Novel Emboli Protection System DuringCardiac Surgery: A Multi-Center, Randomized,Clinical TrialGil Bolotin, MD, PhD, Christoph Hans Huber, MD, Liran Shani, MD,Friedrich Wilhelm Mohr, MD, PhD, Thierry P. Carrel, MD,Michael Andrew Borger, MD, PhD, Volkmar Falk, MD, David Taggart, MD, PhD,Rony-Reuven Nir, PhD, Lars Englberger, MD, Joerg Seeburger, MD, PhD,Etem Caliskan, MD, and Christoph Thomas Starck, MDDepartment of Cardiac Surgery, Rambam Health Care Campus, and Faculty of Medicine, Technion-Israel Institute of Technology,Haifa, Israel; University Clinic for Cardiovascular Surgery Inselspital, Bern, Switzerland; Heart Center Leipzig, University of Leipzig,Leipzig, Germany; University Hospital Zurich, Zurich, Switzerland; and University of Oxford, Oxford, United Kingdom

Background. Stroke is a major cause of morbidity andmortality during open-heart surgery. Up to 60% of intra-operative cerebral events are emboli induced. This ran-domized, controlled, multicenter trial is the first humanstudy evaluating the safety and efficacy of a novel aorticcannula producing simultaneous forward flow andbackward suction for extracting solid and gaseous embolifrom the ascending aorta and aortic arch upon theirintraoperative release.

Methods. Sixty-six patients (25 females; 68 ± 10 years)undergoing elective aortic valve replacement surgery,with or without coronary artery bypass graft surgery,were randomized to the use of the CardioGard (Car-dioGard Medical, Or-Yehuda, Israel) Emboli Protectioncannula (“treatment”) or a standard (“control”) aorticcannula. The primary endpoint was the volume ofnew brain lesions measured by diffusion-weightedmagnetic resonance imaging (DW-MRI), performedpreoperatively and postoperatively. Device safety wasinvestigated by comparisons of complications rate,

Accepted for publication June 17, 2014.

Presented at the Fiftieth Annual Meeting of The Society of ThoracicSurgeons, Orlando, FL, Jan 25–29, 2014.

Address correspondence to Dr Bolotin, Department of Cardiac Surgery,Rambam Health Care Campus, P.O.B. 9602, Haifa 31096, Israel; e-mail:g_bolotin@rambam.health.gov.il.

� 2014 by The Society of Thoracic SurgeonsPublished by Elsevier

namely neurologic events, stroke, renal insufficiencyand death.Results. Of 66 patients (34 in the treatment group), 51

completed the presurgery and postsurgery MRI (27 in thetreatment group). The volume of new brain lesion for thetreatment group was (mean ± standard error of the mean)44.00 ± 64.00 versus 126.56 ± 28.74 mm3 in the controlgroup (p [ 0.004). Of the treatment group, 41% demon-strated new postoperative lesions versus 66% in thecontrol group (p [ 0.03). The complication rate wascomparable in both groups.Conclusions. The CardioGard cannula is safe and effi-

cient in use during open-heart surgery. Efficacy wasdemonstrated by the removal of a substantial amount ofemboli, a significant reduction in the volume of newbrain lesions, and the percentage of patients experiencingnew brain lesions.

(Ann Thorac Surg 2014;-:-–-)� 2014 by The Society of Thoracic Surgeons

ardiac surgery may cause a wide spectrum of

Cneurologic injuries, including ischemic stroke, en-cephalopathy, neurocognitive dysfunction and increasedmortality, hospital costs, and impaired quality of life [1–3].Embolism is considered the main mechanism of post-surgery neurologic injury [4].

Thirty percent to 50% of perioperative strokes detectedwith brain imaging are due to cerebral macroembolism,which can occlude flow in arteries larger than 200 mm indiameter [5–7]. Encephalopathy and neurocognitivedysfunction are believed to result primarily from cerebral

microembolism [8, 9]. In such cases, flow is blocked inarteries of less than 200 mm in diameter, and the emboliare either gaseous or solid particles in composition [2]. Anincreased flow of macroembolism and microembolisminto the cerebral circulation has been detected duringvarious aortic manipulations performed during cardiacsurgery [1, 10–12].Several approaches have been employed to diminish

the number and potential pathologic consequences ofemboli released during on-pump open heart surgery,including extraaortic ultrasonic and intraaortic mechani-cal embolic diverting devices, and an intraaortic filtra-tion device [13–16]. However, these devices have had

Drs Bolotin and Taggart disclose financial relationshipswith CardioGard.

