CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

460 PBR-PET IMAGING OF NEUROINFLAMMATION NOT ELEVATED IN HIV+ PARTICIPANTS Anna Boerwinkle 1 , Jeremy Strain 1 , John Doyle 1 , Sarah A. Cooley 1 , Jon Christensen 1 , Yi Su 2 , Julie Wisch 1 , Beau Ances 1 1 Washington University in St Louis, St Louis, MO, USA, 2 Banner Alzheimer’s Institute, Pheonix, AZ, USA Background: Despite combined antiretroviral therapy (cART), HIV associated neurocognitive disorder (HAND) still develops in people living with HIV (PLWH). Persistent neuroinflammation caused by viral reservoirs in the brain is a potential contributor. This study used the positron emission tomography (PET) tracer [11C]-PBR28 (PBR28) to evaluate neuroinflammation in virologically suppressed (<200 copies/mL) PLWH. Methods: 13 HIV- controls and 24 PLWH underwent neuroimaging (magnetic resonance imaging (MRI) and PET) and cognitive testing. Standard uptake value ratios (SUVRs) were calculated for 20 predefined regions of interest (ROIs) affected by HIV. The whole cerebellumwas used as a pseudo-reference region. SUVRs were compared between the two groups using a Wilcoxon Rank-Sum test after correcting for genotype which can affect the tracer’s affinity for TSPO. Within PLWH, additional analyses compared SUVR with clinical markers (current CD4 cell count, nadir CD4, and duration of infection) and cognition (global deficit score (GDS)). Results: SUVRs in the 20 ROIs were not significantly different (p > 0.05) between PLWH and HIV- controls (Table 1). Within PLWH, GDS correlated with SUVR in the superior parietal and supramarginal white matter (p ≤ 0.05); duration of infection correlated with SUVR in the lateral occipital cortex (p ≤ 0.05). After correcting for multiple comparisons, these three correlations were not significant. No association was seen between other clinical measures (current and nadir CD4 cell count) and SUVR for the 20 ROIs. Conclusion: This study reveals no significant increase of neuroinflammation as measured by PBR28 in PLWH compared to HIV- controls. Within PLWH, neither cognitive status nor clinical disease markers correlated with SUVR. Limitations exist for PBR28 and additional studies using magnetic resonance imaging (diffusion basis spectral imaging) and cerebrospinal fluid markers of neuroinflammation need to be performed in virologically suppressed PLWH.

Poster Abstracts

461 PI DRUG-LEVEL TESTING AS A SCREENING TOOL FOR DRUG RESISTANCE IN 2ND-LINE ART FAILURE Lucas E. Hermans 1 , Kim Steegen 2 , Rob ter Heine 3 , Rob Schuurman 1 , Hugo Tempelman 4 , Robert Moraba 4 , Erik Maarseveen 1 , Taryn Pillay 2 , Derryn Legg- Esilva 2 , Jonathan M. Schapiro 5 , David M. Burger 3 , Sergio Carmona 2 , Annemarie Wensing 1 1 University Medical Center Utrecht, Utrecht, Netherlands, 2 National Health Laboratory Service, Johannesburg, South Africa, 3 Radboud University Medical Center, Nijmegen, Netherlands, 4 Ndlovu Care Group, Groblersdal, South Africa, 5 Sheba Medical Center, Tel Hashomer, Israel Background: An increasing number of patients are on 2nd line PI-based ART in low- and middle-income countries (LMIC). In event of virological failure a switch to individualized 3rd line ART is recommended if PI resistance is present. We hypothesize that qualitative PI drug level testing could identify patients most at risk for harboring PI resistance. Methods: We performed a single-centre pilot study followed by a large regional feasibility study in patients with virological failure of LPV/r-based 2nd line ART. In the pilot, LPV level testing on dried blood spots (DBS) was performed by liquid chromatography mass spectrometry (LCMS). In the feasibility study LCMS was performed as a reference and compared to a low-cost qualitative immunoassay (IA; ARK diagn). LPV levels were defined positive or negative based on prespecified limits of detection (LCMS-DBS: 0.25, LCMS-plasma: 0.01, IA-plasma: 0.04 mg/L). Population sequencing of pol was performed. PI-resistance was defined as presence of at least one major IAS-USA listed PI mutation. Results: 548 patients with confirmed LPV/r based ART failure were included (50 pilot, 498 feasibility). Overall, median age was 41.1 years [IQR: 33.5–48.6], 58.8%was female. Median HIV-RNA was 4.9 [4.3–5.4] c/ml. PI-resistance was detected in 12% of patients in the pilot and 27.2% in the feasibility study. Most common mutation profiles were M46I+I54V+L76V+V82A (26.1%) and M46I+I54V+V82A (25.4%). In the pilot, only 40% of patients had a positive LPV level. Sensitivity and negative predictive value (NPV) of a positive LPV level for presence of PI-resistance was 100%, with a specificity of 68%. In the feasibility study, 52.9% of patients had positive LCMS-LPV level and 54.4% had a positive IA-LPV level. Positive LCMS-LPV level had a sensitivity of 89% [95%CI: 83–94], NPV of 94% [90–97], and specificity of 61% [55–66] for PI-resistance. A positive IA-LPV level had a sensitivity of 89% [95%CI: 82–93], NPV of 93% [89–96], and specificity of 58% [53–63] for presence of PI-resistance. Conclusion: In this largest-to-date analysis of PI-based 2nd line failure, non-adherence was objectively demonstrated in half of cases. PI resistance was infrequent, but extensive when present. Negative LPV levels established either

CROI 2019 169