Chemical feedback refers to the response of Pmus to PaO2, PaCO2, and pH.5–7 In spontaneously breathing and mechanically ventilated patients, this system is an important determinant of respiratory motor output both during wakefulness and sleep.7–11
Mechanical ventilation can influence chemical feedback simply by altering the three variables PaO2, PaCO2, and pH. Hypoxemia, hypercapnia, or acidemia may be corrected by mechanical ventilation and thus modify activity of the medullary respiratory controller via peripheral and central chemoreceptors.5,12 The effects of mechanical ventilation on gas-exchange properties of the lung are beyond the scope of this chapter and are discussed in Chapter 37. In this chapter, the fundamental elements of the response of respiratory motor output to chemical stimuli, their relationship to unstable breathing, and the operation of chemical feedback during mechanical ventilation are reviewed.
Response of Respiratory Motor Output to Chemical Stimuli
Carbon dioxide (CO2) is a powerful stimulus of breathing.5,12 This stimulus, expressed by PaCO2, largely depends on the product of tidal volume (VT) and breathing frequency (f) (i.e., minute ventilation) according to Equation (3):
where is CO2 production, and VD/VT is the dead-space-to-tidal-volume ratio. Because minute ventilation is an adjustable variable in ventilated patients, understanding the relationship between respiratory motor output and CO2 stimuli is of fundamental importance.
Several studies have examined the respiratory motor output to CO2 in ventilated, conscious, healthy subjects.7,13–16 Major findings include
Manipulation of PaCO2 over a wide range has no appreciable effect on respiratory rate. Despite hypocapnia, subjects continue to trigger the ventilator with a rate similar to that of eucapnia. Respiratory rate increases slightly when PaCO2 approaches values well above eucapnia (Fig. 35-2).
The intensity of respiratory effort (respiratory drive) increases progressively as a function of PCO2. This response is evident even in hypocapnic range. The response slope increases progressively with increasing CO2 stimuli, reaching its maximum in the vicinity of eucapnic values (see Fig. 35-2).
There is no fundamental difference in the response to CO2 between various ventilator modes.
Above eupnea, the slope of the response does not differ significantly with that observed during spontaneous breathing, suggesting that mechanical ventilation per se does not considerably modify the sensitivity of respiratory system to CO2.
Schematic of the response of respiratory frequency (open squares) and pressure-time product of the inspiratory muscles per breath (an index of the intensity of patient effort, closed squares), both expressed as a percentage of values during spontaneous eupnea (baseline), to CO2 challenge in conscious healthy subjects ventilated with a high level of ventilator assistance. PETCO2 is end-tidal PCO2, and the dotted vertical line is PETCO2 during spontaneous breathing (eupnea). Contrast the vigorous response of intensity of inspiratory effort to CO2, even in the hypocapnic range, with the response of respiratory frequency, which remains at eucapnic level over a broad range of CO2 stimuli. The response is based on data from references 7 and 13 to 16.
During sleep (or sedation), the response of respiratory motor output to CO2 differs substantially from that during wakefulness, secondary to loss of the suprapontine neural input to the medullary respiratory controller.10,17 In ventilated sleeping subjects, a decrease in PaCO2 by a few millimeters of mercury causes apnea.10 Respiratory rhythm is not restored until PaCO2 has increased significantly above eupneic levels. The difference between eupneic PaCO2 and PaCO2 at apneic threshold, referred to as CO2 reserve,18 depends on several factors (see Response of Respiratory Motor Output to Chemical Stimuli—Chemical Stimuli and Unstable Breathing). This reserve determines the propensity of an individual to develop breathing instability during sleep; propensity increases as CO2 reserve decreases. Similar to wakefulness, the response of respiratory motor output to CO2 is mediated mainly by the intensity of respiratory effort, whereas respiratory rate decreases abruptly to zero (apnea) when the CO2 apneic threshold is reached.19
