Chapter 32. Ventilator Waveforms: Clinical Interpretation

• Pressure and flow waveforms reveal a wealth of information regarding the patient's physiologic derangement.
• Distinguishing the contributions of resistive and elastic pressures allows tailoring and monitoring of therapy.
• Auto-PEEP should be sought in all mechanically ventilated patients.
• Ventilator waveforms show how adequately the physician has accommodated the ventilator to the patient.
• Patient effort confounds interpretation of pressures and flows.
• Attention to ventilator waveforms can improve the accuracy of hemodynamic interpretation.

Intensive care ventilators generate tidal ventilation by applying to the endotracheal tube or mask a pressure higher than the alveolar pressure. This is true whether the mode of ventilation is volume-preset (volume assist-control ventilation [ACV] and synchronized intermittent mandatory ventilation [SIMV]); pressure-preset (pressure-support ventilation [PSV] and pressure-control ventilation [PCV]); or more complex modes (pressure-regulated volume control [PRVC], proportional assist ventilation [PAV], airway pressure release ventilation [APRV], and volume-support ventilation [VSV]). The capability to display waveforms turns modern ventilators into sophisticated probes of patient respiratory mechanics and of patient-ventilator interaction. Respiratory system mechanics and waveform analysis should be integrated into routine ventilator management of the critically ill patient. The fundamental aims are to (1) determine the nature of the mechanical derangement of the respiratory system; (2)~assay the response to therapy; (3) reveal intrinsic positive end-expiratory pressure (auto-PEEP); and (4) determine the patient-ventilator interaction to guide adjustment of ventilator settings. In addition, respiratory muscle activity must be considered when measuring hemodynamic pressures such as the pulmonary artery occlusion pressure (pulmonary wedge pressure, Ppw) or the right atrial pressure (Pra), since these pressures are determined at end-expiration. The time point of end-expiration, as well as the presence of inspiratory and expiratory effort (both of which can greatly confound interpretation of hemodynamic pressures) can be readily discerned by analyzing ventilator waveforms.

It is easiest to derive clinically useful information about the patient's respiratory system when volume-preset modes such as ACV or SIMV are used. At least when the patient is passive, the pressure at the airway opening (Pao) and the pressure vs. time waveform reflect the mechanical properties of the respiratory system, yielding valuable clinical information. During pressure-preset modes, such as PSV and PCV, some information can be derived from the flow vs. time waveform, but this information is generally less readily interpreted than that obtained during volume-preset ventilation. Below we review the determinants of the pressure and flow versus time waveforms during volume-preset, then pressure-preset ventilation, including how to recognize and quantitate auto-PEEP as well as a method for using this information to adjust the ventilator. Volume-pressure loops are reviewed in terms of how they may aid management of the patient with acute lung injury (ALI) or acute respiratory distress syndrome (ARDS). The potentially confounding effect of patient effort on the pressure and flow waveforms is discussed. Finally, examples of problems revealed through careful interpretation of waveforms are presented.

Volume-Preset Modes

Gas is driven to and from the lung by a pressure difference between alveolus and airway opening. The majority of adult patients are ventilated, at least initially, with a volume-preset mode (i.e., ACV or intermittent mandatory ventilation [IMV]),1 allowing ready determination of the respiratory system mechanics. When a muscle-relaxed patient is mechanically ventilated at constant inspiratory flow, the inspiratory Pao consists of three components: one to drive gas across the inspiratory resistance, the second to expand the alveoli against the elastic recoil of the lungs and chest wall, and the third equal to the alveolar pressure present before inspiratory flow begins (PEEP or auto-PEEP) (Fig. 32-1).

The contributions of each of these three components can be found by inserting a 0.3- to 0.5-second end-inspiratory pause,2 briefly stopping flow and allowing the pressure to fall from its peak value (Ppeak) to a plateau pressure (Pplat), in order to quantitate the flow-related pressure, Pres (Pres = Ppeak − Pplat). At any point during passive inspiration, Pao reflects the sum of these three components, as follows:

where Pao is the airway opening pressure, Pres is the resistive pressure component, Pel (Pel = Pplat − Total PEEP) is the elastic pressure, Rrs is inspiratory resistance, ΔV is the increment in lung volume, Ers is elastance of the respiratory system, and total PEEP is applied PEEP or auto-PEEP, whichever is higher. Diagnostic and therapeutic information can be gleaned by distinguishing the individual components of the peak Pao (Ppeak) as follows. First, PEEP is set on the ventilator and this value can be used when auto-PEEP is absent. However, auto-PEEP is present in most ventilated critically ill patients,3 and methods for quantitating it are described below. The Ppeak can be apportioned between its two remaining components, Pres and Pel, by stopping flow (end-inspiratory pause) and allowing the Pres to fall to 0. When flow is 0, Pao drops to a lower Pplat. Then:

