OBJECTIVE: To examine how patient characteristics combined with ART eligibility expansions affect the initiation of antiretroviral therapy (ART) among eligible patients attending a referral center in Senegal from 1998 to 2015. METHODS: This is a retrospective observational study carried out at the outpatient treatment Centre (Centre de Traitement Ambulatoire) in Dakar, Senegal, based on computerized medical records, gathered from 1998 to 2015, of ART-naïve patients over 15 years of age. ART eligibility was defined as (CD4 count below 200) or as (WHO stage 4) or as (WHO stage 3 with (CD4 count below 350 or with unavailable CD4 count)) in 1998-2010; as (CD4 count below 350) or as (WHO stage 3 or 4) in 2011-2013; as (CD4 count below 500) or as (WHO stage 3 or 4) in 2014-2015. Four periods were defined according to ART eligibility expansions and Senegal's HIV care history: 1998-2003 (P 1), 2004-2010 (P 2), 2011-2013 (P3), and 2014-2015 (P4). Patients were expected to participate financially in their treatment during the first period (P1). RESULTS: A total of 3651 patient records were included. The median patient age was 40 years (IQR: 32-48). Women represented 56% of the population. The median CD4 count was 183 cells/mm3. Overall, 53% of patients had CD4 < 200 cells/mm3 at entry. This proportion reached 45% in 2014-2015. 2535 patients (69%) were eligible for therapy, including 1503 (41%) who started ART. The proportion of treated patients among those who were eligible at entry or later increased steadily from 25%, 47%, 75% to 82% in the four periods, respectively. The median time to treatment decreased from 5.6 months (IQR: 3-11) in P1 to 0.8 months (IQR: 0-2) in P4. Eligible patients with more advanced disease (CD4<200 cells/mm3 and/or clinical stage 3 or 4) were more likely to be ART initiated than those with CD4≥200 cells/mm3 and/or clinical stage 1 or 2 at each stage of ART eligibility expansion. CONCLUSION: ART eligibility expansions were marked by a sharp increase in the proportion of eligible patients initiating treatment. These results show that in terms of management, the target of "Test and Treat" can be easily reached but that HIV testing will remain a key element to improve treatment success, as illustrated by the high proportion of people with advanced stage of infection at the time of ART initiation.