Physiology II: Cerebral Blood Flow 

Normal global cerebral blood flow (CBF) is on average ~50ml/100g/min in the resting state.

For the entire brain this is ~750ml/min or 15-20% of the cardiac output.

CBF exhibits wide regional variation.

The grey matter receives proportionally more blood flow than then white matter.

In health cerebral blood flow closely matches cerebral metabolic rate and oxygen demand.

On the neurocritical care unit this relationship can be used to therapeutic effect by reducing cerebral metabolic rate with sedation and by ensuring adequate oxygenation and perfusion.

Any increase in metabolic rate, as seen in pyrexia or seizure activity, can dramatically increase cerebral blood flow. Maintaining normothermia can be another therapeutic target in brain injury.

Any cardiovascular disturbance or intervention which alters cardiac output will also alter steady state cerebral blood flow via flow mediated changes in vascular resistance.

Cerebral blood flow is directly proportional to PaCO2

An increase in brain tissue [H+] concentration from any cause will trigger vasodilatation and increase CBF.


Remember that CO2 exerts an effect on the cerebral vasculature via local buffering with the formation of [HCO3-] and [H+].

Each kPa change in PaCO2 will cause a corresponding dynamic change in CBF by ~15ml/100g/min or 30%. This will change the static cerebral blood volume by up to 10ml. Not much in absolute terms but it may be critical when intracranial compliance is close to exhaustion

The relationship of PaCO2 with CBF holds true until a lower limit is reached, at which point as a result of profound vasoconstriction, ischaemia triggers a marked vasodilatory cascade.

Transcranial doppler data suggests that the cerebral threshold for hypoxic vasodilatation is higher than one might expect at ~8.5kPa or SpO2 89-90%. Some studies have hinted that even normobaric hyperoxia (FiO2 0.85-1.0) may cause vasoconstriction and reduce dynamic CBF by up to 14%. The effect with hyperbaric oxygen may be even more pronounced but this is of little benefit at the bedside.

The Circle of Willis

4 main feeding vessels follow a tortuous path from their origins at aorta and subclavian arteries to where they anastomose to make a circle proper. This path develops centrifugal effects resulting in an axial velocity profile.


This in turn facilitates the mixing of blood and the laminar profile promotes cerebral blood flow.

The circle is congenitally incomplete in 4% of individuals and another 50% may carry anomalous or absent vessels resulting in an increased susceptability to ischaemia.


Many more individuals will acquire atheromatous change leading to carotid artery stenosis and impaired autoregulatory reactivity.

Ischaemia may develop in cerebral watershed zones in low oxygen delivery states (see Anatomy II: beyond the circle of Willis).

The cerebral arterioles and capillaries are primarily responsible for the regulation of vascular resistance and therefore autoregulation. They are more numerous and demonstrate proportionally more reactivity in their diameter than any other vessel.

Autoregulation: the Concept of Cerebral Perfusion Pressure  

CBF is subject to autoregulation.

Autoregulation is a dynamic evolutionary response to protect the cerebral circulation against arterial pressure surges e.g. seen in health during exercise.

When calculating CPP the arterial transducer should be zeroed at the level of the tragus of the ear (the surface anatomy that corresponds to the foramen of Munro). This is recommended in the UK , SBNS NASGBI statement, and is based on practice from the seminal CPP-targeted protocol in traumatic brain injury by Rosner.

Zeroing the arterial transducer at the level of the heart would result in an ~11mmHg underestimation in cerebral perfusion pressure.

The autoregulation of cerebral blood flow is controlled by three main mechanisms:

  • metabolic

  • instrinsic myogenic

  • neurogenic

Autoregulation takes time to establish, as the above mechanisms propagate through the neuronal vasculature in response to a change.

Clinically this can result in an ICP termination spike of ~10-30 seconds duration after homeostasis has been restored.

The homeostatic plateau should prevent ischaemia unless the perfusion pressure falls under the lower limit of autoregulation, seen in the graph above at ~50mmHg. The plateau should also prevent hyperaemia until the upper limit of autoregulation is exceeded, at ~150mmHg.

Even brief periods (2-3 seconds) of systemic hypotension will elicit a marked vasodilatation to maintain CBF and subsequent increase in intracranial blood volume.

Significant cerebral vessel calibre changes typically do not occur until MAP>90mmHg.

Chronic hypertension can shift the autoregulatory plateau to the right.

Many disease states seen in the neurocritical care unit effectively reduce the range at which autoregulation is active i.e. they shorten the plateau part of Lassen's curve. 

The "penumbra" or areas surrounding brain injury have been demonstrated to exhibit abnormal autoregulatory responses when interogated with Positron Emmision Tomography or CT Perfusion scans.  


Severe brain injury may obliterate the normal autoregulatory responses altogether and render the cerebral blood flow directly proportional to cerebral perfusion pressure. This so called pressure passive state carries a high risk of malignant hyperaemia and death, the best approach may involve in applying the Lund Concept (see disease specific teaching: Traumatic Brain Injury).

Cerebral Blood flow after TBI

Following a traumatic brain injury, a transient fall in CBF has been observed. In survivors the CBF will recover over the course of the first 3 days. They may then go through a hyperaemic phase of increased CBF and run the gauntlet of possible vasogenic oedema.

In 1995 Rosner et al published a "complex vasodilatory/vasoconstrictive cascade" model. In their prospective cohort study of 158 patients, it was found that by stabilising CPP at higher levels, ICP could be better controlled without ischaemia. They also reported lower mortality and improved functional outcomes.

The cascade or vortex in which an

optimal CPP will prevent vasodilatation or excessive vasconstriction and optimise CBV and CBF.

The pathogenic vasodilatory cascade:

Rosner described the point at which this cascade is active and efforts should be made to increase CPP as 

the ischaemic response threshold, It corresponds at the bedside to:

  1. ICP values that are spontaneously lower at higher levels CPP levels (indicating the new target CPP)

  2. A trend of increasing ICP over time in an attempt to halt the vasodilatory spiral

  3. the presence of cyclical CSF pressure waves (A or B Lundberg waves) are a priori evidence that CPP is inadequate.

The therapeutic vasoconstrictive cascade:

It can be therapeutically beneficial to utilise the normal autoregulatory response seen as CPP is increased.

Clinically this can be achieved with noradrenaline by infusion.

Noradrenaline used correctly in low dose will recruit unstressed circulating volume through a venoconstrictive action to improve cardiac output.


In high doses profound systemic vasoconstriction can occur resulting in an 'low flow, oversqueezed, underfilled' state. The cerebral perfusion pressure may appear adequate but ICP may be increased due to the cerebral vasodilatory effects of global ischaemia.