Chapter 36: Mechanical Ventilation

Regardless of the indication for endotracheal intubation and initiation of mechanical ventilation, a mechanical ventilator, in and of itself, does not treat much. However, it does have great potential for harm. Effective mechanical ventilation should prioritize limitation of ventilator-induced lung injury (VILI) almost as much as oxygenation and maintenance of normal carbon dioxide levels should take less priority in the majority of patients.¹ Keen knowledge of basic and advanced principles will allow the clinician to use the mechanical ventilator for its maximum therapeutic potential and minimize iatrogenic injury while the process that led to respiratory failure is allowed to heal. It cannot be stressed enough that knowledge of cardiopulmonary physiology, heart-lung interactions, and the effects of mechanical ventilator settings on cerebral oxygenation and blood flow will allow the clinician flexibility when titrating support to an individual patient at the bedside. While the clinician should be intimately familiar with guidelines and large clinical trials, they are only as good as the response a particular patient has to a chosen therapy.

Mechanical ventilation is instituted for a number of reasons (Table 36-1).² Most commonly, these indications are a combination of a failure to adequately oxygenate, ventilate, or meet the metabolic demands of a physiologically stressed patient. Type I respiratory failure is hypoxemic respiratory failure, defined as a partial pressure of oxygen in arterial blood (Pao₂) of less than 60 mm Hg. Type II respiratory failure is hypercarbic respiratory failure, defined as a partial pressure of carbon dioxide in arterial blood (Paco₂) greater than 50 mm Hg.³ Mechanical ventilation may also be instituted to maintain normal blood pH, to decrease work of breathing, to assist left ventricular function in the setting of acute decompensated heart failure, or for airway protection in the setting of toxic overdose, traumatic brain injury, or any other significant acute central nervous system illnesses.

The breath generated by a mechanical ventilator can be separated into four phases ^2,4 (Table 36-2):

• Triggering: Triggering represents the change from expiration to inspiration and occurs either because of a drop in circuit pressure or flow (when patient triggered) or because of elapsed time. Sensitivity refers to a preset threshold of pressure or flow. When this is reached, a mechanical breath is delivered. This can be manipulated by the clinician, and is usually set at −1 to −2 cm H₂O. In the setting of pressure triggering, airway pressure is reduced (by patient effort) in the proximal circuit, the expiratory valve closes, and the patient receives a breath. In the setting of flow triggering, the patient's inspiratory effort induces a disruption of the constant flow in the ventilator inspiratory circuit. This change in flow signals the expiratory valve to close and the ventilator to deliver the next breath.

• Inspiration: Pressurized gas is channeled from the ventilator to the patient after the exhalation valve closes. This phase can be controlled (or targeted, see below) based on volume or pressure. For example, AC/VC is volume control/targeted and AC/PC is pressure control/targeted. Choice of control variable is largely the discretion of the clinician, as either can be manipulated to achieve set goals. It should be noted, though, that in volume-targeted ventilation, excessive airway pressures can arise secondary to worsening pulmonary mechanics. In this situation, the pressure alarm will cause a pressure limit to cycle to expiration. In pressure-targeted ventilation, minute ventilation is not guaranteed. It is a function of the compliance and resistance of the respiratory system. The clinician, therefore, should monitor these physiologic changes closely to avoid untoward changes in either airway pressure or Paco₂ levels.

• The pattern of inspiratory flow can be square, decelerating, or sinusoidal (Figure 36-1). Square inspiratory flow pattern results in a rapid rise to a preset level (set by the clinician) followed by constant flow until cycling occurs. Decelerating flow results in a rapid rise to a maximum level followed by a gradual decrease until cycling. This is secondary to the decrease in pressure difference (Δ pressure) between the ventilator and patient as inspiration progresses during pressure-targeted ventilation. All pressure-targeted breaths are of the decelerating flow pattern. Sinusoidal flow pattern most closely represents normal physiologic breathing. It results in flow that gradually increases, then decreases during inspiration. The choice of inspiratory flow pattern should be based on patient characteristics, and a few common clinical scenarios should be familiar to the clinician. Square flow over time results in a shorter inspiratory time (I time) and, therefore, longer expiratory time (E time). For this reason it can be preferred in patients with obstructive physiology (chronic obstructive pulmonary disease [COPD] or asthma).^5,6 It may also be tolerated better in patients with demand for high minute ventilation, such as those with severe metabolic acidosis or elevated intracranial pressure (ICP). In this situation, there is potential for dyssynchrony to occur during the progression of the inspiratory phase if decelerating flow is chosen. The tradeoff is higher peak airway pressures with a square wave form. Decelerating flow results in a longer I time and a better distribution of flow. With pulmonary pathophysiology involving heterogenous distribution of injury (acute lung injury as the prototype example), decelerating flow is likely the best choice. The length of inspiration depends on several factors. In pressure-targeted modes of ventilation, the clinician directly controls the I time and I-to-E ratio. In volume-targeted ventilation, the I time is a function of set tidal volume (V_T) and flow rate (V_T/inspiratory flow rate).

• Cycling: This is the transition from inspiration to expiration. It can occur because of a decrease in flow, elapsed time, or delivered volume. Flow-cycled breaths are pressure-support (PS) breaths. When a predetermined decrement in flow occurs, inspiration is terminated and the patient cycles to expiration. This is typically set at 25% of initial flow, but can be manipulated based on different clinical conditions. For example, if a large leak occurs (bronchopleural fistula or ventilator circuit leak), it may be wise to set the cycle at a higher percentage of initial flow to terminate inspiration at the appropriate time. It should be noted that for safety reasons, PS can be cycled because of excessive pressures or elapsed time. In time-cycled breaths, inspiration is terminated after a preset interval, regardless of whether preset pressure has been achieved. Pressure-targeted ventilation is time cycled. In volume-cycled breaths, the clinician has selected the targeted tidal volume. Volume is delivered until that volume is reached. Airway pressure is directly dependent on the tidal volume and the mechanical characteristics of the patient's respiratory system. For safety reasons, volume-cycled breaths will be pressure cycled if airway pressures exceed the pressure alarm limit. In this situation, the delivered tidal volume will not reach the set volume target unless the pressure alarm limit is increased (the clinician should exercise caution with this strategy).

• Expiration: Flow from the ventilator is stopped, the exhalation valve opens, and gas is allowed out from the lungs. This occurs passively. Inspiratory flow should cease prior to expiration. If this does not occur, the patient will develop auto-PEEP or intrinsic PEEP (iPEEP). This is most commonly seen in patients with expiratory flow limitation (COPD or asthma) or higher I-to-E ratios (see below).

A ventilator mode describes the set of breath characteristics delivered to the patient and patient- ventilator interactions, and includes such things as control variable/target, phase variables, and mandatory versus spontaneous breaths.^2,7 Two general approaches exist and involve either setting volume/flow as the control variable or pressure. From there, a variety of specific and complicated combinations can be derived. The choice of ventilator mode is largely at the discretion of the clinician. Pressure targeted and volume/flow targeted have more in common than not, can achieve similar frequency and minute ventilation goals, and can equally be employed to limit VILI. For this reason, clinician familiarity and unit and institutional practice patterns determine to a large extent the mode that is employed. Also, data showing definitive improvement in clinically relevant outcomes when comparing one mode with another are lacking. Patient physiology and knowledge of mode differences will allow greater flexibility to titrate safe mechanical ventilation based on bedside evaluation, clinical scenario, and response to therapy. When confronted with a difficult-to-oxygenate patient whose compliance is dynamically worsening, pressure-targeted modes of ventilation may be most effective. Setting a pressure target identifies a priority of minimizing VILI over guaranteed minute ventilation, and the decelerating inspiratory flow pattern also allows for better gas distribution into nonhomogenous injured lung units. Also, the clinician has more direct control of the I-to-E ratio, therefore allowing easier manipulation of these variables to maximize mean airway pressure relative to peak and plateau airway pressures, which is beneficial for oxygenation.

