Chapter 81. Mechanical Ventilation for the Surgical Patient

1. Patients require mechanical ventilation because of apnea, acute or impending acute respiratory failure, or severe refractory hypoxemia.

2. The 2 basic forms of mechanical ventilation are pressure ventilation (peak airway pressure constant, tidal volume variable) and volume ventilation (tidal volume constant, peak airway pressure variable).

3. Although numerous modes of ventilation exist, few data are available to differentiate the benefits of one mode over another, and no mode has been shown to improve patient outcome.

4. A major concern during assisted ventilation is patient–ventilator synchrony—the ventilator should be set to match the patient's ventilatory demands.

5. Auto positive end-expiratory pressure (PEEP) is a common cause of patient–ventilator dyssynchrony; in patients with obstructive lung disease, properly set applied PEEP can improve synchrony and decrease patient efforts to ventilate.

6. It is unnecessary to achieve normal PaO2 and PaCO2. In most critically ill patients, a PaO2 of 60 mm Hg or higher is acceptable, and permissive hypercapnia may be useful in treating some patients.

7. Ventilator-induced lung injury is primarily caused by localized overdistension and the opening and closing of unstable lung units.

8. In most patients who are ventilated, the tidal volume should be 4 to 8 mL/kg predicted body weight (PBW) and the plateau pressure should be less than 30 cm H2O, and PEEP should be set to avoid the collapse of unstable lung units.

9. High-frequency ventilation and airway pressure release ventilation, although useful in managing patients with acute respiratory distress syndrome (ARDS), show no outcome benefits over conventional pressure or volume ventilation.

10. Noninvasive positive-pressure ventilation is useful to transition patients who are at high risk of extubation failure from invasive ventilation to spontaneous breathing.

Most surgical patients do not require support of the respiratory system beyond the immediate postoperative period. However, others with preexistent chronic respiratory diseases, trauma patients, and patients who require extensive surgical procedures may require more lengthy periods of respiratory support. This chapter focuses on mechanical ventilation of the postoperative patient. It primarily discusses invasive ventilation but also reviews selected data for noninvasive respiratory support and the application of noninvasive positive-pressure ventilation (NPPV) and continuous positive airway pressure (CPAP) by mask.

The main objectives of mechanical ventilation in the postoperative patient are to decrease the work of breathing and the load on the cardiovascular system, and to reverse life-threatening hypoxemia or respiratory acidosis. Table 81-1 summarizes some of the specific indications for mechanical ventilation.1,2 In an international survey of 1638 patients requiring mechanical ventilation reported by Esteban et al, acute respiratory failure was the indication for mechanical ventilation in most patients (66%), followed by coma (15%), an acute exacerbation of chronic obstructive pulmonary disease (13%), and neuromuscular disorders (5%).3 Included under the heading of respiratory failure were postoperative respiratory failure, acute respiratory distress syndrome (ARDS), heart failure, pneumonia, sepsis, and complications of trauma.3

Pressure versus Volume Ventilation

During all forms of mechanical ventilation the ventilator must be programmed to deliver gas to achieve specific target values. The format for gas delivery in most ventilator modes is to target either pressure or volume (Table 81-2).1 In volume-targeted ventilation (the original approach to ventilatory support), a specific tidal volume is set by the clinician and the ventilator ensures that the volume is delivered regardless of inspiratory pressure (up to a clinician-set limit). In addition to tidal volume, the flow waveform and peak flow or inspiratory time must be set. With this approach to ventilatory support the focus is on ensuring that ventilation is maintained at a targeted level.

Pressure ventilation is essentially an opposite approach to ventilatory gas delivery. A peak airway pressure is set, with the aim of limiting the peak alveolar pressure. That is, with each breath pressure increases to the set level before the breath terminates. However, the tidal volume and gas flow are allowed to vary from breath to breath. During pressure ventilation the clinician must set the targeted pressure, and in some modes the inspiratory time. Flow is provided rapidly at first, then in an exponentially decelerating pattern where the rate of deceleration is dependent upon the patient's inspiratory demand and lung mechanics.4 Essentially, the ventilator rapidly delivers gas flow to establish the targeted pressure, but once the pressure target is met, the flow must decrease to avoid exceeding the set pressure. This approach to ventilatory support focuses on ensuring that a targeted pressure is met and never exceeded.

As noted in Table 81-3, pressure and volume ventilation respond differently to changes in the impedance to gas delivery. With volume ventilation, pressure increases if the patient becomes more difficult to ventilate, whereas with pressure ventilation the tidal volume is decreased. Most modes of ventilation currently available essentially function based on one of these 2 formats.

Combined Pressure and Volume Targeted Modes

Some of the newer modes of ventilation (see New Modes of Ventilation section) attempt to combine the beneficial effects of volume ventilation (delivering a target tidal volume) and pressure ventilation (the maximum airway pressure is limited).4 With these modes both volume and pressure are targeted. Their gas delivery algorithms are designed to ensure that on average the set tidal volume is delivered, but also that the peak airway pressure does not exceed the set level.4 In other novel ventilatory modes, gas may be delivered initially in a volume ventilation format, but if the patient's demand is increased, the gas delivery mode changes to a pressure format to ensure that ventilatory demand is met.4

The classic modes of ventilation include control, assist/control, assist (pressure support), and synchronized intermittent mandatory ventilation (Table 81-4). All of these modes have been available on mechanical ventilators for more than 25 years and generally can provide the basic approaches to ventilatory support for most surgical patients.

Control Mode

This is the original mode of mechanical ventilation. Controlled ventilation is available in both pressure and volume ventilation formats (Figs. 81-1 and 81-2). In this mode, the ventilator controls all aspects of gas delivery. That is, the patient is assumed to be a passive recipient of mechanical ventilation.5 With the earliest generation of mechanical ventilators, patients were not able to trigger the ventilator to produce an inspiration. However, with today's modern ventilators, even when the airway sensitivity is set at the most insensitive setting, patients with a strong ventilatory drive can still trigger the ventilator. The control mode of ventilation is achieved today by sedating the patient to apnea.

Assist/Control Mode

This mode is essentially the control mode with the sensitivity level set to trigger gas delivery by a spontaneous negative airway pressure. That is, the patient initiates gas delivery (Fig. 81-3). The sensitivity control adjusts the level of patient effort (spontaneous negative airway pressure) required to trigger the mechanical breath. Although the patient determines the ventilatory rate, a "backup" rate is set to ensure a minimum rate of ventilation.5 Worldwide this is the most commonly used mode of ventilation.3 During volume assist/control the following variables are usually set: tidal volume, flow waveform, backup rate, peak inspiratory flow rate or inspiratory time, inspiratory trigger sensitivity, FIO2, and positive end-expiratory pressure (PEEP). With the pressure assist/control mode the peak pressure, inspiratory time, backup rate, inspiratory trigger sensitivity, FIO2, PEEP, and rise time must be set.

Pressure Support

The closest classical mode to true assist ventilation available on all modern intensive care unit (ICU) ventilators is pressure support (Fig. 81-4). With assist ventilation there is no backup rate. In pressure support, backup safety modes of ventilation take over if patients become apneic for 20 seconds or longer, but no true backup rate is available.5 Pressure support is very similar to pressure assist/control.6 The major feature of gas delivery that differs in this mode is the mechanism that terminates inspiration. With pressure assist/control, inspiration is always terminated by time, whereas with pressure support, inspiration is usually terminated by decreasing gas flow. That is, when tracheal flow decreases to a specific level (usually 25% of the peak flow for that breath) the breath is terminated. There are, however, 2 alternate methods of terminating inspiration in pressure support: pressure exceeding the set level after about 300 milliseconds of inspiration or the inspiratory time exceeding the manufacturer's set level, usually 2 to 3 seconds. However, some newer ventilators allow the clinician to set the maximum inspiratory time during pressure support ventilation. Both of these secondary termination criteria are valuable for patient safety and prevent lengthy inspiratory times and prolonged elevation of airway pressures, which can reduce the cardiac output.

Of the classic modes of ventilation, pressure support allows the patient the greatest control over the process of ventilation. Not only does the patient trigger each breath, but the ending of the breath is based on the patient's demand. As with all pressure-targeted approaches to ventilation, the pressure support mode allows the patient to vary the tidal volume of each delivered breath. The only gas delivery variable set other than sensitivity is the pressure level. Pressure support may be used to ventilate any patient with a stable ventilatory drive.

