Article Text

Download PDFPDF

Original research
Severe hemiparesis as a prehospital tool to triage stroke severity: a pilot study to assess diagnostic accuracy and treatment times
  1. Rishi Gupta,
  2. Marissa Manuel,
  3. Kumiko Owada,
  4. Samish Dhungana,
  5. Leslie Busby,
  6. Brenda A Glenn,
  7. Debbie Brown,
  8. Susan A Zimmermann,
  9. Christopher Horn,
  10. Dustin Rochestie,
  11. Joseph T Hormes,
  12. Andrew K Johnson,
  13. Ahmad Khaldi
  1. Wellstar Neurosciences Network, Wellstar Health System, Kennestone Hospital, Marietta, Georgia, USA
  1. Correspondence to Dr R Gupta, Department of Neurosurgery, Wellstar Health System, 61 Whitcher Street, Suite 3110, Marieta, GA 30060, USA; Rishi.gupta{at}


Introduction With the publication of the recent trials showing the tremendous benefits of mechanical thrombectomy, opportunities exist to refine prehospital processes to identify patients with larger stroke syndromes.

Materials and methods We retrospectively reviewed consecutive patients who were brought via scene flight from rural parts of the region to our institution, from December 1, 2014 to June 5, 2015, with severe hemiparesis or hemiplegia. We assessed the accuracy of the diagnosis of stroke and the number of patients requiring endovascular therapy. Moreover, we reviewed the times along the pathway of patients who were treated with endovascular therapy.

Results 45 patients were brought via helicopter from the field to our institution. 27 (60%) patients were diagnosed with an ischemic stroke. Of these, 12 (26.7%) were treated with mechanical thrombectomy and 6 (13.3%) with intravenous tissue plasminogen activator alone. An additional three patients required embolization procedures for either a dural arteriovenous fistula or cerebral aneurysm. Thus a total of 15 (33%) patients received an endovascular procedure and 21/45 (46.7%) received an acute treatment. For patients treated with thrombectomy, the median time from first medical contact to groin puncture was 101 min, with 8 of the 12 patients (66.7%) being discharged to home.

Conclusions We have presented a pilot study showing that severe hemiparesis or hemiplegia may be a reasonable prehospital tool in recognizing patients requiring endovascular treatment. Patients being identified earlier may be treated faster and potentially improve outcomes. Further prospective controlled studies are required to assess the impact on outcomes and cost effectiveness using this methodology.

  • Stroke
  • Thrombectomy

Statistics from

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.

With the recent publications confirming the benefit of mechanical thrombectomy in patients with large vessel occlusions (LVO),1–5 new challenges have arisen with regard to how best to triage patients from the field. The current construct of taking patients to the closest primary stroke centers may delay care for patients harboring an LVO6 and then subsequently require transfer to an endovascular capable facility. Prehospital scales, such as the Los Angeles Motor Scale (LAMS)7 and the Rapid Arterial Occlusion Evaluation (RACE) scale,8 have been proposed and assessed for accuracy in identifying patients with LVO. The RACE score integrates cortical clinical signs, such as aphasia and gaze preference, which may be challenging for prehospital providers to assess, whereas the LAMS score focuses on pure motor weakness.

We worked with emergency medical service (EMS) providers in our region on assessing patients with complete hemiplegia, similar to the LAMS score, to assess the effectiveness in detecting LVO. Patients were brought via helicopter to our facility directly from the scene and we analyzed the accuracy of detection of patients who would require tertiary stroke care, such as an endovascular procedure or open surgical procedure.


