Abstract
1D.1
Campbell B1, Christensen S2, Butcher K3, Parsons M4, Desmond P2, Gordon I5, Barber A, Alan P6, Levi C5, Bladin C7, De Silva D1, Peeters A8, Donnan G9, Davis S1, EPITHET Investigators
1 Department of Neurology, Royal Melbourne Hospital, University of Melbourne, Parkville, Melbourne, VIC, Australia
2 Department of Radiology, Royal Melbourne Hospital, University of Melbourne, Parkville, Melbourne, VIC, Australia
3 Faculty of Medicine and Dentistry, University of Alberta, Edmonton, AB, Canada
4 Hunter Medical Research Institute, Centre for Brain and Mental Health Research, John Hunter Hospital, University of Newcastle, Newcastle, NSW, Australia
5 Department of Mathematics and Statistics, University of Melbourne, Parkville, Melbourne, VIC, Australia
6 Department of Medicine, University of Auckland, Auckland, New Zealand
7 Box Hill Hospital, Monash University, Melbourne, VIC, Australia
8 University Hospital St Luc, Brussels, Belgium
9 Florey Neuroscience Institutes, Melbourne, VIC, Australia
Background: Currently DWI volume is the best MRI predictor of haemorrhagic transformation (HT) after thrombolysis. We compared DWI and ADC with very low cerebral blood volume (VLCBV) using perfusion MRI from the EPITHET study.
Methods: Relative CBV and ADC thresholds were derived from the normal hemisphere along with absolute ADC thresholds. The volume below threshold was calculated within the acute DWI lesion. HT was graded as per ECASS-II.
Results: In 91 patients, parenchymal haematomas (PH) occurred in 13 (four symptomatic), asymptomatic haemorrhagic infarcts (HI) in 31. All VLCBV thresholds predicted PH with optimal ROC parameters at the 25th percentile (AUC 0·73 for HT vs. no HT and 0·78 for PH vs. no PH) compared with DWI (AUC any HT 0·71, PH 0·73). No absolute or relative ADC threshold surpassed VLCBV (ADC < 550 best AUC any HT 0·71, PH 0·71). VLCBV 2·5 > 2 ml had 43% PH risk in tPA-treated patients (LR 16). In multivariate analysis with known clinical HT predictors, each of VLCBV, DWI and ADC < 550 remained significant alone but only VLCBV was significant combined with DWI and ADC < 550.
Conclusions: VLCBV predicts HT following thrombolysis better than DWI/ADC volume in a large patient cohort and could improve pre-thrombolysis decision-making.
1D.2
Tu H1, Campbell B1, Christensen S2, Butcher K3, Collins M4, Parsons M5, Desmond P2, Barber A6, Levi C5, Bladin C7, De Silva D1, Peeters A8, Donnan G9, Davis S1, for the EPITHET Investigators
1 Department of Neurology, Royal Melbourne Hospital, University of Melbourne, Parkville, Melbourne, VIC, Australia
2 Department of Radiology, Royal Melbourne Hospital, University of Melbourne, Parkville, Melbourne, VIC, Australia
3 Faculty of Medicine and Dentistry, University of Alberta, Edmonton, AB, Canada
4 Department of Mathematics and Statistics, University of Melbourne, Parkville, Melbourne, VIC, Australia
5 Hunter Medical Research Institute, Centre for Brain and Mental Health Research, John Hunter Hospital, University of Newcastle, Newcastle, NSW, Australia
6 Department of Medicine, University of Auckland, Auckland, New Zealand
7 Box Hill Hospital, Monash University, Melbourne, VIC, Australia
8 University Hospital St Luc, Brussels, Belgium
9 Florey Neuroscience Institutes, Melbourne, VIC, Australia
Background and Aims: Previous studies have shown that patients with stroke and atrial fibrillation (AF) have worse outcomes. We tested whether differences in stroke size and evolution could explain the adverse effect of AF in the EPITHET trial.
Methods: EPITHET was a randomised-controlled trial investigating the use of IV tPA 3–6 h after onset involving 101 patients, 42 with AF. Patients were assessed with clinical scores (NIHSS, mRS) and multimodal MRI pretreatment, 3–5 days and 3 months poststroke. Haemorrhagic transformation was classified according to the ECASS criteria.
