CROI 2019 Abstract eBook

Abstract eBook

Oral Abstracts

initiation in the study population was 5.12% (95% CI 4.32, 6.10). The cumulative incidence of mortality with and without pre-therapy CD4 at 1 year was 4.54% (95% CI 3.73, 5.60) and 7.06% (95% CI 5.14, 9.98), respectively (Cox test for equality p=0.03). After adjustment for pre-therapy WHO stage, sex, age, facility type, ART initiation date, patients without a pre-therapy CD4 had 1.48 times the hazard of mortality in the first year compared to those with a pre-therapy CD4 determination (95% CI 1.00, 2.17, p=0.046). Advanced WHO stage and male sex were associated with higher probability of early mortality (WHO stage IV, HR, 7.69 (95% CI, 4.19, 14.13 p< 0.001) male sex, HR, 1.62 (95% CI, 1.13, 2.32 p< 0.008)). Conclusion: Despite the possibility of unmeasured confounding, these results suggest that patients initiating ART without pre-therapy CD4 experience a higher risk of early mortality even after adjustment for demographic characteristics and disease stage. Even though pre-therapy CD4 are no longer required to determine eligibility, further research to evaluate the safety of discontinuing pre-therapy CD4 is needed before widespread discontinuation. 149 UTILITY OF CD4 CELL COUNT MONITORING IN BOTSWANA: ANALYSIS OF ROUTINE LABORATORY DATA Tshepo B. Leeme 1 , Madisa Mine 1 , Kwana Lechiile 1 , Mosepele Mosepele 2 , Thongbotho Mphoyakgosi 1 , Charles Muthoga 3 , Julia Ngidi 1 , Bornaparte Nkomo 4 , Dinah Ramaabya 4 , Modiri Tau 4 , Mark W. Tenforde 1 , Richard Hayes 5 , Joseph N. Jarvis 1 1 Botswana Harvard AIDS Institute Partnership, Gabarone, Botswana, 2 University of Botswana, Gaborone, Botswana, 3 Botswana–UPenn Partnership, Gaborone, Botswana, 4 Ministry of Health, Gaborone, Botswana, 5 London School of Hygiene & Tropical Medicine, London, UK Background: Botswana has an adult HIV prevalence of 21.9%. An estimated 317,945 patients (84% of HIV-infected individuals) are on treatment. National guidelines recommend both CD4 count and viral load monitoring. Since the country adopted universal test-and-treat in June 2016, an increasing burden has been placed on the health system. This study aims to assess the ongoing need for regular CD4 monitoring in Botswana. Methods: Data from all HIV-infected patients having CD4 counts at the Gaborone clinics served by the Botswana Harvard reference laboratory during 2015, 2016, and 2017 were analysed. CD4 count and viral load data were assessed to determine the proportion of patients presenting with advanced disease (CD4<200 cells/µL), trends in CD4 cell counts over time, and the proportion of patients presenting without advanced disease experiencing a drop in CD4 count to below 200 cell/µL during follow up. Results: 193,050 CD4 counts were performed on 60,899 patients, with a median frequency of monitoring of 1.48 CD4 measurements per patient per year. 76% (46,474) of patients were established clinic patients, while 24% (14,425) were new to care during the study period. 24.8% (3,571/14,425) of new patients presented with CD4 count<200 cell/µL. Age and sex were strongly associated with advanced disease, with men more likely to present with advanced disease than women (34.9% vs 18.9%, p<0.001). Increased age was associated with lower CD4 cell count. 54% of patients with baseline CD4 cell counts below 200 cell/µL attained a CD4 rise to above 200 cell/µL within 12 months of follow-up. In patients with two or more CD4 counts, a very small proportion (3.6% (180/5060)) of those with a baseline CD4 cell count ≥200 cell/µL experienced a drop in CD4 cell count to <200 cell/µL over the study period. 58.9% (106/180) of patients with a drop in CD4 count had a viral load measurement within 2 months of CD4 measurement, of these, 79.2% (84/106) were virally suppressed. Conclusion: A significant proportion of patients in Botswana still present with advanced disease, demonstrating the ongoing importance of baseline CD4 testing to identify patients at risk of opportunistic infections and in need of interventions including cotrimoxazole prophylaxis and cryptococcal antigen screening. Very few individuals with CD4 counts above 200 cells/µL experienced a drop to below 200 cells/µL, suggesting limited utility for ongoing CD4 count monitoring in individuals without advanced disease in settings with routine viral load testing. 150 TRENDS IN CD4 AND VIRAL LOAD TESTING IN SOUTHERN AFRICA: ANALYSIS OF 6 COUNTRIES Elizabeth Zaniewski 1 , Cam Ha Dao Ostinelli 1 , Nicola Maxwell 2 , Mary-Ann Davies 2 , Jonathan Euvrard 2 , Janneke van Dijk 3 , Samuel Bosomprah 4 , Frank

