than sequences derived from PB and LN after several years of therapy (p ≤ 0.004). The odds of an HIV-1 DNA sequence belonging to EIS increased in PB by 1.09-fold/year of ART (p = 0.003). In tissues, the increase of the proportion in EIS during each additional year of ART was not statistically significant (p = 0.15-0.25). Of note, 37-71% of identical HIV-1 DNA sequences were in EM sorted from PB and tissues. The odds that a viral sequence is part of an EIS in PB-derived EM increased by 1.11fold/year (p = 0.007), whereas other cell types showed no statistically significant trend (p = 0.12-0.46). Conclusions: The proportion of proviruses that were defective did not increase during effective therapy, suggesting that they are established in cells prior to viral suppression. Our data suggest that proliferation of HIV-infected T cells increased the proportion of sequences in EIS within PB over 3-18 years of therapy, and the EM cells were the major contributor.

MOAA0201
Sequential receptor-induced conformational states of native membrane-embedded HIV-1 Env B Ivan; N Friedrich; Z Sun and A Trkola University of Zurich, Institute of Medical Virology, Zurich, Switzerland Presenting author email: ivan.branislav@virology.uzh.ch Background: To fulfil its function of mediating virus attachment and entry, the HIV-1 envelope glycoprotein (Env) trimer undergoes a sequence of conformational changes. Due to their unstable nature, entry intermediates remain challenging to study, leaving several aspects of the entry process unresolved. This includes effects of CD4 and coreceptor engagement on the formation of successive Env intermediates and the stoichiometry of receptor and coreceptor binding. Knowledge on the different conformational states of native Env during the entry process is however crucial to understand its antigenic landscape and vulnerability to neutralization. While Env-directed broadly neutralizing antibodies (bNAbs) are thought to recognize Env preferentially in its closed, unliganded form, we currently lack information on their activity across different entry steps. As activity across intermediate Env conformations could increase the window of action, an antigenic profile of receptor-bound Env needs to be defined to select the best bnAbs for therapy and vaccine design. Methods: We have developed a flow cytometry-based assay to assess and compare antigenic properties and conformational stability of fully native, cell surface-expressed Env trimers using an array of Env-directed Abs and compounds. HIV-1 pseudovirus neutralization activity of a panel of NAbs was determined to explore links between open Env conformation and neutralization sensitivity. Results: Our analysis provides a fine mapping of sequential exposure of shielded Env epitopes upon exposure to increasing doses of soluble CD4 (sCD4). Conformational changes induced by sCD4 binding proved consistent with a model where occupation of a promoter alters conformation of all three simultaneously. Increasing promoter occupancy by sCD4 promoted transition through intermediate, progressively more open Env conformations. According to our data, exposure of the coreceptor-binding site may require at least two CD4 molecules to be bound per trimer. Most intriguingly, upon saturation of CD4-binding sites, the trimer adopts a pre-hairpin conformation even in the absence of coreceptor interaction. Conclusions: We provide here novel insights on antigenic characteristics of unliganded and receptor-bound conformations of native Env, which offer means to decipher critical steps of the entry process with high relevance for Env immunogen selection and design.

MOAA0202
Optimized Env trimer immunization parameters amplify onset, magnitude and consistency of autologous Tier 2 neutralizing antibody development in nonhuman primates Background: The discovery of HIV broadly neutralizing antibodies in HIV-positive individuals has re-galvanized efforts to develop a protective, antibody-based HIV vaccine. However, generation of neutralizing antibody (nAb) responses to clinically relevant strains of HIV (Tier 2) by immunization remains a challenging problem. Furthermore, the immunological barriers to induction of such responses by Env immunogens are unclear. Methods: To directly study the germinal centre (GC) responses induced after immunization, we have now performed over 1000 lymph node fine-needle aspirates (LN FNA) of draining LNs in rhesus macaques immunized with native-like HIV-1 Env trimer proteins, such as BG505 SOSIP. LN FNA sampling captured GC cells and was highly representative of whole LN biopsies. Greater than 95% of samples provided sufficient cells for identifying the major cell types of the GC, GC B cells and GC T follicular helper (Tfh) cells. LN FNAs also afforded an opportunity for gene expression analysis of antigen-specific GC Tfh cells by RNAseq and longitudinal BCR sequencing to track evolution of the Env-specific B-cell response.
Results: A majority of immunized animals developed autologous Tier 2 neutralizing antibodies. Tier 2 nAb development was most strongly associated with the magnitude of the GC response in an initial study (p = 0.007). In a subsequent study, the GC responses predicted Tier 2 nAb development (p = 0.014). Notably, Tier 2 nAbs did not correlate with BG505 SOSIP Ab binding titres. Thus, the GC responses distinguish between nAb responder and nAb nonresponder monkeys, but ELISA binding titre does not. Conclusions: Longitudinal LN FNA sampling has provided strong evidence that GC B and GC Tfh cell responses are central to generating HIV Tier 2 nAbs by immunization.

MOAA0204
Rapid elicitation of broadly neutralizing antibodies to HIV by immunization in cows Background: Anti-HIV-1 non-neutralizing antibodies (nnAbs) capable of antibody-dependent cellular cytotoxicity (ADCC) have been identified as a protective immune correlate in the RV144 vaccine efficacy trial. Broadly neutralizing antibodies (bNAbs) may also mediate ADCC and rely on their Fc region for optimal efficacy in animal models. Methods: Here, we selected 10 bNAbs and 9 nnAbs, targeting various epitopes and conformations of the gp120/41 complex, and analysed the potency of the two types of antibodies to bind and eliminate HIV-1-infected cells through NK engagement in culture. We tested their activity against 18 HIV-1 strains, including primary viruses. Results: The landscape of Env epitope exposure at the surface and the sensitivity of infected cells to ADCC vary considerably between viral strains. The most potent bNAbs, and not the nnAbs, bound to reactivated infected cells from HIV-positive individuals and mediated effective ADCC against those cells. The nnAbs also modestly recognize cells infected with eight different transmitted founder (T/F) isolates. Efficient ADCC requires sustained cell surface binding of bNAbs to Env proteins, and combining bNAbs allows a potent killing activity. Addition of a synthetic CD4 mimetic enhanced the binding and killing efficacy of some of the nnAbs in an epitope-dependent manner, without reaching the levels achieved by the most potent bNAbs. Conclusions: Our study reveals important qualitative and quantitative differences between bNAbs and nnAbs, delineates the parameters controlling ADCC activity of bNAbs and supports the use of the most potent antibodies to clear the viral reservoir.

MOAA0206
Allosteric regulation in human anti-HIV-1 Env ADCCmediating antibodies upon immune complex (IC) formation enhances the binding to FcγRs for the activation of cytotoxicity against HIV-1 virus Background: The field of HIV-1 vaccine research is in rapid development, and a deeper knowledge of the mechanisms of defence against HIV-1 infection is urgent. In this regard, key insights about antibody-mediated immune response were provided by the RV144 vaccine trial that showed moderate protection correlating with Fceffector functions. Historically, the activation of antibody-dependent cellular cytotoxicity (ADCC) has been considered dictated by two distinct and critical steps: optimal binding of viral antigens and engagement of FcγRs, respectively, occurring in two distinguished mAb regions (Fab and Fc domains). Only recently, few studies suggested a functional connection of Fab and Fc regions within the IgG. Methods: Utilizing multiple approaches, such as ELISA, fluorescence correlation spectroscopy, hydrogen/deuterium exchange mass spectroscopy (H/DX MS) and crystallography, we studied whether the immune complex (IC) formation by antigen (Ag) binding may impact the ability of IgG to consequently engage FcγRs and, in turn, activate the cytotoxic immune response against HIV-1sensitized targets. Results: Here, we demonstrate a reciprocal allosteric regulation in anti-HIV gp120 Cluster A mAbs resulting from IC formation with a gp120-CD4 chimera antigen. We establish that the formation of ICs dramatically increases the efficiency of mAbs interaction to lowaffinity FcγRs compared to free IgG. Moreover, antigen binding reverses the effect of FcγRs binding-attenuating LALA mutations in the Fc region, resulting in residual ADCC activity. H/DX MS analysis reveals an allosteric increase in conformational dynamics in the Fc domain of wildtype IgG upon Ag binding, highlighting a putative mechanism for effective Fc receptor engagement. In line with crystallographic data, H/DX MS showed a higher flexibility in the LALA mutant. However, antigen binding further stabilized Fab and Fc regions of LALA mAbs, maintaining an adequate level of flexibility in the same crucial residues observed in the wt-IgG-Ag complex. Of note, the more mobile residues map to non-contact regions for FcγRs but include residues that indirectly affect the IgG-FcγRs affinity and the induction of IgG multimerization. Conclusions: Collectively, these findings provide unique insights into reciprocal allosteric regulation between Fab and Fc domains, as IgG intrinsic factors that dictate the ability to tune selective Fceffector functions, such as ADCC, in vaccine regimens or HIV-1 passive treatments.

MOAB0101
Immunodeficiency at the start of combination antiretroviral therapy in low-, middle-and high-income countries MOAB0104 Forty-eight-week efficacy of a third line based on darunavir plus raltegravir regimen in HIV-infected adults who failed second-line protease inhibitor-based regimen in sub-Saharan Africa, ANRS 12269 THILAO study Methods: In the period of March-June 2016, we provided HTS to children and adolescents 18 months to 19 years using the following models: HTS outreaches to dwelling homes of orphans and vulnerable children (OVC); Know Your child Status Campaigns (KYCS); HTS outreaches to children of female sex workers (FSWs), fisher folks (FFs) and tea plantation workers; and evening HTS points targeting adolescents after work/school hours. We summarized the HIV yield for the different models in proportions and frequencies.

MOAB0202
Impacts of vitamin D and calcium supplementation on bone mineral density among perinatally HIV-infected adolescents: a 48-week randomized clinical trial Background: Raltegravir is an integrase inhibitor, largely used in the recent years, but tolerance data in pregnancy are scarce. Potential teratogenicity has not yet been evaluated for this molecule in a clinical context. We aimed to describe the rates and types of birth defects among children exposed to raltegravir in utero and to study the association with trimester of exposure. Methods: EPF is a multicentre national cohort, which prospectively enrols pregnant HIV-infected women delivering in 90 centres throughout France. Children are followed by paediatricians until two years of age. All births exposed to raltegravir were included. Birth defects were defined using the EUROCAT classification. Rates of birth defects were compared according to timing of exposure to raltegravir (first trimester vs. second and third trimesters) using χ 2 tests. Results: We included 479 foetuses born between 2008 and 2015, exposed to raltegravir, among which 6 stillbirths (1.3%) and 2 late miscarriages (0.4%). There were no terminations of pregnancies for birth defects. Rates of birth defects were 4.2% for all births (20/ 479 [95% confidence interval 2.4-6.0]) and 4.2% among live births (20/471 [2.4-6.1]). This incidence was similar to that reported in a previous study in EPF for live births exposed to any ARV (4.4% [4.0-4.7]). Birth defect rates did not differ significantly between firsttrimester exposure to raltegravir (5.7%; 8/140) and second-or third-trimester exposure (3.5%; 12/339; p = 0.32). The anomalies did not follow any pattern and concerned various organs: seven heart defects, five polydactylies and eight other defects. Other notable adverse outcomes were preterm births (14.2%), and two cases of HIV perinatal infection (0.4%). The follow-up was complete to 24 months for 63% of children. For 15% of children, only the birth questionnaire was available.
Conclusions: This is the largest prospective cohort of children exposed in utero to raltegravir with homogenous evaluation of birth defects. We did not find a significant association between first-trimester exposure to raltegravir and birth defects. This finding is quite reassuring as this molecule is often prescribed to women of child-bearing age, and thus, many children may be exposed in the first trimester of pregnancy.

Screening for sexually transmitted infections (STIs) included
Chlamydia trachomatis, Neisseria gonorrheae and herpes simplex virus-2 (HSV2) infections and syphilis. Multiple logistic regression analysis was used to assess factors associated with presence of any 18,31,33,45,52 and 58) at any body sites.
Conclusions: Half of Asian PHIVA and HIV-uninfected female adolescents in our cohort had high-risk HPV infection. Greater access to HPV vaccination is needed in the region to reduce future HPVrelated cancer risk.

MOAB0301
Hepatitis C care cascade for people living with HIV in the country of Georgia and Gilead Sciences, the country launched a national hepatitis C elimination programme, which provides free treatment with modern direct-acting antivirals (DAAs). Methods: The following steps of HCV care cascade were quantified: (1) HIV/HCV co-infected, (2) Diagnosed for both HIV and HCV, (3) treated for HCV infection and (4) achieved sustained virologic response (SVR). The number of HIV/HCV co-infected persons was estimated using modelling and observed HCV prevalence. Data on diagnosed persons were extracted from the national AIDS health information system as of 1 September 2016. Results: Among estimated 3300 persons living with HIV/HCV coinfection in Georgia, 2201 (67%) were not aware of their HIV status and 1099 (33%) were diagnosed with both HIV and HCV. Of those 1099 diagnosed persons, 697 (63%) were treated for hepatitis C with either PEG/RBV-or DAA-based regimen; 480 (69%) of those treated achieved SVR. Rates of SVR were 44% with PEG/RBV and 89% with DAA. Overall, because of gap in diagnosis stage, only 15% of estimated number of HIV/HCV co-infected persons were cured.
Conclusions: The major gap in the HCV care cascade is at the stage of diagnosis resulting from deficiencies in HIV diagnosis. Reducing the number of people living with undiagnosed HIV/HCV co-infection will be critical for achieving population-level impact of free HCV treatment programme.
times. Deaths due to ESLD declined by approximately twofold and were no longer the most common cause of death in the 2010-2016 time period for either age category. In contrast, smoking-related deaths increased with time and, among those aged 50-80, accounted for the greatest proportion of deaths from 2010 to 2016 (cause-specific HR 2.8; 95% CI, 1.5-5.4). Drug overdose accounted for the greatest proportion of deaths among individuals aged 20-50 from 2010 to 2016. Conclusions: Increased HCV treatment uptake has coincided with decreased liver-specific mortality in HIV-HCV-co-infected patients. However, these gains may be thwarted if modifiable risk factors (tobacco and drug use) are not addressed.

MOAB0303
Efficacy and safety of glecaprevir/pibrentasvir in patients co-infected with hepatitis C virus and human immunodeficiency virus-1: the EXPEDITION-2 study Background: Pangenotypic, once-daily glecaprevir (identified by AbbVie and Enanta)/pibrentasvir (G/P) has demonstrated high rates of sustained virologic response at 12-week post-treatment (SVR12) in patients with hepatitis C virus (HCV) genotype (GT) [1][2][3][4][5][6] infection. This phase 3 study evaluated the efficacy and safety of G/P in patients with chronic HCV GT1-6 infection and HIV-1 coinfection, including patients with compensated cirrhosis. Methods: EXPEDITION-2 (NCT02738138) is a phase 3, multicentre open-label study evaluating G/P (300mg/120mg) treatment in HCV/HIV-1-co-infected adults without or with compensated cirrhosis for 8 or 12 weeks, respectively. Patients were either HCV treatment naive or experienced with interferon (IFN), pegylated IFN ± ribavirin or sofosbuvir + ribavirin ± pegylated IFN. GT3 treatment-experienced patients were excluded. The primary endpoint was the proportion of patients with sustained virologic response (HCV RNA < lower limit of quantification) 12-week posttreatment (SVR12). Results: In total, 153 patients were enrolled, including 16 (10%) with cirrhosis. Baseline demographics are shown in Table 1. In patients with available data, rates of response at end of treatment and post-treatment week 4 were 98.7% (151/153) and 98.6% (144/ 146), respectively. To date, there is one (1/153; 0.65%) virologic failure: a breakthrough in a patient with GT3a infection and cirrhosis. The most common adverse events (AEs) were fatigue (16/ 153; 10%) and nausea (13/153; 8%). Three patients (2%) had serious AEs, and one serious AE of stroke led to treatment discontinuation on day 23 in one patient with cirrhosis; all were unrelated to G/P. All patients maintained HIV-1 suppression (<200 copies/ml) during treatment.
Conclusions: G/P for eight weeks in non-cirrhotic and 12 weeks in cirrhotic patients is a well-tolerated and highly efficacious pangenotypic treatment for HCV/HIV-1 co-infection, regardless of baseline HCV RNA or treatment experience. Full SVR12 rates and prevalence of baseline NS3 and NS5A polymorphisms will be presented.

. (Continued)
Background: Metabolic syndrome (MetS) has become a common finding in HIV-infected patients. However, the severity, risk factors and pathogenesis of liver fibrosis in this population have been poorly documented. From a matched cohort of HIV-monoinfected patients with and without MetS, this study aimed (1) to assess the impact of MetS on the prevalence and severity of liver fibrosis and (2) to analyse the association between liver fibrosis and markers of adipose tissue, insulin resistance and macrophage activation.
Methods: Patients with immune-controlled HIV-1 infection under antiviral therapy (ART) were enrolled in the following exposed-unexposed study. The exposure was defined by the presence of MetS according to international criteria after exclusion of all other causes of chronic liver disease. Fibrosis was assessed using transient elastography (Fibroscan). Adipokines, HOMA index and soluble CD163 and CD14 were measured as markers of fat mass, insulin resistance and macrophage/monocyte activation, respectively. Results: A total of 468 HIV-monoinfected individuals were enrolled (male (89%), mean age 53 (9) years, mean body mass index 24.6 (5.3) kg/m 2 ): 246 with MetS and 222 without MetS. Patients with MetS were older and 49% of them had insulin resistance, that is, HOMA-IR ≥2.5 (compared to 8.5% in patients without MetS). The mean value (SD) of liver stiffness measurement (LSM) was 5.6 (2.2) kPa with a minimum and maximum value of 2.4 and 17.1 kPa. Mean LSM was higher in patients with MetS compared to those without MetS (6.3 (2.6) versus 4.9 (1.5) kPa, p < 0.0001). In multivariable analysis, obesity (odds ratio: 3.9 (95% confidence interval 2.1-7.1)) and insulin resistance (1.1 (1.06-1.2)) were independent factors of significant fibrosis (≥F2) and remained associated after adjustment on MetS. Serum levels of adipokines and sCD163 were significantly associated with the degree of liver fibrosis. When adjusted on MetS, leptin and sCD163 remained strongly associated to fibrosis. HIV parameters and ART regimen were not associated to fibrosis.
Conclusions: In HIV-monoinfected patients, MetS is an important risk factor for liver fibrosis. Obesity and insulin resistance are key factors in the development of liver fibrosis independently of HIV infection parameters. Adipose tissue and macrophage activation certainly play an important role in the development of fibrosis in HIV-monoinfected patients, but the exact mechanisms need to be elucidated.

MOAB0305
Predictor factors associated with liver fibrosis and steatosis by transient elastography in HIV-monoinfected patients under long-term combined antiretroviral therapy: the PROSPEC-HIV study H Perazzo Pedroso Barbosa; S Cardoso; C Yanavich; JC Soares; J Fittipaldi; M Morata; C Cardoso; P Simplicio; C De Almeida; V Veloso and B Grinsztejn Fundação Oswaldo Cruz, LAPCLIN-AIDS, Rio de Janeiro, Brazil Presenting author email: perazzohugo@gmail.com Background: Liver disease remains one of the main causes of non-AIDS mortality in HIV-infected individuals. Transient elastography (TE) is an accurate imaging method to estimate liver fibrosis and steatosis. We aimed to evaluate the risk factors associated with liver fibrosis and steatosis in HIV-monoinfected patients under long-term combined-antiretroviral therapy (c-ART). Methods: This cross-sectional study prospectively included HIVinfected adult patients under c-ART (PROSPEC-HIV; NCT02542020). Liver stiffness measurement (LSM) and controlled attenuation parameter (CAP) by TE were used to estimate liver fibrosis and steatosis, respectively. Exclusion criteria were hepatitis co-infection and c-ART naive. Patients with an unreliable M probe LSM or CAP were excluded from the liver fibrosis and steatosis analyses, respectively. Clinical evaluation, fasting blood tests and TE were performed at the same day. TE exams were performed by a single experimented operator blinded to clinical and laboratory data. Metabolic factors were defined according to the International Diabetes Federation criteria. Alcohol consumption was quantified using the AUDIT score. Presence of liver fibrosis and steatosis was considered when LSM ≥ 8.0 kPa and CAP ≥ 250 dB/m, respectively. Age-and genderadjusted multivariate logistic regression was performed.
Conclusions: Non-communicable diseases (NCD) can play a major role in the development of liver fibrosis and steatosis. NCD prevention and care services need to be integrated to HIV care to decrease the burden of hepatic events in HIV-infected individuals.

MOAB0401
Gaps and opportunities in policy and practice in 20 countries with the highest burden of HIV-associated TB mortality among people living with HIV (PLHIV). In June 2016, Member States adopted the UN Political Declaration on HIV and AIDS, including a target to reduce TB deaths among PLHIV by 75% by 2020. In order to achieve this, gaps in policy, implementation and recording and reporting need to be identified and urgently addressed.
Methods: A total of 20 countries were selected with the highest estimated burden of HIV-positive incident TB in 2015. Data on collaborative TB/HIV activities were extracted from the Global TB Programme Database and the UNAIDS online GARPR tool, and trends and progress analysed. Reviews were further conducted of the latest available policy documents, programme reviews, epidemiological assessments, Global Fund Concept Notes and GARPR reports to identify gaps and opportunities in policy and implementation.
Results: Cascade analysis revealed a 57% gap in notification of TB patients living with HIV, compared with estimated cases, and a 19% gap of ART started among notified HIV-positive TB patients ( Figure 1). Half of countries did not report isoniazid preventive therapy (IPT). Case fatality among TB patients living with HIV was at least three times higher than HIV-negative TB patients in seven reporting countries. Document review revealed the following gaps and barriers: misalignment of policies on ART and IPT; poorly implemented TB screening and IPT; centralized or underuse of Xpert MTB/RIF; centralized ART provision; stock-outs in IPT, HIV testing kits and ART; and separate planning, supervision, health management information systems and procurement and supply.
Conclusions: Considerable gaps and opportunities were identified in this analysis. Countries need to seek ways to resolve barriers, be they policy, implementation or health system-related to ensure access to evidence-informed HIV-associated TB care and to end HIV-related deaths from this preventable disease.

MOAB0402
Genomic epidemiology of extensively drug-resistant tuberculosis in KwaZulu-Natal, South Africa: demographic expansion and genetic determinants of epidemiologic success in a high HIV prevalence setting Background: Extensively drug-resistant tuberculosis (XDR-TB) has emerged over the last decade as a significant public health threat worldwide, particularly among people with HIV. South Africa first reported XDR-TB in 2005 and now has among the highest burden of XDR-TB worldwide, with >1000 cases diagnosed in 2015. The bacterial evolutionary determinants behind the rise of XDR-TB in South Africa are not well understood. Methods: We enrolled persons with newly diagnosed, cultureconfirmed XDR-TB from 2011 to 2014 in KwaZulu-Natal province and performed whole-genome sequencing of their Mycobacterium tuberculosis (Mtb) isolates. Lineage 4 isolates were selected for phylogenetic reconstruction, dating of drug-resistance mutations and estimates of prior demographic history using Bayesian Markov chain Monte Carlo Methods (BEAST v1.8.3).
Results: Among 160 participants with XDR Mtb isolates included in this analysis, 127 (79%) were HIV co-infected. Half (51%) of XDR-TB cases were attributable to a single predominant clade of highly monomorphic isolates (Restriction Fragment Length Polymorphism type HP). There were no significant differences between the proportion of participants with HIV or with CD4 counts <200 cells/μl in the HP vs. non-HP isolates. Both HP and non-HP Mtb populations exhibit evidence of rapid population expansion beginning 25-30 years ago ( Figure 1b). The emergence of key drug-resistance mutations occurred near the historical dates of introduction for their corresponding antibiotics ( Figure 1a). Conclusions: Mtb isolates from the predominant HP clade and from less prevalent XDR-TB strains underwent demographic expansion following the onset of the HIV epidemic in South Africa. Endemic strains in KwaZulu-Natal acquired drug resistance mutations across diverse strain groups and genetic backgrounds, corresponding to the introduction of new antituberculosis medications. The impact of HIV co-infection on pathoadaptive evolution in Mtb remains an important area for further investigation.

MOAB0403
Incidence of tuberculosis in the first year of antiretroviral treatment in West-African HIV-infected adults Background: Despite evidence that isoniazid preventive therapy (IPT) reduces tuberculosis incidence and mortality, its uptake remains very limited in West Africa. Our objective was to assess tuberculosis incidence during the first year of antiretroviral therapy (ART) and identify associated factors in HIV-infected adults in West Africa to support policy decisions in these countries. Methods: We conducted a retrospective observational cohort analysis using data collected in three HIV outpatient centres from Côte d'Ivoire, Burkina Faso and Senegal participating in the IeDEA West Africa Collaboration. We included HIV-infected adults (≥16 years) initiating ART between 2010 and 2014, without tuberculosis diagnosis at ART initiation and with ≥1 follow-up visit. None of them received IPT. Tuberculosis new diagnoses were documented according to national recommendations. We analysed incidence of tuberculosis and identified associated factors using Poisson regression models. Results: Of 4154 patients who started on ART, 3404 had ≥1 followup visit after ART initiation. Of those, 191 (5.6%) had ongoing tuberculosis at ART initiation. Thus, we enrolled 3213 patients in our analysis. Median age was 38.5 (interquartile range (IQR) 32. 0-45.4) years, 67.1% were female and median CD4 count at ART initiation was 211 (IQR 95-343) cells/mm 3 . Overall 170 new tuberculosis cases were reported for 2360.5 person-years (PY) at risk. The crude tuberculosis incidence rate during the first year on ART was 7.2 (95% confidence interval (CI) 6.12-8.28) cases per 100 PY. The adjusted tuberculosis incidence rate was  per 100 PY in women, aged 16-30 years, without prior tuberculosis history, with CD4 ≥500 cells/mm 3 and haemoglobin ≥11 g/dl, followed in Burkina-Faso. A higher tuberculosis risk was significantly associated with male gender (relative risk (RR) 1.87; 95% CI 1.28-2.74), previous history of TB (RR 4.22; 95% CI 2.70-6.42), haemoglobin <9 g/dl (RR 2.26;) and followup in Côte d'Ivoire (RR 4.32; 95% CI 2.90-6.50).
Conclusions: Tuberculosis incidence remains high during the first year on ART in the West African context in the absence of IPT. It is crucial to reinforce implementation of IPT in all HIV-infected adults starting ART, especially in this part of the world.

MOAB0404
Time to treatment initiation for drug-resistant tuberculosis is delayed in a South African prospective cohort BJ Sullivan 1,2 ; J Prvu Bettger 1,2,3  Background: Drug-resistant tuberculosis (DR-TB) is a growing threat to TB management and elimination globally. Early initiation of DR-TB treatment is critical to successful treatment outcomes (cure) and to prevent transmission. The South African National TB Programme (NTP) recommends initiating treatment within five days of diagnosis. This study examined time to treatment initiation and its relationship with patient age among a prospective cohort of individuals living with DR-TB in South Africa. Methods: Patients included subjects aged 13 years or older, enrolled in a DR-TB cluster-randomized trial in two South African provinces. Outcomes were treatment initiation within five days of DR-TB diagnosis and days from diagnosis to treatment initiation. Hierarchical mixed-effects models were employed to examine the association between age and these outcomes, adjusted for patient (sex, TB history and HIV co-infection) and site (rural/urban and province) characteristics and random effects of treatment centre. Results: Of 521 patients, there were 55% male, 75% with HIV coinfection and 53% with prior history of TB. Only 82 patients (16%) received DR-TB treatment within five days of diagnosis. The median time to treatment was 11 days (range = 0-180). Age was not associated with treatment initiation within five days (F = 0.07, df = 1495, p = 0.794) or days to treatment initiation (F = 1.42, df = 1489, p = 0.233). Individuals co-infected with HIV tended to have a greater likelihood of receiving treatment within five days relative to those without co-infection (17% vs. 12%, p = 0.104; odds ratio = 1.749, 95% confidence interval = 0.891-3.433).
Conclusions: Delays in DR-TB treatment increase harm to the patient and risk of disease spread. Only one in six patients with DR-TB received treatment within five days of being diagnosed as recommended by the South African NTP guidelines. Further research is needed to examine what modifiable factors decrease treatment delay and how to most effectively implement treatment guidelines.

MOAB0405
High uptake of antiretroviral therapy among HIV-positive TB patients receiving co-located services in Swaziland HIV services, but the timing of antiretroviral therapy (ART) initiation relative to TB treatment -a marker of programme quality and predictor of outcomes -is unknown. Methods: We conducted a retrospective review of programmatic data from 11 purposefully sampled facilities to evaluate provision of timely ART for adult (≥15 years) HIV-positive TB patients seen and retained in care between July and September 2014. Timely ART was defined as initiated within two weeks of TB treatment initiation for patients with CD4 <50/μl or missing and within eight weeks for others. Descriptive statistics were estimated and logistic regression was used to assess factors independently associated with timely ART.

MOAD0101
Readiness for antiretroviral therapy: implications for linking HIV-infected individuals to care and treatment Background: Antiretroviral therapy (ART) readiness is a key predictor of ART initiation. However, there is a paucity of data on ART readiness among individuals at the time of HIV diagnosis and ART eligibility assessment. Under a test-and-treat approach, understanding factors associated with ART readiness can inform strategies to support early engagement in care and thereby maximize the benefits of ART. This study examined demographic and psychosocial factors associated with ART readiness and potential barriers to linkage to care among individuals referred for treatment from a mobile health clinic. Methods: Between April 2015 and May 2016, 87 individuals (18 years and older) in a resource-limited setting in Cape Town, South Africa, completed a face-to-face survey immediately after referral for ART. ART readiness was assessed using key components of this concept identified in the literature: (1) an awareness that treatment will be beneficial; (2) motivation to initiate treatment; and (3) the intention to start treatment soon. Multiple logistic regression analysis, controlling for age, gender and education, identified factors associated with ART readiness.
Results: Most participants were very ready (84%) and motivated (85%) to start ART, but 28% reported some uncertainty regarding ART initiation. Treatment readiness was lower among those surprised by their diagnosis (adjusted odds ratio (aOR): 0.26, p < 0.05) and among healthier individuals (aOR: 0.44, p < 0.01). In contrast, higher readiness was associated with better ART knowledge (aOR: 4.31, p < 0.05) and knowing someone who had experienced positive health effects from ART (aOR: 2.65, p < 0.05). The three most common self-reported barriers to linking to care were (1) not wanting to be seen at the clinic (31%); (2) no money for transport (29%); and (3) not being able to get time off work (20%).
Conclusions: Results indicate that post-test counselling will need to be designed to deal with surprise at HIV diagnosis and that health messaging needs to be carefully crafted for HIV-positive but healthy individuals to improve ART readiness and to increase likelihood of further linkage to treatment and care. Further research is needed on effective post-test counselling approaches (e.g. motivational-interviewing) and effective framing of health messaging to increase awareness of the positive benefits of early ART initiation and corresponding motivation to engage in treatment.