Results: At baseline, patients with AF were older (78·5 vs. 73 years, P = 0·024), with more severe impairment (NIHSS 16 vs. 11 P = 0·006) and larger infarct volume (29·29 vs. 14·37 ml P = 0·039). At outcome, patients with AF had higher mortality (31 vs. 12·3%, P = 0·041), larger infarcts (53·63 vs. 16·21 ml, P = 0·047), and higher haemorrhagic transformation rates (63 vs. 31%, P = 0·012). However, the effect of AF on mortality became nonsignificant (OR 2·37, CI 0·79–7·10, P = 0·124) after adjusting for baseline variables including age, tPA use, stroke severity, infarct size, and blood glucose.
Conclusion: Although AF is associated with larger infarcts and worse outcomes, further study is needed to confirm whether AF is an independent predictor of the outcome.
1D.3
Gawarikar Y1, Miteff F1, Selmes C1,2, Parsons M1,2, Spratt N1,2, Levi C1,2
1 Department of Neurology, John Hunter Hospital (JHH), University of Newcastle, Newcastle, NSW, Australia
2 Priority Research Centre for Brain 8 Mental Health, University of Newcastle, Newcastle, NSW, Australia
Hypothesis & Aim: This study aimed to evaluate any association between brain imaging findings suggestive of embolism and severity of shunting on contrast transcranial Doppler (cTCD) in patients with patent foramen ovale (PFO) and a history of ischaemic stroke. We hypothesised that severity of shunting would be associated with imaging evidence of prior brain infarction.
Methods: Cohort study (on-going) of 39 patients with recurrent cryptogenic focal neurologic attacks where PFO [identified by contrast transthoracic echocardiography (TTE) or transoesophageal echocardiography (TOE)] was considered implicated in stroke causation. The severity of shunting across the PFO was assessed with cTCD (no/mild vs. moderate/severe) using published criteria [1]. Magnetic resonance imaging (MRI) of the brain was performed to identify ischaemic lesions.
Results: Thirty-one of the 39 patients (79·48%) had a PFO with moderate/severe shunting on cTCD. The presence of MR-detected brain infarction was significantly greater among patients with PFO and moderate/severe shunting on cTCD (18 of 31; 58%) compared with patients with no/mild shunting on cTCD (one of eight, 12·5%; P = 0·0436).
Conclusions: Patients with moderate to severe shunting on cTCD are more likely to have brain infarction on MRI than those with mild shunting. cTCD may be used to more effectively quantify shunting and to identify high stroke risk PFO patients.
1D.4
Lillicrap T1,2,3,5, Hudson S1,2, Stanwell P1,3,6,7, Parsons M1,2,3,4, Spratt N1,2,3, Levi C1,2,3,5
1 Priority Centre for Brain 8 Mental Health Research, Hunter Medical Research Institute, University of Newcastle, Newcastle, NSW, Australia
2 University of Newcastle, Newcastle, NSW, Australia
3 John Hunter Hospital, University of Newcastle, Newcastle, NSW, Australia
4 Stroke Society of Australasia, VIC, Australia
5 National Stroke Foundation of Australasia, VIC, Australia
6 Brigham and Women's Hospital, Boston, MA, USA
7 Harvard Medical School, Boston, MA, USA
Background: Hypothermia is a promising neuroprotective therapy for acute ischaemic stroke and phase II clinical trials are underway using endovascular cooling catheters to induce whole-body hypothermia. While these devices are proven to reduce blood temperature, no study has proven that this method of systemic cooling actually reduces the temperature of ischaemic brain tissue.
Aim: To assess whether MR spectroscopy (MRS) can reliably measure brain metabolite chemical shift as an index of measured temperature change in a commercial standard brain ‘phantom’.
Methods: A commercial phantom containing a number of brain metabolites at physiological concentrations underwent MRS at temperatures between 30° and 40°C. The proton resonance frequency (PRF) of water (temperature dependent) was compared with the PRF of NAA (temperature independent). This provided an internal correction for magnetic field in homogeneity and other factors besides temperature that may influence the PRF of water.