The final component (Pel = Pplat − Total PEEP) is proportional to the elastance of the respiratory system and the tidal volume.

At normal inspiratory flow rates in the range of 1 L/s, Pres is typically between 4 and 10 cm H2O. Elevated Pres is found with high inspiratory flow or increased inspiratory resistance. At constant flow, a rise in Pres may indicate, for example, increased bronchospasm or endotracheal tube obstruction. Conversely, falling Pres may correspond to a response to bronchodilators. Because the Pres depends on ventilator flow rate as well as inspiratory resistance, when interpreting its value one must be careful to take the set flow rate into consideration. The most dramatic example of potential error in this regard is when the inspiratory flow is set to a decelerating profile (Fig. 32-2).

Since Pel = Δ V × Ers, elevated Pel indicates excessive tidal volume or increased elastic recoil of the lungs or chest wall, as in pulmonary fibrosis, acute lung injury, or abdominal distention. Respiratory system static compliance (Crs) is the inverse of Ers:

which is normally about 70 mL/cm H2O. When the tidal volume is a typical 500 mL, Pel should be only about 7 cm H2O (500 mL/70 mL per cm H2O). Thus a ventilated healthy patient should have a Ppeak of roughly 17, consisting of Pres (5~cm H2O), Pel (7 cm H2O), and applied PEEP (5 cm H2O). Often the cause of ventilatory failure has not been determined by the time of endotracheal intubation. If the Ppeak is not increased in a passive ventilated patient, the physician should suspect impaired drive, neuromuscular weakness, or a transient, now resolved, problem (e.g., upper airway obstruction bypassed by the endotracheal tube) as the cause for ventilatory failure. When the Ppeak is high, partitioning its components into the resistive pressure (Pres), the elastic pressure (Pel), and PEEP can aid the physician to narrow the differential diagnosis (Fig. 32-3 and Table 32-1).

In addition, such analysis may allow therapy to be tailored specifically to the cause of ventilatory failure. For example, in a patient with COPD and congestive heart failure who fails extubation following colon resection, bronchodilators will not be helpful if Pres is normal and auto-PEEP is 0. Similarly, if auto-PEEP is greatly elevated, measures to decompress the abdomen are not likely to get the patient off of the ventilator.

Pressure-Preset Modes

The inspiratory pressure waveform during pressure-preset modes, such as PSV and PCV, reflects ventilator settings only and reveals nothing about the respiratory system physiology. These waveforms serve mostly to reveal the current ventilator settings as a snapshot (Fig. 32-4) or to demonstrate the impact of certain complex modes on ventilator actions (Fig. 32-5).

Expiratory Pressure

During either volume-preset or pressure-preset ventilation, analyzing the expiratory pressure [Paw(ex)] is substantially less useful than the inspiratory pressure, since Paw(ex) is largely determined by characteristics of the mechanical ventilator, not the patient.

where PEEP is the applied PEEP (not auto-PEEP), FlowE is expiratory flow rate, and Rexlimb is the resistance of the expiratory limb of the ventilator. It is important to realize that Paw(ex) does not reflect expiratory alveolar pressure or auto-PEEP, and relates to the patient's respiratory system only indirectly through the expiratory flow. Although some ventilators display inspiratory and expiratory pressure-volume plots, only the inspiratory segment gives useful information about the patient.

Expiratory Flow

Expiratory flow depends largely on patient features, such as end-inspiratory lung volume, lung elastic recoil, and characteristics of the airways, rather than ventilator settings. For this reason, expiratory waveforms can be analyzed without respect to mode of ventilation. Look again at Fig. 32-3. Notice the striking difference in the expiratory flow between these two patients, the first having airflow obstruction, the second restriction.

