Rationale and Design for the Remote Ischemic Preconditioning for Carotid Endarterectomy (RIP-CEA) Trial

Abstract

Background:

While the perioperative stroke rate after carotid endarterectomy (CEA) is low, magnetic resonance imaging (MRI) “silent” microinfarctions are common and have been correlated with postoperative neurocognitive decline. Our study will investigate the role of remote ischemic preconditioning (RIPC) as a potential neuroprotective mechanism. RIPC is a well-tolerated stimulus that, through neuronal and humoral pathways, generates a systemic environment of greater resistance to subsequent ischemic insults.

Hypothesis:

Patients undergoing RIPC before CEA will have improved postoperative neurocognitive scores compared to patients undergoing standard care.

Methods:

Patients undergoing CEA will be randomized 1:1 to RIPC or standard clinical care. Those randomized to RIPC will undergo a standard protocol of 4 cycles of RIPC. Each RIPC cycle will involve 5 minutes of forearm ischemia with 5 minutes of reperfusion. Forearm ischemia will be induced by a blood pressure cuff inflated to 200mmHg or at least 15mmHg higher than the systolic pressure if systolic > 185mmHg. This will occur after anesthesia induction and during incision/dissection but before manipulation or clamping of the carotid; thus, patients will be blinded to their assignment. Before carotid endarterectomy, all patients will undergo baseline neurocognitive testing in the form of a Montreal-Cognitive Assessment (MoCA) and National Institutes of Health (NIH) Toolbox. MoCA testing only will be conducted on postoperative day 1 in the hospital. The full neurocognitive testing battery will again be conducted at 1-month follow up in the office. Changes from baseline will be compared between arms at the follow-up time points. Assuming no drop-ins or drop-outs and a 10% loss to follow-up, we would need a sample size of 43 patients for 80% power per treatment arm. The primary endpoint, change in MoCA scores, will be analyzed using a random effects model and secondary outcomes will be analyzed using either linear or logistic regression where appropriate.

Conclusions:

RIPC, if shown to be effective in protecting patients from neurocognitive decline following CEA, represents a safe, inexpensive, and easily implementable method of neuroprotection.

Background:

Stroke is the 5th leading cause of death in the United States.¹ Carotid artery disease (CAD) is responsible for approximately 16.6% of ischemic strokes.² In CAD, optimally treating each patient is paramount to minimizing the future risk of stroke. Carotid endarterectomy (CEA) is a proven modality for stroke-risk reduction in patients with CAD³. The perceived major risks with CEA, which include stroke and MI, are relatively rare (1-3% for stroke and 4% for MI); however, up to 25% of patients experience “silent” microinfarctions demonstrated on postoperative MRI.^4-6 Intra-operative ischemia (microembolic events or hypoperfusion) is thought to be the cause of silent micro infarctions.⁵ “Silent” may be a misnomer as these infarctions have been previously tied to a decline in neurocognitive function and dementia.^6-9 Neurocognitive decline can lead to reduced functioning and overall wellbeing in patients.¹⁰ Thus, there may be a role for neurocognitive protection in patients undergoing CEA to prevent silent microinfarctions and neurocognitive decline.^11,12

An ideal neurocognitive protective intervention would have an excellent clinical profile, be clinically feasible, and lead to a systemic resistance to ischemic insults. Amongst proposed modalities for neurocognitive protection, remote ischemic preconditioning (RIPC) meets these criteria. Ischemic preconditioning is the phenomenon whereby brief periods of ischemia followed by tissue reperfusion confer subsequent protection against ischemia-induced injury. The concept, proposed 30 years ago by Murry et al., demonstrated that brief cycles of ischemia and reperfusion of the coronary arteries protect the myocardium from subsequent prolonged ischemia and reperfusion, leading to a reduction in infarct size.¹³ Remote ischemic preconditioning is the application of this cyclical ischemia-reperfusion stimulus at a site distal from the vasculature of interest. Often, this is via blood pressure cuff insufflation on an extremity.¹⁴ RIPC has been thoroughly investigated at the phase I and II level in the coronary literature and has been shown to be safe.¹⁴ Due to its high metabolic activity, the brain is especially vulnerable to periods of ischemia during carotid cross-clamping.¹⁵ Ischemic tolerance has been demonstrated after direct ischemic conditioning in the brain.¹⁵ However, direct conditioning is difficult and potentially dangerous in carotid interventions making remote ischemic preconditioning an attractive alternative. In animal models, remote ischemic preconditioning (RIPC) has been shown to produce an equivalent response to direct neuronal conditioning at the cellular level.^16,17

