Trigger dyssynchrony is of two types. The first is missed or delayed triggering. One cause for this is an insensitive or poorly responsive triggering system. On most ventilators a patient’s effort is sensed through either a drop in circuit pressure (pressure trigger) or a change in a circuit bias flow (flow trigger) (37, 38). Inherent in all patient-triggering systems is a built-in insensitivity to prevent autotriggering (see below). There are also mechanical triggering delays due to the inherent responsiveness characteristics of a ventilator’s valving systems.

Engineering advances have produced triggering systems that generally require only a small portion of the total effort of inspiration to initiate assisted/supported breaths. However, in the presence of very vigorous patient efforts, even the best systems may not be sensitive or responsive enough to avoid a significant triggering load (37, 38). Clinically these loads can be identified by an observed patient effort either failing to trigger a breath or having the triggering noticeably delayed. On the airway pressure–time graphic, there may be marked airway pressure deflections present before breath triggering. On the expiratory flow–time graphic, there may be evidence of transient flow reversal during missed trigger efforts.

A second cause of missed or delayed triggers occurs in the presence of intrinsic PEEP (PEEPi). This occurs because the patient’s ventilatory muscles must first overcome the PEEPi in the alveoli before any circuit pressure or flow change can occur to trigger a breath (39, 40). This can sometimes be appreciated by noting abrupt expiratory flow termination before a triggered breath or transient expiratory flow reductions or reversals that do not trigger a breath (Figure 1). In some cases, PEEPi can also be detected by measuring airway pressure during an expiratory pause. However, this may not always be evident in the setting of patient inspiratory efforts or in severe airway obstruction with collapsing small airways.

Figure 1. Flow (top) and pressure (bottom) over time (seconds) in a patient with frequent missed triggers. The missed or ineffective triggers (IT, arrows) can be detected on both graphs as transient pressure reductions and expiratory flow reversals. These efforts occur while expiratory flow is still occurring (incomplete exhalation [IE], arrows), which suggests that these missed triggers may be the result of intrinsic positive end-expiratory pressure. Adapted by permission from Reference 49.

Download Figure | Download Powerpoint

[More] [Minimize]

Patient effort and delayed/absent ventilator triggering can be better appreciated if a diaphragmatic EMG or an esophageal pressure (a surrogate for pleural pressure) is available as these techniques directly assess the timing of ventilatory muscle contraction. Coupling these measurements to the onset of flow delivery will clearly demonstrate the missed or delayed trigger (41–43). Moreover, the esophageal pressure–time tracing can be used to quantify any PEEPi present and the muscle PTP related to the imposed triggering load (44). For example, Leung and coworkers used esophageal pressure measurements to show that ventilatory muscle loading (PTP) was 38% higher for missed triggers than for properly triggered breaths (43). In contrast to delayed or missed breaths, a second type of trigger dyssynchrony is excessive triggering. This can be caused by autotriggering, entrainment, or premature cycling of breaths (37, 38). Autotriggering occurs when even small circuit leaks, tube condensation, and/or cardiac oscillations may trigger breaths and produce undesired hyperventilation and/or breath stacking with PEEPi. These extra breaths can result in significant apparent “tachypnea” and hyperventilation. As a consequence, some insensitivity in the triggering system often must be tolerated.

Another mechanism for extra triggering is in the setting of persistent effort after the machine breath has terminated (neural Ti > machine Ti) (37, 38). Under these circumstances, the second breath is tightly linked to the original breath and results in an increase in the measured ventilator rate.

A final mechanism of extra triggering occurs with the entrainment phenomenon described previously (20–22). When this occurs the effects of the stimulated effort depend on the timing and the breath type. If the stimulated effort occurs before the original breath has ended it can result in either an isometric load with airway pressure reductions (flow- and volume-targeted breaths) or as an addition to the Vt (pressure-targeted breath). However, if the stimulated effort occurs after the original breath has ended, a second breath can be triggered. Like the double triggering from a prolonged neural Ti and short machine Ti described previously, the additional triggering from entrainment will often increase the measured breath rate. Similarly, the second breath will graphically be tightly linked to the original breath.

