Background: It is believed that the cessation of isoniazid's early bactericidal activity during the initial phase of antituberculosis therapy is due to the depletion of Mycobacterium tuberculosis in the exponential phase of growth. We examined the veracity of this cornerstone belief.
Methods: We used an in vitro infection model in which M. tuberculosis was exposed to isoniazid concentration-time profiles encountered in human patients. Experiments were performed to examine the time-related changes in the total bacterial population, the isoniazid-susceptible subpopulation, and the isoniazid-resistant subpopulation.
Results: The cessation of microbial kill occurred between days 3 and 4 of isoniazid therapy, as occurs in patients. There were multiple logs of organisms in the exponential phase of growth remaining at the time when bactericidal activity ceased. The isoniazid-susceptible subpopulation was replaced by an isoniazid-resistant subpopulation after 80 h of therapy. The size of the isoniazid-susceptible subpopulation continued to decrease after the total population had ceased to decrease, whereas the resistant subpopulation remained in the exponential phase of growth. Resistance was due to single point mutations in the catalase-peroxidase gene and to reserpine-inhibitable efflux pumps.
Conclusions: The age-old hypothesis that isoniazid's microbial killing of M. tuberculosis during log-phase growth ceases because of the depletion of this bacillary population needs to be modified.