The effects of mechanical ventilation on the response of respiratory motor output to stimuli other than CO2 have not been studied adequately. In a steady state during wakefulness, the effects of oxygen (O2) and pH on breathing pattern are similar qualitatively to that observed with CO2: Changes in O2 and pH mainly alter the intensity of patient effort, whereas respiratory rate is affected considerably less.5,12 There is no reason to expect a different response pattern during mechanical ventilation. Indeed, this is the case regarding the hypoxic response in normal conscious subjects ventilated in assist-control mode during eucapnia.20 Indirect data also revealed that during eucapnia, the sensitivity of respiratory motor output to hypoxia was not modified by mechanical ventilation.20 During mild hypocapnia, however, the response was attenuated, whereas at moderate hypocapnia (end-tidal PCO2 approximately 31 mm Hg) the response was negligible. The latter observations may be relevant clinically because ventilated patients do not always keep PaCO2 at eucapnic levels and can become hypocapnic.16
Chemical Stimuli and Unstable Breathing
The response pattern of respiratory motor output to CO2 during sleep is relevant to the occurrence of periodic breathing in mechanically ventilated patients. Studies indicate that this breathing pattern might increase the morbidity and mortality of critically ill patients because it can cause sleep fragmentation and patient–ventilator dyssynchrony.21–23 Sleep deprivation may cause serious cardiorespiratory,24,25 neurologic,26,27 immunologic, and metabolic consequences.28–31
The following is a brief review of the factors that can lead to unstable breathing. In a closed system governed mainly by chemical control (such as occurs during sleep or sedation), a transient change in ventilation at a given metabolic rate (Δinitial) will result in a transient change in alveolar gas tensions. This change is sensed by peripheral and central chemoreceptors, which, after a variable delay, exert a corrective ventilatory response (Δcorrective) that is in the opposite direction to the initial perturbation32,33 (Fig. 35-3). The ratio of Δcorrective to Δinitial defines the loop gain of the system.32 Loop gain is a dimensionless index that is the mathematical product of three types of gains: plant gain (the relationship between the change in gas tensions in mixed pulmonary capillary blood and Δinitial), feedback gain (the relationship between gas tensions at the chemoreceptor level and those at the mixed pulmonary capillary level), and controller gain (the relationship between Δcorrective and the change in gas tensions at the chemoreceptor level) (Fig. 35-3). Loop gain has both a magnitude and a dynamic component.32,33 In this system, instability occurs when the corrective response is 180 degrees out of phase with initial disturbance (dynamic component) and loop gain is greater than 1 (magnitude component). This instability leads to fluctuation in chemical stimuli, namely, PCO2. If PCO2 reaches the apneic threshold, apnea occurs.
Schematic of the variables that determine the propensity of an individual to develop periodic breathing in a closed system dominated by chemical feedback. Loop gain is the product of three gains: plant, feedback, and controller. Instability occurs when Δcorrective (the final response) is 180 degrees out of phase with Δinitial (the transient initial perturbation) and Δcorrective/Δinitial is greater than 1. Mechanical ventilation, by affecting almost all variables of the system (↑, increase; ↔, no change; ↓, decrease), may change both the magnitude and the dynamic component of loop gain and thus the propensity of an individual to develop periodic breathing. CO, cardiac output; ΔPCCO2 and ΔPCO2, the difference in partial pressures of CO2 and O2 in mixed pulmonary capillary blood, respectively; ΔPchCO2 and ΔPchO2, the difference in partial pressure of CO2 and O2 at chemoreceptors (peripheral and central), respectively; Ers and Rrs, elastance and resistance of respiratory system, respectively; FRC, functional residual capacity; LG, Gplant, Gfeedback, and Gcontroller, loop, plant, feedback, and controller gains, respectively; alveolar partial pressure of CO2; Paw, airway (ventilator) pressure; Pmus, pressure developed by respiratory muscles; , ventilation–perfusion ratio; VD/VT, dead-space fraction.