Inspiratory Termination Criteria

Inspiratory termination criteria, also referred to as E-sens or expiratory sensitivity, is an adjunct to pressure support available on many ventilators as a method of ensuring that the patient and ventilator end their inspiration simultaneously. As Figure 81-5 illustrates, an increase in pressure at the end of a pressure support breath is abnormal and usually indicates that the patient has begun exhalation before the ventilator has allowed exhalation to occur.7 When this increase in inspiratory pressure is observed, the patient is contracting accessory muscles of exhalation during the terminal aspect of the ventilator's inspiratory phase.8 This results in an increased ventilatory drive and ventilatory rate, and the development of patient–ventilator dyssynchrony. E-sens allows adjustment of the percentage of peak flow terminating the breath. Whenever a spike at the end of a pressure support breath is present, a greater percentage of the peak flow should be set to terminate the breath. In this setting the E-sens percentage should be slowly increased until there is a smooth transition from inspiration to expiration.9

Rise Time

Patient–ventilator synchrony at the onset of a pressure-targeted breath (eg, pressure assist/control, pressure support, pressure-regulated volume control) can be improved by adjusting the rise time. As noted in Figure 81-6, the flow rise time varies the slope of the pressure increase at the onset of a breath by varying the time it takes gas flow to increase from zero to peak.7 Rise time should be adjusted to ensure that airway pressure rises rapidly to the peak level without any concavity during the initial airway pressure waveform. If the airway pressure waveform is concave the rise time should be increased (made more rapid); however, if peak pressure exceeds the set level at the onset of inspiration the rise time should be decreased. An increase in rise time generally results in an increase of peak flow and a decreased inspiratory time (with pressure support ventilation). In most patients with a strong ventilatory drive the rise time should be set between the mid and most rapid level. Proper setting of rise time usually results in a decreased mechanical ventilatory rate.

Synchronized Intermittent Mandatory Ventilation

This mode of ventilation combines spontaneous unsupported breathing with the assist/control mode. As with the assist/control mode, the mandatory positive-pressure breaths can be either pressure or volume targeted. Figure 81-7 illustrates the typical volume-targeted synchronized intermittent mandatory ventilation (SIMV). A mandatory respiration rate is set, and between the mandatory breaths the patient can breathe spontaneously.5 As noted in Figure 81-8, an assist/control window is open at specific intervals based on the selected mandatory respiration rate.10 Within each window the patient is able to trigger positive-pressure breaths; if the ventilator does not sense the patient's efforts during this time period a controlled positive-pressure breath is delivered. Ventilator adjustment is the same as that with the control mode except that the sensitivity must also be properly set.

Advocates of SIMV emphasize the benefits of spontaneous breathing between mandatory breaths. As shown in Figure 81-8, spontaneous breathing can improve ventilation/perfusion (V/Q) matching by improving the distribution of ventilation to dependent lung regions.10 In addition, the negative intrathoracic pressure generated during spontaneous breathing decreases the mean intrathoracic pressure, thereby improving cardiac output. However, the work of breathing during SIMV can be excessive. As noted in Figure 81-9, as the mandatory rate is reduced the work of breathing for both mandatory and spontaneous breaths increases.11 The surprising reason for this is that the patient's respiratory center has a difficult time rapidly changing its output based on ventilatory load as the mandatory rate decreases. Essentially, each breath is interpreted as being at the higher load and requiring the higher muscular work output. This is nicely illustrated in Figure 81-10; note that the electromyelogram (EMG) activity of the diaphragm and the sternocleidomastoid muscles, as well as the esophageal pressure changes, are the same for both mandatory and spontaneous breaths.12 In other words, the patient's effort is the same regardless of breath type, although the efficiency of gas delivery may be better during the mandatory breaths. Unfortunately, this set of circumstances increases the patient's ventilatory drive and increases the level of patient–ventilatory dyssynchrony.

With most modern ventilators, the pressure support mode can be applied during spontaneous patient breaths; however, this negates the benefits of unassisted spontaneous breathing and increases the complexity of weaning patients from ventilatory support. In general, the lower the mandatory rate during SIMV the greater the likelihood that the patient's effort will be excessive and dyssynchrony will be present.

Table 81-4 lists the latest modes of ventilation currently available on the newest generation of mechanical ventilators. These modes are designed to take advantage of the beneficial aspects of both volume and pressure ventilation. As noted in Table 81-4, they are divided into modes that adjust gas delivery during a given breath and those that adjust gas delivery on the subsequent breath. In this latter group, adjustment is based on the ability of current settings to achieve gas delivery targets.

Pressure-Regulated Volume Control

This mode of ventilatory support targets both a maximum airway pressure and a tidal volume and is a variation of the pressure assist/control mode. This same mode is referred to on some ventilators as autoflow, volume control plus, or possibly another term.4,5 It accomplishes its goals by varying the pressure applied on the subsequent breath based on the tidal volume achieved on the current breath. All of the ventilators providing this mode of ventilation first deliver a test breath at some low level of pressure and from the volume delivered calculate the pressure required to deliver the targeted tidal volume. The ventilator then automatically adjusts the pressure in 1 or 2 steps to the pressure level needed to deliver the targeted tidal volume. During every subsequent breath, the tidal volume that is delivered is reassessed, and based on this assessment the pressure on the subsequent breath is adjusted ±3 cm H2O to ensure that the targeted volume is delivered on the next breath.4,5 Pressure during pressure-regulated volume control (PRVC) can be increased to a maximum level selected by the clinician and can be decreased with some ventilators all the way to the CPAP/PEEP level, thereby allowing unassisted spontaneous breathing. Herein lies our major concern with PRVC: If a patient has a strong ventilatory drive and the additional stimulus of, for example, hypoxemia, fever, or sepsis, the level of ventilatory support could decrease inappropriately to either a very low level or to no support. As a result, be cautious when using this mode with patients who have a strong ventilatory drive.13 Marked improvement in the operation of this mode could be made by adding a minimum pressure limit, that is, setting a low pressure limit.4 As a result, this mode is most useful in those patients who are receiving control mode ventilation or for those patients who have limited ventilatory demand. To use PRVC, first set all variables used during pressure assist/control, then add the target tidal volume and maximum pressure limit.

Volume Support

The volume support mode operates similarly to PRVC, but is based on the pressure-support mode and not the pressure assist/control format.4,5 Ventilator software must assess the pressure needed to deliver the tidal volume and provide breath-to-breath pressure adjustment to ensure the targeted tidal volume. The same safety concern regarding a patient with a strong ventilatory demand exists with volume support (as with PRVC). The adjustments made to the ventilator during setup are the same as those with pressure support, with the addition of setting the target volume and maximum airway pressure.

Two randomized controlled studies of PRVC and/or volume support have been conducted—one in neonates14 and the other in pediatric patients.15 In the study of newborns,14 PRVC was compared to intermittent mandatory ventilation (without synchronization of the mandatory breath). In this study, neonates who were studied weaned faster with PRVC. In the second,15 volume support was compared to pressure support with and without a protocol for weaning children from mechanical ventilation. No difference in the rate of weaning was observed regardless of the mode of ventilation. As a result, there is little enthusiasm for the use of either of these new modes of ventilation for weaning.

Adaptive-Support Ventilation

This is a very unique new mode of ventilatory support. It attempts to adjust ventilator settings to ensure that an ideal ventilatory pattern is achieved,16 with ideal being defined as the pattern requiring the least patient and ventilator work based on the patient's measured mechanical lung characteristics and breathing effort. The goal of adaptive-support ventilation (ASV) is to provide a preset level of minute ventilation while minimizing the total work of breathing expended by both the ventilator and the patient. It can be applied to both patients who require controlled ventilation and to those who are actively breathing.

During set up the clinician must indicate the patient's ideal body weight (males: predicted body weight [PBW] (kg) = 50 + 2.3 [height (in) (60)]; females: PBW (kg) = 45.5 + 2.3 [height (in) (60)])17 and the percentage of minimal minute volume (% min MV) to be delivered. The % min MV can be adjusted between 25% and 350% of predicted body weight.16 From the PBW and % min MV the maximum rate is calculated as

or

whereas the maximum tidal volume delivered is the targeted MV/5.16 With this information the ventilator determines the optimal breathing pattern that will minimize the work of breathing, that is, a ventilatory pattern that results in minimum work of breathing based on the lung and chest wall mechanics of the patient.

With the initial application of ASV the ventilator provides a series of 5 pressure-limited test breaths at a rate of about 10 breaths/min with a maximum pressure of 15 cm H2O. During these breaths, the ventilator measures dynamic compliance, the respiratory time constant, tidal volume, and the patient's respiratory rate.16 These measurements are used to determine the initial targets for breathing rate and tidal volume. As ventilation continues the ventilator software recalculates the above variables from each breath and determines the new rate and tidal volume targets. Targets are based on the original work by Otis for estimating the patient's minimal ventilatory work.18 If the patient's ventilatory pattern is not that determined to be ideal, the ventilator adjusts pressure level or rate to try to ensure that the ideal pattern is established.

This mode of ventilation is designed to ensure that an ideal ventilatory pattern is maintained and can be used in all phases of ventilatory support. It continually adjusts mechanical ventilation characteristics as the patient's status changes because it monitors the patient's lung mechanics on a breath-to-breath basis. Recently published data indicate that ASV works well in postoperative cardiac surgical patients19,20 and in patients under controlled ventilation.21 Recent lung model data indicate that during controlled ventilation ASV may provide better lung protection than a fixed tidal volume of 6 mL/kg.22 More data during assisted ventilation of patients with lung injury are needed to determine if spontaneously breathing patients will accept the ideal ventilatory pattern this mode tries to impose.

