To the Editor: Dr Rothberg and colleagues1 assessed the relationship between antibiotic therapy and treatment failure in patients hospitalized with acute chronic obstructive pulmonary disease (COPD) exacerbation. I have 3 concerns about the methods used in this study.
First, the authors used a discharge diagnosis of acute exacerbation of COPD to define the patient population from a retrospective cohort. As these authors noted, there is no validation of discharge coding as a method for identifying this category of patients. They excluded an unstated number of patients with pneumonia. However, without knowing what proportion received a chest x-ray and what the results of the chest x-ray were, the proportion who may have had pneumonia as an unrecognized diagnosis is uncertain. Unrecognized pneumonia, if present, would have had a pneumonia severity index (PSI) score2 of at least 70 (class 2, with a mortality risk of 0.6%) for an average patient of their study population based on the characteristics given in Table 1, and a higher score and mortality risk for the 25% who were older than 79 years. The difference in mortality between the groups that received or did not receive antibiotic treatment was less than 0.6% of 84 621 patients in their study. The observed difference could have been accounted for if less than 10% of the 17 392 patients not receiving antibiotic treatment had unrecognized pneumonia and a PSI score greater than 90 (class 4, with mortality risk of 9.5%).