The precise mechanisms underlying the phenomenon of RIPC have yet to be fully elucidated.¹⁴ However, it is likely that both neural and humoral mechanisms are at play. Multiple studies have shown decreased levels of inflammatory markers in brains of animal models undergoing RIPC and then MCA occlusion.¹⁷ Much of the original research in RIPC stems from cardiac literature. Meta-analysis of this literature shows consistent benefits in biochemical outcomes although firm clinical data is still lacking.^18-20

Studies of RIPC in CEA are limited. In one study, 55 patients were randomized to 10 minutes of ischemia on each leg before clamping the carotid.²¹ The primary outcome, significant postoperative deterioration in saccadic latency determined by quantitative oculometry (time taken to respond and fix on a visual stimulus that appears suddenly), lacked clinical relevance. In this small group of patients, deterioration in quantitative oculometry was found in 8/25 patient who underwent RIPC and 16/30 control patients (p=0.11).²¹ Zhao et al. performed a single-center prospective randomized controlled trial assessing whether RIPC is safe and effective in attenuating ischemic injury related to carotid artery stenting (CAS).²² This study had 63 subjects per group and showed an RR for imaging lesions in the intervention group of 0.44 (0.2-0.91 95% confidence interval).²² To date; there are no randomized controlled trials of RIPC for carotid endarterectomy looking at the clinically relevant outcome of neurocognition

Lal et al. recently published results from a prospective neurocognitive study of 144 patients, 82 with carotid stenosis.²³ They found that asymptomatic carotid stenosis is associated with mild to moderate cognitive impairment particularly in the domains of motor/processing speed and learning/memory. They assert that neurocognition has become an important clinical outcome that must be included in future trials testing the efficacy of treatment strategies for carotid stenosis.²³

With this goal in mind, our study will investigate the effect of neuroprotection with remote ischemic preconditioning on neurocognition – what we believe to be a clinically relevant surrogate marker of cerebral microinfarctions. If we are able to prevent decrease cognitive function using RIPC it may have wide applicability in other vascular and cardiac procedures.

Study objectives:

Our primary objective is to investigate a potential neuroprotective mechanism, remote ischemic preconditioning (RIPC), to decrease the impact of subclinical ischemic insults during carotid revascularization on cognitive function. Secondary objectives include testing of the effect of RIPC on post-operative levels of biomarkers of cerebral ischemia (neuron-specific enolase and S100-beta) and coronary ischemia (post-operative troponin levels).

Participants:

The study population will include patients undergoing carotid endarterectomy for accepted indications at any University of Pittsburgh Medical Center hospital.²⁴ Specifically, patients aged 55-85 undergoing elective carotid for symptomatic carotid disease with greater than 50% stenosis by duplex ultrasound or asymptomatic disease with greater than 60% stenosis by duplex ultrasound will be included. Patients who lack radial pulses bilaterally, have a known upper extremity deep venous thrombosis, arteriovenous fistula or graft in both arms, prior diagnosis of hypercoagulable state, pre-existing lymphedema or axillary node dissection, prior history of dementia, intellectual disability or mental illness including depression, anxiety or schizophrenia, or are undergoing simultaneous coronary artery bypass graft will be excluded from the study (Table I).

