Article Text

Download PDFPDF

Is air transport of stroke patients faster than ground transport? A prospective controlled observational study
  1. Rasmus Hesselfeldt1,
  2. Jesper Gyllenborg2,
  3. Jacob Steinmetz3,
  4. Hien Quoc Do1,
  5. Julie Hejselbæk1,
  6. Lars S Rasmussen1
  1. 1Department of Anaesthesia, Copenhagen University Hospital, Rigshospitalet, Copenhagen, Denmark
  2. 2Department of Neurology, Roskilde University Hospital, Denmark
  3. 3Helicopter Emergency Medical Service, Ringsted, Denmark
  1. Correspondence to Dr Rasmus Hesselfeldt, Department of Anaesthesia, Copenhagen University Hospital, Section 4231, Rigshospitalet, Blegdamsvej 9, Copenhagen 2100, Denmark; hesselfeldt{at}


Background Helicopters are widely used for interhospital transfers of stroke patients, but the benefit is sparsely documented. We hypothesised that helicopter transport would reduce system delay to thrombolytic treatment at the regional stroke centre.

Methods In this prospective controlled observational study, we included patients referred to a stroke centre if their ground transport time exceeded 30 min, or they were transported by a secondarily dispatched, physician-staffed helicopter. The primary endpoint was time from telephone contact to triaging neurologist to arrival in the stroke centre. Secondary endpoints included modified Rankin Scale at 3 months, 30-day and 1-year mortality.

Results A total of 330 patients were included; 265 with ground transport and 65 with helicopter, of which 87 (33%) and 22 (34%), received thrombolysis, respectively (p=0.88). Time from contact to triaging neurologist to arrival in the regional stroke centre was significantly shorter in the ground group (55 (34–85) vs 68 (40–85) min, p<0.01). The distance from scene to stroke centre was shorter in the ground group (67 (42–136) km) than in the helicopter group (83 (46–143) km) (p<0.01). We did not detect significant differences in modified Rankin Scale at 3 months, in 30-day (9.4% vs 0%; p=0.20) nor 1-year (18.8% vs 13.6%; p=0.76) mortality between ground and helicopter transport.

Conclusions We found significantly shorter time from contact to triaging neurologist to arrival in the regional stroke centre if stroke patients were transported by primarily dispatched ground ambulance compared with a secondarily dispatched helicopter.

  • Critical Care Transport
  • Helicopter Retrieval
  • Neurology, Stroke
  • Stroke
  • Prehospital Care, Helicopter Retrieval
View Full Text

Statistics from


Fast transport to a dedicated stroke centre may allow earlier thrombolysis and, therefore, improve outcome in stroke patients.1 Centralisation and development of designated stroke centres results in longer transport distances. Thrombolysis should be given within 4.5 h from symptom onset, but the efficacy of this treatment is time dependent.2 This emphasises the importance of fast and direct transport. Helicopters are widely used for interhospital transfers of stroke patients,3 but the benefit is sparsely documented. Moreover, the potential role of helicopters in direct referral from out-of-hospital locations is not well described, and there is a lack of prospective controlled studies assessing the impact of triage and dispatch protocols involving helicopter transport.

In this study, we compared ground-transported patients with helicopter-transported patients, and we hypothesised that the helicopter would reduce time from telephone contact to triaging neurologist to arrival in the regional stroke centre. We also assessed the prognosis for the patients who received thrombolysis.


This was a prospective controlled observational study conducted between 1 January 2010 and 30 April 2011. We evaluated a recently implemented physician-staffed helicopter, based centrally in a rural region, and dispatched as ‘second wave’ to stroke patients. All patients received at the regional stroke centre for thrombolysis assessment were consecutively registered. We included and compared two groups: (1) patients transported by conventional ground ambulance (GRD group) with an actual driving distance from scene to hospital exceeding 30 min, as recorded on the ambulance record, in a 16-month period (1 January 2010–30 April 2011) and (2) all patients transported by a physician-staffed helicopter (AIR group) in a 12-month period (1 May 2010–30 April 2011). No exclusion criteria were defined. Both field triaged and interhospital transfer patients were included.