Tanser 5 , Matthew P. Fox 6 , Nathan Ford 7 , Nosisa Sipambo 8 , Josephine Muhairwe 9 , Matthias Egger 1 1 Institute of Social and Preventive Medicine, Bern, Switzerland, 2 Centre for Infectious Disease Epidemiology and Research, Cape Town, South Africa, 3 SolidarMed, Masvingo, Zimbabwe, 4 Centre for Infectious Disease Research in Zambia, Lusaka, Zambia, 5 Africa Health Research Institute, Durban, South Africa, 6 Boston University, Boston, MA, USA, 7 WHO, Geneva, Switzerland, 8 Chris Hani Baragwanath Hospital, Johannesburg, South Africa, 9 SolidarMed, Maseru, Lesotho Background: Since 2015 the World Health Organization has recommended CD4 testing before starting antiretroviral therapy (ART) to detect advanced disease and routine viral load (VL) testing at 6 months and every 12 months thereafter to detect treatment failure. We assessed trends in CD4 and VL testing in six countries in Southern Africa. Methods: We included adults (≥15 years old) who started ART at one of the HIV treatment programs that participate in the International epidemiology Databases to Evaluate AIDS (IeDEA) Southern Africa region between 2005 and 2017, and had ≥8 months of follow-up time from ART start. We assessed the percent of patients with a CD4 count at ART initiation, the percent with a VL test ≥6 months after ART start and, of those, the percent with virologic failure at the first test ≥6 months after ART start. Virologic failure was defined as VL ≥1000 cells/mm3. The CD4 count at ART start was defined as a CD4 count within a window of 3 months before to 1 week after ART start. Analyses were stratified by sex, age and year of ART start. Results: Our analysis included 520,175 adults from 14 programs in six countries with a median (IQR) age of 34.4 (28.7-41.3) years, of whom 65.0%were female. Median (IQR) follow-up time was 43.6 (23.2-73.0) months and similar across countries. The percent with CD4 testing at ART start has declined over the years from a high of 76.2% in 2005 to a low of 49.4% in 2017. In recent years, the frequency of CD4 testing has also decreased, most notably in Malawi, South Africa and Lesotho (Figure). Women aged 15-24 years had the least CD4 testing (62.5%) and men aged 25-49 years the most (68.3%). Young men aged 15-24 years had the least VL testing (38.4%) and women aged 25-49 years had the most (48.0%). Of those with a VL test, 11.4% had virologic failure with young men aged 15-24 years at greatest risk (19.5%) and women 50+ years at lowest risk (6.2%). Virologic failure has been decreasing in recent years, from 13.7% in 2010 to 8.6% in 2015. Conclusion: CD4 testing at ART start has steadily declined over the years, alongside reduced CD4 testing in general. Virologic failure has been declining; however, without expanded CD4 and VL testing, many patients with advanced disease or with treatment failure may go undetected.

Oral Abstracts

151 HIGH LEVELS OF DRUG RESISTANCE AMONG ART-EXPERIENCED HOSPITALIZED PATIENTS Claire Bossard 1 , Gloria A. Omollo 2 , Patrick Ngimbi Nsuka 3 , Rose Burns 2 ,

Lakshmi Jain 2 , Gisèle Mucinya 3 , Valarie S. Opollo 4 , Stephen S. Wanjala 2 , Gilles van Cutsem 5 , Elisabeth Szumilin 6 , David Maman 7 , Birgitt Schramm 7 , Elisabeth Poulet 7

56

CROI 2019