Results: The relationship between the water–NAA frequency shift and temperature in the phantom was T = −1·4801F +285·77 (R2 = 0·96, P < 0·0001) with F in Hz and T in °C. We were able to estimate temperature to within ±1·1°C (95% confidence intervals) of the measured brain phantom temperature using this scanning method in vitro.
Conclusion: Given that systemic cooling protocols reduce body temperature by 4°C, MRS may be a clinically practical method of noninvasively measuring whether brain temperature drops in parallel with body temperature.
1D.5
Quain D1,2,3, Parsons M1,2,3, Miteff F2,3, Mcelduff P1,3, Levi C1,2,3
1 University of Newcastle, Callaghan, NSW, Australia
2 Department of Neurology, John Hunter Hospital, University of Newcastle, Newcastle, NSW, Australia
3 Hunter Medical Research Institute, John Hunter Hospital, University of Newcastle, Newcastle, NSW, Australia
Background: In MCA occlusion flow diversion from the internal carotid artery to the anterior cerebral artery (ACA) is thought to represent activation of leptomeningeal collateral pathways and can be detected using transcranial Doppler ultrasound (TCD).
Method: We sought to determine the relationship between TCD flow diversion, CTA collateral grade and penumbral volumes in patients with anterior circulation stroke presenting within 6 h of symptom onset. TCD flow diversion was considered present when the ipsilateral ACA mean velocity was at least 20% greater than the contralateral ACA mean velocity. Leptomeningeal collateralisation was measured using CT angiography. Infarct core and penumbral volumes were measured from perfusion CT, with the CTP ‘mismatch ratio’ = mean transit time lesion/cerebral blood volume lesion.
Results: Significant TCD flow diversion (FD) was present in 20/46 (43%) patients. When patients with no vessel occlusion were excluded, 11/14 (79%) in the TCD FD-positive group had a good collateral grade compared with 4/16 (25%) in the FD-negative group (P = 0·009). Patients with FD had substantially larger CTP mismatch ratios (median 120 vs. 1·9 in non-FD group, P = 0·002) and penumbral volumes (median 93·5 vs. 53·8 ml in non-FD group, P = 0·001).
Conclusion: Anterior cerebral artery flow diversion detected by TCD is associated with better collateralisation on CTA, and in turn, with larger penumbral volumes in patients with anterior circulation stroke.
1D.6
Karimi Galougahi K, Stewart T, Choong CYP, Storey C, Yates M, Tofler GH
Departments of Cardiology and Neurology, Royal North Shore Hospital, University of Sydney, NSW, Australia
Introduction: Assessment for cause of stroke is a common indication for transoesophageal echocardiogram (TOE). Although an abnormality is frequently found, it remains uncertain whether the findings alter management. Also, the role of transthoracic echocardiogram (TTE) before or instead of TOE is not well defined.
Aims: We sought to determine the use of TTE before TOE, the outcome of the TOE, and its impact on management.
Methods: We reviewed the records and echocardiography results of 100 consecutive patients who underwent TOE for any reason at a tertiary hospital. In 35 subjects (35%), the indication was to evaluate for source of stroke. Among these, we determined risk factors for stroke, if a TTE was performed before TOE, the results of the TOE, and its effect on management.
Results: The mean age of stroke patients was 64·6 years (17–90) and 49% were women. Eighty per cent had at least one risk factor for stroke and 17% had atrial fibrillation. A TTE was performed before the TOE in 40%, and found an abnormality in 14% (two of 14). The TOE showed an abnormality in 71% of patients; 54% had aortic atheroma; 17% PFO; 14% spontaneous echo contrast; 6% left atrial appendage thrombus, 3% left ventricular thrombus and 3% vegetation. In only one patient (3%), management was altered based on abnormal TOE findings. This was the commencement of anticoagulation in the presence of a PFO and atrial septal aneurysm.
Conclusion: An abnormality on TOE, although common (71%), altered management in only 3% of subjects referred for stroke assessment. Its role requires further consideration.