The Cerebral Circulation and Sepsis-Associated Delirium
The Journal of Intensive Care has newly published a series of sepsis-related organ dysfunction reviews. Additionally, a comprehensive yet concise overview of the cerebral circulation was just disseminated. This summary draws on both of these terrific primary resources as a point-of-departure for discussion of sepsis-associated delirium [SAD].
Cerebral blood flow [CBF] ultimately depends on 1. the arterial pressure head at the entrance to the cranial vault, 2. the cerebrovascular resistance and 3. the down-stream pressure typically estimated as the intra-cerebral pressure [ICP]. For pulmonary physiology enthusiasts, there is a glaring analogy here between alveolar pressure and intra-cerebral pressure. That is, waterfall physiology [or West Zone II in the lungs] tends to occur within the cranial vault as ICP supersedes jugular venous pressure. In other words, the pressure gradient is between the arterial input pressure and the pressure surrounding the veins [i.e. ICP]. This occurs because, normally, the central venous pressure [and therefore jugular venous pressure] approaches atmospheric pressure. Importantly, in states of very high central venous pressure [CVP] [e.g. acute cor pulmonale] and low ICP [e.g. hemicraniectomy, CSF over-drainage], the cerebral perfusion pressure may hold the CVP – rather than the ICP – as its pressure sink. In such a scenario, the transmural pressure of bridging veins will rise giving risk to their rupture.
While often assumed to be equivalent to the arterial blood pressure measured by a radial arterial line, the actual input pressure for the cranial vault lies at the level of the internal carotids and vertebral arteries. The pressure within these arteries is dependent upon the fraction of cardiac output that reaches them and their inherent elastances [i.e. ‘stiffness’]. Note that by this reasoning there is a ‘competition’ between peripheral vascular resistance and cerebral vascular resistance that determines the proportion of cardiac output which accesses the brain. If there is a profound sepsis-related fall in peripheral vascular resistance, vasodilated tissues – in essence – steal flow from the brain.
The Cerebrovascular Resistance
Yet, should arterial input pressure fall, as described above, there are mechanisms by which CBF is protected. Intracerebral arterioles have a thick smooth muscle layer with a tremendous capacity for dilation and constriction. As Donnelly clearly notes, cerebrovascular tone may be affected by: changes in PaCO2 [an inverse relationship], autonomic activity [likely predominantly via alpha-receptors] and direct changes in neuronal activity - this latter effect being mediated by neuroglia. A fall in arterial blood pressure and, by extrapolation, cerebral perfusion pressure [assuming that a fall in peripheral arterial blood pressure reflects a fall in the arterial pressure at the cranial inlet] will lead to cerebral arteriolar vasodilation [see figure 1]. This reflexive fall in cerebral vascular resistance [CVR] defends cerebral blood flow.
By the Monro-Kellie hypothesis, the intracerebral pressure is a function of the volume of each of: intra-cerebral blood, cerebral spinal fluid and brain parenchyma. An increase in volume of any of the aforementioned will raise ICP along the cerebral compliance curve. Intensification of ICP will narrow the pressure gradient along the cerebral vascular bed and, subsequently, diminish CBF [barring any change in CVR to defend blood flow as described above]. Recall that there are therapies which lower ICP in an effort to raise the pressure differential across the cerebrovascular tree. These therapies include mild hyperventilation which curtails intracerebral blood volume, ventricular drainage of CSF and osmotherapy to reduce brain edema/parenchymal volume.
Determining the best cerebral perfusion pressure [CPP] or cerebral blood flow for any given patient is difficult owing to fluctuating and varied cerebral hemodynamic pathophysiologies. One means may be to individualize monitoring by finding a patients most ‘efficient’ CPP. This can be done using the pressure reactivity index [PRx]. The PRx requires continuous monitoring of arterial blood pressure, parenchymal ICP monitoring and specific software to execute the PRx analysis. Simply, when cerebral vessels are dilated, the arterial pressure waveform will be more greatly transduced to the ICP monitor – read as a positive PRx correlation from ABP to ICP. By contrast, an increase in CVR will dampen the transmission of the arterial wave to the ICP – read as a negative PRx correlation. Thus, the CPP where PRx is at a minimum is the ‘most efficient’ driving pressure across the brain for that moment in the patient’s pathophysiology [see figure 1].
Figure 1: The normal cerebral autoregulation curve in orange. The x-axis is cerebral perfusion pressure [CPP] in mmHg, defined as the mean arterial pressure [MAP] less the intra-cerebral pressure [ICP]. Cerebral blood flow [CBF] is on the y-axis and is in litres/minute. The cartoon depicts the change in calibre of cerebral arterioles to defend CBF in response to changes in CPP. The pressure reactivity index [PRx] - in dashed lime green - is described in the text.Sepsis Associated Delirium
Also known as sepsis-associated encephalopathy [SAE], SAD is a multifactorial abnormality which falls under the umbrella of ‘acute brain dysfunction.’ Acute brain dysfunction can be quantified by a RASS score which is non-zero, while the diagnosis of delirium typically excludes coma [RASS -4, -5]. The CAM-ICU is one of the most validated screens for delirium and requires both 1. A change in mental status and 2. Inattention with either 3. Altered consciousness or 4. disorganized thinking. Delirium is defined as a RASS of -3 to +4 and a positive CAM-ICU.
The pathophysiology of SAD is incompletely known and felt to arise within a milieu of neuro-inflammation, neurotransmitter imbalance and cerebral perfusion abnormalities. A plethora of biomarkers including interleukin-6, protein C, TNF-alpha, interleukin-8, among others have been associated with delirium in animal models. Interestingly - in patients - elevated procalcitonin was associated with delirum in both infectious and non-infectious etiologies. Ultimately, cytokine elaboration systemically may alter the blood-brain barrier and lead to relative ischemia, changes in neurotransmitters and anomalous cerebral autoregulation.
Interestingly, a high level of plasma anticholinergic activity [PAA - a measure of cholinergic antagonism] is associated with the onset of delirium and C reactive protein co-varied with PAA; yet short-term administration of oral rivastigmine [a pro-cholinergic agent] did not reduce delirium in post-operative, elderly patients. This, along with the variety of other biological associations with delirium speak to the complexity of SAD pathophysiology.
Circumstantially, impaired cerebral autoregulation may be a contributing factor to SAD. High levels of C reactive protein have been correlated with impaired cerebral autoregulation and SAD – though one cannot conclude causation from these notable findings. Nevertheless, if cerebral vascular reactivity and therefore autoregulation is disturbed, cerebral blood flow will become increasingly sensitive to changes in arterial input pressure and fluctuation in ICP [i.e. the orange autoregulation curve in figure 1, becomes more linear].
Management of SAD
Clearly, source control and appropriate antibiotics are the cornerstone of infectious management, but judicious use of sedation, and early physical therapy [once respiratory and hemodynamic embarrassment has abated] are also imperative. Interestingly, statin therapy – despite its disappointment in many other ICU-associated diseases – is also being tried to mitigate SAD. A phase 2 RCT is currently underway.