The study was conducted in the region of Zealand located in the eastern part of Denmark. The regional thrombolysis centre covers a population of 820 000, and an area of 7.273 km2 (figure 1). The national protocol for acute ischaemic stroke treatment states that all patients should be evaluated by a neurologist (over the phone) and thereafter, if relevant, transported directly to a designated regional stroke centre, bypassing the nearest hospital. Thus, no patients receive intravenous thrombolysis out-of-hospital or in local hospitals. The daylight operating physician-staffed helicopter was implemented as a supplement to the existing system from 1 May 2010. In case an ischaemic stroke was suspected by the Emergency Medical Services (EMS) provider or by hospital staff assessing the patient, the attending neurologist at the stroke centre was contacted by telephone, and after a short interview regarding comorbidity, symptoms and duration of these, he/she decided whether or not the patient was a candidate for thrombolysis. If so, the dispatch centre was contacted and if the expected driving distance with light and sirens to the stroke centre exceeded 30 min, the helicopter was dispatched and rendezvous with the ground ambulance arranged. If weather conditions prohibited helicopter transport or the helicopter was occupied on another mission, the ground ambulance completed the transport as before the helicopter implementation. Data collected included demographics, transport distance by road, pre- and inhospital time intervals, National Institute of Health Stroke Scale (NIHSS)4 at 0 and 24 h, and modified Rankin Scale (mRS)5 after 3 months. Data on 30-day and 1-year mortality were gathered from the Danish Civil Registration System.6

Figure 1

Map showing the Island of Zealand with the helicopter base and regional stroke centre. The area marked in the right upper corner indicate a neighbouring region not described in this study.

The primary endpoint was time from contact to triaging neurologist until arrival at the stroke centre. Secondary endpoints included the change in NIHSS 0–24 h, thrombolysis rate, good functional outcome defined as modified Rankin Scale of 0–2 (no symptom—slight disability) at 3 months follow-up, and 30-day and 1-year mortality.


We reported continuous variables with medians and 5–95% percentiles, and compared groups using Mann–Whitney's test. Categorical data were presented as numbers and percentages. Groups were compared using χ² test, or Fisher's Exact test where appropriate.

The helicopter intervention period was set to 1 year prior to patient inclusion, and the area from which the patients were included was defined beforehand. Thus, a sample size calculation was not done, though we expected 200 patients to be included.

p<0.05 was considered statistically significant.

The data collection and registration were approved by the Danish Data Protection Agency (j. nr: 2009-41-4122) and the National Board of Health (j. nr: 7-604-04-2/128/HKR). According to Danish law, this study did not need approval from an ethics committee, nor was patient consent needed.


A total of 330 stroke patients were included. Ground ambulances transported 265 patients, and 65 patients were transported by helicopter (AIR) (figure 2). The transport distance was shorter in the GRD group than in the AIR group (67 (42–136) vs 83 (46–143), p<0.01) (table 1). Time from triaging neurologist contact until arrival at the stroke centre was shorter in the GRD group (55 (34–85) min), compared with the AIR group (68 (40–85) min) (p<0.01). Ground transport was also significantly faster for the subgroup referred from an out-of-hospital location (field triaged) (table 1).

Table 1

Allocation, time intervals and transport distances by road for stroke patients transported by ground (GRD) or by helicopter (AIR) to a stroke centre

Figure 2

Allocation, transport mode and treatment for stroke patients received at the regional stroke centre (GRD=ground transport; AIR=helicopter transport).

A median delay from stroke centre contact to dispatch of the helicopter of 8 (1–19) min was observed.

