Brain injury after cardiopulmonary bypass (CPB) remains a major concern for cardiac surgery teams, even as circuits, oxygenators, and monitoring systems improve. In this narrative review, Si Zhang and colleagues examine how embolism, disturbed cerebral perfusion, inflammation, and blood–brain barrier injury interact. Their central message is that cerebral protection requires coordinated attention to multiple mechanisms throughout the perioperative period. The review is particularly relevant to perfusionists because it connects circuit management with cerebral physiology while identifying substantial gaps between promising technologies and proven patient benefit.
What the review covers
The authors searched PubMed, Scopus, and Web of Science through March 2026, emphasizing recent literature while including older foundational studies. They also reviewed reference lists. However, this was not a systematic review: it did not follow PRISMA methods, apply strict eligibility criteria, perform a formal risk-of-bias assessment, or pool numerical results. Article selection reflected the authors’ judgment. Its value is therefore an integrated educational overview rather than a definitive estimate of treatment effectiveness.
Adults are the main focus, with additional discussion of pediatric cardiac surgery and genetic susceptibility. The evidence includes clinical trials, observational studies, animal experiments, and broader neurological research. These sources provide different levels of support. Mechanistic plausibility, changes in a laboratory marker, and improved long-term cognitive outcomes should not be treated as interchangeable findings.
Neurological injury extends beyond overt stroke
The review cites stroke rates of approximately 1–5% after conventional on-pump cardiac surgery, delirium rates of 20–50%, and perioperative neurocognitive disorder rates of 10–40%. These are broad ranges drawn from the literature, not pooled estimates or outcomes from a new patient cohort. Risk varies with operation type, patient characteristics, diagnostic definitions, and the timing of assessment. Higher-risk aortic procedures may carry a different neurological burden from routine cardiac operations.
Overt stroke is readily recognized as a major complication, but disturbances in attention, memory, executive function, and mental status also affect recovery. The review emphasizes that postoperative cognitive findings must be interpreted alongside baseline vulnerability. Advanced age, vascular disease, prior neurological injury, and pre-existing cognitive impairment can influence outcomes independently of CPB. Temporal association with surgery does not establish that the bypass circuit alone caused the deficit.
Interacting mechanisms of injury
Embolic injury remains a central concern. Aortic manipulation can release atherosclerotic material, while gaseous and particulate microemboli may arise within the surgical field or extracorporeal circuit. Membrane oxygenators and arterial filtration reduce exposure but do not eliminate every embolic pathway. The review distinguishes stronger clinical evidence for major embolic events from less certain relationships between measured microembolic burden and later domain-specific cognitive decline.
Hemodynamic injury depends on more than a single pressure reading. Cerebral autoregulation normally buffers changes in perfusion pressure, but vascular disease, anesthesia, hypothermia, and other perioperative factors can alter that response. When flow becomes pressure-passive, hypotension or excessive pressure may have different consequences than they would with intact autoregulation. Adequate global pump flow and arterial pressure also do not guarantee uniform tissue-level oxygen delivery.
Inflammation adds another interacting pathway. Blood contact with artificial surfaces, ischemia–reperfusion, and operative stress can activate complement, leukocytes, and inflammatory signaling. The authors discuss microglial activation, oxidative stress, and disruption of the blood–brain barrier as potential contributors to secondary injury. Much of the detailed causal evidence comes from experimental models. Human biomarker associations support further investigation but do not establish that suppressing one inflammatory pathway will improve neurological outcomes.
Monitoring the brain during bypass
The review describes complementary roles for near-infrared spectroscopy (NIRS), electroencephalography (EEG), and transcranial Doppler. NIRS provides regional cerebral oxygenation trends, EEG reflects electrical activity, and Doppler assesses blood-flow velocity and can detect microembolic signals. Each measures a different aspect of cerebral physiology. An abnormal signal should be interpreted in the context of the patient, operative phase, perfusion conditions, and technical limitations of the device.
For perfusion teams, a cerebral desaturation event is a reason to evaluate the broader oxygen-delivery picture, including pump flow, arterial oxygen content, blood pressure, carbon dioxide, and cannula position. The paper discusses numerical intervention thresholds, but its narrative synthesis does not validate a universal target for every patient or monitor. It is more useful as a framework for multidisciplinary assessment than as a stand-alone protocol.
Individualized perfusion: promise and limitations
Autoregulation-guided management aims to identify pressure ranges that better fit individual cerebral physiology. The review distinguishes the NIRS-derived cerebral oximetry index from Doppler-derived flow-velocity indices and the intracranial-pressure-based pressure reactivity index. These measurements are not interchangeable, and their thresholds cannot simply be transferred from one method to another. In particular, an approach requiring invasive intracranial pressure measurement has different practical constraints from routine cerebral oximetry.
Temperature, hematocrit, oxygen delivery, and blood gas management remain linked to this discussion. Cooling reduces metabolic demand, while rewarming increases the need to match oxygen supply with consumption. Alpha-stat and pH-stat strategies affect cerebral physiology differently and must be considered within the relevant adult or pediatric context. The review does not establish one temperature strategy or blood gas approach as universally superior, nor does it justify reducing flow solely to address bleeding risk without preserving adequate oxygen delivery.
Pediatric vulnerability and emerging tools
Children with congenital heart disease require a broader developmental perspective. The review notes that neurological vulnerability can begin before birth and continue through surgery and recovery. Intraoperative protection remains important, but later neurodevelopment cannot be attributed to bypass exposure alone. Pediatric evidence should therefore be evaluated separately from adult findings, with attention to congenital anatomy, developmental stage, surgical complexity, and long-term follow-up.
Genetic associations, including APOE variants, may help explain differences in susceptibility, but the review does not establish routine genetic screening as a clinically validated cerebral-protection strategy. Similarly, biomarkers such as S100B, neuron-specific enolase, GFAP, and UCH-L1 require careful interpretation. Extracranial sources, hemolysis, clearance, and assay differences can complicate the relationship between a blood concentration and actual brain injury.
Artificial intelligence could eventually integrate perfusion and monitoring data to identify evolving risk. The authors emphasize that predictive accuracy is insufficient by itself: externally validated systems and prospective studies must show meaningful improvements in outcomes. AI-assisted cerebral protection, targeted anti-inflammatory drugs, and blood–brain barrier therapies remain research directions rather than established solutions supported by this review.
Implications for perfusion practice
The most useful takeaway is the need to connect embolic prevention, oxygen delivery, temperature control, neuromonitoring, and postoperative neurological assessment. The review supports coordinated risk assessment and clear communication across perfusion, anesthesia, surgery, and critical care. Its limitations include selective literature synthesis, heterogeneous populations and outcome definitions, and reliance on preliminary or indirect evidence for several emerging strategies. Some specific claims also require checking against the cited primary studies before clinical use.
Future research should measure delirium, stroke, cognition, functional recovery, and quality of life with consistent definitions and adequate follow-up. Improvements in a physiological surrogate should be tested against outcomes that matter to patients. The authors report support from the Chongqing Health and Science Joint Youth Fund and declare no conflicts of interest.
Source: Zhang S, Hao J, Fan W, Liu B, Li W, Wen L, Zhang D. Progress in Understanding Brain Injury After Cardiopulmonary Bypass in Cardiac Surgery: A Narrative Review. Reviews in Cardiovascular Medicine. 2026;27(8):50277. doi:10.31083/RCM50277.
Attribution: © 2026 The Author(s), published by IMR Press under CC BY 4.0. This independently written IPA educational summary accompanies the unchanged source article. Cover artwork is conceptual.
