W e read with great interest the mathematical model presented by Fofana et al. (1) on the role of pyrazinamide (PZA) in the emergence of multidrug-resistant tuberculosis (MDR TB), particularly as the results of their model mirror our concerns regarding the amplification of PZA resistance during inappropriate first-line therapy and the dramatic negative consequences that this can have on the subsequent response to second-line therapy. Indeed recent data from Belarus suggest that approximately 50% of PZA resistance is acquired de novo (2). Fofana et al. (1) propose that this problem, a result of using the same important agent in first-and second-line therapy, might be circumvented by using an as-yet-unidentified new agent with characteristics similar to those of PZA in a new second-line regimen. Although considerable effort is being applied to the development of new targets for such agents (3), it is far from certain when or even if these miraculous drugs will be available for routine use. Consequently, PZA is currently a uniquely valuable agent, which we speculate will be very difficult to replace (4, 5). We therefore suggest that in settings where the a priori risk of MDR TB due to transmission is above a certain level, such as in Belarus, prescribing PZA to new patients before they have been screened for rifampin and isoniazid resistance is likely counterproductive. Screening for PZA resistance is still complicated and not in reach for many laboratories. Of course, in an ideal world, a full drug susceptibility test (DST) profile would be available for all patients before TB therapy is prescribed. Unfortunately, in reality, this is seldom the case; at best, there is a considerable delay before DST results are communicated to the physician, and at worst, DST is not performed at all until there is evidence of treatment failure, which may be between 2 and 6 months after starting the patient on first-line agents. As this scenario appears to account for 50% of the PZA resistance seen in MDR-TB cases in many settings (5), action is required now to prevent transmissible PZA-resistant mutants from becoming widely established (6, 7). Thus, we propose that, in settings with a high proportion of MDR-TB patients, PZA be withheld until there is evidence of susceptibility to other first-line agents, because of either a clinical response (acid-fast bacillus [AFB] smear conversion) or a solid laboratory diagnosis. Modeling the exact proportion of primary MDR-TB patients in whom PZA resistance was prevented, a population benefitting from this change, and discussion of the ethics of this approach would be highly informative. Delaying action Citation Anthony RM, Cynamon M, Hoffner S, Werngren J, den Hertog AL, van Soolingen D. 2017. Protecting pyrazinamide, a priority for improving outcomes in multidrug-resistant tuberculosis treatment. Antimicrob Agents Chemother 61:e00258-17. https://doi.