In controlled mechanical ventilation (CMV), the majority of the patient-ventilator interaction that occurs is from the ventilator to the patient, and not vice versa. For example, it is time triggered and is limited and cycled by the ventilator. While this mode of mechanical ventilation is no longer an available selection, this is what is essentially achieved when a patient is not interacting because of heavy sedation or neuromuscular blockade. The term control in this instance should not be confused with the control variable or target, as described below. Assisted breaths can either be volume or pressure targeted (AC/PC or AC/VC, for example). With these types of breaths, the clinician sets the target (tidal volume or pressure) and a respiratory rate. The patient can then trigger the ventilator and receive the set target each time, more frequently than the set rate on the ventilator. In this instance, the ventilator is doing work, specifically on the patient, and therefore assisting the breath after the patient triggers/initiates it. Assisted breaths also encompass intermittent mandatory ventilation, where the ventilator delivers a set number of breaths but the patient can also trigger spontaneous breaths, as well as pressure support ventilation, where the patient triggers all the breaths and the delivered volume is determined by the set pressure, patient effort, and mechanical characteristics of the respiratory system.

Controlled breaths versus spontaneously triggered breaths are undoubtedly different with respect to respiratory muscle function and viability, gas exchange properties, lung expansion, and hemo-dynamics.⁸ A fully controlled breath must fully expand the lungs and chest wall without inspiratory muscle assistance. Spontaneous breathing also serves to maintain venous return by lowering pleural pressure and therefore right atrial pressure.⁹ Diaphragmatic movement is more pronounced ventrally in the supine patient with spontaneous breathing. This favors atelectasis in dorsal lung units and a ventral distribution of ventilation, resulting in increased dead space ventrally and shunt dorsally.¹⁰ In patients with vigorous respiratory muscle activity, a switch to fully controlled ventilation may be beneficial. In this setting, oxygen consumption (Vo₂) can decrease substantially, and by removing expiratory muscle contraction, end-expiratory lung volume can be preserved, which improves oxygenation.⁸ The preservation of spontaneous breathing is of more benefit than harm in the great majority of patients. However, in the setting of very high demand, such as severe, early acute respiratory distress syndrome (ARDS), there may be a role for fully controlled breaths, but this should be titrated to the individual patient.^11,12,13

For a breath to occur, whether it be spontaneous or by positive pressure, a pressure gradient must exist from the airway opening to the alveoli.^2,4,7,14 The volumes delivered and pressures generated largely depend on the mechanical properties of the respiratory system: the lungs and chest wall, as well as the abdomen. Each of these components has mechanical properties that determine the overall behavior of the respiratory system. While the respiratory system can be quite complex, the main variables of interest are really only pressure, flow, and volume. Specifically, the ventilator must generate a pressure to cause flow and therefore increase lung volume.¹⁴ The pressure required to do this reflects a combination of the pressures to inflate the lung and chest wall. This can be illustrated by the equation of motion^14,15:

The equation of motion illustrates several basic principles. A ventilator can only directly control one of the three variables at a time (pressure, volume, or flow). Despite its complexity, a ventilator is simply a machine controlling one of these three variables (control variables).^2,4,7 They are also referred to commonly as the “target” and are the variables that the ventilator holds constant despite changing respiratory system mechanics. For example, in pressure-control ventilation, pressure is the control variable or target and is held constant, while flow and tidal volume will vary based on the patient's respiratory mechanics. Control in this instance should not be confused with a controlled versus assisted breath, as above.

Compliance describes the ability or tendency of the respiratory system to expand in response to a delivered pressure and volume.^14,16 Simplistically, this is Δ volume/Δ pressure, and the compliance of the respiratory system (C_RS) is ΔV/ΔP_alveolar. Elastance is the inverse of compliance, or the ratio of pressure change to volume change, and describes the tendency to recoil. Resistance describes the impedance to airflow through the respiratory system, or the ratio of pressure change to flow change.

A pressure-volume (PV) curve (Figure 36-2) demonstrates key features of respiratory system mechanics.^17,18 The lower (LIP) and upper inflection points (UIP) reveal areas of reduced compliance secondary to low and high volumes, respectively (compliance is volume-dependent). As de-recruited alveoli are opened up, the lower inflection point transitions into the steeper (most compliant) portion of the PV curve. This is theoretically the physiologically most sound and safe portion to ventilate the patient, though where this lies in an individual patient can be difficult to elucidate. At higher tidal volumes, the upper inflection point represents a transition to overdistention, risk for barotrauma, and reduced compliance. This graphically represents a foundation for the concept of PEEP setting and VILI (discussed below).¹⁹ The PV curve also demonstrates hysteresis: greater compliance during expiration versus inspiration. This reflects the fact that recruited alveoli are more compliant than recruiting or unrecruited alveoli and provides the foundation for the concept of recruitment maneuvers in PEEP setting, as well as advanced modes of mechanical ventilation, such as airway pressure release ventilation (APRV) and high-frequency oscillatory ventilation (HFOV). The normal lung and chest wall are remarkably compliant (200 mL/cm H₂O) and the normal compliance of the respiratory system is around 100 mL/cm H₂O (lung compliance and chest wall compliance added in series).¹⁴ The severely injured lung may have compliance as low as 20 mL/cm H₂O or even lower. Unfortunately, C_RS that is displayed by the mechanical ventilator has to be interpreted with caution. This number represents a weighted average of many million alveolar units, and in pathophysiologic states exhibiting heterogenous injury distribution (acute lung injury), compliance, volume distribution, and potential for VILI vary greatly.^20,21 Resistance to airflow is determined by artificial and natural airways, as well as the rate and pattern of flow. Similar to compliance, resistance varies with volume, but also flow, and differs between inspiration and expiration.

The three most common airway pressures displayed for monitoring on a mechanical ventilator are peak airway pressure (P_peak), inspiratory plateau pressure (P_PL), and mean airway pressure (P_mean)¹⁵ (Figure 36-3). P_peak reflects C_RS as well as resistance. Elevated P_peak can lead to barotrauma, increased pulmonary vascular resistance, and right heart dysfunction, as well as decreased venous return.²² Under no-flow conditions (inspiratory hold maneuver on the ventilator), resistance is eliminated, and this pressure reflects P_PL. The distending force of the lung is the transalveolar pressure, or alveolar pressure minus pleural pressure.^14,23 In the setting of elevated pleural pressure, such as elevated intra-abdominal pressure or reduced chest wall compliance, transalveolar pressure may be quite low; therefore, P_PL in the “safe” range may lead to significant underventilation.²³ Currently, without the ability to directly measure pleural pressure, the P_PL is the best reflection of transalveolar pressure and the distending pressure of an alveolus, and is an important reflection of potential for VILI. Of note, pleural pressure can be measured indirectly, via an esophageal balloon, and this pressure can be used in the calculation of transalveolar pressure.²⁴ P_mean is the average airway pressure over the respiratory cycle. Elevating the mean airway pressure should be viewed as a strategy to improve recruitment and oxygenation, with the goal being elevation to a greater extent relative to P_peak and P_PL.²⁵ If a patient has an elevated P_peak relative to P_PL, the problem lies in increased resistance, such as bronchoconstriction or endotracheal tube blockage or kink. If both P_peak and P_PL are elevated, then a compliance problem exists, such as worsening lung injury, pneumothorax, or elevated intra-abdominal pressures.