The most valuable information to come from the expiratory flow tracing is evidence suggesting airflow obstruction, signaled by low or prolonged expiratory flow, often with flow at end-expiration. In addition, there may be two distinct components to the expiratory flow decay, rather a single exponential one (Fig. 32-6), in patients with airflow obstruction.

Inspiratory Flow

Pressure-Preset Modes

It is more difficult to infer the mechanical properties of the respiratory system during pressure-preset ventilation than when using constant-flow volume-preset ventilation, since the Pao waveform becomes meaningless. Information regarding the combined respiratory resistance and elastance (or time constant) can be gained by examining the slope of the inspiratory flow waveform (Fig. 32-7). Assuming a passive patient, during PCV the flow falls throughout inspiration as the rising alveolar pressure reduces the driving pressure for flow (since the Pao is maintained constant by the ventilator). The rate of fall of the flow is related to how fast Palv rises, itself a function of the mechanical properties of the respiratory system. Because the slope of the flow decay cannot be measured readily at the bedside, and because the information it contains lumps features of elastance and resistance, it may be useful to turn patients from PCV to ACV periodically to determine the respiratory mechanics.

Flow may or may not terminate before end-expiration depending on the inspiratory time (TI) and the time constant of the respiratory system (Fig. 32-8). If inspiratory flow terminates in a passive patient, the peak alveolar pressure equals the ventilator inspiratory pressure (PI). Flow waveform displays facilitate adjustments of TI and respiratory rate as discussed below.

When patients are breathing spontaneously on PSV, thereby determining their own TI and rate, waveform analysis aids the identification of patient-ventilator asynchrony. This may be especially important during noninvasive ventilation (NIV). Common problems during NIV include failure of the ventilator to recognize the onset of patient inspiration (generally due to auto-PEEP, as described below) and the onset of expiration (most often due to mask leaks).

Volume-Preset Modes

During ACV and SIMV, flow is set by the physician either directly or indirectly through the choice of minute ventilation and rate. Flow may also be altered by changes in rise time, inspiratory plateau, inspiratory:expiratory ratios, and other settings, depending on the particular ventilator in use. Flow waveforms can reveal the effects on flow of other setting changes, as discussed below, and also whether the flow profile (square, decelerating, or sine) has been inadvertently changed.

Clues in the Waveforms

The waveform indications of increased respiratory resistance are: (1) increased Pres when an end-inspiratory pause has been set; (2) a high shoulder on the early portion of the Pao versus time tracing (Fig. 32-9); (3) low and prolonged expiratory flow, often with persistent flow at end-expiration (see Fig. 32-9); (4) the presence of two components to the expiratory waveform (indicating early airway collapse; see Fig. 32-6); and (5) scooping of the expiratory flow-volume curve (Fig. 32-10).

Determining Auto-PEEP

The auto-PEEP effect occurs when there is insufficient time for the respiratory system to return to functional residual capacity by end-expiration.4,5 Short expiratory times, high minute volumes, and increased expiratory resistance contribute to auto-PEEP, but all of these need not be present. Auto-PEEP is present in the majority of ventilated patients with asthma and COPD (and in many during spontaneous breathing6,7), but it is also seen in ARDS and other settings with high minute ventilation.3 In many regards auto-PEEP acts like PEEP to impede venous return, heighten the risk of barotrauma, and improve oxygenation. In addition, auto-PEEP increases the work of breathing and impairs the patient's ability to trigger the ventilator. For these reasons, it is imperative to monitor routinely the presence and amount of auto-PEEP in mechanically ventilated patients.

Auto-PEEP is present when the expiratory flow tracing reveals persistent end-expiratory flow (Fig. 32-11). Additionally, when the Pres is lower than the height of the early step change in the Pao waveform, auto-PEEP is likely to be present. Several methods for quantifying auto-PEEP are available, but the one typically used clinically is the end-expiratory port occlusion method.4 Several ventilators facilitate this determination by providing an expiratory pause switch. This method will not provide accurate estimation of auto-PEEP if there is a leak in the tubing or around the endotracheal tube cuff, there is gas flow into the circuit during expiration (as during continuous nebulization of bronchodilators), or the patient is not fully passive during the maneuver. In one survey of ventilated patients, quantitation of auto-PEEP was possible by the end-expiratory port occlusion technique in only one-third, because patient effort confounded the airway pressure.3 Serial measurement of auto-PEEP may give information regarding the obstructed patient's response to bronchodilator therapy (if minute ventilation is constant).8

