Can a large intraoperative watershed air or blood emboli stroke be detected during cardiopulmonary bypass (CPB)?

Detection of Large Intraoperative Embolic Stroke During Cardiopulmonary Bypass

Large intraoperative watershed or embolic strokes during cardiopulmonary bypass (CPB) are extremely difficult to detect in real-time and typically go unnoticed until the patient emerges from anesthesia, despite available monitoring technologies having significant limitations in identifying these catastrophic events. 1

Why Detection Is So Challenging

Timing and Masking of Symptoms

• Intraprocedural stroke occurring during general anesthesia or deep sedation is typically only diagnosed when the patient emerges from anesthesia, as focal neurological deficits (weakness, aphasia, visual field disturbances) are completely masked during the procedure 1
• Approximately 40% of cardiac surgery strokes occur intraoperatively, with the remaining 60% occurring postoperatively, peaking at 40 hours after surgery 2
• The mean onset for hemispheric strokes after valve surgery is 1.3 days post-operatively, suggesting many develop after the patient leaves the operating room 2

Mechanisms of Intraoperative Stroke

• Thromboembolism accounts for approximately 70-80% of intraoperative strokes and is often associated with aortic manipulation, while hypoperfusion causes border zone or watershed strokes in only 20-30% of cases 1
• Air emboli, atherosclerotic debris, calcium, fat, platelet thrombi, and CPB tubing fragments can all cause cerebral embolization during bypass 1, 3
• Manipulation of the ascending aorta and aortic arch during cannulation and cross-clamping carries the highest risk of detaching atherosclerotic plaques and calcific deposits 1

Available Monitoring Technologies and Their Limitations

Transcranial Doppler (TCD)

• TCD can detect embolic signals in the carotid arteries during CPB, but has limited sensitivity for very small microemboli (<10 μm) and cannot reliably distinguish between gaseous and solid emboli 4, 5
• Studies show thousands of microemboli <40 μm are transmitted to patients during heart surgery, with increased embolic load at bypass initiation and clamp removal 5
• TCD demonstrates wide interindividual variability in middle cerebral artery flow velocity measurements (range 0.39-2.19), limiting its clinical utility 6

Near-Infrared Spectroscopy (NIRS)

• Bifrontal cerebral oximetry has been available for over 2 decades, but controversy remains about its clinical effectiveness in detecting acute ischemic events 1
• NIRS has low spatial resolution and is significantly influenced by systemic hemoglobin levels, potentially missing focal ischemic events 6
• Two RCTs suggest NIRS may predict early perioperative cognitive decline and stroke, but these studies focused on postoperative outcomes rather than real-time intraoperative detection 1

Electroencephalography (EEG)

• Processed EEG monitoring has limited ability to detect or quantify cerebral ischemia during cardiac surgery, partly due to excessive artifact in the CPB setting 1
• While various EEG variables are accepted markers of cerebral ischemia, current commercial devices (including BIS monitoring) cannot reliably identify acute embolic events intraoperatively 6
• BIS monitoring is primarily aimed at assessing depth of anesthesia and risk of intraoperative awareness, not stroke detection 1

Multimodal Monitoring Approach

• A multimodal neuromonitoring approach combining TCD, NIRS, and EEG is recommended to overcome individual limitations, as each modality has inherent technical constraints 6
• However, even with combined monitoring, the 2019 EACTS/EACTA/EBCP guidelines acknowledge that routine cerebral monitoring effectiveness remains controversial 1
• The 2025 EACTS guidelines recommend continuous monitoring of venous oxygen saturation (SvO2) and hematocrit during CPB, but these are indirect measures of cerebral perfusion 1

Clinical Reality: Should It Be Noticed?

The Harsh Truth

• In practical terms, a large embolic stroke during CPB should theoretically be detectable with comprehensive neuromonitoring, but in reality, most centers lack the resources and expertise to implement effective real-time detection 1, 6
• The 2011 ACC/AHA guidelines note that while cerebral monitoring has been investigated, the ability of current commercial devices to detect or quantify ischemia is limited 1
• Even when monitoring shows concerning changes (such as NIRS desaturation or TCD embolic signals), there is no clear consensus on immediate intraoperative interventions that can reverse an evolving stroke 6

What Can Be Done

• Immediate recognition requires a high index of suspicion when monitoring shows sudden, sustained changes: NIRS desaturation >20% from baseline, sudden loss of TCD signal, or profound EEG suppression 6
• The perfusionist and anesthesiologist should immediately assess for: air in the arterial line, pump malfunction, arterial cannula malposition, or severe hypotension 1
• Transesophageal echocardiography (TOE) is recommended during cardiac surgery and can identify intracardiac air or thrombus that may embolize 1, 7

Prevention Is More Realistic Than Detection

Key Preventive Strategies

• CO2 flushing of the CPB circuit before priming is recommended as standard of care to reduce gaseous microemboli (Class I, Level B recommendation) 1
• Arterial line filters should be considered to reduce microemboli, though routine leucodepletion filters are not recommended with membrane oxygenators 1
• Careful de-airing techniques and avoidance of aortic manipulation (no-touch technique) reduce stroke risk from 1.5% to 0.6% in CABG patients 1
• Continuous monitoring of arterial line pressure, temperature, SvO2, and hematocrit is recommended to optimize cerebral perfusion 1

Bottom Line

A large intraoperative embolic or watershed stroke during CPB will almost certainly go undetected in real-time despite available monitoring technologies, and will only become apparent when the patient fails to wake up appropriately from anesthesia. The focus should be on prevention through meticulous surgical technique, optimal CPB management, and comprehensive de-airing protocols rather than relying on monitoring to catch these events as they occur 1.