Patients who were brought directly via scene flight from our region to our comprehensive stroke center from December 1, 2014 to June 5, 2015, were analyzed. Preceding this pilot project, we performed local outreach efforts with EMS providers and regional EMS directors on assessing patients with complete hemiplegia or severe hemiparesis as being indicative of patients with potential LVO. Patients were flown to our institution if the patient was either equidistant or closer via helicopter compared with ground transportation to the local facility. We reviewed consecutive patients that were directly brought from the scene to our institution via helicopter. Patients were screened by the medics in the field, and if there was evidence of severe hemiparesis, they contacted the helicopter services to transport the patient to our institution, which is the closest comprehensive stroke center for the region. We did not include patients who were brought by ground for this analysis to allow for a homogenous analysis of patients being brought in from rural areas. Additionally, we were unable to determine the number of patients who were not able to be brought to us due to weather concerns.

The helicopter services would contact our emergency room command center to notify the stroke service of the estimated arrival time which would in turn activate our inhouse stroke processes, as we have previously published.9

After institutional review board approval, we retrospectively reviewed the patient demographic information, final diagnosis, endovascular procedures performed, and time metrics for those patients treated for acute ischemic stroke. We analyzed time measures of first medical contact, time of emergency room arrival, time of groin puncture, and time of reperfusion.


A total of 45 patients were brought to our institution via helicopter services from the field. Of these patients, 27 (60%) were identified as having an ischemic stroke, 6 (13.3%) were found to have an intracerebral hemorrhage, 2 of which were due to a dural arteriovenous fistula, 7 (15.5%) were diagnosed with a transient ischemic attack, 1 (2.2%) patient had a subarachnoid hemorrhage and 4 (8.9%) patients had seizures or altered mental status due to an infections or metabolic condition. Thus the overall detection rate of a cerebrovascular etiology was 38/45 (84.5%). Twelve (26.7%) patients were treated with mechanical thrombectomy, 7 of whom were treated with a combination of intravenous tissue plasminogen activator (tPA) and thrombectomy. An additional 6 (13.3%) patients were treated with intravenous tPA alone. When combining patients who underwent a thrombectomy with those who were treated for an aneurysm or dural arteriovenous fistula, a total of 15/45 patients (33%) required an endovascular procedure. If patients who received intravenous tPA alone were also included, then a total of 21/45 (46.7%) patients received an acute treatment.

Table 1 summarizes the demographic information of the 12 patients treated with mechanical thrombectomy. Only one patient experienced an asymptomatic subarachnoid hemorrhage of those treated with mechanical thrombectomy. The last known normal to groin puncture for these patients was 210±83 min. The mean time from first medical contact to puncture was 101±24 min. A total of 8 patients (66.7%) were discharged to home, 2 (16.7%) to inpatient acute rehabilitation, and 2 (16.7%) to a hospice. The patients discharged to home all had a National Institutes of Health Stoke Scale (NIHSS) score of ≤3. For the two patients discharged to inpatient rehabilitation, the first patient had a reduction of their NIHSS from 17 to 8 while the second patient went from 25 to 11. The two patients discharged to hospice did not exhibit a significant improvement in their NIHSS and were both octogenarians.

Table 1

Summary of patients arriving via scene flight treated with mechanical thrombectomy


Prehospital scales to assess stroke severity require simplicity, reproducibility, and optimize the clinical benefit for the patient. Prehospital providers may have limited experience in caring for stroke patients, and the educational forum to disseminate information in an efficient manner is currently challenging. As such, developing a simple elegant tool that prehospital providers can discern severe strokes from less severe ones would potentially benefit patients in being taken to an appropriate facility. Moreover, the prehospital scale may incidentally identify patients with hemorrhagic diseases that may also benefit from being treated at an endovascular capable facility. Our pilot study demonstrated that 26.7% of patients were found to have an LVO that was treated with thrombectomy, while 46.7% received an acute treatment.

Previous studies have focused on refining the processes in the hospital, but have excluded the prehospital component of time lost.10 In our pilot study, we found that the use of hemiplegia alone as a marker of stroke severity allowed for an acute treatment in 46.7% of patients brought from the field. Other scales, such as the RACE score, have shown that of the patients with a score of ≥5, a total of 19% of them underwent an endovascular intervention. The RACE scale correlates with the NIHSS score, and also the authors note that with a higher RACE score, patients were more likely to have an LVO or hemorrhage.8 We found that with the use of a less complex algorithm of hemiplegia, 26.7% of patients underwent thrombectomy for ischemic stroke.

