Cerebral oedema

Though this condition appears constantly wherever neurology or neurosurgery patients are abundant, the college has never really asked their trainees about it in the exams until Question 10 from the first paper of 2019. This was generally well answered, but not too well  (75% of the candidates passed) which suggests that the question may at some stage reappear as a viva station or SAQ repeat. Given that this is a basic science-based topic which is somewhat removed from the pragmatic reality of bedside practice, questions on cerebral oedema may ultimately end up being demoted to Part One content, and if that happens this revision chapter will also migrate in that direction.

For an elaborately detailed overview of the cellular mechanisms of cerebral oedema, one should go to Stokum et al (2016)Mahajan et al (2016) also cover this ground, but in less detail, and they offer additional details regarding clinical manifestations and radiological changes. Some combination of these two articles would be enough to generate a pass mark for Question 10 from the first paper of 2019, but individually they are quite unwieldy and difficult to parse into a ten-minute answer. In search of a short pithy summary, one might be tempted to turn to Oh's Manual, but this would not be a rewarding experience- it has numerous mentions of cerebral oedema in various locations, but if you look it up in their index, the page reference for some reason leads you to a page on bacterial meningitis in children, which is bizarre and unhelpful. 

Definition and classification of cerebral oedema

To borrow a fair-sounding definition from Igor Klatzo, 

"[Cerebral oedema is] ... an abnormal accumulation of fluid within the brain parenchyma producing a volumetric enlargement of brain tissue"

That's brief, memorable and difficult to argue with. Most people would agree with this outline. However, that's where the certainty ends, as far as scholarly agreement and cohesion are concerned. It appears that virtually everybody who has ever written on this topic has come up with their own method of classification for cerebral oedema. Moreover, there is no organisation to govern over this disease entity, and so the lawless landscape of the literature is dominated by opinion articles published by single authors.  A good classification is discussed by Milhorat (1992), and probably represents the most granular description of the various types of cerebral oedema, using "classical" terms to describe them. The evolution of these classical terms is discussed by Klatzo (1994), who lends the subject authenticity given that his publications span across the latter half of the twentieth century. Betz et al (1989) proposed we classify cerebral oedema according to whether or not the blood-brain barrier is intact.  Iencean (2003) proposed a separate scheme, which probably makes a great deal more sense that the classical system, but which must have been viewed as heretical, as nobody appears to have adopted it. The excellent article by Mahajan & Bhagat (2016) contains yet another classification model, which is again slightly different.

In an attempt to reconcile this mess of definitions into something which contains the greatest granularity, one may end up with categories which are not widely recognised

  • Cellular oedema
    • Cytotoxic oedema (due to disturbance of cellular osmoregulation)
    • Metabolic storage (due to abnormal intracellular accumulation of molecular products)
  • Extracellular oedema
    • Vasogenic oedema (due to an increase in brain capillary permeability)
    • Osmotic oedema (due to an osmotic gradient between the plasma and the interstitial fluid, across an intact blood-brain barrier)
    • Interstitial oedema (obstruction of interstitial fluid flow pathways)
    • Hydrocephalic oedema (obstruction of CSF flow pathways)

Mechanisms of cerebral oedema

Cytotoxic oedema occurs as the consequence of broken intracellular osmoregulatory mechanisms. The term was first introduced by Igor Klatzo in 1967. Basically, the central problem is usually a failure of cellular energy supply, which powers the Na/K ATPase, that ever-busy bilge pump of the cellular basement. In the absence of ATP (for any reason), this pump stops sucking ions. In the absence of its equilibrium-disturbing activity, the cell membrane reverts to a passive Gibbs-Donnan "equilibrium", and begins to entrain water (because the intracellular content is significantly hyperosmolar in comparison to the interstitial fluid).

