CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

1130 IMPLEMENTATION OF U=U IN REAL LIFE IN ITALY: RESULTS FROM THE ICONA COHORT Giordano Madeddu 1 , Andrea De Vito 1 , Alessandro Cozzi-Lepri 2 , Antonella Cingolani 3 , Franco Maggiolo 4 , Carlo Federico Perno 5 , Roberta Gagliardini 6 , Giulia Marchetti 5 , Annalisa Saracino 7 , Antonella D'Arminio Monforte 5 , Andrea Antinori 6 , Enrico Girardi 6 , for the ICONA Foundation Study Group 1 University of Sassari, Sassari, Italy, 2 University College London, London, UK, 3 Catholic University of the Sacred Heart, Rome, Italy, 4 Azienda Ospedaliera Papa Giovanni XXIII, Bergamo, Italy, 5 University of Milan, Milan, Italy, 6 Lazzaro Spallanzani National Institute for Infectious Diseases, Rome, Italy, 7 University of Bari, Bari, Italy Background: Zero risk of linked HIV transmission in sero-discordant couples when the HIV-infected partner had viral load (VL)<200 copies/mL (U=U status) was observed in large observational studies. We aimed at estimating the proportion of time in which this status was maintained and identifying factors associated to the risk of losing it among people living with HIV (PLWH) enrolled in a clinical cohort. Methods: We included participants in the ICONA cohort who had reached an established U=U status (VL<=200 copies/mL for >6 months) as of December 2010. The outcome was the number of person days of follow up (PDFU) with a VL>200 copies/ml (cp/ml), relative to the total number of PDFU observed in follow-up. Logistic regression model was used to identify factors independently associated to the risk of losing U=U status. For this analysis, a participant was defined as losing his/her U=U status if he/she spent <90% of his/her PDFU on observation with a VL <=200 cp/mL. The median of VL measurements was 9 (IQR: 4-15). Results: 8,241 PLWH were included in the analysis and contributed 12,670,888 PDFU. Of these, 1,648 (20%) were female, 768 (9%) were people who inject drugs (PWID), 3,786 (46%) men who have sex with men and 3,176 (39%) heterosexuals. Overall, during the entire follow-up, 96.9% of PDFU observed were with a VL<=200 cp/ml. Thus, only 3.1% of PDFU were observed when VL was >200 cp/mL with some evidence for a decrease after 2016. The median time with VL>200 cp/ml was 47.3 days (IQR: 46.3-47.9). Of note, the proportion of PDFU with VL>200 cp/ml was higher than average in females (5.3%), unemployed (5.4%), PWID (4.7%) and in people with>3 previous virological failures (6.3%). At individual level, 617 participants (7.5%) spent <90% of PDFU with a VL<=200 cp/mL and were classified as losing their initial U=U status. Unadjusted and adjusted OR of losing U=U status from fitting the logistic regression model are shown in Table 1. Conclusion: In our population of PLWH meeting the definition of U=U this status was maintained for 97% of the following 10 years of observation with a trend towards an increase in recent years. These findings reinforce the validity of the U=U message in real world settings. However, we identified subsets of our population, including females and foreign-born, at higher risk of not maintaining the U=U status, for whom greater efforts are needed to reduce these infrequent periods of VL>200 cp/ml.