Automatic Tube Compensation

This mode of ventilation has been referred to as electronic extubation. Automatic tube compensation (ATC) is designed to provide sufficient mechanical ventilatory support during inspiration and sufficient decompression of the ventilatory circuit during expiration to maintain the tracheal pressure equal to the set PEEP level. The ventilator software accomplishes this by having the resistance properties of various endotracheal and tracheostomy tube sizes programmed into its memory and continually measuring gas flow. From these data, the pressure needed to overcome the resistance (resistance = change in pressure ÷ flow) of the endotracheal tube is continually calculated and applied.23

Upon activation of the ATC mode, the clinician must indicate the type and size of artificial airway and the desired percentage of automatic tube compensation (20%-100% on most ventilators).4,5 In addition, some ventilators only apply the ATC correction during inspiration, others during both inspiration and expiration. The precise clinical indications for ATC are still unclear. Because this mode responds only to the patient's demand, if demand is low, ventilatory support is low; that is, if the patient generates a low inspiratory effort then little pressure is applied, whereas if the patient generates great effort, the applied pressure is greater.24 ATC, like proportional-assist ventilation, does not force a ventilatory pattern, but is only designed to unload the flow-resistive properties of the artificial airway. ATC may be used on patients ready to wean from the ventilator. If the ATC pressure level is stable and remains at a low pressure (5-7 cm H2O), then extubation is indicated.

The use of expiratory ATC does raise some concerns when used in patients with airways obstruction. The decompression of the airway during early exhalation may precipitate greater airways collapse and obstruction. As with many of these newer modes of ventilation, additional research is required before clinical indications and contraindications can be clearly defined.

Proportional-Assist Ventilation

This mode of ventilation is similar to ATC but the software response is based on both the mechanics of the total respiratory system plus the resistive properties of the artificial airway; that is, the ventilator delivers a pressure assist in proportion to the patient's desired tidal volume (volume assist) and to the patient's instantaneous inspired flow (flow assist).25 The response of these 2 aspects of ventilatory assistance are automatically adjusted to meet changes in the patient's ongoing ventilatory demand. This algorithm is based on the law of motion as it applies to the respiratory system:

where Pmus is pressure generated by the respiratory muscles,25 Pappl is pressure applied by the ventilator, and E and R are elastic and resistance properties of the respiratory system. Assuming that E and R are linear during inspiration, the instantaneous flow and volume to be delivered are proportional to the resistive and elastic work of breathing, respectively. The ventilator continuously measures the instantaneous flow and volume, and periodically measures the E and R. Utilizing this information the ventilator software adjusts gas delivery accordingly by estimating Pmus and assisting Pmus in a proportional manner.26 Thus the patient is the determinant of the ventilatory pattern. Patients are given the freedom to select a ventilatory pattern that is rapid and shallow or slow and deep.26 If the patient desires a small tidal volume a low level of pressure is applied, and if a large tidal volume is desired a high pressure is applied. The ventilator does not force any control variable except the unloading of E and R in a proportional manner.

Proportional-assist ventilation (PAV) is only useful in patients with a stable ventilatory drive who choose an acceptable ventilatory pattern. Of concern is the inability of some mechanical ventilators to reassess E and R on an ongoing basis. With these ventilators the level of unloading may be inappropriate if respiratory system mechanics are dynamically changing. There are a number of clinical reports that compare the effects of PAV with those of pressure support.27-30 In most of these studies PAV is able to minimize patient effort when lung mechanics or demand changes because of the ventilator's ability to assess lung mechanics and the proportional unloading provided. PAV improves patient ventilator synchrony when compared to pressure support.30 PAV is also capable of reducing the sleep disturbances noted in the ICU. This occurs because PAV does not force a ventilatory pattern on the patient, allowing the patient to vary his or her minute volume and PaCO2 during sleep and decreasing the number of periodic apneas as compared with other modes of ventilatory support.31 In addition, the PAV mode has been shown to be useful during noninvasive ventilation.32,33

Neurally Adjusted Ventilatory Support

From a conceptual perspective, neurally adjusted ventilatory support (NAVA) is essentially the same as PAV except that PAV responds to changes in airway pressure and flow, whereas NAVA responds to changes in diaphragmatic EMG activity.34 However, in order for NAVA to function properly a specially designed nasogastric catheter with a 10-cm length of EMG electrodes must be in place.34 Both PAV and NAVA respond to patient effort, providing ventilatory support in a proportional manner. In either mode the clinician does not set pressure, volume, flow, or time. The only setting is the proportion of effort unloaded by the ventilator. In NAVA this is set as the number of centimeters of water applied per microvolt of diaphragmatic EMG activity.

NAVA, when compared to pressure support or pressure assist/control, results in a lower peak airway pressure and generally a smaller tidal volume and more rapid rate.35 NAVA also markedly reduces the level of dyssynchrony.36 NAVA is not affected by the size of the system leak or the presence of auto-PEEP because its response level is only affected by the level of EMG activity.

Airway Pressure Release Ventilation

This mode of ventilation is a combination of pressure-targeted SIMV and inverse-ratio pressure-control ventilation.37 As illustrated in Figure 81-11, airway pressure release ventilation (APRV) applies 2 levels of CPAP and allows spontaneous unassisted breathing at each CPAP level. There are 2 approaches to setting APRV based on the belief that spontaneous breathing should only occur at the high CPAP level38 or at both CPAP levels.39 If the clinician desires spontaneous breaths only at the high CPAP level, the time spent at the low CPAP level is kept brief to prevent complete exhalation, hence an inverse ratio between high and low CPAP levels. In this setting, the high CPAP level is set to maintain oxygenation and the lower level assists ventilation by periodically decreasing the lung volume.

The alternate approach is to set the low CPAP level to maintain oxygenation (as if setting a PEEP level) and the high CPAP level to provide ventilatory assistance, as in SIMV.39 In the latter approach, the time spent at low CPAP levels is sufficient to allow complete exhalation and spontaneous breathing at the lower CPAP levels. Proponents of APRV cite the salutary benefits of spontaneous unsupported breathing to improve ventilation/perfusion matching, increase cardiac output, and reduce the need for sedation.37-39 However, as noted in Figure 81-12, the effort associated with spontaneous breathing under these conditions can result in very high esophageal pressure swings and a high work of breathing for the patient, accounting in part for the greater cardiac output.40 In addition, dyssynchrony is common with APRV. Dyssynchrony occurs at the mandatory transition from high-to-low or low-to-high CPAP levels. If the patient is exhaling when the ventilator increases airway pressure or inhaling when the ventilator decreases airway pressure, dyssynchrony results. Although this mode of ventilation is most commonly used in the management of acute lung injury or the acute respiratory distress syndrome, it has not been shown to be more beneficial than the assist/control mode.39

Bilevel Pressure Ventilation

This is a modification of APRV. As noted in Figure 81-13, pressure support can be added to spontaneous breaths at either the high or low CPAP level or both.4 This reduces the patient's effort and esophageal pressure swings, but it also increases the mean intrathoracic pressure, negating somewhat the increased cardiac output and beneficial augmentation of ventilation/perfusion matching. In addition, the transition from high-to-low and low-to-high CPAP levels is coordinated with the patient's muscular effort (as in SIMV), markedly decreasing dyssynchrony. Like APRV, the bilevel ventilation mode is primarily used in the management of acute lung injury and the acute respiratory distress syndrome.

Patient–ventilator synchrony is a critical issue for every patient triggering the ventilator for a mechanically supported inspiration. A lack of synchrony increases ventilatory effort, respiratory rate, and the patient's work of breathing as the ventilator is not providing flow to match the patient's inspiratory demand. Pressure-targeted modes of ventilation are generally better than volume-targeted modes at reducing the likelihood of dyssynchrony, as pressure ventilation allows the ventilator to deliver a variable flow with each breath based on the patient's inspiratory demand.41 In addition, most ventilators now allow adjusting the slope of the increase of flow delivery (rise time) to ensure that gas delivery is matched to patient's inspiratory flow demand.5,42 Patient–ventilator synchrony is also affected by the inspiratory time. The ventilator's inspiratory time and the patient's inspiratory time should be equal. Trigger sensitivity should always be set to be as sensitive as possible without causing spontaneous autotriggering. In general, flow triggering is to be recommended over pressure triggering.43 A final factor affecting patient–ventilator synchrony is auto-PEEP,44 which normally is only a problem in patients with airways obstruction, although it is a potential concern in all patients. If the patient's inspiratory efforts fail to trigger the ventilator, the problem is usually auto-PEEP.45

Peak Flow and Flow Waveform

If a patient is spontaneously triggering the ventilator, peak flow delivery during volume ventilation should match the patient's inspiratory flow demand.46 Most adult patients with moderate to strong ventilatory demands require a peak flow of 80 L/min or greater. As clearly demonstrated by Marini et al (Fig. 81-14), if the peak flow does not meet the patient's inspiratory demand the work of breathing performed by the patient increases. In this setting the efficiency of the work may be greater than during spontaneous breathing, but the overall patient work may be similar.46 In volume ventilation there is always an indirect relationship between the work provided by the ventilator and the patient's work of breathing.