Though energy failure (i.e. cell death) is usually responsible, theoretically one may produce this sort of effect in energy-replete cells by directly disabling the Na/K ATPase, eg. by using ouabain. That should produce cerebral oedema with totally normal mitochondrial function. Doggett & Spencer (1971) did just that, generating a whole pile of dead albino mice by injecting this toxin directly into their brains. "Ouabain produced a profound, dose related depression of central nervous activity", the authors cackled. The role of broken ion pumps was abundantly demonstrated some years later in an even more hideous experiment by Gazendam et al (1979), who measured fluid dripping from nylon wicks which they inserted into the brains of ouabained cats. Generally this sort of thing is not seen in human clinical practice, even in parts of the world where ouabain is still ethnobotanically relevant, because usually the heart stops well before the brain becomes oedematous.

Metabolic storage oedema is basically the intracellular accumulation of something abnormal and bulky which increases the volume of cells in the presence of initially normal cellular function. It is worth mentioning somewhere, but many might argue that it does not meet the colloquial definition of cerebral oedema as a disorder where water accumulates in brain tissue, adding to its bulk. One might object that metabolic storage diseases cover a vast array of possible metabolic disturbances, and that water is only one among a plethora of different substances which might abnormally accumulate in cells. To quote a nice list from Milhorat (1992), the possibilities include "glycogen (Pompe's disease), mucopolysaccharides (Hurler's disease), GM2 ganglioside (Tay-Sachs disease), glycosyl ceramide (Gaucher's disease), and sphingomyelin (Niemann-Pick disease)". A strict purist might eject this from the list of causes of cerebral oedema because, to be totally precise, the gradual accumulation of intracellular sphingomyelin does not represent oedema per se, and should instead be viewed as a billion tiny space-occupying lesions. On the other hand, many of these molecules add to the intracellular osmoles. For instance, the cerebral oedema which develops due to hyperammonaemia is due to the accumulation of glutamine inside astrocytes, as a metabolic byproduct of ammonia metabolism. 

Vasogenic oedema occurs due to an increase in brain capillary permeability, i.e. a broken blood-brain barrier. In contrast, this barrier may remain intact while all the other mechanisms of cerebral oedema are taking place. The blood-brain barrier may be disturbed by a massively numerous list of pathologies, including hypertension (PRES), infection (meningitis), neoplastic angiogenesis, autoimmune encephalitis, drugs, trauma, hyperthermia, and so on. The fluid which leaks out into the brain parenchyma is typically plasma-like, i.e. highly proteinaceous. 

It is generally said that white matter, where extracellular spaces are larger, is more affected by this mechanism, whereas grey matter (which has higher density) is somehow more oedema-resistant. The most important characteristic from a pragmatic point of view is that the mechanism of this sort of oedema is usually related to endothelial breakdown and vascular hydrodynamics, both of which are relatively susceptible to manipulation. The driving pressure of the blood can be decreased (eg. the aggressive management of hypertension in PRES), and the endothelial integrity can be protected by drugs which act on the inflammatory mediators which disrupt it (dexamethasone, for example).

Osmotic oedema is generated by the sudden appearance of an osmotic gradient between the plasma and the interstitial fluid, or between the extracellular and the intracellular fluid. Excellent clinical examples include the well-meaning abrupt removal of vast amounts of urea by means of a dialysis filter, or the well-meaning abrupt removal of vast amounts of glucose by means of an insulin infusion. Under both circumstances, the cells which had adapted to chronic hyperosmolarity suddenly find themselves in a bath of relatively watery hypoosmolar fluid. That water of course migrates into them osmotically, increasing their volume. The alternative method of producing this effect when one's osmolarity is chronically normal would be to flood oneself with a relatively hypoosmolar fluid, eg. by compulsively drinking six litres of water over a couple of hours.

Compressive oedema supposedly occurs when there is some mechanical obstruction of interstitial fluid flow pathways in the brain, in the presence of normal blood brain barrier and cellular function. Apparently this occurs due to pressure effects, eg. from benign tumours such as meningiomas.  If you've never heard of this sort of oedema before, it may be because it, as a discrete pathophysiological entity, is not widely recognised by the scientific community. The only mention of it is in the article by Milhorat (1992), and the only reference Milhorat gives for it is Milhorat (1987). The idea seems to have originated in the 1980s, with authors such as Gilbert et al (1983) observing extensive histological evidence of oedema in the peri-tumour brain tissue, but with completely intact gap junctions and normal-looking cellular activities in the astrocytes. Surely, this cannot be the usual vasogenic oedema of malignancy, they reasoned; some sort of mechanical effect must be encouraging the buildup of fluid. Subsequent investigations had determined that meningiomas are in fact quite actively involved in destroying the integrity of the blood-brain barrier and that the gap junctions are not in fact intact, suggesting that this "compressive" oedema concept was just vasogenic oedema all along.