The recent published trial showing the benefits of mechanical thrombectomy have considered inhospital processes and worked towards optimization of these steps to reduce times to reperfusion. In particular, the Endovascular Treatment for Small Core and Proximal Occlusion Ischemic Stroke (ESCAPE) trial showed a median time from CT to puncture of 51 min.1 Other studies have shown that approximately 25% of the time from first medical contact to reperfusion occurs during the procedural component. Thus roughly 75% of the time is lost in the prehospital phase or in the hospital phase of care.6

The currently published studies did not report the first medical contact times, but differences in outcomes and times occurred across the published trials. For instance, in the Multicenter Collaboration for Endovascular Treatment of Acute Ischemic Stroke in The Netherlands (MRCLEAN) trial, the median time from stroke onset to groin puncture was 260 min3 while in ESCAPE it was 185 min.1 This is reflected partially in the study design of MRCLEAN where patients were given intravenous tPA at an outside institution first and then shipped for endovascular therapy. In contrast, the ESCAPE trial employed a more rapid treatment algorithm with fewer transferred patients. In part, this may account for why patients in ESCAPE had better clinical outcomes compared with those in MRCLEAN. In our cohort, 66.7% of patients were discharged to home and 83.3% had a significant improvement in their NIHSS score, which in part was due to their early treatment.

The ability to recognize a patient with a large stroke syndrome in the prehospital phase and bring the patient to the right destination the first time will ultimately save time, which in turn will improve outcomes. Currently, the ability to deliver intravenous tPA to a patient meeting the inclusion and exclusion criterion is important, but other factors must be considered in the algorithm. For instance, if one hospital's door to needle time is 55 min while another institution it is 35 min, the additional transportation time loss may be mitigated under such conditions. Transparency of such metrics to EMS may help allay some of the concerns regarding tPA delivery. Moreover, patients with severe strokes who are beyond 3.5 h from last known normal or are currently taking anticoagulants would not qualify for tPA and thus may be candidates to bypass centers unable to offer endovascular treatments.

There are several limitations to our analysis. First, this was a retrospective study with the inherent flaws of this study design. Second, we do not have data on patients who were unable to be air lifted due to weather conditions or other parameters that precluded safe transportation of the patient. Third, patients may have been taken to a primary stroke center with severe hemiparesis and thus there may be a bias in the cohort of patients that were brought to our institution. Fourth, the small sample size that we present may under or over represent the true accuracy of this tool. Lastly, we do not have a control group to assess if outcomes were improved using this methodology.

Despite these limitations, this pilot study shows that training prehospital providers to identify severe hemiparesis may be a simple and easy tool to train with a moderate detection rate of patients with LVO, but a reasonable tool to identify patients who will receive acute therapy. Further study in a prospective controlled manner is required to refine prehospital triage along with its potential impact on outcomes. Additionally, the cost effectiveness of such approaches need to be considered, particularly given the upfront costs of helicopters that may defray longer term care costs of rehabilitation and nursing facilities.



  • Contributors RG, MM, KO, and BAG: conception of the research protocol and writing of the manuscript. BAG, SAZ, CH, DR, and DB: data collection and analysis for the manuscript. AK, JTH, and RG: statistical analysis. SAZ, LB, KO, SD, RG, and AKJ: critical revision of the manuscript.

  • Competing interests RG: consultant for Stryker Neurovascular, Covidien, and Rapid Medical; research funding/grant support from Penumbra, Stryker Neurovascular, Covidien, Zoll, and Wellstar Foundation; associate editor for Journal of Neuroimaging, Journal of Neurointerventional Surgery, and Interventional Neurology; royalties from UpToDate.

  • Ethics approval The study was approved by the institutional review board.

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

  • Data sharing statement Data from this study can be shared on request to the corresponding author with an approved institutional review board protocol.