Using PEEP to Ease Triggering

Auto-PEEP presents an inspiratory threshold load to the spontaneously breathing patient, as discussed in Chap. 36 and Fig. 36-4.9 The work of breathing due to this inspiratory threshold load is roughly equal in magnitude to the excessive resistive work of breathing in patients with COPD exacerbations,10 contributing to distress even when on the ventilator. Thus therapy should be directed at reducing auto-PEEP when it is present. Meanwhile, PEEP can be applied externally, greatly easing the effort required to trigger the breath. In intubated patients with acute-on-chronic respiratory failure, continuous positive airway pressure (CPAP) has been demonstrated to reduce the work of breathing by nearly 50%.11 In patients with COPD and acute respiratory failure, nasal CPAP immediately improves respiratory rate, sensation of dyspnea, and the partial pressure of carbon dioxide (PCO2).12 The amount of auto-PEEP is largely independent of the set PEEP, since the airways of obstructed patients behave more like Starling resistors than like ohmic resistors, much as the rate of flow of water over a waterfall is unrelated to how far the water will fall into the pool below (Fig. 32-12). As long as the set PEEP is not higher than roughly 75% of the auto-PEEP, there is little effect on Palv, expiratory flow, or the magnitude of auto-PEEP,13 although there are occasional exceptions.14

Effects of Therapy

A reduction in airways resistance or a response to bronchodilators may be signaled by (1) reduced Pres (and lower Ppeak), (2) reduced auto-PEEP,8 and (3) a more normal expiratory flow-volume curve. However, expiratory peak flow often does not increase with bronchodilators, because the lowered alveolar pressure (less auto-PEEP) reduces the driving pressure for exhalation.

Flow and pressure waveforms are generally easy to analyze when patients are fully passive. On the other hand, an active patient presents numerous challenges and pitfalls. The discussion above always assumed a passive patient, but this is often not the case.

Inspiratory Effort Preceding Machine Inspiration

The inspiratory threshold load presented by auto-PEEP sometimes leads to a striking delay between the initiation of the patient's inspiratory effort and the onset of machine inspiration, sometimes consisting of several hundred milliseconds (Fig. 32-13). This delay, which signals the presence of auto-PEEP, is often shortened markedly by the addition of externally applied PEEP, a feature that may aid the setting of PEEP in obstructed patients.

Effort during Machine Inspiration

It has long been known that patients perform inspiratory work throughout an assist control breath.15 With modern ventilator management shifted to greater patient effort, lower tidal volumes, less sedation, and only rare use of therapeutic paralysis, most patients exhibit effort on most ventilator breaths, no matter the mode of ventilation. This effort may not be obvious despite careful examination of the patient unless measures of intrathoracic pressure (esophageal, central venous, or wedge pressures) are available. Inspiratory effort may alter Pao (volume-preset breaths), inspiratory flow (pressure-preset breaths), or expiratory flow (any mode). Effort at the end of a volume-preset breath will affect the Ppeak and Pplat, making determination of respiratory system mechanics unreliable. The magnitude of this problem is illustrated in Fig. 32-14, where the degree of patient effort is hidden until therapeutic paralysis reveals it. Some patients have extremely high drive, despite being connected to a ventilator. They may desire inspiratory flow rates much higher than are typically ordered (often in excess of 100 L/min). Since inspiratory flow is set by the physician and ventilator during typical volume-preset ventilation, no amount of patient effort can raise the flow. The effect of this is that Pao may not become positive during the breath or may even become negative (since the Pao reflects the competition between the ventilator, tending to raise Pao, and the patient, tending to lower it). If Pao is not positive during inspiration, the ventilator is not doing work on the patient. That is, the patient's work of breathing would be no lower if the ventilator were disconnected. In extreme circumstances, the ventilator may even impede respiration. This represents a fundamental problem of ACV and SIMV: the greater the patient demand, the lower the Pao. One of the advantages of pressure-preset modes is that Pao is maintained, no matter how high the flow required, and these modes may be more comfortable for the patient with high drive (Fig. 32-15). It would even be sensible to augment pressure when the patient effort is high, and this rationale underlies PAV (see Chap. 36).

