BACKGROUND: In HIV-uninfected individuals, isoniazid preventive therapy (IPT) has been associated with long-term protection against tuberculosis (TB). For HIV-infected/antiretroviral therapy (ART)-naive individuals, high TB rates have been observed following completion of IPT, consistent with a lack of 'cure' of infection. Recent trial data of IPT among HIV-infected individuals on ART in Khayelitsha, South Africa, have suggested that the effect of IPT persisted following completion of IPT. METHODS: Using mathematical modelling, we explored if this increased duration of protection may be due to an increased curative ability of IPT when given in combination with ART. The model was used to estimate the annual risk of infection and proportion of individuals whose latent infection was 'cured' by IPT, defined such that they must be reinfected to be at risk of disease. RESULTS: The estimated annual risk of infection was 4.0% (2.6-5.8) and the estimated proportion of individuals whose latent Mycobacterium tuberculosis infection was cured following IPT was 35.4% (2.4-76.4), higher than that previously estimated for HIV-infected/ART-naive individuals. Our results suggest that IPT can cure latent M. tuberculosis infection in approximately one-third of HIV-infected individuals on ART and therefore provide protection beyond the period of treatment. CONCLUSION: Among HIV-infected individuals on ART in low incidence settings, 12 months of IPT may provide additional long-term benefit. Among HIV-infected individuals on ART in high incidence settings, the durability of this protection will be limited because of continued risk of reinfection, and continuous preventive therapy together with improved infection control efforts will be required to provide long-term protection against TB.