A Nomogram Model to Predict Meningitis Occurrence in Cerebrospinal Fluid Leak Patients

Meningitis following cerebrospinal fluid (CSF) leak is associated with substantial morbidity and mortality. Current strategies for prevention, therapeutic options, and surgical timing lack consensus due to insufficient risk stratification tools. This study aimed to identify meningitis risk determinants and develop a predictive model to facilitate early detection in CSF leak patients.

One hundred seventy-nine patients with CSF leaks were randomly divided into training and validation sets (6:4 ratio). The least absolute shrinkage and selection operator (LASSO) regression was used to screen the relevant factors, and multivariate logistic regression was performed. Verification of the model's accuracy and applicability was conducted with the receiver operating characteristic (ROC) curve, calibration plot, and decision curve analysis (DCA).

Forty patients (22.3%) developed meningitis, predominantly males (70.0%). Smoking and chronic rhinosinusitis were more common in patients with meningitis. The multiple logistic regression analysis revealed that pneumonia and iatrogenic CSF leak are independent risk factors for developing meningitis. A nomogram consisting of four factors-smoking, sex, pneumonia, and etiology of CSF leak-was constructed. The under the curve (AUC) values for the predictive model of meningitis were 0.75 (95% CI: 0.63-0.86) in the training set and 0.76 (95% CI: 0.63-0.89) in the validation set. The calibration plots and DCA curves confirmed the excellent performance of the nomogram in predicting meningitis occurrence in CSF leak patients.

Pneumonia and iatrogenic CSF leaks are independent risk factors for developing meningitis in CSF leak patients. The nomograph model developed in this study effectively predicts meningitis occurrence, aiding in early evaluation.