CROI 2025 Abstract eBook

Abstract eBook

Poster Abstracts

966

Diagnosis and Prognosis of Blood XpertUltra and Myco/F Blood Culture for TB in Advanced HIV Disease Phuong Le Trinh 1 , Dieu Quang Nguyen 1 , Thu Dang Anh Do 1 , Emily Evans 2 , Ly Trieu Vo 3 , Dung Thanh Nguyen 4 , Thach Ngoc Pham 5 , Quang Le Nguyen 1 , Khanh Hoang Dang 1 , Trang Van Dinh 5 , Cuong Do Duy 6 , Dat Quoc Vu 7 , Thuong T. T. Nguyen 8 , H. Rogier van Doorn 8 , Thuy Le 9 1 Oxford University Clinical Research Unit in Vietnam, Hanoi, Vietnam, 2 Emory University, Atlanta, GA, USA, 3 University of Medicine and Pharmacy at Ho Chi Minh City, Ho Chi Minh City, Vietnam, 4 Hospital for Tropical Diseases, Ho Chi Minh City, Vietnam, 5 National Hospital for Tropical Diseases, Hanoi, Vietnam, 6 Bach Mai Hospital, Hanoi, Vietnam, 7 Hanoi Medical University, Hanoi, Vietnam, 8 University of Oxford, Oxford, UK, 9 Duke University School of Medicine, Durham, NC, USA Background: Current TB diagnostics are suboptimal to detect TB in patients with advanced HIV disease (AHD) who are at high risk for disseminated TB and mortality. Blood has been overlooked as a compartment for TB diagnostics in AHD. We determined the diagnostic yield of routine screening for TB in blood using Xpert Ultra (UXpert) and Myco/F Lytic (MFL) culture, and the utility of blood UXpert cycle threshold (Ct) and time to positivity (TTP) of MFL blood culture in predicting mortality. Methods: We prospectively recruited adults with AHD (CD4 ≤100 or WHO stage 3 or 4) who were hospitalized for any reason at 3 hospitals in Vietnam. In addition to standard TB diagnostics (microscopy, UXpert and MGIT culture of non-blood specimens performed when clinically indicated), we screened all patients for TB in blood using UXpert and MFL culture. Patients were followed monthly over 6 months to determine TB outcome. We assessed the association of blood UXpert Ct value of the IS6110-1081 gene (Ct IS ) and TTP of MFL blood culture on 24-week mortality using Cox hazard models. Results: Among 901 patients recruited between February 2021 and July 2023, 72 patients on active TB therapy were excluded. Among 829 patients, 83% were men. Median age was 36 years (IQR: 29-44). Median CD4 cell count was 18 (IQR: 8-43). 208 of 829 patients (25.1%) had microbiologically confirmed TB during the index hospitalization, and 10 developed TB during follow-up. Of these 208 TB patients, blood UXpert and MFL blood culture identified 151 cases (72.6%). This included 62 (30%) additional TB cases that were missed by the standard non-blood approach (Fig. 1A). Blood UXpert alone identified TB (and MDR TB) within 3 hours, enabling rapid initiation of TB therapy in 102 (49%) of 208 TB patients. The Cox models adjusted for sex, age, CD4 count identified the following independent predictors of 24-week mortality: (1) Positive TB in blood (mortality 32% vs. 12%, HR: 3.1, 95% CI:1.4-7, p <0.01) (Fig. 1B); (2) Blood UXpert Ct IS ≤21 (mortality 50% vs. 24%, HR: 2.5, 95% CI: 1.2 – 5.1, p=0.016); (3) Faster TTP MFL blood culture (by one day) (HR: 1.06, 95% CI: 1.02–1.11, p= 0.01). Conclusions: Among hospitalized patients with AHD, screening for TB in blood using currently available diagnostics of blood UXpert and MFL blood culture increased TB diagnostic yield by 30%. UXpert in blood alone made a rapid diagnosis of disseminated TB in 49% of TB patients, allowing rapid treatment and prognosis of patients. Gaps in Latent Tuberculosis Infection Testing Practices: A 15-Year Serial Cross-Sectional Study Jorge R. Ledesma 1 , Yuching X. Ni 1 , Jacek Skarbinski 2 1 Kaiser Permanente Division of Research, Oakland, CA, USA, 2 Kaiser Permanente Oakland Medical Center, Oakland, CA, USA Background: In the US, untreated latent tuberculosis infection (LTBI) is the primary driver of new tuberculosis (TB) cases. Tracking real-world levels and trends in LTBI testing is therefore critical for assessing TB elimination efforts. However, many studies suggest that LTBI testing among people with high TB risk is suboptimal. To evaluate this practice gap, we conducted a serial cross-sectional study using 15 years of electronic health record data from an integrated health system to assess trends in LTBI testing. Methods: In this serial cross-sectional study we included all Kaiser Permanente Northern California (KPNC) members aged ≥18 years with at least 6 months of continuous enrollment between January 2008 and December 2023. The primary outcome was LTBI testing with either a tuberculin skin test (TST) or an interferon gamma release assay (IGRA). The secondary outcome was ever receiving an IGRA among persons tested for LTBI. Using multivariable modified Poisson regression, we assessed factors associated with testing, including TB risk factors (born in TB-endemic country, immunosuppressed, close contact with TB, homelessness) specified in US CDC and California-specific screening guidelines. The figure, table, or graphic for this abstract has been removed.

