CROI 2017 Abstract e-Book

Abstract eBook

Poster and Themed Discussion Abstracts

Methods: We selected from the Dat’AIDS cohort patients with a CD4 T cell count < 200/mm3 starting from 01/01/2010 to 01/01/2015 ART based on 2 NRTIs associated with a bPI, a NNRTI or an INSTI, and admitted to hospital within 6 months. IRIS events were defined as symptoms consistent with an infectious or inflammatory condition associated with a drop of > 2 log10 copies/mL of HIV viral load, not explained by a newly acquired infection, the expected clinical course of a previous infection, or side-effects. Three physicians blinded to the ART regimen evaluated files and determined the classification by consensus. Characteristics associated with IRIS were analyzed in uni-and multivariate analysis. Results: The study population included 2287 patients from 15 centers in France. Median age was 45 years (IQ25-75 37-53), and 63%were men. The third agent was bPI in 65%, NNTI in 12%, and INSTI in 12%. At ART initiation, the median HIV viral load and CD4 T cell count were 5.2 log10 copies/mL (4,8-5,7) and 83/mm3 (31-146). IRIS occurred in 41 patients (1.8%) and was associated with tuberculosis (12 cases), atypical mycobacteria (10), JC virus (6), CMV (5), HHV-8 (4), Toxoplasma (2), Cryptococcus (1) and HBV (1). Patients receiving INSTI-based ART did not differ from those without INSTI regarding pre-ART HIV viral load and CD4 T cell count (table). IRIS occurred in 12/398 (3%) patients receiving INSTI-based ART, compared to 29/1889 (1.5%) patients without INSTI (OR 1.99 (1.1-3.5), p=0.04). Repartition of opportunistic infections did not differ according to ART regimen. Conclusion: Starting ART based on INSTI in exposed patients is associated with a higher risk of IRIS. The homogenous repartition of opportunistic infections among regimen groups argued against a bias of indication linked to mycobacterial infection and co-medication used, although we cannot preclude it formally. While effective control of HIV replication is key, initiation of ART not based on INSTI might be wise in selected patients at high risk of IRIS, deserving further studies. 733 INFLAMMASOMES PLAY A MAJOR ROLE IN TB IRIS PATHOGENESIS Denise C. Hsu 1 , Daniel Sturdevant 2 , Eleanor Wilson 1 , Virginia Sheikh 1 , Ornella Sortino 3 , Silvia Lage 1 , Timothy G. Myers 1 , Stephen Porcella 2 , Irini Sereti 1 1 NIAID, Bethesda, MD, USA, 2 NIAID, Hamilton, MT, USA, 3 Leidos Biomed Rsr Inc, Frederick, MD, USA Background: Tuberculosis (TB) immune reconstitution inflammatory syndrome (IRIS) affects 20% of patients with TB and HIV after antiretroviral therapy (ART) initiation. It is associated with significant morbidity and its pathogenesis is still not fully elucidated. We identified differentially expressed RNA transcripts between TB/HIV co-infected patients with and without IRIS. Methods: Subjects in this cohort study (NCT01611402) were TB/HIV co-infected and had been on TB treatment and ART. Blood was collected in PAXgene tubes at the time of TB-IRIS or, for those without IRIS, at 8-12 weeks after ART initiation, then cryopreserved. RNA was extracted and hybridized on the Affymetrix GeneChip® Human Gene 2.0 ST array. The chip was then stained and scanned. Expression of transcripts was compared using ANOVA with multiple test correction using Benjamini–Hochberg false discovery rate method. Ingenuity Pathway Analysis (IPA) was used to identify canonical pathways and upstream regulators. Results: Patients with (n=6), and without (n=10) TB-IRIS were enrolled. Patient characteristics are listed in Table 1. Patients with TB-IRIS had higher plasma HIV viral load (p=0.01), C-reactive protein (p=0.02) and absolute neutrophil count (p<0.001) and lower hemoglobin (p<0.001). Using p<0.05 as cut off, compared with patients with no IRIS, patients with IRIS had 355 differentially expressed transcripts, of which 167 had ≥±2 fold difference (148 up and 19 down regulated). IPA showed that innate antigen recognition signaling pathways including inflammasome (p=4.27e-6), TREM1 (p=2.70e-5) and Toll-Like Receptor signaling (p=4.13e-5) were the top over-represented pathways in TB IRIS. Upstream regulator analysis predicted Transglutaminase 2 (activated, z-score 2.8, p=9.06e-6), interferon-Υ (activated, z-score 2.3, p=1.22e-4) and interferon regulatory factor 4 (inhibited, z-score -2.0, p=1.63e-4) to be potential upstream regulators of the aforementioned pathways and that the activation of these pathways would lead to an increase in caspase 1 breakdown of pro-interleukin (IL) 1b to IL-1b, resulting in a pro-inflammatory response. Conclusion: Using RNA microarray analysis, transcripts of innate antigen recognition and signaling pathways were over expressed and likely play a crucial role in TB-IRIS. The use of agents that block inflammasome and IL-1b signaling in recalcitrant TB-IRIS merit further investigation.

