Tracheostomy in traumatic cervical spinal cord injury: Early versus late tracheostomy

Abstract

ObjectiveThe purpose of this study was to characterize the relationship between predictors and the time of tracheostomy after traumatic cervical spinal cord injury (TCSCI). MethodsFive hundred twenty-six patients with TCSCI treated between January,2012 and December, 2021 were retrospectively reviewed. Patients were subdivided into two groups: early tracheostomy (≤7 days from initiation of endotracheal intubation) and late tracheostomy. Comparisons between early tracheostomy and late tracheostomy were statistically analyzed. Logistic regression analysis was applied to identify independent predictors of tracheostomy and calculate probability for different grades of combining predictors to predict tracheostomy. Spearman’s correlation coefficient was used to evaluate the association between the grade of combining predictors and the time to tracheostomy. ResultsAmong 526 eligible patients, 63(12.0%) had a tracheostomy performed. Compared with late tracheostomy group, patients in early tracheostomy group had higher ISS, more severe neurological status while fewer In-hospital LOS days and ICU LOS days. By Logistic regression analysis, severe American Spinal Injury Association Impairment Scale (AIS A), the neurological level of injury (NLI>C5), higher Injury Severity Score (ISS>16) and advanced age (over 50 years old) were identified as independent predictors for tracheostomy. Depending on the likelihood of tracheostomy, the combining predictors were graded into five categories. As the value of probability was higher than 50%, Grade I-III made optimistic predictions about tracheostomy. According to Spearman’s correlation analysis, early tracheostomy had a strong association with Grade I, while late tracheostomy was closely correlated with Grade III. ConclusionsFactors related to the decision of tracheostomy were ASIA impairment scale, neurological level of injury, injury severity score and age. The grades of combining predictors could support indication for predicting the time of tracheostomy.