Positive-pressure breathing exerts multiple effects on loop gain by influencing almost all the factors that determine plant, feedback, and controller gains. The effects are complex and at times opposing and variable (Table 35-1; see also Fig. 35-3). Nevertheless, the effect of mechanical ventilation on controller gain exerts the most powerful influence on the propensity to develop breathing instability.8,19,21,23 The magnitude and direction of the change in controller gain depends on the ventilator mode, the level of assistance, the mechanics of the respiratory system, and the Pmus waveform (see the section Interactive Effects of Patient-Related Factors and Ventilator on Control of Breathing).8,16,19,21 Disease states as well as medications (e.g., sedatives) also may interfere with the effects of mechanical ventilation on loop gain. For example, positive-pressure ventilation may increase or decrease cardiac output, causing corresponding changes in circulatory delay depending on cardiac function and intravascular volume (see Chapter 36).34–37 It has been shown that nocturnal mechanical ventilation in patients with congestive heart failure decreases the frequency of Cheyne-Stokes breathing, presumably by causing an increase in cardiac output secondary to afterload reduction.38–40 Sedatives at moderate doses, commonly used in ventilated patients, decrease considerably the loop gain, partly mitigating the effect of mechanical ventilation on controller gain and thus promote ventilatory stability.41
Table 35-1: Effects of Mechanical Ventilation on Gain Factors and Gain Changes |Favorite Table|Download (.pdf)
Table 35-1: Effects of Mechanical Ventilation on Gain Factors and Gain Changes
|Gain Factors (Influence)||Ventilator Effect*||Gain Change|
|Lung volume (stabilizing)||↑||↓Gplant|
|Cardiac output (destabilizing)||↓||↑Gplant, ↑Gfeedback|
|Thoracic blood volume (destabilizing)||↓||↑Gfeedback|
|Paw response to Pmus (destabilizing)||↑||↑Gcontroller|
|Alveolar (stabilizing)||↑||↓Gplant, ↓Gcontroller|
|Respiratory elastance (destabilizing)||↓||↑Gcontroller|
In addition to CO2, O2 and pH can play a key role in producing unstable breathing in ventilated patients during sleep (or sedation). It is well known that hypoxia, acting via peripheral chemoreceptor stimulation, decreases PaCO2. The result reduces the plant gain (stabilizing influence); for a given change in alveolar ventilation, PaCO2 will change less when baseline PaCO2 is low than when it is high.18 Hypoxia, however, increases the controller gain to a much greater extent42 because the slope of ventilatory response to CO2 below eupnea increases,12 a highly destabilizing influence.32,33 Similar principles apply if pH is considered as a chemical stimulus; acidemia decreases the plant gain (lowers PaCO2) and increases, to a much lesser extent, the controller gain.18,42 During mechanical ventilation, the propensity to unstable breathing in the face of changing O2 and pH stimuli depends on a complex interaction between the effects of these stimuli and mechanical ventilation on plant, feedback, and controller gains (Fig. 35-4; see also Table 35-1).
Tidal volume (VT), airway pressure (Pm), integrated diaphragmatic electrical activity (Edi, arbitrary units), and partial pressure of end-tidal CO2 (PETCO2) in a tracheostomized dog during non–rapid eye movement sleep without and with pressure-support ventilation at a pressure level that caused periodic breathing. (A) At a background of 5 hours of metabolic acidosis (pH 7.34, HCO3− 16 mEq/L, 30 mm Hg). (B) At a background of 1 hour of metabolic alkalosis (pH 7.51, HCO3− 35 mEq/L, 44 mm Hg). (C) During hypoxia ( 47 mm Hg, 31 mm Hg). At a background of metabolic acidosis, CO2 reserve was quite high; consequently, the pressure level that caused periodic breathing (20 cm H2O) was significantly higher than the corresponding values (approximately 10 cm H2O) during metabolic alkalosis or hypoxia. Hyperventilation during spontaneous breathing was similar during metabolic acidosis and hypoxia (similar stabilization influence via a decrease in plant gain secondary to low ), indicating that the destabilizing influence of hypoxia was caused by an increase in controller gain (hypoxic increase in the slope of CO2 below eupnoea). (Used, with permission, from Dempsey et al. J Physiol. 2004;560:1–11, based on data from Nakayama H, Smith CA, Rodman JR, et al. Effect of ventilatory drive on carbon dioxide sensitivity below eupnea during sleep. Am J Respir Crit Care Med. 2002;165:1251–1260.)