Poster and Themed Discussion Abstracts

734

THE DIABETES-TUBERCULOSIS COMORBIDITY AMONG PERSONS WITH HIV Monica Sane Schepisi 1 , Jalal Miah 2 , Anna Kaluzhenina 3 , Katerina Manika Manika 4 , Emanuele Pontali 5 , Monica Rios Prego 6 , María Nieves Altet Altet 7 , Diana Sanchez Mellado Sanchez Mellado 8 , Enrico Girardi 1 , for theTB NET 1 Natl Inst for Infectious Diseases L. Spallanzani, Rome, Italy, 2 Newham Univ Hosp Barts NHS Trust, London, UK, 3 Volgograd State Med Univ, Volgograd, Russian Federation, 4 Aristotle Univ of Thessaloniki, Thessaloniki, Greece, 5 Galliera Hosp, Genova, Italy, 6 EComplexo Hospario Univ de Pontevedra, Pontevedra, Spain, 7 Unidad de Prevención y Control de la TB de Barcelona, Barcelona, Spain, 8 IIS-Fundacion Jimenez Diaz, UAM, Madrid, Spain Background: Both HIV and Diabetes mellitus (DM) increase the risk of developing Tuberculosis (TB). While many studies addressed the relationship between HIV infection the risk of DM, the relationship between HIV infection and the co-morbidity tuberculosis-diabetes is still poorly explored. Methods: We performed a retrospective study on 2395 adult patients consecutively diagnosed with TB in 7 European clinical centres (in Italy, Greece, Spain, Russia); patients included in this analysis had either a diagnosis of DM or sufficient clinical or laboratory information to define the presence or to exclude DM. Two distinct multivariable models were fitted to analyze: 1. the association between HIV and the co-morbidity tuberculosis-diabetes among all patients with TB, and 2. the effect of DM on clinical and radiological presentation of TB among HIV TB patients. Results: Of the patients enrolled 292 (12.2%) were infected with HIV. HIV-TB co-infected compared to TB only patients were younger, more likely to be male, to be autochthonous and to have other co-morbidities. The overall prevalence of diabetes among the 2395 TB patients was 7.7%. DM prevalence amongst TB only patients and amongst TB-HIV co- infected patients were respectively 7.8% (164/2103) and 7.2% (21/292). After adjustment for potential confounders –age, gender, country of birth- there was no evidence of a significant association between HIV infection and the DM. Amongst 292 TB-HIV co-infected patients, DM was associated with cavities and night sweating, after adjustment for last CD4+ cell count, HAART, gender and age (table). Other presenting symptoms –persisting cough, weight loss, haemoptysis, fever- and clinical variables explored -TB site, sputum smear and sputum culture- were not significantly associated with DM.

CROI 2017 320