help button home button
AJRCCM
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Online Data Supplement
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Younes, M.
Right arrow Articles by Roberts, D.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Younes, M.
Right arrow Articles by Roberts, D.
American Journal of Respiratory and Critical Care Medicine Vol 166. pp. 21-30, (2002)
© 2002 American Thoracic Society

Original Article

Response of Ventilator-Dependent Patients to Delayed Opening of Exhalation Valve

Magdy Younes, John Kun, Kim Webster and Dan Roberts

Sections of Respiratory and Critical Care Medicine, Department of Medicine, University of Manitoba, Winnipeg, Canada

Correspondence and requests for reprints should be addressed to Dr. M. Younes, 3611-55 Harbour Square, Toronto, ON, M5J 2L1 Canada. E-mail: mkyounes{at}sympatico.ca


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
In conventional mechanical ventilation, the inflation cycle often extends into neural expiration (TEN), potentially exacerbating dynamic hyperinflation (DH). We wished to determine the extent to which patients defend against DH when this happens. Such defense may include prolongation of TEN (timing response) and/or expiratory muscle recruitment (neuromuscular response). Fifty patients were ventilated in the Proportional Assist mode, allowing us to infer these responses noninvasively. At random intervals, exhalation of single breaths was delayed by briefly delaying the opening of exhalation valve (occlusion) (Tocc = 0.78 ± 0.34 seconds). Timing response was assessed from the change in TEN. Neuromuscular response was assessed from the difference between volume exhaled after release of occlusion and volume exhaled in unoccluded breaths over a similar expiratory flow duration ({Delta}Viso-time). There was no evidence of an acute neuromuscular response; {Delta}Viso-time averaged 0.005 ± 0.023 L (NS). Forty-five of 50 patients significantly lengthened TEN. However, the timing response offset only 36 ± 20% of the delay in expiration. Because of absent neuromuscular responses and weak timing responses, DH increased in most patients in postocclusion breaths ({Delta}DH = 0.10 ± 0.08 L, p = 2E-10). We conclude that acute compensatory responses to delays in opening of exhalation value are weak in ventilator-dependent patients. As a result, such nonsynchrony tends to exacerbate DH.

Key Words: Hering–Breuer reflex • control of breathing • patient–ventilator interactions • dynamic hyperinflation • ineffective efforts


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 
In pressure support ventilation (PSV) and assisted volume-cycled ventilation, the end of the ventilator's inflation cycle is not synchronized with the end of the patient's inspiratory effort. As a result, inflation may continue into the phase of neural expiration (1, 2). Although there has been no systematic documentation of the frequency and extent of this occurrence, theoretical analysis using realistic ventilator settings and patient characteristics (1), several published records (e.g., references 3–6), and unpublished personal observations on dozens of patients make it clear that this type of dys-synchrony is very common and often quite pronounced; the ventilator is triggered near the end of inspiratory effort, with the inflation cycle falling almost entirely during neural expiration (e.g., Figure 1 in reference 5). When the ventilator cycle extends into neural expiration, the time available for expiratory flow, before the next inspiratory effort, is reduced. If passive functional residual capacity is not reached during the abbreviated expiratory phase, dynamic hyperinflation (DH) results or is aggravated. DH increases the patient's work of breathing and makes it difficult for the patient to trigger the ventilator (7). The most obvious manifestation of the latter is the occurrence of ineffective efforts (8–11), where a variable number of respiratory efforts fail to trigger the ventilator.



View larger version (59K):
[in this window]
[in a new window]
  		Figure 1. Tracings of Paw, flow, and volume showing one occluded breath (second breath) and the various time markers and primary measurements. Vertical solid lines represent time markers as follows: TendI, end of inspiration defined from point of flow zero crossing; Tendocc, end of occlusion, the time at which Paw and flow decline sharply; ToC, Toocc, ToD, time of onset of inspiratory effort in control, occluded, and postoccluded breaths, respectively (see text for method of identification); TtrC, Ttrocc, TtrD, time of ventilator triggering, the time at which Paw begins rising in "control," occluded, and postoccluded breaths, respectively; Tocc, duration of occlusion; TEM, duration of expiratory flow between end of occlusion and onset of next effort; TIN,C, TEN,C, neural inspiratory and expiratory durations in control breaths, respectively; TIN,occ, TEN,occ, neural inspiratory and expiratory durations in occluded breaths; V·o, expiratory flow at Toocc; V·-0.5, expiratory flow 0.5 seconds before Toocc; VTC, VTocc, inspired volume in control and occluded breaths, respectively; VexpC, Vexpocc, exhaled volume, up to To, in control and occluded breaths, respectively; Vexp@iso-time, exhaled volume in control breaths over a period corresponding to TEM of occluded breaths; {Delta}VDH, difference in EEV between control and occluded breaths. Tdelay (trigger delay) the interval between To and Ttr, is not illustrated.

