Linezolid for multidrug-resistant tuberculosis - authors' reply.

In their Comment in The Lancet Infectious Diseases, Kwok-Chiu Chang and colleagues concluded that expansion of access to linezolid for complicated cases of drug-resistant tuberculosis risks the loss of a potentially useful drug and could promote the emergence and spread of drug-resistant tuberculosis in the community. The rationale behind this idea seems to be the scarcity of controlled clinical trials and, in particular, data for optimum linezolid dose. Although we agree that the evidence base is small, our recent systematic review suggests that good outcomes can be achieved with linezolid among patients who would otherwise have very poor outcomes and high mortality. Among 148 patients treated with linezolid, the rate of treatment success (68%) was at least as good as that expected for multidrug-resistant tuberculosis treatment overall (62%). None of the authors of the studies included in that review advocates large-scale and indiscriminate use of linezolid for drugresistant tuberculosis. Conversely, in view of the poor side-eff ect profi le and treatment complexity, linezolid is recommended for patients with few remaining treatment options, treated in well functioning programmes. The risk of emerging drug resistance is relevant, but no more so than for other second-line drugs presently used for drug-resistant tuberculosis. Existing treatment regimens are lengthy, are associated with sub stantial side-eff ects, and result in overall poor outcomes. Unfortunately, controlled trials to defi ne better regimens are scarce. However, failure to scale up treatment access will lead to continued community transmission and worsening of the epidemic. Newer derivatives of linezolid with improved side-eff ect profi les are under development. Meanwhile, linezolid should be available for patients with few treatment options, owing to either extensive drug resistance or previous treatment failure. We have started fi ve patients on strengthened linezolid and clofazimine-containing regimens in Khayelitsha, South Africa, all of whom produced negative sputum cultures within 3 months. Subsequent linezolid withdrawal was necessary in one patient because of severe peripheral neuropathy. Although side-eff ects need to be monitored and managed carefully, the restricting factor in the use of linezolid in our setting, where Pfi zer holds a patent, is cost, rather than scarcity of evidence. In South Africa, linezolid is available in the public sector at a cost of US$1000 per patient per month. Although improved access to linezolid alone will not solve the worldwide drug-resistant tuberculosis crisis, facilitation of access to new and repurposed drugs for drugresistant tuberculosis will contribute to the overall goal of a shorter, more tolerable, and more eff ective treatment regimen than is available at present, and will off er the hope of cure to patients who would otherwise die.