Hydrocephalic oedema (also called "interstitial" by some authors) is the infiltration of CSF into the interstitial spaces of the brain, resulting from obstruction of CSF flow pathways. This might be likened to "lymphoedema of the brain", except that the brain has no lymphatics. It is seen in hydrocephalus, where raised intraventricular pressure drives CSF into the periventricular parenchyma. The fluid is able to escape the ventricles in this fashion because the ventricular ependymal lining is damaged by the increased pressure.  An early Milhorat et al (1970) determined this by inflating balloons in the fourth ventricles of rhesus monkeys. Gross tears of the ependyma were visible in sectioned specimens, and dye added to the CSF was found to migrate deep into the periventricular white matter, whereas ordinarily it would be confined to the ventricles. This can often be seen on CT as oedema surrounding the lateral ventricles. White matter tracts appear to be the most affected; with chronic oedema myelin sheaths become damaged and local blood supply is compromised. The fluid in this scenario is CSF, and is therefore relatively protein-poor, in contrast to the fluid which accumulates in vasogenic oedema. 

Clinical features of cerebral oedema

Question 10 from the first paper of 2019 asked for clinical manifestations of cerebral oedema, which was a weird thing to ask about, as it is not a discrete disease entity all on its own but rather an epiphenomenon of another disease process. As such, the manifestations of that disease process will be dominant among the clinical findings. It is therefore difficult to separate the clinical manifestations of cerebral oedema on its own from the manifestations of the causative process, and from the clinical manifestations of raised intracranial pressure. This is abundantly demonstrated by the college answer, which, for manifestations of cerebral oedema, basically lists all the clinical features of raised ICP (eg. Cushing’s reflex and papilloedema).

So, what would cerebral oedema feel like, if the ICP remained normal? Well. Surely, the experimental conditions required to determine the answer to this question would be fairly difficult to achieve. To detect the presence of oedema implies either finding cellular swelling or increased interstitial fluid, suggesting that a biopsy of the brain or at least some sort of invasive monitoring would be required. But that is the same brain that is complaining of the symptoms to you as it becomes more oedematous, and so one would be destroying one's measuring instrument while it's reporting the outcome data. Also, the measuring instrument is probably being focally destroyed by a tumour or intracranial haemorrhage, the features of which might obscure those unique to oedema. For the majority of the clinical features listed below, the main source was the UpToDate article on high altitude cerebral oedema.


  • Headache
  • Photophobia
  • Neck stiffness
  • "Lassitude" 


  • Irritability and confusion
  • Ataxia
  • Agitation, delirium
  • Decreased level of consciousness
  • Seizures

With raised ICP:

  • Hypertension
  • Bradycardia
  • Mydriasis
  • Loss of light reflex
  • Papilledema on fundoscopy

Radiological appearance of cerebral oedema

In broad brushstrokes, diffuse cerebral oedema has characteristic radiological features, in addition to which there may also be features of raised intracranial pressure. The college answer to Question 10 from the first paper of 2019 lists uncal and tonsillar herniation as radiological features, which - some might argue - are more related to raised ICP and local mass effects of space-occupying lesions. An excellent resource for this was Weisberg et al (1990), from the journal Computerized Medical Imaging and Graphics.  Ironically, the original CT images available in the scanned online version of the article have taken on a horrible early Xerox sort of quality, to the point where one can barely tell what bodypart was scanned. Fortunately, the writing sparkles, and there is a substantial amount of technical detail. This paper was pillaged and its valuable contents presented below. Additionally, a more recent article by Ho et al (2012) is available, and its digital images are much better preserved.