Operation of Chemical Feedback
The ventilator mode is a major determinant of driving pressure for flow and thus arterial blood gases. Before discussing the operation of chemical feedback, it is useful to review briefly the functional features of three main modes of assisted ventilation, namely, assist-control ventilation (ACV), pressure-support ventilation (PSV), and proportional-assist ventilation (PAV) (for detailed descriptions, see Chapters 6, 8, and 12). Figure 35-5 shows the response of the ventilator to respiratory effort in a representative subject ventilated with each mode in the presence and absence of CO2 challenge.16 With CO2 challenge, Paw decreases with ACV, it remains constant with PSV, and it increases with PAV. Pressure-time product of inspiratory muscle pressure (PTP-PmusI) is an accurate index of the intensity of inspiratory effort.43 With ACV, the ratio of VT to PTP-PmusI per breath (neuroventilatory coupling) decreases with increasing Pmus; the ratio is largely independent of inspiratory effort with PAV. With PSV, VT/PTP-PmusI per breath may change in either direction with increasing Pmus, depending on factors such as the level of pressure assist and cycling-off criterion, change in Pmus, and mechanics of the respiratory system. With PSV, in the absence of active termination of pressure delivery (with expiratory muscle contraction), the ventilator delivers a minimum VT, which may be quite high, depending on the pressure level, mechanics of the respiratory system, and cycling-off criterion.19
End-tidal carbon dioxide tension (PETCO2), airway pressure (Paw), flow (inspiration up), volume (inspiration up), and esophageal (Pes) pressure in a representative subject during proportional-assist ventilation (A, B), pressure-support ventilation (C, D), and volume-control ventilation (E, F) in the absence (A, C, E) and presence (B, D, F) of CO2 challenge. With CO2 challenge, Paw decreases with assist-control ventilation (the ventilator antagonizes patient’s effort); it remains constant with pressure-support ventilation (no relationship between patient effort and level of assist); and it increases with proportional-assist ventilation (positive relationship between effort and pressure assist). (Used, with permission from Mitrouska J, Xirouchaki N, Patakas D, et al. Effects of chemical feedback on respiratory motor and ventilatory output during different modes of assisted mechanical ventilation. Eur Respir J. 1999;13:873–882.)
Assume that in a ventilated patient PaCO2 drops because of an increase in the set level of assistance or decrease in metabolic rate and/or VD/VT ratio.44 During wakefulness, patients will react to this drop by decreasing the intensity of their inspiratory effort, whereas the breathing frequency will remain relatively constant (see “Response of Respiratory Motor Output to Chemical Stimuli,” above). The extent to which a patient is able to prevent respiratory alkalosis via operation of chemical feedback depends almost exclusively on the relationship between the intensity of patient inspiratory effort and the volume delivered by the ventilator (i.e., VT/PTP-PmusI). Similarly, if PaCO2 increases (decrease in assistance level, increase in metabolic rate and/or VD/VT ratio), the patient will increase the intensity of inspiratory effort and, to much lesser extent, respiratory frequency. Thus, VT/PTP-PmusI per breath is critical for the effectiveness of chemical feedback to compensate for changes in chemical stimuli (PaCO2). For given respiratory system mechanics, VT/PTP-PmusI is heavily dependent on the mode of support. Thus, the effectiveness of chemical feedback in compensating for changes in chemical stimuli should be mode-dependent. Modes of support that permit the intensity of patient inspiratory effort to be expressed on ventilator-delivered volume improve the effectiveness of chemical feedback in regulating PaCO2 and particularly in preventing respiratory alkalosis. In normal conscious subjects receiving maximum assistance on the three main ventilator modes,16 the ability of the subject to regulate PaCO2 depends on the operational principles of each mode, specifically in terms of VT/PTP-PmusI (Fig. 35-6). At all levels of CO2 stimulation, preservation of neuroventilatory coupling increased progressively from ACV to PSV to PAV; the ability of subjects to regulate PaCO2 followed the same pattern.16
Ratio (mean ± SD) of tidal volume to pressure–time product of inspiratory muscles (VT/PTP-PmusI) in normal, conscious subjects ventilated with three modes of assisted ventilation in the absence and presence of CO2 challenge (inspired CO2 concentration increased in small steps until intolerance developed). Open and closed bars represent zero and final (highest) concentration of inspired CO2, respectively. AVC, assist-volume control; PAV, proportional-assist ventilation; PS, pressure-support ventilation. Asterisk indicates significant difference from the value without CO2 challenge. Plus sign indicates significant difference from the corresponding value with PAV. With each mode, subjects were ventilated at the highest comfortable level of assistance (corresponding to 80% reduction of patient resistance and elastance with PAV, 10 cm H2O of pressure support, and 1.2-L tidal volume with AVC). With CO2 challenge, VT/PTP-PmusI, decreased significantly when the subjects were ventilated with PS and AVC, but it remained relatively constant with PAV. Without CO2 challenge, VT/PTP-PmusI was significantly higher with PS and AVC than with PAV. This response pattern caused severe respiratory alkalosis with PS and AVC (PETCO2 decreased to approximately 22 mm Hg with both modes) but not with PAV (PETCO2 approximately 30 mm Hg). Unlike with PS and PAV, subjects ventilated with AVC could not tolerate high values of PETCO2 (final PETCO2 was approximately 7, 11, and 13 mm Hg higher than baseline eupnea, respectively, with AVC, PS, and PAV). (Based on data from Mitrouska et al.16)
