Background
It is unclear whether low-level viremia (LV) during antiretroviral therapy (ART), increase the incidence of diabetes mellitus (DM). This study aims to assess the association between HIV viremia exposure during ART and DM using retrospective cohort data.
Methods
People with HIV (PWH) who started ART in 2003 or later were identified from the China’s National Free ART Program database. Participants on ART ≥ 6 months without DM at enrolment were included in this study. According to the two consecutive viral load (VL) measurements after 6 months of ART, participants categorized into three groups: viral suppression (VS), transient episode low-level viremia (Blips), and persistent low-level viremia (LLV). Blips and LLV collectively classified as LV group. We analyzed the incidence of DM depending on viremia exposure using Cox proportional hazard models adjusted for age, sex, baseline VL, CD4 count, ART initial regimen, and WHO HIV stage. Heterogeneous linear mixed models identified fast blood glucose (FBG) trajectory patterns during the follow-up.
Results
During 26,097 person-years of follow-up, we observed 1297 cases of DM in 8731 participants, with median follow-up: 2.4 years [IQR:1.2, 4.5]. Two distinct FBG trajectories, labeled as “Stable” and “Rapid increase”, were identified. The LLV group had a significantly higher proportion of FBG in “Rapid increase” trajectory (OR: 2.53, P < 0.001). Both the Blips (cHR: 1.40, P < 0.001) and LLV (cHR: 1.74, P < 0.001) groups increased the incidence of DM than VS group. After propensity score matching, the LV group showed a higher DM risk (HR: 1.27, P = 0.011). When restricted to the 35–49 age group, the risk of DM was even higher in both the LLV (cHR: 2.24, p = 0.006) and Blips (cHR: 1.43, p = 0.011) groups than the VS group.
Conclusions
Low-level viremia (LV) substantially increased the risk of diabetes mellitus (DM), particularly in middle-aged individuals. Monitoring VL and FBG is crucial to prevent the development of DM and to improve life expectancy among ART patients.