Skip to content

Cerebrovascular biology and the penumbra

TL;DR — Arterial occlusion creates a gradient rather than an instantaneous uniform lesion: critically hypoperfused core is surrounded by dysfunctional but potentially viable penumbra, whose survival depends on collateral flow, systemic physiology, and elapsed time (Ginsberg 2003, PMID 12511777; Guadagno 2004, PMID 15090879). Energy failure drives ionic collapse, glutamate excitotoxicity, calcium loading, oxidative injury, proteolysis, and cell death; later inflammation and blood–brain-barrier (BBB) failure can either extend injury or support repair (Yang 2019, PMID 31611768; Candelario-Jalil 2022, PMID 35387495). Reperfusion is therefore necessary but not biologically neutral: it rescues tissue while exposing damaged microvasculature to hemorrhage, edema, and inflammatory injury. Imaging operationalizes—not directly measures—core and penumbra through tracer- and threshold-dependent surrogates. The largest clinical effect still comes from fast durable reperfusion; decades of pathway-specific neuroprotection have not matched it (Rabinstein 2020, PMID 32224752; Paul 2021, PMID 33144066).

Flow failure is a continuum

Normal brain has high continuous energy demand and little substrate reserve. Falling cerebral blood flow first suppresses function, then membrane homeostasis, and finally structural viability; temperature, glucose, oxygenation, tissue history, and collateral pressure shift every threshold (Macdonald 1998, PMID 9540326; Hossmann 2008, PMID 18222496).

Compartment Operational biology Imaging surrogate Fate without reperfusion
Oligemia Reduced flow without immediate functional failure Mild perfusion delay Often survives
Penumbra Electrical failure with temporarily preserved membrane integrity Perfusion deficit exceeding estimated core Progressively recruited into infarct
Core Severe sustained energy failure and high infarction probability Low ADC/DWI lesion or very low CBF threshold Usually irreversible, but estimates can overcall
No-reflow zone Macrovascular recanalization without adequate microvascular perfusion Persistent perfusion deficit after vessel opening Continued injury despite technical success

In a rat middle-cerebral-artery model reviewed by Ginsberg, core flow was roughly 5–20% of control and penumbral flow 20–40%; these are mechanistic experimental ranges, not validated bedside cutoffs (Ginsberg 2003, PMID 12511777). Human tissue thresholds vary across time and methods.

The ischemic cascade

Stage Dominant events Potential consequence
Seconds–minutes Oxygen/glucose loss, ATP depletion, pump failure Depolarization and cytotoxic edema
Minutes Glutamate release and impaired uptake NMDA/AMPA activation, calcium influx
Minutes–hours Mitochondrial dysfunction, ROS/RNS, acidosis Lipid, protein, DNA, and organelle injury
Hours Protease, phospholipase, nuclease activation Membrane/cytoskeletal breakdown
Hours–days Endothelial activation, leukocyte recruitment, BBB disruption Vasogenic edema, hemorrhage, secondary injury
Days–weeks Debris clearance, angiogenesis, gliosis, plasticity Repair and maladaptive remodeling

Excitotoxicity, oxidative stress, and inflammation overlap rather than follow a clean sequence. This redundancy helps explain why blocking one preclinical target often fails when applied late to heterogeneous human stroke (Siesjö 1991, PMID 1686604; Yang 2019, PMID 31611768; Paul 2021, PMID 33144066).

Excitotoxicity and ionic failure

Loss of ATP disables Na+/K+-ATPase, causing sodium and water influx, potassium efflux, depolarization, and glutamate release. Calcium entry activates calpains, phospholipases, nitric-oxide pathways, mitochondrial permeability, and death programs (Macdonald 1998, PMID 9540326; Yang 2019, PMID 31611768). The cascade is strongest in core but propagating depolarizations increase penumbral demand and lesion expansion.

Oxidative and mitochondrial injury

Mitochondrial electron-transport failure and enzymatic oxidases generate reactive oxygen species. Reoxygenation can transiently increase oxidant production; damaged antioxidant and DNA-repair systems amplify injury (Yang 2019, PMID 31611768). Hyperglycemia adds substrate for acidosis and associates with worse tissue and clinical outcomes, although correcting glucose has not reproduced the effect expected from observational association alone (Arteaga 2022, PMID 36333676).