0003-4975/$36.00http://dx.doi.org/10.1016/j.athoracsur.2014.06.061

Fig 1. The CardioGard aortic cannula is a curved tip 24-Frenchaortic perfusion cannula comprised of 2 hollow tubes, 1 for forwardflow and the other for suction of embolic material.

Abbreviations and Acronyms

AVR = aortic valve replacementCABG = coronary artery bypass graftingCVA = cerebrovascular accidentDW-MRI = diffusion-weighted magnetic

resonance imagingMMSE = Mini-Mental State ExaminationNIHSS = National Institutes of Health Stroke

ScaleRIFLE = risk, injury, failure, loss of kidney

function, andend-stagekidneydiseaseTCD = transcranial Doppler

2 BOLOTIN ET AL Ann Thorac SurgNOVEL EMBOLI PROTECTION SYSTEM 2014;-:-–-

unwarranted side effects and, crucially, none have hadtheir potential efficacy objectively evaluated bycomparing preoperative and postoperative magneticresonance imaging (MRI) scans.

The current randomized, controlled, multicenter trial isthe first human study evaluating the safety and efficacyof a novel aortic cannula that produces simultaneousforward flow and backward suction to extract solid andgaseous emboli from the distal ascending aorta upontheir intraoperative release. The efficacy of the novelcannula was assessed by the volume of new brain lesionsmeasured by diffusion-weighted (DW) MRI, performedpreoperatively and 5 to 7 days postoperatively.

Patients and Methods

The Institutional Review Board of all participating centersapproved the study protocol in accordance with theHelsinki Declaration. All enrolled patients signed awritten informed consent to participate in the studybefore randomization (detailed hereunder).

Aortic Cannula DescriptionThe CardioGard aortic cannula (CardioGard Medical, Or-Yehuda, Israel) is a curved tip 24-French aortic perfusioncannula comprised of 2 hollow tubes (Fig 1). The first tubeis a standard main forward-flow tube administeringarterial blood into the aorta from the cardiopulmonarybypass (CPB) machine. The second tube, attached to anexisting bypass vent port, is a novel element locatedposteriorly to the main tube; its function is to facilitateblood and particle suction, by directing the blood back tothe reservoir of the CPB machine, while the retrievedembolic material is eliminated through the filter of thevenous reservoir. In the current study, a 40-mm filter wasattached to the suction tube lining in order to evaluate thesize and weight of the emboli material captured duringthe procedure.

Study DesignThis multicenter, prospective randomized controlled trialwas the first-in-man experience with the new aorticcannula. Enrolled patients were randomized (1:1) toeither CardioGard (“treatment”) or standard (“control”)

aortic cannula, in 4 medical centers; 1 in Germany, 1 inIsrael, and 2 in Switzerland.The inclusion criteria were patients 50 years of age or

older undergoing elective aortic valve replacement (AVR)or combined AVR and coronary artery bypass grafting(CABG), using CPB. Patients with fixed neurologicimpairment due to previous insult were excluded.After a preoperative assessment, which included

baseline laboratory tests, neurologic evaluation, and DW-MRI, patients were randomized either to the treatmentgroup, in which the CardioGard cannula was used, or thecontrol group, in which a standard cannula was used.Aortic cannulation and decannulation procedures in bothgroups were performed using standard surgical tech-niques. Throughout the procedure blood flow wasmaintained at appropriate values. In the treatment groupa constant suction rate of 1 L/minute was applied. Post-operatively, the external filter was rinsed and thecaptured emboli were measured for size and total weight.

EndpointsThe primary efficacy endpoint of the study was the totalvolume of new brain lesions measured by DW-MRI,performed preoperatively and 5 and 7 days post-operatively. The DW-MRI analysis was performed by anexternal core lab which was blinded to the patients’ groupallocation (BioImage; MRI Research & Consulting, Haifa,Israel). Additional DW-MRI related measures were thenumber of patients with new brain lesions, the number ofnew brain lesions, and the mean volume of new brain

3Ann Thorac Surg BOLOTIN ET AL2014;-:-–- NOVEL EMBOLI PROTECTION SYSTEM

lesions. Secondary endpoints were defined as the diam-eter and total weight of captured solid emboli using thetreatment cannula. Device safety was assessed bycomparing the rates of complications between the treat-ment and control groups, namely death, neurologicevents, cerebrovascular accident, kidney injury, andhemolysis.