Once a breath is patient effort triggered, diaphragmatic contraction continues to occur (45, 46). If flow is synchronous with that contraction pattern, the inspiratory muscle pressure–volume profile conceptually should resemble a near normal pattern (Figure 2). Note from Figure 2 that flow synchrony does not mean the elimination of the work of breathing. Instead it means providing flow to “reshape” the inspiratory muscle’s pressure–time or pressure–volume profile to a more physiologic configuration.

Figure 2. Pressure–volume plots (pressure on the horizontal axis, volume on the vertical axis) depicting various patient–ventilator flow interactions. Pressure to the left of the thick diagonal line (chest wall volume–pressure relationship) is patient generated and the work performed by the patient (the area of the curve) is shaded. Pressure to the right of the thick diagonal line is ventilator generated and the work performed by the ventilator is the open area. (A) A normal subject with normal work of breathing and no ventilator assistance. (B) A patient with respiratory failure resulting in very high work of breathing without any ventilator assistance. (C) This same patient is receiving ventilatory support that has transferred the entire work of breathing onto the ventilator. (D) The patient is performing only enough work to trigger an assisted breath—the ventilator does virtually all of the work. (E) The assistance provided by the ventilator is such that the patient work pattern resembles normal (A). In contrast, F illustrates a ventilator assistance pattern that is placing unphysiologic workloads on the patient. Conceptually, the goal of optimizing patient ventilator flow synchrony is to achieve the pattern shown in E and avoid that shown in F.

Download Figure | Download Powerpoint

[More] [Minimize]

Flow dyssynchrony from inadequate flow delivery can be appreciated clinically by observing inspiratory efforts that appear “flow starved” (vigorous inspiratory efforts unrewarded by adequate flow) and accompanied by marked patient discomfort. Examining the airway pressure–time profile can be useful in assessing flow dyssynchrony (Figure 3). In general, an airway pressure–time tracing that is smooth and consistently positive during inspiration suggests that flow is likely adequate to avoid excessive muscle loading (Figure 3, left). In contrast, a pressure–time waveform being “sucked down” by patient effort suggests that the delivered flow is markedly less than patient demand and excessive muscle loading may be developing (Figure 3, middle). When flow dyssynchrony is severe, the pressure–time waveform during inspiration can actually be pulled below the baseline airway pressure by high patient flow demands (Figure 3, right), an indication that the ventilator is really providing no inspiratory muscle unloading (43, 46–50). Indeed, by calculating the difference between the area under the curve of the pressure–time tracings of the assisted/supported versus the controlled breath, the actual muscle PTP during the assisted/supported breath can be estimated (46, 51). If an esophageal pressure tracing is available, inspiratory muscle loads can be directly calculated (41).

Figure 3. Plotted are airway pressure, flow, and volume over time during three assisted breaths. The dotted airway pressure line represents that observed during a control breath with a similar tidal volume as the assisted breath. Left: The assisted breath airway pressure profile remains smoothly positive and tracks with the control breath airway pressure profile, suggesting that the inspiratory muscle loading is likely not excessive. Middle: The assisted breath airway pressure profile is uneven and appears to be markedly “sucked down” by patient effort during much of the breath. This might suggest that the flow delivery is inadequate for patient demand to the point that inspiratory muscle overload may be present. Right: The assisted breath airway pressure profile goes below the baseline (expiratory) pressure. Flow is thus inadequate to provide any inspiratory muscle unloading.

Download Figure | Download Powerpoint

[More] [Minimize]

Flow dyssynchrony from inadequate flow delivery is more common during acute respiratory failure when inspiratory flow demands are high, vary from breath to breath, and ventilator flow delivery is set inappropriately low. Importantly, as noted previously, unmet flow demands drive further discomfort and inspiratory effort (46, 49, 51). Not surprisingly, flow dyssynchrony appears to be more common with ventilatory settings that deliver a fixed flow (flow-targeted breaths) rather than with a flow that can vary with effort (pressure-targeted breaths (48, 52–55).

