CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

718 ESTIMATING TB TRANSMISSION IN PRIMARY CARE CLINICS IN TB/HIV HIGH-BURDEN SETTINGS Kathrin Zürcher 1 , Carl Morrow 2 , Julien Riou 1 , Simon Bertschinger 3 , Marie Ballif 1 , Keren Middelkoop 2 , Robin Wood 2 , Matthias Egger 1 , Lukas Fenner 1 , for the IeDEA Consortium 1 Institute of Social and Preventive Medicine, Bern, Switzerland, 2 Desmond Tutu HIV Foundation, Cape Town, South Africa, 3 Bern University of Applied Sciences, Bern, Switzerland Background: Tuberculosis (TB) transmission is difficult to measure and its drivers are not well understood. We piloted a novel approach using clinical, environmental and position-tracking data to study the risk of TB transmission in a primary care clinic in Cape Town, South Africa. Methods: We collected risk factors for airborne transmission during 4 weeks on workdays in August 2019. Patient data included characteristics and number of patients, waiting times and anonymous patient movements using video sensors. Environmental data included indoor carbon dioxide levels (CO2 in parts per million [ppm]), relative humidity (RH, associated with Mycobacterium tuberculosis (Mtb) survival in the air), frequency and intensity of patients’ coughing using sound recording (analyses ongoing), and number of Mtb particles in the air using bio-aerosol sampling devices (molecular detection; analyses ongoing). We calculated rebreathed air volume (RAV) based on people density and CO2 levels (indicating airborne transmission). We defined three areas in the clinic: registration desk (1, see figure), waiting room (2), and TB treatment room (3). Results: 14,795 people visited the clinic. The median number visiting per day was 706 (interquartile range [IQR] 622-803), with a median time of 12.4 min (IQR 11.2-13.7) spent in the waiting room. Density of people was highest in the waiting room (see figure). Overall, the median CO2 level was 623 ppm (IQR 501-751); higher in the morning, compared to midday and afternoon (715 vs. 668 vs. 485; p<0.001). The median RAV was 40 L/day (IQR 18-77); higher in the waiting room compared to the registration area and TB room (69 vs. 26 vs 12 L/ day; p<0.001). The ventilation rate (air change) was relatively high with 11.2 l/h per person (typical value for bedrooms: 5.0l/h per person). The proportion of patients’ time spent above 1,000 ppm CO2 indicating poor ventilation was 10% (typical outdoor value: around 400 ppm). The median RH was above 65% in 32% of time. We are in the process of combining these data with clinical data, cough recordings and the number of Mtb particles in the air to construct a mathematical TB transmission model. Conclusion: This pilot study documents the feasibility of a novel approach to the control of TB in a high-risk transmission setting. Mathematical modelling will allow us to identify factors driving the risk of TB transmission and to evaluate interventions such as separating patient flows or improving ventilation.