Cell-death programs

Necrotic membrane failure dominates severe core ischemia; apoptosis-like, necroptotic, ferroptotic, autophagic, and inflammatory death pathways have been described in less severe or delayed injury. Their boundaries overlap, depend on model and time, and have not yielded validated human treatment-selection markers (Sommer 2017, PMID 28064357; Paul 2021, PMID 33144066).

The neurovascular unit

The neurovascular unit comprises endothelial cells, pericytes, basement membrane, astrocytic endfeet, neurons, microglia, oligodendroglia, and perivascular immune cells. It couples neuronal demand to local flow and maintains BBB selectivity (Jiang 2018, PMID 28987927; Liu 2020, PMID 32733433).

Component Normal role Ischemic response
Endothelium Barrier, transport, antithrombotic surface, vascular tone Swelling, adhesion molecules, permeability, thrombosis
Pericytes Capillary stability and tone Constriction/death may contribute to no-reflow
Astrocytes Metabolic support, potassium/glutamate buffering, water handling Swelling, cytokine signaling, later scar and repair
Microglia Surveillance and homeostasis Debris clearance plus context-dependent inflammatory injury
Neurons Information processing Electrical silence, excitotoxicity, death
Extracellular matrix Vessel and parenchymal structure Proteolysis, BBB destabilization, later remodeling

The once-common M1/M2 microglial binary is a useful historical shorthand but compresses time- and context-dependent states into two labels. Reviews emphasize both injurious and reparative microglial functions (Zhao 2017, PMID 28260801; Jayaraj 2019, PMID 31291966).

Blood–brain-barrier failure

Acute ischemia disrupts tight junctions, endothelial transport, basement membrane, and glial support. Oxidative stress, matrix metalloproteinases, cytokines, and infiltrating cells increase permeability; reperfusion pressure then acts on a structurally vulnerable bed (Yang 2019, PMID 30379577; Candelario-Jalil 2022, PMID 35387495).

BBB consequence Clinical/imaging expression
Ionic edema Early water/ion imbalance before bulk protein leakage
Vasogenic edema Extracellular fluid/protein accumulation and mass effect
Hemorrhagic transformation Petechial blood products through parenchymal hematoma
Contrast staining/leakage Imaging marker that can be confused with hemorrhage
Immune-cell entry Secondary injury and later clearance/remodeling

BBB recovery is active, involving endothelial, pericyte, astrocyte, and immune interactions; protecting the barrier indefinitely could interfere with clearance and repair (Jiang 2018, PMID 28987927). Timing is therefore central to any anti-inflammatory or barrier-directed intervention.

Collateral circulation: the tissue clock

Collateral routes include the circle of Willis, leptomeningeal anastomoses, external–internal carotid connections, and smaller adaptive channels. Their anatomy and pressure gradients determine how long tissue remains viable after a proximal occlusion (Lee 2023, PMID 35687300).

Good pretreatment collaterals were associated with more successful reperfusion (RR 1.28, 95% CI 1.17–1.40) and recanalization (RR 1.23, 1.06–1.42) across 24 quantitatively pooled endovascular studies with 2,239 participants. This observational association may reflect easier clot access, smaller thrombus burden, selection, or tissue resilience; it does not prove collateral-targeted treatment improves outcome (Leng 2016, PMID 26579719).

Collateral determinant Direction of concern
Proximal stenosis/occlusion pattern May recruit or exhaust alternate routes
Chronic hypertension Alters autoregulation and small-vessel structure
Systemic hypotension Reduces pressure across collateral gradients
Hyperglycemia Associates with poorer tissue outcome
Age and vascular anatomy Influence route capacity
Time and recanalization Collaterals may fail during persistent occlusion

Collateral grade predicts infarct growth, hemorrhagic transformation, and outcome after thrombectomy in observational analyses, but trials have usually selected with other imaging/clinical criteria. Its independent role in withholding or extending treatment remains unproven (Lee 2023, PMID 35687300).