Transcranial Doppler (TCD) was used in 3 of the studycenters when suitable windows were obtained duringoperation. The TCD counts of detected emboli werecompared between the 2 groups during aortic manipu-lation, namely “cross-clamp on” and “cross-clamp off”conditions, as well as throughout the entire CPB period.

Neurologic evaluation was performed using the Na-tional Institutes of Health Stroke Scale by trained exam-iners who were blinded to the randomization protocol[17]. Evaluation was preoperative, on postoperative day5 to 7, and on postoperative day 30. Stroke, definedas a central neurologic deficit persisting for longer than 24hours, transient neurologic events (ie, transient ischemicattack, delirium, or disorientation, and nonmetaboliccoma) were considered neurologic events. Cognitivedysfunction was assessed by the Mini-Mental State Ex-amination (MMSE) preoperatively and postoperatively.Acute kidney injury was defined according to the RIFLE(risk, injury, failure, loss of kidney function, andend-stage kidney disease) criteria [18].

Data Collection and ManagementMonitoring procedures were performed in accordancewith applicable federal regulations and InternationalConference on Harmonisation of Technical Requirementsfor Registration of Pharmaceuticals for Human Use (ICH)good clinical practice guidelines. On-site research co-ordinators collected the clinical data. Independent studymonitors (Genae Associates NV, Belgium and HaInbal,CRO, Israel) verified all the case report forms. The in-vestigators had full access to the data at their respectivesites.

Randomization ProcessThe randomization plan was established by the studystatistician for all 4 centers. The randomization of eachpatient into the treatment or control group took placeduring the baseline evaluation period; ie, before thesurgery. All 4 sites were provided with sequentiallynumbered randomization envelopes at the beginning ofthe study; these envelopes were stored in the studybinder for the duration of the study. At the time ofrandomization, the study coordinator retrieved andopened the next sequentially numbered randomizationenvelope from the study binder, which determinedwhether the patient belonged to the treatment or controlgroup.

Statistical AnalysisAll analyses were conducted using Statistical Packagefor the Social Sciences, Version 21 (SPSS Inc, Chicago, IL).As all analyses are nonparametric, data are presentedas median (5th percentile to 95th percentile) values.

Comparisons of continuous variables between the treat-ment and control groups were conducted using Mann-Whitney U tests. The comparison of the main outcomemeasures of total lesion volume and mean lesion volumewere 1-tailed in accordance with the hypotheses that thetreatment group would demonstrate smaller valuesthereof; similarly, comparisons of detected TCD signalsbetween the treatment and study groups were 1-tailed.All other analyses of continuous variables were 2-tailed.Comparisons of categoric variables between the treat-ment and control groups were conducted using 2-tailedz-test comparisons of proportions with Bonferroni cor-rections for multiple comparisons. Statistical significancewas set at a p value less than 0.05.Sample size calculation was based on detecting a dif-

ference in the total volume of new brain lesions betweenthe treatment group and the control group. Using thenonparametric Mann-Whitney test for sample sizecalculation (a ¼ 5%), a sample of 26 in each group wasdemonstrated to have a 95% power (b ¼ 5%) to detect areduction of 50% in new brain lesions’ total volume fromthe control group, using a standard cannula (expectedtotal volume: 3.8 � 2 cm3 [19]), to the treatment group,using the treatment cannula (1.9 � 2 cm3). A total samplesize of 52 subjects was considered adequate for the pri-mary performance endpoint. After adjustment for adropout rate of 25% (considering administrative and pa-tients’ issues), the required sample size was 66 patients.

Results

From June 2012 to September 2013, 66 patients (25 fe-males; 68 � 10 years) were included in the study; 34 in thetreatment arm and 32 in the control arm. Fifty-one pa-tients completed the presurgery and postsurgery MRI;27 in the treatment arm and 24 in the control arm. Asdepicted in Table 1, no significant differences wereobserved between the treatment and control groups interms of demographics and preoperative characteristics,including age, body mass index, and the incidence ofhyperlipidemia, hypertension, etc. Likewise, there was nodifference in the logistic European system for cardiacoperative risk evaluation I or baseline MMSE scores. Nodifferences in hemolysis-associated parameters (lacticacid dehydrogenase, bilirubin, free hemoglobin) wereobserved between the treatment and control groups,either at the beginning or end of the surgical procedure,as were additional surgery-related features (Table 2).The mean � standard error of the mean of the new

brain lesions’ volume for the treatment group was 44.00 �64.00 mm3 versus 126.56 � 28.74 mm3 in the control group(p ¼ 0.004; 1-tailed analysis). However, because valueswere not normally distributed, nonparametric statisticalmethods were used for analyses and data are presentedas medians and 5th to 95th percentiles (Table 3). The totalvolume of new brain lesions was significantly smaller inthe treatment group compared with the control group (0.0[0.0 to 1,582.6] mm3 vs 237.3 [0.0 to 3,337.7] mm3, respec-tively, p ¼ 0.028; 1-tailed analysis) (Fig 2). The total vol-ume of new brain lesions normalized by their number

