Extract

(See the major article by Zetola et al on pages 1754–63.)

Before the emergence of molecular epidemiology techniques, the conceptual approach to tuberculosis treatment seemed straightforward. The clinical laboratory sent the patient's isolate (singular) for testing at a reference laboratory, which would perform a drug-susceptibility test (DST) whose findings were reported as a categorical result (ie, susceptible or resistant). This information then guided selection of appropriate antimicrobial therapy, and in clinical trials, this approach worked in the vast majority of cases. With the advent of molecular typing methods, a core premise of the classical paradigm was challenged. Genetic analyses of cultures, and even patient samples, provided evidence that a minority of patients with tuberculosis harbor >1 strain of bacteria at the same time [1–6] and that some of these mixed infections involve both drug-susceptible and drug-resistant organisms [7, 8]. What then are the consequences for patient management? [9].

In previous reports, mixed infection has been identified as the cause of discrepant DST results in pretreatment isolates [8] and of results that change during the course of therapy [7]. Thus, in patients whose initial phenotypic DST identifies only susceptible isolates, mixed infection can explain the subsequent growth of resistant organisms, offering an alternative explanation to reinfection or acquired resistance. In this situation, a patient would typically receive only first-line medications. Given that these antibiotics have little activity against the subpopulation of resistant organisms, one can readily envision how treatment failure could ensue. In the converse situation, the initial DST identifies drug-resistant organisms, but because of mixed infection there is subsequent growth of susceptible isolates. Is it possible that this situation would also result in an adverse treatment outcome if first-line drugs are withheld from patients with a subpopulation of drug-susceptible organisms? In this issue of the Journal, Zetola et al ask precisely this question.

You do not currently have access to this article.