Neurally adjusted ventilatory assist (NAVA) is a new mode of support that, similar to PAV, uses patient effort to drive the ventilator.45–47 The electrical activity of the diaphragm is obtained with a special designed esophageal catheter and serves as a signal to link inspiratory effort to ventilator pressure (see Chapter 13). Because neuroventilatory coupling is preserved, the principles described above also apply to NAVA.46,47
During sleep or sedation, the tendency to develop hypocapnia with ACV and PSV (see Chapter 57 for the effects of mechanical ventilation on sleep) may have serious consequences because a drop of a few millimeters of mercury in PaCO2 leads to apnea and periodic breathing.8,19 Thus, excessive assistance with ACV and PSV promotes unstable breathing secondary to impaired neuroventilatory coupling; controller gain remains high in the face of low inspiratory effort (Fig. 35-7). Unstable breathing, however, during sleep secondary to mechanical ventilation may be prevented or attenuated with PAV and NAVA that does not guarantee a minimum VT.8,19,46,47 Modes that decrease the volume delivered by a ventilator in response to any reduction in the intensity of patient effort enhance breathing stability and may be associated with better sleep quality.48 Nevertheless, if the assist setting during PAV or NAVA is such that controller gain increases considerably, and the inherent loop gain of the patient is relatively high, the patient will be at risk of developing unstable breathing.23,33,41,49,50
Polygraph tracings in a healthy subject during non-rapid eye movement sleep with and without pressure-support ventilation. (A) Spontaneous breathing with continuous positive airway pressure (CPAP). (B) to (D) Pressure support of 3, 6, and 8 cm H2O, respectively. Periodic breathing with central apneas developed with pressure support of 8 cm H2O. C3/A2 and C4/A1, electroencephalogram channels; EMG, electromyogram; EOG, electrooculogram (right [R] and left [L]); Paw, airway pressure; PETCO2, end-tidal PCO2. (Used, with permission, from Meza, et al. Susceptibility to periodic breathing with assisted ventilation during sleep in normal subjects. J Appl Physiol. 2003;167:1193–1199.)
These principles may be altered by disease states and therapeutic interventions. Although little is known about the interaction between disease states and mechanical ventilation on control of breathing, two examples help in illustrating the point. First, in conscious patients with sleep apnea syndrome, a drop in PaCO2 because of brief (40 seconds) hypoxic hyperventilation resulted, contrary to healthy subjects, in significant hypoventilation and triggering of periodic breathing in some patients.51 This hypoventilation was interpreted as evidence of a defect (or reduced effectiveness) of short-term poststimulus potentiation, a brainstem mechanism that promotes ventilatory stability.51 In this situation, a level of assistance that causes a significant decrease in PaCO2 may promote unstable breathing in awake patients with sleep apnea syndrome, a situation closely resembling that observed during sleep. Second, studies in ventilated critically ill patients have shown that when awake patients are unable to increase VT appropriately as a result of the mode used (i.e., PSV), they increase respiratory rate in response to a chemical challenge.52 Behavioral feedback, however, may underlie this response pattern. In sedated patients with acute respiratory distress syndrome (in whom behavioral feedback is not an issue) receiving PSV, considerable variation in PaCO2 elicited a steady-state response limited to the intensity of breathing effort, a response pattern similar to that observed in normal subjects.9,16
Intrinsic Properties of Respiratory Muscles
For a given neural output, Pmus decreases with increasing lung volume and flow, as dictated by the force-length and force-velocity relationships of inspiratory muscles, respectively.53 Therefore, for a given level of muscle activation, Pmus should be smaller during mechanical ventilation than during spontaneous breathing if pressure provided by the ventilator results in greater flow and volume. It has been shown in healthy subjects ventilated with PSV that, compared with spontaneous breathing, the relationship between electrical activity (Edi) and pressure-time product of diaphragm (PTPdi) is shifted to the left; thus, at any given level of Edi, PTPdi is reduced.54
The influence and consequences of mechanical feedback during mechanical ventilation have not been studied satisfactorily. It is possible that this type of feedback is of clinical significance in patients with dynamic hyperinflation (high end-expiratory lung volume), high ventilatory requirements (requirements for high flow and volume), and/or impaired neuromuscular capacity.