Clues to patient effort are often available from the airway pressure tracing when using a volume-preset mode, such as concavity of the rise in Pao, variability of Ppeak, and a dip in Pao before inspiration, indicating a triggering effort (Fig. 32-16). Since every breath of a passive patient should produce identical flow and pressure waveforms, an important clue to effort is simply a lack of uniformity of breaths.

Effort Persisting after Machine Inspiration

One of the paradoxes of mechanical ventilation is that the patient's desired inspiratory duration (neural TI) may bear no relationship to the ventilator inspiratory time (mechanical TI). Mechanical TI is set by the physician, either directly (as with PCV) or indirectly (as with ACV and SIMV), whereas neural TI depends on arterial blood gases, the gas exchange efficiency of the lung, the impact of sedatives, and other central nervous system and patient features. In the usual setting, the patient desires a longer TI than is set so that the patient's inspiratory effort persists, even while the ventilator has cycled to expiration. When neural and mechanical TIs are dissimilar, patients may be uncomfortable. Pressure-support ventilation goes far toward solving this problem by switching off when inspiratory flow falls, since reduced flow often correlates with the end of neural inspiration. In contrast, many patients ventilated with PCV or volume-preset modes rely on happenstance or sedatives to accommodate their neural TI to the machine TI (Fig. 32-17).

Postinspiratory effort can be recognized as loss of the usually rapid initial expiratory flow, as illustrated in Fig. 32-17, or by double- or triple-triggering. The patient shown in Fig. 32-18 was being ventilated with a lung-protective tidal volume (6 mL/kg ideal body weight [IBW]) for ARDS. Every breath actually consists of a double-triggered breath: exhaled total ventilation (VT) alternated between 2 mL/kg and 10~mL/kg, showing that this patient was probably not receiving lung-protective ventilation, despite the set tidal volume. Ventilator TI can be lengthened during volume-preset modes by increasing tidal volume (although this may conflict with lung-protective goals; see Chap. 37) or by reducing inspiratory flow (although this may prompt the patient to exert even more inspiratory effort).

A separate phenomenon is additional attempts to trigger the ventilator during expiration (Fig. 32-19). This is quite common, generally when there is auto-PEEP, and especially during PSV at high levels, and its clinical significance is not known.16 When ventilator-dependent patients were subjected to increasing degrees of ventilator assistance (and demonstrated reduced inspiratory pressure-time product), the rate of ineffective triggering rose even while the total respiratory rate fell.17 It is probably valuable to consider the impact of patient-ventilator asynchrony (PVA) in light of the patient's respiratory drive. When drive is high, PVA should be addressed by manipulation of the ventilator to improve the patient-ventilator interaction. If drive is low, however, PVA may simply indicate unloading of the respiratory system and no changes in ventilator settings are indicated.18

Expiratory Effort

Patients may recruit expiratory muscles during machine inspiration or expiration. Expiratory effort during machine inspiration may raise Pao during ACV or SIMV, even setting off the pressure alarm, and reduce tidal volume on any mode. Expiratory effort at end-inspiration occasionally raises Pplat artifactually (Fig. 32-20). For this reason, it is prudent to view the waveform whenever measuring plateau pressure in order to confirm that there is a true plateau.

A more common problem is expiratory muscle recruitment throughout expiration, even to end-expiration, as is often seen in patients with severe airflow obstruction. Measured values of auto-PEEP may be artifactually elevated by expiratory effort19 (as may hemodynamic pressures, as discussed below).

Ventilation Via Endotracheal Tube or Tracheostomy

Patients vary greatly in their breathing pattern and desire for flow, tidal volume, rate, and TI. Any particular initial ventilator settings are unlikely to coincide with the individual patient's needs.16,17,20 Thus the initial settings should be considered a first approximation. Then, taking into account the patient-ventilator interaction, as judged by subjective patient comfort and waveform displays of flow and pressure, the settings can be tailored to the individual patient. At times, only modest adjustment will improve patient-ventilator synchrony or patient comfort (Fig. 32-21). The beneficial impact of such changes may be evident not just in the flow and pressure waveforms, but in hemodynamic waveforms as well (Fig. 32-22).

