to the editor: As the infection control chair at a 25-bed critical access hospital, I read this article on pyelonephritis with interest. I am concerned about the message that primary care physicians might receive from your choice of wording in the abstract: “Practice guidelines recommend oral fluoroquinolones as initial outpatient therapy if the rate of fluoroquinolone resistance in the community is 10 percent or less. If the resistance rate exceeds 10 percent, an initial intravenous dose of ceftriaxone [Rocephin] or gentamicin should be given, followed by an oral fluoroquinolone regimen.” Although factually accurate, this statement may discourage busy physicians from consulting local antibiograms to ensure excellent treatment. It is counterproductive to mention a specific class of antibiotics when awareness of the local antibiogram is the first decision point in the process.

In our community, as in most, Escherichia coli makes up the majority of pathogens in healthy young women. Our resistance rates are nearly 25 percent for fluoroquinolones and 4 percent for cephalosporins. These rates are similar to those in many other communities in this country, and research is revealing escalating fluoroquinolone resistance.^1,2 Less savvy prescribers often tell me they use fluoroquinolones because that is the recommendation, not reading the subsequent caveat. I suggest an alternative approach—emphasizing the importance of relying on local antibiogram patterns and allowing for appropriateness of multiple classes of drugs. Doing so would improve the quality of care by decreasing the likelihood of inappropriate prescribing and poor antibiotic stewardship.

in reply: We are pleased that our article has stimulated interest in this topic, and agree that local antibiograms can reduce inappropriate antibiotic prescribing.

Dr. Lambke is concerned about the message physicians might receive from our choice of wording on prescribing antibiotics for the treatment of pyelonephritis. As he notes, this is faithful to the guidelines from the Infectious Diseases Society of America (IDSA).¹ We doubt that rephrasing would persuade a physician who is not focused on properly following this guideline to do otherwise. The IDSA guidelines do not offer an alternative approach for physicians who are too busy to check the local antibiogram, as recommended.

We believe that awareness of the local antibiogram is implicitly necessary to follow these guidelines. Unfortunately, the typical local antibiogram is from the local hospital and amalgamates all E. coli and other isolates, regardless of specimen type, host demographics, clinical syndrome, and patient location (e.g., intensive care unit, emergency department, inpatient ward, clinic). Because susceptibility patterns can vary considerably according to these factors, physicians need cumulative susceptibility data derived from the relevant patient population for optimal prescribing. Such data are rarely available, even to physicians who are motivated to incorporate them into their decision making.

Ultimately, we are responsible for our own prescribing habits. How journal articles or guidelines are written is unlikely to change behaviors unless the reader wishes to change.