Control of Breathing in Mechanically Ventilated Patients

Home , Respiratory tract

SERIES "CLINICAL PHYSIOLOGY IN RESPIRATORY INTENSIVE CARE" Edited by A. Rossi and C. Roussos Control of breathing in mechanically ventilated patients

D. Georgopoulos, C. Roussos

Control of breathing in mechanically ventilated patients. D. Georgopoulos, C. Roussos. Pulmonary and Critical Care Dept, General ©ERS Journals Ltd 1996. Hospital "G. Papanicolaou", University of ABSTRACT: During mechanical ventilation, the respiratory system is under the Thessaloniki, Thessaloniki, Greece and influence of two pumps, the ventilator pump and the patient's own respiratory Critical Care Dept, "Evangelismos" Hospital, muscles. Depending on the mode of mechanical ventilatory support, ventilation University of Athens, Athens, Greece. may be totally controlled by the ventilator or may be determined by the interac- Correspondence: D. Georgopoulos tion between patient respiratory effort and ventilator function. In either case, com- General Hospital "G. Papanicolaou" pared to spontaneous breathing, the breathing pattern is altered and this may Pulmonary Dept influence: 1) force-length and force-velocity relationships of respiratory muscles Respiratory Failure Unit (mechanical feedback); 2) chemical stimuli (chemical feedback); 3) the activity of Exochi 57010 various receptors located in the respiratory tract, lung and chest wall (reflex feed- Thessaloniki back); and 4) behavioural response (behavioural feedback). Changes in these feed- Greece back systems may modify the function of the ventilator, in a way that is dependent Keywords: Behavioural feedback on the mode of mechanical ventilatory support, ventilator settings, mechanics of chemical feedback the respiratory system and the sleep/awake stage. mechanical feedback Thus, the response of ventilator to patient effort, and that of patient effort to mechanical ventilation ventilator-delivered breath are inevitably the two components of control of breath- reflex feedback ing during mechanical ventilation; the ventilatory output is the final expression of the interaction between these two components. As a result of this interaction, the Received: May 20 1996 various aspects of control of breathing of the respiratory system may be masked Accepted after revision July 7 1996 or modulated by mechanical ventilation, depending on several factors related both to patient and ventilator. This should be taken into consideration in the management of mechanically ventilated patients. Eur Respir J., 1996, 9, 2151Ð2160.

The act of breathing is a complex process [1, 2]. Brief- the mechanical properties of the respiratory system (elas- ly, the medullary respiratory controller (central control- tance and resistance) (fig. 1). Depending on the mode ler) accepts information from chemical (peripheral and of mechanical ventilatory support, volume-time profile central chemoreceptors) and nonchemical sources. Based and ventilator timing may be totally controlled by the on this information, the central controller activates spinal f R,pt motor neurons serving respiratory muscles, with an inten- Mode sity and rate that may vary substantially between breaths. Mode The activity of spinal motor neurons is conveyed to res- Pmus + Paw = V '×Rrs + V×Ers Ventilator timing Neural timing piratory muscles, which contract and generate pressure (Pmus). Pmus is dissipated to overcome the resistance and Volume-time profile elastance of the respiratory system (inertia is negligible) VT f 'R,vent and this combination determines the volume-time profile and, depending on breath timing, ventilation. Volume- time profile and breath timing via force-length and force- Mechanical Chemical Reflex Behavioural velocity relationships of respiratory muscles affect Pmus, whereas they modify the activity of spinal motor neu- Feedback rons and the medullary respiratory controller via affer- Fig. 1. Ð Schematic representation of the interaction between patient ent nerves from various receptors. On the other hand, respiratory effort and ventilator-delivered breath. Pmus: pressure gen- ventilation and gas exchange properties of the lung deter- erated by respiratory muscles (inspiratory muscles generate positive mine arterial blood gas values, which in turn, via peri- pressure and expiratory muscles negative); Paw: airway pressure; V: pheral and central chemoreceptors, affect the activity of instantaneous volume above passive functional residual capacity (FRC); V': instantaneous flow (inspiratory flow is positive); Rrs: resistance of the medullary respiratory controller, closing the loop. the respiratory system; Ers: elastance of the respiratory system; Ventilator In a mechanically-ventilated patient, the breath deliv- timing: duration of inspiratory and expiratory flow (mechanical inspi- ered by the ventilator has two components, one related ratory and expiratory time); Neural timing: neural (patient) inspira- to the volume-time profile and the other to ventilator tim- tory and expiratory time; VT: tidal volume. fRvent: ventilator (respirator) frequency; fR,pt: patient spontaneous breathing frequency. Depending ing [3, 4]. Volume-time profile, according to the equa- on the mode of ventilatory support, Pmus and neural timing may or may tion of motion [5], is determined by the combined action not affect Paw and ventilator timing, respectively. Note that fR,vent may of Pmus, pressure provided by the ventilator (Paw) and not reflect fR,pt. See text for further details. 2152 D. GEORGOPOULOS, C. ROUSSOS ventilator or may be determined by the interaction bet- (proportional assist ventilation) [3, 4]. With volume-con- ween patient respiratory effort and ventilator function [3, trol modes, the volume-time profile and duration of 4]. In either case, compared to spontaneous breathing, inspiratory flow are predetermined by the ventilator set- the pattern of breathing and ventilation are changed. tings. Thus, changes in Pmus and neural inspiratory time These changes may alter: 1) force-length and force-velo- cannot modify tidal volume (VT) delivered by the ven- city relationships of respiratory muscles (mechanical tilator. Any change in Pmus causes Paw to change in feedback) [6, 7]; 2) chemical stimuli (chemical feed- the opposite direction, because total pressure (Paw+Pmus) back) [8]; and 3) the activity of various receptors locat- is not changed. Therefore, with volume-control modes, ed in the respiratory tract, lung and chest wall (reflex the ventilator antagonizes the intensity of patient effort feedback) [9, 10]. Furthermore, changes in volume-time (fig. 1). Furthermore, the time at which inspiratory flow profile and breathing pattern are readily perceived in is terminated is independent of neural inspiratory dura- awake subjects and may evoke behavioural ventilatory tion. It follows that, with volume-control neither the responses (behavioural feedback) [11, 12]. As a result of intensity of patient effort nor neural inspiratory time mechanical, chemical, reflex and behavioural feedback, are expressed by the output of the ventilator. Pmus and patient neural timing (neural inspiratory and With pressure-control, the ventilator once triggered expiratory duration) are altered and these alterations, causes Paw to increase rapidly to a preset level, remain- depending on the mode of mechanical ventilatory sup- ing at that level until a preset cycling-off criterion (the var- port [3, 4], may or may not influence Paw and ventilator iable that terminates gas delivery) is reached [3, 4, 14]. timing (fig. 1). Thus, the ventilatory output is the final Because Paw is constant, the volume-time profile is under expression of the interaction between patient effort and the influence of Pmus, and any change in the intensity ventilator. It follows that the response of ventilator to of patient effort is expressed by a change in inspiratory patient effort, and that of patient effort to ventilator- flow rate (fig. 1). The cycling-off criterion may be a set delivered breath are inevitably the two components that time or flow. With time-cycling, neural inspiratory time control breathing during mechanical ventilation. An is ignored by the ventilator and the tidal volume is deter- understanding of these two components is essential for mined by Pmus waveform (inspiratory and expiratory) the physician dealing with the issue of control of breath- and mechanical properties of the respiratory system [5] ing in mechanically-ventilated patients. (fig. 1). With flow-cycling, gas delivery is terminated when inspiratory flow reaches a fixed level (usually 0.1 Lás-1) or a value which is proportional to peak inspira- Response of ventilator to patient effort tory flow (usually 25%). This method is called pressure- support (PS) and is widely-used [3, 14]. Theoretically, with PS the patient retains considerable control of the Basic principles of positive pressure ventilators inspiratory volume-time profile and inspiratory flow duration; any change in the intensity and rate of patient effort should be expressed by VT and ventilator timing. Never- Positive pressure ventilators can be characterized by theless, in the face of high ventilatory demands, many various variables, which control the initiation of the me- ventilators are not able to maintain constant Paw, and chanical breath, gas delivery and mechanical inspira- Paw deviates from the target level [15]. Furthermore, it tory time [3]. The response of the ventilator to patient has been shown on theoretical grounds that the ability effort depends on the type of variables that a specific to modulate VT during PS is limited, particularly in pati- mode of ventilatory support uses. ents with abnormal mechanics, for reasons related both to patient and ventilator (for review see [15]). Therefore, Trigger variable. The trigger variable defines when the the ventilatory consequences of a given increase in pati- ventilator initiates gas delivery. This variable may be ent effort might be expressed inappropriately (see below). time, pressure or flow [3, 13]. With time-triggering, the Proportional assist ventilation (PAV) is a new mode ventilator delivers gas at fixed time intervals. With pres- of mechanical ventilation in which Paw is proportional sure- or flow-triggering, gas delivery is initiated when to instantaneous flow and volume [4, 16]. Thus, there the patient decreases airway pressure (Paw) below pos- is not a target level either for pressure or for flow. The itive end-expiratory pressure (PEEP) before the assist proportionality between Paw and instantaneous flow and ventilation begins. In some modes, the ventilator does volume is preset by the ventilator, according to the fol- not provide any flow until Paw decreases to a predeter- lowing equation: mined level (pressure-triggering); whilst in other modes, × × the ventilator allows air to flow in response to the de- Paw = k1 V' + k2 V (1) crease in Paw, and triggering occurs when flow from machine to patient exceeds a set level (flow-triggering). where V ' and V are instantaneous flow and volume, res- Therefore, with time-triggering, the ventilator rate bears pectively, and k1 and k2 are gain factors. To the extent no relationship to patient breathing frequency; while that V' and V depend on the intensity of inspiratory effort with pressure- or flow-triggering, the ventilator rate is, (Pmus), Paw is positively related to Pmus, as opposed to theoretically, set by the patient. being negatively related (volume control) or independent (pressure control). With PAV, the volume-time profile Variables that control gas delivery and mechanical inspi- and breathing pattern are tightly linked to Pmus wave- ratory time. Gas delivery from the ventilator may be form [4, 15, 16]. Any change in the rate and intensity governed by a set flow (volume-control), a set pressure of patient effort should be expressed by ventilatory out- (pressure-control), or instantaneous flow and volume put. CONTROL OF BREATHING IN MECHANICAL VENTILATION 2153

