Antibiotics for respiratory tract infections in primary care

Abstract

Most infections can be managed by watchful waiting

Prescribing antibiotics for community acquired respiratory tract infections in primary care involves balancing the risk of missing pneumonia or serious complications on the one hand and treating infections unnecessarily on the other. Recent studies have shown that using antibiotics causes resistance among respiratory pathogens in individuals.^{1 2} These studies were needed to confirm earlier studies showing an association between antibiotics and resistance at population level and to support prescribing campaigns to combat resistance by optimising antibiotic use.³ But, although optimising the use of antibiotics seems to reduce resistance, it might increase the risk of complications.

In this week's BMJ, Petersen and colleagues assess whether antibiotics protect against serious complications of common respiratory infections.⁴ They identified 3.36 million episodes of respiratory tract infection recorded between 1991 and 2001 in the UK General Practice Research Database and determined whether complications were less common in people who were prescribed antibiotics than in those who were not.⁴ They found that in the month after the original diagnosis, pneumonia after upper respiratory tract infection, quinsy after sore throat, and mastoiditis after otitis media were rare in people not prescribed antibiotics, at 11/10 000, 14/10 000, and 5/10 000 patients, respectively. Prescription of antibiotics was associated with a small absolute reduction in risk for these serious complications. In contrast, the risk of pneumonia in the month after the diagnosis of chest infection was high and substantially reduced by antibiotic prescription (244/100 00 v 66/10 000 patients); the protective effect was greatest in people aged 65 and over.

Although this is one of the better studies to examine the effect of antibiotic prescribing on rare complications of common respiratory tract infections, a major confounding factor is that sicker patients and those more likely to have adverse outcomes were offered antibiotics more often. The findings might have been more meaningful if the participating general practitioners had low rates of antibiotic prescribing for the four conditions studied. But, according to the database used, upper respiratory tract infection, sore throat, otitis media, and chest infection or lower respiratory tract infection were four of the five main indications for prescription of antibiotics. Antibiotics were prescribed in 44.2%, 64.3%, 62.5%, and 82.2% of cases, respectively, between 1998 and 2001,⁵ and the figures were even higher before 1998.⁶

Randomisation eliminates the problem of confounding but, as Petersen and colleagues state, randomised controlled trials generally lack the power to study rare events, and participants may not be representative of those seen in routine practice. Nevertheless, according to meta-analyses of randomised controlled trials in people with colds or upper respiratory tract infections, antibiotics are not effective.⁷ A possible exception is a subgroup of people (20%) with positive nasopharyngeal culture for Haemophilus influenzae, Moraxella catarrhalis, or Streptococcus pneumoniae. However, no rapid point of care tests are currently available to detect these organisms.

Acute rheumatic fever and acute glomerulonephritis are rare in industrialised countries. In people with sore throat, antibiotics slightly improve symptoms and reduce the risk of suppurative complications, such as quinsy.⁸ A recent randomised controlled trial in children with sore throat found no effect of penicillin on the duration of sore throat and lacked power to detect a protective effect. The trial stated that each complication identified was successfully treated without referral to hospital.⁹ According to a meta-analysis of individual patient data from children with acute otitis media, antibiotics are effective in children with otorrhoea (21%) and in those under the age of 2 years with bilateral infections (20%).¹⁰ No cases of mastoiditis were found in children denied antibiotics. In people with acute bronchitis or chest infections, antibiotics modestly reduce cough.¹¹ The reduction in mean days with impaired activities and days feeling ill did not reach significance, and neither did the increase in adverse events. The largest trial to date included 807 people with acute uncomplicated lower respiratory tract infections offered immediate or delayed amoxicillin, or no treatment, and it found little difference in resolution of symptoms.¹²

In summary, the available evidence does not provide clinicians with the guidance they need to prescribe antibiotics effectively for common infections in primary care, except maybe for acute otitis media. For lower respiratory tract infections in particular, clinicians cannot be confident about identifying who will benefit from antibiotics and who will not.

GRACE (genomics to combat resistance against antibiotics in community acquired lower respiratory tract infections in Europe; www.grace-lrti.org), a network funded by the European Commission, is currently undertaking research across Europe to provide answers to these questions. Together with spin offs like TheraEDGE, an integrated platform enabling microbiological diagnosis of lower respiratory tract infections in primary care, GRACE could mould the future management of this condition in primary care.¹²

Simultaneously, the European Centre for Disease Prevention and Control in Stockholm is preparing for an annual European antibiotic resistance day, starting in 2008. This might act as a catalyst for further reductions in antibiotic prescriptions in the member states. In anticipation of this, it might be worth while establishing a surveillance system to monitor complications, to run alongside two existing European surveillance systems (EARSS (www.rivm.nl/earss) and ESAC (www.esac.ua.ac.be)).