Dr Daria Antipova (MBChB, MSc, PhD)

Title of PhD Project: Designing a pre-hospital scoring system for the prediction of haemorrhagic and ischaemic stroke

Dates: April 2017 to March 2020

Description:

Stroke is one of the major causes of death and disability in adults worldwide. Once a stroke patient arrives at hospital they usually have a CT scan of their brain to determine whether the stroke is caused by a blocked artery or bleeding to help choose the right treatment. Some stroke patients might benefit greatly if they were treated before getting to hospital, particularly in rural areas where transit times are long.

The aim of our research was to develop a new type of scoring system to predict stroke type without CT scanning. Stroke patients admitted to Raigmore Hospital (Inverness) and the Queen Elizabeth University Hospital (Glasgow) were recruited.  Their brains were scanned quickly using an ultrasound machine.  They also got a CT scan in line with standard hospital practice. A scoring system was developed based upon what we saw with ultrasound, together with other important demographic and clinical data.

Main findings:

We recruited 115 patients suspected of having suffered a stroke: 66 participants had a stroke caused by blocked brain vessel, among them 15 patients had blockage of a large artery; 19 patients had brain bleed; 15 participants suffered a “mini-stroke” (TIA) and 15 patients had other explanation for their symptoms, for example, a severe migraine, which sometimes can mimic a stroke. No patients withdrew from the study and only one participant was unable to tolerate the ultrasound scan. Age varied from 26 to 92 years, and 57% were males.

It took on average 20 minutes to complete the ultrasound head scan. We recruited 2% of participants within 4.5 hours from stroke onset, 54% - within 6 to 24 hours, and 44% were recruited from 24 to 72 hours. Time from stroke onset to ultrasound varied from 2 hours 28 minutes to 72 hours (on average 24 hours). All patients except one had ultrasound scanning after the CT. Time interval between CT head scan and ultrasound did not exceed 24 hours and on average was 15 hours 53 minutes. 13% of patients had been given clot-busting treatment before the ultrasound scan of the head.

In 17% of patients, of whom 95% were females, ultrasound failed to provide any useful information because of the age-related bone changes.

We developed two models to predict the likely stroke type:

  1. Brain bleed is more likely if the ultrasound scan shows the area of bleed and/or shift of the central parts of the brain to either side, if the patient takes blood thinning medication, has high blood pressure or has signs suggestive of a damage to the covering layer of the brain (speech problems, poor eye movements, poor attention to one side of the body or loss of a part of the vision). We got a higher percentage of correct diagnoses when both ultrasound and clinical assessment were used compared to clinical assessment alone (89% vs. 65%), particularly in the first 24 hours from stroke onset - 50% sensitivity and 100% specificity.
  2. Blockage of a large vessel in the brain is more likely if the ultrasound demonstrates poor blood flow on one side of the brain, the patient is more elderly and in case of the positive signs of damage to the covering layer of the brain. Again, the diagnosis was more accurate when we used both ultrasound and clinical assessment (91% vs. 82%), within the first 24 hours and if the ultrasound was done before clot-busting treatment (71% sensitivity, 97% specificity).

We would like to confirm our findings in a bigger study but these pilot results suggest that this model using ultrasound head scan plus simple clinical examination might be particularly useful in areas with long transport times to the hospitals.

Dissemination:

We presented our project at the following conferences:

  • Conversations from the North, University of Aberdeen, August 2017
  • NIHR Stroke Research workshop, Cambridge, September 2017
  • Scottish Association of Neurological Sciences meeting, Dunkeld, November 2017
  • Edinburgh Stroke Winter School, 19-21 February 2018
  • BSRM Scottish chapter, Aberdeen, November 2019
  • University of Aberdeen Postgraduate Students' conference, Aberdeen, November 2019
  • NHS Highland RD&I Conference, Inverness, 24 November 2017
  • Rethinking Remote, Inverness, May 2018
  • NHS Highland RD&I Conference, Inverness, 4 October 2018
  • NHS Highland RD&I Conference, Inverness, 3 October 2019

Daria received a prize for the best presentation at the PGR SMMSN Winter Conference which took place on 28-29 November. She also received third prize for her poster presentation at the BSRM (Scottish Branch) meeting.

Publications:

Daria Antipova, Leila Eadie, Ashish Macaden, Philip Wilson “Diagnostic accuracy of clinical tools for assessment of acute stroke: a systematic review” BMC Emerg Med. 2019 Sep 4;19(1):49. doi: 10.1186/s12873-019-0262-1.

Daria Antipova, Leila Eadie, Ashish Stephen Macaden, Philip Wilson “Diagnostic value of transcranial ultrasonography for selecting subjects with large vessel occlusion: a systematic review” Ultrasound J. 2019 Oct 22;11(1):29. doi: 10.1186/s13089-019-0143-6.

Supervisory team:

  • Professor Philip Wilson, Director of Centre for Rural Health
  • Dr Leila Eadie, Research Fellow, Centre for Rural Health
  • Dr Ashish Stephen MacAden, Consultant in Rehabilitation Medicine and Stroke, Raigmore Hospital, NHS Highland
  • Dr Stephen Makin, Senior Clinical Lecturer, Centre for Rural Health

Additional Short Study:

Study title

Rapid assessment using transcranial ultrasound: pilot study with novice users

Dates

May – July 2019

Overview and aims

Responding to stroke in remote and rural areas is difficult. It can take a long time for an ambulance to reach patients and then transport them back to hospital to work out whether strokes are being caused by bleeding or by blood clots blocking the arteries.  Acting quickly is very important as the longer the patient waits for assessment and treatment, the poorer their outcome will be.

 

The most common methods of looking for signs of occluded vessel or bleeding in the brain use large scanners which are bulky and not readily portable. Ultrasound scanners can be very portable and can therefore be used in ambulances and other situations outside of a hospital.

 

This study contributes to the ongoing project ‘Designing a prehospital scoring system for the prediction of haemorrhagic stroke and ischaemic stroke’ (https://www.abdn.ac.uk/iahs/research/crh/projects/daria-antipova-phd-1257.php).

It is looking at the practicality and effectiveness of brief training provided for novice transcranial ultrasound users to ensure that they are able to acquire a pre-specified set of images and video clips of satisfactory quality which could then be transferred for expert interpretation and decision making.

 

We are aiming to recruit a maximum of 15 volunteers among qualified clinicians or students of Nursing or Medicine. Participants will receive three-hour training  on transcranial grey-scale and colour-coded duplex sonography to visualise cerebral landmarks and major cerebral blood vessels. This will involve transcranial ultrasound scanning of healthy volunteers through the temporal bone windows.

 

If the training programme is successful, it could be used in the future to provide training for a wider audience of GPs working in remote and rural areas, community physicians and ambulance paramedics. This could also help ensure speedier and more effective prehospital treatment.

Funding

This study is run and funded by the University of Aberdeen.

 

Contact

Daria Antipova on 01463 255902 or email daria.antipova@abdn.ac.uk