 
In laboratory animals (12–14), awake normal humans (15–19), and ambulatory patients with chronic obstructive pulmonary disease (COPD) (20), imposed delays in lung emptying during expiration evoke compensatory responses in the form of prolongation of neural expiration (TEN) and, frequently, recruitment of expiratory muscles (21, 22). These responses mitigate the DH that would otherwise result: the former by making more time available for expiratory flow, and the latter through increasing expiratory flow in the available time. There is only very limited information ([23], see DISCUSSION) about whether, or to what extent, mechanically ventilated patients use these responses to defend against DH when the ventilator cycle extends into neural expiration.

In the present communication, we report on the neuromuscular and expiratory timing responses of ventilator-dependent patients after deliberate delays in the opening of the exhalation valve (end-inspiratory occlusions) in the proportional assist ventilation (PAV) mode (24). In this mode, the end of the inflation phase is automatically synchronized with the end of the patient's inspiratory effort. This makes it possible to obtain the desired information noninvasively. Thus, any prolongation in TEN in response to the occlusions is expressed as an increase in the interval between end of inspiratory phase (i.e., beginning of the occlusion) and onset of next inspiratory effort. Furthermore, by comparing expiratory flow after the release of occlusion with expiratory flow after the opening of the exhalation valve in unoccluded breaths, it is possible to make inferences regarding neuromuscular responses that affect expiratory flow. A preliminary report of these findings has been published (25).


   METHODS
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
RESULTS
 DISCUSSION
 REFERENCES
 
We studied 50 patients who were intubated for acute respiratory failure. Further details of these methods may be found in the online data supplement.

Apparatus
We used a locally constructed, piston-based, multimode ventilator (Winnipeg Ventilator, Winnipeg, MB, Canada) (26). Flow (V·) and airway pressure (Paw) were measured. Data were digitized and stored on a computer disk.

Protocol
Preliminary estimates of passive elastance (E) and resistance (R) were made at the bedside using the interrupter technique (27) during a period of controlled ventilation (CMV). The patient was then switched to the PAV mode with volume assist (VA) and flow assist (FA) at 80% of the passive E and R. Subsequently, brief (0.4–1.6 seconds) end-inspiratory occlusions were applied to single breaths at random intervals.

Analysis
CMV data. The stored data were examined to determine passive E and inspiratory resistance (Rinsp) using the interrupter technique. In addition, we measured expiratory resistance (Rexp) at different volumes using a subtraction technique. The difference between Rexp and Rinsp (i.e., {Delta}R) was used as an index of flow limitation. (See section A1 in the online data supplement).

PAV data. Figure 1 illustrates the various variables measured. Computer files were inspected by one of the investigators (M.Y.). Event markers were manually placed at the onset of inspiratory effort (T0) in the three breaths preceding occlusion and the breath following occlusion. T0 was defined as the time at which the trajectories of flow and Paw deviated conspicuously, in an inspiratory direction, from the trajectories established earlier in expiration (see section B in the online data supplement). Markers were also placed at several other locations as illustrated in Figure 1.

A custom program scanned the files and tabulated several variables from each marked occlusion. With two exceptions, V·exp slope and Tdelay, these variables are defined in Figure 1. V·exp slope is the average rate of change in expiratory flow in the last 0.5 seconds of TEN,C, and was obtained from ([V·0 - V·-0.5]/0.5). Tdelay, or trigger delay, was the time difference between onset of inspiratory effort (T0) and time of triggering (Ttr).

