div epub:type=”chapter” role=”doc-chapter”>
10. Brain Injury with Increased Intracranial Pressure
Mechanical ventilation is frequently applied in the daily management of brain injury patients. The purposes of mechanical ventilation are to protect the airway from the risk of aspiration and to prevent both hypoxemia and hypercapnia, which are two probable causes of secondary brain damage. There are concerns about the potential adverse effects of mechanical ventilation to the brain; for instance, an elevated airway pressure (Paw) may impede the venous return of the brain, which could increase cerebral blood volume (CBV) and intracranial pressure (ICP), especially in the case that a high positive end-expiratory pressure (PEEP) is applied to improve oxygenation. However, recent data suggest that it is not always the case, ICP can increase, unchanged, or even decreased [1–5]. The impact of PEEP on the brain is complicated and there are several factors involved in the brain–lung interaction including respiratory mechanics [2, 4, 5], lung recruitability and PaCO2 , baseline ICP or cerebral compliance [7, 8], etc.
In addition, although the impact of tidal volume setting in patients with lung injury has been extensively investigated, showing that low tidal volume (e.g., the lung-protective ventilation strategy) can reduce ventilator-induced lung injury (VILI) and thus mortality [9, 10]. It is still unclear whether it should also be applied in brain injury patients. Although some ventilation targets had been provided in the recent guideline, it did not detail the setting of tidal volume or PEEP in brain injury patients . Moreover, impaired consciousness in brain injury patients makes the decisions of both weaning from mechanical ventilation and extubation challenging issues. In recent years, some new tools to predict successful weaning have been proposed [12–16]; however, no clear recommendations are currently available owing to the lack of robust evidence in the literature.
In this chapter, the (patho) physiology basis of brain-lung interaction will be discussed, including the impact of positive pressure ventilation on hemodynamics, PaCO2 and its physiological effects on cerebral vessels, and the role of the intracranial Starling resistor and its anatomic basis. The data available in the neurocritical care field regarding respiratory management, from the early phase of mechanical ventilation to the weaning and extubation will be reviewed.
10.2 Physiology and Pathophysiology
10.2.1 Monro–Kellie Doctrine
It was in the seventeenth century when Edinburgh physician Alexander Monro and his student George Kellie first proposed the “science” of ICP, that the skull as a rigid structure containing incompressible brain, and the volume of blood must remain constant unless: “water or other matter is effused or secreted from the blood vessels” in which case “a quantity of blood, equal in bulk to the effused matter will be pressed out of the cranium” .
This doctrine is usually summarized in this way today: with an intact skull, the sum of the volume of intracranial contents (i.e., brain tissue, blood, and cerebrospinal fluid) is constant, an increase in one (e.g., edema, tumor, hematoma, hydrocephalus, etc.) causes a “shifting” in one or both of the remaining two (i.e., some degree of compensation). In this compensated state, the increase in intracranial volume is offset by shifting venous blood out of the intracranial space and cerebrospinal fluid into the spinal subarachnoid space. However, once the capacity to displace cerebrospinal fluid and blood is exhausted, additional increases in any of the intracranial contents are associated with precipitous increases in ICP.
Allied to the Monro-Kellie doctrine is the intracranial compliance that describes the capacity of the intracranial contents to compensate for volume changes. Similar to the calculation of respiratory compliance, it can be calculated as the ratio between the change in ICP (∆ICP) to the change in intracranial volume (∆V). A high compliance represents a good reserve to increase in intracranial volume; in the contrast, compromised intracranial compliance means the corresponding change in ICP to a given increase in intracranial volume becomes greater. Hence, when compliance is reduced, even minor changes in intracranial volume, which can be caused by mechanical ventilation (e.g., increase PEEP, altered tidal volume or respiratory rate that increases PaCO2, etc.; see below for detail), can lead to a rapid increase in ICP.
10.2.2 Transmission of Pressure
Right atrial is the end of venous system and the right atrial pressure (Pra) is thus the downstream pressure of the whole venous system. The driving force of venous return is the pressure gradient between upstream venous pressure (which is the mean systemic pressure of the circulation (Pms) for the whole body and the cranial venous pressure for the brain) and Pra. In this point of view, venous return would decrease for any cause which increases Pra (e.g., increasing PEEP) while the upstream unchanged . In the simplest analysis, the decreased venous return from the brain would thus increase CBV and ICP. In other words, the increased downstream pressure can somehow be “transmitted” into cranium and thus increase ICP. An elevated Paw during positive pressure ventilation can increase pleural pressure (Ppl) and thus Pra, and eventually ICP. It has been suggested by observational data that neurological patients received lower PEEP levels than non-neurological patients . The concern of the potential impact of PEEP on cerebral venous return and ICP might be the reason why physicians prefer lower PEEP levels.
