Column 2: Demographic, behavioral and clinical characteristics of the study population (number [percentages] are given unless otherwise stated); Columns 3 to 8: Results of the logistic regression models for the three outcomes: ART discontinuation, Loss to Follow-up and Initial ART Regimen.

HIV: human immunodeficiency virus, ART: antiretroviral therapy, ADI: AIDS defining illness.
aAge at the start of ART.
bReported mode of HIV risk exposure was categorized injection drug users (IDU) and not IDU.
cPre-treatment CD4+ cell count and HIV RNA were defined as the value closest to the date of start of ART up to 6 months prior.
dInitial ART regimen was classified as NNRTI-based or PI-based. Integrase inhibitor-based regimens were too few (N = 75) to draw consistent conclusions and were thus excluded.
eADI at the start of ART was defined as the presence of any CDC 1993 condition at six months prior to up to one month after start of ART.
fHepatitis B/C co-infection was defined as having chronic infection at the start of ART.
hSignificant cohort and ADI interaction term [aOR 2.02 (1.42, 2.86)].