The following responses were calculated from each occlusion (see online data supplement):

  1. Neuromuscular response: the difference between Vexpocc and Vexp@iso-time, expressed as {Delta}Viso-time. The rationale of this measurement is explained in the detailed METHODS section (see online data supplement).
  2. Timing response: the difference between TEN,occ and TEN,C, expressed as {Delta}TEN.
  3. Impact of occlusion on DH: the difference between VexpC and Vexpocc, expressed as {Delta}VDH in liters (Figure 1). This value was multiplied by elastance to obtain the corresponding increase in elastic recoil pressure.
  4. Impact of occlusion on trigger delay: the difference between Tdelay in the breath following occlusion and Tdelay of breaths preceding occlusion.
  5. Impact of occlusion on respiratory rate (RR): the difference between RR of occluded and control breaths.

In 12 patients, there was a sufficiently wide range of occluded VT, and a sufficiently large number of occlusions, to permit separate analysis of responses at different occlusion volumes in the same patients (see section A2.1 in the online data supplement).

Statistical Analysis
t Test, analysis of variance, and multiple regression analysis were used as needed (see online data supplement).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
DISCUSSION
 REFERENCES
 
The 50 patients ranged in age from 19 to 83 years (63 ± 17), in height from 1.51 to 1.91 m (1.66 ± 0.1), and in body mass index from 14.2 to 41.0 kg/m2 (25.8 ± 6.1). External positive end-expiratory pressure (PEEP) was 5.5 ± 2.2 cm H2O. The passive elastance of the patients ranged from normal (12.9 cm H2O/L) to extremely high (89.0 cm H2O/L) (mean ± standard deviation [SD]: 29.8 ± 13.9), and their Rinsp ranged from a normal 3.6 to a very high 22.8 cm H2O/(L/second) (10.0 ± 4.8). The spectrum of pathophysiology was, accordingly, very wide.

Figure 2 shows the difference between expiratory and inspiratory resistance (i.e., {Delta}R), plotted against Rinsp, in individual patients. Six patients had {Delta}R values that were clearly outside the range observed in the others. Of these, the four patients with the highest values also had other features associated with severe COPD, namely an early expiratory flow spike due to dynamic compression of the airway (28, 29) and a very depressed flow rate throughout expiration, beyond the initial spike (see Figure E1 in the online data supplement). These patients will henceforth be termed to have "severe COPD." In the other two patients in this group, the increase in Rexp was limited to low volumes, and there were no other manifestations of COPD. These patients will be referred to as having "possible COPD," because flow limitation limited to the low volume range need not reflect obstructive disease but may be observed, in the absence of such disease, when FRC is low (e.g., in acute respiratory distress syndrome [30]). In the remaining 44 patients, {Delta}R averaged 0.9 ± 1.8 cm H2O/L/second, and none displayed an early expiratory flow spike. The horizontal dashed lines in Figure 1 represent our estimate of the range of {Delta}R values that could be explained by technical errors.* This estimate is based on the fact that Rexp cannot, physiologically, be less than Rinsp. Thus, negative {Delta}R values represent estimate errors and the most negative value (-4.5 cm H2O/L/second) provides the outer boundary of negative errors in this group. A corresponding positive error boundary was assumed to exist.



View larger version (15K):
[in this window]
[in a new window]
  		Figure 2. Difference between expiratory and inspiratory resistance in individual patients during controlled ventilation, plotted against inspiratory resistance. The horizontal dashed lines are estimates of the boundaries for technical errors in {Delta}R estimates.

 
A total of 1,745 occlusions was analyzed in 50 patients (34.9 ± 14.2 per patient). The average (± SD) duration of occlusions was 0.78 ± 0.33 seconds. Average tidal volume in control breaths was 0.49 ± 0.14 L, with a range of 0.21–0.89 in different patients. RR and minute ventilation in control breaths were 24.2 ± 7.1 per minute, and 11.5 ± 3.7 L/minute, respectively. TIN and TEN in control breaths were 1.12 ± 0.26 and 1.61 ± 0.84 seconds, respectively. The inspiratory phase of occluded breaths (i.e., before the occlusions) did not differ from that in preceding "control" breaths in any respect (i.e., VT, TIN) in any patient. The elastic recoil pressure (including PEEP) during occlusions ranged from 8.4 to 40.2 cm H2O (19.1 ± 5.9 cm H2O) in different patients.

