Introduction Ambulance transport of patients with stroke is common, with rapid sequence intubation (RSI) to secure the airway used regularly. Randomised controlled trial evidence exists to support the use of RSI in traumatic brain injuries (TBIs), but it is not clear whether the RSI evidence from TBI can be applied to the patient with stroke. To this end, we analysed a retrospective stroke dataset to compare survival of patients with RSI compared with patients that did not receive RSI.

Methods This study was a retrospective analysis of 10 years of in-hospital and out-of-hospital data for all patients with stroke attended by Ambulance Victoria, in Victoria Australia. Generalised boosted logistic regression was used to predict propensity scores, with initial vital signs, age and demographic variables as well as measures of illness severity and comorbidity included in the prediction model. This analysis employed a 1:1 nearest-neighbour matching which was applied to generate a dataset from which we calculated the OR of survival to hospital discharge of patients receiving RSI versus no-RSI. The sensitivity of these results to unmeasured confounding was assessed with deterministic sensitivity analysis.

Results The propensity score-matched cohort showed a decreased survival for RSI in strokes with an OR 0.61 (95% CI 0.45 to 0.82; p=0.001) when compared with no-RSI. A subgroup analysis showed no significant survival difference for ischaemic strokes: OR 0.66 (95% CI 40 to 1.07; p=0.09). The survival for haemorrhagic stroke was OR 0.60 (95% CI 0.41 to 0.90; p=0.01) lesser for RSI. Results were likely robust to unmeasured confounding and missing data.

Conclusions Our retrospective analysis shows a decrease in survival when RSI is utilised by paramedics for stroke. Since RSI is commonly used for strokes, controlled trial evidence to support this practice is urgently needed.