Development and internal validation of a nomogram for predicting mechanical ventilation requirement in adult tetanus patients: a dual-center retrospective study

To develop and internally validate a nomogram for predicting mechanical ventilation (MV) requirement in adult patients with tetanus. This retrospective dual-center study included 227 adult tetanus patients admitted to two tertiary hospitals in Southern Jiangxi, China, between January 2012 and December 2024. MV requirement, defined as the initiation of invasive mechanical ventilation for respiratory failure, served as the primary outcome. Thirty-eight candidate predictors encompassing demographics, clinical characteristics, and laboratory biomarkers were evaluated. Missing data (3.2%) were imputed using random forest-based multiple imputation. LASSO regression was employed for feature selection, followed by multivariable logistic regression to construct the prediction model. Internal validation was performed using 10-fold cross-validation and bootstrap resampling (1,000 iterations). Model discrimination was assessed by the area under the ROC curve (AUC), calibration by calibration plots and the Hosmer-Lemeshow test, and clinical utility by decision curve analysis (DCA). Of 227 patients, 123 (54.2%) required MV. LASSO regression selected six variables (age, CKD, neutrophil count, D-dimer, urea, CK-MB), of which five (excluding urea) were retained in the final multivariable model as independent predictors: age (OR 1.032, 95% CI : 1.011–1.054), CKD (OR 4.939, 95% CI : 1.621–18.681), neutrophil count (OR 1.187, 95% CI : 1.078–1.321), D-dimer (OR 1.089, 95% CI : 1.019–1.182), and CK-MB (OR 1.034, 95% CI : 1.015–1.057). The nomogram demonstrated an AUC of 0.813 (95% CI : 0.757–0.868), with an optimal cutoff of 0.576 yielding a sensitivity of 0.675 (95% CI : 0.592–0.758) and specificity of 0.837 (95% CI : 0.765–0.908). The Hosmer-Lemeshow test indicated adequate calibration (chi2 = 11.031; P = 0.200). Bootstrap validation confirmed robustness with an AUC of 0.776 (95% CI : 0.773–0.779), and DCA demonstrated net clinical benefit across threshold probabilities of 0.1–0.8. We developed a nomogram incorporating five routinely available admission parameters that accurately predicts MV requirement in adult tetanus patients. The model demonstrated favorable discrimination, adequate calibration, and positive clinical utility, offering a practical bedside tool for early risk stratification. External validation is warranted to confirm generalizability.