Table 1. Patient Demographics and PreoperativeCharacteristics

VariableTreatment(n ¼ 34)

Control(n ¼ 32) p Value

Sex (female:male) 0.33:0.67 0.44:0.56 0.340Age (years) 67.5 (47.3–81.0) 72 (50.7–85.4) 0.355BMI (kg/m2) 29.5 (23.0–37.2) 31.1 (19.3–39.3) 0.426Hyperlipidemia, n (%) 24 (70.6) 23 (71.9) 0.908Hypertension, n (%) 27 (79.4) 30 (93.8) 0.090Creatinine (mg/dL) 0.83 (0.56–1.49) 0.89 (0.32–1.43) 0.763A1C (mmol/L) 6.2 (4.6–8.1) 6.05 (5.8–7.9) 0.892Logistic EuroSCORE I 4.9 (1.3–16.4) 3.71 (1.2–13.2) 0.379MMSE 31.0 (23.2–34.0) 32.0 (25.5–34.0) 0.247Previous TIA/CVA,

n (%)3 (8.8) 1 (3.1) 0.591

NIH stroke score 0.0 (0.0–1.4) 0.0 (0.0–1) 0.422Modified Rankin Scale 1 (0.0–2.4) 1 (0.0–2.45) 0.907

Data are presented either as proportion or median (5th percentile to 95thpercentile).

BMI ¼ body mass index; CVA ¼ cerebrovascular accident;EuroSCORE ¼ European system for cardiac operative risk evaluation;MMSE ¼ Mini-Mental State Examination; NIH ¼ National Institutesof Health; TIA ¼ transient ischemic event.

4 BOLOTIN ET AL Ann Thorac SurgNOVEL EMBOLI PROTECTION SYSTEM 2014;-:-–-

was also smaller in the treatment group compared withthe control group (0.0 [0.0 to 191.8] mm3 vs 100.2 [0.0 to458.0] mm3 respectively, p ¼ 0.016; 1-tailed analysis).Moreover, the rate of patients demonstrating new brainlesions on the postoperative MRI was lower in the treat-ment group (41% vs 66%; p ¼ 0.03). The median totalvolume and largest diameter of the captured emboli inthe treatment group were 0.2 mm3 (0.01 to 2.40) and 1.5mm (0.67 to 2.74), respectively (Fig 2). The total filteredemboli weight was 0.035 g (0.0014 to 0.585) (Fig 3). Nodifferences were observed in the number of postoperativenew brain lesions or MMSE scores. The TCD analysesindicated no significant differences between the treat-ment and control groups (Table 4).

Table 2. Perioperative Characteristics

Variable Treatment

Isolated AVR procedure, n (%) 22 (64.7AVR þ CABG procedure, n (%) 12 (35.3Free hemoglobinpreoperative (g/dL) 0.05 (7�1Free hemoglobinpostoperative (g/dL) 0.051 (0.01LDHpreoperative (U/L) 223.9 (125LDHpostoperative (U/L) 391.5 (191Bilirubinpreoperative (mg/dL) 0.33 (0.14Bilirubinpostoperative (mg/dL) 0.33 (0.16Atheromatous grade of the ascending aorta 2 (1.0–CPB time (minutes) 108.5 (45.5Cross-clamp time (minutes) 75.5 (20.5Surgery duration (minutes) 229.0 (123

Data are presented either as proportion or median (5th percentile to 95th perc

AVR ¼ aortic valve replacement; CABG ¼ coronary artery bypass graftidehydrogenase.

No death attributable to brain injury was reported ineither group. The treatment and control groups weresimilar with regard to the occurrence of postoperativemajor adverse events, namely cerebral vascular accident,renal failure, peripheral ischemia, or mesentericischemia, as well as minor events; ie, delirium, depres-sion, adjustment disorder, anxiety, or polyneuropathy(Table 5).