However, the transmission of elevated Paw into cranium is not always fully efficient: the effectiveness of the transmission of Paw to the pleural space and the intrathoracic veins is determined by the relative compliance of the lung and chest wall ; while the transmission of pressure from the right atrial to the neck and cranial veins can be impeded by the effect of an intracranial “Starling resistor.”
10.2.2.1 Compliance of the Lung and the Chest Wall
It has been reported that respiratory system compliance may help in predicting how PEEP influences ICP [2, 4, 5, 21, 22]. The effects of PEEP may become evident only in patients with both decreased intracranial compliance and normal lung compliance . In other words, lung disease per se could be a protective factor because less compliant lungs poorly transmit the increased pressure to the pleural cavity. Similarly, cerebral hemodynamics and ICP may be minimally influenced by the application of PEEP in patients whose lung compliance is impaired .
In contrast, a large retrospective study reported a significant relationship between PEEP and ICP only in patients with severe lung injury ; however, these data are retrospective and, in some cases, “respiratory system compliance” is termed “lung compliance.” Indeed, a given decrease in respiratory system compliance might reflect impaired lung compliance (e.g., acute respiratory distress syndrome (ARDS), pulmonary fibrosis), or decreased chest wall compliance (e.g., intra-abdominal hypertension, pleural effusion). Studies differentiated compliance of the lung and the chest wall using esophageal manometry demonstrated that the impact of PEEP on ICP was greater in the subjects (patients or pigs) with lower chest wall compliance [4, 5].
10.2.2.2 Starling Resistor
A Starling resistor is a collapsible tube in which the pressure external to the tube exceeds the outflow pressure. The external pressure determines the degree of collapse of the tube, thereby providing a variable resistor. The anatomic basis for the intracranial Starling resistor is the rigid cranium (sealed box that determines the external pressure) and the rigid artery (upstream tube) and superior sagittal sinus (downstream tube), and the intervening collapsible cerebral veins . This phenomenon is supported by both hemodynamic [7, 24, 25] and imaging studies [26, 27]. Luce and collages conducted a series of studies in dogs [7, 25] in which they measured the pressure in cerebral vein and sagittal sinus with an intravascular catheter. An abrupt drop in pressure was found when the catheter was advanced into the sinus from the cerebral vein  suggested the presence of a Starling resistor between the sagittal sinus and the cerebral veins.
In an imaging study using magnetic resonance venography in intracranial hypertension patients, narrowing of the distal part of brain-bridging veins is clearly displayed . This is exactly the behavior of a Starling resistor: the collapse always begins at the distal end of the flexible tube, and thereby generates an increase in the flow velocity at the level of the partially collapsed segment (so-called Venturi effect), results in the loss of the magnetic resonance venography signal (termed “void sign”).
Luce and collages suggested that this “resistor” regulates cerebral venous outflow when sagittal sinus pressure is increased by PEEP . In this setting, an elevated ICP would occlude the resistor, thereby preventing the transmission of Pra into the cranium; thus, increases in airway pressure will not be transmitted and will not increase ICP. However, steadily elevating the airway pressure will progressively increase Pra and this will eventually open the resistor. In this way, when the resistor is open, a direct (venous) communication exists between the thorax and the ICP, and here, elevating airway pressure would effectively elevate the ICP.
10.2.3 Impact of Mechanical Ventilation on the Arterial Side
10.2.3.1 Cerebral Autoregulation
Cerebral autoregulation is the intrinsic ability of the cerebral vascular to maintain cerebral blood flow (CBF) constant to a fluctuation of blood pressure [28, 29]. Vasoconstriction or dilation of the cerebral vascular in response to elevated or decreased blood pressure helps to maintain a constant CBF across a wide range of systemic arterial pressures (50–160 mmHg) in healthy individuals . However, in a variety of acute neurologic disorders, cerebral autoregulation is impaired or absent. In this case, CBF may simply change parallelly to cerebral perfusion pressure (CPP, defined as the difference between mean arterial pressure and ICP), and results in either inadequate tissue perfusion (in hypotension) or hyperemia (in hypertension). In a recent review, a conceptual framework of the integrated regulation of brain perfusion has been proposed, suggesting that multiple factors rather than blood pressure alone are involved in the regulation of CBF including sympathetic activity, renin–angiotensin action, cardiac output, blood pressure, metabolic products, nitric oxide, etc. .
