External validation of the Baseline Recurrence Risk in Cellulitis (BRRISC) score and the added impact of acute clinical response: a prospective cohort study

Abstract Background The BRRISC score was developed to predict hospital-attended cellulitis recurrence using clinical data routinely available at presentation. In practice, clinicians assess patients’ response to treatment during the recommended 48-72-hour antibiotic review point when deciding on the duration of antibiotic treatment to achieve lasting recovery. We evaluated the performance of the BRRISC score in an external validation cohort and examined whether incorporating markers of acute clinical response could improve it. Methods We recruited adults with lower limb cellulitis attending hospital. From days 0–3 of treatment, we assessed markers of acute clinical response, including physical examination findings (e.g. affected size, oedema), objective measurements of limb temperature taken with a thermal imaging camera, vital signs, blood test results, and patient-reported symptoms. Outcomes included ‘hospital-attended recurrence’ (primary outcome for validation) and ‘any recurrence’ (hospital-attended or community). Using multivariate logistic regression with backwards elimination, we identified response variables independently associated with either outcome that could be included within an extended score. Performance was assessed using the C-index. Results Of 202 patients, 8% ( n  = 17) experienced ‘hospital-attended recurrence’ and 23% ( n  = 46) ‘any recurrence’. In this validation dataset, the BRRISC score had a C-index = 0.75 (95%CI, 0.64–0.86) for predicting ‘hospital-attended recurrence’ vs. 0.65 (0.63–0.68) in the original development population, but only 0.60 (0.51–0.69) for the new outcome ‘any recurrence’. There was weak evidence that an extended score, incorporating day-2/3 severity of skin blistering, improved the C-index for ‘hospital-attended recurrence’ to 0.83 (0.75–0.92). No acute clinical response variables were independently associated with ‘any recurrence’ after adjusting for BRRISC score. Conclusions The BRRISC score can help identify patients with cellulitis at the highest risk for hospital-attended recurrence. Markers of acute clinical response, typically used by clinicians to inform antibiotic treatment decisions during the 48-72-hour antibiotic review period, did not add helpful prognostic value beyond the baseline factors included in the score. Whether treatment response after day-3 improves recurrence prediction could be further explored, but future research should focus on evaluating the potential for baseline risk stratification to personalise antibiotic duration and guide non-antibiotic approaches for preventing recurrence. Clinical trial number Not applicable.