Knowledge of the general concepts of mechanical ventilation is of little use without knowledge of the patient at the bedside and defining the goals of therapy for that individual patient. With all critically ill patients, goals can be defined as oxygenation, limitation of VILI (perhaps just underneath oxygenation in order of importance), and ventilation. However, in the neurologically injured patient, we must add limitation of secondary brain injury.²⁶ This includes hypoxia, hypotension, prolonged periods of hypercapnia or hypocapnia, and elevated ICP. Manipulation of the mechanical ventilator can directly affect all of these.

Mechanical ventilation is common in the management of the neurologically injured patient.^27,28,29 Unfortunately, the direct generalization of many high-quality clinical trials is difficult in this patient population, as TBI is often an exclusion to trial enrollment.³⁰ To add to the complexity, protective ventilation strategies and the management of TBI have often been viewed as having competing and discordant therapeutic targets, such as permissive hypercapnia and PEEP setting to ameliorate VILI.²⁷ To this end, the ventilator must be adjusted to (1) improve brain oxygenation and (2) improve cerebral blood flow. At a very minimum, it must be adjusted to not harm cerebral oxygenation and blood flow. They are intimately linked, and inevitable overlap exists, but they will be discussed separately to provide a balanced approach with seemingly competing and potentially conflicting therapeutic strategies in the neurologically injured patient requiring evolving mechanical ventilation requirements.

Oxygenation should be the primary goal in any mechanically ventilated patient, especially the neurologically injured. Under normal conditions, alveolar ventilation is nearly equal to cardiac output, producing an overall ventilation-to-perfusion (V̇/Q̇) ratio close to unity.³¹ While the exact causes of hypoxemia are innumerable, only a few physiologic mechanisms explain arterial hypoxemia (Table 36-3).³ In the critically ill, mismatch between alveolar ventilation and perfusion and right-to-left shunt are by far the most common mechanisms. Diffusion is rarely the cause of hypoxia, unless severely impaired.

Oxygenation can be assessed several ways.³² A noninvasive pulse oximeter allows assessment of arterial oxygen saturation (Sao₂). Arterial blood gas analysis gives us the Pao₂. Perhaps greater indicators for an oxygenation problem, calculation of the alveolar-arterial oxygenation (A-a) gradient, Pao₂-to-Fio₂ ratio, or oxygenation index (OI) gives the clinician greater insight into the extent of oxygenation failure.

Setting Oxygenation Goals

Normal Sao₂ and Pao₂ are greater than 95% and greater than 80 mm Hg, respectively. In the critically ill patient, these goals can sometimes be unattainable. Setting oxygenation goals in the critically ill can be largely empiric, but patient factors and available data should be considered. The majority of high-quality ARDS trials have allowed for an Sao₂ of 88% and a Pao₂ of 55 mm Hg.^{30,33,34,35,36} It should be noted that at these levels, the patient is most certainly on the steep portion of the oxyhemoglobin dissociation curve and is in danger of a quick decline in arterial oxygenation with any decline in cardiopulmonary physiology, circuit disruption, or daily patient care activities. Hypoxemia not only will decrease cerebral oxygen delivery but also will result in cerebral vasodilation, therefore raising ICP. The Brain Trauma Foundation recommends keeping arterial oxygenation above a Pao₂ of 60 mm Hg. Other guidelines, including the UK transfer guidelines for brain injury, recommend even higher oxygenation targets (97.5 mm Hg).^37,38 Given the well-known association of hypoxemia with worse outcome^{26,39,40,41,42} as well as the consideration of guideline recommendations, the neurologically injured patient should have these goals raised. Hypoxia should be avoided at all costs, and the ARDS Network (ARDSNet) targets are likely too low in this patient population. A target of a Pao₂ of 60% and an Sao₂ of 92% at a minimum is likely reasonable, and these goals should be revisited based on individual patient physiology.

Fraction of Inspired Oxygen

Increasing the Fio₂ is perhaps the simplest and quickest way to improve arterial oxygenation and should be done immediately in the setting of life-threatening hypoxia. Unfortunately, simply increasing Fio₂ may be the least physiologic way to improve oxygenation from a pulmonary mechanics standpoint, and prolonged administration of high levels of oxygen has consequences.

Lung tissue is exposed to the highest concentrations of oxygen in the body. In animal and human studies, high fractions of inspired oxygen can cause lung injury, ranging from mild subjective symptoms to diffuse alveolar damage histologically indistinguishable from ARDS.^43,44 Exposure to elevated Fio₂ can reduce lung volumes secondary to absorptive atelectasis, increase intrapulmonary right-to-left shunt, and directly injure pulmonary parenchyma. The onset and rate of development are likely directly related to level of oxygen administration as well as duration.

Hypoxia is known to cause secondary brain injury, and the delivery of Fio₂ 1.0 has been shown to increase brain tissue oxygenation.⁴⁵ Typically, the partial pressure of brain oxygen tension (Pbto₂) measured by a brain tissue oxygen probe improves almost immediately when Fio₂ is increased.⁴⁵ The administration of high levels of oxygen should be cautioned against, however. Despite some studies demonstrating reduced lactate levels and reduced lactate-to-pyruvate ratios, other studies have not reproduced these findings, and no good clinical studies demonstrate benefit of normobaric hyperoxia.^46,47,48,49 In addition, systemic hyperoxia has not been shown to improve cerebral metabolic rate or functional outcome.⁵⁰ Given these negative findings, as well as the well-known toxicity of hyperoxia, administration of prolonged levels of increased Fio₂ should be discouraged in favor of better ventilator management to maximize recruitment, PEEP effects, and the compliant portion of the PV curve. When Fio₂ exceeds 0.60, the clinician should manipulate the ventilator settings to improve oxygenation.

Positive End-Expiratory Pressure

Positive end-expiratory pressure (PEEP) maintains airway pressure above atmospheric pressure throughout the respiratory cycle by pressurization of the ventilator circuit. A full review of the topic is beyond the scope of this chapter, as the published literature on PEEP is extensive. However, the clinician should be aware of the pathophysiology pertinent to PEEP setting, advantages and side effects, approaches to PEEP setting, and the use of PEEP in the neurologically injured patient.

In patients with hypoxia due to atelectasis, alveolar edema, and/or volume loss (which causes most type I respiratory failure), PEEP serves to restore functional residual capacity (FRC), reduce intrapulmonary shunt, shift ventilation to a more compliant portion of the PV curve, and prevent end-expiratory volume loss (derecruitment).⁵¹ PEEP recruits previously nonaerated lung tissue and homogenizes regional distribution of tidal ventilation. The net effect on gas exchange reflects the balance between recruitment and overdistention. In the setting of obstructive physiology or expiratory flow limitation, PEEP serves less to improve oxygenation and more to improve patient-ventilator synchrony and triggering (discussed below).^5,6,52

PEEP is not without side effects. Given the heterogenous distribution of lung injury in acute lung injury (ALI)/ARDS, PEEP can overdistend more compliant lung units, contributing to VILI.⁵³ If PEEP leads to overdistention, it can augment dead space, increase pulmonary vascular resistance, and cause right heart dysfunction. It can also decrease venous return and, in the setting of volume depletion, decrease cardiac output.

