Citations 0
Editor's Correspondence
April 27, 2009April 27, 2009

Editor's CorrespondenceCOMMENTS AND OPINIONS

Arch Intern Med. 2009;169(8):809-813. doi:10.1001/archinternmed.2009.98

In reply

We appreciate the comments made by Reiter et al about our study of community infections caused by ESBL-producing E coli1 and would like to respond to their concerns. Reiter et al are surprised that we included community-onset, health care–associated infections in our study because these infections have “features that resemble those of nosocomial infections.” This statement, however, has been studied for bloodstream infections2 and pneumonia3 but not for urinary tract or intra-abdominal infections, which accounted for more than 95% of our cases. It is clinically sound to include all community-onset infections caused by ESBL-producing E coli because it reflects the problems faced by physicians in primary care and emergency departments. Hospital admission during the previous year was not an independent risk factor in our study; since 22% of 122 cases and 13% of 244 controls had been hospitalized during the previous year, a β error is improbable. This agrees with our hypothesis that ESBL-producing E coli in outpatients is primarily a community pathogen.4 Obviously, if such organisms are epidemic in nursing homes in a specific area, then living in a nursing home will probably turn out to be a risk factor.