OBJECTIVES: Determine the factors associated with loss to follow-up (LTFU) in people living with HIV on antiretroviral therapy (HAART) in Mexico. METHODS: A case-control study in people living with HIV on HAART was conducted using the national database of the System of management, logistic and monitoring of antiretroviral (SALVAR) of the National Center for the Prevention and Control of HIV and Aids (CENSIDA). Subjects included in the study had begun HAART with at least 18 years age, between January 2009 and December 2013. LTFU was defined as not attending the clinic for a period of 6 months or longer, while not yet classified as dead or transferred-out. Patients in HAART until the end of the study or dead during treatment were considered to be controls. The unmatched sample size was estimated with a 95%CIs and 1:2 ratio of cases-to-control. A simple random sampling technique was used to select the sample cases and controls (330 cases and 660 controls). A logistic regression analysis was run using Stata 11.1. RESULTS: In multivariable analysis the time between the diagnostic and linkage to care ≥50 days (OR 1.87; 95%CIs 1.29-2.73) was associated with LTFU (p<0.05). Once-daily single tablet regimen (OR 0.33; 95%Cis 0.23-0.46), age of start of treatment ≥41 years old (OR 0.57; 95%CIs 0.39-0.83), 4th year since linkage to care (OR 0.11; 95%CIs 0.04-0.29) compared to first year, and last viral load <50 RNA copies/mL (OR 0.24; 95%CIs 0.17-0.33) had reduced odds of being LTFU. CONCLUSIONS: Number of pills per day had a significant impact on the risk to LTFU and an undetectable viral load had a positive feedback in the retention to treatment. The data reflects the importance of continuing research on co-formulated antiretroviral regimens.