In the GRD group, 87 (33%) received thrombolysis versus 22 (34%) in the AIR group (p=0.88). There were no significant differences between these groups in demographics, median door-to-needle time, or symptom onset-to-needle time (165 (104–159) vs 156 (112–248) min; p=0.37). The NIHSS assessed at arrival (0 h), was 8 (3–20) in the GRD group and 10 (2–17) in the AIR group (p=0.84). NIHSS improved in 72% (GRD) and 78% (AIR) of the cases during the first 24 h after arrival; mRS of 0–2 at 3 months was achieved in 47/87 (54%) and 13/22 (59%) patients (p=0.67), respectively (table 2).

Table 2

Demographics, patient characteristics, mortality and modified Rankin Scale for stroke patients transported by either ground (GRD) or by helicopter (AIR) and who received thrombolysis

We did not detect any differences in neither 30-day (8/85 (9.4%) vs 0/22 (0%), p=0.20), nor 1-year mortality (16/85 (18.8%) vs 3/22 (13.6%), p=0.76), for patients who received thrombolysis in the GRD group versus the AIR group. Two patients in the GRD group were lost to follow-up, and none in the AIR group. Time to arrival for ground patients transported during daylight hours (8:00—20:00) (55 (34–90) min), was not different from patients transported during dark hours (20:00–8:00) (57 (35–84) min) (p=0.99).

No patients in the GRD group and 10 patients in the AIR group were treated for hypertension with intravenous labetalol en-route.


This study presents prospectively gathered population based data from a 16-month period in patients with suspected ischaemic stroke. We found ground transport to be faster than helicopter transport irrespective of distance to the stroke centre. No differences in mRS, short- or long-term mortality, were observed.

The data reported in this study reflect actual driving times for the ground ambulance group instead of estimated values, and we included only patients eligible for helicopter transport. Thus, we did not include local patients with short distances to the stroke centre. We believe this strengthens the study design aiming at a direct comparison of transport mode.

A number of limitations should be considered in interpretation of the results. We did not randomise between helicopter and ground transport, and we had no influence on the choices made by the dispatch centre. This could potentially lead to an uncontrolled selection of patients found appropriate for air transport by the dispatcher on call. The longer distance for AIR patients supports the assumption of a selection bias. The fact that the primary endpoint was missing in 25% of all patients is an additional limitation for interpreting the results, although we have no indication of a systematic pattern in these dropouts, mainly due to no registration of ‘contact time’.

The ongoing Phantom-S trial,7 investigates an urban ‘stroke emergency mobile unit’ equipped with CT scanner and neurologist, but only few studies have addressed the important issue of streamlining transport of rural stroke patients. After the European Coorperative Acute Stroke Study (ECASS)8 and the National Institute of Neurological Disorders & Stroke (NINDS)9 trials, other studies have elaborated on the potential role of helicopters for transportation of stroke patients. Some studies focused on the safety of air transport for patients having started thrombolysis treatment,10 and reported patients’ perception of care.3 These studies concluded that a helicopter has the potential to provide fast transport to stroke centres, and patients believed they had a benefit from helicopter transport, but they could not detect superiority over ground ambulance in terms of prognosis or time to definitive care. A study by Thomas et al11 suggested that time to CT could be shorter for patients arriving by helicopter, due to the helicopter crew's competence in stabilising and preparing the patient en-route, but no data were presented to support it. In 2003, Silliman et al12 published a prospective feasibility study evaluating a field-to-stroke centre protocol where the helicopter was dispatched subsequently to EMS providers’ stroke screening on-site. The authors presented data on 105 consecutive helicopter-transported patients with suspected stroke of whom 32% received thrombolysis. Transport distances ranged from 20–150 km, and mean onset-to-needle time was 152 min. These findings are consistent with ours, but, as the authors conclude, did not provide documentation to whether air was faster than ground transport. The Austrian stroke registry study13 found that directly referred patients transported by helicopters and physician-staffed ambulances had shorter symptom onset to stroke centre arrival times compared with ground ambulance without physician. The authors did not report transport distances, and the helicopter was dispatched primarily based on the alarm call, which presumably led to a dispatcher selection of the youngest patients, with most severe symptoms and the highest functional level prior to stroke onset in the air group. Accordingly, air-transported patients in this study had the highest rate of thrombolysis (34%). A recent analysis of 94 interhospital air- and 28 ground-transported patients showed helicopter transport to have longer response time (dispatch to referring hospital arrival), but shorter transport time and a significantly shorter total activation to stroke centre arrival time (53 vs 68 min).14 The transport distance in both groups was approximately 60 miles (∼100 km).

