Isoniazid Activity Is Terminated by Bacterial Persistence

We welcome the application of rigorous in vitro pharmacokinetic-pharmacodynamic methods by Gumbo et al. in their attempt to explain the time course of sputum colony counts during isoniazid monotherapy [1]. However, their conclusion that the rapid emergence of isoniazid-resistant mutants is responsible for the cessation of its action in early bactericidal activity (EBA) studies is not supported by clinical data. Drug resistance occurred in only 1 of 12 patients in the original 14-day EBA study [2] and took up to 12 months to appear in early trials of isoniazid monotherapy in India, with half taking >2 months [3]. Numerous subsequent EBA studies used isoniazid as a positive control, and drug resistance was never encountered in tests routinely set up after the period of monotherapy. The authors hypothesize that resistance might be phenotypic because of up-regulation of efflux and might therefore be lost during primary culture. The data presented in figure 6 do not, however, resolve the wide discrepancy with clinical observations. It is also argued that the 1% end point in many current definitions of resistance would allow small minority populations of resistant mutants to be missed. However, the British Medical Research Council susceptibility test technique used in the first EBA study and in the Indian clinical trial was designed to detect proportions of mutants much smaller than 1% [4].