Missing effort (fig. 2a) and on another to 30 Lámin-1 (fig. 2b). Several important points are illustrated by the figure. At both Ideally, in patients ventilated on assisted modes, all 90 and 30 Lámin-1 values of V'I a significant number of inspiratory efforts trigger the ventilator, which delivers missing efforts occurred. These missing efforts can be gas and supports the patient effort [3]. The level of sup- identified using Paw, flow or Poes waveforms. An abrupt port may range from zero to near maximum and, depend- decrease in Paw and Poes during expiration and hesita- ing on the mode used, may vary from breath to breath tion in expiratory flow, which are not followed by mach- [3]. With zero support, the patient performs the total ine-delivered breath, indicate missing effort. At V'I of 90 work of breathing; whilst with near maximum support, Lámin-1, the rate of machine cycles was 17 breathsámin-1, inspiratory muscles relax after triggering. whereas the patient's spontaneous rate was 22 breathsá High resistance to airflow, low elastic recoil, high ven- min-1. Minute ventilation, determined by ventilator rate tilatory demands, and short expiratory time may not per- and VT, was 9.4 Lámin-1. By changing V'I to 30 Lámin-1, mit the system to reach static equilibrium volume at the it can be observed that there was a decrease in mac- end of expiration [17]. Hence, inspiration begins at vol- hine rate and ventilation to 13 breathsámin-1 and 6.2 umes at which the respiratory system exhibits a posi- Lámin-1, respectively, despite the fact that patient's breath- tive recoil pressure, referred to as intrinsic PEEP (PEEPi) ing frequency increased to 24 breathsámin-1. Furthermore, [18Ð22]. This phenomenon is called dynamic hyperin- note that a considerable portion of inspiratory muscle flation and is a common finding in mechanically-ven- pressure needs to trigger the ventilator; and, in some tilated patients [18Ð21]. In this case, the patients must breaths, all the muscle pressure is dissipated to trigger first generate enough Pmus to overcome PEEPi before the venti-lator and, therefore, neural inspiratory time triggering occurs. There might be a situation where ends when machine inspiratory time starts. It is obvi- pressure generated by the inspiratory muscles to initiate ous that the machine cycles out of phase with the patient a breath is less than PEEPi plus the airway pressure de- and the discrepancy varies substantially from breath to crease required to trigger the ventilator, and, therefore, breath. Finally, observe that when Poes swings are decreased inspiratory effort fails to trigger the ventilator ("miss- the likelihood of missing effort increases. This indicates ing effort") [15, 23Ð26]. Because there is no inflation that, for a given degree of PEEPi, missing effort is more during this breath, lung volume continues to decline, so likely to occur when the Pmus is small, such as when that the elastic recoil is less at the beginning of the next the muscles are fatigued and/or weak or when central patient effort and the patient is in a better position to drive is low (i.e. low Pa,CO2). trigger the ventilator on the next spontaneous cycle. The phenomenon of missing efforts has been studied Figure 2 shows airway pressure (Paw), airflow, and in detail, on theoretical ground, by YOUNES [15, 23], who oesophageal pressure (Poes) in a patient with chronic used a model of the respiratory system to examine the obstructive pulmonary disease (COPD) mechanically- relationship between machine rate and spontaneous breath- ventilated on assist volume-controlled (AVC) mode. In ing frequency during various modes of support (AVC, the example presented in figure 2, VT was set to 0.55 L, PS and PAV). His analysis indicates that, for given given with a square-wave flow-time profile. On one mechanical properties of the respiratory system, the rel- occasion, inspiratory flow (V'I) was set to 90 Lámin-1 ationship is not simple and is influenced by the level of assist ventilation, the intensity of patient effort and a) the spontaneous breathing frequency. Increased assist level, spontaneous patient breathing rate and decreased Paw 25 intensity of patient effort are associated with greater disc- cmH2O 0 repancy between patient and ventilator (fig. 3). The like- Flow 1 lihood of missing efforts was less with PAV, probably L·s-1 0 because neural timing and drive is tightly linked to ventilator timing and VT.