The optimal way to set PEEP is widely debated and controversial, as is the optimal level of PEEP to use.^{34-36,54,55,56,57,58} The literature reveals that PEEP has been set to maximal compliance, a level just exceeding the lower inflection point on a PV curve, the lowest PEEP to achieve adequate oxygenation, or a PEEP-Fio₂ combination table.^34,56,59,60 There is no widely accepted method to enhance clinical outcome. Before implementing any PEEP-setting strategy, an assessment of the risks and benefits and, most important, the potential to recruit alveolar units should be ascertained. Patients with primary pulmonary disorders, such as pneumonia or pulmonary contusions, usually have lung units with much less recruitability compared to patients with an extrapulmonary etiology of hypoxia, such as sepsis.⁶¹ Villar et al set PEEP at 2 cm H₂O above the lower inflection point and demonstrated a mortality decrease from 53.3% to 32%.⁵⁶ Using higher PEEP, Ranieri demonstrated attenuation in cytokine response, suggesting decreased inflammatory signaling with higher PEEP.⁵⁸ These studies were done with simultaneous decreases in tidal volumes, however. The ALVEOLI trial was done to determine the isolated benefit of higher levels of PEEP in patients with ARDS. Despite improvement in oxygenation, the higher PEEP group showed no improvement in clinically relevant outcomes.³⁴ In two recently published trials, high PEEP was set based on a “lung open ventilation” protocol along with recruitment maneuvers and adjusted as high as possible to not exceed a P_PL of 28 to 30 cm H₂O. Both of these trials demonstrated improved oxygenation, but no mortality benefit.^35,36 Another physiologically attractive alternative is PEEP setting under the guidance of an esophageal balloon. The balloon is placed similarly to a nasogastric tube and is used to estimate pleural pressure in order to calculate transpulmonary pressure. Recent data show an improvement in oxygenation and compliance, but again no improvement in clinically relevant outcomes.²⁴ Data have shown wide variability with respect to lung recruitment and PEEP effect on lung recruitment. As such, individualizing PEEP setting is likely the most prudent approach.

PEEP can be set with or without recruitment maneuvers, with the aim being to open collapsed alveoli that may need higher, sustained inflation pressures to open. Clinical characteristics, such as etiology of ALI, can help assess potential for recruitment, but likely the easiest way to assess potential for alveolar recruitment is with a recruitment maneuver. After recruitment, alveoli are more compliant and easier to keep open. Recruitment maneuvers have been applied as sustained inflation pressures (30-40 seconds at 35-50 cm H₂O), as pressure-controlled breaths at increased PEEP levels, or as intermittent sighs.^{35,55,59,62,63,64,65} They are generally very effective at improving oxygenation; however, they should be followed by an incremental increase in PEEP, or the effects will only be transient. In general, recruitment maneuvers are safe and tolerated well, but should likely be limited to patients early in the course of lung injury (more recruitability) with bilateral lung disease and no evidence of hypovolemia.

PEEP can adversely affect ICP and cerebral perfusion pressure (CPP), and PEEP setting in patients with neurologic injury has been a controversial topic.^26,27 By increasing intrathoracic pressure, and therefore juxtacardiac pressure, right atrial pressure can increase. This can impede venous outflow from the brain and lead to elevated ICP. Also, if PEEP decreases cardiac output, this can lead to hypotension and reduced cerebral blood flow. These effects are usually less significant in patients with reduced lung compliance. Also, if PEEP leads to overdistention, as opposed to alveolar recruitment, an increase in dead space can lead to elevated Paco₂ levels. Elevated ICP and reduced CPP have historically been viewed as (relative) contraindications to the use of increasing levels of PEEP. This has led to the observation of a lower threshold to simply increase the Fio₂ to combat hypoxia in this patient population. Fortunately, the literature demonstrates that PEEP is well tolerated in patients with TBI.^26,27,66,67 Early studies show that PEEP causes only modest, clinically insignificant changes in ICP.^68,69,70,71 In fact, PEEP set lower than ICP typically does not result in an increase in ICP, and if PEEP achieves alveolar recruitment, it can reduce ICP. Even in the setting of normal lung compliance (and therefore greater transmission of pressure to the intravascular space), PEEP has been shown to not significantly increase ICP. Furthermore, if PEEP results in alveolar recruitment, brain oxygenation may improve as a result.⁷² Therefore, traumatic brain injury and elevated ICP should not be considered contraindications to PEEP setting in patients who need alveolar recruitment. The data and physiology dictate that the benefits likely outweigh the risks in the majority of patients.

In summary, PEEP serves to improve alveolar recruitment, decrease shunt, and improve oxygenation, as well as improve triggering in the setting of expiratory flow limitation. It has predictable physiologic effects, but the net hemodynamic effect of PEEP is some combination of volume status, ventricular function, and alveolar recruitment versus overdistention. There are several approaches to PEEP setting, but there is no convincing evidence that one method is superior to another. In the setting of TBI and/or elevated ICP, most data show that PEEP is well tolerated, and therefore should be attempted when necessary. PEEP setting and improving respiratory system mechanics should be favored over increasing Fio₂, as this approach has a higher risk versus benefit profile and is less physiologically sound management. When confronted with a neurologically injured patient with either hypoxemia or unacceptable levels of inspired oxygen, the approach given in Figure 36-4 can be used.

Partial Pressure of Brain Tissue Oxygen Tension

Cerebral monitoring has historically focused on ICP and CPP. However, with the knowledge that cerebral ischemia is associated with worse clinical outcomes, Pbto₂ monitoring allows the clinician continuous monitoring of cerebral oxygenation.⁷³ Despite an ongoing clinical debate regarding Pbto₂ monitoring, data suggest worse outcomes associated with low Pbto₂ levels.⁷⁴ Systemic oxygenation, pulmonary function, and cerebral oxygenation are strongly linked, given the basic physiology of oxygen transport from the lungs to the brain.⁷⁵ ALI is also an independent risk factor for cerebral hypoxia in the setting of TBI.⁷⁶ Given these facts, the care of patients with severe TBI should consider brain-lung interactions and that maneuvers that safely increase systemic oxygenation will likely increase Pbto₂ as well.

If monitored, Pbto₂ values of less than 15 mm Hg should be a concern. The first step in addressing persistently low Pbto₂ values should be the assessment of the invasive monitoring probe, as it may reside in dead tissue. Similarly, catheter fidelity must be maintained and assessed. While the absolute cutoff point of acceptable Pbto₂ is debatable, persistently low values have again been shown to be associated with worse outcomes.⁷⁴ Different threshold values likely exist for different patients and conditions, and other clinical questions remain (such as what defines “acceptable” brain oxygenation). Mechanical ventilator strategies to increase Pbto₂ parallel strategies to increase Pao₂: manipulation of Fio₂, PEEP setting, manipulation of the I-to-E ratio to increase mean airway pressure, advanced mechanical ventilation modes, and rescue therapies such as inhaled nitric oxide.