It is important to remember that if the patient is triggering the positive-pressure breaths, the work of breathing is shared between the patient and the ventilator. It is for this reason that patients originally receiving assisted ventilation demonstrate altered gas-delivery patterns after they are sedated to apnea. With the transition to controlled ventilation, the peak airway pressures usually increase during volume ventilation, and tidal volumes decrease in pressure ventilation.47 Because the patient is no longer performing a portion of the work of breathing, the work performed by the ventilator must increase.

Most ventilators during volume ventilation mode can deliver gas flow in a decelerating or square wave flow pattern. If the patient is triggering inspiration, we recommend a decelerating flow pattern, especially when a small tidal volume (VT) is being delivered.4,47 A decelerating flow pattern allows a high peak flow to be delivered, but also ensures that the inspiratory time can be adequately set.

In patients who are sedated and who are not triggering the ventilator, the choice of flow waveform is unimportant, and the setting of peak flow is dependent on the predetermined inspiratory time and tidal volume.

Inspiratory Time

In patients self-triggering every breath, the set inspiratory time should equal the patient's neuroinspiratory time.48 As illustrated in Figure 81-15, when the inspiratory time is decreased to equal the patient's desired inspiratory time and peak flow is increased to match the patient's demand, then the patient's work of breathing and effort correspondingly decrease. It is rare that a patient with a moderate to high ventilatory demand desires an inspiratory time greater than 1 second.48 In fact, many adults with moderate or high ventilatory demands desire an inspiratory time of between 0.6 and 0.9 seconds.47 Carefully matching the patient's inspiratory time with the ventilator's inspiratory time generally markedly improves patient–ventilator synchrony.

In patients who are receiving controlled mechanical ventilation, inspiratory time is set based on the tidal volume and the clinician's perceived optimal time of inspiration. In most settings, an inspiratory time of about 1.0 second is ideal. Some clinicians prefer to lengthen the inspiratory time in ARDS/acute lung injury (ALI), thereby inversing the usual inspiratory-to-expiratory (I:E) time ratio; however, there are no data to indicate that longer inspiratory times result in better oxygenation or better outcomes in these syndromes than an appropriately set PEEP level.49,50 Prolongation of the inspiratory time can have a minor positive effect of increasing CO2 elimination by allowing time for inspired gas to transfer from fast time constant lung units to adjacent slower time constant units ("pendelluft"). In patients with severe asthma, the inspiratory time may need to be increased to 1.2 to 1.5 seconds to ensure that ventilation passes beyond markedly obstructed airways.37 Note that in asthma the airways resistance is not only increased during exhalation but also during inspiration.

Tidal Volume

Over the past 15 years, emphasis has been placed on the relationship between elevated tidal volumes and ventilator-induced lung injury (VILI; see Ventilator-Induced Lung Injury section).51 Recent epidemiologic studies indicate that the most common tidal volume used to manage patients in acute respiratory failure is about 10 mL/kg PBW.52 However, it is also important to emphasize that normal, healthy relaxed mammals breathe at a tidal volume of about 6 to 7 mL/kg of PBW.53 There is no evidence indicating that tidal volumes above 10 mL/kg PBW are useful in the management of critically ill patients. In fact, evidence from a recent retrospective review indicates that the development of ALI is linked to large tidal volume ventilation (>10 mL/kg PBW).52,54 As a general rule, all critically ill patients should be ventilated with tidal volumes below 10 mL/kg PBW regardless of the cause of their acute ventilatory failure. Patients with ALI/ARDS ideally should be ventilated with a VT 4 to 8 mL/kg of PBW.55

Plateau Pressure

The plateau pressure (end-inspiratory equilibration pressure) is a function of the mean peak alveolar pressure and is a reflection of the transpulmonary pressure. Similar to VT, it is unknown if there is a specific plateau pressure that is safe. Because of the high shear forces associated with the ventilation of heterogeneously diseased lung, mechanical lung injury is theoretically possible, even with transpulmonary pressures below 30 cm H2O. However, most clinicians would agree that a plateau pressure less than 25 cm H2O is unlikely to produce an injurious level of alveolar overdistension, and that a plateau pressure in this range and a tidal volume below 10 mL/kg PBW will not produce VILI regardless of the cause of respiratory failure. However, as plateau pressure increases, the delivered tidal volume should be decreased, regardless of the cause of respiratory failure. VT should be kept at about 6 mL/kg PBW if the plateau pressure is 25 to 30 cm H2O and VT should be 4 to 6 mL/kg PBW if plateau pressures are greater than 30 cm H2O. Ideally, all patients should be managed with a plateau pressure of less than 30 cm H2O unless there is a marked decrease of their chest wall compliance (eg, burns, abdominal distension). If the chest wall compliance is decreased, the transpulmonary pressure is decreased and the risk of overdistension is reduced. In patients with stiff, noncompliant chest walls and abdomens because of sepsis, abdominal distension, fluid resuscitation, and the like, a plateau pressure greater than 30 cm H2O may be required and would not be associated with an increased risk of VILI.56 However, in all patients the plateau pressure should ideally be maintained below 40 cm H2O. Alternatively, some have advocated the use of an esophageal balloon to estimate pleural pressure and guide mechanical ventilation strategies to limit transpulmonary pressure.57

Auto-PEEP

Auto-PEEP, also known as occult PEEP, is defined as an alveolar pressure above the airway opening pressure. Auto-PEEP can be caused by flow limitation, dynamic hyperinflation, a high minute ventilation, expiratory muscle activity, and/or a high intra-abdominal pressure (Fig. 81-16).44,58 Occult auto-PEEP is a term that is used to describe those lung units with trapped gas that do not communicate with the airway following a prolonged end-expiratory pause.59 Occult auto-PEEP is likely a function of airway closure or occlusion secondary to mucous plugging, and can be difficult to clinically diagnose. The performance of an end-expiratory occlusion maneuver can be helpful in estimating auto-PEEP levels if the patient is not making expiratory muscle efforts (passive-controlled ventilation). However, the absence of auto-PEEP by this technique should not be completely reassuring, because occult auto-PEEP may still be present. In severe asthma, it is frequently better to monitor the level of plateau pressure, as the plateau pressure will increase or decrease as total auto-PEEP increases or decreases.59

Auto-PEEP is a major cause of dyssynchrony. As noted in Figure 81-17, when auto-PEEP is present, patients frequently cannot generate a sufficient inspiratory effort to decompress the auto-PEEP and trigger a breath.45,60 Whenever the patient's respiratory rate exceeds the ventilator's response rate, this problem is almost always caused by auto-PEEP.

When auto-PEEP is caused by dynamic airflow limitation of exhalation volume, as in chronic obstructive pulmonary disease, the application of applied PEEP reduces the patient's effort to trigger the ventilator45; that is, applied PEEP can be titrated up to approximately 80% of the measured auto-PEEP without increasing the overall level of air trapping. Applying PEEP in the presence of auto-PEEP caused by dynamic airflow limitation increases central airway pressure without increasing auto-PEEP, so that the pressure change needed to trigger the ventilator is reduced, thereby increasing patient–ventilator synchrony, that is, decreasing the pressure difference across the obstruction. From a practical perspective, because auto-PEEP is difficult to measure in the spontaneously breathing patient, the applied PEEP is increased in 1 to 2 cm H2O steps until every inspiratory effort made by the patient triggers the ventilator. In some patients, this may require more than 10 cm H2O of applied PEEP. During controlled ventilation, a controversy exists over the benefit of applying PEEP to balance the auto-PEEP, because no spontaneous ventilatory efforts are present.

Target blood gas tensions during mechanical ventilation are theoretically the normal range for the individual patient. This is usually a PaCO2 of about 35 to 45 mm Hg for all patients except those with chronic CO2 retention. In patients with chronically elevated CO2 levels, the PaCO2 should be maintained at the patient's normal level up to 90 mm Hg.1 If the arterial pH and PCO2 are reset in these patients to "normal" values, the patients may become impossible to wean from ventilatory support, because during spontaneous unassisted breathing they will maintain CO2 levels at their baseline, which now results in an acute acidosis. In patients with ARDS, permissive hypercapnia may be unavoidable, as the effect of increasing the level of ventilation to reduce CO2 levels may produce lung injury.