The characteristics of each breath are influenced by various reflexes that are related to lung volume or flow and mediated, after a latency of a few milliseconds, by receptors located in the respiratory tract, lung, and chest wall.5,6 Mechanical ventilation may stimulate these receptors by changing flow and volume. In addition, changes in ventilator settings, inevitably associated with changes in volume and flow, also may elicit acute Pmus responses mediated by reflex feedback. In sedated patients with acute respiratory distress syndrome, manipulation of ventilator settings altered immediately (within one breath) the neural respiratory timing, whereas respiratory drive remained constant.9,55 Specifically, decreases in VT and pressure support and increases in inspiratory flow caused an increase in respiratory frequency. Depending on the type of alteration, changes in respiratory frequency were mediated via alteration in neural inspiratory and expiratory time; increases in inspiratory flow caused increases in respiratory frequency mainly by decreasing neural inspiratory time; decreases in VT and pressure support caused increases in respiratory frequency by decreasing neural expiratory time. This reflex response was similar, at least qualitatively, to that observed in healthy subjects during wakefulness and sleep.56–60 There was a strong dependency of neural expiratory time on the time that mechanical inflation extended into neural expiration; neural expiratory time increased proportionally to the increase in the delay between the ventilator cycling off and the end of neural inspiratory time (Fig. 35-8).9,55 This finding indicates that expiratory asynchrony may elicit a reflex timing response. A subsequent study in a general intensive care unit population confirmed the dependency of neural expiratory time on expiratory asynchrony.61 The most likely explanation for the timing response is the Herring-Breuer reflex.
Relationship between the changes in the time that mechanical inspiration extended into neural expiration (ΔText, expiratory asynchrony) and neural expiratory time (ΔTen) in mechanically ventilated patients with acute respiratory distress syndrome. Closed circles, open circles, and open triangles represent ΔText induced by changes in volume (at constant flow), flow (at constant volume), and pressure support, respectively. Solid line, regression line. (Based on data from Kondili E, Prinianakis G, Anastasaki M, Georgopoulos D. Acute effects of ventilator settings on respiratory motor output in patients with acute lung injury. Intensive Care Med. 2001;27:1147–1157.)
The final response may be unpredictable depending on the magnitude and type of lung volume change, the level of consciousness, and the relative strength of the reflexes involved. Nevertheless, reflex feedback should be taken into account when ventilator strategies are planned. A few examples may help in illustrating the importance of reflex feedback in patient–ventilator interaction. Assume that the patient is receiving pressure support that is being decreased during weaning. This results in lower VT, which through reflex feedback decreases neural expiratory time, causing an increase in respiratory frequency.9,55 This increase should not be interpreted as patient intolerance to the decrease in pressure support. Consider another patient with obstructive lung disease receiving ACV. VT is decreased at a constant inspiratory flow so as to reduce the magnitude of dynamic hyperinflation (less volume is exhaled over a longer period). The lower VT usually results in less delay in breath termination as compared with the end of neural inspiration, which through vagal feedback will decrease neural expiratory time, limiting the effectiveness of this strategy for reducing dynamic hyperinflation.55 Assume in another patient receiving ACV that inspiratory flow is increased at a constant VT, with the intent of reducing inflation time and providing more time for expiration so as to reduce dynamic hyperinflation. This step causes a reflex decrease in neural inspiratory time and an increase in respiratory frequency. Mechanical expiratory time may change in either direction depending mainly on the relation between neural and mechanical inspiratory time. In patients receiving ACV, expiratory time showed a variable response to changes in flow rate; some patients actually demonstrate a reduced expiratory time with a higher flow,62 which cancels the desired reduction in dynamic hyperinflation.