Skilled management of a mechanical ventilator can influence both venous return and forward blood flow. Unfortunately, when learning mechanical ventilation, most clinicians do not think in terms of blood flow. However, positive pressure ventilation causes predictable physiologic changes, and cardiopulmonary function is intimately linked.^9,77 Cerebral autoregulation maintains cerebral blood flow constant over a mean arterial pressure ranging from around 50 to 150 mm Hg under normal conditions.^78,79 In the setting of traumatic brain injury, this autoregulation may be impaired. Cerebral blood flow can be altered by a change in the venous outflow or arterial inflow as well as changes in mean arterial pressure (and therefore CPP), among other factors (such as blood viscosity, arteriolar resistance, and microcirculatory function). Ventilator management can affect all of these parameters and the effect of ventilator settings on cerebral blood flow is underappreciated. Cerebral blood flow must be maintained in the neurologically injured patient, and with the knowledge of the detrimental effects of hypotension in patients with TBI, every change in the mechanical ventilator must be made with the knowledge of how the change may affect forward flow.^39,40,41 A full review of heart-lung interactions is beyond the scope of this chapter, but the clinician must be aware of some key points.

Venous Blood Compartment

Venous return and right ventricular function

The venous system is a low-pressure reservoir that contains about three-fourths of our total blood volume.^9,80 There is a primacy of preload in determining myocardial performance, and the determinants of venous return have been documented for over 50 years.⁸¹ The mean systemic pressure is the upstream pressure in the venous reservoirs governing venous return, and the right atrial pressure is the backpressure for venous return. The venous drainage of the brain is a system without valves and communicates directly with the right atrium.^78,79 Therefore, venous return and cardiac function are in direct communication with the intracranial compartment. Positive pressure ventilation, by increasing right atrial pressure, can decrease venous return and cardiac output.

The right ventricle's (RV's) main function is to accept venous return and eject the optimum amount of blood into a (usually) highly compliant pulmonary vascular system.^77,80,82,83 This maintains a low right atrial pressure, thereby maximizing venous return and overall myocardial performance. If airway pressure or lung volumes are elevated, pulmonary vascular resistance will increase, inducing RV dysfunction, the most severe manifestation of this being acute cor pulmonale (ACP).⁸⁴ ALI is a syndrome of the alveolar epithelium and capillary endothelium. As such, it is strongly associated with RV function.⁸⁵ RV dysfunction and/or ACP has been shown to be present in up to 35% of patients with ARDS and independently associated with increased mortality rate.^22,86,87 This incidence increases as P_PL increases. Given these findings, RV dysfunction should be suspected in any patient with ALI and ventilator settings should be adapted to unload the right ventricle, especially in the setting of elevated airway pressures.⁸⁸ Also, while not useful to predict preload responsiveness, the value of CVP and waveform morphology (large v wave, signifying tricuspid regurgitation) should both be followed closely, as should echocardiography⁸⁸ (Figure 36-5).

Similar to RV function, the left ventricle (LV) is affected by positive pressure ventilation. If mechanical ventilation induces decreased venous return, this will decrease LV preload several cardiac cycles later.^9,77 By decreasing LV transmural pressure, positive pressure ventilation decreases LV afterload and can therefore improve contractility. This is responsible for the positive effect of positive pressure ventilation in acute decompensated heart failure.

Putting it all together

Respiratory function alters cardiovascular function and cardiovascular function alters respiratory function.^9,77 In turn, they both can alter cerebral blood flow. The right and left ventricles are dependent in series (the left ventricle cannot pump what the right ventricle does not give it) and in parallel (they share a common septum and pericardial space). This is all affected by the mechanical ventilator. A positive pressure breath increases lung volume and increases intrathoracic pressure, which increases juxtacardiac pressure and therefore right atrial pressure. This can decrease venous return and therefore LV preload several cardiac cycles later. Improper use of excessive tidal volumes can elevate airway pressure, increase West lung zone I and zone II conditions, and therefore increase RV afterload.⁸⁵ PEEP, if it serves to recruit alveoli, can lower pulmonary vascular resistance and improve RV function. If it overdistends alveoli, it can increase dead space, increase pulmonary vascular resistance, and induce RV dysfunction.⁸⁵ These deleterious changes can impede venous outflow from the injured brain and further increase ICP.

To minimize the potential deleterious effects of mechanical ventilation on ICP and cerebral blood flow (CBF), secondary to venous return and RV dysfunction, the clinician should employ a low tidal volume ventilation strategy with tidal volumes set at 6 to 8 mL/kg IBW. This will limit the effect of excessive volumes on airway pressures and pulmonary vascular resistance. If this strategy results in permissive hypercapnia, it must be balanced with the effects of hypercapnia on ICP and RV function.⁸⁵ PEEP should be set as previously described. Suspect RV dysfunction in all patients with ALI and follow them with echocardiography. Follow the P_PL closely, and realize that a significant proportion of patients with an “acceptable” P_PL of less than 30 cm H₂O may still have RV dysfunction,²² and continue to have high clinical suspicion of this. To avoid risk of RV dysfunction and/or ACP, if echocardiography is not routinely followed, P_PL should be kept at less than 27 cm H₂O. Realize that an elevated central venous pressure is only present in diseased states, as the primary function of the RV is to keep right atrial pressure low.^9,77 Follow CVP closely when RV dysfunction is suspected. If the patient has RV dysfunction, measures should be taken immediately for treatment, which will likely include some combination of inotropes/inodilators, pressors to maintain mean arterial pressure and CPP, inhaled afterload reducers, and judicious volume management.

Arterial Blood Compartment

The arterial blood flow responsible for cerebral blood volume, and therefore a component of ICP, is also governed by several factors, including blood pressure, viscosity, cerebral metabolic rate, and carbon dioxide and oxygen levels.⁷⁹ Skilled management of a mechanical ventilator can limit the potential for a decrease in mean arterial pressure and CPP, as discussed above. The majority of the following discussion will focus on Co₂ and o₂, however.

Carbon dioxide and cerebral blood flow

The Paco₂ represents a balance between Co₂ production (V̇co₂) and Co₂ elimination (Paco₂ = V̇co₂/V̇_E − V_D, where V̇_E is minute ventilation and V_D is dead space.⁸⁹ In the setting of critical illness, elevated Paco₂ is usually secondary to decreased minute ventilation or increased dead space. While ALI is often viewed as an oxygenation problem, it is dead space that has been shown to directly correlate with mortality in this situation.^31,90 Permissive hypercapnia refers to a deliberate tolerance of elevated Paco₂ levels to limit lung stretch and VILI and is commonly used in a variety of patient populations, such as acute severe asthma and ARDS.⁹¹ Increased Paco₂ has a variety of physiologic effects and data show that it is well tolerated, even at sometimes extreme levels.^89,91

Cerebral blood flow is linearly related to Paco₂. For every mm Hg change in Pco₂, a 3% change in CBF can occur.⁷⁹ In the setting of TBI, cerebral blood flow is altered, and this is not a static phenomenon. Hypocapnia rapidly induces an increase in cerebrovascular resistance and a decrease in CBF (and a subsequent decrease in ICP). If hypocapnia is prolonged, vascular constriction can contribute to cerebral ischemia, promoting secondary brain injury.⁹² For these reasons, prolonged hypocapnia cannot be recommended, and therefore hyperventilation is recommended only for life-threatening intracranial hypertension while other therapies are instituted.