MOAD0102
Factors associated with loss to follow-up in a primary healthcare HIV clinic practising test and treat Conclusions: This study identified testing and initiating on ART instantly being associated with elevated risk of LTFU and as well TB suspicion and advanced disease at enrolment. In a bid to achieve the 90-90-90 campaign therefore, steep ART initiation should be backed by intensive pre-initiation and adherence counselling for better long-term retention of patients.

MOAD0103
Pilot study of a multipronged intervention using social norms and priming to improve adherence to antiretroviral therapy and retention in care among adults living with HIV in Tanzania  Background: Interventions incorporating insights from behavioural economics and psychology have been successfully used in the private sector and have the potential to enhance HIV "treatment as prevention" (TasP). This approach recognizes that decisions are influenced by emotions, contexts and decision-making shortcuts outside of conscious awareness. To test this hypothesis, we evaluated an intervention to improve antiretroviral therapy (ART) adherence and retention in care based on concepts of social norms and priming. Methods: We used patient-centred design to develop a combination intervention using social norms and priming, which is when a stimulus subconsciously or indirectly influences another behaviour. The intervention included visual feedback about cliniclevel retention in care (social norms), a self-relevant prime and useful take-home items with the priming image (calendar, pill box). The intervention was implemented at two HIV primary clinics in Shinyanga, Tanzania, in two-week intervals for six months. We conducted a quasi-experimental pilot study (Clinicaltrials.gov: NCT02938533) by reviewing the medical records of a random sample of 438 adult patients living with HIV infection (PLHIV, 320 exposed and 118 unexposed) to compare retention and the proportion of patients achieving a medication possession ratio (MPR) ≥95% after six months. Intervention acceptability was determined through an in-person survey with a convenience sample of 405 PLHIV at baseline (n = 189) and end line (n = 216). Results: In adjusted analyses, PLHIV exposed to the intervention was significantly more likely to be in care after six months (87% vs. 79%, adjusted odds ratio (OR a ) =

MOAD0205
Effectiveness of comprehensive HIV and stimulant use prevention intervention with Cambodian female entertainment and sex workers Background: HIV prevention services for female entertainment and sex workers (FESW) could serve as a platform for targeting key risk factors for HIV infection, including amphetamine-type stimulant (ATS) use and economic distress. We examined sequentially delivered interventions to decrease ATS use and improve economic well-being to boost HIV risk reduction with FESW.
Methods: This cluster-randomized stepped-wedge trial in 10 Cambodian provinces enrolled 1198 FESW to test the effectiveness of a comprehensive HIV and ATS use prevention package leveraging SMARTGirl, an existing HIV prevention platform for Cambodian FESW. The intervention included a conditional cash transfer with cognitive-behavioural aftercare (CCT+AC) intervention to reduce ATS use followed by a microenterprise (ME) opportunity. The co-primary outcomes assessed in 600 FESW purposively targeted for the 18-month follow-up were: (1) self-reported number of sexual partners (past three months) and (2) positive urine toxicology results for ATS (ATS Tox+). After baseline, FESW with problematic patterns of ATS use were allocated to receive a fourmonth CCT+AC intervention. All FESW who were abstinent from ATS at six months were allocated to receive a ME opportunity. Background: The prevalence of HIV among people who inject drugs (PWID) in Dar es Salaam is 42% compared to 7% in the general population. In 2011, an opioid treatment programme (OTP), using methadone, was established in Tanzania to reduce HIV risk behaviours and transmission. Enrolment of PWID into the OTP programme surged, but linking and sustaining HIV-positive, eligible OTP patients in antiretroviral therapy faced many obstacles. We report the results from the first integrated methadone and antiretroviral therapy (IMAT) programme to improve sustained ART access for PWID in sub-Saharan Africa. Methods: A community engagement process with patients, providers and government stakeholders helped us to collaboratively define an integrated model, including: opt-out HIV screening, OTP providers trained in HIV clinical management and monitoring, multiple ART dispensing models from the OTP clinic and intensive case management for people who were not virally suppressed. We assessed viral suppression 6 and 12 months after implementation of the integrated model and used logistic regression to assess predictors of viral suppression (<1000 copies/ml) at 12 months. Conclusions: We have demonstrated the feasibility and acceptability of HIV self-testing in the UK. It also demonstrated that a high percentage were willing to report their results which allowed for confirmation of linkage to care. We believe that an investment in HIV self-testing will compliment existing options and provide a cost-effective way to scale up our approach to testing.

MOAX0103
The impact of HIV , pharmacy (18%), peer (14%) and/or workplace (13%). FSW indicated they would be willing to pay $0.50-$25 for self-tests, with 35% willing to pay $1 and 30% $5. Conclusions: FSW found HIVST highly acceptable and wanted HIVST to be available to them. A high proportion had a reactive self-test, and importantly, virtually everyone had linked to post-test services by the two-week follow-up questionnaire. Some expressed concern about potential for coercive testing. FSW were willing to pay for HIVST and provided useful insight into how to distribute and promote HIVST during future implementation research. HIVST represents a promising strategy to promote regular re-testing among FSW in Zimbabwe.

TUAA0101
Evaluation of memory CD8+ T cell responses in individuals initiating cART during hyperacute HIV-1 infection Additional cases (Intervention: n = 3, Control: n = 3) were identified from the DBS testing after the 12-month survey.
HIV infection are associated with persistent antigenaemia. Early initiation of combination antiretroviral therapy (cART) confers clinical benefit to HIV-infected persons, but the impact of cART on HIVspecific immune responses and the potential for recall or boosting of these responses are unknown. In this study, we evaluated the maintenance of CD8+ T cell responses longitudinally in early treated individuals with hyperacute HIV-1 subtype C infection. Methods: Samples of young females identified with acute HIV-1 infection (HIV PCR positive, antibody negative) who initiated cART very early and untreated patients were used. The magnitude, breadth and maintenance of HIV-1-specific CD8+ T cell responses were defined using IFN-gamma ex vivo ELISPOT and cultured ELISPOT. Also, the phenotype (HLA-DR, CD38 and CD127) and functional (IFN-gamma) characteristics of tetramer-specific CD8+ T cells were investigated using MHC class I tetramers and intracellular cytokine staining (ICS). Results: Early treated patients with hyperacute infection induced initial CD8+ T cell responses that coincided with a sharp drop in viraemia and an increase in CD4 counts. These early induced CD8+ T cell responses were however low in magnitude (144.3 SFC/million PBMC) when compared to untreated patients (316.6 SFC/million PBMC (p = 0.009)). Interestingly, compared to untreated patients, CD8+ T cells in early treated patients were less activated and had a high expression of CD127, thus suggesting a potential for long-term survival of these responses. Additionally, memory responses specific to HIV-1 measured at later stages (six months onwards) of infection were maintained in these treated patients as indicated by cultured ELISPOT assays. Conclusions: Summarily, our results demonstrate that early initiation of cART led to an induction of CD8+ T cell responses that were less activated and had higher potential for long-term survival. These responses were also maintained as memory responses which may be recalled rapidly upon re-stimulation with HIV-1 antigens. These data may offer insight in implementing novel therapeutic strategies in order to enhance protective immunity and promote control of viral replication post-treatment interruption.

TUAA0102
Early anti-SIV CD8 + T cell antiviral activity is associated with durable elite control of SIV infection in macaques carrying or not protective MHC alleles: the ANRS SIC study Background: How CD8 + T cells control virus during HIV infection is not understood. We hypothesized that the main effect of CD8 + T cells occurs before viral integration, due to minimal direct viral cytopathic effects. We developed a model of viral dynamics with pre-and post-integration stages to study the effect of CD8 + T cell depletion. Model predictions were tested in SIV-infected rhesus macaques (RMs) receiving integrase inhibitor raltegravir (RAL) monotherapy with or without CD8 + T cells.
Methods: Sixteen SIVmac251-infected RMs were treated with both RAL-and CD8-depleting antibody M-T807R1 (RD), or just RAL (R) and followed, with RAL treatment interrupted after 23 days. Plasma viral loads (VLs) were measured by qRT-PCR. T-cell counts and immune activation were monitored flow-cytometrically. We analysed the VLs during the first approximately 12 days following RAL initiation using a viral dynamic model including infected cells pre-and post-viral DNA integration. We fitted the model to the data using a nonlinear mixed-effect model to estimate the death rate of infected cells pre-and post-virus integration and the efficacy of RAL. Results: CD8 + T cell depletion was profound and lasted throughout RAL therapy. Depletion of CD8 + T cells led to an increase in VL prior to the start of therapy. Macaques receiving just RAL treatment had much greater decays in VL than those treated with RAL-and the CD8-depleting antibody. The latter group had small decays or rebounded early during RAL therapy. Background: Burgeoning evidence indicates a broader functional repertoire for NK cells beyond innate immunity including memory and other memory-like functions. One recent example is memorylike NK cells identified by lack of the FcR intracellular γ-signalling chain (FcRΔg-NK cells) which still require antibody to grant antigen specificity but are pre-sensitized and capable of rapid mobilization and more robust responses against viral antigens. Interestingly, FcRΔg-NK cells are initially expanded by huCMV infection as part of innate-priming, but can execute memory-like killing against other pathogens through incompletely understood mechanisms. Methods: Sixty rhesus macaques were used in this study: 21 specific pathogen free, rhCMV-; 10 rhCMV+ but otherwise experimentally naive; and 22 chronically SIVmac-infected macaques. Samples were analysed from 10 naive and 10 untreated HIVinfected human subjects. NK cell analyses were performed using polychromatic and phospho-flow cytometric phenotypic and functional assays.
Results: FcRΔg-NK cells were systemically distributed in mucosal and secondary lymphoid organs but, correlating with viral load, increased two-and fourfold in CMV+ and HIV/SIV-infected individuals, including the GI tract. CD16 and α4β7 were concomitantly upregulated in infection, suggesting that innate memory-like priming is required for both antibody-dependent functional potency and mucosal homing. FcRΔg-NK cells displayed little difference in binding affinity to virus-antibody immunocomplexes compared to traditional NK cells, but exhibit twofold more robust IFN-γ secretion and cytotoxicity, suggesting disparate signalling or activation could account for improved function. To that end, FcRΔg-NK cells showed significantly reduced expression of Helios and Eomesindicative of a broader functional repertoire and/or epigenetic modification, and clustered independently from traditional NK cells in 20-parameter analyses via multidimensional t-SNE. The γchain adaptor, Syk, was reduced or inactively dephosphorylated in FcRΔg-NK cells, but expression of ζ-chain, which is phosphorylated by adaptor Zap70, was significantly upregulated, suggesting that these cells may exploit the ζ-chain/Zap70 pathway in the absence of γ-chain/Syk to achieve greater functional potency. Conclusions: Collectively, our work presents the first description of a combinatorial mechanism of innate-priming and alternative signalling cascade to explain the functional potency of memorylike phenomena of NK cells mobilizing in the mucosae against HIV/ SIV. Future studies harnessing memory-like NK cells could create exciting modalities for both vaccine and curative therapies.

TUAA0105
The human penis is an immunologically active tissue: a preliminary study on the development of an HIV vaccine Background: HIV-1 is primarily sexually transmitted. We have shown that the human penis, including the foreskin but also the urethra, fossa navicularis and glans, is one of the main portal of entry for the virus. Unlike other mucosa, the penile immune system and mechanisms that induce a penile immune response remain unclear, most likely due to the difficulty to access human tissues. Our previous studies demonstrated that the male urethra contains macrophages, the main targets of HIV-1, as well as memory T cells. These studies relied on morphologic analyses and thus failed to provide a comprehensive phenotype. To assess the role of these mucosal immune cells that are a prerequisite to the elaboration of efficient preventive strategies against HIV-1, we characterize extensively the immune profile of immune cells of the different penile regions.
Methods: Single-cell suspensions were prepared for each region of 32 penile tissues collected from individuals undergoing transgender surgery and analysed by multi-parametric flow cytometry. The expression patterns of memory, activating and homing receptors of B and T lymphocytes were evaluated as well as that of NK cells.
In complement, the tissue distribution of each of these immune populations in the different penile compartments was also studied morphologically.
Results: In all penile compartments, CD3-/CD19+ B cells represent around 2% of CD45+ cells and >50% B cells display CD27 and FcRL4 receptors and thus harbour a memory phenotype. However, <5% are IgG+ or IgA+ and thus able to secrete antibodies in the lamina propria. TCD4+ and TCD8+ lymphocytes represent the major populations of CD45+ cells, with 90% with a CD38-/HLADR-/CCR7-/ CD45RA-resting effector memory phenotype (T EM ). These resting T EM cells reside in all penile region epithelium and lamina propria but lack CD103+ resident phenotype. Furthermore, all penile compartments contain low numbers of CD3-/CD56+ NK cells capable of antibody-dependent cell cytotoxicity and surface expressing the NKp44 receptor indicative of activation. Conclusions: Altogether, the human penis is an immunologically active tissue including the cellular machinery required to induce/ produce a specific and effective immune response against mucosal pathogen. This must be taken considered when elaborating efficient vaccine strategies against HIV-1.
Results: A total of 1141 patients were randomized and treated (N = 763 D/C/F/TAF vs. N = 378 control). Baseline characteristics: median age 46 years; 18% women; 25% non-white (21% black); 10% CD4 + <350 cells/mm 3 ; and 71%, 22% and 8% on darunavir, atazanavir and lopinavir, respectively (15% on cobicistat). Cumulative virologic rebound was 1.8% (n = 14 D/C/F/TAF) vs. 2.1% (n = 8 control), of which 10/14 and 5/8, respectively, resuppressed (<50 c/ml) by week 24; there were no confirmed rebounds ≥200 c/ml. At week 24, the FDA snapshot analysis showed that virologic suppression (VL <50 c/ml) was 96.3% (D/C/F/TAF) and 95.5% (control), and virologic failure occurred in 0.5% and 0.8%, respectively, with no discontinuations for virologic failure and no detected resistance to any study drug. Safety was similar between arms through 24 weeks, with low incidences of grade 3-4 adverse events (AEs) (D/C/F/TAF 4.5% vs. control 4.5%), serious AEs (  follow-up of participants on PrEP (range two to four years), which included regular testing for HIV and other STIS, allows assessment on whether effectiveness is maintained in the longer term and the extent of potential exposure to HIV. Methods: PROUD was a pragmatic trial in which MSM were randomized to receive daily TDF/FTC either immediately (IMM) or after a deferral (DEF) period of 12 months. Main efficacy findings were based on follow-up during the deferred phase when IMM had access to PrEP and DEF did not. Since November 2014, when all participants were offered PrEP, the trial has entered a postdeferred phase. We compare incidence rates of HIV and selective STIs during the deferred and post-deferred phases.
Results: A total of 524 (269 IMM, 255 DEF) and 449 (244 IMM, 205 DEF) participants contributed to the deferred and post-deferred phases. Of 368 who attended a clinic in the last six months of follow-up, 327 (89%) had at least one PrEP prescription. HIV and rectal gonorrhoea (rGC)/chlamydia (rCT) incidence in each phase is shown by group in the Table 1. There was no difference in HIV incidence between the groups in the post-deferred phase (p = 0.18), but a significant decrease in the DEF group once they had access to PrEP (p < 0.0001). The rate in the IMM group remained similar in the two phases (p = 0.66). The incidence of rectal infections was high in both groups and phases. rCT was lowest in the DEF group during the deferred phase, and this was driven by those who did not report rCT in the year before enrolment.
Conclusions: The reduction in HIV incidence in the DEF group confirms the remarkable effectiveness of TDF/FTC. The relatively stable incidence in the IMM group indicates that this effect is durable. High ongoing incidence of rCT/rGC shows that participants remained at high risk of HIV, and this needs to be taken into account when planning PrEP provision in public health programmes.

TUAC0102
On-demand PrEP with TDF/FTC remains highly effective among MSM with infrequent sexual intercourse: a substudy of the ANRS IPERGAY trial The ANRS IPERGAY trial demonstrated among MSM a 86% relative reduction of HIV-1 incidence with on-demand PrEP in the TDF/FTC arm (2 infections, 219 person-years (py) of follow-up (FU), incidence: 0.91/100 py) as compared to the placebo arm (14 infections, 212 py of FU, incidence: 6.60/100 py). Participants in this trial used a median of 15 pills/month and had a median of 10 sexual intercourse/month. We wished to investigate whether ondemand PrEP remained effective among participants having less frequent sexual intercourse and using fewer pills. Methods: Assuming that participants with less frequent sexual intercourse would use fewer pills, and because individual patterns of pill use showed large intra-participant variability over time, we focused our analysis on person-time between two consecutive visits when participants used ≤15 pills/month and PrEP was used "systematically or often" during sexual intercourse and not "from time to time or never". We then cumulated in each arm FU time spent with this pattern of pill use. A fourth-generation HIV-1/2 ELISA assay was performed at each visit allowing to date the time of HIV infection. Incidence rates of HIV infection/100 py in both arms were then compared using mid-p exact test. Results: Six HIV-1 infections occurred during FU among participants using ≤15 pills/month taken "systematically or often" during sexual intercourse: 6 in the placebo arm (incidence: 9.3/100 py, total FU time: 64.8 py) and 0 in the TDF/FTC arm (incidence: 0/100 py, total FU time: 68.9 py, p = 0.013). The relative reduction of HIV Background: Long-acting (LA) injectable formulations of antiretroviral agents are being developed for HIV-1 prevention. The MWRI-01 phase 1 study was undertaken to characterize the safety, acceptability, pharmacokinetic (PK) and pharmacodynamic (PD) profile of LA rilpivirine (RPV). Single-dose (SD) data have previously been reported (McGowan I et al. Lancet HIV 2016). We now present data on the multiple-dose (MD) phase of the study. Methods: HIV-1-uninfected participants received three intramuscular doses of 1200mg LA RPV at two-month intervals. We collected plasma, genital/rectal fluids and tissue (rectal (RT), cervical (CT) and vaginal (VT)) before and after exposure to LA RPV for assessment of PK and PD (ex vivo biopsy challenge with HIV-1). Clade B (HIV-1 BaL ) and Clade C (G147-1) viruses were used separately in the explant challenge model. The primary study objective was to characterize product safety, and the analysis included all enrolled participants. Results: We enrolled eight women and four men. There were 195 adverse events reported, of which 193 (99%) were grade 1 (71%) or grade 2 (28%), and the majority were related to injection-site discomfort. Table 1 provides the PK values (geometric mean; 90% confidence interval) for each compartment 56 days after dosing. We found significant suppression of viral replication in RT for both Clade B (p < 0.05) and Clade C (p < 0•0001) viruses at all time points in RT. In contrast, viral suppression was only seen in CT at day 56 after the first dose of LA-RPV and at no time point for VT. Conclusions: MD administration of LA RPV was safe and well tolerated. As with our previous SD data, we saw prolonged suppression of viral replication in RT following exposure to LA RPV. Interestingly, despite modest accumulation of plasma RPV over time, there was minimal evidence of viral suppression in CT or VT.

TUAC0104
Impact of microbiota on female genital tissue and plasma concentrations of dapivirine Interviews were digitally recorded, transcribed and analysed using framework analysis.
Results: The majority of participants reported risk reduction strategies including occasional condom use, strategic positioning or sero-sorting. Participants applied rules to their sexual behaviour, such as using condoms "if it was a one night stand", or not being receptive "outside of a relationship". Typically, PrEP was added to the existing set of "rules". For some participants, PrEP allowed a relaxing of the rules, for example, about strategic positioning: "I have definitely experienced more as a bottom", or about condomless sex: "I have had more unprotected sex than before . . . it doesn't mean that I only have unprotected sex". Other participants insisted PrEP had not changed their rules: "I haven't changed the way I think because I am taking this pill". Participants described PrEP as a "security blanket", an added "defence mechanism" and used analogies such as wearing a "crash helmet . . . on my bicycle". PrEP was described as affording "more intimacy", "reassurance" and giving "added control". By using PrEP, many participants with HIV-positive partners sought to reduce their partner's anxiety about the risk of transmission. The benefits of PrEP were described within the social context of risk environments in cities like London, the chemsex scene and the digitization of sexual contact. PrEP use was viewed as time-limited: "clearly it is a period, a moment . . . it is not going to be a lifetime". Conclusions: These data suggest that PrEP was added to a range of 'rules' already used to mitigate risk, rather than replacing them. PrEP impacted on the boundaries of the rules for some people but not all. In social contexts of high-risk behaviour, PrEP offers added protection and psychosocial benefits that increase individual choice in the mitigation of risk.

TUAC0201
High uptake of community-based HIV testing by adolescent girls and young women aged 15-24: implications and synergies for PrEP rollout?
A Background: HIV incidence among female youth aged 15-24 in South Africa is four times higher than their male counterparts. Recent HIV prevention trials in South Africa documented incidence of 5% to 6% per year in 15-24-year-old adolescent girls and young women (AGYW). HIV counselling and testing is the entry point for treatment and prevention services and is key to implementing effective HIV prevention strategies. Community-based HIV counselling and testing (CBCT) has the potential to increase testing among key populations. We present interim findings of our at-scale CBCT Abstract TUAC0201- Figure 1. Testing uptake by age group and gender. HIV positivity and testing uptakes rates were calculated and stratified by age group, gender and district.
Conclusions: Implementation of our CBCT programme has been extremely successful. Though the largest number of tests was performed on those aged 25-49 years, AGYW had the highest testing uptake. If access to PrEP by AGYW is to be scaled up, innovative supply-side interventions and service delivery platforms must be identified. Stakeholders involved in PrEP implementation should consider the synergies that CBCT programmes may provide for identifying and delivering PrEP services to AGYW.

TUAC0201
Finding the right target population for PrEP: the costeffectiveness of pre-exposure prophylaxis provision to female and male adolescents and young women in South Africa Background: The South African government is in the process of identifying a suitable rollout strategy for providing oral PrEP to young people at risk of acquiring HIV. We were tasked to evaluate the cost-effectiveness of providing PrEP to young women (20-24 years) vs. male or female adolescents (15-19 years), both when covering everyone in these groups or only those self-identifying as being at high risk. Methods: We used Thembisa, an existing HIV transmission model, and cost input from the first PrEP demonstration projects to model the impact of PrEP provision on new HIV infections and cost per HIV infection averted over 20 years, over a baseline of the current HIV programme, including potential cost savings due to reduced treatment need. We compared provision to all young people in a sub-population to targeting via self-selection by high-risk individuals. Target coverage was set to 18% of either population and varied in sensitivity analysis.
Results: PrEP provision to adolescents of both genders is the most cost-effective option, being more effective and less costly than provision to young women (see Table 1). Provision to female adolescents is slightly more effective and cost-effective than provision to male adolescents. At 18% coverage of either all young people in the target population or only those at high risk, the impact on HIV infections averted is similar, but self-selection by high-risk individuals results in much smaller populations on PrEP. As a result, targeted provision is much more cost-effective. PrEP is however not cost saving at any of the coverage rates tested (1-99%) for any of the populations, with or without successful targeting to high-risk sub-populations.
Conclusions: Provision of PrEP to female adolescents is more costeffective than to male adolescents or young women. PrEP, although expensive, can be made more cost-effective if high-risk sub-populations successfully self-select for PrEP.

TUAC0203
State-level school-based sex education policies on sexual orientation are associated with changes in teaching about HIV prevention Oral Abstracts that were discriminatory towards lesbian, gay, bisexual and/or transgender (LGBT) individuals or stated that homosexuality must not be promoted or addressed as a socially acceptable alternative. This was an 83% increase from the number of states in 2014. The implementation of these policies may limit the HIV education available to LGBT students and their heterosexual peers. Methods: We constructed a panel data set for even-numbered years from 2006 to 2014. The independent variable was an indicator from SIECUS State Profiles about whether states' sex education laws were LGBT stigmatizing. The three dependent variables were from the Centers for Disease Control and Prevention (CDC) School Health Profiles system. Based on self-administered questionnaires from the principal and the lead health education teacher in a sample of high schools, the CDC reported for each state the percentage of schools that tried to increase students' knowledge about HIV prevention, human sexuality and sexually transmitted disease (STD) prevention. We conducted ordinary least-squares regression analysis in Stata/SE 14 with state-and year-fixed effects to control for unobserved time-invariant state-level confounders and state-invariant time-varying confounders.
Results: In the regression models, having a state sex education policy that was anti-LGBT was associated with a 17.4%, 21.7% and 16.3% decrease in the percentage of schools that tried to increase student knowledge about HIV prevention, human sexuality and STD prevention, respectively (p < 0.05).
Conclusions: State sex education laws that are negatively biased towards LGBT people may increase reluctance of teachers to try to increase student knowledge about sex education topics in general, including those that are not related to sexual orientation or gender identity. Given the increase in these laws in the past year, we expect that additional teens will be denied instruction on these topics. Further research is necessary to assess how this may affect youth risk behaviours and sexual health outcomes.

TUAC0204
Potential for HIV transmission among adolescents and young adults receiving antiretroviral therapy Background: Adolescents and young adults (AYA) living with HIV have lower rates of virologic suppression and higher rates of sexually transmitted infections (STIs) than older adults, which increases HIV transmission potential. We aimed to identify the proportion of participants with, and risk factors for, high HIV transmission potential within a cohort of HIV-positive AYA.
Methods: Retrospective cohort study of HIV-positive, antiretroviral therapy (ART)-treated, AYA, ages 13-24, at a U.S. adolescent HIV clinic from 2002 to 2015. We included all visits with viral load (VL) measurement after ART initiation. High transmission potential was defined as incident STI (Neisseria gonorrheae, Chlamydia trachomatis or Treponema palladium) with concurrent VL >1500 copies/ ml. Generalized estimating equations (GEE) were used to calculate odds ratios (ORs) and 95% confidence intervals (CI) for hypothesized risk factors for high transmission potential, including age, gender, insurance status, sexual orientation, race and history of STI at entry to care. Results: Participants (n = 251) were followed for a median of 3.2 years (interquartile range (IQR) 1.5-5.3), contributing 2860 visits. Participants were 87% African-American (n = 218), and 73% men and transgender women who have sex with men (n = 182) and 48% (n = 120) had a history of STI at entry to care. The median visit age was 21 years (IQR 19-23). Incident STI was detected in 68% (n = 166) of participants comprising 15% (n = 299) of visits. Participants were viraemic (VL >1500 copies/ ml) at 27% (n = 640) of visits. High transmission potential occurred at least once in 16% (n = 39) of participants and 3% of visits. In the final GEE model, history of STI at or before entry to HIV care conferred a nearly fourfold increased odds of high transmission potential (OR 3.8, 95% CI: 2.0-7.1, p < 0.001).
There was no significant association between age, gender, sexual orientation, race or insurance status and high HIV transmission potential.
Abstract TUAC0203- Table 1. Sex education implementation by LGBT policy.

Oral Abstracts
Conclusions: In this conservative model of transmission potential, 16% of ART-treated AYA in care were episodically at high risk of HIV transmission, demonstrating the limits of treatment as prevention in this setting. A baseline history of STI conferred higher risk of transmitting HIV, emphasizing the need for secondary prevention interventions targeting both ART adherence and sexual risk reduction for HIV-positive youth.

TUAC0205
HIV and HSV-2 risk among young women in age-disparate partnerships: evidence from KwaZulu-Natal, South Africa Background: Young women in sub-Saharan Africa continue to exhibit high HIV prevalence and incidence rates. We explored the role age-disparate partnerships play in HIV infection risk among 15-24year-old women in an endemic setting in South Africa. Methods: During June 2014-June 2015, a cross-sectional household survey was conducted in KwaZulu-Natal Province, South Africa, comprising 9812 individuals aged 15-49 years. Venous blood samples were collected for HIV antibody and viral load tests and herpes simplex virus type 2 (HSV-2) antibody tests. A partnership was defined as age disparate if the age difference between partners was five years or more. Multiple logistic regression analyses were first used to assess the associations between age-disparate partnerships and both HIV and HSV-2 status -HSV-2 may increase a young women's risk of HIV infection -among 15-24-year-old women who reported at least one sexual partner (n = 1557). The second set of analyses used partnership data reported by men -restricted to ongoing partnerships with 15-24year-old women (n = 1078) -to assess whether age-disparate partners of young women were more likely to be HIV positive with a detectable viral load (≥20 copies/ml) and therefore pose a greater level of risk than age-similar partners.
Results: Women who reported any age-disparate partnerships were more likely to test positive for HIV (37% vs. 22%, p < 0.01) and HSV-2 (65% vs. 46%, p < 0.01). After controlling for, inter alia, age and number of lifetime sexual partners, the odds of young women having HIV (adjusted odds ratio ( Oral Abstracts intervention to reduce the risk of HIV infection, but the success of such interventions may be conditional upon changes in gendered power inequalities. Using a randomized experiment in northeast South Africa, we find that a CCT targeted to poor girls in high school reduced the risk of intimate partner violence (IPV) by 34%. The purpose of this study is to understand the pathways through which the CCT affects IPV. Methods: Our study is a phase 3, randomized controlled trial (HPTN 068) in a rural area in Mpumalanga province, South Africa. Eligible young women (aged [13][14][15][16][17][18][19][20] and their parents or guardians were randomly assigned (1:1) to receive a monthly cash transfer conditional on school attendance versus no cash transfer. Participants (N = 2448) were interviewed at baseline and then at annual follow-up visits at 12, 24 and 36 months. We estimate the primary outcome, physical IPV in the past 12 months, using a GEE log-binomial regression model. We examined mediation of direct effects through intermediate pathways using methods designed for nonlinear models under the counterfactual framework. Mediators include sexual behaviours, empowerment and economic well-being measures.
Results: We find evidence that the CCT works through delaying sexual debut or reducing the likelihood of having a sexual partner. The intervention interacts with these mediators leading to a larger reduction in IPV risk. Compared to the direct effect of the CCT on any physical IPV (relative risk (RR) 0.66, confidence interval (CI) (95%):0.59-0.74), the risk of IPV is further reduced when we set the controlled direct effect to either no sexual debut (RR 0.53, CI (95%):0.45-0.63) or to no sexual partners (RR 0.56, CI (95%):0.48-0.63).
Conclusions: Results indicate that a CCT for adolescent school girls has protective effects on girls' experience of violence in part because the intervention reduces the likelihood of debut or having a sexual partner, thereby reducing the opportunity for IPV. Since these behaviours also protect against HIV acquisition, this evidence strengthens the case for CCTs for HIV prevention.

