Work

Study on predictive models for mechanical ventilation in traumatic brain injury patients

Year: 2026
Type: article
Study design: Observational Study
Abstract: Background Early identification of delayed mechanical ventilation after initial clinical stabilization is a major challenge in traumatic brain injury (TBI) management. The red blood cell distribution width-to-total serum calcium ratio (RCR) is an integrated biomarker reflecting systemic inflammation and metabolic dysregulation. This study aims to develop and validate an interpretable 6-h landmark predictive nomogram for delayed mechanical ventilation and evaluate the association between RCR and short- and long-term all-cause mortality in ICU patients with TBI. Methods This retrospective landmark study included adult TBI patients from the MIMIC-IV database (v3.0) who remained alive and unventilated at 6 h post-ICU admission ( N = 1,513). Patients were randomly partitioned (7:3) into a training cohort ( n = 1,060) and an independent holdout testing cohort ( n = 453). Core predictors were identified using the Boruta random forest algorithm exclusively within the training cohort. A multivariable logistic regression nomogram was established and blindly evaluated on the holdout testing cohort. The dose–response relationship was evaluated via restricted cubic splines (RCS). Multivariable Cox proportional hazards models were employed to assess the association of baseline RCR with 30-day and 1-year all-cause mortality. Results Among 1,513 landmark TBI patients, 119 (7.86%) required delayed mechanical ventilation. Six core predictors were selected: RCR, Age, Glasgow Coma Scale (GCS) score, Pneumonia, Sepsis, and Blood Glucose. The derived nomogram demonstrated robust discrimination, achieving an AUC of 0.778 (95% CI: 0.719–0.837) in the training cohort and 0.728 (95% CI: 0.630–0.826) in the independent holdout testing cohort. In multivariable logistic regression, elevated RCR was independently associated with an increased risk of delayed mechanical ventilation (adjusted OR = 1.895, 95% CI: 1.017–3.378, p = 0.036; Quartile 4 vs. Quartile 1: adjusted OR = 3.164, 95% CI: 1.659–6.346, p < 0.001, p for trend < 0.001). RCS analysis confirmed a significant non-linear relationship between RCR and delayed ventilation (Overall p = 0.003, p for non-linearity = 0.013). Furthermore, multivariable Cox regression verified that baseline RCR was an independent predictor for both 30-day mortality (adjusted HR = 2.994, 95% CI: 2.139–4.192, p < 0.001) and 1-year all-cause mortality (adjusted HR = 2.880, 95% CI: 2.253–3.682, p < 0.001). Conclusion Baseline RCR is an independent prognostic biomarker for delayed mechanical ventilation and long-term mortality in adult TBI patients. The 6-h landmark nomogram provides an accessible, interpretable clinical tool to assist clinicians in identifying patients at risk of secondary respiratory deterioration.
Language: English
Field: Medicine
Open Access status: gold
Estimated APC paid: $3,801
OpenAlex ID: W7236289996
FWCI:
Cites: 34
Cited by:
Related to: 10
Parts: 1