A stepwise approach to adjusting the ventilator to the patient during volume-preset ventilation involves changing (1) tidal volume, (2) rate, (3) inspiratory flow rate, (4) TI, itself a consequence of tidal volume and flow rate, and (5) PEEP to counter auto-PEEP. Rarely, rise time may require consideration, as discussed below. For patients on PCV, the steps are to change (1) PI, (2) TI, (3) rate, and (4) PEEP. An example of this process is shown in Figure 32-23. Of course, any of these adjustments can cause problems or create conflict with other goals of ventilation. For example, raising rate (say to match a patient's high drive) may cause undesired auto-PEEP or hypocapnia, or raising tidal volume to lengthen machine TI may violate lung-protective goals. Often additional sedation is required to accommodate the patient to the ventilator, but this is appropriate only after steps have been taken to accommodate the ventilator to the patient.

The Noninvasively Ventilated Patient

During NIV, the patient and ventilator are coupled less tightly than when an endotracheal tube or tracheostomy is used. That is, the patient and ventilator are more easily asynchronous during NIV and it is even more important to carefully adapt ventilator to patient in order to improve the success of NIV.18

Two mechanisms of patient-ventilator asynchrony (PVA) are common. The first is failure of the patient to lower the proximal airway pressure (mask pressure) sufficiently to be able to trigger, due to the presence of auto-PEEP as discussed above. Adding PEEP is helpful to improve synchrony, but it is not generally practical to measure the amount of auto-PEEP in these patients, due to their effort. Rather, the clinician must adjust the machine PEEP upwards while assessing the patient's triggering effort. This modification can be greatly aided by the analysis of flow and pressure waveform displays. We begin routinely with the PEEP set at 2 or 3 cm H2O and find that maximal benefit is reached between 4 and 10 cm H2O.

The second common mechanism for PVA is failure of the ventilator to detect end-inspiration because the patient's subsiding effort is cloaked by a mask leak. Most PSV ventilators terminate inspiration when inspiratory flow falls to a preset threshold, often at an arbitrary low value of flow or at a fixed percentage of the peak inspiratory flow. Mask leaks prevent the flow from falling to this threshold, so the ventilator fails to switch off the inspiratory pressure, even while the patient is making active expiratory efforts. This serves to increase patient discomfort and the work of breathing. Using other methods for terminating inspiration, such as time-cycled pressure-support or ACV, can minimize this problem. Six patients with acute lung injury were studied during NIV with both conventional flow-cycled PSV and time-cycled PSV while the masks were adjusted to facilitate leaks.21 Time-cycled PSV significantly reduced the work of breathing (as judged by the esophageal pressure tidal swing and pressure-time product) and also more dramatically lowered the respiratory rate.

Displays of inspiratory and expiratory flows can be very valuable in detecting PVA and further assist the intensivist in modifying the mask or ventilator to improve synchrony and comfort. Carefully observing the patient-ventilator interaction and modifying the settings accordingly requires substantial time in the first hour on the part of the physician-respiratory therapist-nurse team.

Some intensivists have recommended adjusting the ventilator based on the inspiratory volume-pressure (VP) curve.22 Specifically, the upper end of the curve is examined for an upper inflection point (UIP), thought to signify alveolar overdistention (Fig. 32-24). If a UIP is present the tidal volume should be reduced, according to this view. The bottom part of the curve can be examined in an attempt to identify a lower inflection point (LIP), which may indicate recruitment during inspiration (Fig. 32-25). Such repeated recruitment (and during expiration derecruitment) has been shown in animal models of ARDS to be detrimental.23 One must take care to reduce the inspiratory flow rate to a low level (e.g., 15 liters per minute or 0.25 liters per second) when displaying the VP curve in order to avoid a flow-related artifact. If the flow rate is not reduced, the VP curve appears to show a LIP regardless of whether there is alveolar recruitment (Fig. 32-26).

There remain multiple problems with using VP curves to guide ventilator management in patients with ALI and ARDS. First, obtaining the VP curve is tedious: it requires physician time; detection of inflection points requires very careful analysis; any inflection points probably change as the lung changes, so repeated measures might be needed to guide ventilator settings; any patient effort makes the curve uninterpretable; the flow rate must be greatly reduced during measurement, which in most patients requires a temporary reduction in breathing rate (in order not to drastically curtail TE), during which the patient must remain fully passive; lower inflection points may not correlate with recruitment and derecruitment,24 and most importantly, this method has not been shown to improve outcomes despite its theoretical elegance. A recent study by the ARDS Network failed to show a benefit to a PEEP level higher than "least PEEP'' when applied in patients ventilated with tidal volumes of 6 mL/kg IBW.