Comment

Perioperative stroke remains a major complication asso-ciated with cardiac surgery. The novel CardioGard can-nula has already proved to be highly effective in reducingembolic load in both in vitro and animal studies [20, 21].The current study was designed to investigate the safetyand potential efficacy of this cannula during cardiacsurgery.The main findings of the current study, in terms of ef-

ficacy, are that the use of the treatment cannula wasassociated with a decreased volume of new brain lesionsand a smaller number of patients demonstrating newbrain lesions, as compared with a standard cannula. Interms of safety, the treatment and standard cannulaeyielded comparable hemocompatibility profiles.Current findings corroborate recent in vitro [20] and

animal [21] studies of the treatment cannula, whichdemonstrated retrieval rates of 50% to 83% and 70% to79% of the emboli released into the aorta, respectively.These rates were dependent upon the suction level, suchthat the greater the suction rate the greater the emboliretrieval rate. An additional designated flow regimensimulated the routine lowering of flow rate during aorticmanipulations in a CPB procedure and yielded a retrievalrate of up to 93.0%. As aortic manipulation has beenassociated with the substantial release of embolic mate-rial [1, 10–12], these reports attest to the efficacy of thetreatment cannula to effectively remove solid embolicmaterial from the systemic circulation throughout cardiac

(n ¼ 34) Control (n ¼ 32) p Value

) 27 (83.9) 0.096) 5 (16.1)0-4–0.2) 0.05 (8�10-4–0.2) 0.6565–0.3) 0.07 (0.01–0.3) 0.363.9–485.5) 203.0 (88.8–463.5) 0.681.6–613.0) 346.0 (232.1–723.0) 0.082–1.12) 0.29 (0.12–0.67) 0.066–1.0) 0.29 (0.12–0.52) 0.0924.0) 1 (1.0-3.35) 0.181–202.5) 103.5 (45.7–218.1) 0.868–167.5) 74.5 (14.0–164.8) 0.898.0–379.8) 192.5 (92.7–407.9) 0.311

entile).

ng surgery; CPB ¼ cardiopulmonary bypass; LDH ¼ lactic acid

Table 3. Magnetic Resonance Imaging Outcome Measures

Variable Treatment (n ¼ 27) Control (n ¼ 24) p Value

Total volume of new brain lesionsa (mm3) 0.0 (0.0–1582.6) 237.3 (0.0–3337.7) 0.028b

Patients with new lesions, n (%) 14 (41) 21 (66) 0.030b

Number of new brain lesions (n) 0.0 (0.0–11.5) 0.5 (0.0–11.35) 0.237b

Total volume of new brain lesions/number ofnew brain lesions (mm3/n)

0.0 (0.0–191.8) 100.2 (0.0–458.0) 0.016b

a Primary outcome measure. b 1-tailed nonparametric analysis.

Data are presented either as proportion or median (5th percentile to 95th percentile).

5Ann Thorac Surg BOLOTIN ET AL2014;-:-–- NOVEL EMBOLI PROTECTION SYSTEM

surgery and, with highest efficacy, during the steps ofaortic manipulation. This is in contrast with the embolicprotection devices detailed hereunder in accordance withthis writing.

Various devices, used after aortic cannulation, weredeveloped to diminish the quantity of emboli releasedduring on-pump open heart surgery. One is ultrasoundbased and is placed directly on the aorta during sur-gery [13]. Its disadvantages include adverse heating oftissues and a noncontinuous mode of action. Anothergroup of diverting devices include percutaneousaccess and deployment in the aorta [14], which require

Fig 2. Individual values of total lesion volume among the (A) treatment andtreatment group.

catheterization, insertion and placement, and may beassociated with emboli detachment as well as vascularcomplications. These approaches may also be associatedwith increased ischemic events within the renal, mesen-terial, and peripheral circulation.Other approaches include an intraaortic filtration de-

vice that was reported to be inserted through the aorticcannula to capture particulate emboli during CPB [15, 16].Compared with the treatment cannula described in thecurrent study, the use of this device is more invasive andmay cause plaque disruption during deployment.Furthermore, the particle filtering cutoff value of the

control groups and largest diameter of filtered emboli among the (B)

Fig 3. Example of embolic material retrieved during the use of theCardioGard cannula.

