Association Between Use of a Flying Intervention Team vs Patient Interhospital Transfer and Time to Endovascular Thrombectomy Among Patients With Acute Ischemic Stroke in Nonurban Germany

Abstract

Importance: The benefit of endovascular thrombectomy (EVT) for acute ischemic stroke is highly time-dependent, and it is challenging to expedite treatment for patients in remote areas.

Objective: To determine whether deployment of a flying intervention team, compared with patient interhospital transfer, is associated with a shorter time to endovascular thrombectomy and improved clinical outcomes for patients with acute ischemic stroke.

Design, setting, and participants: This was a nonrandomized controlled intervention study comparing 2 systems of care in alternating weeks. The study was conducted in a nonurban region in Germany including 13 primary telemedicine-assisted stroke centers within a telestroke network. A total of 157 patients with acute ischemic stroke for whom decision to pursue thrombectomy had been made and deployment of flying intervention team or patient interhospital transfer was initiated were enrolled between February 1, 2018, and October 24, 2019. The date of final follow-up was January 31, 2020.

Exposures: Deployment of a flying intervention team for EVT in a primary stroke center vs patient interhospital transfer for EVT to a referral center.

Main outcomes and measures: The primary outcome was time delay from decision to pursue thrombectomy to start of the procedure in minutes. Secondary outcomes included functional outcome after 3 months, determined by the distribution of the modified Rankin Scale score (a disability score ranging from 0 [no deficit] to 6 [death]).

Results: Among the 157 patients included (median [IQR] age, 75 [66-80] y; 80 [51%] women), 72 received flying team care and 85 were transferred. EVT was performed in 60 patients (83%) in the flying team group vs 57 (67%) in the transfer group. Median (IQR) time from decision to pursue EVT to start of the procedure was 58 (51-71) minutes in the flying team group and 148 (124-177) minutes in the transfer group (difference, 90 minutes [95% CI, 75-103]; P < .001). There was no significant difference in modified Rankin Scale score after 3 months between patients in the flying team (n = 59) and transfer (n = 57) groups who received EVT (median [IQR] score, 3 [2-6] vs 3 [2-5]; adjusted common odds ratio for less disability, 1.91 [95% CI, 0.96-3.88]; P = .07).

Conclusions and relevance: In a nonurban stroke network in Germany, deployment of a flying intervention team to local stroke centers, compared with patient interhospital transfer to referral centers, was significantly associated with shorter time to EVT for patients with acute ischemic stroke. The findings may support consideration of a flying intervention team for some stroke systems of care, although further research is needed to confirm long-term clinical outcomes and to understand applicability to other geographic settings.

Figure 1.. Map of Southeast Bavaria, Germany, Where the TEMPiS Network Operates

Blue dots indicate primary stroke centers taking part in the flying team project. Small red dots are referral centers for patient interhospital transfer. The large red dot indicates the helicopter site; it obscures the adjacent and another nearby referral center in the city of Munich. Blue arrows signify the usual pathway of flight or transfer. Additional hospitals that provide stroke care but did not participate in this project are not shown. A, All primary stroke centers receive flying intervention team service from Munich. Flight distance and flying time in minutes is given at each black half circle. B, Transfer directions and transfer time estimates from each primary stroke center to the preferred referral center are shown. Transfer time estimates were calculated taking into account general rates of airborne and ground transfer, a helicopter flying speed of 220 km/h, and estimates of driving times on optimal routes according to the route planner service of the General German Automobile Club (http://maps.adac.de).

Figure 2.. Flow of Patients in a Study of the Effect of a Flying Intervention Team vs Patient Interhospital Transfer on Time to Endovascular Thrombectomy

^aMultiple clinical exclusion criteria were possible. Patients may be included in more than 1 category. ^bGreater than 6 hours for anterior circulation occlusion without possibility of advanced imaging in primary stroke center and greater than 24 hours for posterior circulation occlusion or anterior circulation with mismatch in advanced imaging performed in primary stroke center. ^cThe Alberta Stroke Program Early CT Score is a 10-point quantitative topographic score that measures the extent of early ischemic changes in patients with middle cerebral artery stroke. Starting with 10, a 1-point deduction is made for every region involved. Scores range from 0 (10 affected regions) to 10 (0 affected regions). ^dClinical improvement and/or spontaneous recanalization.

Figure 3.. Functional Outcome at 3 Months in a Study of the Effect of a Flying Intervention Team vs Patient Interhospital Transfer on Time to Endovascular Thrombectomy

Scores on the modified Rankin Scale (mRS) range from 0 (no symptoms) to 6 (death). Severe disability (score of 5) and death (score of 6) were combined in a single category in the ordinal logistic regression model. Of all patients included, follow-up data were missing in 3 patients (2 in the transfer group and 1 in flying team group). In patients who underwent EVT, follow-up data was missing in 1 patient in the flying team group. In patients who received EVT, the adjusted common odds ratio of the flying team group was 1.91 (95% CI, 0.96-3.88; P = .07) for an improved mRS score, 1.64 (95% CI, 0.72-3.72; P = .24) for good clinical outcome, and 1.14 (95% CI, 0.44-2.94; P = .79) for death within 3 months. In the whole study population, including patients who eventually did not receive EVT, the flying team group was significantly associated with an improved mRS score after 3 months (adjusted common odds ratio, 1.91 [95% CI, 1.05-3.50]; P = .04). Multivariable logistic regression models were adjusted for age, sex, and National Institute of Health Stroke Scale score. No missing values were present in the baseline data used in this model.