To the Editors:
We read with interest a meta-analysis by Siempos et al. 1 that showed that macrolides, quinolones and amoxicillin/clavulanate might be considered equivalent for treating acute bacterial exacerbation of chronic bronchitis. Despite meticulous adherence to the methodology of meta-analysis and a comprehensive discussion of the major limitations of their study, the investigators might have used an inappropriate tool for addressing a common clinical problem.
Although meta-analysis has been placed at the pinnacle of the hierarchy of clinical evidence 2, caution is required for clinical scenarios in which targeted pathogens and their drug sensitivity patterns may vary with geographical location and time. One such scenario is the antibiotic treatment of lower respiratory tract infection, including pneumonia and chronic bronchitis, for which the clinical decision is often empirical and heavily dependent upon timely and relevant epidemiological data, as well as the patient’s clinical characteristics 3, 4. Fundamental differences in these major factors that existed between the study populations from different locations and periods would have rendered it meaningless to find summary estimates with meta-analysis.
The only valid finding may be their conclusion about the significantly higher association between adverse effects and amoxicillin/clavulanate in comparison with quinolones, since adverse events may be subject to less variation due to time, place and person. That notwithstanding, the choice of antibiotic for empirical treatment must also take into account the inherent diagnostic uncertainty and long-term implications for resistance profiles. In this regard, fluoroquinones have been incriminated in causing a delay in the diagnosis of tuberculosis 5. Thus, fluoroquinolones have been reserved for use only in certain settings in some tuberculosis-endemic populations.
Statement of interest
None declared.
- © ERS Journals Ltd