Flow dyssynchrony can also occur when excessive flow is delivered (high set flow with flow-targeted breaths or high Pis and/or rapid pressure rise time settings with pressure-targeted breaths), especially in patients with reduced inspiratory efforts (56). Under these circumstances, lung expansion occurs faster than desired by the patient’s ventilatory control center. This can lead to excessive Vts in pressure-targeted modes, which can result in periodic breathing and adversely impact sleep (57). Excessive flow settings can also result in the ventilatory control center abruptly terminating the inspiratory effort (58) and even activation of expiratory muscles as patients “fight” to turn the breath off (a form of cycle dyssynchrony as described below).

The ventilator cycles or terminates flow to end the mechanical inspiratory phase and begin mechanical expiration on the basis of different criteria depending on mode settings. Specifically, in flow–volume-targeted modes, the delivered Vt and cycle time is clinician set and cannot vary with efforts. In pressure-targeted modes, the cycling criteria are either a set Ti in pressure assist control ventilation or a flow cycle setting in pressure support ventilation (PSV).

Cycling dyssynchrony occurs when the neural Ti and the machine Ti are mismatched (Figure 4) (24, 59). Importantly, the mismatch may be because of an abnormal ventilatory drive or because the cycle criteria are set either too short or too long for an appropriate ventilatory drive.

Figure 4. Two examples of cycle dyssynchrony. Depicted are flow ($\stackrel{.}{\mathrm{V}}$), airway pressure (Paw), and esophageal pressure (Pes) over time. Left: Machine Ti is greater than neural Ti. As a consequence, the lung inflation extends into neural exhalation and the patient may activate expiratory muscles to “turn the breath off.” This results in an elevation in airway pressure at the end of the inhalation. Right: Machine inspiratory time (Ti) is less than neural Ti. As a consequence the persistent patient effort “pulls” the airway pressure profile downward and reverses expiratory flow after breath termination. This persistent effort may trigger a second breath. Reprinted by permission from Reference 95.

Download Figure | Download Powerpoint

[More] [Minimize]

A mechanical Ti shorter than neural Ti may be necessary to prevent excessive Vt values if the neural Ti is inappropriately high (e.g., from anxiety, pain, or CNS abnormality). Under these circumstances, in addition to ventilator manipulations, addressing the inappropriate drive should also be done. A mechanical Ti shorter than neural Ti can also result from an inappropriately short set cycle time in a patient with an appropriate ventilatory drive and neural Ti. Regardless of cause, mechanical Ti less than neural Ti can leave the patient uncomfortable (air hungry) as inspiratory muscles continue to contract into mechanical expiratory time (Te) against the sudden elastic recoil of the chest wall (Figure 4, right) (7). Moreover, in the setting of an appropriate neural Ti, an inadequate machine Ti may result in hypoventilation, which can result in the ventilatory control center both increasing rate and neural Ti. Importantly, this persistent effort can also trigger a second breath as noted previously (60, 61).

A machine Ti longer than neural Ti may sometimes be necessary in patients with reduced ventilatory drive to provide an adequate Vt. However, with an appropriate ventilatory drive, an excessive machine Ti can lead to discomfort and expiratory efforts may be evident on the pressure–time and flow–time graphics as patients “fight” to turn off the breath (Figure 4, left) (56, 59, 61, 62). If the excessive machine Ti results in a larger Vt, this can lead to overdistention, which can result in the ventilatory control center reducing rate and neural Ti (63).

A prolonged mechanical Ti can be particularly problematic in patients with obstructive airway disease when using pressure support (23, 64, 65). Under these circumstances, the obstructed airways cause delivered inspiratory flow to decrease slowly and, because pressure support cycles on flow reduction, Ti may be inappropriately prolonged. These factors can lead to PEEPi buildup (dynamic hyperinflation) and consequent triggering dyssynchronies.

The distribution of controlled/assisted/supported breaths (i.e., the mode of ventilation) may also be important in patient–ventilatory synchrony. Specifically, when more than one breath type is being delivered, the patient’s ventilatory control center is unable to anticipate what the loading pattern during the next breath will be, and the potential for all dyssynchronies may go up (7, 18, 19). Thus modes with multiple breath types (i.e., IMV) may be particularly at risk for this. Indeed, as more and more spontaneous breaths with little or no ventilatory assistance are allowed with IMV, ventilatory drive goes up, which can then translate to less synchrony during the assisted mechanical breaths (19, 66).

