Knowledge about brain oedema has changed dramatically with recent developments in MRI (Marmarou et al. 1997; Kuroiwa et al. 2000). Cerebral oedema accompanies many brain pathologies and contributes to the resultant morbidity and mortality (Katzman et al. 1977; Marmarou et al. 2000).
It plays a major role in head injury, stroke, and brain tumour, brain abscess, encephalitis, meningitis, lead encephalopathy, hypertensive encephalopathy, hypoxia, hypo-osmolality, dialysis dysequilibrium, diabetic ketoacidosis, and obstructive hydrocephalus. Brain oedema can now be measured accurately in vivo with MRI (Marmarou et al. 2000).
The distribution of water in neurons, glia, endothelial cells, and in the interstitial spaces can be determined with diffusion-weighted imaging (DWI). The oscillation of water molecules depends upon the space available within and between cells: when space is limited, movement of molecules is limited and this ‘squeeze’ can be measured with MRI and expressed as activated diffusion coefficients (ADC). So, if cells fill with water (cytotoxic oedema), they swell and reduce the interstitial space, thus reducing the activated diffusion coefficient. This ability to oscillate determines the diffusion of water molecules, so that diffusion-weighted imaging measures the interstitial water content.
When the interstitial space becomes filled with water, for example around a brain tumour, the activated diffusion coefficient increases because water molecules are free to oscillate and diffuse over greater distances. This type of intercellular oedema is known as vasogenic oedema and is associated with an increased activated diffusion coefficient.
Brain oedema must be distinguished from engorgement due to an increase in the blood volume of the brain due to venous obstruction or vasodilatation. However, prolonged venous engorgement may lead to brain oedema. If localized or mildly generalized, oedema produces few symptoms and signs. If severe, it may cause major focal signs if it is localized to one cerebral hemisphere, and if generalized, it can give rise to the brain herniation described above.
Cerebral oedema was subclassified initially into vasogenic, cellular or cytotoxic, and interstitial (or hydrocephalic) types (Klatzo 1967; Fishman 1980). A better understanding also comes from classifying brain oedema as open-barrier (vasogenic) and closed-barrier (cytotoxic) oedema (Betz et al. 1989).
Vasogenic (open-barrier) oedema
The vasogenic variety of oedema, associated with increased capillary permeability and an open blood–brain barrier (BBB), is the most common form observed in clinical practice. It occurs in conditions such as tumour, abscess, haemorrhage, infarction, contusion, and purulent meningitis. The oedema is usually localized around the primary lesion. This produces focal symptoms and signs that are often more due to the oedema than to the primary lesion. It is associated with increased activated diffusion coefficients on diffusion-weighted imaging. Open-barrier oedema is much more likely to respond to intervention with steroids such as dexamethasone than the other types of cerebral oedema that are considered below.
Cytotoxic (closed-barrier) oedema
Cellular, or cytotoxic, oedema is characterized by swelling of all the cellular elements of the brain—neurons, glia, and endothelial cells—with an associated reduction in extracellular fluid but with an intact blood–brain barrier. It resembles that due to water intoxication in experimental animals, or that induced experimentally by triethyl tin, in which, however, there are also vacuoles and clefts in the cerebral white matter. It is characterized by swelling of all the cellular elements of the brain, with an associated reduction in extracellular fluid but with an intact blood–brain barrier.
The activated diffusion coefficient on diffusion-weighted imaging is reduced. It occurs clinically in diffuse brain hypoxia, acute hypo-osmolality due to dilutional hyponatraemia, sodium depletion, or excess antidiuretic hormone (ADH) secretion, or in osmotic disequilibrium syndromes, such as in haemodialysis or diabetic ketoacidosis. The clinical manifestations are usually more generalized than in vasogenic oedema, and include drowsiness, stupor or coma, and sometimes convulsions. In ischaemic states a combination of vasogenic and cytotoxic oedema is often seen. The ultrastructural and molecular mechanisms occurring in the cell membrane are now becoming understood. Ischaemia results in the release of lactate and excitatory amino acids, glutamate and aspartate, which open receptor-activated calcium channels. In contrast, lactate does not rise very much or very early in traumatic oedema, which may therefore be different from ischaemic oedema (Eriskat et al. 2000).
The influx of calcium leads to the activation of α-amino-hydroxyl-methyl proprionic acid (AMPA) and metabotropic receptors, with upregulation of genes, which may activate lysozymes, with resultant apoptosis. Massive increases in calcium lead to mitochondrial dysfunction, energy failure, cell membrane rupture, and necrosis. In Reye's syndrome, the oedema is cytotoxic and resembles that of triethyl tin intoxication.
Interstitial or hydrocephalic oedema simply identifies the increased water content of the periventricular brain (largely extracellular), which is seen in hydrocephalus. The main site of accumulation of water is periventricular and manifests itself on CT scan as periventricular lucency.
Recognition of the type of oedema has implications with respect to treatment. High doses of steroids (dexamethasone, betamethasone) are of proven efficacy in most forms of vasogenic oedema but not in cytotoxic oedema. In cytotoxic oedema, osmotherapy with hypertonic mannitol or diuretics such as frusemide may be useful. While neuroprotective drugs appear to protect animals from the most severe effects of cerebral ischaemia, their role in the management of brain oedema is still uncertain. This applies particularly to the oedema seen with head injury, where prospective randomized controlled trials of steroids, barbiturates, N-methyl-D-aspartate (NMDA) receptor antagonists, and calcium antagonists have been shown not to improve outcome (Ward et al. 1985; European Study Group on Nimodipine 1994; Yates et al. 1999; Iannotti 2000).