TUAC0402
The effect of school attendance and school dropout on incident HIV and HSV-2 among young women in rural South Africa enrolled in HPTN 068 Background: Education may protect against sexually transmitted infections but has primarily been studied as educational attainment in adults or using measures of prevalent rather than incident infection. Few studies have explored schooling as a measure of time spent in a structured school environment. We hypothesize that low versus high attendance in school and school dropout versus staying in school are associated with a higher risk of incident HIV and HSV-2 infection among young women Methods: We used longitudinal data from the HPTN 068 randomized trial in Agincourt, South Africa, to determine if percentage of school days attended between annual surveys and school dropout affect incident HIV and HSV-2 in young women aged 13-23. We examined inverse probability of exposure-weighted survival curves and used them to calculate one-, two-and three-year risk differences and risk ratios for the effect of school attendance on incident HIV and HSV-2, accounting for confounding. A marginal structural cox model was then used to estimate the hazard ratios for the effect of school attendance and school dropout on incident HIV and HSV-2.
Results: Over the study period, 107 incident HIV cases occurred among the 2328 women without HIV at baseline and 208 HSV-2 incident cases occurred among the 2238 women without prevalent HSV-2 at baseline. Risk of HIV and HSV-2 increased over time and was lower for young women who had high attendance (≥80% school days) versus low attendance (<80%)  Oral Abstracts and GRADE to evaluate the quality of the evidence. Where appropriate, random-effects meta-analysis was conducted.
Results: Of 1742 citations identified, four randomized controlled trials (RCTs) and six observational studies totalling 5150 index patients from eight countries were included. Meta-analysis of three individually randomized trials provided moderate-quality evidence that assisted partner notification services increased HTS uptake among partners 1.5 times compared to passive referral (relative risk [RR] = 1.46; 95% confidence interval (CI): 1.22-1.75; I 2 = 0%). Overall, studies reported that 13-72% of partners were HIV positive, and between 12% and 66% of partners were newly diagnosed. The proportion of HIV-positive partners was 1.5 times higher with assisted partner notification than with passive referral (RR = 1.47; 95% CI: 1.12-1.92; I 2 = 0%; moderate-quality evidence). Few instances of violence or harm were reported. Conclusions: Assisted partner notification improved partner testing and increased diagnosis of people with HIV infection, with few reports of harm. The WHO strongly recommends offering voluntary-assisted HIV partner notification services as part of a comprehensive package of testing to all newly diagnosed persons and periodically to all HIV-positive persons throughout their care and treatment.

TUAC0404
Association between HIV and sexually transmitted infections and partner circumcision among women in uMgungundlovu District, South Africa: a cross-sectional analysis of HIPSS baseline data Background: Randomized controlled trials and observational data have demonstrated that circumcision partially protects men from acquiring HIV and some sexually transmitted infections (STIs) through heterosexual sex. They also suggest that female partners receive some protection, possibly indirectly through lower infection prevalences among men. However, population-level data outside experimental settings are lacking. The HIV Incidence Provincial Surveillance System (HIPSS) is a longitudinal study in Vulindlela and Greater Edendale sub-districts, South Africa, which collected population-level baseline data in 2014 and 2015.
Methods: Female HIPPS participants were aged 15-49 years. Those with at least one past or current male sexual partner who were able to report his circumcision status were analysed. Participants were assessed for HIV status via double fourth-generation ELISA testing with confirmatory Western blot; N. gonorrhoeae, C. trachomatis, T. vaginalis and human papillomavirus infection with standard testing of self-collected vulvovaginal swabs; T. pallidum, HSV-2 and hepatitis B with serology; and STI diagnosis history and current STI symptoms with self-report. They were grouped by circumcision status of their most recent partner, stratified into age below or at least equal to 25 years and compared on presence of STI outcomes by chi-square testing, weighted for selection probability and nonresponse.
Conclusions: Partner circumcision was associated with decreased prevalence of HSV-2 in all female HIPSS participants, decreased prevalence of HIV in younger women and decreased prevalence of syphilis in older women. Its positive association with self-reported STI history in older participants may derive from differential ascertainment; circumcision typically involves STI screening in men, potentially leading to partner notification. Findings support community-level protection against HIV and some other STIs among women from male circumcision.

TUAC0405
Effects of syringe distribution policy change at a syringe services programme in Baltimore, MD: a forecast analysis S Allen 1 ; J Park 2 ; B Weir 1 ; D Holtgrave 1 and S Sherman 1 1 Johns Hopkins University, Health, Behavior and Society, Baltimore, USA. 2 Johns Hopkins University, Epidemiology, Baltimore, USA Presenting author email: sean.travis.allen@gmail.com Background: Syringe services programmes (SSPs) are associated with decreases in prevalence and incidence rates of blood-borne diseases (e.g. HIV and hepatitis C virus [HCV] infections), soft-tissue infections and other morbidities among people who inject drugs (PWID). The core goal of SSPs is to decrease the circulation time of contaminated syringes and to increase the volume of new syringes, effectively increasing the "coverage" of sterile needles and syringes for every injection. In 2014, the Baltimore City SSP shifted from a strict one-to-one syringe exchange policy to a needs-based distribution policy whereby PWID could receive as many syringes as they require. The purpose of this research is to examine the impact of this policy change on syringe distribution among PWID. Methods: Syringe distribution data from April 2012 to November 2016 were abstracted from the Baltimore City SSP and divided into monthly observations. These data were used to build an ARIMA model that forecast the estimated number of syringes that would have been distributed had the syringe distribution policy not changed in the 26-month period following the policy change.
Results: There were significant (p < 0.05) differences in the mean number of syringes distributed per month in the pre-and postpolicy change periods (44,410 and 96,187, respectively Background: To know if HIV acquisition among migrants living in Spain takes place before or after migration to the country. Methods: Cross-sectional study in 18 centres/hospitals following patients with HIV. We selected patients diagnosed with HIV in the last five years, aged 18 and more, born outside Spain and able to complete the survey in one of the 15 available languages. Two questionnaires were used: one self-completed by participants in a computer/tablet and another about patients' clinical data. We collected epidemiological, socio-economic, clinical, behavioural, migratory trajectory, previous HIV serology, CD4 determinations, viral load and viral subtype data. To estimate timing of HIV acquisition, Bayesian methods with longitudinal measurements of CD4 and viral load were used, and estimates were made applying mixed linear models based on HIV natural history (data from the European collaboration of seroconverters CASCADE). Results: Of 710 participants, 685 (97%) had data to estimate time of HIV acquisition (TA). Participants with information on TA had a median of nine years [RI: [6][7][8][9][10][11][12][13] of residence in Spain, were mainly male (77%) and were from Latin America and the Caribbean (LAC) (64%), Other European Countries (OEC) (17%) and sub-Saharan Africa (SSA) (13%). Most common route of transmission was the relationship between men who have sex with men (MSM) (60%), followed by heterosexual transmission (34%); injecting drug users (IDUs) were minority (3%) and in 3% were unknown. Seventy-six per cent of patients acquired HIV after migrating to Spain. HIV post-migration acquisition was higher among MSM (85%) than in heterosexuals (67%) and in people from LAC (77%) and OEC (75%) compared to those from SSA (62%). HIV was acquired after migrating in greater proportions among those aged more than 50 years (82%), with secondary or university studies (77%) and with legal residency status (79%). Conclusions: A significant proportion of migrants living with HIV in Spain acquired HIV after migration. The proportion of post-migration HIV acquisition is higher among MSM and in people from LAC and OEC. These results show a failure in preventive interventions in specific migrants' groups and emphasize the need to design interventions taking into account migrants' heterogeneity. Background: Debate continues with respect to whether transactional sex among men who have sex with men (MSM) constitutes increased risk of HIV transmission. We sought to identify factors associated with transactional sex using event-level data from a population-based longitudinal behavioural cohort of MSM in Metro Vancouver, Canada. Methods: Sexually active MSM aged ≥16 years were recruited using respondent-driven sampling (RDS), and from February 2012 to February 2016, participants completed study visits every six months, which included a computer-assisted self-interview. At each visit, participants provided event-level data by reporting on their last sexual encounter with their five most recent partners (e.g. participant and partner substance use, partner's relative age). We used a four-level mixed-effects model (RDS recruitment chain; participant; visit; and event) to evaluate temporal trends and factors associated with transactional sex. We built a multivariable model to compare events where money, drugs or goods was received for sex to events with no transaction reported using backward selection with minimization of QIC and type III p-values. Results: Of 690 participants, 8990 sexual events were reported across 2792 study visits (median follow-up time of 3.62 years). Although 11.7% of participants reported any transactional sex, event-level reports were rare: 2.4% of events included receiving money/drugs/goods, 1.5% of events included providing money/ drugs/goods and 96.0% of events did not include transactional sex. Transactional sex decreased over time (3.8% to 1.0% from first to eighth visit: odds ratio [OR] = 0.82, 95% confidence interval [CI]:0.73-0.92). After controlling for significant individual-level factors, the following event-level factors were associated with greater odds of transactional sex: having met online (adjusted OR [aOR] = 3.32, 95% CI:1.32-8.37), having a shorter relationship (aOR = 0.99 per month, 95% CI:0.99-1.00), having an older partner (aOR = 2.42, 95% CI:1.02-5.75) and having a partner who used crystal methamphetamine (aOR = 3.85, 95% CI:1.12-13.17) or used gamma-hydroxybutyrate (aOR = 6.79, 95% CI:2.19-21.01). Across all events, sexual HIV risk was not associated with transactional versus non-transactional sex: 23% versus 24% reported condomless anal sex with a sero-concordant partner (p = 0.212) and 17% versus 12% reported event-level condomless anal sex with a sero-discordant/unknown status partner (p = 0.946). Conclusions: Transactional sex events were rarely reported. Partner substance use was strongly associated with transactional sex, but we found no significant associations with HIV risk behaviour.

TUAC0503
Is on-demand HIV pre-exposure prophylaxis (PrEP) a suitable tool for men who have sex with men (MSM) who participate in chemsex? Results from a sub-study of the ANRS-IPERGAY trial  Background: Chemsex -the use of psychoactive substances during sexual encounters -is a growing concern among men who have sex with men (MSM). It includes the even riskier practice of "slamming", which consists in injecting a substance before having sex. On-demand HIV pre-exposure prophylaxis (PrEP) may be a suitable tool to prevent HIV transmission in "chemsexers" (i.e. those practicing chemsex). We used a sub-study of the ANRS-IPERGAY trial to describe chemsexers and their PrEP use. Methods: Among the 361 MSM (3051 visits) enrolled in ANRS-IPERGAY's open-label extension study, we selected the 331 (1657 visits) who reported drug use during at least one sexual encounter. A two-monthly web questionnaire over 12 months collected sociobehavioural data including the use of PrEP and practicing chemsex. We compared sexual behaviours between visits where chemsex was reported and those where it was not. A GEE logistic regression was used to study whether practicing chemsex was associated with PrEP use. Results: Among the 331 participants, during follow-up, 29% reported chemsex while 8% reported slamming at least once. Chemsex was reported in 16% of all visits (12% and 4%, respectively, with one and multiple partners) mainly involving the use of GHB/GBL (51%) and synthetic cathinones (46%). Chemsexers were not significantly different from non-chemsexers regarding sociodemographic characteristics, although they reported higher use of anxiolytics during the previous 12 months (p < 10 -2 ). When MSM reported chemsex during their most recent sexual encounter, it was associated with a greater likelihood of receptive anal sex (p < 10 -3 ), hardcore sexual practices (p < 10 -3 ), casual partner(s) (p < 10 -3 ) and a higher perception of risk (p < 10 -3 ). Those who reported chemsex at their most recent encounter were more likely to use PrEP than those who did not (p < 10 -3 ) and less likely to use condoms (p < 10 -3 ). After adjustment for other potential correlates, chemsex remained associated with PrEP use (odds ratio [95% confidence interval] = 2.18 [1.04; 4.59]). Conclusions: Our findings show that chemsexers are more likely to report high-risk sexual practices but also have a higher perception of risk. Consequently, they are also more likely to use PrEP when practicing chemsex; PrEP may be a suitable tool to reduce HIV risk transmission among chemsexers.

TUAC0504
Assessing HIV prevalence and health outcomes of children of female sex workers in Port Elizabeth, South Africa, to guide PMTCT programming for vulnerable populations Methods: FSW in Port Elizabeth, South Africa, were recruited at mobile clinics and within the community to bring their children ≤12 years to the study site. A cross-sectional interview with the mother, followed by health assessments for the mother and her children, was completed. HIV testing was completed for mothers and children; children were tested using rapid antibody tests (≥18 months) or polymerase chain reaction (<18 months). HIV outcomes and health status of mothers and children are characterized. Stunting and wasting were estimated using WHO 2006 Child Growth Standards. Results: From July 2015 to February 2016, 114 mothers and 200 children were enrolled. Overall, 77/114 (68%) mothers were living with HIV, of which 53% (41/77) were on antiretroviral therapy. On average, FSW continued sex work for a median of five months during their last pregnancy (interquartile range (IQR) 4-7). The median age of children attending was 6 years (IQR 3-9). The majority (73%, n = 145/200) of children were breastfed; the median and mean duration of breastfeeding were 6 (IQR 3-24) and 12 months (SD 12), respectively. Just over half (108/200) of children had ever been tested for HIV, including 95/133 (71%) of children with HIV-positive mothers. HIV prevalence among children was 3% (5% among those with HIV-positive mothers, n = 6/133). Seventy-three per cent of children were reported as up-to-date on their vaccinations. Among the 79 children under 5, 29% were stunted according to height-for-age and 13% wasted per weightfor-age. Conclusions: The majority of FSW mothers were HIV positive, and many were not on treatment. HIV prevalence among children was higher than the national average among children, and nearly onethird of children with HIV-positive mothers had never received HIV testing. Aside from children's HIV risks, substantial chronic nutritional deficiencies were identified through height-for-age, alongside acute nutritional deficiencies (weight-for-age) and immunization gaps. Programmes for FSW should address vertical transmission risks including treatment support during pregnancy and breastfeeding and consider catch-up HIV testing and vaccination campaigns to promote children's health.

TUAC0505
Low HIV incidence but high HCV incidence among people who inject drugs in Haiphong, Vietnam: results of the ANRS 12299/NIDA P30DA011041 DRIVE-IN study on the rate of HIV new infections, but also on HCV transmission, remains unknown. Methods: We carried out a community-based respondent-driven sampling (RDS) survey among 'active PWID' (i.e. with positive urine test for heroin and presence of injection marks) in Haiphong, with HIV and HCV testing. Then, HIV-negative participants and HCVnegative participants not on methadone maintenance therapy (MMT) were eligible for one-year follow-up. HIV/HCV was tested at six months and one year along with routine harm reduction activities from community-based organizations (CBO) and support to access MMT. We estimated HIV and HCV incidence and risk factors associated with HCV seroconversion. Results: Among 603 RDS participants, 90% were males, and their median age was 36.5 years. HIV prevalence was 25% (95% confidence interval (CI): 22-29) and HCV prevalence was 66% (95% CI: 63-70). A total of 204 RDS participants were enrolled in the cohort, including 94 HIV negative/HCV negative, 5 HIV positive /HCV negative, 105 HIV negative /HCV positive. The cohort participants were mainly males (90%)  Conclusions: In Haiphong, harm reduction activities and high levels of antiretroviral treatment likely contributed to the reduction of HIV epidemic among PWID. However, HCV incidence is still unacceptably high. Current harm reduction activities cannot control the HCV epidemic; additional strategies such as universal ART and large-scale HCV treatment should be urgently evaluated to end the HIV epidemic and tackle HCV transmission among PWID in Vietnam.

TUAD0101
Wide-ranging real-world impacts of a policy change on treatment eligibility on ART initiation and retention in care in Zambia A Mody 1 ; A Zanolini 2 ; K Sikombe 2 ; P Somwe 2 ; I Sikazwe 2 ; C Bolton 2 ; C Holmes 2,3 ; N Padian 4 and E Geng 1  Background: Incomplete uptake as well as unforeseen consequences often accompanies changes in public health guidance. In April 2014, Zambia increased the CD4 threshold for HIV treatment from 350 to 500 cells/μl. We evaluated the effect of this guideline change on ART uptake and retention in patients both targeted and not targeted by this change to evaluate for unintended consequences using a regression discontinuity design. Methods: We analysed non-pregnant ART-naive patients in Zambia who newly enrolled within six months of the guideline change (1 September 2013 to 1 November 2014), excluding patients enrolled within 30 days of the implementation date to account for imprecise rollout. We utilized a quasi-experimental regression discontinuity design with local linear regression to estimate the effects of this policy change on ART initiation within three months of enrolment and retention in care at six months (defined as clinic attendance between three and nine months after enrolment) in all new enrolees, stratifying by enrolment treatment eligibility. Results: A total of 20,513 patients (53.1% female, median age 34 years (interquartile range (IQR) 28-41)) were eligible for our analysis. Newly eligible patients (CD4 350-500, 15.5% of patients) saw a significant increase in ART initiation within three months (risk difference [RD] +35.3%, 95% CI 28. 1-42.4, p < 0.001) and retention at six months (RD +7.4%, 95% CI 0.5-14.3, p = 0.034) with the policy change. Additionally, never-eligible patients (CD4 >500, 17.0% of patients) also saw an increase in ART initiation with the guideline change (RD +16.3%, 95% CI 9.8-22.7, p < 0.001), though retention was unaffected (p = 0.203). Among always-eligible patients (CD4 ≤350 or WHO stage≥3, 67.6% of patients), ART initiation at three months (p = 0.955) and retention at six months (p = 0.600) were unaffected. Conclusions: Policy increasing CD4 threshold for treatment eligibility led to rapid changes in ART initiation practices as well as enhanced retention in the group targeted by the guidelines. ART initiation also improved amongst treatment-ineligible patients with the policy change, a positive spillover effect perhaps due to expansion of ART supply, while initiation and retention among those always eligible was not compromised. Real-world implementation of evidence-based practice often has broader impacts than those directly targeted that should be routinely evaluated to guide policy interventions.

TUAD0102
Pre-ART peak and plateau: early lessons from Zimbabwe on operational impact of "pre-ART mop-up" on ART initiation rates under Treat All Background: The WHO 2015 test and treat guidelines recommend ART initiation for all people living with HIV regardless of CD4 count or WHO clinical staging. Implementation of Treat All will require identification and return of pre-ART patients previously clinically ineligible for ART into care for timely initiation. Little is known about the feasibility and timelines required to conduct pre-ART mop-up in resource-limited settings. Our objective was to establish the proportion of clients initiated on ART from pre-ART mop-up following start of Treat All in 92 public health facilities in seven districts of Zimbabwe. Methods: We purposively selected 92 health facilities implementing Treat All learning phase. We analysed routinely reported data from April-December 2016, comparing proportion of new ART initiations to patients newly diagnosed as HIV positive using chisquare testing. A facility-based survey of pre-ART register data and healthcare worker (HCW) perceptions and experiences of pre-ART mop-up during initial stages of Treat All was conducted to identify key operational themes. Results: Over the period of interest, 9875 clients newly initiated on ART. The proportion of new ART initiations vs. newly diagnosed peaked at 160% after start of Treat All and plateaued back to below 100% of new diagnoses after six months, rates significantly higher than before Treat All (69.4% vs. 91.9%, p < 0.0001).
Additional cell phones, air time and staff for identifying and contacting pre-ART patients were crucial for tracing large numbers of pre-ART clients at start of Treat All. Conclusions: We demonstrate success of Treat All at returning previously ineligible HIV-positive clients for ART initiation in a resource-limited, high-burden setting. Provision of additional resources to support pre-ART mop-up phase was required and recommended for replication as Treat All scales-up. Following return to care of traceable pre-ART clients, ART initiations stabilize to rates above 90%, indicating the value of Treat All for progress towards the 2nd 90.

TUAD0103
Evaluating the feasibility of implementing UNAIDS' 90-90-90 strategy, achieving universal access to treatment and eliminating HIV in Malawi S Blower and L Palk UCLA Medical School, Semel Institute, Santa Monica, USA Presenting author email: sblower@mednet.ucla.edu Background: Malawi has a severe HIV epidemic: prevalence in the general population is approximately 11%. The majority of the population of approximately 16 million live in rural communities. We estimate the total number of HIV-infected individuals (both diagnosed and undiagnosed) in Malawi and determine their geographic location. We use these results to evaluate the feasibility of implementing UNAIDS'90-90-90 strategy and of achieving universal access to treatment. Methods: We constructed an epidemic surface prevalence (ESP) map using geo-referenced HIV-testing data from approximately 14,000 individuals (15-49 years old) who participated in a nationally representative population-level survey: the 2010 Malawi Demographic and Health Survey. We constructed a density of infection (DoI) map by combining the ESP map with gridded demographic data from the WorldPop database and a census-based agestructure map. The DoI map shows the estimated number of HIVinfected individuals (15-49 years old) in each square kilometre in Malawi, diagnosed and undiagnosed individuals. We calculated the total number of HIV-infected individuals by aggregating the mapped estimates.
Results: The ESP map shows that prevalence (in 15-49 year olds) varies from approximately 1% to 25%, and there is a strong northsouth trend in increasing prevalence and a substantial urban-rural difference. Prevalence is highest in cities in the south (Blantyre and Zomba), Lilongwe in the central region, Mzuzu in the north and villages along Lake Malawi. The DoI map reveals the geographic dispersion pattern of all infected individuals. DoI ranges from one infected individual/km 2 in rural areas to more than 1000 infected individuals/km 2 in the four major urban centres. The map shows substantial regional differences in the DoI; our results show that these reflect differences in settlement patterns, population density and prevalence. We estimate that there were approximately 692,000 HIV-infected individuals (15-49 years old) in 2010 in Malawi, and only approximately 25% were living in urban areas. Conclusions: Our results show the vast majority of HIV-infected individuals in Malawi live in rural areas in small communities where the DoI is low. This indicates that implementing the 90-90-90 strategy and achieving universal access to treatment may not be feasible. We recommend that HIV elimination strategies for resource-constrained countries are designed based on DoI and not on prevalence.

TUAD0104
Generic treatments for HIV, HBV, HCV and TB could be mass produced for <$90 per patient Background: High prices to treat HIV, viral hepatitis and TB can limit treatment access. This analysis aimed to determine prices currently feasible for HIV, HBV, HCV and first-line (1L) DS-TB treatment, assuming competitive generic manufacture. Methods: Data on API exported from India were collected from an online database (www.infodriveindia.com) for July 2014-July 2016. Linear regression was used to plot API cost/kg versus export date, weighted by export volumes: the generated model was used to calculate current average cost/kg of API. Target prices were calculated based on the per-pill cost of API, plus costs of manufacture ($0.01/pill), 10% profit margin and assumed 27% tax on profit.

Oral Abstracts
Results: Table 1 shows current prices of antiretrovirals for HIV, entecavir (ETV) for HBV (per person-year), HCV treatments (per 12week course) and 1L DS-TB treatment (RHZE, per six-month course). API costs/kg were $1189 for ATV, $182 for TDF, $241 for 3TC, $109 for EFV, $380,965 for ETV, $1224 for SOF, $4448 for LDV and $852 for DCV. EFV, 3TC, ETV and RHZE are already generic in the USA. The US substance patents on atazanavir expire in 2017, TDF in 2018, sofosbuvir in 2030 and daclatasvir in 2031. Sofosbuvir + ledipasvir combination patents expire in 2032. Conclusions: Treatment of HIV, HBV, HCV and TB could be achieved for <$90 per person globally if robust generic competition is enabled. In most countries, generic TDF/3TC/EFV, TDF/3TC, ETV or 1L DS-TB treatment could be available for <$90 by early 2018 after patent expiry. Most HCV direct-acting antivirals (DAAs) will remain on patent for ≥12 more years. Voluntary licensing or other mechanisms will be required to enable access to HCV DAAs at low prices.

TUAD0105
Comparative analysis of ARV costs before and after the Clinical Protocol and Therapeutic Guidelines for the management of adult HIV infection (PCDT) was adopted in 2013 in Brazil Background: The PCDT of 2013 established standardization of treatment, practice of early treatment and treatment for all. These initiatives have diverse and opposing consequences on the costs of antiretrovirals (ARVs). Objectives were to know and ponder these consequences by comparing costs between the antiretroviral treatment structure used until 2013 and the PCDT changes applied since 2014. The problem evaluated was if the number of patients and the resulting increase in total costs would be counterbalanced by a significant reduction in average costs as consequence of recommendations for standardization and simplification of treatment, considering the 2020 timeline.

Methods:
The study covered 2009-2013 and 2014-2015 scenarios before and after the adoption of the PCDT in Brazil by the Institute of Economics of the Federal University of Rio de Janeiro with the Ministry of Health. The study design estimated the total and average costs for the total adult patients treated in December 2009 and 2015, comparing them to estimate the observed cost difference between the two strategies adopted, followed by a simulation of results for 2020. Scenarios of the evolution of costs of HAART took into account treatment targets of the analysed population and changes of schemes. The study population considered were all adult patients under treatment for both periods, 207.014 and 444.093 respectively. Data collection used Systems for Control of Laboratory Tests (CD4 and viral load), for ARV Dispensation, and price records of ARV. Results: Findings indicate decrease in the average costs due to a higher concentration of patients in first-line regimens in 2015. The difference is even more important in the incoming population. The significance is that the total cost of patients in 2015, with an average cost of 41.46% lower, allowed savings of more than 250 million US$. With the average cost of 2009, the total annual cost of patients in 2015 would be 70.83% higher if the PCDT had not been adopted. This is enough to cover 675,000 new patients and more than compensate for the adoption of treatment as prevention. Conclusions: Results become even more relevant since cost savings were accompanied by a better clinical outcome of patients with simplification of treatments.

TUAD0201
Association between user fees and dropout from methadone maintenance therapy: results of a cohort study in Vietnam Oral Abstracts maintain MMT attendance in other provinces has not been assessed. The primary objective of this study is to assess the association between user fees and client dropout from MMT, exploiting the fact that some provinces had implemented user fees in 2015 while other had not. The secondary objective was to estimate the catastrophic payments associated with MMT.
Methods: An observational cohort study of 1021 MMT clients in seven provinces from May/June 2015 to May/June 2016. Three provinces implemented user fees, while four provinces did not. Provinces and facilities were randomly selected, while all MMT clinics in the selected provinces were included. Box Cox proportional hazard models were used to assess the association between user fees and (i) dropout and (ii) being inactive.
Results: About 85% of the cohort was actively on MMT at the end of the observation period, including 10% of clients had transferred facilities and 1% of clients quit with facility staff permission. About 14% of the cohort had stopped MMT care, about 8% dropped out, 3.5% were incarcerated, 1.5% died and 2% stopping for other reasons. The hazard ratio for paying user fees compared to not paying user fees ranged from 0.70 (unadjusted, p = 0.26) to 0.29 (adjusted, p = 0.33) for dropout and from 0.75 (unadjusted, p = 0.24) to 0.50 (adjusted, p = 0.48) for being inactive. However, 29% of clients at facilities paid more than 40% of non-subsistence expenditures for MMT-associated user fees and transportation.
Conclusions: Over the course of one year, we do not find evidence that user fees increased dropout or reduced retention in MMT.
Overall, dropout rates in Vietnam are very low compared to other countries. However, catastrophic payment rates remain a concern, and longer-term follow-up is needed.

TUAD0202
A combination intervention strategy to enhance outcomes across the HIV care continuum and support epidemic control: data from a cluster-randomized trial in Mozambique Conclusions: This RCT provides evidence that in addition to increasing HTC access, policymakers, implementers and external donors should consider providing in-kind incentives as they are cost-effective at increasing CHTC uptake and identifying HIV-positive persons.

TUAD0204
The effects of short-term cash and food incentives on food insecurity and labour force participation among HIVinfected adults initiating antiretroviral therapy in rural Tanzania Background: We previously demonstrated that short-term cash and food incentives increased antiretroviral therapy (ART) possession and retention in HIV services in Tanzania. To elucidate potential pathways that led to these achievements, we examined whether these incentives also improved food security and labour force participation. Methods: At three clinics, 805 food-insecure adults who recently initiated ART (≤90 days prior) were randomized to receive cash or food transfers (approximately $11/month for ≤6 months, conditional on visit attendance) or standard-of-care (SOC) ART. After six months, we re-assessed food security (Household Hunger Scale), body mass index (BMI), employment and functional limitation (illness prevented work or housework) among patients returning for ART. The incentives' effectiveness at increasing retention contributed to differential loss-to-follow-up (13% overall; 20% SOC; 8% cash; 15% food; p < 0.01), which we accounted for in the current analysis using inverse probability weighting. Results: After six months, food security improved from 0% to 37% (p < 0.01) and did not differ comparing the SOC (33%) to cash (36%, p = 0.64) or food (39%, p = 0.37) groups. Oral Abstracts understand the mechanisms through which incentives may increase and sustain retention in HIV services.

TUAD0205
Using conjoint analysis to model hospital directors' decision-making in adoption of an evidence-based stigmareduction intervention C Lin 1 ; L Li 2 ; S-J Lee 2 and Z Wu 3  The authors used a real-life intervention with efficacious outcome to reduce HIV-related stigma in healthcare settings as a "product" to study adoption of EBI. Conjoint analysis was conducted among 60 hospital directors recruited from 30 hospitals of different levels and types in Fujian Province, China. The directors evaluated their willingness to adopt the evidence-based stigma reduction intervention in their hospitals by rating across eight hypothetical scenarios with preferred and non-preferred levels of seven attributes, including (1) administrative support, (2) cost, (3) personnel involvement, (4) format, (5) duration, (6) technical support and (7) priority alignment with the hospital. A mixed-effect model was fit to the likelihood of intervention adoption for the eight scenarios, and the seven attributes (categorized as preferred = 1 or not preferred = 0) served as independent variables in the model. Results: Monetary cost of intervention implementation (impact score = 24.8) had the greatest impact on the directors' willingness to adopt a certain EBI, followed by duration of the intervention (impact score = 10.0), availability of technical support (impact score = 7.5) and flexibility of format (impact score = 4.6). The majority (88.3%) of the hospital directors perceived the conjoint administration process as clear and easy to understand. The data collection time was relatively short, which was approximately 30 minutes. Background: It is recently revealed that an HIV-1 accessory protein Nef plays an essential role in virion infectivity by antagonizing a host restriction molecule SERINC5. However, it remains elusive whether Nef's ability to counteract SERINC5 influences viral fitness in vivo. Because Nef is a highly polymorphic protein due to the selective forces by host cellular immunity, we hypothesized that certain immune-escape polymorphisms might affect the Nef function and thereby plasma viraemia. Methods: We collected plasma viral RNA from HLA-typed, treatment-naive, chronically HIV-1-infected subjects (n = 375) and analysed DNA sequences of Nef-encoding region. Immune-associated Nef polymorphisms were analysed by a phylogenetic network model. We also introduced several mutations to a control strain and patient-derived Nef clones and tested various Nef functions in vitro, including downregulation of CD4 and HLA class I as well as enhancement of virion infectivity and counteraction of SERINC5. Results: We identified 112 Nef polymorphisms that were overrepresented within patients sharing the same HLA genotypes. Specifically, two mutations, Tyr-120 to Phe and Gln-125 to His, were overrepresented in patients carrying HLA-B*51:01 and HLA-C*14:03, and the number of the two mutations correlated inversely with plasma viral load (p = 0.004). Nef functional assays demonstrated that the double-mutant Nef impaired in SERINC5 counteraction and enhancement virion infectivity whereas other Nef functions such as CD4 and HLA class I downregulation remained unchanged. Jurkat cells lacking SERINC5 expression lost such functional difference between the parental and mutant Nef clones.
Conclusions: Taken together, these results suggest that naturally occurring immune-associated mutations impair Nef's ability to counteract SERINC5 and enhance virion infectivity, associating with reduced plasma viral load in vivo.