Poes 0 To summarize the observations on missing efforts: 1) cmH O -25 the rate of the machine's cycles does not reflect the pat- 2 ient's spontaneous breathing frequency. 2) at constant b) patient breathing frequency, the rate of the machine's cycle may be influenced by Pmus, an index of VT demand Paw 25 (i.e. drive). Any factor that affects Pmus may also affect cmH2O 0 ventilator frequency and thus, paradoxically, stimuli that Flow 1 increase drive may actually affect machine rate; 3) at L·s-1 constant patient breathing frequency and Pmus, manipula- 0 tion of the assist level (pressure or volume assist), machine inspiratory time and cycling-off criteria may change the Poes 0 machine's rate; and 4) decrease in patient breathing fre- cmH O -25 quency may decrease the proportion of missing effort 2 by prolonging expiratory time. This may increase the Fig. 2. Ð Airway pressure (Paw), flow and oesophageal pressure machine's rate and vice versa. (Poes) in a patient with chronic obstructive pulmonary disease venti- It is obvious from the above considerations that the lated on assist volume-controlled mode with two different inspiratory phenomenon of missing effort has a considerabe affect flow rates (V'I): a) 90 Lámin-1; and b) 30 Lámin-1. Tidal volume was kept constant (0.55 L). Missing efforts are indicated by arrows. See on the interpretation of ventilatory output in relation to text for further details. the control of breathing during mechanical ventilation. 2154 D. GEORGOPOULOS, C. ROUSSOS

a) 1 0 Flow 0.5 -1 L·s 1

Volume above FRC L 0 2 Volume 0

-1 L 0 1 L·s Flow 5 ETCO -2 2 % 0 6 O 2 1 s

mus 3 Paw 10 P cmH 0 cmH2O 0 30 Fig. 4. Ð Flow (inspiration down), volume (inspiration down), end- O 2 tidal CO2 (ETCO2) and airway pressure (Paw) in a normal subject ven- PS 15

cmH tilated on assist volume-controlled mode. Note double-triggering (arrow) 0 when inspiratory flow was 1 Lás-1. This occurred because mechanical 0 1 2 3 4 5 6 7 inspiratory time, which was preset by ventilatory settings, was con- Time s siderably shorter than neural inspiratory time. In this case, Pmus imme- b) diately after inflation, decreased Paw below the threshold for triggering 1 and caused the ventilator to recycle. The actual tidal volume (VT) delivered to the subject and the ventilator rate are double the prede- 0.5 termined VT (note the expired VT) and spontaneous subject breathing frequency, respectively. Changing inspiratory flow from 60 to 50 Volume above FRC L 0 Lámin-1 (VT was kept constant) increased mechanical inflation time 2 from 0.6 to 0.8 s and double-triggering did not occur. Observe, also, the difference in ETCO2 between the breaths with and without double- -1 0 triggering. L·s Flow

-2 at the end of mechanical inspiration, Pmus continues to 6 increase and, because inspiratory flow is zero or is rev- O 2 ersed, it is dissipated to overcome the elastic recoil mus 3 P

cmH alone. Thus, there might be a situation where Pmus is 0 greater than elastic recoil, causing airway pressure to 30

