Predictive score for cervicofacial cellulitis severity: development and validation of a multidisciplinary tool in...

Cervicofacial cellulitis represents serious infectious emergencies with high morbidity and mortality in resource-limited settings. This study aimed to develop and validate a predictive score for cervicofacial cellulitis severity specifically adapted to the African context.

A retrospective observational study was conducted on 127 patients hospitalized for cervicofacial cellulitis between January 2020 and December 2023 in Yaoundé Central Hospital, Cameroon. Multivariate logistic regression identified independent predictive factors of severe complications, defined as respiratory distress requiring mechanical ventilation, septic shock, mediastinitis, or in-hospital death. A point-based scoring system was developed and internally validated using bootstrap methodology.

Mean age was 42.3 ± 16.8 years with male predominance (54.3%). Five independent predictive factors emerged: age over 55 years (OR = 2.4), dyspnea (OR = 6.8), trismus (OR = 3.2), white blood cell count over 12,000/mm³ (OR = 2.8), and cervical extension (OR = 4.1). The predictive score demonstrated excellent discriminative capacity with area under the ROC curve of 0.89 (95% CI: 0.83-0.95), sensitivity 88.9%, and specificity 82.6%. Three-tier risk stratification identified low-risk patients (2.3% severe complications), moderate-risk patients (28.6% severe complications), and high-risk patients (71.4% severe complications). Overall mortality was 7.9%.

This validated predictive score enables systematic risk stratification using readily available clinical and laboratory parameters in resource-limited settings. The tool demonstrates superior performance compared to clinical judgment alone and facilitates evidence-based resource allocation. External validation across diverse African populations is warranted before widespread implementation.