WEAA0102
HIV-1 concentrates and shelters cell-associated infectivity a "viral biofilm" Background: HIV-1 replication is restricted in resting CD4 T lymphocytes. Stimulation of these cells via either the T-cell receptor (TCR) or gamma-cytokine receptors up-regulates HIV provirus formation at levels of reverse transcription (RT) and nuclear import. However, the enhancements of both RT and nuclear import are not fully explained by activation-induced changes in known restriction factors. Here, we study a "master regulator" activated by both TCR and gamma-cytokine signalling: the mechanistic target of rapamycin (mTOR) kinase. Methods: Resting CD4 T cells purified from blood donor PBMCs using immunomagnetic cell separation were stimulated with anti-CD3/anti-CD28 beads, PHA/IL2 or IL7/15. Inhibitors were used before activation to study the role of mTOR. qPCR quantified HIV-1 RT products and 2-long terminal repeat (2-LTR) circles. Flow cytometry after infection with a single-round HIV-1-GFP reporter virus monitored productive infection. Quantitation of dNTPs used ultra-sensitive LC-MS/MS detection. Flow cytometry and immunoblotting assessed effects of treatments on mTOR activity.
Results: mTOR activity induced by engagement of either T cell or gamma-cytokine receptors coordinates expression of transporters for glucose (GLUT1), glutamine (ASCT2) and transferrin (CD71), as well as rate-limiting enzymes for pyrimidine (CAD), purine (IMPDH2) and deoxyribonucleotide (dNTP) synthesis (RRM1). mTOR has been previously reported to govern the expression of nutrient transporters and pyrimidine biosynthetic genes, but this is the first demonstration of this global mTORdependent programme in activated CD4 T lymphocytes.
Pharmacological ablation of mTOR activity suppressed dNTP pool expansion after activation. Multiple chemically distinct catalytic inhibitors of mTOR were found to reduce HIV-1 RT products after TCR stimulation. Moreover, both TCR and gammacytokine-activation induced mTOR inhibitor-sensitive accumulation of 2-LTR circular forms of HIV-1 DNA, indicating that mTOR activity also regulates active, energy (GTP/ATP)-dependent HIV-1 nuclear import.
Conclusions: CD4 T lymphocyte activation-induced mTOR "metabolic reprogramming" drives increased susceptibility to HIV-1 by expanding key nucleotide substrate and energy pools necessary for both reverse transcription and nuclear import. This adds mechanistic understanding, confirms earlier reports that catalytic inhibitors of mTOR hold promise for improving HIV-1 chemotherapy and prevention and suggests continued investigation of mTOR's role in establishment as well as reactivation of HIV-1 infection.

WEAA0104
Membrane-associated RING-CH (MARCH) 1 and 2 are other members of MARCH proteins that inhibit HIV-1 infection Methods: We intravenously infected six animals with SHIV-1157ipd3N4 and left them untreated for six months, followed by six months of cART (raltegravir, PMPA and FTC). Three animals remained untransplanted, and three animals received myeloablative total body irradiation (1020 Gy) followed by haplo-identical BMT without ART discontinuation. Donor chimerism was monitored, and viral DNA and RNA were measured by qPCR in approximately 35 tissues following necropsy. Results: All animals showed peak viral titres in the range of approximately 10 7 copies/ml approximately two-week post-infection, which subsequently reached steady state. One animal controlled viraemia before ART initiation (day 184). In the other animals, plasma viral RNA became undetectable two to three weeks post-cART initiation. The three transplanted animals were euthanized at day 47, 29 and 9 post-transplant, due to infection, acute graft-versus-host disease and renal complications, respectively. Analysis of whole blood and sorted CD4+ T cells showed rapid acquisition of 100% donor blood chimerism, with lower chimerism in multiple tissues. Despite undetectable viraemia post-transplant, viral DNA quantification in 35 tissues, including haematopoietic and major organs, CNS and reproductive organs, revealed an increased reservoir in transplanted non-controllers versus untransplanted controls.

Conclusions:
Our results indicate that the DNA reservoir significantly increases early after transplant, despite control of peripheral viraemia. This suggests that transplant represents a significant initial "shock" with loss of anti-HIV immunity contributing to reservoir enlargement, potentially explaining the rebound observed in the Boston patients. We believe that full recovery of anti-HIV immune control may be restored if additional HIV resistance factors and/or anti-HIV strategies are incorporated post-transplant to enhance the targeted killing of infected cells.

WEAA0202
Investigating clinical therapeutics to target infected cells and promote HIV clearance Oral Abstracts with the pre-clinical BH3-mimetic ABT-737, which targets pro-survival proteins Bcl-2, Bcl-xL and Bcl-w, resulted in preferential killing of HIV-infected cells in a concentration-dependent manner. A similar outcome was observed using the FDA-approved, clinical-stage Bcl-2 inhibitor Venetoclax. Our humanized mouse model successfully recapitulates key stages of human disease, including chronic infection, viral suppression, as well as subsequent rebound following treatment interruption. Suppressed mice were administered Venetoclax over three treatment cycles before therapy was interrupted to assess for viral rebound. Humanized mice treated with Venetoclax experienced a substantial delay in rebound up to one week longer than vehicle-treated controls.
Conclusions: These results pave the way for further in vivo studies using a humanized mouse model of HIV latency. Efforts are underway to optimize treatment regimens as well as to investigate the combination of BH3-mimetics with traditional latency-reversing agents. Overall, these findings represent a novel approach to killing latently infected cells and purging the proviral reservoir. Background: Long-lived memory CD4 + T-cells with high selfrenewal capacity such as central memory (T CM ) T-cells and T memory stem cells (T SCM ) are major contributors to the viral reservoir in HIV-infected individuals on antiretroviral therapy (ART). The Wnt/beta-catenin signalling pathway primarily regulates the balance between self-renewal and differentiation of T SCM /T CM , and pharmacological manipulation of this pathway offers an opportunity to reduce persistence of latently infected cells.

WEAA0203
Methods: To block self-renewal and promote T SCM /T CM differentiation, we used an inhibitor targeting the Wnt/beta-catenin pathway initially developed to target cancer stem cells (called PRI-724). We evaluated the safety of PRI-724 with a dose-escalation study in healthy rhesus macaques (RM). We then conducted an efficacy study in 12 SIV mac251 -infected RM treated with ART. After suppression of viraemia, 8/12 RM received PRI-724 at 10 or 20mg/kg/day, while 4/12 RM were maintained on ART only. The gene profile of CD4 + memory T-cell subsets was assessed longitudinally by RNAseq analyses. RM were sacrificed on ART following 12 weeks of PRI-724 treatment for assessment of SIV reservoirs. Results: PRI-724 was well tolerated in both healthy and ART-treated SIV-infected RM, with blood counts, liver function and renal function within normal limits. No alteration of tri-lineage haematopoiesis was observed in bone marrow. The transcriptomic profiling of genes associated with T-cell differentiation showed an increased expression of the EOMES transcription factor in the CD4 + T SCM of PRI-treated animals compared to controls. Moreover, Gene Set Enrichment Analyses showed that gene sets distinguishing effector versus memory were significantly enriched in the CD4 + T SCM compared to the naive T-cells. Interestingly, this CD4 + T SCM enrichment in effector-like genes was significantly higher in the PRI-treated than in the control RM. These results suggest an effect of PRI-724 on CD4 + T SCM differentiation. PCR analyses in the sorted subsets of memory CD4 + T-cells showed no significant difference in cell-associated SIV DNA levels between PRI-724-treated and PRI-724-untreated RM.
Conclusions: This work provides novel support for the strategy of promoting CD4 + T SCM differentiation by pharmacological modulation of the Wnt/beta-catenin pathway. However, combination strategies targeting stem cell pathways and/or additional anti-reservoir strategies are likely necessary to reduce SIV/HIV persistence. Background: HIV persists in latently infected CD4 + T cells in infected individuals even after prolonged periods on suppressive antiretroviral therapy (ART). Transcriptional reactivation from latency is not inhibited by current ART, highlighting the need for novel approaches. The HIV-1 Tat protein exponentially activates viral transcription, and limited Tat-transactivation correlates with HIV-1 latency establishment. We identified didehydro-cortistatin A (dCA) as a potent Tat inhibitor. Over time, treatment of infected cells with dCA drives HIV-1 gene expression into an induced state of persistent latency, refractory to viral reactivation by a standard panel of latency-reversing agents (LRAs), suggesting that the HIV promoter is epigenetically repressed. We postulated a strategy for a functional cure, dubbed "block-and-lock", where a specific HIV-1 transcriptional inhibitor would promote a durable state of latency, reduce ongoing viral replication during ART and inhibit spontaneous reactivation during ART or when ART is discontinued.

WEAA0204
Methods: We will present our long-term studies of the activity of dCA in primary human CD4 + T cells isolated from aviraemic infected individuals, using an optimized approach that maintains cell cultures up to 10 weeks. To explore the in vivo efficacy of dCA, we used the bone marrow-liver-thymus (BLT) humanized mouse model of HIV-1 latency. We will also discuss epigenetic mechanisms regulating dCA-induced deep latency in CD4 + T memory T cells, as well as mechanisms of viral evasion to dCA. Results: We demonstrate in CD4+ T cells from five suppressed infected individuals that not only combining dCA with ART promotes faster HIV-1 suppression, but more importantly that prior treatment with dCA significantly reduces viral rebound up to 25 days of treatment interruption, even when viral reactivation is stimulated by LRAs. In the BLT mouse model of latency, we demonstrate that adding dCA to ART-suppressed mice, for a period of 14 days, results in a significant 1.5-10.5-fold reduction in viral RNA production in all tissues. Experiments assessing the ability of dCA to inhibit viral rebound upon treatment interruption are ongoing and will be presented.

Conclusions:
In combination with ART, dCA abrogates residual HIV production from cellular reservoirs, blocks viral reactivation and may ultimately reduce reservoir replenishment and latent reservoir size.

WEAA0206
Myeloablative conditioning is dispensable for transplantdependent HIV cure HIV cure was likely a combined result of the conditioning regimen, an allogeneic "graft versus reservoir" effect, and the CCR5Δ32 donor cells. We studied the impact of conditioning and CCR5edited cells in simian-human immunodeficiency virus (SHIV)infected macaques suppressed by combination antiretroviral therapy (cART). The goal of this study was to identify the mechanisms by which each facet impacted the latent viral reservoir. Methods: Pigtailed macaques were challenged with CCR5-tropic, HIV-enveloped SHIV and suppressed by cART following viral set point. In one cohort, autologous HSCT was performed with unmodified stem cells. In a second cohort, animals received CCR5-edited stem cells. Control animals were infected and suppressed, but did not undergo transplantation. Flow cytometry and ELISA were used to monitor changes in immune homeostasis, and quantitative PCR and viral reservoir assays were used to identify virologic changes between experimental and control animals.
Results: The conditioning regimen resulted in a dramatic, but incomplete depletion of CD4 + , CD8 + T-cells and B-cells, disrupted T-cell homeostasis, elevated markers of microbial translocation, a significant loss of SHIV-specific antibodies and increased viral rebound, relative to untransplanted controls. Although CCR5 editing did not reach threshold levels for post-cART control of viraemia, correlates of protection were observed: quantitative viral outgrowth, Tat/Rev-induced limiting dilution and tissue viraemia assays showed that the size of SHIV reservoirs was impacted.
Conclusions: Our data identify perturbations of the immune system as a mechanism for the failure of autologous transplantation (without gene-protected cells) to eradicate HIV, highlighting the importance of an intact immune system for viral control after cART withdrawal. We conclude that reduced-intensity conditioning, which is safer and less toxic, should be a focus for transplantbased approaches. Analogous to cutting-edge therapies for cancer patients, next-generation HIV cure strategies should balance killing of virus-infected target cells with retention of greater immune function, for example, with immune modulators. High-efficiency gene therapy/gene editing to protect transplanted cells and actively target the viral reservoir during ongoing cART will be essential. Background: HIV-infected adults on ART experience high rates of non-communicable diseases (NCDs). These comorbidities can accumulate, and older HIV-infected adults often suffer from multi-morbidity. Little is known of the burden of multimorbidity in HIV-infected adults in low-and middle-income countries.

WEAB0101
Methods: HIV-Brazil Cohort is an observational, multi-site study of HIV-infected adults initiating ART between 2003 and 2014. We studied NCDs and multi-morbidity in patients from seven clinical sites in six Brazilian cities. NCDs included coronary artery disease, cerebrovascular disease, high-grade hyperlipidaemia (HLD), diabetes, chronic kidney disease, cirrhosis, osteoporosis, osteonecrosis, venous thromboembolism (VTE) and non-AIDS-defining cancers. Multi-morbidity was defined as the incident accumulation of two or more unique NCDs. We examined incidence trends using Poisson regression and predictors of multi-morbidity using competing risk and Cox regression models.
Results: Of the 5786 adults included, 388 (7%) developed multimorbidity during the study period. From 2003 to 2014, parallel to the rise of patients over the age of 50 in the cohort, the incidence of multi-morbidity rose to 21 patients per 1000 person-years ( Figure 1). HLD and VTE incidence decreased while diabetes and osteoporosis rates significantly increased from 2003 to 2014. In adjusted Cox models, female sex, age and low CD4 nadir at baseline were significantly associated with risk of multi-morbidity (Table 1, also adjusting for education, race, year, hepatitis C). Among all patients with multi-morbidity, the most common NCDs were HLD (87%) and diabetes (59%); however, women with multi-morbidity were more likely to have osteoporosis than men (15.4% vs. 6.8%

Oral Abstracts
Background: Peripheral arterial disease (PAD) affects approximately 8-10 million U.S. adults annually and is the second most common clinical manifestation of atherosclerosis after acute myocardial infarction (AMI). While the increased risk of AMI and ischemic stroke among HIV-infected (HIV positive) compared to uninfected people is well documented, data linking HIV to incident PAD events are sparse. We, therefore, compared PAD risk among HIV-positive and uninfected veterans. Methods: We analysed data on 91,457 veterans (33% HIV positive) without prevalent cardiovascular disease from the Veterans Aging Cohort Study (VACS). VACS is an observational, longitudinal cohort of HIV-positive veterans matched 1:2 with uninfected veterans on age, gender, race/ethnicity and clinical site. Participants were followed from their first clinical encounter on or after 1 April 2003 until a PAD event, death, their last follow-up date or 30 September 2012. We used ICD-9 and CPT codes to identify participants with incident PAD. Cox proportional hazard regression models were utilized to assess the association between HIV, CD4+ T cell count and PAD adjusting for atherosclerotic risk factors (Table 1). Finally, we constructed cumulative incidence curves to examine PAD risk stratified by HIV status and CD4+ T cell count.
Results: During a median follow-up of seven years, there were 5091 PAD events. See Table 1 and Figure 1 for rates and risk of PAD stratified by HIV status and CD4+ T cell count.
Conclusions and Relevance: HIV-positive veterans have a significantly higher risk of PAD than uninfected veterans.

WEAB0103
Impact of exposure to each antiretroviral treatment (ARV) on the risk of fracture in HIV-1-infected individuals: an analysis from FHDH ANRS CO4 Abstract WEAB0102- Figure 1. Cumulative incidence of PAD stratified by HIV status and CD+ T cell count.

WEAB0105
SHIV infection and drug transporters influence brain tissue concentrations of efavirenz that inadequate ARV concentrations may contribute. However, information on brain tissue concentrations is sparse. This study compared the concentration of ARVs in four regions of brain tissue with CSF in uninfected and SHIV-infected rhesus macaques. Methods: In 12 male macaques (6 uninfected; dosed to steady-state condition, concentrations of 6 ARVs -tenofovir (TFV), emtricitabine (FTC), efavirenz (EFV), raltegravir (RAL), maraviroc (MVC) and atazanavir (ATZ) -were measured by LC-MS/ MS in the CSF (LLOQ = 0.5 ng/ml) and cerebrum, cerebellum, basal ganglia and parietal cortex regions of the brain (LLOQ of homogenate ranged from 0.002 to 0.01 ng/ml). Tissue concentrations were converted into ng/g using density of 1.06. To assess the influence of drug transporters on ARV concentration, brain tissue was analysed for Pgp and BCRP efflux transporter proteins by LC-MS proteomics (LLOQ = 0.1 pMol/mg protein). Data are presented as median (range); statistical analysis was by Kruskal-Wallis test.

Results:
CSF concentrations did not differ by infection status (p > 0.1). Since there was no difference in ARV concentration in the various regions of the brain (p > 0.1), these data were combined. Concentrations in brain tissue were significantly greater than CSF for TFV, FTC and EFV: ranging from 5-times (FTC) to 769-times (EFV) higher. Brain tissue concentration of EFV was 4.1 times higher in uninfected animals. BCRP concentration was 1.7 times higher in infected animals (p = 0.02); Pgp concentration did not differ with infection status (p = 0.06).
Conclusions: In this study, brain tissue concentration of EFV was fourfold lower in infected macaques, and this may be due to increased BCRP concentrations. Further, we have shown that ARV CSF concentrations may need cautious interpretation when used as surrogate for brain tissue exposure. Based on these data, further investigations are needed to determine how ARV brain tissue concentrations influence HAND prevalence.

WEAC0101
Barriers to uptake of pre-exposure prophylaxis among respondents to the Flash! PrEP in Europe survey Methods: MSM undergoing anonymous HIV testing completed a DCE with 12 "choice sets" by selecting their preferred option within each set. Each set included "usual methods to prevent HIV infection" (excluding pre-exposure prophylaxis) as one option and two hypothetical NPT options which differed according to HIV prevention efficacy (50%, 65%, 80% or 99% risk reduction), route of administration, side effects (none or mild) and risk of drug resistance (none, low or moderate). We used mixed logistic regression to infer relative preferences for NPT attributes and latent class analysis to determine patterns of responses. Results: Of 306 participants, 54% were white and median (interquartile range) age was 30 (25, 38). Participants reported 6 (3, 10) partners and 0 (0, 2) condomless receptive anal sex-acts in the preceding six months. Most had heard of post-exposure prophylaxis (80%) and pre-exposure prophylaxis (91%), but only 11% and 5%, respectively, had used them. We excluded 40 participants who had all missing data or gave invariant responses. An on-demand pill was the most preferred NPT, followed by a daily pill, monthly injection and on-demand rectal gel. Resistance was an important determinant of NPT preference if the risk was moderate, but not if low. The minimum NPT efficacy required for an on-demand pill to be preferred over usual methods was 52.8% (95% confidence interval (CI) = 46.9-58.7); for a daily pill, injections and rectal gel, estimates were 60.1% (95% CI = 53.8-66.5), 67.0% (95% CI = 61.0-73.0) and 78.3% (95% CI = 70.9-85.7), respectively. Latent class analysis identified one subset of participants clearly favouring ondemand PrEP (40.5%) and three others preferring usual methods but with an aversion to injections (20.7%), aversion to rectal gels (21.9%) or relative indifference to NPTs ( Background: PrEP is now recommended for high-risk persons in Africa. There are limited data on PrEP uptake in Africa outside of clinical efficacy trials. "Early adopters" can provide insights for programme strengthening. We report on early PrEP adopters in SEARCH (NCT01864603), an ongoing population-based combination prevention study of 320,000 persons in rural Uganda and Kenya. Methods: Following mobilization and community education, two groups were offered PrEP: (i)HIV-uninfected adults at high risk (R) based an HIV risk score that maximized observed seroconversion coverage under a minimized number of persons needed to treat and (ii)those who perceived themselves at risk (S) including being in serodiscordant relationship. "Early adopters" were defined as those who started PrEP within 30 days of being offered. To estimate predictor coefficients for early PrEP uptake, we used generalized linear models with binomial distribution. Results: Of 24,709 HIV-uninfected individuals in six communities, 4622 were identified for PrEP: 2995 based on risk score (Rs) and 1627 as self-referrals (Ss). A total of 2374 (51%) scheduled an appointment with 946 (20%) initiating PrEP; 916 (97%) of these were "early adopters" with a vast majority 712 (78%) starting PrEP on the same day. "Early adopters" tended to be Ss (64%), women (52%) and married (68% 2) and a serodiscordant spouse (OR 2.5, 95% CI 1.0-6.1). Conclusions: Among the 916 PrEP "early adopters", most started the same day as offered, two-thirds were married and perceived themselves as high risk. Low participation among certain crucial groups such as youth (18-25 years) emphasizes the need for more effective mobilization.

WEAC0104
Health systems and study design features permitting rapid enrolment of individuals at high-risk of HIV acquisition into a pre-exposure prophylaxis study in Melbourne, Victoria, Australia Background: Australia's Medicare system provides clinicians feefor-service and residents receive free or low-cost healthcare. Australia's Pharmaceutical Benefits Scheme (PBS) subsidizes medication costs. Tenofovir/emtricitabine (TDF/FTC) is registered in Australia for HIV pre-exposure prophylaxis (PrEP) but is not PBS subsidized; hence, individuals must import generic TDF/FTC or pay A$800 monthly. In 2016, we implemented a 2600-person PrEP demonstration study in Victoria, hypothesizing a resultant 33% decline in new HIV infections in men who have sex with men (MSM). We describe the health systems and processes that facilitated rapid study enrolment and concomitant increases in HIV and sexually transmitted infection (STI) testing. Methods: Victoria's population is approximately 6 million, including an estimated 37,000 HIV negative, sexually active MSM. From January 2016, individuals registered study interest online and nominated which of seven study clinics in Melbourne they would attend, whether they already attended that clinic and whether they were currently using PrEP. At study commencement, on 26 July 2016, 2198 individuals had registered interest. Key community stakeholders, study clinics and retail pharmacies were engaged in the study design and service system planning. Clinics were incentivized with A$100/participant or a study nurse. Australian PrEP guidelines specified eligible individuals at high risk. Participants were enrolled electronically. HIV/STI test results were extracted automatically using a sentinel surveillance system (ACCESS), extant in five clinics. We report HIV and syphilis testing rates in three of seven clinics across five-month pre-intervention (26 July 2015-26 December 2015) and PrEP-intervention (26 July 2016-26 December 2016) periods. Results: A total of 1000 participants were enrolled within 21 days of the study commencing, and one in three participants were using PrEP; 2350 participants were enrolled in six months. Six clinics chose the A$100 payment per patient. HIV tests increased from 3009 to 4952, and syphilis tests increased from 2926 to 4704 compared to the same five-month period in 2015, respectively. Conclusions: In a free healthcare system that provides clinicians fee-for-service, rapid enrolment into PrEP programmes appears feasible. A detailed registry of interest, prior use of PrEP, clinic remuneration, electronic enrolment and data extraction and collaborative planning were features of the study's rapid enrolment rate. A substantial rise in HIV and syphilis testing accompanied the study rollout.  Background: Oral, daily pre-exposure prophylaxis (PrEP) prevents HIV acquisition in optimally adherent men who have sex with men (MSM). Given the importance of adherence in PrEP-related outcomes, accurately and affordably monitoring adherence is a priority during implementation. We evaluated two low-burden measurements, self-report (SR) and medication possession ratio (MPR), for concordance with the well-established method of determining tenofovir diphosphate (TFV-DP) levels in dried blood spot (DBS). Methods: We reviewed behavioural and DBS data on patients presenting to the Washington University in St. Louis (USA) PrEP Clinic between November 2015 and August 2016. Optimal adherence was defined as TFV-DP ≥700 fmol/punch and was compared to patient seven-day SR and three-month MPR using pharmacy refill data. Sensitivity, specificity and negative and positive predictive value (NPV, PPV) for SR and MPR in relation to DBS were calculated. Results: From 88 MSM, 137 DBS TFV-DP levels were analysed. Their median age was 27 years; 58% were white, 30% black, 6% Latino; 69% graduated college; and 71% reported condomless receptive anal sex in the last three months. Ten patients had a DBS <700 fmol/punch. Drug concentration was not related to demography and did not significantly decline over time. By SR, five patients took <4 doses/week, four of whom had sub-optimal DBS (NPV 80%), and of the 83 reporting ≥4 doses/week, 77 had optimal DBS (PPV 93%), resulting in 99% sensitivity and 40% specificity. For MPR, three patients had an MPR <0.60 (indicating <4 doses/week), all of whom had sub-optimal DBS (NPV 100%), and of the 84 with MPR ≥0.60, 77 had optimal DBS (PPV 92%), resulting in 100% sensitivity and 30% specificity. MPR and SR correlated with DBS TFV-DP levels (r = 0.55, p < 0.001; r = 0.48, p < 0.001). More stringent cutoffs to the strategies produced higher specificity -60% for SR ≥6 doses/week and 70% for MPR at 0.70. Conclusions: In a real-world clinical setting, SR and MPR correlated with optimal DBS concentrations despite different measurement windows (past 7, 30 and 90 days). Specificity in this sample was improved when more stringent SR and MPR cutoffs were used. Results provide evidence for using low-burden measurements for PrEP adherence monitoring. Background: Late-presenter pregnant women need aggressive antiretroviral therapy to reach a plasma viral load (PVL) <50 copies/ml before delivering. We compared the safety and efficacy of LPV/r Abstracts of the 9th IAS Conference on HIV Science (IAS 2017) Journal of the International AIDS Society 2017, 20 (Suppl 5) Oral Abstracts and raltegravir (RAL) in decreasing PVL in late-presenter pregnant women in Salvador, Brazil. Methods: in this open-label, pilot trial (N = 40), we included drugnaive pregnant women who started antiretroviral therapy (ART) at a gestational age ≥28 weeks. They were randomly assigned to receive AZT+3TC+LPV/r or AZT+3TC+RAL. We measured time to reach undetectable PVL and compared the proportion of women with PVL <50 copies/ml at delivery, in each group. PVL was measured weekly by real-time PCR, up to delivery. Frequency of side effects and MTCT rate were assessed. Babies were tested for HIV-1 plasma RNA at four weeks of age. Results: we already enrolled 28 women (14/arm). Groups were comparable for age, education, smoking/alcohol use and number of previous gestations/miscarriages. Most (73%) were black/racially mixed and single (82%). Twenty-five women already completed the trial. Median baseline PVL was similar for LPV/r (4.26 log 10 , interquartile range (IQR): 4.02-4.04) and RAL (4.05 log 10 , IQR: 3.55-4.31) groups, as well as mean gestational age (32.8 ± 11.4 vs. 32.9 ± 10.4 weeks, respectively). At delivery, only 1/11 (9%; 95% confidence interval (CI): 0.5-37%) women in group LPV/r had PVL <40 cps/ml, compared with 9/14 (62%; 95% CI:38-86%) in RAL group (p = 0.01). The median time to reach undetectable PVL was significantly shorter for RAL (44 days) in comparison with LPV/r arm (69 days, p = 0.009). In RAL arm, 2 (14%) patients reached PVL <50 cps/ml at week 1 and 3 (21%) at week 2 of ART. In contrast, in LPV/r arm, the first PVL <50 copies/ml was reached only after six weeks of therapy. More gastrointestinal adverse events were observed in LPV/r arm (5/11) than in RAL one (1/14). No case of MTCT was detected. Conclusions: use of RAL was associated with significantly higher rate of undetectable PVL at delivery and lower incidence of adverse events, in comparison with LPV/r. Time to reach undetectable PVL was significantly shorter in RAL group. RAL should be the preferential option to treat late-presenters pregnant women to minimize the risk of HIV-1 MTCT.

WEAC0202
Intensification of antiretroviral treatment with raltegravir for late-presenting HIV-infected pregnant women Background: The risk of HIV perinatal transmission in HIV-infected pregnant women who are started late on antiretroviral therapy during the third trimester is estimated to be up to 6-10%. Raltegravir, HIV integrase inhibitor, has rapid viral reduction and is recommended by the British HIV Association (BHIVA) guideline for late-presenting HIVpositive pregnant women. This study aims to describe HIV perinatal transmission from high-risk HIV-positive pregnant women who have received raltegravir intensification antiretroviral treatment. Methods: A prospective cohort study was conducted at the Thai Red Cross AIDS Research Centre. Inclusion criteria were HIV-positive pregnant women with high risk of HIV vertical transmission defined as (1) having been started on antiretroviral therapy (ART) at gestational age (GA) >32 weeks or (2) having received ART but having HIV-RNA >1000 copies/ml at GA 32-38 weeks. Pregnant women received standard three-drug ART regimen plus raltegravir 400mg twice daily until delivery and then were continued on threedrug ART after delivery. Plasma HIV-RNA was performed before adding raltegravir and at the time of delivery. The HIV status of infant was determined by HIV-DNA PCR at birth and one, two and four months. Results: From January to December 2016, 57 pregnant women were enrolled. Median CD4 count was 307 cell/mm 3 (interquartile range (IQR) 155-507). Median plasma HIV-RNA before initiation of raltegravir was 3.6 log 10 copies/ml . Median GA at time of starting raltegravir was 35 weeks (IQR 33-37). Combinations of ART were 32 EFV based (56%), 21 LPV/r based (37%) and 4 others (7%). Median duration of raltegravir was 18 days (IQR 7-28). The proportion of pregnant women who had plasma HIV-RNA <50 and <1000 copies/ml at time of delivery were 47% and 81%, respectively. To date, 50 infants were born, 40% by caesarean section and 8% preterm (GA <37 weeks). HIV perinatal transmission rate was 0% (95% confidence interval 0-6.3). No rash, hepatitis and jaundice in mothers or infants have been reported. Conclusions: No HIV vertical transmission occurred among high-risk HIV pregnant women who received raltegravir intensification ART. This strategy is feasible and effective, supporting elimination of HIV mother-to-child transmission. Background: Using spatial-temporal analyses to understand coverage and trends in elimination of mother-to-child transmission of HIV (e-MTCT) may be helpful in understanding effectiveness of interventions while refocusing e-MTCT programme efforts to the right places to achieve epidemic control. We measured MTCT rates using early infant diagnosis (EID) programme data collected from January 2007 to November 2013 in Western Kenya and assessed associated HIV transmission risk factors within a spatial context irrespective of treatment guideline changes. Methods: We performed trend analysis for 102,116 HIV-exposed infants (HEIs) using extended Cochran-Mantel-Haenszel stratified test and logistic regression models to determine associations of infant HIV status with maternal and infant characteristics recorded on EID laboratory test request forms. We fitted spatial and spatialtemporal semi-parametric Poisson regression models with infant and maternal covariates to explain MTCT rates. We used R-Integrated Nested Laplace Approximation (INLA) package and Quantum GIS to map raw and fitted estimates. Results: Median age of HEIs was two months, interquartile range (IQR) 1.5-6 months. Pooled positivity was 11.8% in the seven-year period, which significantly spatial-temporarily declined from 17. Oral Abstracts model with covariates was better in explaining geographical variation in MTCT (deviance information criterion (DIC 296)) than either a non-temporal spatial model (DIC 326) or temporal model without covariates (DIC 311). Conclusions: Improved EID uptake and reduced MTCT rates are indicators of success of the e-MTCT programme in this low-resource setting. Adding both time and covariates in spatial-temporal analysis provides a robust approach for explaining programmatic impact over time. Geographical disparities in programme achievements may signify gaps in spatial distribution of e-MTCT efforts and indicate areas needing further resources and intervention.

