Clinical Question

Should we prescribe antibiotics for sore throat?

Evidence-Based Answer

Compared with placebo, antibiotics can shorten the duration of sore throat symptoms by about 16 hours and can reduce complications. In countries where the absolute rates of complications are higher, antibiotic therapy is more likely to be effective. The effectiveness of antibiotic therapy is greatest in persons with streptococcal pharyngitis. (Strength of Recommendation: A, based on consistent, good-quality patient-oriented evidence.)

Practice Pointers

Sore throat is commonly encountered in primary care, accounting for approximately 1.3% of outpatient visits, and is often treated with an antibiotic.^1,2 Although antibiotics are useful for treating sore throat with bacterial etiology, the cause of sore throat is not always confirmed at the time of treatment, and most cases are caused by nonbacterial agents.³ Antibiotic prescribing rates vary considerably among physicians, and high prescribing rates increase costs and microbial resistance.

In this Cochrane review, the authors identified 27 studies comparing antibiotics with placebo. They examined the effect on symptom duration, the likelihood of clinical response, and the likelihood of secondary outcomes such as headache, acute rheumatic fever, acute glomerulonephritis, peritonsillar abscess, acute otitis media, and acute sinusitis. They found a mean 16-hour reduction in sore throat symptoms treated with antibiotics. Symptom resolution after three days was greatest in persons who were culture positive for group A streptococcus (number needed to treat [NNT] = 3.7), yet antibiotics also modestly benefited patients who were culture negative (NNT = 6.5) and those who were never tested (NNT = 14.4). A benefit of antibiotic treatment on development of secondary complications was also noted. Compared with no treatment, antibiotic therapy decreased the incidence of acute rheumatic fever, acute otitis media, acute sinusitis, and peritonsillar abscess, although the absolute risk reduction for each of these was modest.

Although these data are compelling, the dates of the studies included in the review should be considered. Most were conducted before 1975, when there were much higher rates of secondary complications, making the benefits of antibiotics seem more dramatic. As an example, the review found that the incidence of acute otitis media as a secondary complication of sore throat was 3% before 1975, compared with 0.7% in 2013. This difference increases the NNT from 50 to nearly 200 to prevent a single case of acute otitis media.

This systematic review found a modest reduction in the duration of sore throat symptoms and complications with antibiotic treatment, even among patients who had a negative culture for streptococcus. However, the impact on complications in contemporary developed nations is much smaller. Some patients who had a negative culture for group A streptococcus might have had group C streptococcus or may have had a false-negative culture. Limitations in design (e.g., inadequate blinding and allocation concealment, loss to follow-up) may have created a bias in favor of treatment. In addition, treating nonstreptococcal pharyngitis with antibiotics increases costs as well as antimicrobial resistance, and unnecessarily exposes patients to potential adverse effects. Special consideration should be given if the clinician is practicing in a location with a high incidence of acute rheumatic fever. The Infectious Diseases Society of America recommends that confirmed cases of streptococcal pharyngitis be treated with an appropriately selected antibiotic for a duration sufficient to eliminate the infection (typically a 10-day course).³