One factor limiting an increase in PaCO2 with a permissive hypercapnic ventilatory strategy is the acidosis induced by the elevated CO2.61 Most healthy young adults can tolerate a pH as low as 7.20 without adverse effect.62 In fact, in acute lung injury or asthma, the increased cardiac output induced by the acidosis may be beneficial. Some clinicians have proposed that the acidosis associated with permissive hypercapnia protects the cell from hypoxemic injury.63 In older or cardiovascularly unstable patients, the level of acceptable acidosis must be individually gauged based on the response of the patient's cardiovascular system. In all patients the more gradual the permissive hypercapnia is allowed to develop, the greater the likelihood it will be well tolerated.

Tissue oxygenation is dependent on appropriate blood levels of oxygen and cardiac output. Assuming a normal cardiac output, PaO2 values in critically ill patients need only to be maintained higher than 60 mm Hg, whereas for some patients, because of a lack of alternatives, a PO2 of higher than 50 mm Hg may be acceptable.1 There is no value in most patients to maintaining a PaO2 of 100 mm Hg or higher.

Although the application of mechanical ventilation can be lifesaving, it has become increasingly clear that the mechanical ventilator can induce lung injury. Laboratory data have defined 2 primary settings in which the mechanical ventilator can induce lung injury, by producing overdistension of the lung at end inspiration and by collapse and reexpansion of unstable lung regions during ventilation.

Overdistension

Figure 81-18 depicts the impact of a large tidal volume without PEEP on a rat lung as compared to a rat lung ventilated with a small tidal volume and one with 10 cm H2O PEEP and a large tidal volume.64 One hour of ventilation with a large tidal volume at zero PEEP caused marked hemorrhagic edema of the lung. This type of injury was termed volutrauma by Dreyfuss and Saumon because it is believed that the injury is primarily caused by an excessive tidal volume.65 However, as described above, an excessive tidal volume is accompanied by an excessive transpulmonary pressure, making the distinctions between pressure and volume injury somewhat semantic. VILI is defined as an increase in the permeability of the alveolar capillary membrane, the development of pulmonary edema, the accumulation of neutrophils and protein within the lung parenchyma and airway, the disruption of surfactant production, and a decreased lung compliance.65 VILI is indistinguishable from human ARDS and can be produced in rats within minutes by large tidal volume ventilation.66 Indeed, VILI caused by inflation with large tidal volumes has been demonstrated in many small- and large-animal models. Parker et al,67 in isolated perfused dog lungs, demonstrated that ventilatory pressures greater than 30 cm H2O peak alveolar pressure resulted in VILI. Hernandez et al,68 Egan,69 and Kolobow et al70 also demonstrated similar injuries in large-animal models. Ventilation of many species with large tidal volumes results in the development of VILI and an increase in the severity of preexisting lung injury.

The single factor most responsible for the development of VILI appears to be the transpulmonary pressure (pressure inside minus pressure outside the lung or alveolar minus pleural pressure).66 In a now classic experiment, Dreyfuss et al66 demonstrated that rats with strapped chest walls were protected from VILI as compared with those ventilated with the same peak pressure but without chest wall binding. In fact the lung injury was similar regardless of the use of positive- or negative-pressure ventilation, provided high transpulmonary pressures were established. As a result, we reason that higher ventilating pressures can be used in the presence of a decreased chest wall compliance without the development of VILI, as the reduced chest wall compliance decreases transpulmonary pressure at peak airway pressure and results in a lower delivered tidal volume.

Recruitment/Derecruitment Injury: "Atelectrauma"

As noted in Figure 81-18, the application of 10 cm H2O PEEP in spite of a peak airway pressure of 45 cm H2O decreased the level of VILI as compared with similar animals ventilated with the same peak pressure but at zero PEEP.64 Only mild interstitial edema developed in animals ventilated at 10 cm H2O PEEP. Similar findings have been reported by Corbridge et al in dogs with hydrochloric acid–induced lung injury.71 Dogs ventilated with 12.5 cm H2O PEEP and a VT of 15 mL/kg after lung injury had far less pulmonary edema at autopsy than dogs ventilated with 3.2 cm H2O PEEP and 30 mL/kg tidal volumes, in spite of the fact that peak alveolar pressure (33 cm H2O) was the same for both groups during the 5 hours of ventilation.

Biotrauma

Slutsky and Tremblay72 have referred to the activation of pulmonary and systemic inflammatory mediators by an inappropriate ventilatory pattern as biotrauma. Increasing data from both animal and clinical studies indicate that a traumatic ventilatory strategy can augment the level of the pulmonary and systemic inflammatory response. Tremblay et al,73 in an ex vivo healthy and injured rat lung model, demonstrated that both pro- and anti-inflammatory mediators are activated by ventilatory patterns associated with high peak alveolar pressure and zero PEEP. For ventilated animals in which PEEP was set greater than Pflex (lower inflection point on the respiratory system pressure-volume curve) and overdistension was avoided, minimal mediator activation was observed. Similar data were reported by Bethmann et al in an ex vivo perfused mouse lung model.74 Hyperventilation at 2.5 times the normal transpulmonary pressure using either positive- or negative-pressure ventilation resulted in a 1.75-fold increased expression of tumor necrosis factor α (TNF-α) and interleukin-6 messenger RNA (mRNA). Imai et al75 and Takata et al,76 studying rats, noted greater TNF-α mRNA expression with conventional ventilation at low peak pressure (30 cm H2O) and low PEEP (5 cm H2O) as compared to high-frequency oscillation at a mean airway pressure of 15 cm H2O (same as with conventional ventilation). Ranieri et al showed the same effect of ventilatory pattern on pulmonary lavage and plasma TNF-α level in ARDS patients.77 In a randomized comparison, patients ventilated with PEEP set greater than Pflex and peak alveolar pressure kept below the upper inflection point of the pressure-volume curve had reduced TNF-α levels in both plasma and lung lavage fluid after 36 hours. Patients randomized to PEEP based on oxygenation with VT set to produce eucapnia increased their TNF-α levels after 36 hours. In addition, the ARDSnet45 demonstrated a lower systemic proinflammatory mediator response in a low tidal volume group of patients as compared to a high tidal volume group. Most recently, Chu et al78 demonstrated that lung maintained at a high lung volume either by CPAP or ventilated at the same high peak pressure resulted in the same marked activation of inflammatory mediators.

Based on these data both Slutsky and Tremblay72 and Dreyfuss and Saumon79 proposed that the lung is the "engine" that drives the development of multisystem organ failure in ARDS patients. The additional lung injury resulting from inappropriate mechanical ventilation causes the movement of inflammatory mediators into the systemic circulation, increasing the systemic production of mediators influencing and damaging the function of distal organs and tissues.

As a result, it seems reasonable to conclude, as discussed by Marini,56 that "excessive mechanical forces produce lung damage by at least three mechanisms: (a) the signaling of proinflammatory mediator release by mechanical stress, (b) trauma to the epithelium of the lung parenchyma by repetitive opening at high pressures and subsequent closing with each breath, and (c) physical stress on the parenchyma by high peak alveolar pressures." Marini also points out that there is some evidence that high ventilatory frequency80 (higher frequency, greater injury), high pre- and postalveolar microvascular pressures,81 decreased surfactant production,82 supine body position,83 high body temperature,84 and immune suppression85 all enhance the development and severity of VILI.

As already discussed, a primary aim of ventilatory support of the surgical patient is to achieve "lung protection," that is, to ensure that the process of mechanical ventilation does not amplify the extent of lung injury. Care should always be exercised to ensure that the plateau pressure and tidal volume do not result in overdistension. In general, based on current information the tidal volume should be maintained less than 10 mL/kg in all acutely ill patients or patients undergoing anesthesia. In ARDS, as presented in the previous section, it is certain that smaller tidal volumes are required. At least 3 clinical trials have demonstrated a benefit of ventilating at 6 to 7 mL/kg (PBW) tidal volumes in ARDS.55,86,87 In conjunction with these small tidal volumes, the plateau pressures should be maintained below 30 cm H2O unless the chest wall is stiff.56 However, the lower the plateau pressure, the less likely tidal volumes of 8 to 9 mL/kg will present concern.

Outside of the clinical setting of ARDS the data are less clear; however, Hubmayr et al performed a retrospective analysis of data on mechanical ventilation in the ICU44 and intraoperatively88 at the Mayo Clinic, and on data from an international ventilation survey89 indicate that small tidal volume (<10 mL/kg) ventilation should be the norm used in most critically ill patients. In 2 of these analyses54,89 they found that patients in the ICU had a greater probability of developing acute lung injury when ventilated with large tidal volumes (≥10 mL/kg). In their most recent review88 they demonstrated that patients undergoing pneumonectomy at the Mayo Clinic who were ventilated intraoperatively with larger tidal volumes (median: 8.3 mL/kg) versus smaller tidal volumes (median: 6.7 mL/kg, P < 0.001 groups differ) were more likely to develop postoperative respiratory failure and require ventilatory support for longer than 48 hours postoperatively.