Hypercapnia has the exact opposite effect. For patients with increased ICP, hypercapnia may be the cornerstone of the discordant and sometimes competing goals of safe mechanical ventilation, aimed at limiting tidal volumes and airway pressures.^26,27 These concerns are real, as increasing Paco₂ can certainly increase ICP. Examination of the high-quality protective ventilation clinical trials in ARDS reveals the Paco₂ levels from day 0 to day 7 to be in the normal to high 40s mm Hg range.³⁰ Also, high tidal volume ventilation in the setting of TBI has been shown to be associated with the development of ARDS, which not only has been shown to be associated with brain hypoxia but also increases morbidity and mortality rates in this patient population.^76,93 Given these findings, it is likely that the majority of patients with ALI and elevated ICP can safely be managed with a low tidal volume strategy. The balance between permissive hypercapnia and ICP will likely be a fine line and should be managed on a case-by-case basis. Knowing the physiology governing Paco₂ levels, V̇co₂ can be decreased with proper sedation, fever and seizure control, and appropriate nutrition goals. Dead space in the ventilator circuit can easily be removed and is a simple solution, albeit with varying results. While alveolar recruitment and PEEP are often viewed as a strategy to ameliorate shunt and hypoxia, effective recruitment also decreases dead space, allowing more effective ventilation and less minute ventilation requirements. In the setting of ALI, manageable ICP, and adequate CPP, a Paco₂ high threshold of 55 mm Hg is a reasonable target, as this level will likely facilitate protective lung ventilation and can be acutely corrected should elevated ICP become an issue. In the setting of difficult-to-control ICP, normocapnia or slight hypocapnia should be maintained. If this is accomplished with a nonprotective lung ventilation strategy, the patient should be managed with the lowest tidal volumes and airway pressures as possible, and every attempt should be made for the safest ventilatory strategy as soon as possible. All patients with an ICP monitor and dynamically changing respiratory system mechanics and/or ventilator settings should also be managed with an end-tidal Co₂ detector as well. Similarly, a direct, invasive CBF monitoring probe is available and can give readily available data on the effects of changing Paco₂ on CBF.

The majority of patients with ALI do not die of hypoxia, but rather multiple organ dysfunction syndrome (MODS).^{94, 95,96,97} While mechanical ventilation can be a life saving intervention, it has great potential for harm, and the clinician's focus should extend beyond normalization of gas exchange to providing safe and physiologically sound mechanical ventilation. Perhaps nowhere is this demonstrated more significantly than with the concept of VILI. Despite a uniform appearance on chest radiography, the distribution of ventilator defects and the distribution of injury is very heterogenous in ALI.^53,98 As such, applied airway pressures have a heterogenous distribution, leaving more compliant lung units overdistended and others collapsed and nonrecruited. CT scans have provided valuable insight into the pathoanatomy of ALI, showing alveolar units with normal gas tissue density (ventrally), alveolar units totally collapsed and potentially nonrecruitable (dorsally), and others in between, which likely remain recruitable and prone to repetitive opening and closing.^53,98 Respiratory system compliance is related to the amount of normally aerated lung tissue that remains, the so-called baby lung.²⁰ This heterogeneity leads to a maldistribution of ventilation, with some alveoli overdistended, others collapsed throughout the respiratory cycle, and others cyclically opening and closing. VILI is a spectrum, including barotrauma, atelectrauma, and volutrauma (stretch injury).⁹⁹ All of these can lead to biotrauma, or the decompartmentalization of the inflammatory response secondary to increased alveolar epithelial–capillary endothelial permeability. This results in the release of biologic inflammatory mediators and the spread of injury to distant organs, causing MODS and increased mortality. In an individual patient, it is likely some combination of these factors contributing to VILI.

Barotrauma represents extra-alveolar air from pneumothoraces, pneumomediastinum, subcutaneous emphysema, pneumoperitoneum, and pulmonary interstitial emphysema. While some controversy exists as to the exact pathophysiology of its development, it is likely a combination of excessive airway pressures and tidal volumes, coupled with at-risk patients, namely, those with expiratory flow limitation (asthma, COPD) and/or severe ALI. Atelectrauma refers to the repeated cyclic opening and closing of alveoli. Theoretically, this closure would occur at end expiration at low lung volumes, as the volumes drop below FRC and the LIP on a PV curve. Reopening occurs with each inspiratory cycle and this repetition is injurious to alveoli, causing shear injury. Overinflation of alveoli with high pressures and/or volumes (likely more significant) has been established to cause injury indistinguishable from ALI. Alveolar epithelial injury, capillary endothelial injury, and diffuse pulmonary edema can occur. Barotrauma, atelectrauma, and volutrauma likely all lead to some extent to biotrauma.

So is it excessive pressure or volume, and what is the evidence?¹⁰⁰ Lung protective ventilation has been shown to improve outcome, but the exact mechanism of lung protective ventilation is not known.³⁰ Excessive stretch is known to cause lung injury, whether it is due to pressure or volume, and the addition of PEEP has been shown to ameliorate lung injury, likely by preventing atelectrauma.¹⁰¹ By banding the chest and abdominal wall in rats (and therefore controlling pressure), Dreyfuss et al showed that it is perhaps higher tidal volumes that are most important, as the higher tidal volume group developed ALI regardless of airway pressures.¹⁰² This evidence, coupled with clinical trials of low tidal volume ventilation, points to the importance of tidal volume as perhaps the greatest contributor to VILI. The effects of flow rate and respiratory rate on VILI has been described, but with less convincing evidence.^103,104,105 Numerous studies lend support to the concept of biotrauma, showing increased cytokines in bronchoalveolar lavage and serum in ALI. Data show increased levels of these markers with higher tidal volumes and airway pressures and without PEEP, and a decrease with the use of protective lung ventilation.^{106,107,108,109}

Human studies support the concept of VILI and the importance of a protective ventilation strategy with low tidal volumes and the use of PEEP. Low tidal volume ventilation is the only intervention shown to consistently improve outcome in ALI, and support for protective lung ventilation to decrease VILI from several well-conducted clinical trials exists.^30,57,58 This includes data showing a decrease in inflammatory mediators in patients ventilated with a protective lung strategy. In the absence of a patient-specific factor to suggest otherwise, a protective ventilation strategy, consisting of low tidal volume ventilation, PEEP setting, and limitation of P_PL, should be attempted at all times. Emerging data suggest that these strategies should be employed not just in patients with ALI, but in all patients receiving mechanical ventilation, especially those with known risk factors for development of ALI.¹¹⁰

The brain and peripheral nervous system are affected by critical illness. Encephalopathy and critical illness polyneuropathy are well described in the literature. VILI leading to MODS is also well described; however, this organ dysfunction has historically focused on organs other than the brain, namely, the kidneys, lungs, and heart. This is despite the neurologic complications of mechanical ventilation existing in the literature for years and the extracranial complications of acute brain injury being well known.²⁹ Mounting data suggest that this focus should also include the brain, and clinicians should view the brain as an end organ in MODS.¹¹¹ Patients with brain injury have a known predisposition both to requiring mechanical ventilation and to progressing to ALI, and are therefore innately at risk for additional brain dysfunction secondary to mechanical ventilation and VILI. The brain as an end organ for VILI should be considered, and brain-lung interactions should be viewed as crucial. While it would be difficult to ascribe the effects of VILI as being more important than the sequelae of the neurologic injury itself, the effects of ventilator settings on brain function should be considered. Also, patients with acute neurologic injury frequently present with preexisting comorbidities. Given these facts, a paradigm shift should occur as no patient should be considered to have “isolated head injury.”