 Vasogenic oedema

  • Primarily affecting white matter
  • Indentations of edema with frond-like projections into normal gray matter
  • Enhancement following contrast
  • Usually focal, i.e. due to some local lesion

Cytotoxic oedema

  • Usually diffuse
  • Does not enhance with contrast
  • Involves subcortical nuclear gray matter structures (e.g., basal ganglia, thalamus).

Hydrocephalus-associated "interstitial" oedema

  • Transependymal fluid: low attenuation periventricular changes around the lateral ventricles. These are usually most prominent surrounding frontal and occipital horns.

Oedema in general

  • effacement of the sulci
  • effacement of the ventricles
  • loss of grey-white differentiation


Stokum, Jesse A., Volodymyr Gerzanich, and J. Marc Simard. "Molecular pathophysiology of cerebral edema." Journal of Cerebral Blood Flow & Metabolism 36.3 (2016): 513-538.

Mahajan, Shalvi, and Hemant Bhagat. "Cerebral oedema: Pathophysiological mechanisms and experimental therapies." Journal of Neuroanaesthesiology and Critical Care 3.04 (2016): S22-S28.

Iencean, S. M. "Brain edema–a new classification." Medical hypotheses 61.1 (2003): 106-109.

Milhorat, Thomas H. "Classification of the cerebral edemas with reference to hydrocephalus and pseudotumor cerebri." Child's Nervous System 8.6 (1992): 301-306.

Klatzo, Igor. "Evolution of brain edema concepts." Brain Edema IX. Springer, Vienna, 1994. 3-6.

Reichardt, M. "Zur Entstehung des Hirndrucks bei Hirngeschwülsten und anderen Hirnkrankheiten und über eine hei diesen zu beobachtende besondere Art der Hirnschwellung." Deutsche Zeitschrift für Nervenheilkunde28.2-4 (1905): 306-355.

Klatzo, Igor. "Neuropathological aspects of brain edema." Journal of Neuropathology & Experimental Neurology 26.1 (1967): 1-14.

Doggett, N. S., and P. S. J. Spencer. "Pharmacological properties of centrally administered ouabain and their modification by other drugs." British journal of pharmacology42.2 (1971): 242.

Gazendam, Jurjen, K. Gwan Go, and Annie K. Van Zanten. "The effect of intracerebral ouabain administration on the composition of edema fluid isolated from cats with cold-induced brain edema." Brain research 175.2 (1979): 279-290.

Betz, A. L., F. Iannotti, and J. T. Hoff. "Brain edema: a classification based on blood-brain barrier integrity." Cerebrovascular and brain metabolism reviews 1.2 (1989): 133-154.

Philippe, Geneviève, and Luc Angenot. "Recent developments in the field of arrow and dart poisons." Journal of ethnopharmacology 100.1-2 (2005): 85-91.

Walters, R. J. L., et al. "Haemodialysis and cerebral oedema." Nephron 87.2 (2001): 143-147.

Maccario, Micheline, and ChrystopoulosP Messis. "Cerebral oedema complicating treated non-ketotic hyperglycaemia." The Lancet 294.7616 (1969): 352-353.

Ligtenberg, J. JM, et al. "A lethal complication of psychogenic polydipsia: cerebral edema and herniation." Intensive care medicine 24.6 (1998): 644-645.

Milhorat, Thomas Herrick. Cerebrospinal fluid and the brain edemas. New York Society of Neurosurgery, New York (1987)

Gilbert, Joseph J., et al. "Cerebral edema associated with meningiomas." Neurosurgery 12.6 (1983): 599-605.

Milhorat, Thomas H., et al. "Structural, ultrastructural, and permeability changes in the ependyma and surrounding brain favoring equilibration in progressive hydrocephalus." Archives of Neurology 22.5 (1970): 397-407.

Weisberg, Leon, Jack Greenberg, and Antonio Stazio. "Computed tomographic findings in brain swelling." Computerized Medical Imaging and Graphics 14.4 (1990): 263-268.

Ho, Mai-Lan, Rafael Rojas, and Ronald L. Eisenberg. "Cerebral edema." American Journal of Roentgenology 199.3 (2012): W258-W273.