Respiratory muscle activity greatly affects intrathoracic pressure, which alters measured hemodynamic values. By convention, hemodynamic values such as Pra and Ppw are measured at end-expiration since the respiratory muscles are most likely to be passive at end-expiration. It can be quite difficult to determine the point of end-expiration from a hemodynamic tracing, mostly because of respiratory activity (Fig. 32-27). This can lead to incorrect measurement of important pressures, perhaps prompting incorrect treatments. In the modern era of low tidal volume ventilation, reduced reliance on sedatives, and sparing use of therapeutic paralysis, effort is more the rule than the exception.

End-expiration can often be detected in the hemodynamic waveforms by paying attention to inspiratory-to-expiratory ratios, the nature of the respiratory rise in pressure (which differs between the ventilator-induced rise in the passive patient and the spontaneous expiratory rise in the active patient), and the abruptness of the falls in pressure, as discussed in Fig. 32-27. Additional confidence can be gained at the bedside by examining the patient and ventilator while simultaneously displaying the hemodynamic tracing in question. Perhaps the simplest and most accurate approach, however, is to connect the ventilator circuit to a pressure transducer and display this on the same time scale as the hemodynamic waveform. End-expiration is readily identified on the ventilator pressure waveform (Fig. 32-28). One then moves 200~milliseconds earlier on the time scale (since patient effort begins before machine inspiration, especially in the patient with auto-PEEP, as discussed above) and measures the hemodynamic pressure there.

Modern ventilators allow remarkable control over many aspects of ventilation beyond those related to tidal volume, rate, and PEEP. The ability to adjust flow profiles, flow rates, and rise times gives the intensivist tools for improving patient-ventilator synchrony and patient comfort. At the same time, however, changing some of these parameters has unanticipated consequences, many of which are revealed by comparing flow or pressure waveforms before and after the intervention. For example, the first set of pressure and flow tracings shown in Fig. 32-29 were seen in a patient ventilated with ACV and a rise time of 10% (meaning that 10% of the total respiratory cycle is devoted towards raising flow towards its peak value). Pressure does not rise above the PEEP value for the first two-thirds of the breath. The basis for this is readily evident from the flow tracing, which shows that flow is positive, but excruciatingly slow, for the first two-thirds of the breath. Flow is maintained low because the ventilator is not allowed to give the peak flow until 10% of the respiratory cycle has elapsed. If the patient tries to get more flow before this time, the ventilator simply throttles flow down further. Obviously, this is counterproductive, forcing the patient to work excessively. Yet no alarm will sound; the only clue to this problem is found in the ventilator waveforms. When the rise time was set appropriately (see Fig. 32-29), patient effort was reduced as the ventilator took on more of the burden of the work of breathing.

Lengthening the rise time also affects peak flow, since as the percentage of time at peak flow is reduced, the amount of flow must be raised in order to keep TI constant (Fig. 32-30). Therefore changing rise time can alter Ppeak, potentially confounding assessment of the respiratory system mechanics or response to therapy.

Many ventilators are provided with an option during volume-preset modes to give constant flow (the standard approach) or a decelerating (or sine wave) flow profile. Although the decelerating flow profile most closely mimics the flow pattern seen with PSV and PCV, there is no evidence that this pattern is either better or more comfortable than constant flow. The decelerating profile is most often selected when Pao is very high and there is concern for barotrauma. The effect of this is to lower Ppeak since at end-inspiration, when lung recoil is greatest, flow is much reduced. Yet an unavoidable consequence of lowering the flow late in the breath is that TI is lengthened (and TE shortened). This almost always serves to cause or exacerbate auto-PEEP (see Fig. 32-2). We strongly discourage the use of any flow profile other than square.

Even when inspiratory flow is kept constant throughout the breath, lowering its value can have a marked impact on auto-PEEP (Fig. 32-31). There are few reasons to ever ventilate a patient with the inspiratory flow set lower than 50 L/min (0.83 L/s).