Griffith et al [1] comprehensively review the problem of drug shortages as they pertain to antimicrobial agents and the treatment of infectious diseases, and they highlight an extremely important problem in the everyday clinical practice of infectious disease specialists and many other types of physicians, from surgeons to internists, and critical care specialists to dermatologists. By doing a simple analysis of antimicrobial shortages, obtained from the Web site of the American Society of Health-System Pharmacists, they have brought to attention the frequent widespread lack of availability of antibiotics, and that this problem reflects trends in other pharmaceutical agents. They evaluate the variety of reasons for antibiotic shortages, provide examples, and, most importantly, point out how they can affect patient care, sometimes with serious consequences. Their analysis was little more than a listing of antimicrobial agents, reasons for each shortage, and the time periods that the drugs were unavailable, but their investigation covered 2005–2010, a long enough time frame to evaluate trends and draw some conclusions. As they noted, while drug shortages in any pharmacologic area can be an inconvenience for pharmacists, clinicians, and others; lead to increased costs for patients, hospitals, and insurance companies; and have an easily discerned anecdotal impact on individual patients, it is difficult to systematically measure the resulting clinical problem or draw quantitative conclusions about differences in outcomes. Many broad clinical syndromes, such as pneumonia or various types of skin infections, can be treated with several antibiotics or combinations that are equally good. Guidelines published by the Infectious Diseases Society of America and other professional groups generally do not specify a single best antibiotic treatment for most diseases. Nearly all large prospective antimicrobial studies are done as noninferiority trials, in which the superiority of one antibiotic over another cannot be established. Sometimes a new antibiotic ‘‘fails’’ the test of noninferiority, so we know about some antibiotics to avoid in certain situations, such as daptomycin for pneumonia, because of its inactivation by pulmonary surfactant. Other antibiotics are not expected to work in certain clinical situations, such as bacteriostatic agents for endocarditis, and are also generally avoided. However, when clinicians are used to using a particular antibiotic for a commonly encountered clinical situation, and are then no longer able to use it because of a shortage, they may not select an equally good alternative. This may be due to lack of knowledge of the local antibiogram data, or even the lack of a recent local antibiogram, or for other reasons, and may be a particular problem for clinicians who do not use antimicrobials on a frequent, regular basis. As Griffith et al [1] point out, studies of several clinical syndromes have shown higher mortality when inadequate empiric treatment is used, so the choice of an empiric antibiotic with antimicrobial activity against the infecting bacteria is not merely an academic exercise. However, for diseases such as pneumonia or cellulitis, the infecting bacteria and their antimicrobial susceptibilities are usually not identified, so concluding with certainty that clinical failure was because of the selection of an incorrect antibiotic can be done only in a small subset of patients from whom a culture from a normally sterile site grows the causative bacteria. The initial antibiotic is frequently the only antibiotic unless a change is needed because of deterioration, lack of improvement, or other clinical reasons. Received 2 November 2011; accepted 8 November 2011. Correspondence: Ira Leviton, MD, Division of Infectious Diseases, Montefiore Hospital, 111 E 210th St, Bronx, NY 10467 (ileviton@montefiore.org). Clinical Infectious Diseases The Author 2012. Published by Oxford University Press on behalf of the Infectious Diseases Society of America. All rights reserved. For Permissions, please e-mail: journals.permissions@ oup.com. DOI: 10.1093/cid/cir942