Recanalization is not reperfusion

  • Recanalization means reopening the occluded artery.
  • Reperfusion means restoring flow to downstream tissue.
  • Tissue rescue means preventing infarction.
  • Clinical recovery additionally requires viable functional networks and avoidance of complications.

After IV thrombolysis alone, pooled partial-or-complete early recanalization was 33% (95% CI 27–40) across 2,063 patients, but varied by site: 52% distal MCA, 35% proximal MCA, 13% intracranial carotid, and 13% basilar; complete recanalization was only 4% for intracranial carotid and basilar occlusions (Seners 2016, PMID 27462117). This anatomy helps explain thrombectomy's value in proximal occlusion.

No-reflow can result from distal emboli, capillary plugging, endothelial/pericyte constriction, edema, and microthrombosis. Conversely, successful angiographic reperfusion may arrive after core completion and produce little functional gain—“futile reperfusion” is an outcome description, not evidence that reperfusion was biologically irrelevant.

Imaging the penumbra

Modality Putative core Putative at-risk tissue Main caveat
CT perfusion Very low relative CBF Delayed contrast transit/Tmax Vendor, motion, coverage, and threshold dependence
MR diffusion/perfusion DWI/low ADC Perfusion–diffusion mismatch DWI can reverse or be initially negative
PET Metabolic/oxygen-extraction measures Directer physiological characterization Too slow and unavailable for routine triage
Collateral CTA Indirect estimate Downstream vessel filling Timing and grading dependence

Perfusion–diffusion mismatch is an approximation: transit-time maps do not directly equal tissue viability, and some DWI lesions reverse after early reperfusion (Guadagno 2004, PMID 15090879). Nevertheless, DAWN and DEFUSE 3 showed that validated clinical–core or perfusion criteria can identify late presenters who benefit from thrombectomy (Nogueira 2018, PMID 29129157; Albers 2018, PMID 29364767).

DWI–FLAIR mismatch is another operational tissue clock for unknown-onset stroke, but scanner strength, lesion size, reader judgment, and biological variability affect classification (Emeriau 2013, PMID 23640823).

Reperfusion injury and hemorrhagic transformation

Reperfusion limits ischemic duration but introduces oxygen, pressure, inflammatory cells, and proteases into injured tissue. Risks rise with large core, severe BBB disruption, hyperglycemia, delayed treatment, and anticoagulant/coagulopathic states. Hemorrhagic transformation ranges from asymptomatic petechiae to space-occupying parenchymal hematoma (Candelario-Jalil 2022, PMID 35387495; Jiang 2018, PMID 28987927).

The net randomized effect of reperfusion remains strongly beneficial in eligible patients. “Reperfusion injury” should therefore motivate better selection and adjuncts, not delay vessel opening (Goyal 2016, PMID 26898852; Emberson 2014, PMID 25106063).

Why neuroprotection repeatedly fails

Translation gap Consequence
Young homogeneous animals versus older multimorbid humans Different inflammation, collaterals, medications, and reserve
Controlled occlusion/reperfusion versus heterogeneous emboli Mismatch in timing and tissue evolution
Infarct volume versus disability endpoint Surrogate benefit may not affect function
Pretreatment in animals versus post-onset treatment in humans Effective window missed
One pathway targeted in a redundant cascade Compensation and incomplete target engagement
Publication and model bias Efficacy appears more robust than it is

Rodent filament, photothrombotic, endothelin, and embolic models each reproduce different features; photothrombosis produces reproducible cortical injury but little penumbra, while embolic models better resemble human occlusion at the cost of variability (Sommer 2017, PMID 28064357). Contemporary reviews still identify promising anti-inflammatory, oxidative, conditioning, and BBB targets, but no adjunct has reperfusion's replicated clinical effect (Liaw 2020, PMID 32551094; Paul 2021, PMID 33144066).

