BACKGROUND: HAART greatly reduces the risk of developing tuberculosis in HIVinfectedrnpersons. However, individuals who initiate HAART may still be reported as havingrnTB, either because they are developing active TB due to persistent immunodeficiency orrnbecause they have sub clinical TB that becomes apparent in the immune reconstitutionrninflammatory syndrome (IRIS). Despite the severity of the problem, little information isrnavailable on the extent to which HAART restores TB specific immune response or CD4 cellrnand characteristics of patients who develop TB while taking HAART and proportion ofrnimmune reconstitution syndrome associated with TB.rnOBJECTIVES: To assess specific immune response to tuberculosis in HIV-1 infectedrnpatients before and during HAART and to assess the characteristics of patients whorndeveloped ART associated TB and to determine the incidence rate of TB IRIS.rnMETHODS: In a longitudinal prospective cohort study, 177 study participants naive tornHAART were enrolled and followed for six months after starting HAART at ALERTrnHospital, Addis Ababa Ethiopia. The study period was from June 2006 to September 2008.rnBlood samples were collected before initiation of HAART (at baseline), at 3rd and 6th monthrnof HAART and at occurrence of suspected TB IRIS. T lymphocyte sub set enumerated andrnthe immune response to tuberculosis was assessed in vitro, using ELISPOT assay in PBMCrnstimulated with TB antigens (PPD, ESAT-6, and CFP-10), and in vivo using tuberculin skinrntest (TST) at the specified time points.rnRESULTSrnRecovery of TB specific immune response: The proportion of TST positive responsesrnincreased significantly from baseline 17.5% to 26.8% at 3rd month and to 28.9% at the 6thrnmonth (p= 0.02). TST response increased significantly among groups with CD4 cell countrn 200 cells/mm3rndecreased significantly with decreasing baseline CD4 cell count. Thus, individuals whorninitiated HAART with CD4 < 50 cells/mm3 are likely to have high risk of morbidity andrnmortality for longer period. This underscores the need of designing strategy to identify andrntreat patients before they get into a stage of profound CD4 lymphcytopenia. Our studyrnindicates that TB specific T cell repertoire may not be totally lost in HIV infectedrnindividuals with CD4 < 50 cells/ mm3 and their levels can still be improved by HAART.rnOur findings provide useful baseline information for further research to assess overallrnpotential of HAART in the restoration of immunity to tuberculosis. Finally, in this study, wernobserved a high rate of tuberculosis in our cohort during the initial months of HAARTrnindicating the need for careful screening for TB before initiation of HAART. More sensitivernand specific diagnostic methods for TB may assist in early diagnosis. In addition, our studyrnconfirmed that a sub set of patients who developed TB during HAART manifested as IRIS.rnKey words: Tuberculosis, CD4 cell count, IRIS, HAART, immune response and Ethiopia