Objectives To determine whether antibiotics are prescribed appropriately for acute tonsillitis in an emergency department (ED).

Methods Cross-sectional observational study in large district general hospital in London. Patients diagnosed and coded with ‘acute tonsillitis’ in the ED over a 3-month period in 2015. Medical records were reviewed for Centor criteria, which is a clinical scoring system to guide antibiotic prescribing in UK general practice. Drug charts were reviewed for the specific antibiotic(s) prescribed, and throat swab (TS) cultures were recorded.

Results 273/389 patients with tonsillitis were analysed—186 children, 87 adults. Exclusions were missing patient records (86), patients had/awaiting tonsillectomy (22), receiving antibiotics (6) and immunocompromised (2). Centor score (CS) was not recorded for any patient. Based on derived CS from documented signs/symptoms, antibiotics were prescribed inappropriately to 196/273 patients (80%; 95% CI 74% to 85%) including broad-spectrum antibiotics to 25%. These included co-amoxiclav (18%), amoxicillin (6%), azithromycin (0.5%) and ceftriaxone (0.5%). TSs were taken in 66/273(24%) patients; 10/66 were positive for group A streptococcus (GAS). However, 48/56 GAS negative patients were prescribed antibiotics.

Conclusions CS was not being used in the ED to guide antibiotic prescribing for acute tonsillitis. Antibiotic prescribing was based on clinical judgement. Based on derived CS (<3), 80% of patients were inappropriately prescribed antibiotics, particularly broad-spectrum antibiotics. Further studies need to assess use of CS to guide antibiotic prescription in ED. TSs were commonly performed in the ED but did not either improve diagnosis or guide antibiotic prescription.