CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

1112 HIGH HIV-RELATED TUBERCULOSIS RISK IN COUNTIES WITH LOW HIV PREVALENCE− KENYA, 2015 Herman O. Weyenga 1 , Abraham Katana 1 , Thomas Achia 1 , Brian Baker 1 , Lucy Ng’ang’a 1 , Evelyne Ngugi 1 , Tai Ho Chen 1 , Maureen Kamene 2 , Emily C. Zielinski- Gutierrez 1 , Enos Masini 2 , Dunstan Achwoka 1 , George K. Kadondi 2 , Christine Wambugu 2 , Kenneth Masamaro 1 , Kevin M. De Cock 1 1 US CDC Nairobi, Nairobi, Kenya, 2 Ministry of Health, Nairobi, Kenya Background: Infection with human immune deficiency virus (HIV) is an independent risk factor for tuberculosis (TB). Antiretroviral therapy (ART) reduces TB risk and mortality at both patient and population levels. Kenya has geographically differentiated HIV and TB epidemics and control efforts have to-date been focused on the counties with the highest HIV prevalence. However, the risk of people living with HIV (PLHIV) developing and dying from TB in different Kenyan counties might better inform TB/HIV prevention efforts. Methods: We conducted a retrospective analysis of 2015 data from the national TB surveillance database (TIBU) and Ministry of Health HIV estimates. Variables included county smear positive, TBHIV and total TB case notification (CN), population estimates, estimated HIV prevalence, number of PLHIV (including children), ART coverage and poverty index. Eleven counties with HIV prevalence >5.6% (national estimate) were defined as high HIV prevalence and the rest as low HIV prevalence. Using MS excel and STATA statistical software, county CN rate for TBHIV and deaths were computed. Poisson regression models with and without spatially-structured and unstructured random effects were fitted controlling for potential confounders. Deviance information criterion was used to identify the best fitting model. Correlation analysis was done to determine the association between TB risk and death among PLHIV and HIV prevalence. Results were presented as counts, CN rate, relative risk and Pearson correlation coefficient. Results: Overall, of 82,406 reported TB cases and 1,517,707 estimated PLHIV in all 47 counties 26,615 were co-infected. Case fatality rates among HIV-positive and HIV-negative TB cases were 10.4% and 3.4%, respectively. Fifty-three percent of deaths among persons with TB/HIV were in high HIV prevalence counties. The median TBHIV CN rate was 1,866 (IQR, 1,538-2,606) per 100000 PLHIV. TB risk among PLHIV correlated negatively with HIV prevalence (Spearman’s r=-0.40, p=0.005). Low HIV prevalence counties had higher TB risk among PLHIV when compared to counties with high HIV prevalence, RR=1.56, 95% CI (1.26–1.97). The TBHIV case fatality rate was not associated with HIV burden, smear positive TB rate, ART coverage or poverty index. Conclusion: PLHIV in low HIV prevalence counties have a higher relative risk of TB than PLHIV in high HIV prevalence counties in Kenya. Resource allocation for comprehensive HIVTB prevention in low burden counties is recommended.

1111LBSAME-DAY ART INITIATION ASSOCIATED WITH POORER RETENTION BUT HIGHER VIRAL SUPPRESSION Dvora Joseph Davey , Claire Serrao, Madaline Feinberg, Marlien Prins, Ernest Darkoh BroadReach Corporation, Cape Town, South Africa Background: South Africa rolled out universal HIV testing and treatment in September 2016 including ART initiation on the same-day as HIV diagnosis. Apart from the anticipated improvement in ART uptake, it is essential to understand the impact of same-day ART initiation on loss to follow-up (LTFU), as well as on viral suppression (VS) and mortality. Methods: We analysed national data from five high HIV prevalence Provinces in South Africa to evaluate the impact of same day ART initiation on treatment outcomes including LTFU, VS (<400 copies/mL) and mortality in adults on ART for >90-days who initiated on ART between September 2016 and August 2017 who had a diagnosis date on file using multivariable regression models controlling for a priori confounders (e.g. age, gender, baseline CD4, District, time on ART) censoring patients who transferred or moved. Results: In total 20,847 patients >15-years old initiated ART and were on ART for >90-days. Overall, 2013 patients initiated ART on the same day as diagnosis (11.6%) of which 725 (36%) were pregnant at ART start. Median baseline CD4 cell count was higher in same-day ART patients (386, IQR=228-560), vs. other patients (323, IQR=163-529; z=-13.8; p<0.001). LTFU was higher in the same-day ART patients (22% vs. 15%; p<0.001). Adjusting for age, time on ART, baseline CD4, District and gender, odds of LTFU at >6-months on ART were 1.5-times higher in patients with same-day ART (aOR=1.45; 95%CI=1.18-1.78). Stratifying by pregnancy status, odds of LTFU were 1.9-times higher in pregnant women (aOR=1.86; 95% CI=1.23-2.80) and 1.4-times higher in non-pregnant adults (aOR=1.41; 95%CI=1.06-1.86) vs. non-same day initiators. However, mortality was lower in the same-day ART group (n=7 [0.5%] vs. n=182 [1.2%]; aOR=0.46;95%CI=0.26-0.82). In patients with a viral load done (n=17,704; 85%), VS was higher in the same-day ART group (87% vs. 85% of patients with a viral load done, p=0.02). Adjusting for the same factors, same-day ART initiators had greater odds of achieving VS (aOR=1.16; 95% CI=1.02-1.33). Conclusion: This is one of the first studies to demonstrate treatment outcomes in a large cohort of patients who initiated ART the same day as diagnosis. Same-day ART initiation was associated with poorer retention, but was associated with lower mortality and viremia compared with non-same-day patients. Additional research on targeted interventions to improve counselling and ART readiness in patients who initiate ART on the same-day as their diagnosis are needed.

Poster Abstracts

1113 BURSTING MYTHS: PROGRAMMATIC SCALEUP OF ISONIAZID PREVENTIVE THERAPY, KENYA 2014-2016

Evelyn M. Karanja 1 , Herman O. Weyenga 2 , Abraham Katana 2 , Martin Sirengo 1 , Enos Masini 3 , Lucy Ng’ang’a 2 , Evelyne Ngugi 2 , Shobha Vakil 1 , Caroline Olwande 1 , Agnes Langat 2 , Christine Wambugu 3 , Brian Baker 4 , Kevin M. De Cock 2 1 National AIDS and STD Control Programme, Nairobi, Kenya, 2 US CDC Nairobi, Nairobi, Kenya, 3 Ministry of Health, Nairobi, Kenya, 4 US CDC Kisumu, Kisumu, Kenya Background: Infecting a third of the global population, M tuberculosis has a large pool of individuals fromwhommillions of active tuberculosis(TB) cases emerge annually. HIV infection is associated with a tenfold life-time risk of

CROI 2018 428