Table 5. Postoperative Characteristics

VariableTreatment(n ¼ 34)

Control(n ¼ 32) p Value

Major adverse events:Mortality, n (%) 2 (5.9) 2 (6.3) 1.0CVA/TIA, n (%) 1 (2.9) 0 (0.0) 1.0Renal failure, n (%) 2 (5.9) 4 (12.5) 0.420Creatinine atdischarge (mg/dL)

0.86 (0.62–1.5) 0.9 (0.69–1.68) 0.282

Peripheral/mesentericischemia, n (%)

0 (0.0) 0 (0.0) 1.0

Minor adverse eventsMinor events, n (%) 5 (14.7) 10 (31.3) 0.145MMSE at baseline 31.0 (23.2–34.0) 32.0 (25.5–34.0) 0.247MMSE at discharge 32.0 (22.2–34.0) 32.0 (24.3–34.0) 0.585MMSE after 30 days 32.0 (25.3–34.0) 32.0 (26.8–34.0) 0.853Postoperative characteristics:NIH Stroke Scoreat discharge

0.0 (0.0–2.0) 0.0 (0.0-4.0) 0.112

NIH Stroke Scoreafter 30 days

0.0 (0.0–5.2) 0.0 (0.0–1.85) 0.305

MRS at discharge 1.0 (0.0–3.6) 1.0 (0.0–3.0) 0.328MRS after 30 days 0.0 (0.0–4.3) 1.0 (0.0–2.85) 0.128ICU stay (days) 2.0 (1.0–10.8) 2.0 (1.0–10.4) 0.487Hospitalization (days) 10.0 (6.0–35.3) 12.5 (4.7–38.7) 0.093

Data are presented either as proportion or median (5th percentile to 95thpercentile).

CVA ¼ cerebrovascular accident; ICU ¼ intensive care unit;MMSE ¼ Mini-Mental State Examination; MRS ¼ modified rankingscale; NIH ¼ National Institutes of Health; TIA ¼ transientischemic event.

6 BOLOTIN ET AL Ann Thorac SurgNOVEL EMBOLI PROTECTION SYSTEM 2014;-:-–-

intraaortic filter device is 120 mm in diameter, while thatof the treatment cannula is only limited by the aortic filterof the CPB machine, which is approximately 40 mm.Importantly, compared with all other existing devices forintraoperative reduction of emboli, the current studyobjectively confirmed the efficacy of the treatment can-nula using comparisons of preoperative and post-operative DW-MRI scans.

The MMSE screening test demonstrated comparableand normal scores in both groups at baseline, discharge,or 30 days postoperatively. This observation may stem

Table 4. Summary of Transcranial Doppler Data Gathered inThree Study Centers

Center(n) Condition

TreatmentGroup:DetectedSignals(Median)

ControlGroup:DetectedSignals(Median) p Value

1 (18) Right clamp on 1 0 0.414Left clamp on 0 0 0.488Right clamp off 7.5 31 0.444Left clamp off 21 27 0.421Total 315 278.5 0.401

2 (18) Right clamp on 3 3.5 0.187Left clamp on 5 6 0.206Right clamp off 3 4.5 0.136Left clamp off 3 6 0.359Total 83 169 0.184

3 (11) Right clamp on 7 5.5 0.810Left clamp on 15.5 10 0.522Right clamp off 6 10 1.000Left clamp off 15.5 11.5 0.646Total 134 53.5 0.646

from the MMSE test being only moderately sensitive toacute post-stroke cognitive impairment, namely 66% [22].However, its specificity to such impairment was found tobe 97%. This statistic holds more pertinency to the currentstudy in light of the reported normal MMSE scores. Theexhibited comparability and normality of the MMSEscores in the treatment and control groups seem to be inkeeping with previous prominent studies by Selnesand colleagues [23, 24] that concluded the following:(1) early post-CABG cognitive decline is generally mildand reversible within 3 months; and (2) late post-CABGcognitive changes, investigated for 6 years, are not spe-cific to the operation, or more specifically, to the use ofCPB. In this context it must be stressed that the primaryendpoint of the current study was the volume of newbrain lesions measured by DW-MRI, rather than the po-tential yet still unestablished consequences thereof,including cognitive decline or verifiable chronic braininjury; these variables are immensely intricate to quantifydue to their vast variation and still equivocally under-stood underlying pathophysiology [23, 24].Several limitations of the current study must be

considered. First, a relatively small number of patientscompleted the second MRI. Second, the TCD measure-ments varied in terms of equipment and derived vari-ables, such as emboli detection threshold and resolution.

