Use of highly active antiretroviral treatment (HAART) has had a major impact on HIV-associated morbidity and mortality in industrialized countries. Access to HAART is now expanding in low-income countries where tuberculosis (TB) is the most important opportunistic disease. The incidence of TB has been fueled by the HIV epidemic and in many countries with high HIV prevalence current TB control measures are failing. HAART reduces the incidence of TB in treated cohorts by approximately 80% and therefore potentially has an important role in TB control in such countries. However, despite the huge beneficial effect of HAART, rates of TB among treated patients nevertheless remain persistently higher than among HIV-negative individuals. This observation raises the important question as to whether immune responses to Mycobacterium tuberculosis (MTB) are completely or only partially restored during HAART. Current data suggest that full restoration of circulating CD4 cell numbers occurs only among a minority of patients and that, even among these, phenotypic abnormalities and functional defects in lymphocyte subsets often persist. Suboptimal restoration of MTB-specific immune responses may greatly reduce the extent to which HAART is able to contribute to TB control at the community level because patients receiving HAART live much longer and yet would maintain a chronically heightened risk of TB.