The patients displayed a wide range of breath-by-breath variability in breathing pattern. The SD of VT ranged from 0.02 to 0.17 L (0.08 ± 0.04) in different patients. This resulted in coefficients of variation ranging from 0.05 to 0.39 (0.17 ± 0.08). In 12 patients, the SD of VT was more than 0.1 L, and there were more than 40 observations in total, thereby permitting, after separation into VT quartiles, analysis of the effect of occlusion VT on responses in the same patients. None of these patients had evidence of COPD.

Neuromuscular Response to Delaying Mechanical Expiration
The volume exhaled after release of the occlusion before onset of the next inspiratory effort (Vexpocc) averaged (± SD) 0.37 ± 0.14 L. The volume exhaled at iso-time in control breaths (Vexp@iso-time) was 0.37 ± 0.13 L. There was no significant difference between the two values, indicating that, on average, there was no effective neuromuscular response to increase expiratory flow in the aftermath of the occlusion. Figure 3 is a scatter plot of the two values in individual patients. It can be seen that the data points are scattered very close to the line of identity; the difference between the two values (i.e., {Delta}Viso-time) was 0.005 ± 0.023 L.



View larger version (13K):
[in this window]
[in a new window]
  		Figure 3. Relation between volume expired after release of occlusion and volume expired during control breaths over the same duration as the duration of flow postocclusion (see Figure 1). The diagonal line is the line of identity. Each dot is a separate patient. A value above the line of identity signifies an increase in forces that promote expiratory flow in the postocclusion period relative to control breaths.

 
Figure 4 (dashed line) shows {Delta}Viso-time as a function of occlusion volume in the 12 patients in whom this response could be determined over a range of occlusion volumes. {Delta}Viso-time remained very small at all volumes in all patients, with the average response being essentially 0 at all volumes.



View larger version (18K):
[in this window]
[in a new window]
  		Figure 4. Average responses in 12 patients in whom responses were determined at different occlusion volumes. Solid line: timing response, expressed as the ratio of prolongation in TEN to occlusion duration. Dashed line: neuromuscular response, expressed as the difference between volume exhaled after occlusion and volume exhaled in control breaths over the same duration ({Delta}Viso-time, right axis). Bars are standard errors of the mean.

 
Response of TEN to Delaying Mechanical Expiration
The average TEN in occluded breaths (TEocc) was 1.91 ± 0.91 seconds. The corresponding value for control breaths was 1.61 ± 0.84 seconds. The difference (0.31 ± 0.27 seconds) was significant (p = 4E - 11). TEocc was significantly longer than control TEN in 45 of 50 patients.

The relation between {Delta}TEN and the duration of occlusion (Tocc) is shown in Figure 5A . In general, the more the onset of expiratory flow was delayed (Tocc), the more TEN was prolonged ({Delta}TEN). However, there was considerable scatter, with the ratio of {Delta}TEN to Tocc ranging from -0.12 to 0.93. The average (± SD) ratio was 0.36 ± 0.20, indicating that, on average, TE prolongation neutralized only a third of the delay in onset of expiratory flow. Figure 6 contrasts the results of two patients, one (Figure 6A) with a strong timing response ({Delta}TEN/Tocc = 0.8) and the other with a weak response ({Delta}TEN/Tocc = 0.08).



View larger version (15K):
[in this window]
[in a new window]
  		Figure 5. (A) Relation between the duration of occlusion and the extent of prolongation of neural expiratory time in different patients. Isopleths describe different strengths of the TE response, expressed as the ratio of {Delta}TEN/Tocc. A ratio of 1.0 indicates complete compensation for the occlusion. (B) The magnitude of reduction in the duration of expiratory flow as a function of occlusion time in individual patients. The solid line is the line of identity and indicates the response expected in the absence of changes in TEN. Note that all patients sustained a reduction in the duration of expiratory flow, although the magnitude was quite variable among patients. Open symbols (n = 4) are patients with severe COPD. Shaded symbols (n = 2) are patients with flow limitation at low lung volumes.

 


View larger version (77K):
[in this window]
[in a new window]
  		Figure 6. Examples from two patients illustrating very strong (A) and very weak TEN responses. Note that in panel A the duration of expiratory flow was little affected by the long occlusion. The extent of DH (difference between exhaled volume of the occluded and pre-occluded breaths) is fairly small in both cases. In panel A this is because of the substantial prolongation of TEN, which preserved TEM. The minimal change in DH in panel B was because expiratory flow rate in the terminal part of unoccluded breaths was very low, so that reduction in the duration of expiratory flow had little consequence in terms of volume.