There are neural reflexes that inhibit inspiratory muscle activity if lung distension exceeds a certain threshold, which is well below total lung capacity (Hering-Breuer reflex).6,63,64 These reflexes protect the lung from overdistension, which is associated with lung injury.65,66 Pressure-control or volume-control modes of assisted ventilation considerable interfere with the ability of these reflexes to regulate tidal volume.16,67 With these modes, as a result of neuroventilatory uncoupling (high VT/PTP-PmusI), overassistance may result in high tidal volume leading to regional or global lung overdistension. Conversely, recent evidence indicates that ventilator modes that permit reflex feedback to regulate the tidal volume and respiratory rate (viz., NAVA, PAV) may protect against or lessen ventilator-induced lung injury.
Brander et al68 randomized anesthetized rabbits with early experimental acute lung injury into three ventilator strategies: NAVA (nonparalyzed), volume control with tidal volume of 6 mL/kg (paralyzed, protective strategy), and volume control with tidal volume of 5 mL/kg (paralyzed, injurious strategy). Animals randomized to NAVA selected an average tidal volume of 2.7 ± 0.9 mL/kg and respiratory rate up to three times higher than that in both controlled ventilation groups—a breathing-pattern response that can be explained by vagally controlled reflexes.6,63,64 Compared to the 15 mL/kg group, animals ventilated with either NAVA or volume control at 6 mL/kg exhibited less ventilator-induced lung injury, as indicated by lung injury scores, lung wet-to-dry ratio, and lung and systemic biomarkers (Fig. 35-9). These results indicate that the use of NAVA, which allowed the animals to choose their own respiratory pattern, was at least as effective in preventing various manifestations of ventilator-induced lung injury as conventional, volume-controlled ventilation using a tidal volume of 6 mL/kg.
Parameters indicative of ventilator-induced lung injury (VILI) in rabbits with induced acute lung injury (ALI) and ventilated with three strategies: NAVA, volume control with tidal volume (VT) of 6 mL/kg, and volume control with VT of 15 mL/kg. (A) There were no differences in partial pressure of arterial oxygen to fractional inspired oxygen concentration ratio (/FIO2) among groups before and 30 minutes after induction of ALI. The increase in /FIO2 shortly after switching to the assigned ventilation mode (i.e., after randomization into the treatment groups) was more pronounced with NAVA than with volume control (VC) 6-mL/kg (p < 0.05 post hoc analysis), although /FIO2 was not different between NAVA and VC 6-mL/kg at the end of the protocol. With VC 15-mL/kg, /FIO2 remained below 200. (B) The lung wet-to-dry ratio with NAVA and with VC 6-mL/kg was lower than with VC 15-mL/kg (albeit not significantly for the dependent lung in VC 6-mL/kg animals). (C) and (D) Interleukin 8 (IL-8), tissue factor, and plasminogen activator inhibitor type 1 (PAI-1) concentration in bronchoalveolar (BAL) fluid was higher in all study groups compared to healthy controls and was higher with VC 15-mL/kg than with the other two groups (except for PAI-1 in VC 6-mL/kg). Lung tissue IL-8 concentration was increased in all groups as compared to nonventilated controls and was highest in the nondependent lung regions with VC 15-mL/kg. In the VC 6-mL/kg and NAVA groups, lung tissue IL-8 concentration was lower compared to VC 15-mL/kg (albeit not significant for the dependent lung region). Groups are shown as mean ± standard deviation (SD) for A and B, or as median (quartiles) for C and D. Symbols represent group mean; bars indicate standard deviation. e–g, time–group interaction (two-way analysis of variance). Post hoc pairwise comparison procedure between groups: †p <0.05 NAVA versus VC 6-mL/kg; ‡p <0.05 NAVA versus VC 15-mL/kg; §p < 0.05 VC 6-mL/kg versus VC 15-mL/kg. (Used, with permission, from Brander L, Sinderby C, Lecomte F, et al. Neurally adjusted ventilatory assist decreases ventilator-induced lung injury and non-pulmonary organ dysfunction in rabbits with acute lung injury. Intensive Care Med. 2009;35:1979–1989.)
