To the Editor.
—Dr Saint and colleagues1 have written an article in support of treating exacerbations of chronic obstructive pulmonary disease (COPD) with antibiotics. It is always nice to know one's applied standard of care is supported by the medical literature. Critical analysis revealed an exhaustive project leading to a conclusion with severe limitations.The selection criteria produced nine of 239 studies for metaanalysis. Randomized studies comparing treatment groups (antibiotics) vs control groups (placebo) were analyzed. A variety of outcomes were compared and averaged. In essence, averaged apples were compared with averaged oranges. No range of outcomes results were defined. For comparison between studies of effect size to be useful, the outcomes of a given study all must fall within a narrow range, with the mean defined as a representative sample of the population of outcomes studied. Even then, the mean of the means is suspect when the result