Lanz et al. aimed to investigate the impact and clinical significance of low-level viremia (LLV), accounting for its dynamic nature on the highly granular longitudinal data set from the SHCS spanning >2 decades, enabling a detailed and dynamic assessment of LLV without relying on aggregated or cumulative data.
The authors analysed participants in the Swiss HIV Cohort Study, starting ART between July 1999 and April 2023, with HIV RNA values <200 copies/mL 6 months after ART initiation. Using longitudinally collected data, they applied a time-updated Cox proportional hazards model to determine the association of LLV with the risk of subsequent viral failure, defined as ≥200 copies/mL. LLV was quantified by the time-updated area under the curve (AUC) of HIV RNA values, categorized as undetectable or, based on AUC tertiles, low, intermediate, or high.
They included 8’132 participants with a total of 49’579 person-years of follow-up. The median follow-up time was 4.7 years, and the median number of HIV RNA measurements was 16. Participants had a median age of 38 years, 75.9% were male, 74.4% were white, and 45.9% had HIV-1 subtype B. In total, 625 participants (7.7%) experienced viral failure during the observation period. LLV was associated with an increased risk of subsequent viral failure, with the highest LLV category showing the strongest association (hazard ratio, 3.3 [for comparison with undetectable viral load]) among all included variables, including race/ethnicity, age, and ART. Treatment modifications were common; 3’134 (participants 38.5%) had no modification, 2’202 (27.1%) had 1, and 2796 (34.4%) >1 (median number of modifications, 2 [IQR, [1–3]). Among participants with LLV, most (1873 of 2341 [80.0%]) achieved subsequent undetectable viral loads regardless of ART modification (7.9% with modification, 92.1% without modification. Starting the observation period in 2008 or 2014 led to stronger effects for high LLV (aHR [95% CI], 5.3 [3.0–9.4] for 2008 and 10.9 [5.0–23.8] for 2014), in line with the observation of a decreasing proportion of LLV in those without viral failure.
To summarize, this study demonstrates that LLV is a strong predictor for subsequent viral failure. Although the incidence of LLV has declined over time, its association with subsequent viral failure—particularly for high levels—has become progressively stronger, highlighting its clinical relevance in the INSTI era. Further research is needed to elucidate the mechanisms linking LLV and viral failure and identify clinically relevant patient subgroups for whom LLV is most concerning. In conclusion, individuals experiencing LLV, especially those with prolonged episodes of LLV, are at increased risk of viral failure. While current guidelines state that viral loads <200 copies/mL do not necessitate treatment changes, these findings suggest that LLV warrants close monitoring, with treatment adjustments considered in clinical practice.