O decrease below PEEP and this triggers the ventilator 2 15 (fig. 4). Retriggering may occur with PS or AVC. On PS cmH 0 the other hand, retriggering does not occur with PAV because, with this mode, Pmus is the variable that con- 2 3 6 0 1 4 5 7 trols gas delivery. Short mechanical inflation time may Time s promote retriggering (fig. 4). With the phenomenon of Fig. 3. Ð Volume, airflow, muscle pressure (Pmus) and airway pres- retriggering, machine rate overestimates patient sponta- sure in a simulated patient with obstructive lung disease ventilated on pressure-support (PS) mode. a) The ventilator is triggered every other neous breathing frequency. Furthermore, as with miss- spontaneous inspiratory effort. Ventilator rate is 20 cyclesámin-1, while ing efforts, retriggering might change the patient effort the patient's spontaneous breathing frequency is 40 breathsámin-1 (arrow if alterations in various feedback systems occur (figs. 1 indicates missing effort). b) Keeping the same Pmus, patient's rate and 4). decreases from 40 to 30 breathsámin-1, allowing more time for expiration. This causes a reduction in the magnitude of dynamic hyperinflation and, as a result, each inspiratory effort triggers the ventilator. Response of patient effort to ventilator-delivered The ventilator rate increases to 30 cyclesámin-1, while the patient's rate has actually decreased. Notice in both figure 3a and b the dis- breath crepancy between neural and machine inspiratory and expiratory time. (Vertical dotted line indicates the beginning of inspiratory effort). Mechanical feedback (From YOUNES [15], with permission). Mechanical feedback describes the well-known effects Furthermore, with missing efforts, significant alteration of length (volume) and velocity of contraction (flow) of in patient effort occurs due to changes in feedback loop. respiratory muscles, as well as of geometrical factors on Failure of the ventilator to respond to patient inspiratory Pmus [6, 7, 27]. For a given level of muscle activation, effort may alter mechanical, chemical, reflex and beha- Pmus decreases with increasing lung volume and flow. vioural feedback, thus, secondarily affecting the inten- Thus, for similar neural output to respiratory muscles, sity and rate of the patient's respiratory effort (fig. 1). Pmus should be smaller during mechanical ventilation than during spontaneous breathing if pressure provided by the ventilator results in greater flow and volume. The Retriggering consequences of mechanical feedback in mechanically- ventilated patients are not known. However, the effects With retriggering the ventilator is triggered more than of mechanical feedback on Pmus would be small if vol- once during the same inspiratory effort [15, 23]. This ume and flow are low relative to their maximum val- may occur if the patient inspiratory effort is vigorous ues [28, 29]. During mechanical ventilation, the operating and longer than mechanical inflation time. In which case, volume and flow are relatively low [3, 4], indicating that CONTROL OF BREATHING IN MECHANICAL VENTILATION 2155 mechanical feedback is not very important for mech- changes in respiratory muscle pressure and, thus, in anically-ventilated patients. Nevertheless, it is possible patient effort [16]. Furthermore, oesophageal pressure that this type of feedback may be of clinical signifi- was related to static recoil pressure of the chest wall cance in patients with high ventilatory requirements and was not corrected for flow resistance [16, 32]. It and/or impaired neuromuscular competence. Furthermore, follows that with high inspiratory flows, observed with in the presence of dynamic hyperinflation inspiratory constant pressure, inspiratory effort was underesti- muscles are forced to operate at high lung volume, which mated, making the interpretation of the results compli- is a disadvantageous position for pressure generation cated. [30Ð32]. Therefore, mechanical feedback, by reducing Recently, using a rebreathing method, we studied the Pmus, might increase the number of missing efforts in response of neuromuscular output to CO2 with and patients with dynamic hyperinflation. with-out unloading of the respiratory system [38]. The unloading was achieved using PAV. At similar PCO2 in Chemical feedback peripheral and central chemoreceptors, neuromuscular output, expressed by transdiaphragmatic pressure and One of the main objectives of mechanical ventilation total pressure generated by all respiratory muscles, rem- is to unload the respiratory muscles [3]. It would be ained virtually unchanged by an approximately 50Ð60% interesting to see the effects of respiratory muscle unload- reduction of the normal mechanical load; the neuromus- ing on control of breathing. Theoretically, the respira- cular output was tightly linked to CO2 and not to load tory system can follow one of three courses in response reduction. These results indicate that increasing the ass- to unloading: 1) respiratory muscle activation is down- ist level in mechanically-ventilated subjects unloads the regulated, so that the same ventilation as before the respiratory muscles only to the extent that PCO2 decr- unloading is obtained; 2) respiratory muscle activation eases. The degree of downregulation should depend on remains unchanged and, therefore, ventilation increa- the sensitivity to CO2 and the magnitude of PCO2 reduc- ses according to the degree of unloading; and 3) there tion. Notwithstanding that the response to unloading may be an intermediate response, whereby ventilation might be related, to some extent, to baseline mechani- is higher at a lower level of respiratory muscle activity. cal load or to the mode of mechanical ventilation [37], It is generally believed that the respiratory system fol- these results emphasize the importance of chemical feed- lows the third course; with unloading, ventilation is high- back during mechanical ventilation. Paradoxically, the er and respiratory motor output is lower [16, 33, 34]. role of chemical feedback has been largely ignored by Whilst these findings indicate that reflex feedback rel- studies dealing with the effect of mechanical ventila- ated to the load per se plays a role in determining the tion on respiratory muscle activity. level of respiratory muscle activation, the results of such The effectiveness of chemical feedback to compensate studies fail to provide information about the relative for changes in chemical stimuli in mechanically-ven- importance of such feedback. This is because these stud- tilated patients is a complicated issue. During controlled ies were performed using an open loop system and, mechanical ventilation (CMV), an increase in chemical therefore, chemical feedback was not strictly compara- stimulus (Pa,CO2 or arterial oxygen tension (Pa,O2)) can- ble with and without unloading. Thus, the observed not elicit any ventilatory response, because the ventila- downregulation of respiratory muscle output could have tor does not increase its rate or its VT in response to been related to associated reduction of chemical feed- patient effort. With assist modes of mechanical venti- back produced by the higher ventilation. In an open sys- latory support, the patient, theoretically, has the option tem, chemical feedback cannot be discounted on the to change ventilation as a result of chemical feedback. grounds that partial pressure of oxygen or carbon diox- With constant flow SIMV, a change in chemical stim- ide (PO2 or PCO2) did not change "significantly". uli may elicit a ventilatory response only through alter- The ability of respiratory muscle unloading to down- ations in the characteristics of spontaneous breaths, while regulate respiratory motor output has been questioned mandatory breaths are independent of patient effort, a by several pieces of evidence. Data from patients dur- situation similar to that during CMV. In patients venti- ing constant flow synchronized intermittent mandatory lated on AVC mode, the respiratory system can compen- ventilation (SIMV) have shown that for a given level sate for changes in chemical stimuli through breathing of assist, inspiratory effort did not differ between spon- frequency, but not through the intensity of patient effort. taneous and mandatory breaths [35, 36]. These results On the other hand, with pressure assisted modalities of indicate that inspiratory output is preprogrammed and ventilatory support (PS or PAV) the ventilator delivers is relatively insensitive to breath-by-breath changes in a VT which varies with the intensity of patient effort. In load seen during SIMV. Chemical feedback could be a this case, the ventilator has the ability to respond to both critical factor for this breath programming. These results components of the ventilatory response to change in have recently been challenged by GIULIANI et al. [37], chemical stimuli (fig. 5). It follows that with pressure- who showed that the mode of mechanical ventilation is assist, chemical feedback may better control arterial important to show an effect of unloading on respiratory blood gas values. effort. They demonstrated that inspiratory effort was small- What happens in reality? We are all familiar with a er in mandatory breaths than in spontaneous only if patient ventilated on assist modes, who although hav- SIMV is applied with flow-triggering and constant pres- ing an intact central drive and normal or near normal sure. However, in this study, respiratory effort was quan- respiratory system mechanics, develops respiratory alka- titated using oesophageal pressure, which complicates losis or acidosis as a result of an inappropriate assist the interpretation of the results. With pressure assist ven- level, a change in metabolic demands, or a change in tilation, changes in oesophageal pressure do not reflect gas exchange properties of the lung. Notwithstanding the 2156 D. GEORGOPOULOS, C. ROUSSOS

Volume-assist Pressure-assist a) a,CO 5 P 2 Pa,CO2 ● ∆ 4 P ,70

O ● 2 ● ∆ f R R P ,40 f Respiratory muscles 3 ● ●

cmH ● contraction ● ∆ ● ● P ,20 Pa,CO Pa,CO mus 2 ● ● 2 2 P ● ● ● ∆ ● ● ● ● ● ● ● 1 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● f R f R Respiratory muscles ● ● 0 ● contraction b) Fig. 5. Ð Pathways that the respiratory system can use in order to

-1 12 ● compensate for change in arterial carbon dioxide tension (Pa,CO2) during volume-controlled and pressure-assist modes of synchronized (pres- O·s 2 sure- or flow-triggering) mechanical ventilatory support. : decrease; 8 ● : increase; f R: respiratory frequency. Pmus: pressure generated by ● ● respiratory muscles. See text for details. ● ● 4 ● ● ● ● dp/dt cmH ●●● ● ● ● differences between volume and pressure assist, as far ∆ 0