In general, rapid respiratory rates (20-40 breaths/min) are needed to maintain gas exchange at low tidal volumes. The smaller the tidal volume and the more severe the lung injury, the greater the probability that patients will require permissive hypercapnia.54,89,90

The most appropriate method of setting PEEP in ARDS is still highly controversial.55,76,77,90 However, there is no question that PEEP is necessary in lung injury to prevent derecruitment. The ARDSnet55,90 recommends the use of a PEEP/FIO2 table for setting PEEP, but these tables do not take into consideration the patient's lung mechanics. Incremental PEEP trials have been used for years to provide a reasonable estimate of the necessary PEEP for many patients91 but may not be the best approach to setting PEEP in lung injury. Three prospective randomized trials have demonstrated positive outcomes when PEEP was set at Pflex + 2 cm H2O.77,86,87 In 2 studies,86,87 mortality was decreased in the group ventilated with PEEP at Pflex + 2 cm H2O along with a small delivered tidal volume, and in one, pulmonary and systemic inflammatory mediators were attenuated when this ventilatory strategy was used.77 Most recently, a number of meta-analyses of the 6 trials77,86,87,90,92,93 that evaluated different approaches to setting PEEP indicated that high PEEP in ARDS does reduce mortality.94,95 In general, in ARDS patients a medium PEEP of about 15 cm H2O is needed with the usual range from 10 to 20 cm H2O of PEEP. However, some patients require greater than 20 cm H2O PEEP.

In severe lung injury, the method of setting PEEP that appears best is a decremental PEEP trial following a lung recruitment maneuver.96 That is, PEEP is titrated downward from a level higher than is needed to determine the minimum level that maintains the benefit of the lung recruitment maneuver.

A number of groups have demonstrated the ability of a lung recruitment maneuver to open the lung in ARDS/ALI patients and the ability of a decremental PEEP trial to keep the lung open.97-99 However, no outcome benefit has been attributed to lung recruitment maneuvers. The 2 lung recruitment approaches that have received the most study are the use of a sustained CPAP level of 40 to 50 cm H2O for 30 to 40 seconds97,99 and the use of pressure-controlled ventilation at a peak pressure of 40 to 50 cm H2O along with PEEP of 20 to 30 cm H2O for 1 to 3 minutes.98 Lung recruitment maneuvers can open the lung of patients with lung injury and should be applied in the first few days after lung injury, but only in patients who are hemodynamically stable with appropriate fluid resuscitation without the presence of, or an increased likelihood of developing, barotrauma.

An alternate approach to setting PEEP in patients with severe lung injury was recently reported by Talmor et al.57 They randomized patients to be managed with the ARDSnet protocol PEEP table55 or a table based on esophageal pressure. Their goal when setting PEEP was based on esophageal pressure. PEEP was increased to ensure that end-expiratory esophageal pressure was not negative and as a result reducing lung collapse at end expiration. They demonstrated marked physiologic improvement and a trend toward improved mortality; however, the study was underpowered to identify a mortality benefit. Additional data from multicenter studies are needed before this approach can be recommended.

Various approaches to high-frequency ventilation (HFV) have been used over the last 30 years to manage patients intraoperatively, as well as to ventilate critically ill ICU patients. Although these techniques have demonstrated their ability to ventilate in both of these settings, limited beneficial outcome data regarding any of these techniques are available. The 3 common methods reported are high-frequency jet ventilation (HFJV), high-frequency percussive ventilation (HFPV), and high-frequency oscillatory ventilation (HFOV).

High-Frequency Jet Ventilation

HFJV is the approach to HFV that has been used most commonly intraoperatively.100,101 With this technique, gas under high pressure is injected via a small-bore nozzle placed within the endotracheal tube or in the airway itself. The velocity of the gas delivery stream creates a "jet drag effect" entraining secondary gases. This technique can provide ventilatory support with or without a sealed or closed airway. As a result, it has been used successfully during upper-airway surgery, tracheal reconstruction, unilateral surgical procedures, and other procedures in which the airway is not sealed.100,101 In addition, simple HFJV systems have been used to provide emergency ventilatory support in settings where immediate access to the airway via the pharynx is impossible.102 This is accomplished by placing the injector through the cricothyroid membrane; however, an escape pathway for injected gases via the upper airway must be present.

High-Frequency Oscillatory Ventilation

HFOV is the most commonly used high-frequency technique in the ICU. However, most of the successful applications of this ventilatory mode have been in neonates.103 Randomized controlled trials evaluating the use of HFOV during ARDS/ALI in pediatric104 and adult105-107 populations have been negative, and HFOV is considered equivalent to conventional ventilation in these settings. With HFOV a high mean airway pressure is established because gas delivery is achieved by oscillating about the mean airway pressure. Tidal volumes as low as 1 mL/kg are delivered at high frequencies (8-12 Hz) in neonates. In adults, generally larger tidal volumes (2-4 mL/kg) are delivered at lower frequencies of 4 to 8 Hz. All clinical approaches to HFV require the use of ventilators specially designed to provide HFV.

A recent meta-analysis of the adult and pediatric randomized controlled trials indicates a mortality benefit from the use of HFO.108 However, this meta-analysis failed to recognize and discuss the conventional ventilation arm in all of the studies that were evaluated. In none of these studies was conventional ventilation provided in what is considered an ideal lung-protective approach.

High-Frequency Percussive Ventilation

HFPV is a unique form of ventilation that combines the effects of conventional mechanical ventilation with HFO.91 That is, a typical pressure-targeted breath is delivered but during the breath, oscillations are provided. Oscillations can also occur during the expiratory phase. The goal is to increase the distribution of ventilation and mobilization of secretions by adding the oscillations to conventional ventilation. Most of the data on HFPV have been accumulated in burn or trauma patients.109,110 As with HFO this approach to managing critically ill patients is effective, but thus far no benefit over conventional mechanical ventilation has been demonstrated.111

Prone positioning of critically ill patients has been shown to improve oxygenation in a number of case series.112 The reasons indicated for this improvement are lung recruitment by removing the weight of the heart from the lungs and the positioning of two-thirds of the lung in nondependent regions, drainage of secretions, and better ventilation/perfusion matching. However, the primary concern with proning patients is the need to routinely reposition patients to the supine position. With turning the patient a number of adverse events have been identified including tracheal extubation, loss of vascular access, hemodynamic instability, and arrhythmias. A number of randomized controlled trials have compared prone positioning to continued supine positioning.113-115 None of these trials were positive; overall mortality is almost exactly similar regardless of position. However, a recent meta-analysis indicates a mortality benefit for the sickest of patients. Prone positioning seems to benefit those patients with a PaO2/FIO2 of less than 100 mm Hg.116 As a result it is considered a last line of therapy for treating severe hypoxemia.

Normal Preoperative Pulmonary Function

Surgical procedures that require general anesthesia and invade the thorax or abdominal cavity can result in impaired ventilatory function. As a result, well over 50% of cardiac and thoracic surgical patients show postoperative radiographic evidence of atelectasis.94 In patients with normal preoperative pulmonary function, this normally does not result in any compromise. But in patients with preexisting pulmonary disease, acute respiratory failure is more common.117

Initial Management

In patients without prior pulmonary disease, ventilatory management is normally uneventful. Pressure or volume assist/control is preferred with tidal volumes as large as the patient desires (6-10 mL/kg PBW) because lung function is normal and the period of total mechanical ventilation time is usually brief.37 Five cm H2O PEEP is applied to maintain the functional residual capacity, and the FIO2 and rate are adjusted to maintain adequate gas exchange.

Maintenance

Ventilatory management is short term in most patients. The FIO2 is titrated to maintain PaO2 above 70 mm Hg, and the rate or tidal volume is adjusted to maintain baseline PaCO2. Patients can frequently be rapidly transitioned to pressure support and extubated quickly.

Chronic Obstructive Pulmonary Disease

Pathophysiology

Postoperative chronic obstructive pulmonary disease (COPD) patients are commonly encountered in the ICU and are challenging to ventilate. Of primary concern is patient–ventilator synchrony.1 The single most significant factor affecting synchrony in these patients is air trapping and auto-PEEP.58,60 In COPD patients, auto-PEEP is caused by dynamic airway compression; that is, unstable airways dilate during positive-pressure inspiration but are compressed by the increased intrathoracic pressure during exhalation. Because auto-PEEP occurs distal to airways obstruction, the airway pressure proximal to the obstruction at end exhalation is normally equal to baseline ventilator circuit pressure. This means that during spontaneous inspiration the patient must first decompress the auto-PEEP level, and then trigger the ventilator. This results in a marked increase in patient effort to trigger the ventilator and dyssynchrony. Many of the patient's inspiratory efforts may not produce sufficient negative force to trigger the ventilator. The patient's actual respiratory rate is then greater than the ventilator's response rate. This increases ventilatory drive, patient respiratory rate, and the patient's overall ventilatory effort. The dyssynchrony associated with auto-PEEP can best be managed at the bedside by applying PEEP as discussed earlier (see Ventilation Settings and Patient–Ventilator Synchrony—Auto-PEEP section).45

A second major concern is ensuring that the ventilator's gas delivery rate meets the patient's inspiratory demand. COPD patients who are spontaneously triggering the ventilator have their ventilatory demands best supplied with pressure, as opposed to volume ventilation.47 However, in either case gas delivery (rise time during pressure ventilation, and peak flow and flow waveform during volume ventilation) should be set to match inspiratory demand. Inspiratory time should coincide with the patient's desired inspiratory time. In assist/control ventilation, the inspiratory time is generally set from 0.6 to 1.0 seconds, and with pressure-support ventilation, the inspiratory termination criteria should be adjusted to ensure a synchronous ending of inspiration.48

COPD patients require ventilatory support because of their inability to perform the work needed to maintain normal ("for them") gas exchange. As a result, the ventilator should ensure adequate rest but should always allow the patient to trigger the ventilator. In general, it is advisable not to use ventilatory modes that do not support every inspiratory effort, because they do not provide sufficient rest for the patient. Thus synchronized intermittent mandatory ventilation, airway pressure-release ventilation, and bilevel ventilation should not be used in COPD.

