Optimal strategies for oxygen delivery in brain injury

Optimal strategies for oxygen delivery in brain injury

Critically ill Walker Translation Group

I.
Preface

Optimizing oxygen delivery is an important priority for acute brain injury (ABI), namely subarachnoid haemorrhage (SAH), intracranial hemorrhage (ICH), acute ischaemic stroke (AIS), and traumatic brain injury (TBI
).
Severe ABI activates a series of undesirable processes in neurons, leading to excitatory neurotransmitter release, protease activation, lipid peroxidation, free radical production, and ultimately induce apoptosis and necrosis
.
These cascades underlie the clinical manifestations of secondary brain injury and are the main drivers of neurological prognosis deterioration of brain injury
.
In particular, oxygen imbalance (i.
e.
, excessive or insufficient oxygen supply to oxygen demand) may induce secondary brain injury through a variety of mechanisms, with a worse
prognosis for all diseases associated with ABI.
Therefore, oxygen-based treatment goals are promoted to be included in
the treatment of patients with ABI.

Traditionally, ABI interventions have focused on intracranial pressure (ICP), but strategies to control ICP do not always guarantee optimal oxygen delivery and can sometimes exacerbate the imbalance between oxygen supply and demand in the brain as a whole or locally
.
Therefore, recent clinical algorithms seek to combine the target ICP to optimize oxygen delivery
.
This is a rapidly evolving area of research, and this article will focus on the optimal strategy of oxygen transport in ABI patients, focusing on the basic physiological theories and latest evidence
guiding clinical treatment.

Oxygen transport to the damaged brain: why it matters and how (where) to monitor it

Although a healthy brain accounts for only 2% of total body weight, it receives about 15% of cardiac output (CO) and a blood flow rate of 700 ml/min (or 50-60 ml/100 g)
to brain tissue.
Because the brain has little energy storage, neurons rely on aerobic metabolism to continuously provide oxygen to produce energy
.
This highly dependent situation is most clinically highlighted, where there may be a complete loss
of consciousness within 10 seconds of a complete cessation of cerebral blood flow (CBF).
Under the condition of continuous decrease in cerebral blood flow, the amount of neuronal oxygen extraction is increased to maintain a stable oxygen supply; However, this process is depleted when the CBF is below 25-30ml/100g/min, at which point neurons are converted to anaerobic metabolism, and with it, lactic acid and hydrogen ions
are produced.
When the flow rate is 10-12 ml/100 g/min, the sodium-potassium pump fails and the nerve transmission stops
.
Blood flow is further reduced, leading to cell death
.

Therefore, oxygen optimization strategies need to maintain a dynamic balance
between cerebral oxygen supply and demand.
In clinical practice, this matching is guided
by systemic and brain-specific oxygenation targets.
Arterial blood gas analysis can systematically measure oxygen, such as partial pressure of oxygen (PaO2) or arterial oxygen saturation (SaO2), which reflects the oxygenation status of the body as a whole but is not a direct marker
of oxygen supply to the brain.
Measurements can also be made at the brain level by a brain tissue partial pressure of oxygen (PbO2) monitor that measures free dissolved oxygen
within a 1mm3 region of the sampling point.
Various catheters are available for PbO2 monitoring, including Licox and Neurovent
.
Direct insertion into the brain parenchyma by drilling and well tolerated and accurate
.
PbO2<20 mmHg in patients with severe TBI is a common threshold<b16> for intervention to improve oxygenation.
Other strategies to guide cerebral oxygenation include: jugular venous oxygen saturation (SjO2), a probe placed in the jugular bulb that measures the overall ratio of oxygen delivery to oxygen consumption; Brain microdialysis, which can measure lactic acid, pyruvate and other metabolites in brain tissue and indicate insufficient oxygen supply by the ratio of lactic acid to pyruvate; Finally, near-infrared spectroscopy (NIRS) can noninvasively measure cerebral oxygen saturation at a depth of about 2 cm from the skin surface on the forehead
.
Table 1 further elaborates on the technical points, advantages and limitations
of each method.

For each of these sampling methods, oxygenation indicators can be used to guide oxygen supply
.
Table 2 lists the results of neurological studies
based on PaO2, PbO2, jugular venous oximetry, microdialysis, and near-infrared spectroscopy (NIRS) monitoring targets.