Determining the prevalence of patient–ventilatory asynchrony is difficult as studies examining this question have involved varying patient populations, definitions of dyssynchrony, methods of detection, duration and timing of observation, and ventilatory modes (43, 60, 67–69). Triggering dyssynchronies have been the most well studied. Depending on patient population, ventilator settings, and measurement techniques, triggering dyssynchronies have been reported in 26–82% of mechanically ventilated patients (67). Among these patients, anywhere from 20 to 63% have more than 10% of their efforts associated with trigger dyssynchrony (67). Importantly, as many as 20% of patients with triggering dyssynchronies were not detected without measurements of esophageal pressure or diaphragm electrical activity (67). Not surprisingly, trigger dyssynchronies were more common in patients with chronic obstructive pulmonary disease and at risk for PEEPi development (43, 67). Double triggering is the other commonly reported triggering dyssynchrony, but this occurs in generally fewer than 10% of patients in these various studies (67).

The incidence of other forms of dyssynchrony (flow dyssynchrony and cycle dyssynchrony) has not been as well characterized. However, a retrospective evaluation of the National Institutes of Health (NIH) ARDS Network small Vt study reported cycling dyssynchronies associated with double triggering in 9.7% of all breaths analyzed (70, 71). Indeed, it is likely that patient–ventilatory dyssynchrony is ubiquitous if any patient is observed long enough during assisted/supported mechanical ventilation.

Although there is no doubt that many dyssynchronies are subtle and of little clinical relevance, significant dyssynchronies can produce patient discomfort and are a frequently cited indication for the administration of sedatives in many intensive care units (ICUs) (9, 72, 73). This may impact ventilator duration as high sedation usage is linked to longer ventilator use (4).

de Wit and colleagues demonstrated a longer duration of mechanical ventilation, worse 28-day ventilation-free survival, and longer ICU and hospital stays but no differences in ICU or hospital mortality in an observational study of 60 patients with various dyssynchronies studied during the first day of mechanical ventilation (68). Thille and colleagues (60) found that efforts associated with trigger dyssynchronies more frequently than 10% of the time (seen in 24% of their patients) were associated with substantially longer durations of mechanical ventilation and even a trend toward worse mortality. Whether the relationship between these additional adverse outcomes and patient–ventilatory dyssynchrony suggests causation or only represent a common link of a poor prognosis remains unclear.

The clinician should choose the trigger sensor (flow vs. pressure) that is most sensitive and responsive to patient effort (37, 38). Importantly, some ventilators have both types of effort sensors present and will respond to whichever signal is detected first. With either sensor, the clinician should adjust the sensitivity of the triggering system to be as sensitive as possible without producing autotriggering (37, 38).

In the setting of PEEPi trigger dyssynchrony, there are several clinical strategies. First, clinicians should try reducing the PEEPi as much as possible by reducing minute ventilation (e.g., reduce set rate, reduce set Pi, reduce set Vt, reduce ventilation needs driving patient efforts), lengthening the Te, or improving airway mechanics (17). The triggering load from PEEPi can also be reduced by applying judicious amounts of applied circuit PEEP, which serves to narrow the gradient between circuit (extrinsic) and PEEPi (39, 40). This could be guided by an esophageal pressure tracing with the goal of providing about 70–80% of measured PEEPi as circuit PEEP (19, 65, 73, 74). If an esophageal balloon is not available, an alternative approach is to empirically titrate PEEP and monitor the patient’s response (75). If the application of PEEP is benefiting the patient, the delay between effort and ventilatory triggering will shorten and the patient will be observed to be more comfortable. Ironically, the ventilator breathing frequency may actually increase (as will minute ventilation) because more efforts that were previously missed are now being triggered. This may require subsequent adjustments to avoid excessive ventilation. An important sign to look for is the amount of pressure required for the Vt. As long as the applied PEEP is less than the PEEPi this Pi/Vt relationship will not change (19, 65). Excessive PEEP above the PEEPi, however, will either drive the end-Pi up in flow–volume-targeted ventilation or reduce the Vt in pressure-targeted ventilation.