Elevated Ppl induced by positive pressure ventilation would not only (to some extend) impede the venous return of the brain, but also the global venous return, which determines preload and thus cardiac output. If a high Ppl reduces systemic arterial pressure, CPP decreases. In the case that cerebral autoregulation is intact, CBF may be maintained despite a lower CPP; but if impaired, decreased CPP may lower CBF and CBV, and thereby decrease ICP. In a recent experimental study, the authors reported that ICP is reduced by gradual increases of PEEP, probably due to the lowered CPP . It is worth pointing out that although ICP can be decreased by elevated Paw, it is actually in the price of decreased CBF. If hypoperfusion/ischemia is sustained, secondary brain injury is inevitable, and the outcome would be poor.
10.2.3.2 Impact of CO2 and pH
PaCO2 is a powerful physiological modulator of ICP via intracranial vasodilation. The relationship between PaCO2 and CBF is an asymmetric sigmoid curve that is approximately linear within a physiological range of PaCO2 . Hypercapnia increases CBF and CBV, leading to an increase in ICP, whereas hypocapnia results in vasoconstriction and decrease ICP. While deliberate hyperventilation is applied to reduce elevated intracranial hypertension, the vasoconstrictive effects of hypocapnia may result in cerebral hypoxia and ischemia. Adaptation can occur and the vasoconstrictive effects dissipate within hours [32, 33], but it seems clear that the effects of hypocapnia on intracranial blood volume are even more short-lived than its effects on CBF (possibly due to differential effects of hypocapnia on intracranial resistance and capacitance vessels) . In addition, rebound elevation of ICP may occur once hyperventilation is discontinued . Standard guidelines on the management of traumatic brain injury recommend hyperventilation as a temporizing measure only for the emergency reduction of acute neurologic deterioration (e.g., herniation, sudden ICP elevation, etc) only; in addition, prolonged “prophylactic” hyperventilation is no longer recommended .
A contrasting concern is that hypercapnia may increase ICP, especially in those who are receiving lung-protective ventilation involving permissive hypercapnia. Petridis et al. reported despite hypercapnia (PaCO2 50–60 mmHg) there was no increase in ICP in patients with subarachnoid hemorrhage who were ventilated with lung-protective ventilation . However, the dominant cause of intracranial hypertension in many patients with subarachnoid hemorrhage is hydrocephalus, and once this is treated with CSF diversion, the intracranial compliance may be much less of a problem; indeed, in the series from Petridis et al., the ICP was 10–18 mmHg following CSF drainage . Consequently, these results cannot be safely extrapolated to patients in whom the primary cause of intracranial hypertension is cerebral edema or a space-occupying lesion. Another study suggested that ICP increased significantly only in patients with non-recruitable lung where PEEP was associated with increased PaCO2 due to alveolar overdistension .
Collectively, the impact of positive pressure ventilation on CBV and ICP is determined by multiple factors . In brief, an increment in Paw will transmit to pleural cavity and intrathoracic major vessels, elevating both Ppl and Pra, where the effectiveness of the transmission is determined by the relative compliance of the lung and the chest wall. The elevated Pra impedes the venous return from both cerebral veins, which decreases cerebral venous outflow and results in an increase in CBV (Starling resistor involved here) and the whole body, which reduces cardiac preload and output. Reduced cardiac output (and blood pressure) may result in a decrease in CBV due to decreased cerebral artery inflow, and this is regulated by cerebral autoregulation. An increment in Paw may also result in hypocapnia, which can cause cerebral vasoconstriction and reduce CBV. Consequently, the net impact on ICP reflects the balance of changes in CBV caused by altered (venous) outflow and (arterial) inflow, and here, the intracranial compliance is another key determinant.
10.3 Management of Mechanical Ventilation
In the early phase following brain injury, hypoxemia and hyper/hypocapnia lead to secondary brain insults, which alter the outcome . Traditional ventilation strategy in brain injury patients, especially those with intracranial hypertension, includes airway protection (intubation), optimization of brain oxygen delivery, strict control of PaCO2, and minimizing the postulated adverse effects of positive pressure ventilation on ICP. Such a brain-centered ventilation strategy lead to the use of larger tidal volumes, high inspired O2, and low or zero PEEP .
Treatment of hypoxemia can be modulated via increasing the inspired oxygen fraction (FiO2) to ensure a PaO2 >60 mmHg . Moreover, PaO2 could be further modulated if brain hypoxia is diagnosed with low tissue oxygen tension (PtiO2) and jugular venous oxygen saturation (SvjO2) . However, it has also been shown that a supranormal PaO2 level may aggravate secondary brain damage after both severe traumatic brain injury  and cardiac arrest . In addition, the optimal setting of tidal volume and PEEP is still controversial, and the recent guideline did not provide any recommendation on this . Nonetheless, no consensus is available to set the tidal volume and PEEP, and in daily practice, one should always assess the risk/benefit relationship before altering any mechanical ventilation setting.