1131 IMPACT OF PrEP AND TasP ON INCIDENCE OF HIV DIAGNOSES IN 48 HIGHEST-BURDEN US AREAS Robertino Mera Giler 1 , Staci Bush 1 , Trevor Hawkins 1 , Moupali Das 1 , Julius Asubonteng 1 , Scott McCallister 1 1 Gilead Sciences, Inc, Foster City, CA, USA Background: Use of Tenofovir Disoproxil/Emtricitabine (TVD) for Pre-Exposure Prophylaxis (PrEP) has significantly reduced the HIV diagnosis rate in many US states, independent of the effect of treatment as prevention (TasP). Methods: Using publicly available HIV surveillance data on HIV diagnoses from 105 US metropolitan statistical areas (MSAs) (2012-2017), virologic suppression data from 38 US states and DC as a proxy for Treatment as Prevention (TasP), and TVD for PrEP drug utilization obtained from a national pharmacy and medical claims database, we evaluated the independent impact of PrEP and TasP on HIV diagnosis rates in 48 counties and localities in the End the Epidemic Initiative (48-ETE). We calculated the person time at risk of HIV infection excluding time of those taking PrEP as well as those who became HIV positive. Incidence rates, rate ratios and 95% confidence intervals were assessed using a multilevel Poisson regression model for the 48-ETE and overall after adjusting for the effect of PrEP and TasP. Results: Over this 6-year analysis, the US rate of HIV diagnoses in the 48-ETE locations decreased at a rate of 7.1% (95%CI –6.9 to –7.3%) per year while PrEP use in those with a CDC-defined PrEP indication increased 9.9-fold in the same locations from a mean 1.31/100 individuals (95% CI 0.3-2.3) in 2012 to 13.1/100 (95% CI 12.1-14.1) in 2017. HIV viral suppression (viral load <200 c/mL) increased by 1.4% per year (95% CI 1.1 to 1.7%) during the same time among HIV treated subjects. A multivariate poisson model showed that PrEP use was significantly associated with the decline in the rate of new HIV cases in the 48-ETE localities, independent of a significant TasP effect. 48-ETE localities with an average PrEP use rate of 17.4 per 100 subjects at risk could expect a decline of 15.5% in the rate of new HIV diagnoses. 48-ETE localities had significantly higher new HIV diagnosis rate than the rest of the US MSAs (IRR 2.0, 95% 1.61 – 2.58), but had a significantly lower PrEP use (-2.1 per 100 subjects at risk, 95% CI -0.93 to -3.2), and TasP proportion (-1.30 %, 95% CI -0.41 to -2.2%) than those MSAs not selected for intervention. Conclusion: From 2012-2017, HIV diagnoses declined significantly in the 48 counties and localities selected for intervention where PrEP use was the highest. The effect of PrEP use was significantly associated with this decline and was independent of treatment as prevention. Improvements in PrEP and TasP coverage in these localities could yield important declines in the rate of new HIV diagnoses. 1132 IMPACT OF FOOD INSECURITY ON THE HIV EPIDEMIC IN SUB-SAHARAN AFRICA (2015–2017) Andrea Low 1 , Elizabeth Gummerson 1 , Amee M. Schwitters 2 , Rogerio Bonifacio 3 , Nicholus Mutenda 4 , Karampreet K. Sachathep 1 , Choice Ginindza 5 , Avi Hakim 6 , Danielle T. Barradas 7 , James Juma 8 , Samuel Biraro 1 , Keisha Jackson 6 , Neena M. Philip 1 , Steven Hong 9 , Sally Findley 10 1 ICAP at Columbia University, New York, NY, USA, 2 CDC Lesotho, Maseru, Lesotho, 3 World Food Programme, Johannesburg, South Africa, 4 Ministry of Health and Social Services, Windhoek, Namibia, 5 Central Statistical Office, Mbabane, Swaziland, 6 CDC, Atlanta, GA, USA, 7 CDC Zambia, Lusaka, Zambia, 8 Tanzania Ministry of Health, Community Development, Gender, Elderly, and Children, Dar es Salaam, Tanzania, United Republic of, 9 US CDC Windhoek, Windhoek, Namibia, 10 Columbia University, New York, NY, USA Background: To assess associations between food insufficiency (FI) and HIV-related outcomes, including infection, we used data from nationally representative population-based HIV impact assessment (PHIA) surveys in Zambia, Eswatini, Lesotho, Uganda, Tanzania, and Namibia (2015–2017). Methods: We collected FI data, defined as having any time with no food in the house in the past 4 weeks, from the household head. We also offered household-based HIV testing. Recent infection (<130 days) was measured using the HIV-1 Limiting Antigen (LAg) Avidity assay combined with lack of viral load suppression (VLS, <1000 copies/mL) and antiretroviral (ARV) testing data. Recent infection indications were those with LAg<1.5 normalized Optical Density (ODn), VL>1000 copies/mL, and no detectable ARVs. We performed pooled analyses to determine the association between FI and several HIV-related outcomes on weighted data on adults aged 15–59 years using logistic regression adjusted for age, sex, urban/rural residence, wealth quintile, and education, fitting an interaction term between country and FI. We stratified by sex for transactional

Poster Abstracts

CROI 2020 427