Managing an extra-triggering phenomenon depends on the cause. Ventilator autotriggering can be managed with a careful search for reversible causes (e.g., water in the circuit, small leaks) and/or adjustments to the trigger sensitivity settings (37, 38). Prolonged efforts with short mechanical Tis that trigger second breaths can be addressed by adjusting cycling criteria (mechanical Ti; see below). Managing entrainment effects can be more problematic as this phenomenon is less well studied. Additional sedation seems counterproductive as entrainment is associated with the use of heavy sedation (20). Conceptually, a reduction in sedation and mandatory breath delivery might be useful, but this has not been studied.

Ventilator setting adjustments for achieving flow synchrony depend on whether flow-targeted volume-cycled breaths or pressure-targeted breaths are being used (18). Flow-targeted volume-cycled breaths are the most common breath type used in modern ICUs (76), and they give the clinician direct control over the flow magnitude, flow delivery pattern, Ti, and the ultimate volume delivered. This can be useful in guaranteeing that a safe and effective Vt is provided. Unfortunately, the fixed flow delivery pattern cannot interact with the patient’s ventilatory drive and thus achieving flow synchrony can be a challenge.

When using flow-targeted breaths, the magnitude and the shape of the flow can be adjusted (sinusoidal vs. square vs. decelerating) to enhance synchrony (18, 77, 78). Inspiratory pause times can also be used to help with synchrony. When flow rates are properly titrated, Kallet and colleagues have shown that comfort with flow–volume-targeted breaths appears comparable to variable flow, pressure-targeted breaths (54). Of note is that careful attention to the flow settings with flow-targeted breaths in the NIH ARDS Network small Vt study resulted in no increased sedation use when using small Vts compared with large Vts (70).

Pressure-targeted breaths may offer synchrony advantages over flow-targeted breaths. This is because pressure targeting allows the ventilator to deliver whatever flow is needed to attain the set pressure target. Flow thus varies with patient effort, and this feature has been shown in many clinical studies to thereby enhance flow synchrony (Figure 5) (18, 19, 48, 52–55, 79).

Figure 5. Enhanced flow synchrony with a variable flow, pressure-targeted breath. Depicted are flow ($\stackrel{.}{\mathrm{V}}$), volume (Vt), and airway pressure (Paw) over time (seconds). Left: A flow-targeted breath is delivered but the flow is inadequate for patient demand, and dyssynchrony is manifest by the deeply downward coved airway pressure profile approaching the baseline pressure (solid arrow). In contrast, the pressure-targeted breath on the right is set to deliver a similar tidal volume. However, the variable flow feature (dashed arrow) synchronizes better with patient effort, providing a smoother, more constantly positive airway pressure profile. Reprinted by permission from Reference 48.

Download Figure | Download Powerpoint

[More] [Minimize]

The pressure-targeted breath also has several additional features that can further enhance flow synchrony. For example, the pressure rise time adjustment (flow acceleration adjustments also known as “pressure slope,” “inspiratory percent,” and other proprietary names) allows manipulation of the initial flow delivery, thereby increasing or decreasing the rate of rise of Pi (56). In theory, vigorous efforts might synchronize better with a rapid pressurization pattern; less vigorous efforts might synchronize with a slower pressurization pattern. Observational trials suggested this might be the case (56, 80, 81), but no study has shown this manipulation to alter outcome. Another commonly available feature is to have the ventilator adjust the circuit pressure profile to compensate for calculated endotracheal tube resistance and thereby produce a more “square wave” pressure profile in the trachea. Observational trials have suggested this might reduce muscle loads imposed by the artificial airway but, again, no study has shown this feature to alter outcome (82).

A concern with pressure targeting is that Vt control is lost (83). One way to address this is with feedback modes that allow the clinician to set a target Vt and then have the ventilator automatically adjust the pressure to maintain that volume. Although this has theoretical appeal, it is possible that changes in effort from anxiety or pain may create high Vts that then result in inappropriate lowering of the Pi (84, 85).

