Background: The outcome of patients on ART is, usually categorized, as patients alive onrntreatment, transferred out to other health facility, dead or lost to follow up. “Lost to follow up†isrnconsidered if he/she has missed more than three scheduled clinic visits or has not been seen for atrnleast 3 months after the initiation of cART. (After single or more visits). As ART programs scalesrnup attrition of patients from the treatment creates a great challenge for developing countries such asrnEthiopia and the ART clinic have limited capacity to track effectively lost to follow up and addressrnthe pressing question why patients drop out and what condition needed to retain them.rnObjectives: To assess reasons for defaulting from ART program at hospital based ART sites inrnAddis Ababa.rnMethods: A case-control study, on reasons for defaulting from ART carried out between Februarysrnto June 2009 in three public hospital- based ART sites in Addis Ababa. Both cases and controlsrnwere selected among patients who started receiving ART during a period of March 1, 2005 tornMarch 31, 2008. Cases were patients who were absent from clinic visit for>= 3 months after therninitiation of HAART & controls were those who are alive and on ARTrnCases were traced actively through telephone calls or home visits .Controls selected by pairedrnmatching of sex and age using lottery method from the list of active ART users who came to thernhospital to collect their medication at the day of their appointment. Two controls were matched torneach case. Both case and controls were interviewed by trained interviewers using pre testedrnstructured questionnaires. The two groups were compared by social, demographic and healthrn8rnservice variables. The over all lost to follow up rate and specific rate by month of defaulting&rnhealth institutions were calculated.rn.rnResults: 1499 lost to follow ups were included in the study. 87%were on WHO clinical stage 3rnand 4 and had a median CD4 count of 50 at the start of treatment .The over all lost to follow uprnrate from ART found to be 21% among 6992 ART patients who started treatment between a periodrnof March 2005to March 2008. During active tracing 471(31.4%) were reported dead and 198(13%)rntraced were alive while the rest (55%) could not be traced. Defaulting occurred within 3 monthsrnafter the start of ART. Traced alive and interviewed 103cases (defaulters) matched with 206rncontrols. Unemployment (AOR=2.162) with CI of 1.08 to 4.22) and monthly household incomern(AOR=0.09 with CI 0.011, 0.73) was found to be significantly associated with defaulting. Optingrnout for religious/alternative healing and clinic transfer were the reasons mentioned for treatmentrninterruption in most cases.rnConclusions: The majority of lost to follow ups were enrolled for treatment at advanced stage ofrnHIV illness. Large proportion could not be traced due to the wrong address included in the ARTrnregister. Death is found to be the main reason for treatment default .Unemployment andrneconomical constraints are the main determinant factors of treatment interruptions.rnRecommendations: ART clinics with effective pre ART programs that can identify pool of eligiblernpatients for ART and enroll timely into the ART cohorts, that have active patient trackingrnmechanisms & linkage with community and faith based organizations should be strengthened.