 
The response of the four patients with severe COPD was significantly lower than that of the others ({Delta}TEN/Tocc= 0.09 ± 0.18 versus 0.38 ± 0.19, p = 0.002) (open symbols, Figure 5). The response of the two patients with possible COPD was not different (shaded symbols, Figure 5).

The 12 patients in whom analysis of the effect of different occlusion volumes was possible showed a wide range of timing responses; {Delta}TEN/Tocc (computed from all occlusions in the same patient) ranged from -0.01 to 0.93. Figure 4 (solid line) shows that, on average, the timing response increased progressively with occlusion volume (p = 0.008 by analysis of variance for repeated measures). There was also a good correlation between individual responses at low and high occlusion volumes (r = 0.79, p = 0.02), indicating that patients with weak responses at low volumes continued to have weak responses at high volumes and vice versa.

Determinants of Strength of the Timing Response
Correlation analysis was performed between {Delta}TEN/Tocc as the independent variable and the following 10 variables: age, sex, body mass index, VT, VE, TI/TE ratio of control breaths, the rapid shallow breathing index (RR/VT), elastance (Ers), {Delta}R (i.e., Rexp - Rinsp), and Paw during the occlusion (after subtracting PEEP), representing the change in elastic recoil pressure associated with the occluded VT (i.e., Pocc). Only RR/VT, TI/TE, Pocc, and {Delta}R showed significant correlations (p < 0.05) with {Delta}TEN/Tocc in univariate analysis. When all 10 variables were entered in stepwise regression analysis, the only variables that were not eliminated were RR/VT, which showed a negative correlation (coefficient = -0.003 fraction of Tocc per unit change in RR/VT, p < 0.01), {Delta}R, which also showed a negative correlation (coefficient = -0.008 fraction of Tocc per unit difference in resistance, p < 0.02), and Ers, which was positively correlated (coefficient = 0.005 fraction of Tocc per unit elastance, p < 0.05). The overall correlation coefficient for the final equation was 0.53 (r 2 = 0.28).

Effect of Occlusion on RR
The prolongation of TEN necessarily resulted in a reduction of RR of the occluded breath. Figure 7 is a scatter plot of the change in RR as a function of Tocc. Virtually all patients sustained a reduction in RR. {Delta}RR was quite variable even at the same Tocc. In 10 patients (20%) it exceeded 6 breaths/minute. On average, {Delta}RR was -2.4 ± 2.4/minute (p < 0.0001).



View larger version (17K):
[in this window]
[in a new window]
  		Figure 7. Change in RR ({Delta}RR) during occluded breaths, relative to control breaths, in individual patients. Results are plotted as a function of occlusion time.

 
Effect of Occlusion on Mechanical Expiratory Time
Because the prolongation in TEN was less than the duration of occlusion, the time available for expiratory flow (TEM) was invariably reduced. Figure 5B shows the magnitude of reduction in TEM in different patients as a function of Tocc. The solid line is the line of identity and shows the relation in the absence of prolongation in TEN. In virtually all patients the reduction in TEM was less than Tocc, reflecting the prolongation in TEN described earlier (Figure 5A). Although the reduction in TEM was correlated with Tocc (r = 0.62, p < 0.0001), there was much scatter, particularly with long Tocc. Thus, with a Tocc of approximately 1.4 seconds, the reduction in TEM was as little as 0.09 seconds or as large as 0.9 seconds (Figure 5B). On average, TEM decreased by 0.46 ± 0.20 seconds for an average Tocc of 0.77 ± 0.34 seconds.

Expiratory Flow in the Terminal Part of Expiration
The extent to which a given reduction in TEM results in DH depends critically on the magnitude of expiratory flow in the terminal part of expiration (V·expterminal). For example, if V·expterminal is close to 0, a 0.5-second reduction in TEM would not appreciably increase DH. Conversely, if V·expterminal is high, small reductions in TEM may result in an important increase in DH. Thus, V·expterminal determines the susceptibility to DH in response to changes in TEM.