as the response of the ventilator to patient effort is con- -1 c) cerned, the pati-ent with both modes of ventilatory sup- ● 1 ● port, through chemi-cal feedback, should be able to ● ● ● ● maintain a constant Pa,CO2, by appropriate adjustments ● ● ● ● ● ● ● ● 0 ● ● in rate, intensity of respiratory effort, or both. Assum- breaths·min R ing normal activity of nonchemical inputs (mechano- f ∆ 26 28 30 32 34 36 38 40 42 receptors, irritant receptors, cortical influence), failure PET,CO mmHg to maintain a constant Pa,CO2 may imply either a defect 2 or reduced effectiveness of chemical feedback during Fig. 6. Ð Changes in a) pressure generated by respiratory muscles mechanical ventilation. This issue is of paramount impor- (∆Pmus) at different fractions of mechanical inflation time, b) rate of decline of airway pressure before triggering (∆dp/dt) and c) breath- tance in understanding the relationship between chem- ∆ ing frequency ( fR) as a function of end-tidal CO2 tension (PETCO2) ical feedback and mechanical ventilation, and the following in normal subjects mechanically-ventilated on assist volume-controlled observations may help us to clarify it. mode. ∆Pmus and ∆dp/dt reflect changes in the intensity of respira- Awake normal humans ventilated on AVC or PS with tory effort. ∆P,70, ∆P,40 and ∆P,20: ∆Pmus at 70, 40 and 20% of mechan- a relatively high VT were seen to develop hypocapnia ical inflation time, respectively. Note that the intensity of respiratory [39Ð42]. This was due to the fact that the subjects con- effort increases with increasing PCO2 stimulus even at low PET,CO2. f 'R remains relative constant over a wide range of PET,CO2. (From tinued to trigger the ventilator rhythmically despite high PATRICK et al., [39], with permission). tidal volumes and hypocapnia. Manipulation of PCO2 over a wide range had no appreciable effect on breath- defending respiratory alkalosis. To the extent that breathing frequency [39, 40]. On the other hand, the intensity ing frequency is insensitive to CO2 [39, 40], mechani- of respiratory effort, quantified from changes in airway cally-ventilated awake subjects may easily develop pressure at constant flow and volume, rate of decline of respiratory alkalosis due to inappropriately high assist airway pressure prior to triggering, VT and mouth occlu- levels, reduced metabolic rate, or improvement in gas sion pressure at 0.1 s from the onset of inspiratory effort exchange properties of the lung. (P0.1) increased as a function of PCO2 [39, 40]. It is of Secondly, because the ventilatory response to CO2 is interest to note that the response was evident even in expressed mainly by the intensity of respiratory effort the hypocapnic range [39] (fig. 6). These results indi- [39], PS mode, which permits VT to change in response cate that in mechanically-ventilated, awake humans: 1) to patient effort [14], may, in contrast to AVC, com- breathing frequency is relatively insensitive to CO2 over pensate for changes in PCO2. We should mention, how- a wide range of PCO2; 2) the intensity of respiratory ever, that the compensation during PS is partial because effort increases with increasing PCO2, even below eucap- of the minimum VT delivered (see above) and the lim- nic levels; and 3) the ventilatory response to CO2 is ited ability of respiratory effort to modulate VT, partic- expressed mainly by intensity of respiratory effort. ularly in patients with abnormal mechanical properties These observations have at least two important con- of the respiratory system [15, 23]. PAV may permit bet- sequences that should be taken into consideration in the ter control of patient effort to ventilator-delivered vari- management of mechanically-ventilated patients. Firstly, ables [43], due to the fact that the volume-time profile PS and AVC modes of mechanical ventilation greatly and ventilator timing are tightly linked to Pmus [4, 15]. compromise the ability of chemical feedback to control Indeed, it has been demonstrated in patients ventilated PCO2. This is because, with AVC, the ventilator once on PAV, that ventilation and breathing pattern did not triggered delivers a fixed VT [3, 4], whilst with PS, in change appreciably as assist level was varied from near the absence of active termination of inspiration, the VT maximum to the lowest tolerable [43]. However, this has a minimum value which depends on PS level, mechan- mode is currently under investigation and we cannot ical properties of the respiratory syste, and the cycling- comment on it further. off criterion [14, 23]. It follows that, with both modes Although the above studies used CO2 as a stimulus, of support breathing, frequency plays a key role in similar principles should apply if PO2 is altered. In CONTROL OF BREATHING IN MECHANICAL VENTILATION 2157 steady-state, the effects of CO2 and O2 on breathing pat- Periodic or irregular breathing during sleep as a result tern are qualitatively similar; increasing the O2 or CO2 of mechanical ventilatory support may be prevented or stimulus affects mainly the intensity of respiratory effort, attenuated with PAV, which does not guarantee a min- while the response of breathing frequency is signifi- imum VT. Indeed, MEZA et al. [52] have shown, in nor- cantly less [8]. mal sleeping subjects, that mechanical ventilation with We should be aware that the effect of mechanical ven- PAV was not associated with periodic or irregular breath- tilation on control of breathing as far as chemical feed- ing. They observed that VT, respiratory frequency (fR) back is concerned might be modulated by various disease and end-tidal carbon dioxide tension (PETCO2) remained states. It has been shown in awake patients with obstruc- relatively stable, even at the highest assist level. These tive sleep apnoea [44], and in patients with brain damage results indicate that during sleep PAV permits chemi- [45], that a drop in PCO2 because of brief (40 s) hypo- cal feedback to regulate breath-by-breath arterial blood xic hyperventilation, resulted, in contrast to normal sub- gas values. jects, in significant hypoventilation. When hypoxia was In summary, the effectiveness of chemical feedback sustained for much longer (25 min), hypoventilation was to compensate for changes in chemical stimuli during observed even in normals [46]. This hypoventilation was mechanical ventilation depends on: 1) the mode of mech- interpreted as evidence of a defect or reduced effec- anical ventilatory support; and 2) the sleep/awake stage. tiveness of short-term poststimulus potentiation, a brain Failure to appreciate the role and limitations of chemi- stem mechanism promoting ventilatory stability [46, 47]. cal feedback during mechanical ventilation may lead to To the extent that mechanical ventilation may repre- serious consequences for patient management. Diseases sent a type of forced hyperventilation, these results could that may alter the response to mechanical ventilation apply in mechanically-ventilated patients. High assist should always be a consideration. levels in these patients may decrease Pa,CO2 and trigger periodic breathing. Nevertheless, further studies are need- Reflex feedback ed that will test the ventilatory response to chemical stimuli during various modes of ventilatory support, as Reflex feedback plays an important role in control of well as the effects of disease. breathing [1, 2]. The characteristics of each breath are So far, we have discussed chemical feedback during influenced by various reflexes, which are related to lung wakefulness. The picture is completely different during volume or flow and mediated by receptors located in the sleep or under anaesthesia. Several studies have shown respiratory tract, lung and chest wall [9, 10]. Most of our that, under these circumstances, the maintenance of res- knowledge about the effects of these reflexes on control piratory rhythm is critically dependent on chemical feed- of breathing has been obtained from animal studies [53Ð back [41, 48Ð50]. Reducing Pa,CO2 by only a few mmHg 55]. Very little is known about the relevance of these re- causes apnoea. In the face of mechanical ventilatory sup- flexes to mechanical ventilation and much work needs to port with assist modes, there are compensatory changes be done. A few points, however, deserve some comment. in breathing pattern, so that PCO2 is forced to remain Static and dynamic changes in lung volume elicit around the CO2 set-point [41]. Thus, if pressure support responses mediated by vagal and chest wall receptors or volume assist are set at values higher than those [9, 10, 15, 23, 53, 54]. In addition, it has been shown required for eucapnia or for Pa,CO2 set-point, periodic or that controlled mechanical ventilation results in the gen- irregular breathing may be caused (fig. 7). These epi- eration of a VT-dependent inhibitory input to inspiratory sodes may be associated with significant hypoxaemia, muscles, mediated by an unidentified pathway [56, 57]. an issue of great importance for critically ill patients. All of these reflexes related to lung volume influence However, it should be mentioned that in the presence the breathing pattern in a complex way. The final resp- of active lung disease, input to the respiratory controller onse depends on the magnitude and type of lung vol- from nonchemical sources [51] may not permit chemi- ume change, the level of consciousness, and the relative cal feedback to prevent respiratory alkalosis during sleep strength of the reflexes involved [15, 23]. At present, the or under anaesthesia. role of the above reflexes on mechanically-ventilated 10 patients is unclear. Pmask cmH O Currently, in mechanically-ventilated patients inspira- 2 0 tory flow rates are adjusted mainly for the purpose of Flow 1 enhancing patient-ventilator interaction and of changing -1 inspiratory time, and, thus, affecting airway pressures, L·s 0 dynamic hyperinflation, haemodynamic status and distr- ibution of ventilation [37, 58, 59]. However, inspiratory 35 flow rates may affect respiratory output in a way that PET,CO2 has been largely ignored in patient management. It has mmHg been shown, in mechanically-ventilated normal subjects, 0 that increasing inspiratory flow rate exerted an excitatory reflex effect on respiratory output; increasing inspi- 1 min ratory flow was associated with an increase in central Fig. 7. Ð Airway pressure (Pmask), flow, and end-tidal carbon diox- drive and breathing frequency, and a decrease in expirat- ide tension (PET,CO2) in a normal subject ventilated noninvasively ory time [60, 61]. This effect was complete in one breath (nose-mask) on pressure-support during non-rapid eye movement (NREM) sleep. Note that this subject exhibited periodic breathing as after a change in flow rate, and persisted, although to a a result of mechanical ventilatory support. (From MORRELL et al., [41], lesser degree, during non-rapid eye movement (NREM) with permission). sleep [60] (fig. 8). The strength of this reflex was not 2158 D. GEORGOPOULOS, C. ROUSSOS O a)6 Awake 2 * 40

