Abstract Body

HIV-infected patients on combination antiretroviral therapy (ART) may experience episodes of low-level viremia (LLV). Persistent LLV is associated with antiretroviral (ARV) drug resistance and regimen failure, but its etiology and optimal management are unclear. We evaluated the association between LLV and ARV concentration in hair, a measure of long-term ARV exposure.

HIV-infected women in the Women’s Interagency HIV Study receiving protease-inhibitor (PI)-containing ART for ≥1 year with plasma viral load (VL) <500 copies/mL were evaluated during semiannual visits from 2002-2014 for persistent LLV (≥2 consecutive detectable VL <500 copies/ml), transient LLV (detectable VL <500 copies/mL on non-consecutive visits), or sustained viral suppression (VS, all VL below limit of detection). Participants with virologic failure (any VL ≥500 copies/ml) were excluded. Concentrations of lopinavir, atazanavir, and darunavir were measured from hair samples using liquid chromatography/tandem mass spectrometry and divided into quartiles by drug. Multivariable logistic regression and generalized linear mixed models examined factors associated with persistent or transient LLV (versus VS) including age, race, enrollment site, year of first ART, pre-regimen VL, self-reported adherence, and hair ARV quartile.

Among 730 participants contributing 6266 person-visits of follow-up, 90 (12%) had persistent LLV and 167 (23%) experienced transient LLV. Maximum VL during follow-up was similar for women with persistent and transient LLV (median 132 [IQR 78, 240] and 119 [IQR 119, 210] copies/mL, p=0.12). In models including the above factors except hair ARV quartile, only non-Hispanic African American race was associated with persistent LLV (OR 1.9, 95%CI 1.2–3.0) and transient LLV (OR 1.8, 95%CI 1.0–3.3). Among 440 participants with hair PI levels, mean hair ARV concentration was lowest in the persistent LLV group for all 3 drugs (Figure 1). After adjusting for self-reported adherence, race, and the above factors, hair ARV quartile was significantly associated with persistent LLV (OR 2.5, 95%CI 1.1–5.8) but not transient LLV (OR 1.5, 95%CI 0.8–2.9).

In this cohort of HIV-infected women receiving PI-containing ART, one-third experienced either transient or persistent LLV during >4 years follow-up. Persistent LLV was more likely to occur among women with lower hair PI levels, suggesting that improving ARV exposure could prevent persistent LLV and its adverse consequences.