Table 1 Overview of oxygen measurement methods

ICP, intracranial pressure; PaO2, arterial oxygen partial pressure; PbO2, brain tissue oxygen; SaO2/SpO2, percent saturation of hemoglobin; TBI, traumatic brain injury

Table 2 Oxygen monitoring methods and prognostic studies related to disease conditions

AIS, acute ischemic stroke; CBF, cerebral blood flow; CI, confidence interval; CPP, cerebral blood flow; DRS, disability grading; GCS, Glasgow Coma Score; GOS, Glasgow Prognostic Score; GOS-E, Glasgow Prognostic Score Extension; ICH, intracranial hemorrhage; ICP, intracranial pressure; L: P, lactic acid/pyruvate ratio; MRS, improved Rankin score; NIRS, near-infrared spectroscopy; OR, odds ratio; PaO2, arterial partial pressure of oxygen; PbO2, oxygen content of brain tissue; SAH, subarachnoid hemorrhage; SjO2, jugular venous oxygen saturation; TBI, traumatic brain injury; TCD, transcranial Doppler

Third, the optimization strategy of oxygen transport in acute brain injury

By controlling each determinant of cerebral oxygen supply and oxygen consumption, oxygen delivery
can be optimized.
Cerebral oxygen supply depends on arterial oxygen content (CaO2), cerebral perfusion pressure (CPP), and cerebrovascular reactivity, while oxygen consumption is mainly dependent on cerebral oxygen metabolic rate (CMRO2).

Another factor to consider is diffuse injury, microvascular collapse, perivascular edema, and endothelial cell swelling, which limits the oxygen utilization of neurons, a phenomenon that occurs in traumatic brain injury and has recently been observed
in cardiac arrest patients.

Arterial oxygen content (CaO2) is related to hemoglobin level (Hb), hemoglobin saturation (SaO2) percentage, and has a small relationship with arterial oxygen partial pressure (PaO2), summarized as follows:

CaCO2(ml/100ml)=(1.
34×Hb×SaO2)+(PaO2×0.
003)

At the same time, CPP depends on systemic mean arterial pressure (MAP) and ICP, which are worth the difference between two values:

CPP（mmHg)=MAP-ICP

Cerebrovascular reactivity relies on arterial partial pressure of carbon dioxide (PaCO2), with elevated PaCO2 leading to cerebral vasodilation and decreased PaCO2 leading to cerebral vasoconstriction
.
When PaCO2 is in the 20-80mmHg range, CBF increases by about 3%
for every 1mmHg increase in PaCO2.
Cerebrovascular reactivity is also inhibited by hypoxemia, and cerebral blood flow increases significantly linearly when PaO2 drops below 50-60 mmHg (cerebral blood flow is almost unchanged when PaO2 is at physiological levels).

CMRO2 is dependent on cellular activity, can be increased by factors such as fever, epilepsy, or sympathomimetic, and can decrease
due to factors such as sedation or hypothermia.
Figure 1 discusses each of the factors associated with oxygen supply and consumption and briefly illustrates their importance
.
Strategies
to optimize oxygen transport using these relationships are then discussed.

3.
1 Arterial blood oxygen content

Although there is no strong evidence to guide the choice of specific haemoglobin targets, oxygen delivery
to the injured brain can theoretically be improved by increasing haemoglobin levels and arterial oxygen levels.
Two international surveys of critical care practitioners found that the haemoglobin threshold for blood transfusions from acute brain injury was very inconsistent: in one survey, more than 50% of practitioners believed that haemoglobin levels above 80 g/l in patients with TBI, SAH and stroke required blood
transfusions.
The second survey found that the transfusion threshold varied between 70 and 100 g/l based on the acuity of performance, and more than 50% of participants listed low PbO2, CPP and CBF as the most important physiological factors
contributing to the decision to transfuse.

Figure 1: Determinants of brain oxygen supply and demand

In patients with TBI, a randomized clinical trial (RCT) investigating the effects of erythropoietin and two hemoglobin transfusion thresholds (70 g/lVS.
100 g/l) found no difference in neuroprognosis between groups at 6 months post-injury, while thromboembolic complications increased
in the high-threshold group.
Another ongoing RCT comparing neurological outcomes in patients with moderate to severe head injury (NCT03260478) based on restrictive (<70 g/L) and liberal (< 100 g/L)
transfusion targets.
Another randomised controlled trial comparing transfusion threshold less than 80 g/L with less than 100 g/L in people with SAH is also ongoing (NCT03309597).

