Low-level HIV viremia during antiretroviral therapy

Abstract: In most cases, antiretroviral therapy (ART) results in undetectable plasma HIV viral load (VL). Still, up to 25% of ART recipients may have detectable low-level viremia (LLV) of different amplitude and persistence. This thesis explores the impact of LLV during ART on virologic and clinical outcomes. Paper I–III are retrospective analyses based on InfCare HIV, a national quality register for people with HIV in Sweden. Time-updated viremia categories were used. Paper I included all participants in Malmö and Gothenburg between 1996 and 2016. Compared with viral suppression, individuals with LLV of 200–999 c/mL, but not 50–199 c/mL, had increased risk of future virologic failure (adjusted hazard ratio [aHR], 3.1; 95% confidence interval [CI], 1.4–7.0). LLV was associated with increased all-cause mortality, although this was not statistically significant in multivariable analysis. For paper II and III, we linked the nationwide InfCare HIV cohort (1996–2017) to national health registers. After 49 986 person-years of follow-up (median 5.7 years), 4177/6956 (60%) were classified as viral suppression, 339 (5%) as LLV of 50–199 c/mL, 258 (4%) as LLV of 200–999 c/mL, and 2182 (31%) as non-suppression. LLV of 50–999 c/mL was associated with increased all-cause mortality when compared with viral suppression (aHR, 2.2; 95% CI, 1.3–3.6). In subanalysis, LLV of 50–199 c/mL had an aHR of 2.2 (95% CI, 1.3–3.8) and LLV of 200–999 c/mL of 2.1 (95% CI, 0.96–4.7). LLV was not associated with AIDS, but individuals with LLV of 200–999 c/mL had increased risk of serious non-AIDS events (SNAE; cardiovascular disease being the most common diagnosis). Neither time-updated viremia category nor cumulative viremia during ART had statistically significant associations with cancer incidence. Higher pre-ART VL was associated with cancer (adjusted subhazard ratio, 1.4; 95% CI, 1.0–1.8). In subanalysis, the association between pre-ART VL and cancer was restricted to AIDS-defining malignancies and infection-related non–AIDS-defining cancer. In paper IV, we measured the levels of nine biomarkers in people with LLV (≥3 VLs in the range 50–999 c/mL) and matched controls with viral suppression. We found no difference in markers of inflammation and immune activation, but patients with LLV had higher levels of growth differentiation factor 15 (GDF-15) and D-dimer. Lastly, we analyzed 21 blood biomarkers and measures of cardiovascular function and structure in participants of a South African research cohort (paper V). We observed similar cardiovascular profiles among individuals with detectable viremia (50–999 c/mL in one measurement) and those with viral suppression (<50 c/mL). In conclusion, this thesis adds to mounting evidence that LLV is associated with inferior clinical outcomes in ART recipients. Specifically, we observed associations between LLV and virologic failure, all-cause mortality, and SNAE, respectively. Our findings suggest that this is likely not mediated through inflammation or immune activation, but elevated GDF-15 and D-dimer for people with LLV in repeated VL measurements could suggest higher cardiovascular risk. We found no evidence of increased risk of cancer or AIDS for people with LLV.