Statistics from Altmetric.com
Macfarlane et al present a comprehensive and thorough review of the multiplicity of factors affecting therapeutic decision making by general practitioners for patients presenting with acute lower respiratory tract symptoms.1 They do, however, pass very briefly over the evidence base for the use of antibiotics in this common and important clinical situation, citing only one original study, one review, and one meta-analysis to justify the statement that “antibiotics have little impact on the duration of symptoms of acute bronchitis”. For such an important and fundamental cause of morbidity in primary care there is an extraordinary dearth of studies to inform evidence-based decision making; the published studies are small, variable in quality, and use various antibiotics, dosage regimes, and outcome measures. In the quoted meta-analysis by Fahey et al 2 of randomised controlled trials comparing antibiotics with placebo, only nine studies investigating a total of 700 randomised patients were found for analysis. Only six of these studies were suitable for the analysis of some of the key outcomes. The authors' conclusion that antibiotic treatment has no effect on the resolution of acute cough was subsequently criticised.3-5
Although the clinical improvements analysed in the antibiotic treated group failed to reach statistical significance, quite narrowly for some outcomes, the results did favour antibiotics for an effect on both resolution of cough and clinical improvement at re-examination, suggesting a trend favouring the use of antibiotics over placebo. The wide confidence limits and the small numbers point to the need for further data. The Cochrane meta-analysis of the same data6reached very different conclusions, commenting that “the review confirmed the impression of clinicians that antibiotics have some beneficial effects in acute bronchitis”. The benefits are probably small and confined to certain patient subgroups, but the quantification of benefit and the definition of the characteristics of responder groups need further studies to delineate.
All responsible clinicians must be in favour of appropriate use of antimicrobial drugs and efforts to “raise the trigger line” for the use of such agents are laudable. The assertion that the majority of British GPs and their European colleagues are ignoring a good evidence base when they prescribe antibiotics in this situation would, however, appear to be premature. Clarification of which patients with acute lower respiratory symptoms will benefit and by how much can only assist us in targeting and restricting the use of antimicrobials. Increasingly well informed patients and GPs attempting to practice evidence-based medicine need such information to make rational decisions on appropriate management options. There is a need for well designed prospective placebo controlled, randomised trials performed in real world primary care settings with adequate power to provide definitive answers.
authors' reply We are grateful to Dr Thomas for interest in our review1-1 and pleased that he found it comprehensive and thorough. In his letter he debates the evidence base for the use of antibiotics for acute bronchitis or lower respiratory tract illness. There are problems with studies in this area relating to size of the studies, differing definitions of acute bronchitis, and identification of easily measurable and clinically important end points. There does seem to be a consistent message from the different studies that, overall, there is not much clinical benefit from antibiotics for acute bronchitis. This does not mean that all patients with acute bronchitis will not benefit from antibiotic use and the view that antibiotics are never indicated is unhelpful and impractical. However, we suspect that the proportion who need antibiotics is nearer to the 25% of patients in our studies in whom the GP stated that antibiotics were definitely clinically indicated than the 75% of patients consulting with acute bronchitis who are actually given antibiotics. We agree with Dr Thomas that the challenge is identifying that small group of patients in whom antibiotics are clinically indicated and it is here that further research is indicated, along with clearly described illness definitions and clinically relevant end points.