To date, no RCTs have assessed transfusion thresholds
in people with intracerebral haemorrhage or AIS.
To balance the available evidence, a recent guideline recommends haemoglobin below 90 g/L with hypoxia of brain tissue (i.
e.
, PbO2<20 mmHg) as the final measure<b14> to improve oxygenation in transfusions in patients with severe TBI.
Whether this recommendation should be applied more broadly to other ABI populations is unknown, and until further evidence is available, a sound clinical strategy is to transfuse haemoglobin below 70 g/l, while carefully considering the benefit
of transfusion at 70 to 100 g/l of haemoglobin at low PbO2 levels.

Arterial oxygen levels can also be increased by increasing PaO2, such as hyperbaric oxygen
.
In phase II clinical trials, severe TBI received 1.
5 atmospheric hyperbaric oxygen followed by normobaric hyperoxi, lower intracranial pressure (ICP) at 6 months, improved markers of oxidative metabolism, and better neurological outcomes compared with usual care
.
Hyperbaric oxygen studies have also been conducted in patients with subarachnoid haemorrhage, adenocarcinoma in situ (AIS), and preclinical models of intracerebral hemorrhage, all of which have seen an outcome benefit (possibly due to factors other than increased PaO2).

However, despite these results, the current evidence for hyperbaric oxygen is weak and should be considered an experimental approach
in patients with ABI.

3.
2 Cerebral perfusion pressure

CPP can be regulated
by controlling mean arterial pressure (MAP) or intracranial pressure (ICP).
Since MAP is a product of heart rate, stroke volume, and systemic vascular resistance, these factors can be further controlled to optimize cerebral perfusion pressure (CPP) and thus PbO2
.
Currently, severe TBI guidelines recommend maintaining CPP at 60-70mmHg
.
With this overall goal in mind, fluid response patients can be increased by intravenous fluids
.
Rational use of vasopressors increases systemic vascular resistance
.
Keep in mind that higher vasopressor doses may increase oxygenation
by increasing cerebrovascular resistance.
If bradycardia decreases MAP, it should be corrected (e.
g.
, use of a time-lapse or increase pacing rate in patients with a permanent pacemaker).

These strategies should be aware of previous comorbidities (e.
g.
, heart failure) or fatal acute brain injury (e.
g.
, neurogenic pulmonary edema, stress cardiomyopathy), which may limit their use
.

Care must also be taken to optimize ventilation in patients receiving mechanical ventilation to minimize adverse effects
of CPP, ICP, and PbO2.
Excessive positive end-expiratory pressure (PEEP) may reduce preload and stroke volume, which in turn leads to lower
MAP and CPP.
These effects are amplified when high PEEP is used in hypovolemic patients, and PEEP can be increased to improve volume status
if appropriate.
PEEP also has a complex and variable impact on ICP through the Starling mechanism; The general strategy is to keep the head elevated, reduce the effect of PEEP on ICP (minimize the effect of PEEP on ICP), and set PEEP below ICP levels
.
To avoid the haemodynamic and intracranial effects described above, there may be a preference to minimize PEEP, but this is undesirable: PEEP improves alveolar recruitment and V/Q matching, thereby improving systemic oxygenation and PbO2
.
Initially, PEEP should be set at the same level
as without ABI in patients with ABI and without ICP elevation.
For patients with acute brain injury and high ICP, PEEP is carefully adjusted to optimize and balance ICP
and related parameters (e.
g.
, MAP, CPP, PbO2).

Strategies to control ICP include maintaining the head of the bed elevated 30°, ensuring adequate jugular venous drainage, draining cerebrospinal fluid through ventricular drainage, and osmotherapy
.
Decompression of the bone flap is the last option
.
For a given MAP, lowering ICP through the above measures can improve CPP and generally PbO2
.
A recent consensus guideline on intracranial pressure therapy in patients with severe TBI suggests the sequencing of these treatments
.