* cmH 0

* aw P

5 -1 1 0 Flow L·s s

4 L tot

t 0.5 0 Volume Volume 3

Fig. 9. Ð Airway pressure (Paw), flow, and volume (inspiration positive) in a patient with obstructive lung disease ventilated on assist 0 volume-controlled mode. The arrow indicates the point at which constant inspiratory flow (V'I) increased from 30 to 90 Lámin-1 (tidal vol- 30 40 50 60 70 6050 40 30 ume was kept constant). Notice that within one breath after a change -1 V 'I L·min in V'I, total breath duration decreased considerably. This excitatory b)6 NREM sleep effect of V'I on the rate of inspiratory effort counterbalances the ben- eficial effect of high V'I on expiratory time. Behavioural feedback 5 * The effects of behavioural feedback on control of * * * * * breathing during mechanical ventilation are unpredictable, depending on several factors related to an indi- s 4 vidual patient, ventilatory settings and intensive care tot

t unit (ICU) environment [11, 12]. Ventilatory strategies intended to achieve a particular goal might be ineffec- tive in awake patients due to behavioural responses. For 3 example, it has been shown in mechanically-ventilated normal subjects, that both higher and lower than spontaneous inspiratory flow increases the sense of dysp- 0 noea [64]. Thus, in awake patients a change in inspiratory 30 40 50 60 70 6050 40 30 flow may cause discomfort and alteration in patient -1 effort. Similarly, increasing the assist level, which V 'I L·min inevitably increases the airway pressure [3, 4], may force Fig. 8. Ð Total breath duration (t 'tot) as a function of constant inspi- URBAN ratory flow (V'I) in normal subjects ventilated on assist volume-con- the patient to fight the ventilator. Indeed, J et al. trolled mode, during: a) wakefulness; and b) non-rapid eye movement [65], in patients with COPD, increased the pressure- (NREM) sleep. Both during wakefulness and NREM sleep, t tot decreas- support level and observed expiratory efforts, while the es in a graded and reversible manner as V'I increases and decreases, ventilation was still inflating the thorax. This neural- respectively. The response was attenuated by NREM sleep. (From mechanical dyssynchrony can be very uncomfortable, GEORGOPOULOS et al., [60], with permission). Bars are SEM. *: significant diference from ttot at 70 Lámin-1 (p<0.05). as is well-recognized with the use of inverse-ratio ventilation. Furthermore, active expiratory efforts in pati- affected by breathing route (nose or mouth), temperature ents with flow limitation during passive expiration cause and volume of inspired gas and anaesthesia of upper and dynamic compression in the airways downstream and lower airways [61]. Presumably, the excitatory effect of an unpleasant sensation [66]. Discomfort related to ven- inspiratory flow is mediated through receptors located tilatory settings may be manifested with rapid shallow deep in the airway mucosa or chest wall. breathing (panic reaction) leading to a vicious cycle There are at least four implications for the mechani- [67]. Finally, we should recognize that ventilatory set- cally-ventilated patient, as far as this reflex is concerned. tings that seem satisfactory during sleep, where behav- Firstly, an increase in assist level intended to decrease ioural feedback is absent, may become a source of respiratory effort is likely [37, 59] to be less effective discomfort during wakefulness with unpredictable effects than planned because of the stimulating effect of the on patient status. concomitant increase in flow. Secondly, high inspiratory flow rates may cause hyperventilation and respiratory Conclusion alkalosis, an important cause of various arrhythmias and weaning failure [62, 63]. Thirdly, the desired effect of To summarize, mechanical ventilation considerably flow on expiratory time [58] may not be achieved (fig. influences the control of breathing, as well as its expres- 9). Fourthly, the ventilatory consequences of flow are sion. During mechanical ventilation, the respiratory sys- likely to be different depending on sleep/awake stage. tem is under the influence of two pumps, the ventilator Collectively, these observations indicate that the excita- pump (Paw) and the patient's own respiratory muscles tory effect of flow rate may modify expected responses (Pmus). The physician dealing with a mechanically-ven- to change in ventilatory settings, thus, affecting thera- tilated patient should be aware that: 1) ventilatory out- peutic decisions. put may not reflect patient effort; and 2) various aspects CONTROL OF BREATHING IN MECHANICAL VENTILATION 2159 of control of breathing may be masked or modulated by 18. Pepe PE, Marini JJ. Occult positive end-expiratory pres- mechanical ventilation. Guidelines for mechanical venti- sure in mechanically-ventilated patients with airflow latory support must take into consideration the interact- obstruction. Am Rev Respir Dis 1982; 126: 166Ð170. ion between patient feedback (chemical, reflex, mechanical 19. Fleury B, Murciano D, Talamo C, Aubier M, Pariente and behavioural) and ventilator-delivered breath. R, Milic-Emili J. Work of breathing in patients with chronic obstructive pulmonary disease in acute respiratory failure. Am Rev Respir Dis 1985; 132: 822Ð827. 