WEAC0204
Cost and cost-effectiveness analysis of a randomized controlled trial evaluating perinatal home visiting among South African mothers/infants A Wynn 1 ; M Tomlinson 2 ; MJ Rotheram 1 and I Le Roux 3 1 University of California, Los Angeles, Psychiatry, Los Angeles, USA. 2 Stellenbosch University, Psychology, Stellenbosch, South Africa. 3 Philani Nutrition Centers, Cape Town, South Africa Presenting author email: mrotheram@mednet.ucla.edu Background: South Africa faces a high antenatal HIV prevalence and infant mortality rate. Community-based programmes involving home visits contribute to reductions in neonatal mortality. However, most low-and middle-income countries lack the budget to deliver such preventive interventions by nurses. Therefore, paraprofessional interventionists may be an innovative alternative strategy to healthcare workers. This study assesses the costs and benefits of implementing a home visiting programme utilizing community health workers (CHWs). Methods: We conducted an economic evaluation alongside a cluster RCT in Cape Town, South Africa, called Philani+. The trial assessed the impacts of training CHWs to deliver antenatal and postnatal home visits to address maternal and child health risks.
Financial costs were collected from the perspective of the health system. We calculated incremental cost-effectiveness ratios by dividing the costs of the intervention by the number of low birthweight newborns and cases of infant undernutrition averted among intervention participants compared with controls. These measures are strong indicators of maternal and newborn health. Numbers of averted cases were modelled as the product of intervention subjects and differences in the rates of adverse outcomes between intervention and control groups. Results: The total cost of the intervention over 24 months was estimated at US$91,574. The average cost of supporting 12 CHWs was $7631 per CHW and the cost per mother was US$142. The intervention group had higher HIV treatment adherence and longer breastfeeding duration. The intervention was associated with averting an estimated 55 (90% CI: 41-74) cases of low birthweight and 59 (90% CI: 42-83) cases of undernutrition. The estimated cost per low birthweight averted was $1664, and the estimated cost of averting an undernourished child was US$1552. Conclusions: Philani+ was innovative because it integrated HIV care and prevention with activities to improve maternal and child health. The employment of CHWs provides cost savings compared to use of nurses and builds capacity in a country with a high unemployment rate and shortage of healthcare workers. Finally, Philani+ was able to improve child health at a relatively low cost considering the costs to the health system caused by low birthweight and undernutrition.

WEAC0205
A community-based, household survey to determine mother to child HIV transmission rates and HIV-free survival in Swaziland Background: The Joint United Nations Programme on HIV/AIDS renewed efforts to virtually eliminate mother-to-child HIV transmission (MTCT) with a target of reducing the mother-to-child transmission rate to 5% or less among breastfeeding populations by breastfeeding cessation and to 2% or less among non-breastfeeding populations. In Swaziland, although data are available on MTCT rates at 6 weeks, no study has been performed to determine MTCT and HIVfree survival through the end of breastfeeding. Methods: The Elizabeth Glaser Pediatric AIDS Foundation performed a national, cross-sectional study of children born 18-24 months prior to the study launch among HIV-infected mothers to determine MTCT rates and HIV-free survival through a community survey in randomly selected constituencies in all four regions of Swaziland. At the time of this cohort's birth, Swaziland had been implementing World Health Organization Option A for prevention of MTCT (PMTCT). We also evaluated the relationship between both maternal and child characteristics and child infection or death. Results: Most HIV-positive mothers (91.8%) received antiretroviral prophylaxis for PMTCT or antiretroviral treatment during pregnancy. Among 724 known HIV-exposed children between 18 and 24 months, 26 children were HIV-positive and 694 were HIV-negative and alive. Four (all with unknown HIV status at time of death) HIV-exposed children died by 24 months of birth. The overall 18-24-month HIV-free survival among this cohort was 95.9% (95% CI: 94.1-97.2). At 18-24 months, the estimated proportion of HIV-positive children among known HIV-exposed children was 3.6% (95% CI: 2.4-5.2). Older maternal age, delivering in a health facility, high maternal CD4 count and receiving antenatal antiretroviral drugs were associated with reduced risk of child infection or death. Child hospitalization was associated with higher rates of child HIV infection or death.

Conclusions:
The Swaziland PMTCT programme under Option A was largely effective with a high HIV-free survival of 95.9% and low MTCT at 18-24 months of 3.6%. This would be expected to improve further under current Option B+ (universal maternal antiretroviral therapy).

WEAD0101
Trends in pediatric HIV testing across six African countries

Oral Abstracts
Conclusions: The rate of HIV positivity in most countries remained relatively stable, although there was large decrease between Q3 2015 and Q4 2015 in Malawi. Intensified pediatric HIV testing dramatically increased the numbers tested and identified 9688 HIV-positive children. Overall HIV positivity was relatively low, likely due to effective prevention of mother-to-child HIV transmission programmes in these countries combined with mortality among undiagnosed older children. Testing of higher-risk children (e.g. TB clinics, malnutrition wards, risk-screening in OPDs) as opposed to broad testing may be a more effective way to identify HIV-positive children in an era of maternal treatment.  Table (FIT) and FIT chart audits; community outreach testing; and text message reminders for HIV-exposed infants. We compared the number of children tested monthly and the number of HIV-positive children between intervention and control sites using negative binomial generalized estimating equations. Analyses adjusted for repeated measures, geographic location, health facility tier and test kit stock-outs. Results: Mean number of children tested monthly increased across all age groups: from 2.8 to 7.2 (p < 0.0001) in infants <18 months; from 44.8 to 142.0 (p < 0.0001) in children 18 months to 9 years; and from 30.1 to 123.3 (p < 0.0001) in adolescents 10-14 years. Identification of HIV-positive children increased: 0.06 to 0.37 (per month per facility; p < 0.0001) in infants; 0.34 to 0.62 (p = 0.002) in children; and 0.17 to 0.26 (p = 0.03) in adolescents. Use of the FIT was significantly associated with increased HIV testing in infants, incidence rate ratio (IRR) = 2.89 (95% confidence interval (CI) = 1.53-5.49; p < 0.001) and identification of HIV-positive infants, IRR = 8.71 (95% CI = 1.45-52.4; p < 0.02). Among children, FIT chart audits were significantly associated with increased testing, IRR = 2.15 (95% CI = 1.36-3.40; p < 0.001). Among adolescents, HTC space was significantly associated with increased HIV testing, IRR = 1.45 (95% CI = 1.09-1.93; p < 0.01). Conclusions: Targeted testing of family members of HIV-positive adults increased both testing and identification of HIV-positive children. Our findings suggest that the one-time investment in improving HTC space may be an effective approach for increasing HIV testing among adolescents in this context. Significant increases in number of children tested resulted in only a modest number of new children identified with HIV, highlighting the need for multiple testing approaches.

WEAD0103
Disclosure of HIV status to children living with HIV in Malawi: needs assessment and formative evaluation of an intervention intended to help with the disclosure process Organisation recommends that disclosure of HIV status should take place between the ages of 6 and 12 years, very little is known about the practice of HIV disclosure in Malawi. This study aimed to evaluate the current practice of HIV disclosure to Malawian children and to assess the acceptability of a series of age-appropriate, culturally acceptable story books intended to help with the disclosure process. Methods: Questionnaires, interviews and focus group discussions were used to collect data from caregivers, healthcare workers, school teachers, adolescents living with HIV and community leaders across the three administrative regions of Malawi. Data on disclosure of HIV to the child, reasons for non-disclosure, the need and acceptability of the proposed series of story books, the child's mental health and the family psychosocial characteristics were collected using reliable instruments. Data were analysed using chisquare test, multiple logistic regression and thematic analysis. Results: The response rate was 99%: 600 questionnaires, 19 interviews and 12 focus groups were completed. The prevalence of nondisclosure was 64%. Non-disclosure of HIV status was more likely for younger children (aOR 3.8; 95% CI: 2.1-6.8), those in a farming family (aOR 3.4; 95% CI: 1.2-9.3) and those whose healthcare workers lacked training about disclosure (aOR 7.7; 95% CI: 3.4-11.6). The lack of disclosure guidelines and materials (33%), the child's capacity to cope with the diagnosis (29%) and a lack of confidence to disclose appropriately (19%) were cited as the main reasons for non-disclosure. Ninety-eight per cent of participants supported the idea of developing the proposed series of story books. More than threequarters of the participants emphasized the need for all stakeholders involved in caring for children living with HIV to work together towards promoting effective HIV disclosure. Conclusions: The rate of non-disclosure in Malawi is high. The results of this study support the need for the development and rigorous evaluation of disclosure materials and the involvement of all stakeholders to promote effective disclosure and meet the evolving needs of children.

WEAD0104
An assessment of the effectiveness of reaching undiagnosed HIV-positive children through communitybased testing in Lesotho Oral Abstracts the country adopted the WHO-recommended "Treat All" policy in 2016. However, the antiretroviral treatment coverage for children, 0-14 years, is less than 60%, suggesting many children are not regularly accessing services at facilities. Recognizing that testing is the critical first step in initiating HIV-positive children to lifesaving treatment, community-based PITC strategies were piloted to understand the effectiveness of identifying undiagnosed children in Lesotho beyond the facility. Methods: From December 2015 to December 2016, four community-based strategies were utilized: (1) mobile outreach clinics; (2) door-to-door testing; (3) household-based index patient testing; and (4) targeted testing events conducted at venues thought to have high-risk children. A mobile application was designed for use by healthcare workers for real-time data collection at point-of-care to capture data on newly identified HIV-positive children; this included residence, gender and age, as well as HIV testing history and health facility attendance history. Results: Out of 36,121 children tested, 161 were HIV positive (0.44% positivity yield), and 123 were enrolled in the programme with a 1:1.2 male-to-female ratio. Only 23.5% of all enrolled patients were known to be previously tested. Overall, 12% were 0-2 years, 25% were 2-5 years and 63% were 5-14 years. Of the 52% of patients with known facility attendance history, 20% had not been in over one year, 17% had attended within four months to one year and 15% had visited a facility within the previous three months.
Conclusions: The majority of children testing HIV positive through community-based PITC were 5-14 years, nearly half had unknown facility attendance history and 9.5% who had been to a facility in the previous year did not receive an HIV test. This establishes the need for investigation into causes of Lesotho's shortcomings in offering facility-based PITC. Additionally, it proves that community-based PITC is a necessary tool in closing the access gap by bringing these lifesaving services to children in need, particularly older children. National scale-up is essential in reaching UNAIDS' first 90 target and is recommended for other countries struggling to achieve widespread coverage for children through facility-based testing.

WEAD0105
The clinical impact and cost-effectiveness of incorporating point-of-care (POC) assays into early infant HIV diagnosis (EID) programmes at 6 weeks of age in Zimbabwe: a model-based analysis SC Frank 1,2 ; J Cohn 3,4 ; L Dunning 5 ; E Sacks 6 ; RP Walensky 1,2,7 ; S Mukherjee 6 ; E Turunga 3 ; KA Freedberg 1,2,7 and AL Ciaranello 1,2 Background: Many EID programmes use laboratory-based total nucleic acid (conventional) assays. New POC EID assays are costlier, but may increase access to testing and shorten time to resultreturn and ART initiation. Methods: We used the CEPAC-Pediatric model to examine the clinical benefits, costs and cost-effectiveness of using POC EID assays at 6 weeks of age in Zimbabwe. We simulated two EID strategies: conventional and POC. Positive results led to ART initiation; ART was stopped if a confirmatory assay of the same type and a third conventional assay (all sent pre-ART) were negative. Modelled assays differed in sensitivity (conventional: 100%; POC: 96.9%), specificity (conventional: 98.8%; POC: 100%), time and probability of result-return (conventional: one-month delay, 71%; POC: immediate, 97%) and cost (conventional: $15; POC: $21). Model outcomes included early survival, life expectancy (LE), and average lifetime per-person cost for HIV-infected infants and all HIV-exposed infants. We calculated incremental cost-effectiveness ratios (ICERs) using discounted (3%/year) costs and LE for all HIV-exposed infants, defining ICERs ≤$930/life-year saved (Zimbabwe per-capita GDP) as cost-effective. Results: With conventional EID, projected undiscounted LE was 24.95 years (HIV-infected infants) and 60.16 years (all HIVexposed infants), at $1050/HIV-exposed infant (Table 1). POC EID improved projected undiscounted LE (HIV-infected: 26.58 years, HIV-exposed: 60.27 years) at $1120/infant, and increased survival by 4.5% in months 1-2 of life. The ICER of POC vs. conventional was $730/life-year saved (LYS). This ICER remained <$930/LYS if POC specificity was >95% or POC sensitivity was >85%. Large improvements in conventional assay result-return were needed to offset slightly lower POC assay sensitivity ( Figure 1). Conclusions: POC assays for HIV-exposed infants improve survival and LE and are cost-effective compared to conventional assays. EID programmes in Zimbabwe should replace conventional testing with POC assays.
Abstract WEAD0105 - Background: From 2008 to 2014, USAID-funded organizations were responsible for directly delivering HIV care and treatment via "Comprehensive Care, Management and Treatment" facilities (CCMTs) in South Africa. Despite USAID having communicated its plans to phase out funding for direct service delivery (DSD) and instead focus primarily on technical assistance, there has been a sense that public clinics were not adequately prepared for the transition. The aim of this study was to examine the impact on financial and human resources and workloads immediately after the withdrawal of funds for a CCMT at a clinic in Johannesburg.
Methods: In late 2016, we conducted a natural experiment in which trends in budgets and expenditure, clinical staff complements, patient loads and services rendered were compared at the study clinic before ( Despite increased workload, service volumes for primary healthcare services at the clinic (HIV counselling and testing, tuberculosis testing and family planning) did not decrease after the departure of the CCMT.
Conclusions: The phasing out of funding for DSD by USAID at a clinic in Johannesburg negatively impacted on human resources and staff workload while decreasing per capita expenditure. The volume of primary healthcare services delivered at the clinic did not decline; however, the impact on service quality is unknown.

WEAD0202
How changes in United States funding policies could impact the HIV epidemic in sub-Saharan Africa the IMF to estimate the US share of the total HIV/AIDS response historically and in future. We then removed this US share from the total funding available, or changed the way it could be allocated to future prevention efforts, to explore a series of alternative policy strategies that the administration might adopt

WEAD0203
Estimating the size of the pediatric antiretroviral (ARV) market in 26 low-and middle-income countries (LMICs) through 2025 as prevention of mother to child transmission (PMTCT) initiatives continue to succeed VR Prabhu 1 ; S Mcgovern 1 and P Domanico 2 Background: PMTCT initiatives have significantly reduced HIV infections in children, a trend expected to accelerate with the "Start Free, Stay Free, AIDS Free" initiative. Its "Super Fast-Track" targets aim, by 2020, to reduce new perinatal infections to <20,000 annually and achieve universal pediatric ART coverage. Population projections are needed to help optimize supply of pediatric ARVs, especially new formulations.
Methods: Provisional single age band estimates for HIVinfected children (ages 0-14) for 26 high-burden LMICs, provided by UNAIDS, representing approximately 77% of the global pediatric HIV burden in 2015, were used. For each forecast year, age cohorts were moved from one age bracket to the next, new perinatal HIV infections were added to the age 0 cohort and the age 14 cohort "aged out" of the market. Annual decreases in AIDS-related deaths were assumed to be similar to 2011-2015, and evenly distributed across age groups. Three scenarios of the "Super Fast-Track" targets for new infections were modelled: "Aggressive" (targets met), "Moderate" (two-year delay) and "Conservative" (similar decrease as 2011-2015). Resulting cohorts were converted to weight bands using published tables.
Results: The number of HIV-infected children will continue to rise until 2019 before decreasing. By 2025, there could be between 350,000 and 500,000 children living with HIV, needing ART, across the 26 high volume countries. The more aggressive scenarios suggest an increasing proportion of >20 kg children.
Conclusions: As pediatric ART coverage increases, more children will need second-or third-line treatment. These population estimates can inform discussions on the development of new pediatric ARV formulations.
Abstract WEAD0203 - Figure 1. Total ped. PLWHA. Intact proviruses, capable of infectious virus production, were identified using viral outgrowth assays (VOAs). The levels of viral RNA present in infected cells were determined for the archival drug-sensitive population, the recently infected DR population and for clones carrying intact and defective proviruses. Results: We analysed a total of 77 million PBMC, of which 10,450 contained HIV pro-pol sequences: 7137 were WT, 1714 were DR and 1599 were defective (contained stop codons). The median fraction of proviruses that expressed RNA in cells more recently infected with DR virus was 25%, whereas in cells with WT or hypermutant proviruses, it was 14% (p = 0.0008). Levels of expression in single cells with DR proviruses were higher than in cells with WT proviruses (p = 0.002). The median fraction of cells in apparent clonal populations carrying intact proviruses (N = 3) expressing HIV RNA was 2.3% (1.2-8.8%). For clones carrying defective proviruses (N = 5), the median fraction expressing was 3.5% (0.9-7.0%), and for clones carrying proviruses that did not have obvious defects in the pro-pol region but were not recovered in the VOA (N = 26), the median was 6.6% (1.3-66.7%).
Conclusions: A small fraction of the proviruses in clones of HIVinfected cells expressed HIV RNA. The fraction and levels of proviral expression were significantly higher in more recently infected cells than in those that persisted during long-term ART. These findings show that ART can select both for cells infected before ART initiation that either do not express HIV RNA or express at low levels and for cells infected recently with drug-resistant viruses that express higher levels of HIV RNA.

MOPDA0102
Evidence of production of HIV-1 proteins from "defective" HIV-1 proviruses in vivo: implication for persistent immune activation and HIV-1 pathogenesis Background: Greater than 95% of proviruses detected in circulating peripheral blood mononuclear cells are referred to as "defective" and are unable to encode intact viruses. They have been thought to represent a silent graveyard of viruses with little contribution to HIV-1 pathogenesis. We have recently shown that these "defective" proviruses are capable of transcribing novel mRNA species. In the present study, we demonstrate that these "defective" proviruses are also capable of producing HIV-1 proteins.
Methods: CD4+ T cell clones were obtained from an HIV-infected individual who had recently been placed on suppressive combination antiretroviral therapy (cART). The CD4+ T cells were initially plated in 96-well cell culture plates at a cell density of 100 cells/ well and expanded for 2 weeks in the presence of autologous feeder cells. Positive wells by HIV-DNA PCR were further expanded in 48-well plates for another week. The identification of single-cell clones harbouring "defective" proviral DNA was confirmed by combining 5´LTR-to-3´LTR single-genome amplification and direct amplicon sequencing of the genomic DNA. Cellular expression of HIV-1 proteins was analysed by western blot and flow cytometry. Results: Two months after suppression of plasma viremia to <40 copies/ml, the estimated frequency of CD4+ T cells containing HIV-DNA was 1%. Multiple cell lines harbouring defective proviruses ranging from 6.5 to 8.2 kb in length were derived from the CD4+ T cells. Most prominent among these were cells containing an identical 6.5 kb provirus. Sequencing of this 6.5 kb provirus revealed a 2.4 kb internal deletion affecting the region encoding the HIV-1 accessary proteins and the gp120 portion of Env protein.
The Gag, Pol and Nef regions remained intact in the 6.5 kb provirus. Consistent with the DNA data, western blots revealed the presence of the Gag and Nef proteins. Conclusions: These data indicate that "defective" proviruses in successfully treated HIV-infected patients are not silent dead-end products but rather capable of producing HIV-1 proteins in vivo. The proteins encoded by these defective "zombie" proviruses may be responsible for persistent seropositivity and immune activation in most patients with controlled HIV-1 infection during suppressive cART.

MOPDA0103
HIV reservoirs in the brain and association with sex and neurocognition HIV-DNA was detected in 62.5% of brain and 100% of lymphoid tissue. Lymphoid tissue has higher HIV-DNA levels than brain (85.6 vs. 14.2, p < 0.001). BG has higher HIV-DNA levels (20.3 copies/10 6 cells) compared to FC (13, p = 0.018) and OCC (9.3, p = 0.005). Female sex and younger age are associated with higher HIV-DNA in brain (p = 0.026 and p = 0.06, respectively), but not in lymphoid tissue (p = 0.31). When evaluating NC sub-domains, higher HIV-DNA (any brain region) was associated with worse speed of information processing (p = 0.012) and better verbal fluency (p < 0.05). No sex differences in Clinical Rating were observed.
Conclusions: HIV-DNA was detected in most of brains despite virologic suppression. While levels of HIV-DNA were comparable in lymphoid tissue, women had higher brain HIV-DNA levels than men. Higher brain HIV-DNA levels negatively affected the speed of information in both sexes. The negative association between HIV-DNA and verbal fluency requires more investigation.

MOPDA0104
Monocyte-derived reactive oxygen species impair CD4+ T cell restoration in HIV-1 patients under therapy Background: Antiretroviral therapy is highly efficient at suppressing viral replication in over 90% of HIV-1-infected patients. However, 5-25% of these virologic responders do not restore correctly their CD4 count. This suboptimal immunologic response is often correlated with a persistent hyperactivity of the immune system. However, the molecular mechanisms linking residual immune activation to impaired CD4 cell recovery remain to be identified. We tested the hypothesis that peripheral blood mononuclear cells (PBMCs) from aviremic HIV adults might induce DNA damage in CD4+ T cells resulting in their apoptosis. Methods: We probed by immunofluorescence the presence of g-H2AX, 53BP1 and 8-hydroxy-2-deoxyguanosine, markers of genostress, DNA double-strand break and oxidation, respectively, in primary fibroblasts co-cultured in transwells with PBMC from virologically suppressed patients. PBMC sub-populations were sorted using magnetic beads. The amount of reactive oxygen species (ROS) expressed in the monocytes and CD4+ T cell apoptosis were measured by flow cytometry using dichloro-dihydro-fluorescein diacetate and fluorescent Annexin V, respectively. DNA-dependent protein kinase (DNA-PK) and p53 phosphorylation were analysed by western blot.
Results: PBMC of 56 out of 103 virologic responders (54%) induced g-H2AX nuclear foci in cocultured fibroblasts. Cell sorting and inhibition of this phenomenon by a ROS scavenger and an NADPH oxidase-inhibitor established that this genostress, characterized by DNA oxidation and double-strand break, was due to ROS released by monocytes. In cocultured CD4+ T cells, this resulted in DNA-PK as well as p53 phosphorylation, and finally in apoptosis. Patients with PBMC able to damage DNA, a phenotype that we found stable over time, presented with lower CD4 recovery than patients whose PBMC did not induce DNA damage (p = 0.003 Background: Chronic T cell activation and dysfunction are hallmarks of HIV infection. Taking into consideration that T cell metabolism influences T cell functionality, we hypothesized that CD4+ T cell dysfunction during HIV infection could be associated with virus-induced metabolic alterations. A critical transcription factor in the coordination of T cell metabolism, differentiation and effector function is hypoxia inducible factor-1α (HIF-1α). Herein, we analysed the expression, activity and function of HIF-1α in CD4+ T cells.
Methods: CD4+ T cells isolated from the blood of healthy donors were activated with anti-CD3/CD28 antibodies and infected in vitro with HIV. HIF-1α activity was evaluated by using a reporter cell line, expressing GFP under the control of the Hypoxia-Responsive-Element. Cytokine production was evaluated by CBA kit. Cell viability was evaluated by 7-AAD staining and Annexin V binding. Silencing of HIF-1α expression was achieved by transduction with lentivirus-encoded shRNAs. To analyse ex vivo the relationship between HIF-1α levels and cell death in CD4+ T cells, a total of seven HIV-1-infected patients on cART and six healthy donors were recruited.
Results: We show that HIV-1 infection triggers HIF-1α expression and activity, promoting aerobic glycolysis and the production of proinflammatory cytokines in CD4+ T cells infected in vitro. We also observed that the promotion of aerobic glycolysis by HIV is associated with a higher rate of CD4+ T cell death. Remarkably, silencing HIF-1α expression in CD4+ T cells reverted the promotion of cell death and production of proinflammatory cytokines induced by HIV-1 infection. Finally, we also analysed HIF-1α expression in samples from HIV-1-infected patients on cART. Interestingly, these patients also exhibit higher levels of HIF-1α expression compared to healthy donors. Moreover, the expression levels of this transcription factor presented a negative correlation with CD4+ T cell counts.
Conclusions: In conclusion, we show that HIV infection induces the activity of HIF-1α in productively infected cells promoting glycolytic activity, a proinflammatory phenotype and cell death. These results pave the way to explore the possibility of targeting HIF-1α and/or T cell metabolism to restore CD4+ T cell physiology in HIV-1-infected individuals.

MOPDA0106
Toll-like receptor activation modulates inflammation and HIV-1 infection in the female reproductive tract (FEMINIVI study)    -4) and presence of NCI, defined as >1 SD below the mean on two or more tests or >2 SD below the mean on one test. We used linear and logistic regression models to determine the association between cardiovascular risk and the primary outcomes.
Results: Of 988 participants (30% black, 21% Hispanic/Latino, 20% women), mean age was 52 years and education 14 years. Median ART duration was eight years, mean CD4 count 661 cells/mm 3 and 90% of participants had viral load <40 copies/ mL. Current smoking (26%), statin (27%) and anti-hypertensive (36%) use were common, while stroke (2%), myocardial infarction (3%) and injection drug use (<1%) were uncommon. Mean LDL and HDL cholesterol were 109 and 49 mg/dL, respectively, and systolic blood pressure was 126 mmHg. One hundred and eighty participants (18%) had NCI. In demographics and education-adjusted models, higher HDL was associated with better NPZ-4 (+0.04, p = 0.040) and lower odds (OR 0.88, p = 0.043) of NCI per 10 mg/dL higher HDL, as was statin use (+0.15 NPZ-4, p = 0.037). An association between smoking and worse NPZ-4 (−0.15, p = 0.053) became non-significant after controlling for anti-depressant use and hepatitis C. In a multivariable model including factors significant at p < 0.10 in demographics and education-adjusted analyses, older age, female sex, Hispanic/ Background: Chronic pain is common among older HIV-infected (HIV positive) adults and contributes to substance use, reduced physical function and disability. We designed a pilot randomized controlled trial of cognitive behavioural therapy (CBT) + tai chi + motivational texting vs. standard of care to reduce pain, disability and substance use in older HIV-positive adults.
Methods: Evidence-informed chronic pain and substance abuse CBT protocols were adapted and combined with group tai chi and motivational text messaging (EXP). HIV-positive adults ≥50 years of age with chronic pain and substance use (n = 55) were randomized to EXP (n = 18), support group control (cINT, n = 19) or no intervention (noINT, n = 18) for 8 weeks plus a 4-week post-study follow-up. All participants also completed daily diary assessments. Effects were compared within and between groups. Linear regression models assessed factors, including treatment assignment, associated with physical function, pain, substance use and quality of life. Results: Participants had mean age 55 years, 17 years HIV positive, 11 years chronic pain; 84% were non-white; and 76% male and 9% transgender women. Approximately 1/3 each reported alcohol, marijuana and stimulants as their preferred substance. At baseline, all participants had physical and mental health (SF-12) scores below population means, and 87% had reduced physical function (short physical performance battery [SPPB] ≤10). After 8 weeks, only EXP participants had significantly improved SPPB scores (+31%, within-group p < 0.001, between-group p = 0.04) and physical function (60% less reduced physical function) that persisted at the post-study follow-up visit and after controlling for age, education, HIV and baseline pain severity. After 12 weeks, both EXP (−55%, p = 0.04) and cINT (−120%, p = 0.03) participants demonstrated reduced 30-day preferred substance use. However, only EXP participants experienced reduced overall substance use (−51%, p = 0.02) and improved percent pain relief (−76%, p = 0.03).
Conclusions: An eight-week combined CBT, tai chi and motivational text messaging intervention significantly improved physical function in older HIV-positive adults with chronic pain and substance use. After 12 weeks, substance use decreased and achievable pain relief also improved, possibly reflecting delayed onset of improvement in these outcomes. Further study will determine whether CBT + tai chi + motivational texting can provide sustained improvements in this vulnerable population.

MOPDB0105
Capacity to screen and manage mental health disorders at HIV treatment sites in low-and middle-income countries Results: Most sites (n = 40, 75%) were in urban areas. Although 34 sites (64%) reported screening for depression, only 14 (26%) sites had guidelines for screening. Screening for depression and PTSD was more common in low-income countries than middleincome countries. This trend was reversed for SUD screening (Table 1). Depression, PTSD and SUD were managed on site (defined as having services provided at the HIV clinic or the same health facility) in 62%, 43% and 36% of sites, respectively. Selective serotonin reuptake inhibitors (SSRIs) were available in all sites from upper-middle income countries, but only in 22% and 35% of low-income and lower-middle income countries, respectively.
Conclusions: Interim findings suggest that most of the HIV treatment facilities surveyed have integrated some mental health services and that screening for depression is commonly reported in low-income countries. However, on-site management of depression is less common in these settings. Additional research is needed to understand individual, organizational and contextual factors that may influence availability of mental health interventions in LMIC.

