Predictors of tracheostomy in patients with spontaneous intracerebral hemorrhage

Background: One third of patients with intracerebral hemorrhage (ICH) require mechanical ventilation; in most, tracheostomy may be necessary. Limited data exist about predictors of tracheostomy in ICH. The aim of our study is to identify predictors of tracheostomy in ICH. Methods: We reviewed medical records of patients seen in our institution between 2005 and 2009, using ICD-9 codes for ICH, for admission clinical and radiological parameters. A stepwise logistic regression model was used to identify tracheostomy predictors. Results: Ninety patients with ICH were included in the analysis, eleven of which required tracheostomy. Patients requiring a tracheostomy were more likely to have a large hematoma volume (≥30 mL) (63.4% vs. 29.1%, p = 0.037), intraventricular hemorrhage (81.8% vs. 27.8%, p < 0.0001), hydrocephalus (81.8% vs. 8.8%, p < 0.0001), admission GCS < 8 (81.8% vs. 5.1%, p < 0.0001), intubation ≥ 14 days (54.5% vs. 1.27%, p < 0.0001) and pneumonia (63.6% vs. 17.7%, p = 0.003). Stepwise logistic regression yielded admission GCS (OR = 80.55, p = 0.0003) and intubation days (OR = 87.49, p < 0.006) as most important predictors. Conclusion: We could potentially predict the need for tracheostomy early in the course of ICH based on the admission GCS score; duration of intubation is another predictor for tracheostomy. Early tracheostomy could decrease the time, and therefore risks of prolonged endotracheal intubation and length of hospital stay.