20. Rossi A, Gotfried SB, Higgs BD, Zocchi L, Grasino A, References Milic-Emili J. Respiratory mechanics in mechanically- ventilated patients with respiratory failure. J Appl Physiol 1. Younes M, Remmers J. Control of tidal volume and res- 1985; 58: 1849Ð1858. piratory frequency. In: Hornbein TF, ed. Lung Biology 21. Georgopoulos D, Giannouli E, Patakas D. Effect of in Health and Disease. Regulation of Breathing New extrinsic positive end-expiratory pressure on mechani- York, Marcel Dekker, 1981; 17: pp. 621Ð671. cally-ventilated patients with chronic obstructive pul- 2. Berger AJ. Control of breathing. In: Murray JF, Nadel monary disease and dynamic hyperinflation. Intensive JA, eds. Textbook of Respiratory Medicine. Philadelphia, Care Med 1993; 19: 197Ð203. W.B. Saunders, 1988; pp. 49Ð166. 22. Rossi A, Polese G, Brandi G, Conti G. Intrinsic posi- 3. Slutsky AS. Mechanical ventilation. ACCP consensus tive end-expiratory pressure (PEEPi). Intensive Care conference. Chest 1993; 104: 1833Ð1859. Med 1995; 21: 522Ð536. 4. Younes M. Proportional assist ventilation, a new approach 23. Younes M. Interactions between patients and ventila- to ventilatory support: theory. Am Rev Respir Dis 1992; tors. In: Roussos C, ed. Thorax. 2nd edn. Lung Biology 145: 114Ð120. in Health and Disease. Vol. 85, Chapter 81. New York, 5. Mead J, Agostoni E. Dynamics of respiration. In: Fenn Marcel Dekker, 1995; pp. 2367Ð2420. WO, Rahn H, eds. Handbook of Physiology. Section 3. 24. Fabry B, Guttmann J, Eberhard L, Bauer T, Haberthur Vol. I. Respiration. Washington, DC, American Physio- C, Wolff G. An analysis of desynchronization bet- logical Society, 1964; pp. 411Ð427. ween the spontaneous breathing patient and ventilator 6. Bigland B, Lippold OCJ. The relation between force, during inspiratory pressure support. Chest 1995; 107: velocity and integrated electrical activity in human mus- 1387Ð1394. cles. J Physiol (Lond) 1954; 123: 214Ð224. 25. Rossi A, Appendini L. Wasted efforts and dyssynchrony: 7. Grassino A, Goldman MD, Mead J, Sears TA. Mechanics is the patient-ventilator battle back? (Editorial). Inten- of the human diaphragm during voluntary contraction: sive Care Med 1995; 21: 867Ð870. statics. J Appl Physiol: Respirat Environ Exercise Physiol 26. Nava S, Bruschi C, Rubini F, Palo A, Iotti G, Braschi 1978; 44: 829Ð839. A. Respiratory response and inspiratory effort during 8. Cunningham DJC, Robbins PA, Wolff CB. Integration pressure support ventilation in COPD patients. Intensive of respiratory responses to changes in alveolar partial Care Med 1995; 21: 871Ð879. pressures of CO2 and O2 and in the arterial pH. In: 27. Younes M, Riddle W. Relation between respiratory Cherniack NS, Widdicombe JC, eds. Handbook of neural output and tidal volume. J Appl Physiol: Respirat Physiology. The Respiratory system. Vol. 2. Bethesda, Environ Exercise Physiol 1984; 56: 1110Ð1119. MD, American Physiological Society, 1986; pp. 475Ð 28. Agostoni E, Fenn WO. Velocity of muscle shortening 528. as a limiting factor in respiratory airflow. J Appl Physiol 9. Coleridge HM, Coleridge JCG. Reflexes evoked from 1960; 15: 349Ð353. tracheobronchial tree and lungs. In: Cherniack NS, 29. Agostoni E, Mead J. Statics of the respiratory system. Widdicombe JC, eds. Handbook of Physiology. The In: Fenn WO, Rahn H, eds. Handbook of Physiology. Respiratory System. Vol. 2. Bethesda, MD, American Section 3. Vol. I. Respiration. Washington, DC, American Physiological Society, 1986; pp. 395Ð430. Physiological Society, 1964; pp. 387Ð409. 10. Shannon R. Reflexes evoked from respiratory muscles 30. Roussos C, Macklem PT. The respiratory muscles. N and cortovertebral joints. In: Cherniack NS, Widdicombe Engl J Med 1982; 307: 786Ð797. JG, eds. Handbook of Physiology. The Respiratory sys- 31. Tobin MJ . Respiratory muscles in disease. Clin Chest tem. Vol. 2. Bethesda, MD, American Physiological Med 1988; 9: 263Ð286. Society, 1986; 431Ð438. 32. Roussos C, Campbell EJM. Respiratory muscles ener- 11. Killian KJ, Campell EJM. Dyspnea. In: Roussos C, getics. In: Macklem PT, Mead J, eds. Handbook of Macklem PT, eds. The Thorax. Lung Biology in Health Physiology. The Respiratory System. Mechanics of and Disease. Vol. 29. New York, Marcel Dekker, 1985; breathing. Vol. 2. Bethesda, MD, American Physiological pp. 787Ð928. Society, 1986; pp. 481Ð509. 12. Altose MD. Dyspnea. Curr Pulmonol 1986; 7: 199Ð226. 33. DeWeese EL, Sullivan TY, Yu PL. Ventilatory and 13. Sassoon CSH. Mechanical ventilation design and func- occlusion pressure responses to helium breathing. J Appl tion: the trigger variable. Respir Care 1992; 36: 815Ð828. Physiol: Respirat Environ Exercise Physiol 1984; 54: 14. MacIntyre NR. Respiratory function during pressure sup- 1525Ð1531. port ventilation. Chest 1986; 89: 677Ð683. 34. Hussain SNA, Pardy RL, Dempsey JA. Mechanical 15. Younes M. Patient-ventilator interaction with pressure- impedance as determinant of inspiratory neural driving assisted modalities of ventilatory support. Semin Respir during exercise in humans. J Appl Physiol 1985; 59: Med 1993; 14: 299Ð322 365Ð375. 16. Younes M, Puddy A, Roberts D, et al. Proportional assist 35. Marini JJ, Smith TC, Lamb VJ. External output and force ventilation: results of an initial clinical trial. Am Rev generation during synchronized intermittent mechanical Respir Dis 1992; 145: 121Ð129. ventilation. Am Rev Respir Dis 1988; 138: 1169Ð1179. 17. Marini JJ. Should PEEP be used in airflow obstruction? 36. Imsand C, Feihl F, Perret C, Fitting JW. Regulation of (Editorial). Am Rev Respir Dis 1989; 140: 1Ð3. inspiratory neuromuscular output during synchronized 2160 D. GEORGOPOULOS, C. ROUSSOS

