Controversies in Management: Are antibiotics appropriate for sore throats? Costs outweigh the benefitsBMJ 1994; 309 doi: http://dx.doi.org/10.1136/bmj.309.6960.1010 (Published 15 October 1994) Cite this as: BMJ 1994;309:1010
- P S Little,
- I Williamson
- Department of Primary Medical Care, Aldermoor Health Centre, Southampton University, Southampton SO1 6ST.
General practitioners prescribe antibiotics for sore throat for various reasons including to prevent complications (rheumatic fever, glomerulonephritis, sinusitis, otitis media, etc), to relieve symptoms, and for psychosocial reasons. However, the benefit is marginal and the costs are great.
Do antibiotics prevent complications?
Studies on the prevention of rheumatic fever were carried out using penicillin injections in military personnel in barracks after the second world war.1 The attack rates were high (0.3-5%), and the results may not be generalisable to a modern community setting with lower attack rates and where the likelihood of developing rheumatic fever or glomerulonephritis is the same in those who have and have not had oral antibiotics.2,3 The incidence of rheumatic fever has been falling since the turn of the century - well before antibiotics were discovered.4 General practitioners in Britain have about a one in five chance of every seeing a patient with either post-streptococcal glomerulonephritis or rheumatic fever after a sore throat.2,3
*This is the seventh in a series of articles examining some of the difficult decisions that arise in medicine
The main problem of prescribing to prevent these problems is that most patients with sore throat never attend their general practitioner.2,3,5 Even if the benefit of oral antibiotics in the community were proved general practitioners' surgeries would need to be overwhelmed with patients or antibiotics would need to be freely available in the community to prevent such complications effectively.
Some evidence exists for a small protective effect of antibiotics on the development of otitis media and sinusitis.1 However, these studies are old, included small numbers of complications, and were mainly conducted in institutionalised servicemen. Studies in general practice had very wide confidence intervals for the odds ratio for developing complications (greatly overlapping 1 for prevention of otitis media).6,7 Thus it seems doubtful whether oral antibiotics prevent suppurative complications of sore throat. Even if large modern studies supported these results at least 29 subjects with sore throat would have to be treated to prevent one case of otitis media,1 which is usually a self limiting condition.
Other reasons for prescribing
The evidence for relief of symptoms in sore throat is also marginal. Results from the few placebo controlled trials in general practice suggest there may be a small increase in the number of patients well after three days among those taking penicillin.1 However, the largest trial (n=528) showed this benefit for only a small subgroup of the study population.8 Furthermore, the illness was not shortened at all irrespective of initial presentation with fever, purulent tonsils, or lymphadenitis (figure).
Psychosocial factors for both the doctor and the patient are important determinants of prescribing,9,10 and it is important to acknowlege and explore them. General practitioners probably perceive more pressure to prescribe than exists, since 41% of patients entering consultations expect a prescription but 67% leave with one.11 Even if patients receive an antibiotic for sore throat a 10 day course would be needed to eradicate streptococci, and the evidence suggests that only half of children complete such a course.12 An uncontrolled report of the acceptability of no antibiotic treatment in otitis media13 and a controlled trial of no antibiotic prescription with advice in sore throat14 indicate that most patients will find explanation and treatment of symptoms an acceptable alternative even in painful upper respiratory conditions.
Could a subgroup be targeted to improve outcome?
Given that the evidence for antibiotic prescribing in sore throat is not good for the normal range of presentations, could particular subgroups be identified? Unfortunately symptom clusters do not seem to be a good indicator of streptococcal infection or antibiotic responses, and the sensitivity and specificity of the throat swab are low–26–30% and 73–80% respectively.15 Although a rise in streptococcal antibody titres would be definitive, the delay, cost, and inconvenience of serial titres rule out their routine use.
Costs of prescribing
The probable marginal benefit of prescribing in sore throat must be weighed against the possible costs. Routine prescribing for sore throat encourages patients' dependence and reattendance at surgery,14 taking up valuable time of the doctor and the patient for a self limiting condition. In addition there are financial costs to the patient, surgery, and health service and side effects of antibiotic use such as allergy (3.8%),16 and diarrhoea (10% to 60% of children).17 18 The estimated incidence of anaphylaxis with penicillin is 1.5-4 cases per 10000 patients with two deaths per 100000.16 If every case of acute pharyngitis and acute exudative tonsillitis were treated with penicillin—that is, about 500 cases per general practitioner per year19—in the average working lifetime a general practitioner would have roughly a one in three chance of having a patient die from anaphylaxis after treatment for sore throat. This is slightly higher than the chances of nephritis or rheumatic fever after a sore throat, neither of which have a high death rate.
We argue that the evidence for benefit of prescribing for sore throat is marginal, and the costs to the patient and health service are likely to outweigh any possible benefit. Until evidence for the use of antibiotics in sore throat comes from randomised clinical trials, general practitioners should continue to explore the psychosocial reasons behind consultations and negotiate with their patients to improve the management of the symptoms of sore throat without relying on antibiotics.