Determinants of the Need for Tracheostomy in Neurocritical Patients

Abstract

BACKGROUND: Given the difficulties in predicting the need for prolonged intubation and the timing of tracheostomy, the stroke-related early tracheostomy score (SETscore) was developed, and this tool has demonstrated moderate accuracy in predicting intensive care unit (ICU) length of stay (LoS), ventilation duration, and need for tracheostomy. We aim to assess the usefulness of SETscore in a more heterogeneous population that includes trauma patients to whom this score has not yet been applied. MATERIAL AND METHODS: A retrospective consecutive analysis of all neurocritical patients who were admitted to our medical-surgical ICU between 2016 and 2018 and who required endotracheal intubation within 48 h of admission was performed in this study. Clinicodemographic data, as well as tracheostomy timing, imaging results, and SETscore were evaluated. RESULTS: The medical records of 732 neurocritical patients were reviewed, but only 493 patients were included, 68 of whom were tracheostomized (TR). These TR patients presented longer LoS and ventilation and antibiotic duration, lower Glasgow Coma Scale (GCS) score at admission, and more respiratory comorbidities. Severity scores, including SETscore, were higher in the TR group. A SETscore of >10 demonstrated 92.6% sensitivity and 79.1% specificity in predicting the need for tracheostomy. The majority of patients were tracheostomized after the seventh day of ICU admission. No significant differences in SETscore as well as in severity scores, age, and gender were observed between the early and late TR groups. However, the need for tracheostomy was significantly associated with lower ICU death rate even after controlling for GCS at admission, gender, age, and duration of invasive mechanical ventilation. CONCLUSION: SETscore can be applied to a heterogeneous population. However, more data and prospective analyses are needed to validate their clinical usefulness on a daily basis. Nevertheless, the present data are expected to contribute to the management of neurocritical patients, particularly in the setting of ICUs managing a broad spectrum of critically ill patients.