Achieving breath cycling synchrony involves delivery of an appropriate Vt in accordance with patient demands and matching of neural and machine Ti. With flow–volume targeting, adjusting the Vt and machine Ti is relatively straightforward as these are set independent variables that produce the machine Ti. With pressure targeting, adjusting the Vt and machine Ti is more complex and involves the interactions of applied Pi, respiratory system mechanics, patient effort, and cycling criteria. Altering any of these parameters often results in changes in others. In general, higher Pi settings, better mechanics, increased effort, and longer cycling criteria settings (higher set Ti in pressure assist control ventilation, lower expiratory flow criteria with PSV) extend the machine Ti (56, 59–62, 86). The pressure rise time in PSV can also affect machine Ti depending on its effects on the resulting patient ventilatory drive and its impact on peak flow and the flow cycling criteria (56).

One common cycling management problem is the patient on a pressure-targeted mode with a vigorous inspiratory effort who, despite a low applied Pi, still demands Vts that may be considered excessive (e.g., above 8–10 ml/kg ideal body weight) (70). Assuming this patient does not have a reversible cause for excessive inspiratory drive (e.g., pain), is on as low a Pi setting as possible, and is not ready for ventilator withdrawal (e.g., has high fraction of inspired oxygen or PEEP needs), many would argue that the high Vt should be tolerated and not suppressed with sedation.

Two new approaches to improving patient ventilatory interactions have been introduced: proportional assist ventilation (PAV) and neurally adjusted ventilatory assist (NAVA) (87). PAV requires that “test breaths” (controlled breaths with fixed flow and volume) be given. This allows for the calculation of respiratory system mechanics, which can be coupled with the measured ventilation to calculate work of breathing (resistive and elastic ventilatory muscle loads). These load calculations are repeated at regular intervals to maintain reliable inputs for the PAV algorithm.

PAV breaths are patient-initiated breaths triggered in a conventional way using circuit pressure or flow sensors. Thereafter, the ventilator continues to monitor flow and volume demanded by the patient and puts a clinician-set “gain” on this demand to augment flow and pressure in proportion to the desired reduction in the patient’s work of breathing. The PAV breath cycles when sensed flow demand has ceased.

Like pressure-targeted breaths, PAV flow delivery varies with patient effort; unlike pressure-targeted breaths, pressure also varies with patient effort. The conceptual upside to PAV is that flow and cycle synchrony should be enhanced over conventional assisted/supported breaths. Another conceptual upside is that patient-driven Vt variability with its theoretical lung protective benefits may be enhanced. The downside, however, is that, unlike conventional pressure-targeted breaths, there is no minimal pressure or flow provided. Thus, PAV must be used with caution in patients with unreliable ventilatory drives from either disease or drugs. Indeed, with all patients on PAV, careful monitoring and backup support modes should be available.

Most clinical studies with PAV have shown enhanced synchrony compared with conventional modes (88–90). However, it is not clear what the ideal PAV gain(s) should be in various clinical settings. Moreover, to date, there have been no good randomized trials looking at important outcome benefits (e.g., ventilator duration, sedation needs, mortality) when PAV is compared with properly provided conventional assisted/supported breaths.

NAVA requires a unique esophageal catheter with an array of diaphragm EMG sensors (87). These sensors detect the onset, intensity, and termination of inspiratory efforts directly. Like PAV, a clinician-set gain is then applied that determines flow and pressure delivery in proportion to the EMG signal.

The conceptual upside to NAVA is that synchrony with all three phase of breath delivery (trigger, gas delivery, and cycle) should be enhanced over conventional assisted/supported breaths. Like PAV, another conceptual upside is that patient-driven Vt variability with its theoretical lung protective benefits may be enhanced. Also like PAV, the downside is that there is no minimal pressure or flow provided. Thus, like PAV, NAVA must be used with caution in patients with unreliable ventilatory drives from either disease or drugs. Moreover, with NAVA there is also concern about the stability of the EMG signal coming from a catheter that can move within the esophagus. Thus all patients on NAVA require careful monitoring and backup support modes.

Most clinical studies with NAVA have shown enhanced synchrony compared with conventional modes (91–94). However, like PAV, it is unclear what the optimal EMG gain setting(s) should be in various clinical settings. To date there have been no good randomized trials looking at important outcome benefits (e.g., ventilator duration, sedation needs, mortality) when NAVA is compared with properly provided conventional assisted/supported breaths.