[Facteurs associés au délai d'accès à la thrombectomie mécanique dans la prise en charge des accidents vasculaires cérébraux ischémiques en Gironde, France, 2017-2018].

BACKGROUND AND PURPOSE: When an ischaemic stroke due to a large vessel occlusion occurs, the sooner Mechanical Thrombectomy (MT) is performed, the better the functional prognosis. However, the organisation of care does not systematically allow rapid access to MT. The aim of our study was to determine the clinical and organisational factors associated with the time to access to MT.

METHODS: We conducted a cohort study in Gironde County, France. Patients admitted for MT and regulated by the Gironde Emergency Medical Services (EMS) between 01/01/2017 and 31/12/2018 were included. The time to access to MT was the difference between the first call to EMS and groin puncture for MT. The main explanatory variables were: type of pathway (mothership (MS), drip and ship (DS) with cerebral imaging performed in the local hospital centre (LHC), and DS without imaging in the LHC); NIHSS score; driving distance to MT; time of stroke onset (weekend or holiday, school holidays, other); age and sex. Linear regression models were used to explain time to access to MT. Missing data were handled using a multiple imputation procedure (Full conditional specification, Mice R-Package) carried out in our multivariable linear regression model. A quantitative bias analysis was performed by weighing the imputed time to access to MT and identifying the weight changing the conclusions of our analysis.

RESULTS: Among the 314 included patients, 152 were women (48.4%), and the mean NIHSS score was 16.4. Two hundred and two (64.3%) patients were managed through the MS pathway. The average time from onset to femoral puncture was 251 minutes. In the multivariate analysis, the time to MT was longer when patients were managed DS with imaging in the LHC pathway (+106 min, p = 0.03), and even longer in the DS without imaging in the LHC pathway (+197 min, p = 0.002), compared with MS. Time from onset to MT decreased with increasing NIHSS score (-6 min per NIHSS point, p <.0001). In our quantitative bias analysis, we multiplied the imputed time in access to MT in the DS pathways only (with or without imaging in the LHC) by weights varying from 0.9 to 0.2 (imputed delays reduced from 10% to 80%). With reduction of 40% or more, there was no longer any difference in time to access to MT between the three studied pathways.

CONCLUSIONS: The DS pathway can be shortened by generalizing access to cerebral imaging in LHCs. Optimizing pre-admission orientation toward MT is a major issue in LVOS management.