Neurologic dysfunction in mechanically ventilated patients is a spectrum that includes delirium, dementia, cognitive decline, and loss of IQ, as well as mood and memory disorders.^112,113 These types of brain injuries can occur with alarming frequency and are likely underappreciated in the critical care population as a whole. Multiple studies show persistent cognitive deficits that exist years after liberation from mechanical ventilation. Similar findings exist with respect to quality of life, IQ loss, and mood and memory deficits.^{114,115,116,117,118} Patients with ALI have brain atrophy and ventricular enlargement, suggesting neuronal cell damage of some etiology.¹¹²

The pathophysiology of this neurologic dysfunction is multifactorial. This includes the initial injury, hypoxia, hypotension, loss of central nervous system connectivity, persistent inflammation from nonprotective ventilation strategies, and metabolic abnormalities, such as hyper- and hypoglycemia.¹¹¹ Sedation regimen, specifically benzodiazepines, likely plays a factor as well.^113,119

There is no specific treatment for VILI, but the clinician should focus on prevention, have a heightened awareness of the at-risk patient, and provide sound evidence-based and physiologic-centered care. Although specific recommendations on how to set the mechanical ventilator in brain-injured patients are lacking, care must taken to protect all organs with respect to VILI. Patients with ALI should have low tidal volume ventilation at 6 to 8 mL/kg of IBW. PEEP should be set as previously recommended, and P_PL should be limited to at least less than 30 cm H₂O. Given the heterogenous distribution of lung injury, there is likely no safe P_PL, and at any given value, there are likely some alveolar lung units that remain collapsed, while others are overdistended. P_PL should be kept as low as possible in patients with ALI. If monitored with an esophageal balloon, transalveolar pressure should be limited to at least less than 25 cm H₂O. Brain-injured patients may also benefit from continuous brain tissue oxygen monitoring, and optimization of brain oxygenation should be considered a potential aim for therapy. When appropriate and safe, all patients should be awakened daily, have daily spontaneous breathing trials, have target-based sedation to allow appropriate limitation of sedative infusions, have limited benzodiazepine exposure, and be monitored for delirium.¹¹³ These interventions as a whole should result in decreased mortality, a decrease in all mechanisms of VILI, increased ventilator-free days, fewer ICU days, less sedation use, and less delirium.

Patients with obstructive lung disease and expiratory flow limitation (most commonly asthma and COPD) present unique challenges with respect to mechanical ventilation. Regardless of the specific disease that caused the pathology, the increase in airway resistance, loss of normal lung elastance, and airway narrowing lead to alveolar regions that have difficulty emptying and returning to resting end-expiratory lung volume, as well as increased inspiratory muscle workload.^5,6,52 Because of inadequate expiratory time, these patients are prone to air trapping and iPEEP, increased respiratory pump workload, and a variety of gas exchange abnormalities.

In the setting mentioned above, iPEEP can occur, as alveolar units with prolonged time constants are not allowed to empty to resting end-expiratory lung volume. This can have several cardiopulmonary consequences. By raising intrathoracic pressure and therefore right atrial pressure, venous return may be impaired. Of note, the effects of iPEEP and extrinsic PEEP (set on the ventilator) are similar from a hemodynamic, alveolar recruitment, and alveolar overdistention standpoint, as long as they produce similar changes in end-expiratory lung volume.^120,121 Because of compression of alveolar capillaries and increases in West zone I and zone II conditions, the high lung volumes may result in increased pulmonary vascular resistance and RV failure.¹²² Overinflation also puts inspiratory muscles at a mechanical disadvantage, substantially increasing work of breathing and contributing to patient-ventilator dyssynchrony and inability to effectively trigger. In contrast to parenchymal pathology, the ventilation/perfusion mismatch is one largely of increased dead space, as opposed to shunt. This is secondary to capillary loss, as well as overinflation, which leads predominantly to hypoventilation and type II respiratory failure.^5,6,52 Hypoxemia may also occur due to ventilation and blood flow redistribution, but is not usually the etiology of the respiratory failure in the setting of expiratory flow limitation. If it does occur, it can contribute to increased pulmonary vascular resistance and RV failure as well.

These changes are in contrast to the patient with alveolar filling and consolidation, parenchymal pathology, and reduced compliance. The primary etiology of respiratory failure is type I, with hypoxia driven by shunt physiology. As outlined above, this patient population should benefit from alveolar recruitment with PEEP setting and judicious tidal volume management.

The goals of managing patients with expiratory flow limitation and obstructive physiology are largely the same as for patients with parenchymal lung injury and include oxygenation, protection against VILI, ventilation, preservation of respiratory muscle function, and prevention of secondary neurologic injury. Because of the physiology outlined above, the clinician should also add the following goal: allow for adequate expiratory time to limit the effects of iPEEP.

Unlike ALI, there are no large randomized controlled trials to guide the clinician in how to best employ mechanical ventilation in this setting. Instead, basic knowledge of mechanical ventilation and physiologic rationale help guide therapy and patient response. When mechanically ventilating this population, a reasonable approach to ventilator settings can have similarities to ALI. Tidal volume should be 6 to 8 mL/kg IBW and P_PL limited to at least less than 30 cm H₂O. Expiratory time should be long enough to limit iPEEP. The most effective way is to decrease respiratory rate. Expiratory time can also be extended by decreasing tidal volume and increasing flow rate (in volume-targeted ventilation), but the clinician should be aware that this is less effective and an increased flow rate results in increased airway pressure. Intrinsic PEEP should be monitored closely. This can be done by monitoring the flow/time waveform to assess for no flow before inspiration begins. If expiratory flow remains when inspiration begins, iPEEP is present. Also, an expiratory hold maneuver can stop flow in the ventilator circuit, allowing detection of iPEEP. The clinician should also make note of the purpose of extrinsic PEEP in this patient population. Unlike parenchymal lung injury, in which PEEP is delivered to maximize alveolar recruitment and eliminate shunt, PEEP in the setting of expiratory flow limitation is usually not needed for hypoxia/shunt. For a breath to occur, a pressure gradient must be overcome from the ventilator to the patient. This gradient is increased in the setting of iPEEP. The purpose of extrinsic PEEP is to therefore overcome this inspiratory threshold load imposed by iPEEP, assist in triggering, and ameliorate dyssynchrony. Intrinsic PEEP should be measured and the clinician should set extrinsic PEEP to some value less than this (perhaps as high as 80% of iPEEP).^120,121 This will facilitate triggering while not adversely raising alveolar pressures. If set higher, extrinsic PEEP can raise alveolar pressure and produce overdistention. The net effect of tidal volume, respiratory rate, and iPEEP limitation can result in hypercapnia, which is generally very well tolerated in this patient population. However, despite a decrease in minute ventilation, if this strategy allows for decreased dead space, then ventilation may actually improve. In the neurologically injured patient, if ICP is a concern, permissive hypercapnia may not be tolerated. As in all instances, the balance will be a very fine one and the benefit of limiting iPEEP and tolerating higher Paco₂ levels must be weighed against secondary brain injury and elevated ICP. With these concerns, volume-targeted ventilation may be best in this patient population, given the strict control of minute ventilation it allows.

The majority of patients, regardless of the etiology of their respiratory failure, can be managed with conventional modes of mechanical ventilation, such as volume- or pressure-targeted assist control. However, in certain clinical situations, other modes of mechanical ventilation can be used. The clinician should recognize that clinical outcome data for these modes of mechanical ventilation are not as robust. However, the physiologic rationale does make sense in certain clinical situations.

