HEart and BRain interfaces in Acute ischemic Stroke (HEBRAS) – rationale and design of a prospective oberservational cohort study

The Center for Stroke Research Berlin (CSB) is the sponsor of the investigator initiated HEBRAS study, which is supported by a grant from the German Federal Ministry of Education. The study was approved by the scientific advisory board of the CSB in September 2013, and by the Ethics Committee of the Charité - Universitätsmedizin Berlin, Germany (EA2/033/14) in March 2014. All study procedures are carried out in accordance with the principles of Good Clinical Practice and the Declaration of Helsinki. All study patients have to fulfill the study entry criteria listed in Table 1 including written informed consent.

Study patients will receive routine diagnostic procedures in hospital (defined as stroke unit monitoring (according to clinical needs), echocardiography, ultrasound of the brain-supplying arteries, and 24-hour Holter-ECG (if indicated according to the treating physician)). Moreover, study patients will undergo an enhanced MRI work-up (comprising brain as well as cardiac MRI, MR angiography of the aortic arc and the supra-aortic, brain-supplying arteries) combined with Holter-ECG of up to 10 days (including a maximum of 5 days after hospital discharge). For the primary outcome, the results of routine diagnostic care will be compared to study-related diagnostics. Results of the ECG core lab analysis will be provided to study patients and treating physicians. Study patients will be followed up for cardiovascular outcomes e.g., recurrent stroke, major vascular events, death and degree of dependency at 3 and 12 months after enrollment.

Etiology will be determined according to TOAST [18] and ESUS criteria [4]. The following diagnostic findings will be regarded as pathologic and relevant to stroke etiology: atrial fibrillation of any type, sick sinus syndrome, bacterial or nonbacterial thrombotic endocarditis, ipsilateral carotid or intracranial stenosis of ≥ 50 % according to NASCET [19], left atrial or left ventricular thrombus, atrial myxoma, acute myocardial infarction resulting in wall motion abnormalities, akinetic left ventricular segment, dilated cardiomyopathy, systolic dysfunction with reduced left ventricular ejection fraction (<45 %), aortic plaque of ≥ 4 mm.

In HEBRAS, enhanced diagnostic MRI work-up is defined as brain MRI, cardiac MRI and MR-angiography of the aortic arc and the brain-supplying arteries. All MR examinations are performed on a 3 Tesla MR scanner (TIM TRIO, Siemens, Erlangen, Germany). Cerebral MRI mainly consists of a T2*-sensitive sequence (TR 620 ms, TE 20 ms, flip angle 20°, FOV 220 mm, slice thickness 5 mm) to exclude intracerebral hemorrhage, a diffusion weighted (DWI) sequence (TR 7600 ms, TE 93 ms, FOV 230 mm, slice thickness 2.5 mm) for detection of ischemic tissue, and a 3D time-of-flight (TOF) angiographic sequence (TR 22 ms, TE 3.86 ms, flip angle 18°, FOV 200 mm, slice thickness 0.65 mm) for imaging of intracranial vessels. For cardiac imaging, initially a dark-blood–prepared HASTE (half-Fourier acquisition single-shot turbo spin-echo) sequence (TR 750 ms, TE 49 ms, Flip angle 160°, Slice thickness 5 mm) is acquired, followed by 2 chamber as well as 4 chamber view acquisitions using a cine SSFP sequence (TR 40.56 ms, TE 1.48 ms, flip angle 50°, slice thickness 5 mm) with prior scouting for individual frequency adjustment. Post-contrast imaging is performed after intravenous injection of 0.1 mmol/kg body weight Gadobutrol (Gadovist®, Bayer Schering Pharma, Berlin, Germany) with post-contrast cine imaging parameters held constant. For angiographic imaging of the aortic arch as well as extracranial cerebral arteries, a three-dimensional TWIST (Time-resolved angiography With Interleaved Stochastic Trajectories) sequence was used (TR 2.36 ms, TE 0.99 ms, Flip angle 20°, Slice thickness 1.40 mm). Finally, myocardial late enhancement is characterised employing a turbo FLASH (Fast Low Angle SHot) sequence (TR 750.00 ms, TE 1.97 ms, Flip angle 20°, Slice thickness 8 mm).

The following stratification of lesion localization is planned: left / right anterior cerebral artery, left / right middle cerebral artery, left / right posterior cerebral artery, left / right anterior choroideal artery, left / right insular cortex, left / right basal ganglia, left / right thalamus, left / right cerebellum, left / right brainstem, left / right medulla oblongata.

To bypass restrictive a priori assumptions on a specific relationship between stroke lesion and laboratory changes, an additional voxel-by-voxel analysis of lesion location and troponin-elevation will be conducted employing statistical principles developed for functional brain imaging [20, 21].