Initial Setting

Initially, intubated, mechanically ventilated COPD patients should be managed with pressure support or assist/control (volume or pressure with pressure preferred). The delivered tidal volume should be 6 to 10 mL/kg PBW, dependent on plateau pressure and patient demand. Most COPD patients can be ventilated with a plateau pressure of 25 cm H2O or less. The respiratory rate is patient determined; however, in assist/control, a backup rate sufficient to maintain the appropriate PaCO2 should be set. In volume ventilation, the peak flow should be set at 80 L/min or higher with a decelerating waveform. With pressure ventilation, the rise time and inspiratory termination criteria (pressure support only) should be properly set. Inspiratory time is set equal to the patient's neuroinspiratory time (0.6-1.0 s) and the PEEP level adjusted to ensure that all of the patient's inspiratory efforts trigger the ventilator.48

Maintenance

In many COPD patients the key medical management issue is to provide rest for the patient. After 48 to 72 hours of ventilatory support, many patients are ready for ventilator discontinuation. In all cases, the focus should be on providing a level of ventilatory support consistent with minimal overall ventilatory effort to ensure recovery from ventilatory muscle dysfunction. As patients recover their PEEP level, pressure support or pressure assist/control level and FIO2 can be decreased, provided that ventilation targets can be met without markedly increasing the patient's work of breathing.

Patients requiring long-term ventilation should be considered for extensive rehabilitation, including generalized muscle retraining. Patients should be weaned with a spontaneous breathing trial, and most COPD patients can be extubated after a successful 30- to 120-minute trial. However, those requiring long-term ventilatory support (2-4 wk) should be given a tracheotomy and normally require successful 12- to 16-hour spontaneous breathing periods before being allowed to breathe without ventilatory support during the night.

For COPD patients who are ventilated for short periods, who meet all the criteria for weaning, and who clinically seem ready for ventilator discontinuation but continue to fail spontaneous breathing trials, carefully consider extubation to NPPV.118-120 Some COPD patients can be successfully transitioned to ventilator independence using NPPV. (See Weaning and Noninvasive Positive Pressure section.)

Acute Respiratory Distress Syndrome

Pathophysiologic Issues

ARDS and ALI are acute lung diseases characterized by atelectasis, edema, decreased compliance, and severe arterial hypoxemia. As a result, an important issue during mechanical ventilation is to avoid inducing greater lung injury either by end-inspiratory overdistension or by the repetitive opening and closing of unstable lung units. Other important issues are the recruitment of lung, the reversal of atelectasis, and the assurance of adequate systemic oxygenation and CO2 exchange while avoiding pulmonary oxygen toxicity.

The end-inspiratory plateau pressure should be less than 30 cm H2O to avoid lung injury unless chest wall compliance is decreased. This means that for most ARDS/ALI patients the tidal volume should be set to 4 to 8 mL/kg PBW.55,86,87 However, if the plateau pressure can be maintained below 25 cm H2O and the patient desires a tidal volume up to 10 mL/kg PBW, it may be more acceptable to meet the patient's ventilatory demands than to force the patient to receive a low tidal volume (which may require intravenous [IV] sedation to apnea).

Some of the atelectasis in ARDS/ALI patients is recruitable by the use of short-term application of a "high" airway pressure (a recruitment maneuver).97-99 These maneuvers work best on the initial day of identification of ARDS/ALI. The longer the patient has ARDS/ALI, the less likely that recruitment maneuvers will succeed and the greater the likelihood of hemodynamic compromise. Following a recruitment maneuver, sufficient levels of PEEP should be applied to maintain the benefits of lung recruitment.

Initial Settings

Assist/control mode (Table 81-5), either pressure or volume, can be used with pressure targeting recommended if the patient is triggering each breath and has a variable ventilatory demand. A maximum plateau pressure of 30 cm H2O should be set at a tidal volume of 4 to 8 mL/kg PBW unless the plateau pressure is less than 25 cm H2O. The rate should be set to ensure adequate CO2 elimination. Respiratory rates can be set as high as 40 breaths/min. The rate-limiting factor is the development of auto-PEEP. In most patients, a normal PaCO2 of 35 to 50 mm Hg can be established by adjusting the rate. In some patients with a low compliance, permissive hypercapnia may be necessary. Tolerance of hypercapnia depends on the pH; it is better to allow the PaCO2 to gradually increase over a day or 2 to prevent the development of a marked respiratory acidosis.

FIO2 should initially be set at 1.0 and PEEP at 10 to 12 cm H2O until hemodynamic stability is achieved. Once the patient is hemodynamically stable, a lung recruitment maneuver should be considered, using either a CPAP or pressure-assist/control approach as discussed earlier (in the section Lung-Protective Ventilation). Before recruitment the patient should be stable hemodynamically, sedated to near apnea, and breathing 100% oxygen. Careful monitoring of the hemodynamic response and oxygenation level during the recruitment maneuver should be performed, and the maneuver should be stopped if hemodynamic compromise or desaturation is observed. Following the maneuver, the minimum PEEP level maintaining the oxygenation benefit of the recruitment maneuver should be applied. This is best determined by a decremental PEEP trial.96 First, the PEEP level is set at 20 cm H2O and the dynamic compliance determined. Then the PEEP level is decreased by 2 cm H2O and after the compliance has stabilized, the PEEP is decreased again by 2 cm H2O. In general it takes about 3 to 5 minutes for the compliance to stabilize. The decremental trail is stopped once the PEEP level associated with the best compliance can be identified. The PEEP is then set 2 cm H2O above this level because the best compliance PEEP underestimates the best oxygenation PEEP by about 2 cm H2O. At this point another recruitment maneuver is performed, after which PEEP is set at the identified level plus 2 cm H2O. If another method of setting PEEP is used and the oxygenation benefit of the recruitment maneuver is lost over a brief period, the PEEP level is set too low and should be increased. Many ARDS patients will require a PEEP level of 10 to 20 cm H2O and ALI patients often require a PEEP level of 8 to 15 cm H2O.

Regardless of the use of pressure or volume mode ventilation, the inspiratory time should be set at about 0.6 to 1.0 seconds, depending on the patient's ventilatory demand. In patients who are breathing spontaneously, the inspiratory time should be set equal to the patient's neuroinspiratory time. In volume ventilation, adequate peak flow should be provided to meet the patient's ventilatory demand (≥80 L/min) using a decelerating waveform.

A recently published report by Papazian et al121 demonstrated that paralysis as compared to standard care for the first 48 hours of ventilation in severe ARDS patients resulted in decreased mortality. It has been speculated that this was a result of decreased oxygen consumption and lack of fighting the ventilator that may have precipitated atelectasis and increased the likelihood of developing ventilator-associated pneumonia.122 In addition, paralysis was obtained with cis-atracurium, which has been shown to have anti-inflammatory properties.123 However, before this approach can be universally recommended additional studies are needed.

Maintenance

The FIO2 should be decreased before PEEP when oxygenation improves. PEEP generally should not be decreased until the FIO2 is less than 0.50. If the PaO2 decreases as PEEP is lowered, the prior PEEP level should be reestablished, and the FIO2 should not be increased. When this occurs, decreasing PEEP resulted in derecruitment of lung, indicating that the higher PEEP level is needed.

Once the patient's ventilatory drive has normalized and spontaneous triggering of the ventilator can be maintained, patients with ARDS/ALI can usually be maintained with pressure-support ventilation. As with assist/control, the peak pressure should be maintained as low as possible and always less than 30 cm H2O.

In patients with reduced chest wall compliances as a result of abdominal sepsis, ascites, obesity, thoracic deformity, or massive fluid resuscitation following trauma, a plateau pressure greater than 30 cm H2O may be necessary for adequate ventilation. This is acceptable, as the transpulmonary pressure is lower with the stiff chest wall preventing overdistension of the lung. However, when the plateau pressure is greater than 30 cm H2O, the tidal volume should be reduced to less than 6 mL/kg PBW.