10.3.1 Tidal Volume
The effectiveness of lung-protective ventilation strategy, including lower tidal volume to attenuate VILI, limiting plateau pressure, and using PEEP to limit atelectasis [9, 40], has been well-established in the treatment of ARDS patients. Such ventilation strategy even also has shown benefits in patients without lung injury (but non-brain injury) .
Recently, the safety and efficacy of lung-protective ventilation in brain injury patients were tested in two studies [42, 43]. One study included 499 patients with TBI, SAH, and stroke, and tidal volume was set to 6–8 mL/kg predicted body weight combined with a PEEP level of >3 cmH2O . A significant increase in the number of ventilatory-free days was observed . In the other study that included 749 brain injury patients, tidal volume was set to ≤7 mL/kg of predicted body weight and PEEP was set between 6 and 8 cmH2O. No differences in ventilatory-free days were observed . Nonetheless, in both studies [42, 43] lung-protective ventilation did not lead to clinically relevant effects on ICP. It also should be noted that the level of PaCO2 was carefully monitored and manipulated within normal ranges in both studies. These data highly suggest that protective ventilation with a moderate tidal volume (6–8 mL/kg predicted body weight) could be safely applied in brain injury patients, with the careful monitoring and manipulation of PaCO2 within normal range by adjusting the respiratory rate.
Impacts of positive end-expiratory pressure on intracranial pressure in patients with brain injury
PEEP increases ICP
Neurosurgery (n = 25)
0 vs. 10 PEEP
Increase in ICP was only observed in the patients who manifested decreased cerebral compliance.
TBI (n = 16), SAH (n = 2)
PEEP increased ICP and decreased CPP in patients with abnormal cerebral compliance and normal lung compliance. No impact in patients with normal cerebral compliance or those who with both decreased cerebral and lung compliance.
Neurosurgery (n = 30)
5 vs 15 PEEP
Patients with greater increment in ICP had lower chest wall compliance.
TBI (n = 33)
Baseline vs. additional 5–15 PEEP
PEEP increases ICP, but patients with elevated baseline ICP experienced no significant increase.
Neurosurgery (n = 24)
Between different PEEP levels
PEEP increased ICP.
ICH (n = 25)
Between different PEEP levels (from 0 to 14 cm H2O)
PEEP increased ICP. No impact on CPP.
Neurosurgery (n = 10)
PEEP increased ICP but without intracranial hypertension. In patients with increased ICP, the combination of head flexion and rotation with institution of PEEP caused a dangerous increase in ICP.
TBI (n = 10)
Between different Paw levels
Elevated Paw increased ICP; increasing peak Paw increased variability of ICP, CPP and VmMCA
TBI & ARDS (n = 20)
ICP, VmMCA, and SjO2 remained constant in recruiters but increased in non-recruiters.
Neurosurgery (n = 18)
In patients with normal ICP, PEEP increased ICP. In patients with increased baseline ICP, PEEP did not change ICP or CPP.
SAH (n = 10)
Baseline to 20 PEEP
Stepwise elevation of PEEP increased ICP; decreased rCBF and PtiO2.
TBI (n = 12)
0 vs. 4–8 PEEP
In six patients PEEP increased ICP. CPP was less than 50 mmHg in six patients given PEEP.
Neurosurgery (n = 20)
PEEP increased ICP. No impact on CPP.
No impact on ICP
TBI, SAH (n = 21)
PEEP reduces CPP and VmMCA only in patients with normal respiratory system compliance. No impact on ICP or SjO2.
Coma (n = 7)
0 to up to 40 PEEP
PEEP did not increase ICP in patients with either normal or low intracranial compliance and did not increase ICP in the absence of pulmonary disease.
TBI & ARDS (n = 7)
Standard ventilation vs. mid to end expiratory tracheal gas insufflation
No changes in hemodynamic or cerebral parameters
TBI & ARDS (n = 20)
PEEP increases PtiO2, no impact on ICP or CPP
Pedeatric brain tumor (n = 21)
No change in ICP, CPP or VmMCA
ICH (n = 39)
Baseline vs. 15 PEEP
PEEP had no adverse effect on CPP and led only to clinical insignificant increase in ICP.
TBI & ARDS (n = 9)
RMs using different PEEP levels
No impact on ICP and CPP 5 min after RMs
Neurosurgery (n = 20)
CPP significantly changed depending on the various PEEP levels. Three distinct reaction patterns of ICP and other parameters observed