Additional ECG recording will be performed by using the portable CardioMem®4000 (GETEMED AG, Teltow, Germany). Study ECG will be started as soon as possible after enrolment and will be continued for a maximum of 10 days, i.e., two consecutive ECG measurements of five days. When patients are discharged before the first five days are completed, ECG recorder will be changed at discharge, resulting in an overall duration of less than ten days in Holter-ECG.

Study ECG data will be transmitted to the cardiac core lab at Department of Cardiology and Pneumology, Georg-August-University Göttingen, Germany for analysis. Raters analysing ECG-data will be blinded for clinical data including infarct localization. Core lab analysis will be provided to the respective study patient and the patient’s treating physician (if indicated by clinical assumption).

The following definitions will be used throughout the study: Atrial fibrillation: Absolute arrhythmia lasting ≥ 30 s as defined in the current ESC guideline [22]. Short atrial tachycardia: Atrial tachycardia lasting < 30 s and consisting of ≥ 6 conducted premature activations. Excessive supraventricular ectopic activity: More than three premature atrial complexes per hour [23]. Atrial flutter will be reported separately, and the study specifies to initiate anticoagulation for atrial flutter in the same way as for AF. The analysis will include: heart rhythm; PQ intervall [s], minimal [bpm], maximal [bpm] and mean heart rate (HR) [bpm], and measures of Heart rate variability (HRV) such as the standard deviation of beat-to-beat (NN) intervals (SD-NN), Root Mean Square of Successive Differences (rMSSD), HRV-Triangular Index (HRV-TI), low and high frequency (LF and HF, respectively) band and LF/HF Index [24].

Baseline assessment of all study patients will include: a detailed analysis of patient demographics (also including time of stroke onset, hospital admission as well as admittance to the ward), clinical characteristics (including the National Institutes of Health Scale Score [25] and the modified Rankin Scale score [26], vital signs, medical history and concomitant diseases, extend and results of inpatient ECG recording, laboratory results (all data of clinical routine assessment plus high-sensitivity (hs) troponin and NT-proBNP), results of imaging (head CT/MRI, carotid ultrasound of brain-supplying arteries, echocardiography) and TOAST classification at hospital discharge (before evaluation of the additional ECG monitoring), in hospital treatment (i.e., thrombolysis, ventilation etc.) as well as in hospital complications (such as pneumonia, recurrent stroke or myocardial infarction).

Follow-up will be done 3 months and one year after the index stroke by a telephone interview conducted by trained personal of the Center for Stroke Research Berlin, Germany. The following data will be assessed: medication, recurrent ischemic stroke and other major vascular events, major bleeds, and the mRS score. Additionally, (all-cause) death and date of death will be registered using information from registration offices.

The primary outcome of the HEBRAS study is the rate of pathologic findings relevant to stroke etiology in patients with acute ischemic stroke obtained by enhanced diagnostic MRI work-up combined with prolonged Holter-monitoring in comparison to findings obtained by the routine diagnostic work-up in this cohort of patients. The primary hypothesis is that the detection rate of pathologic findings can be increased by undergoing enhanced diagnostic MRI work-up combined with prolonged Holter-monitoring. The primary outcome and all secondary outcomes are listed in Table 2, and a flowchart of study specific data acquisition is provided in Table 3.

The sample size calculation is based on the aforementioned primary hypothesis (Table 2), that an enhanced diagnostic work-up consisting of cerebral and cardiac MRI (including contrast-enhanced angiography of the aortic arch and the carotid arteries) combined with a prolonged ECG-Monitoring (of up to 10 days duration) leads to a reduction in the number of strokes classified as cryptogenic from 30 to 25 % when compared to the routine diagnostic work-up at the Stroke Unit. A sample size of 396 patients which receive both the enhanced as well as the routine diagnostic work-up would provide a power of 79 % to detect a difference in the number of strokes classified as cryptogenic (from 30 % after routine diagnostic work-up to 25 % after enhanced diagnostic work-up, and with assumed 13 % discordant pairs: 9 % diagnosed as cryptogenic stroke by routine work-up but not according to enhanced diagnostic work-up, and 4 % diagnosed as cryptogenic stroke by enhanced work-up but not according to routine diagnostic work-up). The hypothesis will be tested using the McNemar-Test (two-sided, α = 0.05). An interim analysis (with a significance level of α1 = 0.004) will be performed when 264 patients (67 %) are enrolled in the study. If p is ≤ α1 at the interim analysis, the study will be terminated. When the study will be continued, the primary endpoint will be tested with α2 = 0.046 (two- tailed); the overall significance level will be 5 %.  Taking into account a drop-out rate of about 20 %, 475 patients will have to be enrolled in the study.

The first patient was enrolled in May 2014. We anticipate a recruitment period of 18 months, and a recruitment period of additional 12 months is intended to enroll 475 patients With regard to the planned follow-up duration, first patient in to last-patient out will be 4.5 years in total.