An analysis of the Clinical Assessment of Strep Throat (CAST) to evaluate the likelihood of streptococcal infection: a randomised trial of patients with upper respiratory tract infection

Abstract

Background: The symptoms of viral and bacterial sore throats are similar, making it difficult to distinguish between them using clinical features alone. We evaluated the Clinical Assessment of Strep Throat (CAST) for predicting beta-haemolytic streptococcal infection. Results were compared to throat cultures that provide a definitive streptococcal diagnosis. Methods & Materials: A total of 440 patients with recent-onset sore throat due to upper respiratory tract infection were randomised to receive a single dose of flurbiprofen 8.75 mg delivered as a spray or lozenge. Using the patient's history, symptoms and physical findings, investigators performed the CAST to make a clinical judgement on the likelihood of streptococcal infection using a 4-point categorical scale (unlikely [-1], uncertain [0], likely [1], very likely [2]). Throat swabs for culture identified patients with Strep A or C infection (n = 433). Results: Streptococcal infection was considered ‘very likely’ in one patient, ‘likely’ in 78 patients (18.0%), ‘uncertain’ (ie a lack of diagnosis) in 123 patients (28.4%) and ‘unlikely’ in 231 patients (53.3%). Overall, 25 patients had Strep A- or C-positive throat cultures; of these only 8 were considered ‘likely’ or ‘very likely’ to have a streptococcal infection using the CAST. The sensitivity and specificity of the CAST was 32% and 55.9%, respectively. The misclassification rate (clear errors in prediction) was 17.1%. However, a trend test showed the ‘likely’ and ‘very likely’ categories were associated with higher probability of streptococcal infection (p = 0.0003). Conclusion: While the number of Strep A or C cases was relatively low, these results clearly show the difficulty in predicting the likelihood of infection based on clinical features alone. Of the 79 patients identified as ‘likely’/‘very likely’ to have a streptococcal infection using the CAST, 71 patients could have potentially received antibiotics inappropriately. Such a large proportion is critical, as inappropriate antibiotic usage contributes to the global problem of antibiotic resistance. Flurbiprofen is an alternative to inappropriate antibiotics and in this study provided symptomatic relief in patients with or without streptococcal infection. Conflicts of interest: Reckitt Benckiser funded this study. Adrian Shephard and Valeria Bychkova are employees of Reckitt Benckiser, and Natalia Burova's organisation received Reckitt Benckiser funding for the study.