Changes to inflammatory markers during 5 years of viral suppression and during viral blips in people with HIV initiating different integrase inhibitor based regimens

Abstract

Background: Heightened levels of inflammatory markers are linked to increased morbidity/mortality in people with HIV (PWH) and often remain elevated after virologic suppression by antiretroviral therapy (ART). As new combinations of ART become available, an evaluation of their effects on immune activation and inflammation is warranted. Additionally, it remains unknown whether transient increases in viral load ("blips") during ART are associated with increases in inflammation.

Methods: We utilized cryopreserved samples from treatment-naïve PWH enrolled in two Phase 3 clinical trials investigating the efficacy and safety of bictegravir, emtricitabine and tenofovir alafenamide (B/F/TAF) or dolutegravir, abacavir, and lamivudine (DTG/ABC/3TC) or DTG + F/TAF over a 5-year window (GS-US-380-1489/1490). At week 144, participants were offered the option to switch to open label B/F/TAF for an additional 96 weeks. We measured levels of interleukin-6 (IL-6), C-reactive protein (hsCRP), D-dimer, soluble CD14 (sCD14), and tumor necrosis factor-α receptor 1 (TNFR1) from available baseline, week 24, 48, 144, and 240 samples (B/F/TAF, N=123; DTG/ABC/3TC, N=62; DTG+F/TAF, N=58). Additional samples from PWH who experienced a viral blip (n=44, defined as a single HIV-1 RNA >50c/mL) were also analyzed and paired with the most recent available suppressed sample before the blip. Longitudinal biomarker changes were assessed using a constrained mixed effects linear regression model adjusting for covariates.

Results: Baseline demographics and selected laboratory characteristics were similar across groups. Levels of D-dimer, sCD14, and TNFR1 decreased significantly from baseline in all treatment arms, with no significant differences between arms at any timepoint. Biomarker levels also remained stable following ART-switch at week 144. No significant changes in hsCRP or IL-6 were observed versus baseline in any arm at any timepoint. A significant association was observed between sCD14 and increasing viral load (p=0.022) in viral blips; D-dimer also increased with blips in the B/F/TAF arm.

Conclusions: Viral suppression was associated with reductions in most inflammatory markers in PWH, with no significant differences among the three ART regimens during the 144-week randomized period. These decreases were sustained after the open label switch to B/F/TAF. Viral blips were associated with increases in monocyte activation (sCD14). Further analysis is needed to confirm these findings and determine the potential impact on clinical outcomes.

Keywords: HIV-1; antiretroviral therapy; inflammation; intermittent viremia; monocyte activation.

Figure 1

GS-US-380-1489/90 clinical study design and plan for retrospective biomarker analysis. We utilized cryopreserved samples from a sampling of treatment-naïve PWH in the United States who are enrolled in two Phase 3 clinical studies investigating the efficacy and safety of B/F/TAF, DTG/ABC/3TC and DTG+F/TAF over a 5-year window. Samples from sustained VS participants with baseline viral loads (VL) > 100,000 cp/mL were over selected. Participants with an underlying comorbid condition that complicated their inflammatory profiles were excluded.

Figure 2

Significant decreases in biomarkers observed across all ART treatment arms in PWH. Longitudinal changes in serum biomarkers associated with inflammation (Least-square means ± 95%CI; statistical differences against baseline and between Week 144 and Week 240 were determined from log2 transformed data and a constrained mixed effects linear regression model adjusted for sex and baseline viral load). Levels of (A) sCD14, (B) D-Dimer, (C) TNFR1, (D) hsCRP, and (E) IL-6 were measured by enzyme linked immunosorbent assay (ELISA). Black borders on the symbols denote a significant difference from baseline at a false discovery rate (FDR) < 0.05.

Figure 3

Changes in inflammatory markers were associated with increases in CD4+ T cell percentages. Association of changes in CD4+ T cell percentages and inflammatory biomarker levels derived from a linear mixed effect model adjusting for difference to baseline viral load, treatment arm, and time. (A) CD4+ percentage increases overtime in each treatment arm. (B) Changes in CD4+ T cell percentages (least-square means ± 95%CI) in relation to levels IL-6, D-Dimer, TNFR1, sCD14, or hsCRP that were measured by enzyme linked immunosorbent assay (ELISA).

Figure 4

Transient viral elevations (Blips) were associated with increases in sCD14 and D-dimer levels. Data from PWH who experienced a viral blip (n=44, defined as single 50c/mL> HIV-1 RNA < 1000c/mL) during treatment were also analyzed and paired with the most recent suppressed sample before the blip. Participants with viral blips in each arm were included (B/F/TAF, n=23, DTG/ABC/3TC, n=13, and DTG+F/TAF, n=8). Associations between inflammation biomarkers and viral load were tested with a B-spline LMEM adjusting for baseline viral load, treatment arm, and time. Least-squares means and confidence intervals shown for (A) sCD14 and median (IQR) shown for (B) D-Dimer, with p-values tested by Wilcoxon Signed Rank Test.