PWID had increased TB rate, most likely due to high HIV prevalence in this key population (eg. 27% in Haiphong, 2 million inhabitant city, Vietnam). The record of a high numbers of deaths due to TB during the implementation of large project aiming at ending HIV transmission among PWID in Vietnam, prompted the evaluation of the TB rate in this population. Methods: We implemented a cross-sectional assessment of active TB during a follow-up visit of 2 open cohorts of HIV-negatives and HIV-positives PWID in Hai Phong. Cohort participants were recruited through 2 community-based Respondent-Driven-Sampling surveys carried out at 1 year interval (N=1383 and 1451, respectively). Adult PWID with heroin detected in urine and recent injection skin marks were available. During a cohort follow-up visit, community- based organization (CBO) members systematically assessed TB symptoms using a standardized questionnaire. If any symptomwas recorded, then a Chest X-Ray (CXR) was done at the local TB hospital, followed by a Xpert® MTB/RIF test on sputum if the CXR was abnormal. Results: Among the 581 HIV positives and 672 HIV-negative participants expected, 484 and 457 PWID completed their cohort visit. Overall, 93%were males, their median age was 42 years; 75% and 51%were using methadone, respectively. Among HIV-positives, 90%were on ART and 82% had a viral load < 1000 copies/mL, with a median CD4 count of 472 cells/µL. Among the 451 HIV-positive PWID screened for TB, 293 (65%) had a least one symptom, 84/253 (33%) had an abnormal CXR, and among the 38 who had a Xpert® MTB/ RIF result available, 8 were positive. Assuming all PWID who dropped from the screening cascade had no TB, the conservative TB prevalence was 1.8% [0.6; 3.0]. Very similar figures were found among HIV-negative PWID, with 7 active TB cases for a TB prevalence of 1.6% [0.4; 2.8]. Conclusion: In this high TB burden setting, the active TB prevalence among PWID is more than 10 times higher than the annual TB incidence in the general population, with no increased risk due to HIV. This very high TB rate suggests transmission of M. tuberculosis within PWID. Urgent interventions targeting PWID are required to reach the objective of ending the TB epidemic. 720 PREVALENCE OF TB SYMPTOMS, DIAGNOSIS, AND TREATMENT AMONG HIV PATIENTS NOT ON ART Alana T. Brennan 1 , Mhairi Maskew 2 , Bruce Larson 1 , Isaac Tsikhutsu 3 , Margaret Bii 3 , Lungisile Vezi 2 , Matthew P. Fox 1 , Willem D. Venter 4 , Peter Ehrenkranz 5 , Sydney Rosen 1 1 Boston University, Boston, MA, USA, 2 Health Economics and Epidemiology Research Office, Johannesburg, South Africa, 3Henry M Jackson Foundation, Bethesda, MD, USA, 4 Wits Reproductive Health and HIV Institute, Johannesburg, South Africa, 5 Bill and Melinda Gates Foundation, Seattle, WA, USA Background: Current WHO guidelines recommend that HIV-positive patients who report >1 symptom of tuberculosis (TB) require further investigation for TB disease prior to antiretroviral treatment (ART) initiation. This requirement for ruling out active TB before initiating ART may preclude same-day treatment initiation for many patients who do ultimately not have TB, and, by requiring extra clinic visits, contributes to loss-to-follow-up. We compared the prevalence of TB symptoms, which can delay ART initiation, to the prevalence of TB diagnosis and treatment in intervention arm patients enrolled in the Simplified Algorithm for Treatment Eligibility clinical trials (SLATE I and II) in South Africa and Kenya. Methods: We used intervention arm screening data to describe prevalence of TB symptoms (cough, weight loss, fever, night sweats), diagnosis, and treatment in patients presenting for HIV care not currently on ART in South Africa (n=594) and Kenya (n=240). Data for SLATE I and II in South Africa were combined. Results: 38%(95%CI:32-44%) of patients in Kenya and 41%(37-45%) in South Africa had >1 symptom of TB when presenting for HIV care. 70% of patients in both countries who presented with >1 TB symptomwere tested for TB disease. 13%(7-22%) tested positive for TB in Kenya and 6%(4-10%) tested positive in South Africa. All 27 patients who tested positive for TB disease in both countries reported having >3 symptoms. In both countries, patients with TB symptoms had lower CD4 counts at study enrollment than did those with no symptoms of TB (Kenya: median 152 cells/mm 3 (IQR:64-329) vs. 357(191-632); South Africa: 205(104-391) vs. 351(172-513)). The lowest median CD4 counts were recorded among those with active TB disease (Kenya 124(12-150); South Africa 193(56-223)). Among the 493 asymptomatic patients in SLATE I and II, 4(3%) of patients in Kenya and 151 (44%) of patients in South Africa were tested for TB. One patient tested positive for TB in South Africa and commenced TB treatment; no adverse events (e.g. immune reconstitution inflammatory syndrome) were reported.

Poster Abstracts

719 ALARMING TUBERCULOSIS RATE AMONG PWID IN VIETNAM

Nicolas Nagot 1 , Vu H. Vinh2, Khuat T. Oanh 3 , Delphine Rapoud 1 , Hoang T. Giang 4 , Catherine Quillet 1 , PhamM. Khue 4 , Roselyne Vallo 1 , Thanh T. Nham 3 , Jean-Pierre Molès 1 , Don Des Jarlais 5 , Duong T. Huong 4 , Phuong N. Lan 6 , Thuy T. Dong 7 , Didier Laureillard 8 1 INSERM, Montpellier, France, 2 Viet Tiep Hospital, Hai Phong, Vietnam, 3 Center for Supporting Community Development Initiatives, Hanoi, Vietnam, 4 Hai Phong Medical University, Hai Phong, Vietnam, 5 New York University, New York City, NY, USA, 6 Friends for international Tuberculosis Relief, Gräfelfing, Germany, 7 Friends for international Tuberculosis Relief, Hai Phong, Vietnam, 8 CHU de Nimes, Nimes, France Background: Vietnam belongs to the 30 high TB burden countries according to WHO, with an annual TB incidence of 129/100.000. A few reports suggested that

CROI 2020 264