Table 1:

Inclusion and exclusion criteria for participation in Remote Ischemic Preconditioning for Carotid Enterectomy (RIPCEA) study

Inclusion Criteria	Exclusion criteria
Patients must meet the indication for Carotid endarterectomy: Patients must be symptomatic with >50% stenosis by duplex OR Have asymptomatic with >60% stenosis Patient’s age must be between 55 – 85	Lack of radial pulse on the arm Known DVT in upper extremities:   The patient’s other arm maybe be used if DVT free Arteriovenous fistula or graft in arm Prior diagnosis of a hypercoagulable state Pre-existing lymphedema or axillary node dissection Diagnosis of dementia, intellectual disability, or mental illness The patient will undergo simultaneous coronary artery bypass grafting

DVT, Deep venous thrombosis

Summary of Study Design and Interventions:

RIP-CEA is an ongoing, randomized, single-blinded, two-arm, controlled clinical trial. Subjects are patients undergoing carotid endarterectomy for accepted indications at any University of Pittsburgh Medical Center hospital and are followed from the time of enrollment until one-month post-operative follow-up. Surgeons from the University of Pittsburgh perform all surgeries. Trained research assistants within the Division of Vascular Surgery perform all study-related procedures.

Potential patients are identified by providers in the office or by review of the operating room schedules by the study team. If the patient appears to be an appropriate candidate for the trial, the trial is discussed with the patient, informed consent is obtained, and the patient is randomized.

Patients who meet the eligibility criteria and provide informed consent are randomized in 1:1 to receive either Remote Ischemia Preconditioning plus standard care or standard care alone (Figure 1). The randomization scheduled is generated by a fixed block algorithm stratified by symptom status in fixed blocks of four. Sequentially numbered, opaque, sealed envelopes for both symptomatic and asymptomatic patients are utilized to implement the allocation sequence. Given a small planned sample size, we stratified based on the prognostic factor of symptom status to ensure equal distribution of the intervention between the two groups. The complete randomization sequence was computer-generated by a biostatistician who is otherwise uninvolved with the study.

Fig.

Flow chart of study events. CEA, Carotid Endarterectomy. DVT, Deep venous thrombosis. RIPC, Remote ischemic preconditioning.

After consent, patients undergo baseline neurocognitive testing in the form of the Montreal Cognitive Assessment (MoCA) and completion of the selected batteries from the NIH Toolbox to evaluate general cognition as well as specific cognitive domains. Of note, use of the NIH Toolbox is consistent with recommendations by National Institute of Neurological Disorders and Stroke (NINDS) for standardized assessment of neurocognitive function in clinical trials.²⁵ This baseline testing will be completed prior to surgery. Baseline functional status is assessed through the Modified Rankin Scale.

Subjects then undergo carotid endarterectomy under general anesthesia with technical details of the procedure and postoperative care at the discretion of the treating surgeon.

Remote Ischemic Preconditioning Intervention

Patients, but not surgeons, are blinded to the intervention. Patients randomized to RIPC will undergo a standard protocol of 4 cycles of 5 minutes of forearm ischemia with 5 minutes of reperfusion requiring 35 minutes for an application. Forearm ischemia will be induced by a blood pressure cuff inflated to 200mmHg or at least 15mmHg higher than the systolic pressure if systolic > 185mmHg or until the radial pulse is obliterated. This will be done after anesthesia induction and during incision/dissection before manipulation or clamping of the carotid. The treatment will be discontinued if the patient becomes hemodynamically unstable or requires an emergent resuscitation procedure. The primary surgeon will also be able to discontinue procedure if it is determined that patient safety is compromised.

Standardized anesthesia protocol

All patients undergo a standardized anesthesia protocol developed by cardiac anesthesia at UPMC. Highlights are as follows: Arterial line and IV are placed in the same arm so that one arm is free for the intervention. Midazolam is used for anxiolysis. Induction is achieved with propofol, fentanyl, and lidocaine. Anesthesia is maintained with oxygen and sevoflurane, muscle relaxant, and intermittent fentanyl for analgesia.

Post-operative testing

On postoperative day one, patients undergo a brief neurocognitive assessment with the MoCA. Bloodwork is also collected for testing troponin I levels and neurobiomarkers (Neuron-specific enolase and S100-beta). Patients are generally discharged on postoperative day one unless they have a medical reason for a prolonged stay.