7Ann Thorac Surg BOLOTIN ET AL2014;-:-–- NOVEL EMBOLI PROTECTION SYSTEM

Third, the proportion of treatment patients who under-went combined CABG and valve operations was higherthan in the control group; however, as combined CABGand valve operations are prone to produce more embolithan isolated AVR operations, the patients of the treat-ment group were potentially at greater risk of developingnew brain lesions. Therefore, the current findings mayactually have underestimated the true efficacy of thecannula.

Current findings indicate the superiority of thetreatment cannula in terms of efficacy and support itsintroduction in the daily practice of cardiac surgery as ithas the potential to decrease one of the most severepostoperative complications, namely postoperativestroke.

CardioGard sponsored the current study.

References

1. Hogue CW Jr, Palin CA, Arrowsmith JE. Cardiopulmonarybypass management and neurologic outcomes: an evidence-based appraisal of current practices. Anesth Analg 2006;103:21–37.

2. Alexander KP, Anstrom KJ, Muhlbaier LH, et al. Outcomesof cardiac surgery in patients > or ¼ 80 years: results fromthe National Cardiovascular Network. J Am Coll Cardiol2000;35:731–8.

3. Hogue CW, Gottesman RF, Stearns J. Mechanisms of cere-bral injury from cardiac surgery. Crit Care Clin 2008;24:83–98.

4. Blauth CI. Macroemboli and microemboli during cardio-pulmonary bypass. Ann Thorac Surg 1995;59:1300–3.

5. Djaiani G, Fedorko L, Borger M, et al. Mild to moderateatheromatous disease of the thoracic aorta and new ischemicbrain lesions after conventional coronary artery bypass graftsurgery. Stroke 2004;35:e356–8.

6. Likosky DS, Marrin CA, Caplan LR, et al. Determination ofetiologic mechanism of strokes secondary to coronary arterybypass graft surgery. Stroke 2003;34:2830–4.

7. Arnold JV, Blauth CI, Smith PL, Jagoe JR, Wootton R,Taylor KM. Demonstration of cerebral microemboli occur-ring during coronary artery bypass graft surgery usingfluorescein angiography. J Audiov Media Med 1990;13:87–90.

8. Blauth CI, Cosgrove DM, Webb BW, et al. Atheroembolismfrom the ascending aorta: an emerging problem in cardiacsurgery. J Thorac Cardiovasc Surg 1992;103:1104–11.

9. Barbut D, Hinton RB, Szatrowski TP, et al. Cerebral embolidetected during bypass surgery are associated with clampremoval. Stroke 1994;25:2398–402.

10. Borger MA, Taylor RL, Weisel RD, et al. Decreased cerebralemboli during distal aortic arch cannulation: a randomizedclinical trial. J Thorac Cardiovasc Surg 1999;118:740–5.

11. Stump DA, Rogers AT, Hammon JW, Newman SP. Cerebralemboli and cognitive outcome after cardiac surgery.J Cardiothorac Vasc Anesth 1996;10:113–9.

12. Goto T, Maekawa K. Cerebral dysfunction after coronaryartery bypass surgery. J Anesth 2014;28:242–8.

13. Sauren LD, la Meir M, Bolotin G, et al. The EmBlocker: ef-ficiency of a new ultrasonic embolic protection deviceadjunctive to heart valve surgery. Ann Thorac Surg 2009;88:253–7.

14. Carpenter JP, Carpenter JT, Tellez A, Webb JG, Yi GH,Granada JF. A percutaneous aortic device for cerebralembolic protection during cardiovascular intervention. J VascSurg 2011;54:174–81.

15. Banbury MK, Kouchoukos NT, Allen KB, et al. Emboli cap-ture using the Embol-X intraaortic filter in cardiac surgery: amulticentered randomized trial of 1,289 patients. Ann ThoracSurg 2003;76:508–15.

16. Christenson JT, Vala DL, Licker M, Sierra J, Kalangos A.Intra-aortic filtration: capturing particulate emboli duringaortic cross-clamping. Tex Heart Inst J 2005;32:515–21.

17. Meyer BC, Hemmen TM, Jackson CM, Lyden PD. ModifiedNational Institutes of Health Stroke Scale for use in strokeclinical trials: prospective reliability and validity. Stroke2002;33:1261–6.

18. Roy AK, Mc Gorrian C, Treacy C, et al. A comparison oftraditional and novel definitions (RIFLE, AKIN, and KDIGO)of acute kidney injury for the prediction of outcomes inacute decompensated heart failure. Cardiorenal Med 2013;3:26–37.

