The appropriate fluid therapy in neurosurgical patients remains an are
a of disagreement between neurosurgeons and anaesthesiologists. Fluid
restriction has long been practiced in patients with brain pathology,
in order to reduce or prevent the formation of cerebral oedema. This g
rows from a fear that rapid administration of fluids, particularly non
colloidal fluids, can enhance cerebral oedema, although there is a lac
k of experimental evidence to substantiate this belief. On the other h
and, fluid restriction can lead to relative hypovolaemia, causing haem
odynamic instability during anaesthesia and influence defavourably cer
ebral perfusion. The appropriate fluid management of patients with bra
in pathology requires a careful review of the Starling's law and a cle
ar understanding of osmolality, oncotic pressure (OP) and the nature o
f the blood-brain barrier (BBB).The Starling equation of ultrafiltrati
on states that the net movement of fluid between the intra- and extrav
ascular compartments is the result of the summated influences of the p
ressure gradients (hydrostatic pressure, OP, and osmotic pressure) bet
ween those compartments and the properties of the barriers (capillary
endothelium) that separate them. In most peripheral tissues this barri
er is freely permeable to small molecules and ions and net fluid movem
ent depends on intravascular hydrostatic pressure and OP. Under normal
circumstances, intraluminal hydrostatic pressure is higher than inter
stitial pressure, favouring water egress. By contrast, intraluminal OP
is higher than interstitial OP, favouring water retention. These forc
es do not balance exactly, and fluid accumulation is prevented by the
lymphatics. If this net movement exceeds the capacity of the lymphatic
clearance mechanisms, fluid accumulates, which is the definition of o
edema. In the brain, the capillary endothelial cells are connected by
continuous tight intercellular junctions and there is no intracellular
gap. Moreover, the brain interstitium is made up of a tightly interwo
ven matrix of glial cells and neurons which is quite non compliant and
which limits fluid influx. The BBB is normally impermeable to large m
olecules (such as proteins) and small polar solutes enter the brain ve
ry slowly. The colloid contribution to the osmotic gradient is negligi
ble and osmolality becomes the overriding force. When the BBB is widel
y open, neither OP nor osmolar gradients can be established and oedema
formation appears to be driven entirely by hydrostatic forces. Haemod
ilution increases CBF ; this improvement however can be due either to
the decrease in viscosity or to the reduction in the oxygen carrying c
apacity of the blood, which would necessitate an increase in flow to m
aintain oxygen delivery. Oxygen delivery may well be the dominant fact
or responsible for the increase of CBF during haemodilution therapy. A
s haemodilution, while reducing viscosity, decreases also the oxygen c
arrying capacity of the blood, only moderate levels of haemodilution a
re aimed for in order to maintain adequate cerebral oxygen delivery :
indeed oxygen carrying capacity seems to be maximal at a Hct of 30-33
%, the dominant factor being the increased CBF at this level of Hct. I
n conclusion, brain oedema is defined as a net increase in the water c
ontent of cerebral tissue leading to an increase in brain mass. Fluid
therapy can defavourably influence cerebral (and peripheral) oedema fo
rmation when it reduces plasma osmolality. A decrease of OP with maint
enance of plasma osmolality is associated with peripheral oedema but n
ot with cerebral oedema. What should be kept in mind is that most comm
on crystalloid solutions - such as Ringer-lactate (but not normal sali
ne) - are hypoosmolar, and large amounts of such solutions will reduce
plasma osmolality and induce brain oedema. From above, it follows eas
ily that one of the main objectives of fluid management in patients wi
th brain pathology is to avoid any reduction in plasma osmolality. Fin
ally, for haemorheological reasons, the patients at risk of cerebral i
schaemia should be kept with a Hct of 30-33 %.