Airway Pressure Release Ventilation

Airway pressure release ventilation (APRV) has been described in the literature since 1987.¹²³ It is a time-triggered, pressure-targeted, time-cycled mode of mechanical ventilation that exactly resembles AC/PC with very extended I-to-E ratios if the patient is not spontaneously breathing.^10,25 It has been described as continuous positive airway pressure (CPAP) with an intermittent, time-cycled release phase to a set lower pressure. Whereas conventional modes of mechanical ventilation elevate airway pressure up from a set baseline to accomplish tidal ventilation, APRV employs very extended I-to-E ratios to maintain mean airway pressure and brief deflations to accomplish ventilation. The expiratory valve in APRV also facilitates spontaneous breathing throughout the ventilatory cycle, facilitating ventilation. Pressure high (P_HIGH) is the baseline airway pressure that occupies most of the ventilatory cycle. Pressure low (P_LOW) is the release pressure and is set at 0 cm H₂O. Time high (T_HIGH) is the length of time for which P_HIGH is maintained, and time low (T_LOW) is the length of time for which P_LOW is maintained.

While outcome data for APRV are lacking compared to conventional ventilation using protective lung strategies, the reported advantages of APRV include sustained alveolar recruitment with improved oxygenation, higher mean airway pressures accomplished with lower P_peak and P_PL, spontaneous breathing throughout the ventilatory cycle, and decreased use of sedation and neuromuscular blockade.²⁵ The maintenance of spontaneous breathing may improve V̇/Q̇ matching by preferential ventilation of dependent lung regions, and the higher mean airway pressure, relative to P_peak and P_PL, may limit VILI.

Alveolar recruitment is a pan-inspiratory phenomenon, and alveoli that are recruited are more compliant than recruiting or nonrecruited alveoli. With prolonged elevated pressures, APRV likely recruits alveoli that require a longer inflation with higher threshold opening pressures.^10,25 Sustained inflation maintains recruitment, decreases shunt and dead space, and uses the more compliant expiratory limb of the PV curve. Ventilation is determined by the stored kinetic energy at the high pressure, the intermittent release phase, and is augmented by spontaneous breathing. While minute ventilation is decreased with this mode of ventilation, ventilation is also improved by a decrease in dead space.

Clinical data are limited but have shown improved oxygenation, less shunt and dead space, and a decreased need for sedation and neuromuscular blockade.^{25,124,125,126,127} Data on clinically relevant outcomes, such as mortality, mechanical ventilation days, and ICU days, are very limited and conflicting, with no high-quality trials specifically looking at the neurologically injured patient. This mode remains physiologically attractive for its effects on oxygenation and potential to limit VILI; however, no specific recommendations can be made given the lack of good outcome data.

High-Frequency Oscillatory Ventilation

Fundamentally, high-frequency oscillatory ventilation (HFOV) is very different from conventional bulk flow ventilation and uses respiratory frequencies much higher and tidal volumes much lower than conventional modes. It is similar to APRV in that it aims to elevate mean airway pressure to maximize recruitment and oxygenation while limiting P_peak.¹²⁸ Given the fact that tidal volume is often less than anatomic dead space, HFOV relies on gas transport mechanisms much different than those with bulk flow ventilation. These include some convective gas transport, but also molecular diffusion, pendelluft, coaxial flow, and Taylor dispersion.¹²⁹

When initiating HFOV, the mean airway pressure should be set 2 to 5 cm H₂O higher than that on conventional ventilation. Prior to this, the patient's endotracheal tube should be verified to be patent, as heavy secretions or kinks in the endotracheal tube will significantly harm ventilation. The bias flow delivers fresh gas into the ventilator circuit at 40 to 60 L/min and helps maintain mean airway pressure.¹²⁹ HFOV relies on a piston-type mechanism to apply pressure, and this amplitude (Δ pressure), as well as frequency (hertz), helps determine ventilation. The Δ pressure is set to observe vibration down to the patient's thighs, and initial frequency is set at 5 hertz.

Similar to APRV, clinical outcome data on HFOV are limited. It is also physiologically attractive with respect to limitation of VILI and improvement in oxygenation, and promising data do exist with respect to oxygenation and perhaps improved outcome with the application of HFOV. This effect seems most beneficial when applied early in the course of ALI.^128,130,131 Limited data suggest that HFOV is safe in patients with elevated ICP and is well tolerated in this subgroup of patients.^28,132,133 In patients with mean airway pressures greater than 24 cm H₂O in whom protective lung ventilation strategies are failing with respect to oxygen requirements and airway pressures, HFOV should be considered. The clinician should note that these patients will require heavy sedation and/or neuromuscular blockade.

Pressure-Regulated Volume Control

Pressure-regulated volume control (PRVC) is a dual-control mode of ventilation that is very similar to AC/PC. It is a pressure-targeted breath, time cycled, with square pressure waveform and decelerating variable flow. Tidal volume is set on the ventilator, and a “test breath” is delivered. The ventilator then calculates the C_RS and adjusts the needed pressure delivered to achieve the set tidal volume. In this regard, tidal volume is used as a feedback control, which changes the pressure limit (usually breath to breath) as respiratory mechanics change. It maintains the advantage of guaranteed minute ventilation while limiting pressure and using a decelerating flow pattern.⁷

Proportional Assist Ventilation and Neurally Adjusted Ventilatory Assistance

Proportional assist ventilation (PAV) is an alternative mode of partial support, where the ventilator delivers pressure in proportion to the patient's effort.¹³⁴ The theory behind PAV is based on the equation of motion. Through a sophisticated software algorithm, the ventilator is able to know the flow, volume, airway resistance, and lung compliance, and therefore is able to calculate patient work of breathing. The ventilator responds to the mechanical output of the patient and will amplify patient effort with a preset proportional amount of pressure. The clinician sets the percent effort for the ventilator to provide support based on patient effort and desire.

Whereas PAV responds to a mechanical output, neurally adjusted ventilatory assistance (NAVA) responds to the neural input from the electrical activity of the diaphragm (Edi), and the patient retains all control of ventilator pattern. Pressure is delivered in proportion to the electrical activity of the diaphragm. This requires the placement of an esophageal electrode (similar to nasogastric tube placement). The ventilator is triggered based on Edi, therefore improving synchrony, as the time delay from patient effort to breath is very brief. In addition, the amount of ventilator assistance varies based on Edi. As such, to achieve greater assistance, the patient must increase Edi. Clinical data on NAVA are limited but have shown improved synchrony, preserved ventilatory pattern over a wide range of PEEP, and successful offload of respiratory muscles.^{135,136,137,138}

Like many clinical situations involving choice of mechanical ventilator mode, the decision to use PAV or NAVA should be tailored to the individual patient. Patients with a high amount of dyssynchronous breaths and iPEEP secondary to obstructive physiology may benefit from a switch to these modes. These modes do not, however, ameliorate iPEEP, and extrinsic PEEP should be delivered appropriately. In patients recently intubated (< 48 hours), severely hypoxemic, or hemodynamically unstable, these modes should be avoided. Both modes require an intact ventilator drive, and ongoing assessments of pulmonary mechanics are a necessity. If sudden changes in respiratory system mechanics occur, the percent assist must be adjusted. Finally, as neither mode is controlling any factors in the equation of motion, neither mode can change the ventilator pattern (only assist level is being changed), and the clinician must give up a significant amount of control when using PAV and NAVA.

I would like to thank Dr. William A. Knight IV from the University of Cincinnati for his thoughtful and very helpful review of this chapter, and Dr. Nicholas Mohr from the University of Iowa for aid with table and figure preparation.