19. Stolz E, Gerriets T, Kluge A, Kl€ovekorn WP, Kaps M,Bachmann G. Diffusion-weighted magnetic resonance im-aging and neurobiochemical markers after aortic valvereplacement: implications for future neuroprotective trials?Stroke 2004;35:888–92.

20. Shani L, Cohen O, Beckerman Z, Nir RR, Bolotin G.A novel emboli protection cannula during cardiac surgery:in vitro results. J Thorac Cardiovasc Surg 2014 [Epub aheadof print].

21. Shani L, Cohen O, Beckerman Z, Nir RR, Bolotin G. Novelemboli protection cannula during cardiac surgery: first ani-mal study. Asian Cardiovasc Thorac Ann 2014;22:25–30.

22. Godefroy O, Fickl A, Roussel M, et al. Is the MontrealCognitive Assessment superior to the Mini-mental StateExamination to detect poststroke cognitive impairment?A study with neuropsychological evaluation. Stroke 2011;42:1712–6.

23. Selnes OA, Gottesman RF. Neuropsychological outcomesafter coronary artery bypass grafting. J Int Neuropsychol Soc2010;16:221–6.

24. Selnes OA, Gottesman RF, Grega MA, Baumgartner WA,Zeger SL, McKhann GM. Cognitive and neurologic outcomesafter coronary-artery bypass surgery. N Engl J Med 2012;366:250–7.

DISCUSSION

DR PAUL GRUNDEMAN (Utrecht, Netherlands): Dr Bolotin,I think we have to congratulate you, this is an excellent paper,very good results, and also proof that we have to improve oncardiopulmonary bypass conduct.

My question would be, if I see this, there must be a differencebetween a coronary case and a valve case. So how is the divisionin the groups between the type of operation that were subjectedto this technique?

DR JOHN CONTE (Baltimore, MD): We are going to go back andshow the slide that actually shows that there were more patients

who underwent coronaries in the investigative group than in thecontrol group.

DR BOLOTIN: All patients have AVR [aortic valve replacement].Some of the patients have AVR plus CABG [coronary arterybypass grafting]. All patients have AVR. In the CardioGardgroup there was a trend higher percentage of AVR plus coronary.This is in a way even those patients supposed to be sicker andperhaps will have higher percentage of embolic event, but in ourresults they have actually less. So I think it actually strengthensour results, but this is what we have.

8 BOLOTIN ET AL Ann Thorac SurgNOVEL EMBOLI PROTECTION SYSTEM 2014;-:-–-

DR GRUNDEMAN: And the second question is: What was thenature of the particulate matter that you found in the filter? Wasit originating from the heart-lung machine, or was it biologicalmaterial?

DR BOLOTIN: Oh, no, it is, as you can see, it is actually calciumprobably from the aorta, from the valve. It is particle calcium thatwe all think that we wash them very carefully, but apparently weare not doing the job good enough. But all of them are biologicalparticles. I guess most of them from the aortic valve anddescending aorta.

DR GRUNDEMAN: So at least there were no glass particles, nopolymers?

DR BOLOTIN:No.

DR GRUNDEMAN: So it is originating from the body.

DR BOLOTIN: Yes, biological.

DR LOUIS PERRAULT (Montreal, Quebec, Canada): Veryinteresting study. My question is related to the prior one. Couldyou say whether the patients that had CABG have a little bitmore emboli, if you separate them from the isolated AVRs andyou look at them? Were the proximals done under just one cross-clamp?

DR BOLOTIN: It depends on the center, some with side-bitingclamps, some without. The number is too small, you see, 12patients compared to 5 with the CABG, so I do not have the datato tell you. I assume that there is more disease in the patientswith CABG, but I do not have the data to tell you.

DR PERRAULT: Well thank you, this is very interesting deviceand study. As you may know, the NHI [National Institutes ofHealth]-sponsored CTSN network [Cardiothoracic Surgical Tri-als Network] is very interested in perhaps mounting a bigger trialwith this device.

DR CONTE: These are very impressive results. Were you able toquantify a difference in the amount of particulate material in thecoronary group versus the group that just had aortic valves?

DR BOLOTIN: No, we did not look at it. It is small numbers.

DR CONTE: And a technical question. When you set up yourcircuit, you had suction on one limb of the cannula. Was thebypass circuit set up so that you used suction on the venousdrainage cannula as well?

DR BOLOTIN: No, the venous cannula was without suction.

DR CONTE: Just used gravity?

DR BOLOTIN: Yes.