MOPDC0101
Methods of gestational age (GA) assessment influence the observed association between ART exposure and preterm delivery (PTD): a prospective study in Cape Town, South Africa In 1037 live-singleton births (mean birthweight 3124 g; 10% SGA <10th centile), PTD risk was doubled for HIV-infected compared to uninfected women for ultrasound-based GA (odds ratio = 2.01, 95% confidence interval = 1.15-3.51); but for LMP/SFH-based GA, the association was not significant (Figure 1). These differences between GA assessment methods persisted after adjustment for age, parity, height and previous PTD; PTD risk did not vary by ART initiation timing for any GA method.
Abstract MOPDB0105- Background: In the era of universal HAART, concerns remain that triple therapy may increase adverse pregnancy outcomes. We compared low birth weight (LBW), preterm birth (PTB) and survival in infants born to ART-experienced women and to uninfected women in Blantyre, Malawi, where first-line ART includes tenofovir, lamivudine and efavirenz. Methods: We enrolled HIV-infected and HIV-uninfected women and their babies at delivery into a one-year prospective study at four health facilities. Eligibility included confirmed HIV status, consent, singleton births, CD4 >350 cells/ml 3 and no stage 3/4 HIV. We documented sociodemographic data, clinical and reproductive history, maternal and infant survival, birth weight and gestational age using Ballard score. We applied logistic regression to measure the association between HIV and LBW and PTB. Odds ratios and 95% confidence intervals (CIs) are presented. Results: We enrolled 459 HIV-uninfected and 335 HIV-infected women on ART from January to December 2016; 1.4% were virally suppressed at baseline (<40 copies per/ml). Rates of LBW among HIV-infected and HIV-uninfected women were 4.2% and 3.4%, respectively; p = 0.69; 11.9% of HIV-infected women and 12.0% of HIV-uninfected women delivered PTB; p = 0.99. In multivariate analyses for LBW and PTB (Table 1), there was no association between HIV status and adverse outcomes. Among all women, having more than one pregnancy (odds ratio (OR) = 0.36, 95% CI (0.18, 0.73)) or more than one birth (OR = 0.39, 95% CI (0.19, 0.82)) was protective against LBW. PTB was 8.9% among women starting ART before or during first trimester; 15.5% among those starting in second trimester; and 17.9% among women starting in third trimester (p = 0.06 for trend). Two infants died: one HIV exposed and one HIV unexposed. No maternal deaths were observed. Conclusions: It is reassuring to observe that adverse outcomes were not different between healthy HIV-infected women and HIV-uninfected women. It appears that near-universal ART can eliminate mother-to-child transmission of HIV without significant negative impact on other pregnancy outcomes. the duration of viral suppression, affecting HIV viral load and potentially HIV transmission from mother to infant.

MOPDC0103
Methods: A case cohort study was conducted using data from the Breastfeeding, Antiretrovirals and Nutrition study to (1) examine correlation between plasma and breastmilk ARV drug concentration, (2) estimate associations between ARV drug concentration and HIV viral load and (3) compare time to breastmilk HIV transmission by plasma drug concentration status. We included mothers randomized to 28 weeks of postpartum maternal ARVs or infant nevirapine who had ≥1 plasma or breastmilk (maternal ARV arm only) specimen available 2-24 weeks postpartum. Among these, we included all mothers who transmitted HIV to their infants between 2 and 28 weeks and 15% of mothers who did not (n = 27 and 227, respectively). Plasma and breastmilk drug concentrations for nevirapine, nelfinavir and lopinavir were dichotomized using the median effective concentration (EC 50 ) as a cutoff (>EC50 vs. ≤EC50). Plasma and breastmilk viral load were dichotomized as detectable (plasma: >40 copies/ml, breastmilk: >56 copies/ml) or undetectable. Spearman correlation coefficients were used to assess correlation between plasma and breastmilk ARV concentration. Associations between drug concentration and viral load cutoffs were assessed using mixed-effects models. Cox models were used to estimate the association between plasma drug concentration and breastmilk HIV transmission between 2 and 28 weeks. Results: All ARV compounds exhibited substantial correlations between maternal plasma and breastmilk concentrations (rho: 0.85-0.98, p-value <0.0001). Having plasma drug concentration above the EC 50 was associated with lower odds of having detectable HIV RNA (plasma odds ratio (OR) 0. Background: Stillbirth (SB) has multifactorial and incompletely understood causes. We aimed to assess the SB rate and associated risk factors in HIV-infected women delivering in UK/Ireland between 2007 and 2015.
Methods: We analysed data from singleton deliveries and defined an SB as a baby delivered at ≥24 gestational weeks (GW) showing no signs of life. We performed multivariable logistic regression of SB risk factors, adjusted for maternal age and country of origin (grouped, see Table 1), year, injecting drug user history, parity, first antenatal CD4 count ≤350 cells/μl, antenatal ART regimen, late antenatal care start (≥13 GW) and newborn gender with random effect for repeated pregnancies in the same mother.
Conclusions: SB rate in HIV-infected women in UK/Ireland was 0.9% in 2007-2015, around twice that in the general population (<0.5%). Further research is needed to understand circumstances around SB in this population in order to identify possible interventions.

MOPDC0105
Usefulness of HBV rapid tests to identify pregnant women at high risk of HBV mother-to-child transmission: the pilot ANRS 12328 study in Cambodia Background: Mother-to-child transmission (MTCT) of HBV is mainly associated with high maternal HBV DNA viral load (VL) levels.
International guidelines recommend the use of antiviral drugs (such as tenofovir (TDF)) during pregnancy if maternal VL is >200,000 IU/mL. In Southeast Asia (SEA), many pregnant women are not screened for hepatitis B surface antigen (HBsAg) and VL testing is not accessible for those who are HBsAg-positive. Here, we report among pregnant women from Cambodia, the performance of a serial algorithm using two HBV rapid diagnostic tests (RDTs), in which samples reactive for HBsAg were further tested for hepatitis B e antigen (HBeAg), as a surrogate marker for HBV replication.
Methods: In 2015, we prospectively collected plasma samples of 250 pregnant women from the Calmette Hospital (Phnom Penh), including women with a known positive HBsAg status. All specimens were initially tested with the SD BIOLINE HBsAg RDT (Standard Diagnostics), and results were compared to the Murex HBsAg ELISA v3.0 (Diasorin) (gold standard). HBeAg status was then blindly assessed among all ELISA HBsAg-positive samples using the SD BIOLINE HBeAg RDT, and results were compared to HBV DNA levels (PUMA HBV kit, Omunis) (gold standard). Analysis was done according to thresholds of 200,000 and 2,000,000 IU/mL.

MOPDD0101
HIV treatment and care services for adolescents: a situational analysis of 218 facilities in 23 sub-Saharan African countries were living with HIV globally. The extent to which health facilities provide appropriate treatment and care was unknown. To support understanding of service availability in 2014, Paediatric-Adolescent Treatment Africa (PATA), an NGO supporting a network of health facilities across sub-Saharan Africa, undertook a facility-level situational analysis of adolescent HIV treatment and care services in 23 countries.
Methods: Two hundred and eighteen facilities, responsible for an estimated 80,072 HIV-infected adolescents in care, were surveyed. Sixty per cent of the sample were from PATA's network, with the remaining gathered via local NGO partners and snowball sampling. Data were analysed using descriptive statistics and coding to describe central tendencies and identify themes.
Results: Respondents represented three sub-regions: West and Central Africa (n = 59; 27%), East Africa (n = 77, 35%) and Southern Africa (n = 82, 38%). Half (50%) of the facilities were in urban areas, 17% peri-urban and 33% rural settings. Insufficient data disaggregation and outcomes monitoring were critical issues. A quarter of facilities did not have a working definition of adolescence. Facilities reported non-adherence as their key challenge in adolescent service provision, but had insufficient protocols for determining and managing poor adherence and loss to follow-up. Adherence counselling focused on implications of non-adherence rather than its drivers. Facilities recommended peer support as an effective adherence and retention intervention, yet not all offered these services. Almost two-thirds reported attending to adolescents with adults and/or children, and half had no transitioning protocols. Of those with transitioning protocols, 21% moved pregnant adolescents into adult services earlier than their peers. There was limited sexual and reproductive health integration, with 63% of facilities offering these services within their HIV programmes and 46% catering to the special needs of HIV-infected pregnant adolescents.
Conclusions: Results indicate that providers are challenged by adolescent adherence and reflect an insufficiently targeted approach for adolescents. Guidance on standard definitions for adherence, retention and counselling approaches is needed. Peer support may create an enabling environment and sensitize personnel. Service delivery gaps should be addressed, with standardized transition and quality counselling. Integrated, comprehensive sexual reproductive health services are needed, with support for pregnant adolescents.

MOPDD0102
Factors associated with viral suppression among adolescents living with HIV in Cambodia selected from 11 ART clinics in the capital city and 10 provinces, utilizing a structured questionnaire. The most recent viral load test result was retrieved from medical records obtained from the ART clinics. Adolescents were categorized as having viral suppression if the viral load count was <1000 copies/ml. Multivariate logistic regression analysis was performed.
Results: This study included 328 adolescents with a mean age of 15.9 years (SD = 0.8); of whom, 48.5% were female. Mean duration on ART was 97 months (SD = 40.2). Proportion of adolescents with viral load suppression was 76.8%. In bivariate analyses, viral suppression was significantly associated with older age, duration on ART, higher CD4 count, better family socio-economic status, living with parents, parental education, having parents as main caregivers, no experience of negative attitude from healthcare providers, being aware that they were receiving ART, knowing that HIV is transmitted through unprotected sex with people living with HIV, understanding that there is no cure for AIDS, receiving treatment from a paediatric clinic and type of ART (first or second line). After adjustment, viral suppression remained significantly associated with longer duration on ART, higher CD4 count, receiving treatment from a paediatric clinic, being aware that they were receiving ART and better HIV-related knowledge including knowing that HIV is transmitted through unprotected sex with people living with HIV and understanding that there is no cure for AIDS.
Conclusions: Viral load suppression rates among adolescents living with HIV in this study were considerably high, but fall short of the global target of 90% viral suppression among people living with HIV on ART. Our findings indicate the need to strengthen treatment literacy and understanding of HIV among adolescents as they prepare for transition from paediatric to adult HIV care.

MOPDD0103
Evaluating the implementation and impact of the adolescent package of care at health facilities in former Nyanza province, Kenya Conclusions: We observed increased FP uptake and VL suppression during scale-up of the APOC. However, increased utilization of the APOC checklist was not associated with these improved outcomes.
The checklist is only one element of APOC. Further investigation of additional elements is needed to assess full implementation and impact of APOC on adolescent outcomes.

MOPDD0104
Economic context and HIV vulnerability in adolescents and young adults living in urban slums in Kenya: a qualitative analysis based on scarcity theory Background: Urban slum adolescents and young adults have disproportionately high rates of HIV compared to rural and non-slum urban youth. Yet, few studies have examined youth's perceptions of the economic drivers of HIV. This study applied principles from traditional and behavioural economics, in particular the theory of scarcity as defined by Mullainathan and Shafir in 2013, in understanding the duality that impoverished youth face in making sexual decisions both in response to direct money shortages and under the cognitive load of financial distress. Methods: Twenty focus group discussions were conducted with 120 adolescents, aged 15-17, and young adults, aged 18-22, living in one of two urban slums in Nairobi, Kenya. Using a semi-structured discussion guide, we asked youth to describe how their economic scarcity, in the form of financial, material, and physical deprivation, contributed to sexual risk behaviours and influenced their capacity to prevent HIV acquisition. All discussions were conducted in Kiswahili and translated and transcribed into English. Data were then coded and analysed using interpretive phenomenology.
Results: Results indicated that slum youth made many sexual decisions considered rational from a traditional economics perspective, such as acquiring more sex when resources were available, maximizing wealth through sex, being price sensitive to costs of condoms or testing services and taking more risks when protected from adverse sexual consequences. Youth's engagement in sexual risk behaviours was also motivated by scarcity phenomena explained by behavioural economics, such as compensating for sex lost during scarce periods (i.e., risk-seeking), valuing economic gains over HIV risks (i.e., tunnelling, bandwidth tax), and transacting sex as an investment strategy (i.e. internal referencing). When scarcity was alleviated, young women additionally described reducing the number of sex partners to account for non-economic preferences (i.e. slack). These findings further revealed two implications for prevention interventions relating to gender-specific economic interests and reduced perceived costs of HIV infection.
Conclusions: Scarcity theory draws attention to the role of financial insecurity in altering how individuals and couples approach sexual decision-making. Combination prevention interventions, including biomedical technologies relying on adherence, should not ignore traditional and behavioural economic drivers of sexual decisions in urban poor settings.

MOPDD0105
The effectiveness and cost-effectiveness of communitybased support for adolescents receiving antiretroviral treatment in South Africa G Fatti 1 ; N Shaikh 1 ; O Oyebanji 1 ; F Chirowa 2 ; D Jackson 3,4 ; A Goga 5,6 ; T Magubu 7 ; JB Nachega 8,9,10  Background: Adolescents receiving antiretroviral treatment (ART) in sub-Saharan Africa are at particular risk of suboptimal adherence, inadequate viral suppression and high loss to follow-up (LTFU). Sub-Saharan Africa also has critical shortages of professional healthcare workers. Community-based support (CBS) programs are task-shifting interventions to address these shortages. Effectiveness and cost-effectiveness data of interventions improving ART outcomes amongst adolescents are very limited. We measured the effectiveness and cost-effectiveness of a large CBS programme in South Africa, the country having the highest HIV burden globally. Methods: A cohort study included ART-naïve adolescents and youth (ages 10-24 years) who initiated ART at 47 facilities. CBS workers conducted regular home visits providing ART-related and adherence education, psychosocial support, symptom screening for opportunistic infections and traced patients defaulting clinic appointments. Outcomes were compared between adolescents who received CBS plus standard clinic-based care versus adolescents who received standard care only. Cumulative incidences of LTFU, mortality, CD4 count increases, viral suppression and pharmacy refill data using the medication possession ratio (MPR) were analysed using multivariable competing-risks regression, generalized estimating equations and multilevel mixed-effects models over five years of treatment. Costs of CBS were compiled and incremental cost-effectiveness ratios (annual cost per additional patient retained in care for patients receiving CBS versus not receiving CBS) were calculated. Background: Men who have sex with men, sex workers and people who inject drugs are considered key populations (KPs) due to their high risk of HIV infection and transmission. Adolescent KPs (less than 19 years) are at higher risk of HIV than their older counterparts yet programmes have not prioritized them. Legal barriers affect their access to sexual reproductive health (SRH) and HIV services. A study was conducted to identify risk perceptions and experiences that heighten risk of HIV among adolescents aged 10-19 years who engage in either sex work, same sex relationships or intravenous drug use and their uptake of SRH services.
Methods: A qualitative exploratory study was conducted between October 2015 and April 2016 in Nairobi, Mombasa and Kisumu Counties. Nine (9) focus group discussions and 18 in-depth interviews were conducted with 37 adolescent girls reporting sex work, 36 adolescent boys in same sex relationships and 35 adolescents involved in intravenous drug use. The participants were purposively selected from various KP community-based organizations. Data were coded thematically and analysed using Nvivo 10. Ethical approval was obtained from AMREF ESRC. Results: 108 adolescent KPs (51 female and 57 male) ranging from 10 to 19 years were interviewed. They reported similar experiences that placed them at heightened HIV risk; being forced to have "condomless" sex for more money, selling sex to sustain themselves and their dependents (22 of the 37 adolescent FSWs had 78 Abstracts of the 9th IAS Conference on HIV Science (IAS 2017) Journal of the International AIDS Society 2017, 20 (Suppl 5) children), sexual abuse and physical abuse from clients and police. Self-perceptions of HIV risk were mixed; those practicing sex work had multiple sexual partners perceived themselves to be at high risk while those having one partner perceived themselves at low risk. All groups preferred accessing SRH services in private unlike public facilities due to stigma and discrimination.
Conclusions: The findings demonstrate the challenges faced by the often forgotten adolescent KPs highlighting the need for adolescent KP friendly policies, services and structural interventions to address poverty, legal barriers to service access, violence, stigma and discrimination. HIV prevention interventions should address knowledge of and self-perception of HIV risk to prevent HIV acquisition and transmission.

TUPDA0101
Combined IL-21 and IFNα treatment limits residual inflammation and delays viral rebound in SIV-infected macaques Background: Although antiretroviral therapy (ART) suppresses HIV replication, immune dysfunctions and chronic inflammation critically contribute to non-AIDS-related morbidity and mortality in infected subjects. Furthermore, inflammation facilitates HIV persistence during ART. We previously demonstrated that addition of Interleukin (IL)-21, an immunomodulatory cytokine, reduces chronic inflammation and HIV persistence in ART-treated, SIVinfected rhesus macaques (RMs). This study sought to combine the anti-inflammatory functions of IL-21 with the antiviral properties of IFNa to reinvigorate antiviral responses. We hypothesize an impact on viral rebound following ART treatment interruption (ATI). Methods: 15 RMs were infected with SIV mac239 IV. RMs started a triple formulation of TDF, FTC and Dolutegravir (DTG) day 35 postinfection and continued for at least 12 months. Eight RMs received Macaquized (M)-IL-21-IgFc (100 mg/kg, SC, once weekly for four weeks) at initiation and mid-way thru ART. Additionally, this group received M-IFNa-IgFc (500,000 IU, SC, once weekly for five weeks) prior to ART interruption. Upon ART discontinuation, the eight IL-21-treated RMs received PEGylated-IFNa-2a (PEGASYS), 7 mg/kg, SC, once weekly for seven weeks; while the remaining seven RMs were ART-treated controls. Blood (PB), lymph nodes (LN) and colorectal (RB) biopsies were longitudinally collected to assess the effects of IL-21 and IFNa on inflammation, T-cell subsets and viral persistence. Results: ART fully suppressed plasma viremia (<30 RNA copies/mL) in all RMs. During ART, IL-21 reduced levels of activated (HLA-DR + CD38 + ) and proliferating (Ki-67 + ) T cells in PB, RB and LN in comparison to ART-only controls (p < 0.01). Levels of inflammation remained significantly lower also during and after addition of IFNa (p < 0.01). Upon ART interruption, IL-21/IFNa-treated RMs exhibited delayed viral rebound with a median of 21 days as compared to 9 days in the controls (p = 0.0009). Moreover, IL-21/IFNa-treated RMs maintained reduced viremia in comparison to controls up to 45 days after ATI (p = 0.0004).
Conclusions: These data support the safety of a combined IL-21 and IFNa treatment for HIV infection. While IL-21-treatment effectively reduces inflammation, addition of IFNa prior-and after-ART discontinuation resulted in a prolonged and more effective control of viral rebound. The synergy of such therapeutics may promote reinvigoration of host responses toward reduction of latent HIV reservoirs.

TUPDA0102
Interleukin ( Background: We have reported that IL-27 inhibits HIV replication in macrophages, T cells and dendritic cells (DC) and proposed that IL-27 is a potent novel anti-viral cytokine. We have also demonstrated that IL-27 induces HIV resistance in macrophages via downregulating SPTBN1 (Spectrin-beta chain brain 1) expression; however, the mechanism whereby IL-27 induces HIV resistance in T cells or DC has not been described.
Methods: T cells and monocytes were isolated from healthy donor PBMCs. T cells were stimulated with PHA. Monocytes were differentiated to DC in the presence of IL-4 and granulocyte macrophage colony-stimulating factor (GM-CSF) with or without IL-27. HIV replication was monitored using a commercial p24 ELISA kit. CD4 and chemokine receptor expression were analysed by FACS. Posttranslational modification (PTM) of proteins were analysed using 2-Dimentional Difference in Gel analysis (2D-DIGE) and Western blotting. Genes of interest were knocked down using siRNA.
Results: IL-27-treated PHA-stimulated T cells (27-T) and IL-27induced DC (27-DC) resisted HIV infection by 70% and >90%, respectively, without significant change in the expression of CD4, CCR5 or CXCR4. To define the mechanism of these anti-HIV effects, microarray and Western blotting were performed. In 27-T cells, four genes were upregulated by >3-fold and no genes were significantly downregulated by >2-fold in microarray analysis of 20,000 genes. 2D-DIGE revealed that the amounts of PTM Y box binding protein 1 (YB1: a Y-box transcription factor) was diminished in 27-T by 60%. Knockdown of YB-1 in control T cells led to 70% resistance to HIV infection. In 27-DC, over 100 genes were upregulated or downregulated by >2-fold. A series of qPCR and Western blot analyses confirmed that the expression of ANKRD22 (ankyrin repeat domain 22) was consistently induced in 27-DC. Knock down of ANKRD22 reversed resistance to HIV infection in 27-DC. These results indicated that IL-27-induced HIV resistance in T cells and DCs by a decrease in PTM-YB1 and an increase in ANKRD22, respectively. Conclusions: PTM-YB-1 and ANKRD22 were identified as novel host dependency factors in T cells and DC, respectively. Taken together with our previous report, these data demonstrate that IL-27 induces HIV resistance in macrophages, T cells and DC in celltype-dependent manners.

TUPDA0103
Increased effector cytotoxic lymphocytes in lymph nodes of hetIL-15 treated macaques suggest potential to disrupt SIV/HIV reservoirs A Valentin 1 ; E Moysi 1 ; DC Watson 1 ; C Petrovas 2 ; C Bergamaschi 1 ; BK Felber 1 and GN Pavlakis 1 1 National Cancer Institute, Frederick, USA. 2 VRC, NIAID, Bethesda, USA Presenting author email: george.pavlakis@nih.gov Background: Heterodimeric interleukin-15 (hetIL-15) activates and expands cytotoxic T and NK cells, which suggests that the cytokine could be useful for the treatment of malignancies and HIV infection. Based on these properties, hetIL-15 is currently evaluated in humans for the treatment of cancer. We also study the effects of hetIL-15 in infected macaques to evaluate its use in HIV infection. Methods: Twelve rhesus macaques received six subcutaneous injections of hetIL-15 over two weeks using increasing doses of cytokine (step-dosing). At the end of the treatment, the animals were sacrificed and the hetIL-15 effects on different lymphocyte populations isolated from tissues collected at necropsy were monitored by multi-parametric flow cytometry and quantitative multiplexed confocal microscopy (Histo-cytometry). Results: This protocol was safe in rhesus macaques and resulted in systemic expansion (Ki67+) of CD8 + T lymphocytes and NK cells with higher granzyme B content. These expanded cell populations were found in both effector sites, such as liver, vagina and rectum, and secondary lymphoid tissues. Among the gut-resident CD8 + T lymphocytes, we found a gradient, based on anatomical location, of the IL-15-induced T-cell proliferation, which follows the proliferation pattern found in untreated animals. Importantly, a significant increase in cytotoxic effector memory CD8 + T cells was found in lymph nodes from all hetIL-15-treated macaques, and this increase was bigger than that in SIV-infected animals. CM9 tetramer staining demonstrated that the increase of CD8+ effector T cells in lymphoid organs included actively proliferating SIV-specific T cells with higher granzyme content. Imaging analysis by Histocytometry revealed that these effector CD8 + T cells infiltrated the B-cell follicles where chronically infected follicular helper CD4 + T cells are located.

Conclusions:
Step-dose administration of hetIL-15 is a well-tolerated regimen that results in systemic activation and expansion of cytotoxic leukocytes that infiltrate areas where chronic HIVinfected cells reside. These results suggest that hetIL-15 could be useful in disrupting or eliminating long-term viral reservoirs in HIV-1-infected individuals, thus contributing to a functional cure of the infection. Work assessing the long-term impact of hetIL-15 on the size of the viral reservoir is currently ongoing.

TUPDA0104
The human IL-15 superagonist ALT-803 activates NK and memory T cells, reactivates latent SIV and drives SIV-specific CD8 + T cells into B-cell follicles Background: There is an urgent need to develop alternate approaches to activate and clear the latent HIV reservoir that do not negatively impact immune function and are independent of viral sequence. IL-15 is a key cytokine for homeostatic maintenance, proliferation and expansion of memory CD4 + T cells, the primary HIV cellular reservoir. Here, we explored the human IL-15 superagonist, ALT-803, as an immunostimulatory molecule and potential latency reversing agent (LRA) in cART-suppressed SIVinfected macaques. Methods: SIV-naïve and SIV-infected macaques were administered ALT-803 IV and subsequently monitored for NK and T-cell proliferation. SIV-infected macaques treated with ALT-803 were assessed for intrafollicular migration via in situ staining of lymph nodes with MHC class I tetramers. ALT-803 was tested as an LRA in vitro with primary CD4 + T cells from cART-suppressed macaques, and in vivo in SIV-infected, cART-suppressed macaques. Results: ALT-803 activated and induced robust proliferation in NK cells and in both effector and central memory T cells. ALT-803 redirected activated cells to secondary lymphoid tissues, a known anatomical location of the viral reservoir, and in situ with MHC class I tetramer staining showed increased migration into B-cell follicular sanctuaries. ALT-803 did not affect viral loads in macaques with uncontrolled SIV infection; instead, ALT-803 potentiated low-level viral replication in elite controllers. In experiments using CD4 + T cell from cART-suppressed macaques, ALT-803 induced robust viral replication in vitro. When administered to macaques with cART suppression of plasma viremia, ALT-803 treatment resulted in plasma viral "blips" and unlike previous reports of other common g-chain cytokines like IL-7, ALT-803 did not cause an increase in the size of the latent viral reservoir. Conclusions: ALT-803, an IL-15 superagonist, is well tolerated in SIV-infected, cART-suppressed macaques and induces virus reactivation both in vitro and in vivo. In addition to reactivating quiescent virus, ALT-803 potently activates NK and memory CD8 + T cells, which traffic to lymph nodes, specifically entering B-cell follicles where latently infected CD4 + T follicular helper cells reside. The ability of ALT-803 to potentially mediate both the "shock" and "kill" make it an appealing candidate for further studies aimed at durable cART-free HIV remission.

TUPDA0105
Tissue factor pathway inhibitor Ixolaris improves disease outcome in progressive SIVsab-infected pigtail macaques HIV-infected subjects face high risk of non-AIDS comorbidities, which is often associated with elevated immune activation and inflammation (IA/INFL) and hypercoagulability. Tissue factor (TF) may bridge IA/INFL and hypercoagulation via protease-activated receptor signalling. We hypothesized that a TF pathway inhibitor can reduce the hypercoagulation and IA/INFL associated with HIV/ SIV infection, and thus improve the disease outcome. Methods: We compared the dynamics of TF expression on monocyte isolated from progressive (pigtail macaques, PTMs) versus non-progressive (African green monkeys, AGMs) models of SIVsab infection to establish its role in SIV pathogenesis. We tested the specificity and potency of a new TF inhibitor, Ixolaris, ex vivo on non-human primate (NHP) plasma and monocytes stimulated with LPS. Ixolaris was administered to five SIVsab-infected PTMs, with three SIV-infected, untreated PTMs as controls. IA/INFL markers (CD38/HLA-DR and proinflammatory cytokines) and hypercoagulation markers (D-dimer) were monitored throughout Ixolaris administration in treated and control PTMs. Results: Monocyte TF expression increased postinfection in PTMs while it remained at baseline levels in chronically infected AGMs, confirming the role of TF in exacerbating SIV pathogenesis. Ixolaris specifically inhibited the extrinsic coagulation pathway and strongly inhibited TF activity by monocytes stimulated with LPS. In vivo administration of Ixolaris resulted in significantly lower inflammation (IL-17) and immune activation (HLA-DR + -and CD38 + -expressing CD4 + and CD8 + T cells) during early chronic infection in treated SIVsab-infected PTMs compared to controls. Ixolaris also reduced hypercoagulation levels (D-dimer) in acutely SIVsab-infected PTMs. While viral loads and CD4 counts were minimally impacted by treatment, Ixolaris improved PTM survival, with no PTMs developing disease during the first 100 days postinfection, which is rarely seen in untreated SIVsab-infected PTMs. Conclusions: Our results support a role of TF pathway in promoting disease progression and cardiovascular comorbidities in SIVinfected NHPs. In vivo TF inhibition by Ixolaris resulted in reduced IA/INFL and hypercoagulation in SIVsab-infected PTMs independent of CD4 counts and plasma viremia and improved the outcome of the SIVsab infection. Therefore, targeting TF pathway in HIVinfected subjects may represent an effective approach to prevent the deleterious consequences of HIV infection.

TUPDA0106
Engineering HIV immunity with a chimeric antigen receptor in the non-human primate model Background: HIV-1-specific cytotoxic T lymphocytes are crucial for the elimination of HIV-infected cells. We have previously demonstrated using humanized mice that hematopoietic stem cells (HSCs) modified with a protective CD4 chimeric antigen receptor (CD4CAR) successfully differentiated into CD4CAR expressing T cells that significantly suppressed HIV replication. These results demonstrated the feasibility of engineering HIV-specific T-cell immunity with a HSC-based approach. Methods: We tested the safety and feasibility of engineering T-cell immunity via HSC in a non-human primate (NHP) model of SHIV infection. We utilized CD4CAR vectors that also carry an anti-HIV protective gene (C46) that would inhibit infection. Two pigtailed macaques (Macaca nemestrina) were transplanted with C46CD4CAR-modified autologous HSC, and two were transplanted with HSC modified with control vector C46CD4CARdeltaZeta that lacks the signalling Zeta chain. After hematopoietic recovery, the animals were challenged with SHIV and went through combined anti-retroviral drug treatment and withdrawal. Animals were monitored for viral load, CAR cell detection and immune function for over a year. Results: We determined that engraftment of pigtailed macaques with C46CD4CAR-modified HSCs was safe and the animals had normal transplant recovery. We observed long-term engraftment and stable production of C46CD4CAR expressing cells without any significant toxicities and found that C46CD4CAR-modified HSCs could differentiate into multiple hematopoietic lineages. Following challenge of the animals with SHIV, we observed significant expansion of C46CD4CAR expressing cells after infection and wide distribution of CAR-expressing cells in multiple lymphoid tissues. Importantly, we found lower viral loads in lymphoid tissues in C46CD4CAR-containing animals than in control animals, suggesting viral suppression by C46CD4CAR-expressing cells.