intermittent mechanical ventilation. Anesthesiology 1994; 52. Meza S, Giannouli E, Younes M. Ventilatory response 80: 13Ð22. to inspiratory muscle unloading with PAV during sleep. 37. Giuliani R, Mascia L, Recchia F, Caracciolo A, Fiore Am J Respir Crit Care Med 1995; 151 A: 639. T, Ranieri VM. Patient-ventilator interaction during syn- 53. Clark FJ, Von Euler C. On the regulation of depth and chronized intermittent mandatory ventilation. Am J Respir rate of breathing. J Physiol 1972; 222: 267Ð295. Crit Care Med 1995; 151: 1Ð9. 54. Grunstein MM, Younes M, Milic-Emili J. Control of 38. Georgopoulos D, Mitrouska I, Webster K, Bshouty Z, tidal volume and respiratory frequency in anesthetized Younes M. Effects of respiratory muscle unloading on cats. J Appl Physiol 1973; 35: 463Ð476. the ventilatory response to CO2. Am J Respir Crit Care 55. Pack AI, Delaney RG, Fishman AP. Augmentation of Med 1995; 151: A639. phrenic nerve activity by increased rates of lung infla- 39. Patrick W, Webster K, Puddy A, Sanii R, Younes M. tion. J Appl Physiol: Respirat Environ Exercise Physiol Respiratory response to CO2 in the hypocapnic range in 1981; 50: 149Ð161. conscious humans. J Appl Physiol 1995; 79: 2058Ð2068. 56. Altose MD, Castele RJ, Connors AF Jr, Dimarco AF. 40. Sceid P, Lofaso F, Isabey D, Harf A. Respiratory response Effects of volume and frequency of mechanical venti- to inhaled CO2 during positive inspiratory pressure in lation on respiratory activity in humans. Respir Physiol humans. J Appl Physiol 1994; 77: 876Ð882. 1986; 66: 171Ð180. 41. Morrell MJ, Shea SA, Adams L, Guz A. Effects of inspi- 57. Simon PM, Skatrud JB, Badr MS, Griffin DM, Iber C, ratory support upon breathing during wakefulness and Dempsey JA. Role of airway mechanoreceptors in the sleep. Respir Physiol 1993; 93: 57Ð70. inhibition of inspiration during mechanical ventilation 42. Mancedo J, Isabey D, Lorino H, Lofaso F, Lemaire F, in humans. Am Rev Respir Dis 1991; 144: 1033Ð1041. Brochard L. Comparative effects of pressure support 58. Georgopoulos D, Mitrouska I, Markopoulou K, Patakas ventilation and intermittent positive pressure breathing D, Anthonisen NR. Effects of breathing patterns on (IPPB) in nonintubated healthy subjects. Eur Respir J mechanically-ventilated patients with chronic obstruc- 1995; 8: 1901Ð1909. tive pulmonary disease and dynamic hyperinflation. 43. Marantz S, Patrick W, Webster K, Roberts D, Oppen- Intensive Care Med 1995; 21: 880-886. heimer L, Younes M. Respiratory response to different 59. Ward ME, Corbeil C, Gibbons W, Newman S, Macklem levels of proportional assist (PAV) in ventilator-depen- PT. Optimization of respiratory muscle relaxation dur- dent patients. Am Rev Respir Dis 1992; 145: A525. ing mechanical ventilation. Anesthesiology 1988; 69: 44. Georgopoulus D, Giannouli E, Tsara V, Argiropoulou 29Ð35. P, Patakas D, Anthonisen NR. Respiratory short-term 60. Georgopoulos D, Mitrouska I, Bshouty Z, Webster K, poststimulus potentiation (after-discharge) in patients Anthonisen NR, Younes M. Effects of breathing route, with obstructive sleep apnea. Am Rev Respir Dis 1992; temperature and volume of inspired gas and airway anes- 146: 1250Ð1255. thesia on the response of respiratory output to varying 45. Georgopoulus D, Mitrouska I, Kolestos K, et al. Post- inspiratory flow. Am J Respir Crit Care Med 1996; 153: stimulus ventilation in patients with brain damage. Am 168Ð175. J Respir Crit Care Med 199; 152: 1627Ð1632. 61. Georgopoulos D, Mitrouska I, Bshouty Z, Anthonisen 46. Georgopoulus D, Bshouty Z, Younes M, Anthonisen NR, Younes M. Effects of NREM sleep on the response NR. Hypoxic exposure and activation of after-discharge of respiratory output to varying inspiratory flow. Am J mechanism in conscious humans. J Appl Physiol 1990; Respir Crit Care Med 1996; 153: 1624Ð1630. 69: 1159Ð1164. 62. Ayres SM, Crace WJ. Inappropriate ventilation and 47. Younes M. The physiologic basis of central apnea. Curr hypoxemia as causes of cardiac arrhythmias: the con- Pulmonol 1989; 10: 265Ð326. trol of arrhythmias without antiarrhythmic drugs. Am J 48. Fink BR, Hanks EC, Ngai SH, Papper EM. Central reg- Med 1969; 46: 495Ð505. ulation of respiration during anaesthesia and wakeful- 63. Pingleton SK. Complications of acute respiratory fail- ness. Ann NY Acad Sci 1963; 109: 892Ð899. ure. Am Rev Respir Dis 1988; 137: 1463Ð1493. 49. Skatrud JB, Dempsey JA. Interaction of sleep state and 64. Manning HL, Molinary EJ, Leiter JC. Effect of inspi- chemical stimuli in sustaining rhythmic respiration. J ratory flow rate on respiratory sensation and pattern of Appl Physiol: Respirat Environ Exercise Physiol 1983; breathing. Am J Respir Crit Care Med 1995; 151: 55: 813Ð822. 751Ð757. 50. Datta AK, Shea SA, Horner RL, Guz A. The influence 65. Jurban A, Van De Graaff WB, Tobin MJ. Variability of of induced hypocapnia and sleep on the endogenous res- patient-ventilator interaction with pressure support ven- piratory rhythm in humans. J Physiol 1991; 440: 17Ð tilation in patients with chronic obstructive pulmonary 33. disease. Am J Respir Crit Care Med 1995; 152: 129Ð136. 51. Rebuck AS, Slutsky AS. Control of breathing in dis- 66. O'Donnell D, Anthonisen NR, Sanii R, Younes M. Effect eases of the respiratory track and lungs. In: Cherniack of dynamic airway compression on breathing pattern NS, Widdicombe JC, eds. Handbook of Physiology. The and respiratory sensation in severe COPD. Am Rev Respir Respiratory System. Control of Breathing. Vol. II. Part Dis 1987; 135: 912Ð919. 2. Bethesda, MD, American Physiological Society, 1986; 67. Marcy TW, Marini JJ. Respiratory distress in the ven- pp. 771Ð791. tilated patient. Clin Chest Med 1994; 15: 55Ð73.