Extensively drug-resistant tuberculosis in South Africa.

Keertan Dheda and colleagues (May 22, p 1798) report extremely poor treatment outcomes for patients with extensively drug-resistant (XDR) tuberculosis in South Africa and conclude that prevention of XDR tuberculosis through strengthening of tuberculosis pro grammes overall should be prioritised. Most patients diagnosed with XDR tuberculosis in this cohort (72%) had previously been diagnosed with multidrug-resistant (MDR) tuberculosis. Presumably, these patients received treatment for MDR tuberculosis before diagnosis of XDR tuberculosis. The standard treatment available for MDR tuberculosis at the time of the study, as described, is a relatively weak regimen likely to lead to amplifi cation of resistance—patients with strains of Mycobacterium tuberculosis already resistant to one of the second-line injectable agents or a fl uoroquinolone are likely to develop XDR tuberculosis even with good treatment adherence. Thus, another important strategy to prevent XDR tuberculosis is to ensure adequate treatment of MDR tuberculosis. A regimen for all cases of MDR tuberculosis that contains moxi fl oxacin and at least three other drugs that are likely to be eff ective will no doubt substantially reduce the creation of XDR tuberculosis during treatment. Contrary to what some suggest, moxifl oxacin is well tolerated and unlikely to contribute to additional adverse events. Such a regimen needs to balance tolerability against effi cacy, both in terms of cure and prevention of further resistance. The contribution of previous treatment of MDR tuberculosis to poor outcomes in this study is unclear, but might well explain the disparity between this study and the relatively good outcomes reported from Peru in patients treated with more robust second-line regimens. Given extensive previous treatment in the South African cohort, outcomes might not be generalisable to patients initially diagnosed with XDR tuberculosis and treated appropriately.