TUPDB0102
Impact of T-cell homeostasis and thymic output on the seeding of the HIV reservoir in infants M Massanella 1,2 ; J Ananworanich 3,4 ; L Leyre 1,2 ; T Jupimai 5 ; P Sawangsinth 5 ; M de Souza 3 ; P Suntarattiwong 6 ; P Kosalarksa 7 ; T Puthanakit 8,9 ; N Chomont 1,2 ; HIVNAT209 and HIVNAT194 study groups 1  Background: Early antiretroviral therapy (ART) limits the size of the HIV reservoir and immune activation levels in adults; however paediatric data are limited. We evaluated the impact of immune parameters on the size of the HIV reservoir in early treated vertically infected infants. Methods: Virologically non-suppressed HIV-infected Thai infants (n = 12, <2 months old) and fully suppressed children (for >1 year) who started ART within the first six months of life (n = 57, median 4 years old) were included. We quantified total HIV DNA in CD4 T cells by RT-PCR. Immune activation and proliferation markers in CD4 and CD8 naïve, stem cell, central, transitional and effector memory T cells as well as frequencies of effector, recent thymic emigrants (RTE) and T-regulatory (Treg) cells were assessed by flow cytometry. Results: High frequencies of naïve CD4 + T cells were observed in non-suppressed and ART-treated children (median 78 and 64, respectively). Interestingly, the frequency of RTE was significantly increased in ART-treated children despite their older age (median 81 vs. 89, p = 0.009), suggesting that HIV replication in nonsuppressed infants may limit thymic production. Frequencies of T cells undergoing proliferation (Ki67 + ) were significantly lower in CD4 + (median 4.8 vs. 3.6, p = 0.04) and CD8 + (median 29 vs. 4.6, p < 0.0001) T cells from ART-treated children. Similarly, frequencies of activated CD8 + T cells (HLA-DR + CD38 + ) were significantly lower in ART-treated children (median 29 vs. 4.6, p < 0.0001). High frequencies of Treg were observed in both groups (>5.1% of CD4 + T cells) with no significant differences. In non-suppressed infants, the frequency of cells harbouring HIV DNA was negatively correlated with the frequency of RTE (r = −0.9, p = 0.01) and positively correlated with the frequency of memory Treg cells (CD45RA − CD27 − CD25 bright FOXP3 + , r = 0.7, p = 0.04). Similarly, HIV DNA levels in ART-treated children negatively correlated with RTE (r = −0.3, p = 0.02) and positively correlated with Treg (r = 0.3, p = 0.04). Conclusions: The high frequencies of RTE resulting from enhanced thymic production in paediatric population may limit the establishment and persistence of the HIV reservoir. In contrast, the positive association between the frequency of HIV-infected cells and Treg suggests that this subset may play a prominent role in the establishment of the HIV latent reservoir during infancy. Background: It is unknown if extremely early initiation of ART may be curative. We describe an individual who started ART an estimated 10 days after infection with plasma HIV RNA of 220 copies/ ml. After extensive blood and tissue sampling, he underwent an analytical treatment interruption (ATI) following 34 months of ART. Methods: Colorectal and lymph node tissues, bone marrow, cerebral spinal fluid (CSF), plasma and large numbers of PBMC obtained longitudinally were studied for HIV persistence in several laboratories using molecular and culture-based detection methods including a humanized mouse outgrowth assay. Results: The individual initiated PrEP (TDF/FTC) during very early Fiebig stage I (HIV-1 RNA+ EIA-) followed by ART intensification (TDF/FTC/r/DRV/RAL) 8 days later. HIV RNA levels were 120 and <40 copies/mL, 7 and 22 days after PrEP initiation, respectively, followed by no detectable level. Low-level cell-associated HIV RNA (3.2 copies/million CD4+ T cells) was detected 32 days after infection. Over the following two years, no further HIV could be detected, despite massive sampling from ileum, rectum, lymph nodes, bone marrow, CSF, CD4 + T-cell subsets and plasma. 530 million CD4 + T cells were also assayed in a humanized mouse outgrowth assay. One mouse (53 million input cells) experienced very low-level viremia (201 copies/mL) after 5.5 weeks, but sequence confirmation was unsuccessful. The participant stopped ART and remained aviremic for 7.4 months, rebounding with HIV RNA of 36 copies/mL that rose to 59,805 copies/mL 6 days later. ART was restarted promptly. Rebound plasma HIV sequences were identical to those obtained during acute infection by single-genome sequencing. Mathematical modelling predicted that the latent reservoir size was approximately 200 cells prior to ATI and that only around 1% of individuals with a similar HIV burden may achieve lifelong ART-free remission. Conclusions: We report HIV relapse despite initiation of ART at one of the earliest stages of acute HIV infection possible. Near complete or complete loss of detectable HIV in blood and tissues did not lead to indefinite ART-free HIV remission. However, the small numbers of latently infected cells in individuals treated during hyperacute infection may be associated with prolonged ART-free remission.

TUPDB0104
HIV RNA persists in rectal tissue despite rapid virologic suppression in blood plasma with dolutegravir-based combination antiretroviral therapy in treatment-naïve patients Background: Despite virologic suppression in blood plasma (BP) with combination antiretroviral therapy (cART), HIV eradication remains elusive, largely due to incomplete suppression in reservoir sites including gut-associated lymphoid tissue. We compared HIV RNA dynamics within BP and rectal tissue (RT) following initiation of cART with integrase strand transfer inhibitor dolutegravir (DTG). Methods: Treatment-naïve HIV-positive volunteers began DTG-based (50 mg daily) cART and serial sampling of BP and RT through 84 days as per Figure 1. HIV RNA and DTG concentrations were quantified using Abbott Real-Time HIV-1 assays and tandem mass spectroscopy, respectively, in both compartments. Median DTG concentrations were compared between undetectable and detectable RT HIV RNA groups using Mann-Witney non-parametric tests. Results: Eight participants were enrolled: 4 (50%) females, 6 (75%) black, median age 39.2 years (IQR 32. 9-52.7). Median baseline CD4 and HIV RNA were 208 cells/mm 3 (IQR 104-320) and 24,847 copies/ mL (IQR 6237-50,269), respectively. All participants (100%) had undetectable BP HIV RNA (<40 copies/mL) by Day 42; only three achieved RT viral suppression at any time ( Figure 1). DTG quantitation was performed on 22 paired BP and RT steady-state samples collected Days 7 through 84. The undetectable RT RNA group had higher median DTG RT concentrations than those with detectable RNA: 1340 ng/g (IQR 683-2100) versus 624 ng/g (IQR 377-939), p = 0.03. There were no significant differences in BP DTG or RT:BP DTG ratios between undetectable and detectable groups: median 2810 ng/mL (IQR 1720-4140) versus 1400 ng/mL (IQR 687-3250) and median   Background: PrEP can dramatically reduce HIV acquisition risks, particularly for transgender women and other sexual and gender minorities. However, access to PrEP in developing countries remains very limited. Brazil has one of the largest and oldest HIV treatment programmes in the world and will soon integrate PrEP in the national public health system. We explored PrEP willingness among transgender women. Methods: We recruited 127 transgender women using Respondent Driven Sampling in Salvador, Brazil's third largestand one of its lowest-income and highest afro-descendants-cities. Participants were interviewed about PrEP. Latent class analyses were used to identify those willing to take PrEP from the following list of variables: 1disposition to use PrEP; 2willingness to use PrEP if available in the public health system; 3willingness to use PrEP even if had to pay; 4interest in using PrEP even if it is not 100% effective; 5being less afraid of contracting HIV if using PrEP; 6willingness to take one pill a day; 7knowledge about PrEP. Entropy was 0.992 indicating good distinction of classes. Results: Only 18.4% of women knew about PrEP. However, upon becoming aware, willingness to use PrEP was reassuring. Two latent classes were identified: Class 1-willingness to use PrEP (91.3%); Class 2-non willingness to use PrEP (8.6%). Most participants noted that PrEP was an important HIV prevention tool for transgender women. Correlates of Latent Class 2 were: Sociobehavioural factors including not being black, having a monthly income greater than R$ 900.00 and regular use of condoms in sexual intercourse. When asked about implementation strategies, participants suggested integrating conversations about PrEP into a suite of HIV prevention services, addressing health system services that address the broader social and structural factors influencing transgender health risks. Conclusions: PrEP willingness was very high as 91% of transgender women wanted it in Northeast Brazil. While access to PrEP is still limited, uptake among transgender women will likely be high when Brazil releases PrEP. However, it is important to take into account socio-behavioural factors and combination prevention as those regular users of condom may not see and additional benefit of PrEP.

TUPDC0102
How are transwomen acquiring HIV? Insights from phylogenetic transmission clusters surveillance systems and surveys, transwomen and their male sexual partners are classified as "men who have sex with men" (MSM), irrespective of gender identity and sexual orientation. Little is known about how transwomen acquire HIV, which may be due in part to this misclassification as "MSM". We sought insights on sexual and use of non-sterile needle networks as potential sources of HIV among transwomen by examining phylogenetic transmission clusters. We also assessed a new transmission risk paradigm that re-classifies males closely linked to transwomen as non-MSM. Methods: San Francisco residents diagnosed with HIV (2000HIV ( -2015, in care at public facilities and with available viral pol sequences were included in the analysis. Transmission clusters with ≥2 cases were identified by bootstrap values ≥90% and mean pairwise genetic distances ≤0.025%. Results: Transwomen were 275 of 5200 cases with viral sequences; 86 transwomen were in 70 clusters; 44 (51%) had injection risk. Many clusters with transwomen contained MSM-persons who inject drugs (MSM-PWID) (47% of clusters) and non-MSM PWID (26%); whereas MSM were in 54% of clusters and heterosexual men in 1%. After re-classification, the profile of clusters shifted: 16% of clusters contained MSM-PWID, 57% had non-MSM PWID, 16% had MSM and 47% had heterosexual men. Similarly, among 130 clusters containing cis women, 27% had MSM-PWID, 41% had non-MSM PWID, 28% had MSM and 58% had heterosexual men. Conclusions: Applying the new paradigm for classifying the transmission risk of transwomen and their partners suggests that transwomen may stand apart from the MSM epidemic. The risk profile of transwomen's transmission clusters is highly sensitive to whether or not male partners are classified as MSM. Under the new paradigm, transmission clusters containing transwomen closely resemble clusters containing cis women, with a strong presence of PWID and heterosexual men. There is increasing recognition that transwomen should be considered by their gender identity for health services and research. The same consideration perhaps should apply to male sexual partners of transwomen. Examining transmission clusters may bring new insights to the applicability of MSM-focused research to transwomen and their partners.

TUPDC0103
Discrepancy between risky sexual behaviour and perceived HIV risk among transgender women in community-based test and treat cohorts in Thailand AR Ramautarsing 1 ; T Nakpor 2 ; R Janamuaysook 1 ; S Pengnonyang 1 ; J Jantarapakde 1 ; D Trachunthong 1 ; T Sungsing 1 ; P Rodbamrung 1 ; P Mingkwanrungruangkit 1 ; W Sirisakyot 1 ; S Tongmuang 1 ; S Jarupittaya 1 ; N Mahachokchai 2 ; P Yokchawee 2 ; S Samalu 3 ; M Uthaikorn 3 ; W Champa 4 ; P Patpeerapong 5 ; S Mills 6 ; S Chareonying 6 ; P Phanuphak 7 ; M Cassell 8  Background: Transgender women (TG) are a high-burden population for HIV. Globally, their risk of HIV is 49 times higher compared to the general population, indicating that HIV prevention and care services for TG are critical and urgent. To understand the facilitators and barriers for TG to access HIV testing and antiretroviral therapy (ART) at community-based clinics, we studied socio-demographic characteristics, behavioural risk and knowledge and attitudes towards HIV and ART. Methods: Thai TG aged ≥18 years were recruited from six community-based clinics in Thailand. Demographic characteristics, behavioural risk, HIV knowledge and attitudes towards ART were assessed using questionnaires. Trained community health-workers provided same-day result HIV testing and sexually transmitted infection (STI) screening, as well as CD4 testing and linkage to care for HIV-infected participants. Results: Of 734 TG participants enrolled, 56.1% were between 18 and 25 years old. Only 15.7% had a college degree or higher, and 36.8% earned less than 280 USD monthly. Prevalence of any STI was 32% (syphilis 3.5%, chlamydia 23.3%, gonorrhoea 14.6%). Nobody (0.0%) reported a negative attitude towards people with HIV, 42.0% of participants knew that ART can reduce HIV transmission, and 65.4% thought all people with HIV should use ART. Almost 90% said they would start ART immediately if they were diagnosed with HIV. Multiple sexual partners in the last six months was reported by 53.5%. While 77.4% of TG reported unprotected sex in the last six months and 91.1% identified proper condom use as a manner of decreasing HIV-risk, only 17% perceived their HIV risk as high. Almost half (45.0%) had never tested for HIV. HIV prevalence was 9.0%, and among 66 HIV-infected participants, median (IQR) CD4 count was 406 (306-562), 84% initiated ART within two weeks of diagnosis. Conclusions: Among these young Thai TG presenting for HIV-testing at community-based clinics, the prevalence of HIV and STIs was high. Attitudes towards HIV and ART were positive, and ART uptake was high. However, there was a remarkable discrepancy between risky sexual behaviour and perceived HIV risk. Our findings are crucial to informing HIV education and prevention programs targeting TG in Thailand.

TUPDC0104
Engaging transgender women of colour in HIV prevention: findings from mixed methods community-based research T Poteat 1 ; M Malik 1 ; A Wirtz 1 and E Cooney 2 1 Johns Hopkins School of Public Health, Epidemiology, Baltimore, USA. 2 Johns Hopkins Bloomberg School of Public Health, International Health, Baltimore, USA Presenting author email: tpoteat123@gmail.com Background: Transgender women (TW) across the globe are highly vulnerable to HIV. TW of colour (TWOC) in the United States are particularly burdened, with an estimated HIV prevalence over 50% in some studies. Effective HIV interventions are needed to prevent HIV acquisition and onward transmission and to improve health outcomes. We used mixed-methods approaches to understand how to develop and implement accessible, acceptable, and effective HIV interventions for TWOC. Methods: Qualitative, in-depth key informant (KI) interviews (n = 20) were conducted with TW and health and social service providers. Interviewer-administered surveys and rapid HIV tests were implemented among TWOC (n = 46) in Baltimore, USA. Results: Among TWOC, HIV prevalence was 48%, with 18% unaware of their HIV status and 33% who had not been tested for HIV in the last 12 months. History of sex work (78%) and condomless anal intercourse (47%) were high. Most participants had heard of PrEP (89%); of those, 78% knew where to get it and 23% had ever taken it. Of those who had ever taken PrEP (n = 9), 67% had done so in the prior 12 months. Despite low uptake, 81% of HIV-negative TWOC stated they would take PrEP if it was made available to them. Interviews elicited diverse reflections on how to better engage TWOC in HIV programmes. KIs emphasized the importance of embedding HIV services within social service programmes that are responsive to community needs (e.g., job readiness, mental health support, housing), to improve access and acceptability of HIV programmes for TWOC. KIs also recommended regular staff training in transgender competent care (e.g., using correct gender pronouns), recognizing that experiencing discrimination in health venues deters TWOC from future care seeking. Other suggestions included: offering services where TWOC regularly visit, hiring TWOC to lead programs, and doing tailored outreach and advertising. Conclusions: Acceptable high-impact HIV prevention and care interventions for TWOC are urgently needed as evidenced by an elevated HIV prevalence and low PrEP uptake. Noting the disconnect between willingness and uptake of PrEP among TWOC, HIV prevention programmes could better bridge this gap by incorporating strategies voiced by TW and responding to identified access barriers.

TUPDC0105
Factors associated with HIV infection among transgender women in Cambodia: results from a national integrated biological and behavioural survey S Yi 1,2 ; S Chhim 3 ; P Chhoun 1 ; S Tuot 1 ; P Mun 4 ; K Pal 1 and C Ngin 1  Background: Transgender women are at high risk for HIV infection, and little is known about the burden of HIV infection and its related factors in this population worldwide. This study was conducted to examine factors associated with HIV infection among transgender women in Cambodia. Methods: This cross-sectional study was conducted between December 2015 and February 2016 in the capital city of Phnom Penh and 12 HIV high-burden provinces. Respondent driven sampling was used to recruit sexually active transgender women aged 18 and over. A structured questionnaire was used for a behavioural survey, and rapid finger-prick HIV testing was performed using Determine™ antibody test. Multivariate logistic regression analysis was conducted to identify factors associated with HIV infection using STATA. Results: A total of 1375 transgender women participated in the study with a mean age of 25.9 years (SD = 7.1). The overall HIV prevalence among this population was 5.9%. In multivariate logistic regression, participants living in urban areas were twice as likely to be HIV infected as those living in rural areas. Participants with primary education were 1.7 times as likely to be infected compared to those with high school education. HIV infection increased with age; compared to those aged 18-24, the odds of being HIV infected were twice among transgender women aged 25-34 and 2.8 times higher among those aged ≥35. Self-injection of gender affirming hormones was associated with a fourfold increase in the odds of HIV infection. A history of genital sores over the previous 12 months increased the odds of HIV infection by threefold. Transgender women with stronger feminine identity dressing up as a woman all the time were twice as likely to be HIV infected compared to those who did not dress up as a woman all the time. Having never used online services developed for transgender women was also associated with higher odds of being HIV infected. Conclusions: Transgender women in Cambodia are at high risk of HIV. To achieve the goal of eliminating HIV in the country, effective combination prevention strategies focusing on the above risk factors among transgender women are urgently needed.

TUPDC0106
Prevalence and correlates of police harassment among transgender women in Jamaica: implications for HIV prevention and care CH Logie 1 ; A Lacombe-Duncan 1 ; K Levermore 2 ; N Jones 2 ; A Neil 2 ; T Ellis 2 ; Y Wang 1 and PA Newman 1 1 University of Toronto, Toronto, Canada. 2 Jamaica AIDS Support for Life, Kingston, Jamaica Presenting author email: carmen.logie@utoronto.ca Background: Criminalization of homosexuality limits access to HIV prevention and care. Little is known about police harassment targeting transgender women in contexts where homosexuality is criminalized, such as Jamaica. We examined prevalence of police harassment and its association with HIV infection and risk factors among transgender women in Jamaica. Methods: We implemented a cross-sectional survey with transgender women in Kingston, Ocho Rios and surrounding areas in Jamaica using chain referral sampling. Unadjusted and adjusted logistic regression analyses were conducted to identify health (e.g. HIV status), intrapersonal (e.g. sex work), interpersonal (e.g. social support) and structural (e.g. transphobic violence) factors associated with ever experiencing police harassment perceived to be due to transgender identity. Results: Participant (n = 137) mean age was 24.0 years (SD: 4.5; range 18-44); two-thirds (n = 92; 67.2%) lived in Kingston, and half (n = 71; 51.8%) reported sex work involvement. Three-quarters (n = 103; 75.7%) had received an HIV test; of these, onequarter (n = 26; 25.2%) were HIV positive. Almost half (43.8%; n = 60) reported experiencing police harassment due to their transgender identity. Of participants with complete data (n = 127), 16.5% (n = 21) reported a history of incarceration due to transgender identity. Of those, 71.4% (n = 15) reported being incarcerated one to three times and 28.6% (n = 6) were incarcerated four to six times. In unadjusted analyses, police harassment was associated with sociodemographic (<high school education vs. ≥high school [OR: 3.04, 95% CI: 1. Background: Despite a 30% decline in HIV-related deaths over the past decade, HIV remains still the leading cause of death among African adolescents. Our objective was to summarize performance along the continuum of HIV care and identify predictors of loss to follow-up (LTFU) among youth enrolled in HIV care in rural Mozambique. Methods: We analysed a retrospective cohort of youth (15-24 years) accessing care at one of 89 PEPFAR-supported clinics in Mozambique between 2012 and 2015. We determined the proportion of patients pre-antiretroviral therapy (ART) LTFU at 6 months, cumulative incidence of post-ART LTFU and mortality one-year postinitiation. We defined LTFU as being >60 days late for the last scheduled visit (pre-ART), or ART pick-up (post-ART). We used logistic and multivariable Cox regression models to identify predictors of pre-and post-ART LTFU, respectively, in the year after enrolment. Results: Of 23,322 patients in the cohort, 19,287 (83%) were female. Primary referral source was prevention of mother-to-child transmission clinics for females (49%, n = 8639), and voluntary counselling and testing sites for males (65%, n = 2314). Females enrolled in care at earlier HIV disease stage (median CD4 469 vs. 363 cells/mm 3 , p < 0.001) and initiated ART more expeditiously than males (median  days, p < 0.001). Pre-ART LTFU at 1 year was 36% overall, and lower for females vs. males (33% vs. 56%, OR = 0.28; 95%CI:0.24-0.33). The cumulative incidence of post-ART LTFU was 36% overall (95% CI:35-36%), with small differences by sex (F:M 35% vs. 37%, aHR 0.66 95%CI:0.58-0.75). Adjusted one-year mortality rate for ART patients retained in care was 12.6% (95%CI:6-22%). Conclusions: In the cohort of youth enrolled in care four-fifths were female. Females were enrolled in care earlier in their disease course, initiated on ART more quickly, and less likely to experience pre-ART LTFU than young males. After ART initiation, one-third of all patients were LTFU, and mortality rates were high. While outcomes were poor overall in this setting, young women may require enhanced prevention efforts, while young men need targeted testing and treatment interventions.

TUPDD0102
Gender age considerations and likelihood of viral load suppression at sub-national level in five counties, Kenya S Muga 1 ; J Onyalo 1 ; M Obuya 1 ; P Njoka 1 ; R Kithuka 1 ; C Komen 1 ; L Muyumbu 1 ; G Obanyi 1 ; E Ashiono 1 ; C Muturi 1 ; R Odhiambo 1 and P Mwarogo 2 1 FHI360, USAID-APHIAPlus Nuru ya Bonde, Nakuru, Kenya. 2 FHI360, Country Office, Nairobi, Kenya Presenting author email: smuga@fhi360.org Background: Sub-Saharan Africa accounts for 66% of new HIV infections globally. Program responses need to focus on subnational contexts if epidemic control is to be achieved. The objective of this paper is to describe the likelihood of viral suppression at sub-national level, considering age and gender in five counties in Kenya. Methods: This retrospective correlational survey reviewed the significance of age and gender on the likelihood of viral load suppression among clients on antiretroviral therapy (ART) who had accessed at least one viral load test in the period October 2015 to September 2016. ART was initiated as per the national guidelines in Kenya. Raw viral load data was assessed from the national viral load database for 139 facilities in five counties (Baringo, Kajiado, Laikipia, Nakuru and Narok) and 38,113 records were analysed. The survey used logistic regression to assess relationships between gender, age and viral load suppression (<1000 copies/ml), with p < 0.05. Results: Overall, females were 16% more like to be virally suppressed compared to males. Females aged 1-9 years,10-14 years, 20-24 years were (62%, p = 0.00), (26%, p = 0.024), (79%, p = 0.001) respectively more likely to be virally suppressed than males. There was no significant difference in the likelihood of viral suppression between females and males for clients under one-year-old, (p = 0.230), 15-19 years (p = 0.254), and in 25 years or older (p = 0.079). In Laikipia, males aged 10-14 years were 25% less likely to be virally suppressed. In Nakuru and Narok, males aged between one and nine years were 53% less likely to be virally suppressed. Further, in Nakuru those aged 20-24 years were 48% less likely to virally suppressed compared to female counterparts. In Kajiado, Baringo and Laikipia counties, the likelihood of clients being virally suppressed increased with age. While in Narok and Nakuru counties, it was not dependent on increase in age. Conclusions: The likelihood of achieving viral suppression seemed to agree with literature that female clients are likely to be suppressed. However, there are sub-national differences in the suppression rates. There are also differences in suppression rates through age groups in the counties. This information can be used for further research and effective programming.

TUPDD0103
An exploratory study to determine the survival period to switching for adult ART patients (15+ years) in Swaziland using routinely collected data Background: It has been observed over the past five years that there is limited use of data stored in electronic data systems for making targeted programmatic decisions and conducting operational research. To demonstrate how routine data stored in these electronic systems can be used to inform HIV programming, SIAPS worked with the Ministry of Health (MoH) to conduct a study to determine the survival period to switching for adult patients (15+ years) on antiretroviral therapy (ART). Methods: This was a retrospective cohort design study. The study population consisted of 117,586 adult (≥15 years) ART patient records identified as active between years 2010 to 2015 in the national ART database. The survival rates from time of ART initiation to time of regimen switch were evaluated according to gender and age using Kaplan-Meier models, that is, outcome variable was switching (binary event) and explanatory variables were WHO staging I; II; III; IV, Age categories: 15-24; 25-34; 35-44; and 45 years and above, and Sex: Male; Female. Results: 3.6% (n = 4266) of the studied ART patients were found to have switched at some point during the course of treatment. Median survival time for all ART patients switching from first-to second-line regimen was found to be 607 days (95%CI: 601-613) days or 19 months. Patients aged 45 years and older at ART initiation remained on the first-line regimen for longer periods than other age groups at 93.5%. Survival times for males were less than those of females. Only 83% of patients initiated at WHO stage IV remained on first-line regimen by end of a five-year followup period. Conclusions: There were significant differences in survival periods with men exhibiting a poorer survival period. Also, for the 18-25year age group, we found that women actively switch regimen more often than men. Further studies would be required to understand the factors contributing to these findings. This can inform policies in HIV programming that target the unique needs of males and female clients.

TUPDD0104
Tanzanian men more successful than women in referring sexual partners to HIV testing via partner notification Background: Partner notification (PN) is effective at increasing uptake of HIV testing services (HTS) and identifying previously undiagnosed individuals. 2016 Guidance from WHO recommends inclusion of PN into HTS programmes. PN and referral to HTS can be conducted by the index client (passive notification) or with provider-assisted strategies. Passive PN involves HIV status disclosure and is impacted by socio-cultural dynamics surrounding sex and gender. This abstract describes the success of male versus female index clients in listing and referring sexual partners for testing in Tanzania. Methods: A cross-sectional observational study was conducted in three hospitals in Njombe region, Tanzania, from June to September 2015. Individuals newly diagnosed with HIV in VCT or PITC were offered their choice of passive or provider-assisted referral for partner HIV testing. Odds ratios were calculated for likelihood of male and female index clients to successfully refer partner(s) to HTS or list multiple partners, and in-depth interviews were conducted with 40 index clients and partners. Results: Almost all (91.6%) of the 390 (183 males and 207 females) index clients chose passive rather than provider-assisted referral. Of the 439 listed partners, 249 (56.7%) were successfully referred to HTS (63.4% of female partners; 49.8% of male partners). Male index clients were 2.2 (1.4-3.5) (p < 0.001) times more likely than female index clients to successfully refer at least one partner, and 6.2 (2.7-14.1) (p < 0.001) times more likely to list more than one partner. In qualitative analysis, both gender-specific (feeling undervalued if the relationship had not produced children; lacking contact info for commercial sex partners) and non-gender-specific (difficulty communicating with past partners) challenges were described. Conclusions: PN has been shown to be effective and is being scaled up in multiple African countries, but little discussion has addressed gender and PN. In this study, men were more likely to list multiple partners and to successfully refer at least one partner to HTS. This strength could be built upon in programmatic approaches which target men for PNS. Women's more limited ability to successfully refer their partner(s) could signal a need for different approaches to support women in the PN process.

TUPDD0105
Male engagement strategies effective in improving Option B+ retention in rural Mozambique Background: Retention in antiretroviral therapy services initiated during antenatal care (ANC) in Mozambique is less than 30% at one year. To nurture supportive prevention of mother-to-child transmission (PMTCT) services we introduced a Male Engagement Strategy (MES) that involved partnering with Traditional Birth Attendants and training a new cadre of maleto-male community health agents, "Male Champions." Together they counselled expectant couples to change community norms around male engagement in spousal/partner pregnancies and uptake of HIV services. Methods: We conducted a retrospective analysis of women (>15 years) enrolled in HIV care and treatment through PMTCT services at 112 sites in rural Zambézia province, Mozambique. We compared clinical retention rates among sites receiving MES versus. those receiving standard of care (SOC) using chi-squared tests, Wilcoxon rank sum tests, and Cox regression models. In addition, we assessed the effect of MES on retention by implementation time using a Cox regression model. Results: Six thousand five hundred women were enrolled in PMTCT care at MoH-run clinics receiving Friends in Global Health technical support from January 2015 to November 2016. Median age was 24 years (IQR: 20-29), 84% were married or living with a partner, median CD4 cell count was 463 cells/mm 3 , and 51% were enrolled in sites supported by MES. Cumulative incidence of ART loss to follow-up (LTFU) at six months was 38.1% (36.4%, 39.8%) among those enrolled at MES sites vs. 43.3% among those who received SOC (p = 0.001). Controlling for clinical (e.g. CD4 cell count) and social (e.g. education, marital status) characteristics, those who attended MES clinics had a 33% lower risk of being LTFU at six months versus those receiving SOC (p < 0.001). Longer duration of MES exposure at a clinic was associated with increased retention: covariate-adjusted hazard ratios for late ART pick-up decreased from 0.75 (0.65, 0.86) at 12 months to 0.47 (0.38, 0.58 Background: Swaziland's HIV prevalence remains the world's highest at 26%. However, Swazi men often report fear and suspicion towards VMMC, mostly caused by solely biomedical approaches. In response, KI's innovative project, Lihawu Male Mentoring Camp (LMMC), offers 15-29-year-old men a comprehensive package of adolescent male mentoring and health services, sensitization and interventions, with the following objectives: • To increase number of circumcisions in the pivot age.
• To create a conducive space for VMMC clients to engage with mentors and peers in a fun way, clarifying myths and misconceptions about male health issues, masculinity and MC.
• To increase adherence to post-operative MC complementary care, conduct and lifestyle choices.
Methods: LMMC is a three-day residential camp of activities aimed at age pivot adolescents, combining behaviour change tenants of traditional Bantu initiation rites of passage with clinical best practice in VMMC. Activities include challenges, games and cultural observances as well as sensitization on masculinity and gender awareness, goal setting, HIV and male health issues and services. At the end of the camp, participants are offered a comprehensive package of male health services including HTC and VMMC.
Results: Increased VMMC (86%) and HTC (87%) uptake amongst pivot-age clients (national av. 19%) exposing them to a package of interventions and life skills marking a transition from childhood to adulthood. Pre-and post-camp surveys show a dramatic increase in gender equitable beliefs and acceptance of gender deviance, as well as increases in post-circumcision care and conduct, and in condom usage and efficacy knowledge. Conclusions: Contrary to common beliefs, traditional approaches may be utilized as effective methods to sensitize men on positive masculinity, SRH issues, and the benefits of VMMC, as well as dramatically increase uptake in prevention services.

WEPDB0101
Immediate vs. delayed oral etoposide (ET) among HIVinfected individuals with limited-stage KS initiating ART: A5264/AMC-067 study incidence and increased time to KS progression with immediate ET, but no effect on mortality or need for additional, non-ET chemotherapy.

WEPDB0102
Implementing CRAG screening in HIV patients initiating ART in rural HIV clinics with regular absence of CD4 testing services in rural Tanzania Background: The World Health Organization (WHO) recommends screening for cryptococcal antigen (CRAG) in blood of HIV-infected antiretroviral therapy (ART)-naïve patients with CD4 <100 cells/μL. CRAG+ persons who receive ART but not antifungal therapy are at a high risk of death. However, absence of reliable or prompt CD4 testing services in rural clinics jeopardizes implementation of CRAG screening. Methods: We implemented CRAG screening in all primary health HIV clinics in the Kilombero district, southern Tanzania. Point-ofcare CRAG lateral flow assay testing was recommended for all ARTnaïve HIV-infected persons with criteria for ART initiation or with headache for >5 days. All CRAG+ persons were transported to the Referral Hospital in Ifakara for a meningitis diagnostic workup and antifungal therapy. Patient transport costs, antifungals, and incentives to clinicians were provided. Results: From November 2015 to November 2016, 723 ART-naïve patients were tested for CRAG in 8 HIV clinics. Of these, 45 (6.2%) were CRAG+, and 26 (58%) were diagnosed at peripheral clinics and referred to Referral Hospital for evaluation. The median age of the CRAG+ patients was 35 years (interquartile range [IQR], 21-55), and 60% (27/45) were women. Lumbar punctures were performed in 41 consenting (91%) patients, and 51% (21/41) of patients were CRAG+ in cerebrospinal fluid (CSF). Among these 21 CSF CRAG+ persons, 3 were asymptomatic (7% of overall CRAG+ persons). Conclusions: Our CRAG screening algorithm tailored for rural HIV clinics was effective in maximizing cryptococcal detection in advanced HIV patients at a district level in the absence of regular CD4 testing. The high CRAG prevalence found highlights the importance in the absence of CD4 testing of extending CRAG screening to all HIV-infected persons enrolling in care in order to reduce early mortality.