Patients undergo a final neurocognitive assessment identical to their preoperative assessment (MoCA and NIH Toolbox) at their one-month follow up appointment, which is conducted at their surgeon’s office. Functional status is again assessed using the Modified Rankin Scale.

Endpoints:

The primary endpoint of the study is neurocognitive score difference (pre-and post - CEA) between the RIPC and the standard care cohorts. Secondary endpoints include:

1. Post-operative levels of serum biomarkers of the brain (neuron-specific enolase and S100-beta) and coronary (troponin) ischemia;

2. Rate and severity of postoperative stroke/TIA

3. Coronary ischemia (Troponin I levels and adverse cardiac events). (Table II).

Table II.

Outcome measures

Baseline Characteristics	POD 1 outcomes	1 month outcomes
Cognitive:  -  MOCA  -  NIH Toolbox	Cognitive:  -  MOCA  -  NIH Toolbox Biomarkers:  -  S100 beta  - Neuron-specific enolase  - Troponin I Adverse Cardiac Events	Cognitive:  -  MOCA  -  NIH Toolbox Adverse Cardiac Events Stroke/TIA

MOCA, Montreal Cognitive Assessment:

Primary Endpoint Assessment

We hypothesized that patients undergoing RIPC will have increased improvement in their neurocognitive score post-operatively compared to patients undergoing usual care. To investigate, we will utilize the Montreal-Cognitive Assessment (MoCA) & National Institutes of Health (NIH) Toolbox. Patients, prior to carotid endarterectomy, will undergo baseline neurocognitive testing in the form of a MoCA & NIHToolbox. Both are validated methods of neurocognitive assessment. The MoCA can be administered rapidly and gives an overall assessment of neurocognitive functioning. The NIH Toolbox is a multidimensional set of measures assessing cognitive domains developed to meet the need for a standard set of measures to be used as a “common currency” across diverse study designs and settings. From the NIH toolbox, subjects will be assessed with the Flanker Inhibitory Control and Attention Test which assesses executive function and attention, the Dimensional Change Card Sort Test which also assesses executive function, and the Pattern Comparison Processing Speed Test which assesses processing speed.²⁵ Both methods of neurocognitive testing will be conducted preoperatively; MoCA testing only will be conducted on postoperative day one, and the full neurocognitive battery will be assessed again at one-month follow-up.

Secondary Outcome Assessment

Our secondary outcome measures include other markers of neuronal ischemia and cardiac ischemia. To explore if RIPC can decrease the levels of neuronal ischemia as evidenced by biomarkers compared to usual care, we will use S100-beta and neuron-specific enolase. Both S100-beta and neuron-specific enolase are established biomarkers of neuronal ischemia detectable shortly after an ischemic insult.^26-29 Levels of both can be obtained from serum. While not reliable enough to be diagnostic (and thus not used in routine clinical practice), they have been shown to correlate well will presence of neuronal injury.^27,29 Some data shows that they may correlate with severity of stroke and degree of cognitive dysfunction.

In addition, patients will be assessed for stroke or TIA occurring within the 30-day post-operative time frame. Severity of symptoms and disability will be graded with the NIH stroke scale and the modified Rankin Scale.

To explore if RIPC can have a systemic impact, we will also examine post-operative cardiac ischemia. We hypothesize that RIPC can decrease the number of perioperative (<30 days) adverse cardiac events following CEA as well as post-operative troponin levels compared to usual care. Postoperative myocardial infarction (MI), new congestive heart failure (CHF), and new arrhythmia rates will be recorded. Additionally, troponin I levels will be drawn on postoperative day one. If an event does not occur before discharge, patients will be assessed by history and electronic medical record at follow-up to ascertain if any adverse cardiac events have occurred within 30 days of surgery.

Data Management and Analytical Methods:

All data are collected by trained research assistants and residents/fellows. Baseline characteristics, clinical data, procedural data, and postprocedure outcomes will be entered into a secure online database for maintenance and analysis. Data analysis will be conducted with an intention to treat only. We expect very low drop-out and very low nonadherence.

