Increased risk of default among previously treated tuberculosis cases in the Western Cape Province, South Africa

OBJECTIVE: To investigate, in two urban communities with high tuberculosis (TB) incidence and high rates of TB recurrence, whether a history of previous TB treatment is associated with treatment default. METHODS: Retrospective cohort study of TB cases with an episode of treatment recorded in the clinic-based treatment registers between 2002 and 2007. Probabilistic record linkage was used to ascertain treatment history of TB cases back to 1996. Based on the outcome of their most recent previous treatment episode, previously treated cases were compared to new cases regarding their risk of treatment default. RESULTS: Previous treatment success (adjusted odds ratio [aOR] 1.79; 95%CI 1.17-2.73), previous default (aOR 6.18, 95%CI 3.68-10.36) and previous failure (aOR 9.72, 95%CI 3.07-30.78) were each independently associated with treatment default (P < 0.001). Other factors independently associated with default were male sex (P = 0.003) and age 19-39 years (P < 0.001). CONCLUSIONS: Previously treated TB cases are at increased risk of treatment default, even after previous successful treatment. This finding is of particular importance in a setting where recurrent TB is very common. Adherence to treatment should be ensured in new and retreatment cases to increase cure rates and reduce transmission of TB in the community.