In ARDS/ALI, care should be exercised to never disconnect the ventilator circuit. Whenever the circuit is disconnected from the airway, lung derecruitment occurs within seconds. Inline suction catheters should always be used, and the use of manual resuscitators (Ambu bags) should be avoided. Suctioning should only occur when the presence of secretions is observed, and when performed, the suction pressure should be regulated to 120 mm Hg and performed gently to avoid lung derecruitment.

If lung recruitment and an appropriate PEEP setting are achieved but the FIO2 remains greater than 0.60 to achieve adequate oxygenation, then consider prone positioning (if the PaO2/FIO2 is <100 mm Hg) of the patient. Although neither prone positioning nor lung recruitment have been shown to improve survival rates, both can recruit lung and may improve oxygenation. The recruitment of regions of atelectasis reduces the plateau pressure and FIO2, improves surfactant function, and decreases the risk of pneumonia. At this time, the use of high-frequency oscillation or airway pressure-release ventilation does not appear to improve outcome in ARDS, but both modes can provide adequate management of ARDS/ALI patients. Neither mode has been shown to be superior to conventional assist/control ventilation.

Unilateral Lung Disease

Patients with unilateral lung disease present a particularly difficult challenge for mechanical ventilation. The postoperative patient after single-lung transplantation illustrates these problems. One lung has relatively normal pulmonary mechanics (the transplanted lung), whereas the native lung has mechanics reflecting either obstructive or restrictive disease. In these patients, the ventilator should be set to ensure maximum function of the native lung, as this lung will present the greatest ventilatory challenge. If the native lung has chronic obstruction, ventilate with moderate volume and slow rates. With pulmonary fibrosis of the native lung, a smaller VT and more rapid rate are indicated. In the case of pulmonary fibrosis, there is less concern about air trapping. However, peak alveolar pressure may be high because of the reduced lung compliance.

The greatest ventilatory challenge is the patient with a single-lung transplant where the native lung is obstructed and the transplanted lung has become stiff (low compliance) as a consequence of infection, rejection, or acute lung injury. In this setting, it is difficult to determine the ideal ventilator settings because of the differing pathology of each lung. Careful attention needs to be paid to 2 variables as adjustments are made: first, concern about peak alveolar pressure because of ventilator-imposed lung injury, damaging the transplanted lung and its bronchial anastomosis, and second, air trapping in the obstructed lung resulting in grossly compromised ventilation/perfusion ratios, overdistension, and the like. In this setting, permissive hypercapnia is often necessary with the final ventilator settings being a compromise between the conflicting needs of each lung.

Recent guidelines recommend the use of spontaneous breathing trials for weaning all patients from ventilatory support.124 These guidelines also recommend that protocols be developed to allow assessment for weaning from the first day of ventilatory support. Table 81-6 lists the criteria for initiation of a spontaneous breathing trial for any patient requiring ventilatory support. Table 81-7 lists guidelines to identify those patients having failed a trial of spontaneous breathing.

If a patient meets the guidelines in Table 81-6 by midnight, sedation should be adjusted to ensure spontaneous breathing by early morning. A 30- to 120-minute spontaneous breathing trial is performed, and if the patient does not fail the trial based on the criteria in Table 81-7, consider the patient for extubation.124 After short-term ventilatory support, most clinicians do not recommend measuring lung mechanics. Even the rapid shallow breathing index has not proven universally useful in identifying those patients who are ready for weaning. All of the current data indicate that the use of other approaches to weaning simply results in prolongation of the ventilatory support period.95 Patients failing weaning trials should be carefully assessed for the specific factors causing failure. Table 81-8 from Ely lists areas of concern that should be assessed in the patient falling multiple weaning trials.125 Finally, in some patients the use of NPPV may facilitate the process of transition to unsupported spontaneous breathing.118-120

Over the last 20 years increasing emphasis has been placed on the use of noninvasive (without tracheal intubation) positive pressure to support patients developing respiratory failure.126 Although much of the emphasis in such studies has focused on the patient with COPD in an acute exacerbation,127 a number of trials over the years have focused on the use of NPPV128 or CPAP for the management of postoperative patients.103-107,129-133

The primary mechanism producing arterial hypoxemia after surgery is impaired ventilation/perfusion matching as a result of the development of atelectasis.107 Atelectasis in this setting is a result of recumbent positioning, ventilation with high oxygen concentrations, diaphragmatic dysfunction, and incisional pain producing impaired secretion clearance.129 Several studies show that mask CPAP ventilation in this setting improves gas exchange, minimizes the development of atelectasis, and increases functional residual capacity.129-133 Most of these studies were case series129-133; the most recent study, however, was a randomized controlled trial.133 Squadrone et al133 demonstrated that mask CPAP at 7.5 cm H2O at 0.5 FIO2 for 6 hours in all patients following major elective abdominal surgery with a PO2/FIO2 less than 300 mm Hg prevented intubation compared to those receiving 50% oxygen by a Venturi mask (1% vs 10% intubation rate; P < 0.005). In addition, significant decreases in subsequent pneumonia and sepsis rates were observed.

In postoperative lung resection patients with hypoxemic respiratory failure, Auriant et al128 demonstrated that the use of NPPV versus standard care prevented intubation (P = 0.030) and decreased hospital mortality (P = 0.045). Although CPAP and NPPV are useful in treating postoperative respiratory failure, they should be cautiously applied, and tracheal intubation for patients not responding to these therapies should not be delayed.108 These concerns were recently illustrated by Delclaux et al.134 They randomized patients with hypoxemic respiratory failure to either CPAP or standard therapy. They found no difference in the subsequent intubation rate or the mortality rate, but a greater incidence of cardiac arrest and stress ulcers in those managed with CPAP.

A number of recent studies have focused on the use of NPPV in the weaning process. The most commonly studied patients are those with COPD. At least 2 groups have shown beneficial results with COPD patients who were failing weaning trials and who were extubated and ventilated with NPPV; that is, NPPV provided a bridge from invasive ventilation to spontaneous breathing in these patients failing weaning trials.118,119 However, it must be emphasized that these COPD patients were not postoperative patients.

The most encouraging use of NPPV during weaning are the results reported by Nava135 and Ferrer.110,136,137 In these studies patients passing their weaning trials and who were then extubated but at risk for failing extubation were studied. Those managed postextubation for 24 to 48 hours with NPPV had a significant decrease in the likelihood of their needing reintubation. The increased risks were defined as COPD, congestive heart failure, an Acute Physiology and Chronic Health Evaluation (APACHE) score above 12, age above 65 years, ineffective cough, and excessive secretions, 1 or more weaning failures, upper airway obstruction, or more than 1 comorbid condition. Of note, however, very few of these patients were postoperative patients.

There is considerable concern over the use of NPPV for the patient developing postextubation hypoxemic respiratory failure. Keenan et al138 demonstrated no benefit from the use of NPPV as compared to standard therapy in about 160 randomized patients. No differences were observed in the length of mechanical ventilation, hospital stay, or hospital mortality. Of even more concern was the increased mortality reported by Esteban et al139 in patients with postextubation hypoxemic respiratory failure who were managed with NPPV. These authors argue that the primary reason for the increased mortality was the delay in intubating those who were treated with NPPV (12 vs 2 h).

Based on the current data from these randomized controlled trials, NPPV is best used for patients passing weaning trials who are at an increased risk of developing respiratory failure as described by Nava135 and Ferrer.136,137

• Smith TC, Marini JJ. Impact of PEEP on lung mechanics and work of breathing in severe airflow obstruction. J Appl Phys. 1988;64:1488-1496.
• Dreyfuss D, Saumon G. Ventilator-induced lung injury: lessons from experimental studies. Am J Respir Crit Care Med. 1998;157:294-323.
• The Acute Respiratory Distress Syndrome Network. Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung injury and the acute respiratory distress syndrome. N Engl J Med. 2000;342:1301-1308.
• Villar J, Kacmarek RM, Perez-Mendez L, Aguirre-Jaime A, for the ARIES Network. A high positive end-expiratory pressure, low tidal volume ventilatory strategy improves outcome in persistent acute respiratory distress syndrome: a randomized, controlled trial. Crit Care Med. 2006;34:1311-1318.
• Fernandez-Perez ER, Keegan MT, Brown DR, Hubmayr RD, Gajic O. Intraoperative tidal volume as a risk factor for respiratory failure after pneumonectomy. Anesthesiology. 2006;105:14-18.
• Borges JB, Okamoto V, Gustavo M, et al. Reversibility of lung collapse and hypoxemia in early acute respiratory distress syndrome. Am J Respir Crit Care Med. 2006;174:268-278.
• Ferrer M, Esquinas A, Arancibia F, et al. Noninvasive ventilation during persistent weaning failure. Am J Respir Crit Care Med. 2003;168:70-76.