3.
3 Cerebrovascular reactivity

PaCO2-mediated cerebrovascular reactivity can be used to optimize cerebral blood flow and PbO2
.
In patients receiving mechanical ventilation, while recognizing the effects of ICP, the respiratory rate and tidal volume can be carefully adjusted to control PaCO2
.
In the BOOST-II pilot trial, patients with PbO2 less than 20 mmHg and ICP less than 20 mmHg were allowed to raise PaCO2 to greater than 45 mmHg
in order to improve PbO2.
A recent guideline on the management of PbO2 and ICP in patients with severe TBI allows PaCO2 elevations to 45 to 50 mmHg as a last option
to improve PbO2 under normal ICP conditions.
In patients with poorly graded SAH and late-onset cerebral ischemia, short-lived and controlled elevations of PaCO2 improve CBF and PbO2
.
However, PaCO2-guided PbO2 optimization is still an experimental strategy that may have a significant impact on ICP.
Without direct PbO2 and ICP monitors (without which titration-to-endpoint and hazard detection can be difficult), they should generally not be attempted
.

3.
4 Cerebral embolism

Brain metabolism is a major determinant
of oxygen consumption in the brain.
CMRO2 is usually tightly coupled to cerebral blood flow, and lowering CMRO2 reduces cerebral blood flow and intracranial pressure
.
If systemic MAP is maintained, CPP may be improved and thus oxygen delivery
may be improved.
In practice, the most common method of CMRO2 reduction is sedation, and commonly used drugs include propofol, benzodiazepines, and barbiturates
.
CMRO2 can also decrease
as the temperature decreases.
For every 1°C reduction in temperature, CMRO2 and CBF decrease by 7%, while at 27°C, CBF is almost halved
.
In the ongoing BOOST-III trial, lowering the temperature to 35-36°C or 32-35°C will be used as a higher-level measure in patients with elevated ICP and low PbO2 after conservative treatment (NCT03754114).

Therefore, in clinical practice, when ICP is also high, controlled hypothermia can be considered to improve PbO2, but it should be recognized that this method has not been widely validated
.

Fourth, optimize oxygen delivery and control intracranial pressure

In clinical practice, PbO2 optimization is not an isolated end goal, but part of a comprehensive monitoring plan, and other clinical goals must be addressed at the same time, the most important of which is the goal ICP
.
Three major RCTs are currently ongoing, and whether PbO2 and ICP combined improve outcomes compared with ICP-guided treatment alone (NCT03754114, NCT02754063, CTG1718-05).

However, before these results are available, a practical algorithm for optimizing both PbO2 and ICP has been proposed
.
The algorithm provides a step-by-step approach
to implementing various treatments for different levels of PbO2 and ICP.
Practical recommendations based on this algorithm and other relevant physiological principles are summarized in Figure 2, with the caveat that these recommendations are specific to TBI and may not be applicable to other acute brain injury states
.

Figure 2: As described herein, oxygen consumption or oxygen supply
can be optimized with respiratory, neurological, cardiac, and general critical care interventions.
ABI, acute brain injury; CPP, cerebral perfusion pressure; ICP, intracranial pressure; MAP, mean arterial pressure; PaCO2, arterial blood carbon dioxide partial pressure; PaO2, arterial partial pressure of oxygen; PbO2, brain tissue oxygen; PEEP, positive end-expiratory pressure; TBI, traumatic brain injury

V.
Conclusion

The premise of oxygen monitoring in neurocritical care is simple: if oxygen delivery can be closely tracked, it may reduce secondary brain injury due to hypoxia or hyperoxia, although this approach has significant limitations
.
Treatment with improved oxygenation is not without risks: CPP enhancement may additionally add fluids and vasopressor drugs; Blood transfusions carry a risk of infection and inflammation; Invasive oxygen monitoring devices (eg, PbO2, jugular venous oximetry) require additional techniques and troubleshooting techniques that may not be available
in many cases.
In addition, increasing oxygen delivery often does not address potential limitations in oxygen diffusion and utilization that may prevent injured brains from using oxygen
.
On the other hand, the current oxygen supply measures in neurocritical care are mainly derived from small physiological studies and retrospective observational data
with confounding outcomes.
This heterogeneity may reflect important differences in oxygen metabolism and utilization between different diseases, so "one-size-fits-all" may not be appropriate
.
Therefore, until more reliable data are available, the strategies discussed here to improve oxygen delivery should be used with their limitations in consideration and higher quality recommendations
should not be abandoned.