The observational data presented by Macfarlaneet al on the aetiology of acute lower respiratory tract illness in the community1 confirm that the often stated assertion that these illnesses are usually caused by viral infection is incorrect. The high prevalence of bacteriological and atypical pathogens and, in particular, the high prevalence ofChlamydia pneumoniae in these patients is of interest and points to the need for further studies to clarify the clinical significance of these isolates. The lack of correlation between indirect evidence of infection (radiographic and CRP levels), GP assessment of the need for antibiotics, and pathogen isolation is also of great interest and has important messages for those working in the community.

The conclusions from this study do, however, need to be treated with some caution. The authors state that outcomes were similar whether or not antibiotics were used but, as this was a non-randomised observational study, we cannot say that the groups of patients who were and were not given antibiotics were comparable. The experienced GP researchers in this study may well have had particular reasons for giving or withholding antibiotics, and the significance of similar reconsultation rates in these groups is open to interpretation.

In the accompanying editorial2 the authors state that systematic reviews of randomised controlled trials of antibiotic prescription for acute bronchitis do not support antibiotic treatment, and evidence based educational initiatives aimed at GPs are advocated as one of the strategies to alter clinical behaviour. Having recently reviewed the literature on this important clinical topic myself,3 I cannot agree with their assessment of the current evidence. The more recent review quoted4 has been criticised on methodological grounds, and the most recent and extensive systematic review of this clinical problem published on the Cochrane database5 (not referred to by Macfarlaneet al) comes to 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 delineation. The world literature currently consists of eight small randomised controlled trials of variable quality, some 20 years old, that use different antibiotic regimes and different outcome measures. Several of these studies have concluded that the antibiotic regimes used did improve outcomes.

The recent enquiry into deaths from community acquired pneumonia in young adults published in this journal6 revealed that the primary care management of these patients at the severe end of lower respiratory tract infection was deficient in many cases—three quarters of patients had seen their GP for the illness without a correct diagnosis and few had received antibiotics from the GP. Many areas of uncertainty remain in this field and, while observational studies such as that by Macfarlane et al help to bring some clarity into this confused area of daily clinical practice, well designed randomised controlled trials are still needed to produce the evidence based guidance that GPs require. The current evidence is inadequate to meet the challenge identified by Macfarlaneet al—that of identifying the cohort of patients who will benefit from antibiotics.