Results: Among 6,572,839 persons in the study, 1,405,768 (21.4%) were tested for LTBI; 76% of all testing during the study period was among people without any TB risk factors. Of all people tested, 373,659 (26.5%) received an IGRA. The overall LTBI testing rate did not increase substantially during the study period: the rate per 100 population was 5.2 in 2008 and 6.7 in 2023. People born in a TB-endemic country had the lowest rate of LTBI testing with a rate of 3.3 in 2008 and 5.0 in 2023 (Figure). Persons born in a TB-endemic country had a 25.2% (24.8-25.5%) lower prevalence of LTBI testing compared to US-born persons. Persons with TB risk factors including close contact with TB case (PR=2.66 [2.62-2.70]), homelessness (PR=1.29 [1.26-1.32]), and HIV (PR=3.58 [3.53-3.63]) were more likely to receive LTBI testing. Among persons who received any LTBI testing, persons born in TB-endemic countries were 61.7% (60.5-62.9%) more likely to have been tested with an IGRA than US-born persons. Conclusions: We identified a substantial gap between evidence and practice. Despite data indicating that 82% of all TB cases in California are among persons born in TB-endemic countries, current LTBI testing practices are disproportionately undertesting this population while over testing persons without any TB risk factors. Effect of Alcohol Use on Active TB Incidence Among PWH With Prior Receipt of TB Preventive Therapy Judith A. Hahn 1 , Robin Fatch 1 , Sowmya R. Rao 2 , Adah Tumwegamire 3 , Christine Ngabirano 3 , Julian Adong 3 , Nneka Emenyonu 1 , Gabriel Chamie 1 , Winnie Muyindike 3 , for the International URBAN ARCH Center 1 University of California San Francisco, San Francisco, CA, USA, 2 Boston University, Boston, MA, USA, 3 Mbarara University of Science and Technology, Mbarara, Uganda Background: High-risk alcohol use is associated with increased risk for active TB, the leading cause of death among persons with HIV (PWH). We hypothesized that the incidence rate of active TB after receipt of TB Preventive Therapy (TPT) would differ by level of alcohol use among PWH. Methods: We conducted a review of HIV/TB clinic charts to find active TB diagnosis among 988 PWH with prior positive tuberculin skin tests in a sample enriched for persons with high-risk alcohol use in South-western Uganda. These persons engaged in our prior research and received 6 months of TPT, with electronic pill-cap monitoring. We used person-time methods and defined time at risk as months from 6-months after TPT start to date (t0) until TB diagnosis, last known TB-free clinic or study visit, or death not known to be due to TB. We conducted Cox modeling to obtain the hazards ratio (HR) and 95% confidence intervals (CI) of high-risk alcohol use at t0, defined as Alcohol Use Disorders Identification Test – Consumption (AUDIT-C, modified to cover 3 months) ≥6 or the alcohol biomarker, phosphatidylethanol (PEth) ≥200 ng/mL, versus not high-risk, adjusting for gender and age. We further adjusted for the following variables if they were individually associated (in models with alcohol use, gender, and age) with active TB (p<0.10): HIV viral non-suppression (≥200 copies/ml), ART regimen, CD4 count, smoking, TPT non-completion (<90% pills taken), body mass index, and number of people/household room. Results: 968 (98%) persons had total follow-up of 3229 person-years (py). There were 21 cases of TB, confirmed by Gene Expert (n=13) or lipoarabinomannan (LAM, n=8), for a TB incidence rate of 0.65% (95% CI: 0.42-1.00%) per py (ppy). 410 (42.4%) had high-risk alcohol use, 634 (65.5%) completed TPT, and 32 (3.3%) were HIV non-suppressed. The TB incidence rate was 1.28% (95% CI: 0.79-2.09%) ppy among those with high-risk alcohol use, and 0.25% (95% CI: 0.11-0.61%) ppy among those without high-risk alcohol use (Figure). The adjusted hazard ratio (aHR) for high-risk alcohol use yes versus no was 3.74 (95% CI: 1.29-10.79). Cigarette smoking (aHR: 2.42, 95% CI: 0.99-5.92) and HIV non-suppression (aHR: 5.58, 95% CI: 1.64-18.98) were also independent predictors of active TB. There was no interaction by gender. Conclusions: These results suggest ongoing TB acquisition and/or reactivation despite high rates of viral suppression and receipt of TPT among PWH, particularly among those with high-risk alcohol use and cigarette smoking. The figure, table, or graphic for this abstract has been removed.

968

Poster Abstracts

967

CROI 2025 305