WEPDB0103
High mortality despite high-dose oral fluconazole (1600 mg) and flucytosine, and serial lumbar punctures, for HIV-associated cryptococcal meningitis: ANRS 12257 study in Burundi and Ivory Coast Background: Mortality from HIV-associated cryptococcal meningitis (CM) remains unacceptably high in low-income countries, where applicable and effective antifungal strategies are needed. In a context where Amphotericine B (AmB) is unavailable, in hospitals lacking intensive care units, we evaluated prospectively the safety and efficacy of an oral combination of fluconazole 1600 mg and flucytosine associated with serial lumbar punctures (SLP) in HIVassociated CM. Methods: Eligible HIV-infected patients presenting a first episode of CM were enrolled in a one-arm open-label clinical trial in Burundi and Ivory Coast from 2012 to 2015. After inclusion, patients received fluconazole 1600 mg per day in combination with flucytosine 100 mg/kg per day for 2 weeks, followed by fluconazole alone, 800 mg per day for 8 weeks and then 200 mg until the end of follow-up (24 weeks). Intracranial pressure was treated with SLP, according to IDSA recommendations. The primary endpoint was 10-week mortality, expected at 35% ±15% precision. Secondary endpoints were 2-week and 24-week mortality, early fungicidal activity (EFA) determined by serial quantitative cerebrospinal fluid (CSF) cultures, and safety. Results: Forty-one (22F/19M) patients were included, 59% being antiretroviral therapy (ART)-naive; 14 (34%) had reduced level of consciousness and 24 (59%) had elevated intracranial pressure at presentation. Overall 10-week mortality was 48.8% (95% CI = 32.9-64.9); 2-week and 24-week mortality were 26.8% (14. 2-42.9) and 58.5% (42.1-73.7), respectively. Mean EFA was −0.27 +/-0.20 log CFU/ml per day, and 16 patients had sterile CSF after two weeks of treatment. The treatment appeared to be well tolerated with no study drug discontinuation. For naive patients, ART was started at a median of 28 days, with no cryptococcal immune reconstitution inflammatory syndrome observed. Conclusions: Mortality with high dose oral fluconazole and flucytosine associated with SLP was at the upper range of expected values. Oral treatment seems to be less effective than AmB-based combinations, probably due to lower fungicidal activity. Nevertheless, in low-income countries where AmB is not available, this combination appears to be a well-tolerated therapeutic option.

WEPDB0104
Comparison of various anal intraepithelial neoplasia screening strategies including standard anoscopy, anal cytology and HPV genotyping in HIV-positive men who have sex with men Abstract WEPDB0104- Table 1 Background: Some have expressed concern that gay and bisexual men (GBM) who use pre-exposure prophylaxis (PrEP) will engage in more condomless anal sex (CAS) and acquire/transmit STIs more frequently while on PrEP. Others, however, argue that increases in STIs among men on PrEP result from required STI screening and testing. There has been little longitudinal data to support either conclusion.
Conclusions: BIA and GI may be early indicators or risk factors for genital ulceration; importantly, partners' non-STI GI is also a strong risk factor, and screening of both partners for BIA and GI is indicated. Uncircumcised men with foreskin smegma were at increased risk for genital ulceration. Interestingly, HSV-2-positivity was only predictive of genital ulcer for HIV+ men. Targeted screening among HIV+ individuals with more advanced stage of disease may be worthwhile.

WEPDD0101
Community-based testing strategies among sex workers in the transport corridor in Mozambique

WEPDX0102
Transmission cluster-specific pattern of adaptive evolution of the HIV-1 envelope gp120 protein sequence in a Japanese MSM population Analyses were restricted to the first sample collected from each patient and drug resistance codons were censored from the alignment. Phylogenetic trees were inferred using FastTree2. Phylogenetic clusters of five or more participants were identified using a tip-to-tip distance cutoff <0.02 substitutions per site. Diversification rate and phylogeographic analyses were conducted in R and BEAST, respectively. Results: We observed variation among provinces in the proportion of non-subtype B infections, with the Prairies displaying significantly greater numbers of non-B infections (p < 0.05). We recovered 285 clusters of size ≥5 (Figure 1). Cluster size was associated with proportion of people who inject drugs (p < 0.004). Most provinces contain large, primarily province specific, clusters dominated by transmission through injection drug use. Some betweenprovince clustering is observed (n = 55 clusters including 3 or more provinces). Association of clusters with more than one province was associated with proportion MSM risk factor (p < 0.05). Consistent with other evidence, the Prairies had the highest rates of HIV diversification.
Abstract WEPDX0104- Figure 1. Phylogenetic clusters of HIV in Canada. Background: Next-generation sequencing (NGS) has transformed genomics for many pathogens, increasing the availability of sequence data for studying evolution, epidemiology, vaccines and therapeutic design. However, on the comprehensive Los Alamos National Laboratory database, more than 90% of HIV sequences were generated with the older method of Sanger sequencing. Widespread adoption of NGS may have been hindered by the technical difficulty of reconstructing and interpreting quasispecies data from reads (short fragments of DNA), given the high diversity of HIV between and within hosts. Methods: Our public computer program 'shiver' reconstructs whole HIV genomes using two commands, by mapping (aligning) reads to a reference genome constructed specifically for each sample to maximize accuracy, then finding the consensus of the reads. Our program 'phyloscanner' interprets mapped reads with a single command. This extracts all patients' reads in a sliding window of the genome, processes and aligns them, constructs a phylogeny with RAxML, analyses the diversity within and between hosts, performs ancestral host-state reconstruction and produces a per-patient summary. Contamination is detected by comparing reads between patients and by finding phylogenetic outliers. We used these two programs on new whole-genome NGS data for 24 seroconverters from European and African cohorts: 5 each from subtypes A1, B, C and D, and 1 each from 01_AE, 01_AG, F1 and G. Results: In our data, we found donor-recipient pairs with the direction of transmission suggested by the host-state reconstruction, identifying ancestry between quasispecies. (Unsampled intermediate patients are always possible.) Variability across the Abstract WEPDX0105- Figure 1. Understanding NGS data for HIV.
genome highlighted the importance of whole genomes and many genomic windows for detecting transmission from sequence data. Dually infected individuals had two distinct phylogenetic clusters of reads. Phylogenetic intermediates showed where these two strains recombined in the host. Conclusions: Raw NGS data for HIV can be analysed easily and powerfully with our public tools phyloscanner and shiver. Methods: Reverse transcriptase, protease and integrase sequences were obtained by SS and UDS from 70 HIV-1 infected, treatmentnaive MSM diagnosed between January 2012 and July 2013 in Paris. Maximum likelihood phylogenies were estimated using FastTree and RAxML, with SH-like tests and 1000 bootstrap replicates, respectively, with both data sets. Transmission events were identified as clades with branch support ≥70% and intra-clade genetic difference <2.5%. TDR mutations were recognized from the consensus list of TDR surveillance. Results: Median time between the HIV-1 diagnosis and date of the sample used for genotypic analysis was 12 days; 5 diagnosis occurred during the acute infection stage. SS and UDS data concurred in the identification of 7 transmission pairs and 1 cluster of 3 patients. With UDS, direct linkage and direction of transmission was unambiguously inferred in 3/8 and 1/8 clusters, respectively. Sequences from the putative recipient were polyphyletic for 4/8 clusters, suggesting multiple founder viruses and excluding unobserved, intermediary links. By SS, the prevalence of TDR mutations was 5.7% in the unlinked patients and 0.0% in the linked patients (13.2% and 35.3% by UDS, respectively). Minority-resistant variants were not shared among the transmissions chains.

WEPDX0106
Conclusions: While SS and UDS identified the same transmission chains, UDS provided extra information on founder viruses, linkage and levels of TDR. No mutation within the clusters was associated with reduced efficiency of PrEP, even by UDS. Moreover, no sharing of minority-resistant variants was observed among the chains of transmission. These results highlight the benefits of UDS data in the phylogenetic identification of transmission chains, allowing the inference of direct linkage and multiplicity of founder viruses in the recipients, and potentially of direction of transmission.
Background: CD32 has been identified as a marker of a substantially enriched HIV reservoir. Here, we explore the relationship of CD32 expression on CD4 T cells with other correlates of reservoir size including time to viral rebound after treatment interruption.
Methods: CD32 expression was measured by flow cytometry on PBMCs (n = 39) and tonsillar tissue (n = 1) from individuals who initiated ART during primary HIV infection (PHI), and uninfected controls (n = 10). Co-expression with immune checkpoint receptors (ICRs), lineage, memory and T follicular helper (Tfh) markers was measured. HIV DNA was quantified in bulk and sorted CD32+ and CD32-populations.

Results:
One-year post-ART initiation, the frequency of CD32+ CD4 T cells was 1.5% (range 0.2-6.4), and did not differ from controls. CD32+ CD4 T cells were found predominantly within differentiated memory subsets (transitional, effector memory and T EMRA ) compared with CD32-CD4 T cells (all p < 0.001) for HIV+ (n = 20) and controls. CD32+ CD4 T cells were highly enriched for HIV DNA compared with CD32-cells (average 103fold, n = 6, p = 0.03), although CD32 percentage did not correlate with reservoir size (n = 29). In a subset of individuals (n = 19) who interrupted ART after 48 weeks, CD32+ CD4% did not predict viral load rebound, although all three individuals with persistently undetectable viraemia had CD32+ CD4% below the median. CD32+ CD4 T cells from blood had higher expression of PD-1, Tim-3 and TIGIT (all p < 0.0001) and a higher density of CD2 (p = 0.001) than CD32-cells in HIV+ participants (n = 20) and controls. Tonsil CD32+ CD4 T cells (n = 1) showed a similar pattern of memory distribution and ICR expression as the periphery. Although tonsillar CD32+ CD4 T cells had higher individual expression of Bcl-6, ICOS and CXCR5 than CD32cells, the co-expression pattern was not consistent with a Tfh phenotype.
Conclusions: We confirm the role of CD32 as a marker of the HIV reservoir, and show that this may occur early during PHI on more differentiated CD4 T cells and is highly co-expressed with ICRs. That expression is similar between HIV+ and HIV-individuals and suggests that preferential infection or survival of CD32+ cells, rather than CD32 up-regulation, is responsible for the observed enrichment. Background: Adherence clubs, where groups of 25-30 patients stable on antiretroviral therapy (ART) meet for counselling and medication pick-up, are an innovative model to retain patients in care and facilitate task-shifting. Adherence clubs can be organized at a clinic or community venue. We performed a randomized controlled trial to compare club retention between community and clinic-based adherence clubs.

MOAX0203LB
Weekly oral MK-8591 protects male rhesus macaques against repeated low-dose intrarectal challenge with SHIVC109P3 Methods: Two groups of 8 male RM were given 5 mL/kg of 10% Tween80 with (treated) or without (placebo) 3.9 mg/kg MK-8591 by oral gavage on day 0, day 7 and weekly thereafter for a maximum of 14 doses or until SHIV infection was confirmed. All animals were challenged intrarectally (IR) with 1 ml of 50 TCID 50 of SHIVC109P3, a viral stock derived from the third passage in RM of the molecular clone SHIVC109F.PB4, which contains an HIV Env initially derived from a newly HIV-infected Zambian. Challenges occurred on day 6 and weekly thereafter for a maximum of 12 challenges or until infection was confirmed. Prior to weekly challenge, blood was drawn to determine infection status and drug levels. Infection was confirmed by real-time RT PCR amplification of viral gag sequences in plasma on two consecutive samples. Proviral DNA was measured by PCR and virus-specific antibody responses were assessed. Intracellular levels of MK-8591-triphosphate (TP) were measured by LC/MS/MS. Results: All placebo animals became infected after 1-4 challenges (median 1, mean 2). All treated animals remained uninfected after 12 challenges and were followed through Week 24 without evidence of infection as determined by the absence of plasma viremia, proviral DNA and seroconversion. MK-8591-treated macaques had a 41.5-fold lower risk of infection (95% C.I 7.3, 237.9) compared with placebo macaques (p < 0.0001, log-rank test). Mean trough concentrations of the active MK-8591-TP at the time of challenge were 4.07mM (range: 2.26-5.17) and compare favourably with the level achieved by a weekly oral dose of 10mg in HIV-1-infected humans.
Conclusions: MK-8591 is a potent NRTTI that completely protected against repeated low-dose IR challenge in this SHIVC109P3/RM model with intracellular active drug concentrations readily achieved in humans. These results support the potential use of MK-8591 for HIV prophylaxis.
Abstract MOAX0202LB- Conclusions: INSTI is unique for its use of a "hanging" fingerstick blood drop, without the need for a collection device. This first field study of such a fingerstick blood-based self-test provides strong evidence that the INSTI HIV Self-Test is accurate, acceptable and easy to use by self-testers with diverse backgrounds in sub-Saharan Africa. Modifications to the kit instructions to include a visual of a weak positive result would provide a more consistent interpretation.

TUAA0106LB
Efficacy of epithelial stem cell-based AIDS vaccine to induce mucosal immune responses offering protection against SIV challenge in macaques Background: A key obstacle limiting the development of an effective AIDS vaccine is the inability to deliver antigen for a sufficient period of time resulting in weak and transient protection. HIV transmission occurs predominantly across mucosal surfaces; therefore, an ideal vaccine strategy would be to target HIV at mucosal entry sites to prevent infection. Methods: We developed a SIV single cycle vaccine under the control of the involucrin promoter (pINV-SIVsc), which was tested for its ability to drive SIV expression in terminally differentiated epithelial cells, induce mucosal immune response and offer better protection against SIV challenge. A total of 20 naive young Rhesus macaques were selected (10/20 expressed MHC class I Mamu-A*01 allele). The pINV-SIVsc vaccine was administrated intravaginal (n = 12) at Week 0. Animals were monitored overtime for specific immune responses in blood and various tissues (n = 4). Eight animals were challenged at Week 12 (n = 4) or at Week 24 (n = 4) using repeated pathogenic SIVmac239 and monitored for specific immune responses in blood and various tissues. Eight additional animals were infected with repeated SIVmac239, and served as unvaccinated Controls. Complementary approaches were used to characterize SIV-specific immune responses in blood, vaginal secretions, LN and vaginal biopsies collected at various times.
Results: This vaccine induced strong mucosal IgA and IgG responses and specific T cells expressing a4ß7 homing to the mucosa. Repeated challenges revealed significant delay and lower viremia with 3-4 logs-reduction at peak, 4-5 logs-reduction at set-point and undetectable viremia by Week 10-14 post-SIV in vaccinated females compared to Controls. Following challenge, we demonstrated a positive correlation between the generation of mucosal and systemic T cell responses and control of viremia, an inverse association between viremia and post-infection vaginal IgA/IgG responses.

Conclusions:
We have obtained evidence, within the limitation of the small animals' number studied, that macaques vaccinated with pINV-SIVsc can generate strong mucosal SIV-specific T cell responses and local antibody responses (IgA/IgG). We demonstrated the efficacy of an epithelial stem-cell-based SIV vaccine to serve as antigen delivery system suggesting an important role in protection against mucosal infection.

TUAB0104LB
Fixed dose combination of doravirine/lamivudine/TDF is non-inferior to efavirenz/emtricitabine/TDF in treatmentnaive adults with HIV-1 infection: Week 48 results of the Phase 3 DRIVE-AHEAD study Background: DAWNING is a non-inferiority study conducted to compare a protease inhibitor-sparing regimen of DTG+2NRTIs with a current WHO-recommended regimen of LPV/RTV+2NRTIs in HIV-1 infected subjects failing first-line therapy of a non-nucleoside reverse transcriptase inhibitor (NNRTI) + 2 NRTIs (ClinicalTrials. gov: NCT02227238). An Independent Data Monitoring Committee (IDMC) performed periodic reviews of data to protect the ethical and safety interests of subjects. Methods: Adult subjects failing first-line therapy, with HIV-1 RNA ≥400 copies(c)/mL, were randomized (1:1, stratified by Baseline plasma HIV-1 RNA and number of fully active background NRTIs) to 52 weeks of open-label treatment with DTG or LPV/RTV combined with an investigator-selected dual NRTI background, including at least one fully active NRTI. An IDMC review was performed, which included data from 98% (612/627 randomized) of subjects through 24 weeks on therapy.
Results: At Week 24, 78% of subjects on DTG versus 69% on LPV/ RTV achieved HIV-1 RNA <50 c/mL (adjusted difference 9.6%, 95% CI: 2.7% to 16.4%, p = 0.006 for superiority). The difference was primarily driven by lower rates of Snapshot virologic non-response in the DTG group. The safety profile of DTG+2NRTIs was favourable compared to LPV/RTV+2NRTIs with more drug-related adverse events (AEs) reported in the LPV/RTV group, mainly due to higher rates of gastrointestinal disorders. Following review of Week 24 data and large subsets of data from Weeks 36 and 48, the IDMC recommended discontinuation of the LPV/RTV arm due to persistent differences in rates of Snapshot virologic non-response and protocol-defined virologic failure (PDVF) favouring the DTG arm.
Conclusions: The IDMC recommended discontinuation of the LPV/ RTV arm due to superior efficacy of DTG+2NRTIs and the potential to harm subjects on LPV/RTV based on available data. Final Week 24 results of this study will be presented. DAWNING provides important information to help guide second-line treatment decisions in resource-limited settings.

TUAB0106LB
HIV-specific broadly neutralizing monoclonal antibody, VRC01, minimally impacts time to viral rebound following treatment interruption in virologically suppressed, HIVinfected participants who initiated antiretroviral therapy during acute HIV infection Conclusions: Introducing HIVST for 3 months in communities already exposed to door-to-door HIV-testing services for 3 years increased the proportion of the population who knew their HIV status. This effect was seen most markedly in men.

TUAC0106LB
Safety, tolerability and pharmacokinetics of long-acting injectable cabotegravir in low-risk HIV-uninfected women and men: HPTN 077 Background: Prospective data on the association of HIV transmission and undetectable viral load (UVL) in homosexual male HIVserodiscordant couples (HM-SDC) are extremely limited. We report the final results from the Opposites Attract cohort study of HM-SDC in Australia, Bangkok and Rio de Janeiro. Methods: HM-SDC were recruited through clinics. Information on sexual behaviours was collected at each visit from the HIVnegative partner (HNP). HNPs were tested at baseline and follow-up for HIV antibodies/sexually transmitted infections (STIs), and positive partners (HPPs) for HIV viral load/STIs. Phylogenetic analysis of pol and env genes was performed to identify linked HIV transmissions within couples based on genetic distance and monophyletic grouping. Incidence was calculated per coupleyear of follow-up (CYFU), and stratified by pre-exposure prophylaxis (PrEP) use and by whether different forms of condomless anal intercourse (CLAI) were reported. UVL was defined as <200 copies/mL. One-sided upper 95% confidence limits (UCL) were calculated. Results: By end 2016, 358 HM-SDC were enrolled: 157, 105 and 96 from Australia, Thailand and Brazil, respectively. There were 591 CYFU in 343 couples with at least one follow-up visit of whom 57.4% reported anal sex with outside partners during any point in follow-up. At baseline, 79.9% of HPPs were on anti-retroviral therapy (ART) and 77.9% had UVL; STI prevalence was 14.3%/ 11.7% in HPPs/HNPs, respectively. There were 318 CYFU in periods where CLAI was reported with a total of 16,889 acts of CLAI. There were 3 new HIV infections but no linked transmissions. The overall UCL of the transmission rate when CLAI was reported was 1.16/100 CYFU, and it was 1.56/100 CYFU when there was UVL.
Conclusions: There were no linked HIV transmissions in almost 600 CYFU involving close to 17,000 acts of CLAI in HM-SDC. Our results provide strong support for the hypothesis that undetectable viral load prevents HIV transmission in homosexual men.

TUAC0206LB
Safety and acceptability trial of the dapivirine vaginal ring in US adolescents Background: Young women aged 15-25 years are disproportionately affected by the HIV epidemic. Two Phase 3 trials of a 25mg dapivirine vaginal ring demonstrated HIV-1 risk reduction in adult women over 21, but not in those aged 18-21. Lack of protection was correlated with low adherence. Methods: A Phase 2a randomized, double-blind, placebo-controlled trial of the dapivirine ring was conducted in sexually active females, aged 15-17. Participants were randomized 3:1 to a dapivirine or placebo ring to be inserted monthly for 6 months. Safety endpoints included Grade 2 product-related adverse events (AE) and Grade 3 and higher AEs. Adherence to ring use was assessed through self-report, plasma dapivirine concentrations and residual levels in used rings. A plasma dapivirine concentration >95 pg/mL was used to define short-term adherence (hours); a dapivirine residual level <23.5 mg was used to define long-term adherence (monthly). Acceptability was assessed through computer-assisted self-interviews. Results: Ninety six participants were enrolled across six US sites. The mean age was 16.3 years; 59% were black and 34% white. Adherence to study visits was 97%. There were no differences in safety outcomes between treatment arms. By self-report, 42% (95% CI 32, 52) of participants reported that they never removed the ring Abstract TUAC0506LB- except to replace it monthly. In the dapivirine group, drug levels indicated adherence in 87% of plasma samples and 95% of rings. Participants noted no discomfort due to the ring at 87% of visits and "liking" the ring at 93% of visits. The most frequently cited concern (28%) involved their primary sex partner feeling the ring during sex. Conclusions: The dapivirine vaginal ring, a promising microbicide approach, is safe and acceptable in this population. By self-report and objective measures, adherence was high. Discussing potential partner concerns with participants prior to ring use may positively influence adherence.

TUAC0207LB
Pluspills: an open-label, safety and feasibility study of oral pre-exposure prophylaxis (PrEP) in 15-19-year-old adolescents in two sites in South Africa remained constant and HIV incidence was low. SA adolescents need access to PrEP with tailored adherence support and potentially augmented visit schedules.

TUAD0206LB
Willingness and ability of OST clients to pay for some OST services in Odesa region in the light of the Ukraine crisis Background: Odesa region is one of the highest burden of HIV in Ukraine. HIV epidemic is mainly concentrated among key populations (PWIDs, CSWs and their partners, MSMs). The international donors' support of OST programmes for PWIDs are currently in the process of financing reduction. The economic crisis in Ukraine complicated the nationwide state financing of OST programmes in Ukraine. The implementation of the alternative financing models such as co-payment is optimal way to secure critical for PWIDs services. The aim of the study is to assess the willingness and ability of current OST clients to pay for some OST services.
Methods: Cross-sectional study was conducted in Odesa city in 2016 among current OST clients in Odesa city. Random sampling strategy was used for respondent recruitment. Structured questionnaire was used for data collection. The study protocol received an approval of IRB at Ukrainian Institute on Public Health Policy. Bivariate analysis was used to estimate correlates of clients' willingness to pay for some OST services.
Results: A total of 161 clients of OST programmes (125 (77.6%) males and 36 (22.4%) females) were interviewed. The average age of the respondents was 41.98 years (SD = 8.0). 66.5% of interviewed OST clients in Odesa city had a secondary education and 14.9% had a higher education. Half of the respondents (49.7%) had a single marital status. 74% of respondents uptake OST programme for more than 12 months. Overall, 89% of respondents (n = 143) demonstrate willingness to pay for some OST services in Odesa. The willingness to co-pay for some OST services is higher among clients listed "possibility to uptake additional medical care at OST site" as personal benefit of OST participation. The willingness is also correlated with income per family member (p = 0.009), self-assessment of financial wellbeing (p = 0.008) and perceived social support (p = 0.014).
Conclusions: Clients of OST programme in Odesa region demonstrated willingness and ability to pay for some OST services despite an economic crisis in the country. Their willingness depends on socio-economic status of PWID and his/her family. These important correlates of willingness and ability to pay should be considered for design and implementing OST co-payment service delivery model in Odesa region.

TUAD0106LB
Index partner testing and targeted case finding in northern Haiti edPrEP (n = 38). Infection risk was also considered: dPrEP was preferred for unplanned and/or frequent sexual risk behaviour (n = 79), while edPrEP was chosen when risk was more predictable (n = 57). Some chose for, or switched to, edPrEP to inhibit sexual risk behaviour (n = 4), while others chose for, or switched to, dPrEP to gain more sexual freedom (n = 17). Other reasons to switch to edPrEP included experiencing side effects (n = 14), having less sex than anticipated (n = 20), experimenting with another regimen (n = 2) and receiving negative reactions from the environment (n = 1). Doubts about edPrEP's safety (n = 2), inability to plan sex (n = 13) and desire for more structure (n = 9) were motivators to switch to dPrEP. Motives to temporarily stop dPrEP (n = 99) were situational (e.g. medical issues or vacations). Changed life circumstances (n = 2) and reduced sexual risk (n = 6) were motives to completely stop with PrEP use (n = 12).
Conclusions: A great variety of personal and contextual factors determine the choices for PrEP regimens, related switches and stops. In order to successfully support future PrEP users, a tailored approach, addressing choices for PrEP regimens as a continuum of flexible and changeable choices, is essential.

WEAD0206LB
Cost-effectiveness of routine viral load monitoring in lowand middle-income countries: a systematic review Background: Routine viral load (VL) monitoring for HIV-1 management of persons on ART has been recommended by the WHO to identify treatment failure. However, VL testing represents a substantial cost in resource-constrained healthcare systems. The central challenge is whether and how VL monitoring may be delivered such that it maximizes health gains across the population for the costs incurred. We hypothesized that key efficiency assumptions about programme design and cost drive the cost-effectiveness of programmes with viral load monitoring. Methods: We conducted a systematic review of studies on the cost-effectiveness of viral load monitoring in low-and middleincome countries (LMIC). We followed the Cochrane Collaboration guidelines and the PRISMA reporting guidelines. Results: We identified 18 studies that evaluated the cost-effectiveness of viral load monitoring in HIV treatment programmes. Overall, we identified three key factors that make it more likely for viral load monitoring to be cost-effective: (1) Use of lower cost, robust approaches to viral load monitoring; (2) Ensuring the pathway to health attainment is established and that viral load results are acted upon; (3) Viral load result is used to simplify HIV care in patients with viral suppression (i.e. differentiated care, with fewer clinic visits and longer prescriptions); viral load monitoring in differentiated care programmes provides evidence that reduced clinical engagement, where appropriate, is not impacting health outcomes.
Conclusions: The cost-effectiveness of viral load monitoring critically depends on the context. To achieve this goal of being cost effective, viral load monitoring will need to facilitate scale up of differentiated care approaches to HIV treatmentintroducing viral load monitoring without differentiated care is unlikely to be costeffective in most settings and results in lost opportunity for health gains through an alternative use of limited resources. As countries scale up differentiated care programmes, data on clinical outcomes and cost are essential to evaluate the ongoing cost-effectiveness of viral load monitoring as utilized in practice.

WEAD0106LB
The Eligible women had weaned infants from breastfeeding before enrolment, but were able to pump breast milk. Participants were instructed to wear the VR continuously for 14 days. Milk and blood plasma samples were collected (pre-insertion, Hour 3, Hour 6, Hour 24, Day 7, Day 14 after ring placement, and two days after ring removal) and analysed for DPV using validated ultra-performance liquid chromatography-tandem mass spectrometry assays, with lower limits of quantification of 10 pg/ mL and 20 pg/mL for milk and plasma, respectively. We estimated infant DPV intake assuming 150 mL/kg/day milk ingestion. Adverse events (AEs) were collected at all participant contacts. Here, we describe antiviral efficacy and tolerability of single doses of 0.5 mg to 30 mg MK-8591 over 7 to 10 days in HIV-1 infected subjects in an ongoing Phase 1b monotherapy proof-of-concept efficacy study.

Methods:
In an open-label study in HIV-1-infected subjects naive to antiretroviral treatment (ART), subjects are being administered a single dose of MK-8591 across a range of doses. Blood samples are being collected for viral load (VL), MK-8591 PK and MK-8591-TP PK at pre-specified time points up to 7 to 10 days post-dose. Following completion of Day 7 or Day 10 procedures, subjects are being offered standard of care ART. Safety, PK and VL data from the doses of 0.5 mg, 1 mg, 2 mg, 10 mg and 30 mg (N = 6/panel) are available.
Results: Single doses of MK-8591 across the entire tested range were associated with a rapid and robust reduction in VL. At 168 hours postdose, a mean (95% CI) placebo-corrected VL reduction of 1.18 log10 (0.95, 1.46) was observed for 0.5mg. For the 30mg dose, mean VL continued to decline through Day 10 with a mean placebo-corrected reduction of 1.57 log10 (1.34, 1.85), with no evidence of recrudescence at any dose. In samples with sufficient VL for testing (14/24), no common mutant strains, including M184V/I, were detected. All doses were generally well tolerated, with a limited number of mild/moderate adverse experiences reported. MK-8591 plasma and MK-8591-TP PBMC PK were similar to previously observed data in healthy subjects. Conclusions: MK-8591 suppressed HIV replication for at least seven days when administered as a single dose as low as 0.5mg. The antiviral potency, human pharmacokinetics (PK) and physical properties of MK-8591 have the potential to open new paradigms for extended duration HIV treatment and prophylaxis approaches.

TUPDB0203LB
Pharmacokinetics, pharmacodynamics and pharmacogenomics of efavirenz 400mg once-daily during pregnancy and postpartum Background: Antiretroviral dose reductions may compensate for the finite global manufacturing capacity and allow access programmes to reach larger numbers of HIV-infected patients. The ENCORE-1 study showed that efavirenz 400mg (EFV400) is noninferior to the standard adult dose. WHO clinical guidelines now recommend EFV400 as an alternative first-line agent, however with a disclaimer that no data on EFV400 during the third trimester of pregnancy (TT) exist. This study investigated the pharmacokinetics (PK), efficacy and CYP2B6 pharmacogenetics of EFV400 in women living with HIV (WLWH) during TT and post-partum (PP) with a view to removing the disclaimer and allowing wider EFV400 use in firstline.
Conclusions: Cmax, AUC and C 24h in TT were 14%, 26% and 38% lower compared to PP but within ranges of those measured for EFV600 during TT by Schalkwijk et al. (2016) and those measured in ARV-naive patients on EFV400 in ENCORE- 1 (Dickinson et al. 2015). All subjects maintained a VL <50, suggesting that EFV400 can be used in pregnant WLWH.

TUPDB0204LB
Universal sputum testing versus symptom-based testing for tuberculosis (TB) in HIV-infected pregnant women: a cluster-randomized implementation trial in South Africa Background: TB in HIV-infected pregnant women is a leading cause of maternal and infant morbidity and mortality. Currently-recommended symptom-based screening of HIV-infected pregnant women may be insensitive. Methods: We conducted a cluster-randomized trial to compare universal sputum TB testing of HIV-infected pregnant women against standard symptom-based testing. Sixteen public-sector antenatal clinics in two health districts were assigned to either