In a human study employing randomized design, Xirouchaki et al69 ventilated 108 critically ill patients, most of whom had acute lung injury or acute respiratory distress syndrome, with PAV+ (PAV with automatic estimation of elastance and resistance of the respiratory system; see Chapter 12) Even with high assistance, tidal volume and end-inspiratory plateau pressure were comparable to these observed during protective controlled mechanical ventilation. Examination of individual end-inspiratory plateau pressures during PAV+ showed that out of a total of 744 measurements only on nine occasions (1.2%) and in five patients (4.6%) were plateau pressures above 30 cm H2O (Fig. 35-10). Ninety-four percent of the end-inspiratory plateau pressures were below 26 cm H2O, a value associated with lung protection.70 Similar to the findings of Brander et al, these results can be explained by the operation of reflex feedback (vagally controlled reflexes).6,63,64
Individual values of quasi-static airway pressure obtained by 300 msec pause maneuver at the end of selected inspirations (PPLATpav) as a function of time in 108 critically ill patients randomized (zero time) to proportional assist ventilation with load-adjustable gain factors (PAV+). PAV+ was continued for 48 hours unless the patients met predefined criteria, either for switching to controlled modes or for breathing without ventilator assistance. Closed black circles connected by solid thick line represent mean values. Each patient is denoted by a single color. For comparison the mean ± standard deviation (SD) values of static end-inspiratory airway pressure, obtained within 8 hours before randomization during controlled mechanical ventilation (CMV), is shown (closed black square). Notice that in the majority of the patients PPLATpav was below 26 cm H2O. (Used, with permission, from Kondili et al. Patient–ventilator interaction. Br J Anaesth. 2003;91:106–119.)
Is it possible to use this reflex feedback into a clinical scenario? Recent studies suggest that in patients with acute respiratory distress syndrome titration of tidal volume based on individual lung mechanics may be a better strategy than using a fixed tidal volume (i.e., 6 mL/kg).65,66,70–72 Obtaining lung mechanics, however, necessitates the use of cumbersome techniques not easily available at bedside. Theoretically, tidal volume selected by the patient should be based on individual lung mechanics, which serve as a guide for setting the ventilator.73,74 Although studies support this hypothesis,68,69 caution should be exercised in patients with strong signals of nonrespiratory origin (acidosis, brain dysfunction) that drive ventilation. Notwithstanding the limitations and feasibility of this approach, this hypothesis deserves further studies.
Mechanical ventilation at relatively high tidal volume and ventilator frequency results in a non–chemically mediated decrease in respiratory motor output.75–77 This decrease, referred to as neuromechanical inhibition, is manifested both in respiratory frequency and in amplitude of respiratory motor output. Neuromechanical inhibition lasts for several breaths after termination of mechanical ventilation, thus constituting a type of control system inertia and resetting of the spontaneous respiratory rhythm.78 Although the mechanism underlying neuromechanical inhibition is not entirely clear, the Hering-Breuer reflex is the most plausible explanation. In addition, Sharshar et al79 showed that mechanical ventilation reduces the excitability of cortical motor areas representing respiratory muscles. It is possible that mechanoreceptor feedback accounts for the depression of the motor-evoked potential of the diaphragm via vagal and other proprioceptive afferents to the respiratory center. The clinical relevance of neuromechanical inhibition is currently unknown. Available evidence suggests that its contribution to respiratory motor output in ventilated critically ill patients is rather minimal.9,11,55
Entrainment of Respiratory Rhythm to Ventilator Rate
Entrainment of respiratory rhythm to the ventilator rate implies a fixed, repetitive, temporal relationship between the onset of respiratory muscle contraction and the onset of a mechanical breath.80–82 Human subjects exhibit one-to-one entrainment over a considerable range above and below the spontaneous breathing frequency.83,84 Cortical influences (learning or adaptation response) and the Hering-Breuer reflex are postulated as the predominant mechanisms of entrainment. Theoretically, one-to-one entrainment should facilitate patient–ventilator synchrony, but studies of the entrainment response in critically ill patients are lacking.
The effects of behavioral feedback on control of breathing in ventilated patients are unpredictable, depending on several factors related to the individual patient and surroundings. Alteration in ventilator settings, planned to achieve a particular goal, might be ineffective in awake patients because of behavioral feedback.85,86 Inappropriate ventilator settings may cause breathing discomfort in awake patients. Consequent panic reactions further aggravate the unpleasant breathing sensation and create a vicious cycle. Behavioral feedback also may be altered considerably from time to time secondary to changes in the level of sedation, sleep–awake state, patient status, and environmental stimuli. The many factors involved in behavioral feedback complicate its study and the interpretation of its effects on the system that controls breathing in mechanically ventilated patients.