CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

the IMMY CrAg LFA and the novel SQ-CrAg LFA. The sensitivity and specificity of the SQ-CrAg in the reflex CrAg screening cohort were determined relative to the standard CrAg LFA. To further validate the SQ assay known CrAg+ EDTA blood samples from a prior CrAg-screening study and a CM treatment trial were also tested with both assays. Serial dilutions were performed for all CrAg+ samples and re-tested with the standard LFA to determine titres. SQ titers and conventional titers were compared. All testing was performed by two independent blinded investigators and inter-rater reliability assessed using the Kappa coefficient. Results: 692 sequential samples were screened using both assays; 43 (6.2%) were IMMY CrAg LFA-positive. Using this standard CrAg LFA as a reference, the overall sensitivity and specificity of the novel SQ-CrAg LFA were 93.0% (95%CI 80.9 – 98.5%) and 93.8% (95%CI 91.7-95.6%) respectively. A further 180 known CrAg+ samples were tested and the combined results used to evaluate the SQ-CrAg quantification. Median (IQR) CrAg titers for SQ-CrAg 1+, 2+, 3+, and 4+ bands were 1:10 (1:5 – 1:20), 1:40 (1:20 – 1:80), 1:640 (1:160 – 1:2560), and 1:5120 (1:2560 – 1:20480) respectively (Figure 1). Inter-rater agreement in titer assessment was excellent at 98.2%, with a kappa coefficient of 0.96, p<0.001. Conclusion: Overall sensitivity and specificity of the novel IMMY SQ-CrAg assay were high in a cohort of HIV-positive individuals with CD4 counts ≤200 cells/ml undergoing reflex CrAG screening. An SQ titre of 3+ or greater corresponded to a titer of >1:160 which has previously been shown to be associated with increased mortality.

of 169 (IR: 17.8 per 100 person-years) CrAg- participants developed cryptococcal meningitis or died for a rate of death or cryptococcal meningitis that was 6.3- times higher for those who were CrAg+ (CI: 2.7-14.6). Conclusion: Although few PLHIV with moderate immunosuppression screened CrAg positive, a positive CrAg test was predictive of increased risk of cryptococcal meningitis or death. Systematic CrAg screening may reduce morbidity and mortality in PLHIV with CD4 100-200 cells/mm 3 .

Poster Abstracts

746 HIGH RATES OF MENINGITIS OR MORTALITY AMONG CrAg+ PLHIV WITH CD4 100-200 CELLS/MM³ James H. Wykowski 1 , Paul K. Drain 1 , Sean Galagan 1 , Sabina M. Govere 2 , Connie L. Celum 1 , Mahomed-Yunus Moosa 3 , Carole Wallis 4 1 University of Washington, Seattle, WA, USA, 2 AIDS Healthcare Foundation, Durban, South Africa, 3 University of KwaZulu-Natal, Durban, South Africa, 4 Lancet Labs and BARC SA, Johannesburg, South Africa Background: Cryptococcal antigen (CrAg) screening with fluconazole prophylaxis has been shown to prevent cryptococcal meningitis and mortality for people living with HIV (PLHIV) with CD4 <100 cells/mm 3 . While cryptococcal meningitis occurs in individuals with CD4 100-200 cells/mm 3 , there is limited evidence that CrAg screening predicts cryptococcal meningitis or mortality among this group with moderate immunosuppression. Current IDSA and WHO clinical guidelines recommend restricting CrAg screening to PLHIV with CD4 <100 cells/mm 3 . Methods: We conducted a prospective cohort study of PLHIV >=18 years who had not initiated ART in South Africa. We followed participants for 14 months to determine onset of cryptococcal meningitis or all-cause mortality. At study completion, we retrospectively tested stored serum samples for CrAg using an enzyme immunoassay (EIA). We calculated CD4-stratified incidence rates of outcomes and used Cox proportional hazards to measure associations between CrAg positivity and outcomes. Results: We enrolled 2,383 PLHIV, and 1,309 participants and had serum samples tested by CrAg EIA. The median CD4 was 317 cells/mm 3 (interquartile range: 173-491 cells/mm 3 . By CD4 count at baseline, there were 209 individuals with a CD4 count of 100-200 cells/mm 3 with available CD4 test results and four (1.9%) tested positive. Among this group, two of four (IR: 58.8 per 100 person-years) CrAg+ participants and 11 of 205 (IR: 5.6 per 100 person-years) CrAg- participants developed cryptococcal meningitis or died for an overall rate of death or cryptococcal meningitis that was 10.0-times higher for those who were CrAg+ (CI: 2.2-45.3) (figure). Among those with CD4 <100 cell/mm 3 and CrAG EIA test results (n=179), ten (5.6%) participants tested CrAg+. Among this group, seven of ten (IR: 137.6 per 100 person-years) CrAg+ participants and 26

747 CSF CYTOKINES AND CHEMOKINES ASSOCIATED WITH MORTALITY IN CRYPTOCOCCAL MENINGITIS Elizabeth C. Okafor 1 , Liliane Mukaremera 1 , Nicole Wyman Engen 1 , Katherine Huppler Hullsiek 1 , Lillian Tugume 2 , Kenneth Ssebambulidde 2 , Abdu Musubire 2 , Edwin Nuwagira 3 , Edward Mpoza 2 , Darlisha A. Williams 1 , Conrad Muzoora 3 , David Meya 2 , Joshua Rhein 1 , David R. Boulware 1 , for the ASTRO-CM Team 1 University of Minnesota, Minneapolis, MN, USA, 2 Infectious Disease Institute, Kampala, Uganda, 3 Mbarara University of Science and Technology, Mbarara, Uganda Background: Cryptococcal meningitis causes substantial mortality globally. Our understanding of the role of the host immune system in patient outcomes is limited. We investigated the cytokine and chemokine environment at the site of infection, the CNS, to better understand the impact of immune cell activation and tissue inflammation on mortality. Methods: We prospectively enrolled Ugandans presenting with first episode cryptococcal meningitis fromMarch 2015 to May 2017, as part of a larger study focused on drug treatment with amphotericin + fluconazole +/- sertraline to improve neurological outcomes. We analyzed the CSF of 321 subjects at diagnosis for soluble biomarkers of immune cell activation utilizing a luminex assay. Statistical analysis grouped each biomarker into quartiles (Q1, Q2+Q3, Q4) and compared each group for 14-day mortality via logistic regression, adjusted for Glasgow Coma Scale and CSF quantitative culture. We compared Q1 (low) and Q4 (high) to the reference Q2+Q3 group. Results: Participants with Q1 (low) levels of markers indicative of cytotoxic cell function such as TRAIL (p=0.004),Granzyme-B (p=0.03), and IP-10 (p=0.007) had significantly increased risk of 14-day mortality compared to middle two quartiles (Q2+Q3) reference group levels. Participants with Q1 (low) levels of markers associated with naïve T cell activation and recruitment such as CXCL2 (p=0.003), PDL1 (p=0.013), and CCL19 (p=0.013) had increased risk of 14-day mortality while those with Q4 (high) levels of CCL19 (p=0.009) had decreased mortality , when compared to the Q2+Q3 ref group. Inflammatory mediators such as TNF-alpha, IFN-gamma, IL-6, and IL-1beta were not associated with

CROI 2020 275