Open questions

  • Can collateral augmentation preserve penumbra safely during transfer without increasing edema or hemorrhage? (Lee 2023, PMID 35687300)
  • Which imaging measure distinguishes irreversibly injured tissue from transient diffusion abnormality at the individual-patient level? (Guadagno 2004, PMID 15090879)
  • When does post-stroke inflammation transition from predominantly injurious to reparative, and can treatment be state- rather than clock-triggered? (Candelario-Jalil 2022, PMID 35387495; Jayaraj 2019, PMID 31291966)
  • What explains no-reflow after technically successful recanalization, and which microvascular intervention improves functional outcome? (Jiang 2018, PMID 28987927)
  • Can neuroprotective trials demonstrate target engagement in tissue selected to remain salvageable long enough for the drug to act? (Paul 2021, PMID 33144066)

References

  1. Ginsberg MD. Adventures in the pathophysiology of brain ischemia: penumbra, gene expression, neuroprotection. Stroke. 2003. PMID 12511777
  2. Guadagno JV, et al. Imaging the ischaemic penumbra. Curr Opin Neurol. 2004. PMID 15090879
  3. Macdonald RL, Stoodley M. Pathophysiology of cerebral ischemia. Neurol Med Chir (Tokyo). 1998. PMID 9540326
  4. Siesjö BK. Neurocytotoxicity: pharmacological implications. Fundam Clin Pharmacol. 1991. PMID 1686604
  5. Hossmann KA. Cerebral ischemia: models, methods and outcomes. Neuropharmacology. 2008. PMID 18222496
  6. Sommer CJ. Ischemic stroke: experimental models and reality. Acta Neuropathol. 2017. PMID 28064357
  7. Yang Q, et al. Potential neuroprotective treatment targeting excitotoxicity, oxidative stress, and inflammation. Front Neurosci. 2019. PMID 31611768
  8. Candelario-Jalil E, et al. Neuroinflammation, stroke, BBB dysfunction, and imaging. Stroke. 2022. PMID 35387495
  9. Jiang X, et al. Blood-brain barrier dysfunction and recovery after ischemic stroke. Prog Neurobiol. 2018. PMID 28987927
  10. Yang C, et al. Neuroinflammatory mechanisms of BBB damage in ischemic stroke. Am J Physiol Cell Physiol. 2019. PMID 30379577
  11. Jayaraj RL, et al. Neuroinflammation: friend and foe for ischemic stroke. J Neuroinflammation. 2019. PMID 31291966
  12. Zhao SC, et al. Regulation of microglial activation in stroke. Acta Pharmacol Sin. 2017. PMID 28260801
  13. Liu LR, et al. Interaction of microglia and astrocytes in the neurovascular unit. Front Immunol. 2020. PMID 32733433
  14. Lee JS, et al. Collateral status and outcomes after thrombectomy. Transl Stroke Res. 2023. PMID 35687300
  15. Leng X, et al. Collateral status and successful revascularization: systematic review and meta-analysis. Cerebrovasc Dis. 2016. PMID 26579719
  16. Seners P, et al. Early recanalization after intravenous thrombolysis: systematic review and meta-analysis. Stroke. 2016. PMID 27462117
  17. Arteaga DF, et al. Collateral status, hyperglycemia, and functional outcome after stroke. BMC Neurol. 2022. PMID 36333676
  18. Rabinstein AA. Update on treatment of acute ischemic stroke. Continuum. 2020. PMID 32224752
  19. Liaw N, Liebeskind D. Emerging therapies in acute ischemic stroke. F1000Res. 2020. PMID 32551094
  20. Paul S, et al. Emerging neuroprotective strategies for ischemic stroke. Exp Neurol. 2021. PMID 33144066
  21. Emeriau S, et al. DWI-FLAIR mismatch at 3 Tesla and stroke age. Stroke. 2013. PMID 23640823
  22. Nogueira RG, et al. Thrombectomy 6 to 24 hours after stroke. N Engl J Med. 2018. PMID 29129157
  23. Albers GW, et al. Thrombectomy 6 to 16 hours with perfusion selection. N Engl J Med. 2018. PMID 29364767
  24. Goyal M, et al. Endovascular thrombectomy meta-analysis. Lancet. 2016. PMID 26898852
  25. Emberson J, et al. Effect of treatment delay on alteplase outcomes. Lancet. 2014. PMID 25106063