Missing data will be managed following multiple imputation methods assuming data is missing at random. Data analysis will be performed using Stata Statistical Software: Release 14 (College Station, TX: StataCorp LP.).

Demographic variables that are continuous will be compared between groups using t-tests for parametric data analysis and Mann-Whitney U tests for nonparametric data analysis. Demographic variables that are categorical will be analyzed between groups using chi-squared tests. These comparisons of baseline demographics will be carried out at the time of outcome analysis to uncover differences between groups that may need to be adjusted for in the outcome analysis.

The primary endpoint will be analyzed using a random effects model with patient as a random effect in the model adjusting for covariates to compare neurocognitive scores at the multiple time points between groups. This will account for the correlated nature of multiple observations per subject. Subgroup analysis will be carried out on octogenarians and by symptomatic status.³⁰ This analysis will be conducted using the same adjusted random intercept model and employ the Bonferroni correction to account for the increated risk of false positives.

The secondary outcome of serum biomarker levels will be compared between groups using a simple linear regression model. Finally, the secondary outcome of rates of adverse neurologic and cardiac events will be compared between groups using a logistic regression model including adjusting for baseline comorbidities known to predict perioperative events such as symptomatic status, coronary disease, diabetes, and age.

Sample size estimation:

The changes in neurocognitive testing after carotid endarterectomy are ill-defined. In general, the lack of standardization of specific cognitive tests and timing of assessment of cognitive function after carotid endarterectomy does not allow for definite conclusions to be drawn regarding expected outcomes after the intervention to a baseline score [25]. One recently completed prospective study by Wang et al. followed patients before carotid endarterectomy and at six weeks postoperatively with MoCA scores and found that average preoperative MoCA score was 20.48 ± 1.79 vs. 22.04 ± 1.48 at six weeks postoperative [26]. A minimally clinically important difference has not been established, but we believe a two-point difference would have substantial clinical relevance as suggested in other studies [26]. We used the 6-week data from Wang et al. which should be comparable to our one-month period data (especially because patients usually follow up in the 4-6-week range in reality). Alpha was set at 0.05 and beta at 0.8. The standard deviation of 3.1 for MoCA scores was derived from mocatest.org. This is the standard deviation for patients with mild cognitive impairment (MoCA score < 26). Assuming no drop-ins or drop-outs and a 10% loss to follow-up (which is generous given the high 1-month follow-up rates), we would need a sample size of 43 patients per treatment arm.

Ethical considerations:

This study was approved by our Institutional Review Board before enrollment of the first patient. All potential risks and benefits (those related specifically RIPC) are discussed with potential study subjects, and participants provide appropriate informed consent before enrollment in the study. Patients are made aware that they can withdraw from the study at any time. All patient data collected as part of this study will remain secure and confidential. By our study protocol, no raw data, even deidentified, will be disseminated to groups outside of our institution. All blood samples are stored securely until they are tested for biomarkers, after which time they are safely disposed of according to institutional protocol. No samples are stored long term.

This study is being funded by two grants. First, this study was awarded a grant by the University of Pittsburgh Medical Center Heart and Vascular Institute. Also, this study was given a pilot award grant from the University of Pittsburgh’s Clinical and Translational Science Institute. Both sources of funding have no financial interests in the results of our trial. No industry funding has been obtained for the conduct of our research, and the study investigators have no relevant financial interests in the outcome of our study.

Future directions: We have chosen to focus on a clinically relevant primary endpoint of neurocognition for this initial study. In the future, we hope to demonstrate the relationship between neurocognitive outcomes and imaging findings of microinfarction on MRI.

Conclusions:

Despite the mounting evidence that carotid artery stenosis and carotid endarterectomy significantly affect neurocognitive function, this relationship remains poorly understand and neuro-protective mechanisms have not been well studied. The RIP-CEA trial has been actively enrolling since December 2016 and will be completed during the winter of 2019. The trial will examine the effectiveness of RIPC as a neuroprotective mechanism for carotid endarterectomy. If effective, RIPC represents a safe, inexpensive, and easily implementable method of neuroprotection.