CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

Conclusion: In the first prospective study of its kind from southern Africa, we demonstrate that it is feasible to implement UTT in correctional settings in Zambia and South Africa, and that such an intervention can achieve high early ART uptake and excellent viral suppression for HIV-positive inmates during incarceration. However, frequent facility transfer and release threatens to undermine UTT by limiting access to timely ART and fragmenting care for inmates living with HIV.

of TB status, calendar year, and CD4 count at enrollment. The association between CD4 count and time to ART initiation changed dramatically over time (p<0.0001). The association between TB status and time to ART initiation varied extensively based on CD4 count (p<0.0001) and to a lesser extent the date of enrollment (p=0.06). For a person enrolling with TB and a CD4 count of 500 cells/mm3, the estimated median time to ART initiation was approximately 500 days in 2006, compared to under 50 days in 2015. Conclusion: In recent years in Latin America, there has been a dramatic shortening in the time from enrollment to ART initiation, for both those with and without TB, particularly among those with high CD4 counts.

Poster Abstracts

1022 UPTAKE OF ANTIRETROVIRAL THERAPY IN THE “TREAT ALL” ERA IN RIO DE JANEIRO, BRAZIL Lelia H. Chaisson 1 , Juliana Domenico 2 , Valeria Saraceni 2 , Richard D. Moore 3 , Jonathan Golub 3 , Betina Durovni 4 1 Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA, 2 Secretaria Municipal de Saúde do Rio de Janeiro, Rio de Janeiro, Brazil, 3 Johns Hopkins University School of Medicine, Baltimore, MD, USA, 4 Oswaldo Cruz Foundation - Fiocruz, Rio de Janeiro, Brazil Background: Randomized controlled trials have proven the efficacy of early antiretroviral therapy (ART) for reducing HIV morbidity, mortality, and transmission. As a result, guidelines recommending treatment for all, regardless of CD4 count, are being scaled up worldwide. The Brazilian Ministry of Health has recommended treatment for all since 2013 and offers free antiretrovirals (ARVs) for all HIV-infected patients. We evaluated uptake of “treat all” guidelines in Rio de Janeiro, Brazil, from 2014-2017. Methods: HIV has been a notifiable disease in Brazil since 2014. We included all patients diagnosed with HIV and reported to the Rio de Janeiro Health Secretariat from 2014-2016, with follow-up through 2017. HIV notifications and comorbidities were obtained from the national notifiable diseases information system; ARV prescriptions from the Rio de Janeiro pharmacy information system; and death notifications from the Rio de Janeiro mortality registry. We joined databases using a novel probabilistic linkage strategy. We assessed HIV notifications, prevalence of opportunistic infections (OIs), and median time to ART initiation over time. We used Nelson-Aalen cumulative hazard estimates to construct risk curves comparing 3-month ART initiation by diagnosis year and estimated the hazard of ART initiation by diagnosis year using Cox proportional hazards regression. Results: From 2014-2016, 6,454 persons were diagnosed with HIV and notified to the Rio de Janeiro Health Secretariat. Of these, 2,009 (31%) were female and median age was 34 years (IQR 26-43). 1,725 (27%) had a documented OI, including 417 (6%) with pulmonary tuberculosis. Of 2,628 (41%) patients reported to have initiated ART, 2,028 (77%) did so within 3 months of HIV diagnosis; median time to ART initiation was 42 days (IQR 15-94) and decreased from 51 days [IQR 21-135] in 2014 to 46 days [IQR 20-92] in 2015 and 31 days [IQR 8-68] in 2016. Patients diagnosed in 2015 had an increased hazard of 3-month ART initiation compared to those diagnosed in 2014 (aHR 1.31, 95% CI 1.18-1.46). There was a non-significant increased hazard of 3-month ART initiation for those diagnosed in 2016 compared to 2014 (aHR 1.06, 95% CI 0.95-1.19). Conclusion: The rate of ART initiation in Rio de Janeiro was low, despite the availability of free ARVs and guidelines recommending treatment for all. “Treat

1021 IMPROVEMENT IN TIME TO ART INITIATION REGARDLESS OF TB STATUS IN LATIN AMERICA Bryan E. Shepherd 1 , Serena Koenig 2 , Ahra Kim 1 , Cathy Jenkins 1 , Pedro Cahn 3 , Beatriz Grinsztejn 4 , Marcelo Wolff 5 , Denis Padgett 6 , Juan Sierra-Madero 7 , Eduardo Gotuzzo 8 , Catherine McGowan 1 , Jean William Pape 9 , Timothy R. Sterling 1 , for the The Caribbean, Central and South America Network for HIV Epidemiology (CCASAnet) 1 Vanderbilt University, Nashville, TN, USA, 2 Harvard Medical School, Boston, MA, USA, 3 Fundación Huésped, Buenos Aires, Argentina, 4 Instituto Nacional de Infectologia Evandro Chagas, Rio de Janeiro. Brazil, 5 Fundación Arriarán, Santiago, Chile, 6 Instituto Hondureño de Seguridad Social, Tegucigalpa, Honduras, 7 Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán, Mexico City, Mexico, 8 Universidad Peruana Cayetano Heredia, Lima, Peru, 9 GHESKIO, Port-au- Prince, Haiti Background: In 2006, the World Health Organization recommended antiretroviral therapy (ART) for persons with CD4 count <200 cells/mm3 (<350 if co-infected with tuberculosis [TB]) or Stage 4 HIV disease. Subsequent guidelines recommended earlier ART initiation (2009: CD4 <350 cells/mm3 or Stage 3/4 disease [including TB]; 2013: CD4 <500cells/mm3; 2015: universal ART). The recommended timing of ART initiation relative to TB medications also changed during this period. We characterized temporal trends in the time to ART initiation and compared trends between HIV+ persons with and without TB. Methods: The study included data from HIV clinical sites in Brazil, Chile, Haiti, Honduras, Mexico, and Peru participating in CCASAnet. We included all persons ≥18 years old who were ART-naïve at first clinic visit from 2006 to 2015. We estimated median time to ART initiation as a function of baseline TB status (within 30 days before or after enrollment), CD4 count, and year of enrollment from a multivariable Cox regression model that included these variables, two-way interactions between these variables, sex, education, and age, and stratified by study site. Continuous variables were fit with natural splines to relax linearity assumptions. Results: Of 19,197 patients, 1306 (7%) were diagnosed with TB at enrollment. Patients with TB were more likely to be male, older, less educated, with lower CD4 counts, and living in Haiti or Peru. A total of 17,183 (93%) initiated ART during a median of 3.6 years of follow-up; 96% of those with TB compared to 93%without TB (p<0.001). The median time to ART initiation was 42 days for those without TB, and 43 days for those with TB (p=0.94). The Figure shows the estimated median adjusted time from enrollment to ART initiation as a function

CROI 2019 400