CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

include poor health care access, lack of diagnostics and severe shortage of health care providers. Therefore, it is important to understand differences in the risk of TB between these 2 populations. To address this knowledge gap, we evaluated the risk of TB among HIV+ patients on cART. Methods: We performed a retrospective cohort study on a sample of HIV patients who started cART between 2005 and 2014 within the Zambia National ART Program. We estimated the Incidence Rates (IR) of TB were person-time at risk of TB was accrued from the date of starting cART until diagnosis of TB. To assess the risk factors associated with incident TB, Cox proportion hazard regression was performed. Results: Overall 1,518 patients met the eligibility criteria (rural: 33%; urban:67%). At the time of initiating cART 82 patients (5.4%) were diagnosed with prevalent TB. New cases of TB were diagnosed for 44 patients (2.4%) over 21,209 person-years of observation (PYO). The overall IR was 2.07/1000PYO (95% CI: 1.8–3.7). The IR was 2.6/1000PYO (95% CI: 1.6–4.4) in urban health settings and 1.9/1000PYO (95%: 1.3–2.7) in the rural health settings. Within the first year of cART the IR was 7.6/1000PYO (95% CI: 5.3–10.7), 1.9/1000PYO (95% CI: 0.8– 4.2) in the second year and 0.43/1000PYO (95% CI: 0.2–1.1) after 5 years. In the adjusted analysis, the incidence of TB was not associated with rural/ urban health care setting (aHR =0.9, 95% CI: 0.4 – 1.7) (table 1). As compared to patients with prevalent TB, patients not diagnosed with TB at the start of cART were 90%more likely to be diagnosed with TB during follow up on cART (aHR = 1.9, 95% CI: 1.1 – 2.7). Conclusion: Incidence of TB is substantially high in both rural and urban HIV care settings especially during the first year of cART. HIV treatment programs must develop effective TB screening mechanisms and robust use of isoniazid prophylaxis when TB has been ruled out.

gathering at the clinic. We evaluated if facilities with a larger proportion of clients in CARGs had fewer ART clients initiating TB treatment. Methods: This analysis used data from two six-month time periods: October 2016 to March 2017 and October 2017 to March 2018. The exposure of interest was the proportion of ART clients at each facility who were in CARGs. The outcome was the number of ART clients who started TB treatment in the six-month period, and the number of ART clients at each facility was used as an offset to estimate rates. To evaluate the association, we used a mixed-effects generalized linear model with random effects for each facility, a negative binomial family, log link, and robust standard errors. Results: 181 facilities were included in the analysis. In the earlier 6-month period 2.0% (3,401/170,114) of ART clients were in CARGs compared to 14.6% (28,595/195,443) in the later 6-month period, and 0.6% (2,016/365,557) of ART clients started TB treatment per 6-month period. We found that within any given site, the rate at which ART clients initiated TB treatment when the site had 10-30% of ART clients in CARGs was 0.85-times (95% CI, 0.62-1.15) the rate compared to having <10% of clients in CARG. When any given site had more than 30% of its ART clients in CARGs, it had 0.54-times (95% CI, 0.36-0.79) the rate compared to having <10% of clients in CARGs. This multivariable model adjusted for facility type, facility size, and time period. Conclusion: Sites with a larger proportion of ART clients in CARGs experienced a lower rate of ART clients starting TB treatment. This may reflect a decline in active TB cases among CARG members due to improved ART adherence and/ or reduced TB exposure at clinic waiting areas. However, it is also possible that reduced frequency of clinic visits among CARG members is resulting in undiagnosed TB cases. Community-level TB screening among CARG members may be one solution to address this possibility. 733 INCIDENCE OF TUBERCULOSIS IN THE BOTSWANA NATIONAL ARV PROGRAMME Lucy Mupfumi 1 , Sikhulile Moyo 1 , Qiao G. Michan 2 , Sanghyuk S. Shin 3 , Judith Nawa 4 , Botshelo T. Kgwaadira 4 , Tony Chebani 4 , Thato Iketleng 1 , Tuelo Mogashoa 1 , Rosemary Musonda 1 , Ishmael Kasvosve 1 , Nicola M. Zetola 1 , Simani Gaseitsiwe 1 1 Botswana Harvard AIDS Institute Partnership, Gabarone, Botswana, 2 University of California Irvine, Irvine, CA, USA, 3 Botswana–UPenn Partnership, Gaborone, Botswana, 4 Ministry of Health, Gaborone, Botswana Background: We previously reported a high incidence of TB in a small cohort of HIV-infected patients initiating antiretroviral therapy (ART) and sought to examine whether TB incidence remains high among a national sample of HIV- infected patients receiving ART in Botswana Methods: We analyzed a dataset of 66,382 adult patients (≥18years) who initiated ART between 2011 and 2016. We estimated the incidence and risk factors for TB using Kaplan-Meier survival curves and Cox regression analysis adjusting for gender, age and baseline CD4+ T-cell counts Results: We excluded records from 8098 patients with missing ART initiation dates. Of 58,284 patients, 65%were women with a median age of 37 years (IQR 31-45) and baseline CD4+ T-cell count of 272cells/μl (IQR 146-403). Two thousand and eleven patients developed TB over a median of 1.9 years (IQR 0.6-3.5) of follow-up (IR 2.02 per 100py; 95% CI 1.94-2.11). The risk of TB was greatest in the first 6months of ART (IR 31.36/100py; 95%CI 29.43-33.42) and decreased to 3.08/100py and 0.90/100py by 18 and 36 months post ART initiation respectively. When we excluded cases occurring within 6 months, the overall incidence rate decreased to 1.09/100py (95% CI 1.03-1.17). The risk of TB was high in men, adjusted hazard ratio 2.78 (95% CI 1.84-4.18), and low in those with high baseline hemoglobin levels [aHR 0.83 (95% CI 0.78-0.90)] and formal employment [aHR 0.97 (95% CI 0.67-1.40)]. Conclusion: TB incidence is highest in the first 6 months of ART suggesting a need for active TB case finding and a possible utility for the use of preventative therapy during this time period.

Poster Abstracts

732 COMMUNITY ART REFILL GROUPS AND TUBERCULOSIS RATES IN ZIMBABWE Aaron Bochner 1 , Claudios Muserere 2 , Blessing Wazara 2 , Batsirai Makunike 2 , Gloria Gonese 2 , Elizabeth Meacham 1 , Frances Petracca 2 , Ponesai Nyika 3 , Shirish Balachandra 3 , Ruth Levine 4 , Ann Downer 4 , Stefan Z. Wiktor 4 1 University of Washington, Seattle, WA, USA, 2 I-TECH, Harare, Zimbabwe, 3 CDC, Harare, Zimbabwe, 4 I-TECH, Seattle, WA, USA Background: Community ART Refill Groups (CARGs) are an antiretroviral (ART) differentiated service delivery model in which stable clients on ART form into groups, with a single individual collecting ART for all group members. In focus group discussions, healthcare workers suggested that CARGs may reduce rates of diagnosed tuberculosis (TB) either due to reduced frequency of TB screening by healthcare workers or through reduced transmission with clients no longer

CROI 2019 282