In our study, time from contact to stroke centre arrival was shorter in the GRD group. Since only 7% were interhospital transfers, the data mainly included field triaged patients. The AIR group clearly has the disadvantage of being activated when the ground ambulance may be ready to drive from the scene and the delay due to take-off and response time was apparently not compensated by the faster transport. Night time GRD patients did not arrive later than those transported in daylight. Hence, the diurnal selection did not seem to bias the results. Median times were comparable, despite significantly longer road distances in the AIR group. However, the number of interhospital transfers is low in this study, and the issue needs to be further explored before conclusions can be made.

There are several possible explanations for the delay in the AIR group. One factor could be the longer transport distance, but as illustrated by figure 3, the GRD was faster for all distances.

Figure 3

Relationship between distance (kilometre intervals) to stroke centre (X-axis) and time from contact to arrival at stroke centre (Y-axis) for ground (GRD) or helicopter (AIR) transported patients. N-values indicate numbers of patients in each group.

It must also be taken into account that the stroke centre is located in a relatively small city (population approx. 48 000) with good road accessibility and, therefore, with less advantage of air transport as compared with stroke centres in densely populated cities.

A high blood pressure can disqualify patients for thrombolysis upon arrival, and the helicopter physician-administered labetalol en-route in 10 cases. This could theoretically have led to reduced door-to-needle time, but this has not been shown in studies addressing this issue,15 and neither was it detectable in our study.

There were no deaths at 30-day follow-up in the AIR group, but mortality was not significantly different between groups. Neither did we observe significant differences in stroke severity or in any of the other recorded outcome variables. Thus, the shorter time to arrival in the GRD group did not seem to translate into an improved prognosis in our study. The exact clinical significance of a difference in median time of 13 min is uncertain, however it is believed that every effort should be made to minimise delay to thrombolysis treatment since a considerable number of neurons are lost every minute of ischemia.16 The lack of adjustment for additional confounding variables, such as comorbidity, is a clear limitation to the clinical outcome analysis, and needs to be explored in an appropriately powered study.

This is the first prospective and controlled study comparing prehospital time intervals for ground and air transport of stroke patients, at least to our knowledge. It provides a contribution to the discussion of streamlining stroke care, and indicates that dispatch protocols and stroke centre location are of foremost importance when the potential benefits of helicopter transport are assessed. Some questions are left unanswered in this study. It is not possible to extrapolate whether it was the actual mode of transport or the presence of a physician that has affected the transport times or clinical outcomes. Perhaps patients may benefit from a helicopter emergency medical service if this was primarily dispatched, thereby overcoming the many delaying steps in secondary dispatched air transport. It is also possible that a helicopter will allow thrombolysis in patients living in very remote places or on islands. We suggest future studies to elucidate optimisation of helicopter dispatch criteria and new techniques for prehospital identification of patients eligible for thrombolysis and thrombectomy.

In conclusion, we found that primarily dispatched ground ambulance transport was associated with significantly shorter time from contact to triaging neurologist to arrival in the regional stroke centre as compared with a secondarily dispatched helicopter.


View Abstract


  • Contributors RH, JS and LSR were responsible for conception, study design, analysis and interpretation of data. RH, JG, HQD, JH gathered data. RH drafted the manuscript and all authors revised it critically and approved final version to be submitted for publication.

  • Funding This study was funded by the TrygFoundation.

  • Competing interests RH and LSR's institution have received research grants from the TrygFoundation as funder of the submitted work. LSR's institution has received research grants from AMBU. JS has done consultancy as national investigator in the Euromax study for Medicines Company. The other authors have no conflict of interests to declare.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.