Expiratory flow at the onset of inspiratory efforts before occlusion (V·0, Figure 1) ranged from 0 to -0.42 L/second (-0.17 ± 0.12 L/second). Expiratory flow normally decreases in the latter part of expiration (e.g., Figures 1 and 6). To determine the average expiratory flow in that part of expiration corresponding to {Delta}TEM (i.e., V·expterminal), we calculated the slope of expiratory flow in the last 0.5 seconds of expiration{dagger} in control breaths (V·expslope) and estimated V·expterminal from (V·0 - 0.5 [{Delta}TEM x V·expslope]). The average slope was 0.24 ± 0.18 L/second2. V·expterminal ranged from -0.02 to -0.63 (-0.23 ± 0.14) L/second, reflecting a wide range in susceptibility to DH among patients.

Effect of Occlusion on DH
Figure 8 shows the effect of delaying onset of expiratory flow on the magnitude of {Delta}DH in individual patients. {Delta}DH is expressed in volume units ({Delta}VDH, Figure 8A), and in pressure units ({Delta}PDH, Figure 8B). End-expiratory lung volume immediately after the occluded breath increased, on average, 0.10 ± 0.08 L (Figure 8A), corresponding to 2.2 ± 1.9 cm H2O in elastic recoil (Figure 8B). There was no significant correlation between Tocc and {Delta}DH whether the latter was expressed as volume or as pressure.



View larger version (12K):
[in this window]
[in a new window]
  		Figure 8. Change in DH as a function of the duration of occlusion. In panel A this is expressed in volume units, whereas in panel B it is expressed in pressure units. Note the lack of correlation between either variable and Tocc.

 
Effect of Delaying Expiratory Flow on Trigger Delay
The average trigger delay (interval between onset of inspiratory effort and ventilator triggering, Tdelay) in control breaths was 0.23 ± 0.10 seconds. In 17 patients, some of the inspiratory efforts after the occluded breaths failed to trigger the ventilator. In these patients, the percent of ineffective postocclusion efforts ranged from 3 to 81% (18 ± 19%). The average percent ineffective postocclusion efforts in the entire group was 6 ± 13%. For all triggered postocclusion efforts the average trigger delay was 0.29 ± 0.12 seconds. This was 0.06 ± 0.05 seconds longer than in control breaths (p < 0.0001). Figure 9 is a scatter plot of the magnitude of change in trigger delay ({Delta}Tdelay) as a function of Tocc.



View larger version (15K):
[in this window]
[in a new window]
  		Figure 9. Change in trigger delay ({Delta}Tdelay) following occluded breaths, relative to control breaths, in individual patients. Results are plotted as a function of occlusion time. Trigger delay is the interval between onset of inspiratory effort (T0, Figure 1) and ventilator triggering and includes the time required to reduce expiratory flow to 0 in the presence of DH.

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
REFERENCES
 
The main findings are that when, in ventilator-dependent patients, the onset of expiratory flow is delayed relative to the onset of neural expiration, (1) there is no effective neuromuscular response to speed up expiratory flow in the abbreviated time available for expiration; (2) there is compensatory prolongation in TEN, but the response is generally weak offsetting, on average only 36% of the initial delay in onset of expiratory flow; (3) as a result of weak responses, most patients sustain an increase in DH and trigger delay; and (4) the timing and DH responses are highly variable among patients.

Technical Considerations
The extent of muscular response to the expiratory occlusion was assessed from the difference between volume exhaled after release of occlusion and volume exhaled over a corresponding duration, beginning with the onset of TEM, in un-occluded breaths. The two durations occurred at different times within TEN (late TEN in occluded breaths and early TEN in control breaths, Figure 1), and volume history was different in the two cases. Because muscular pressure (Pmus) may normally change in the course of neural expiration,{ddagger} and viscoelastic behavior may have resulted in different elastic recoil pressure (Pel) at the onset of expiratory flow, it can be argued that the index we used may have missed an important increase in expiratory muscle pressure in response to the occlusion. We believe this is extremely unlikely because the two responses (change in Pmus and in passive recoil) tend to cancel each other in the course of occlusion. Thus, the decrease in inspiratory Pmus should, at the same volume, cause a progressive increase in Palv, whereas viscoelastic behavior should result in a lower Palv. In a separate analysis of the same data (32), we calculated the time course of change in alveolar pressure (Palv) in the course of these occlusions. The change in Palv during occlusion represents the net effect of changes in Pmus and Pel. We found that at high levels of assist, such as those used here (~ 80%), the two changes cancel out; the change in Palv beyond the first 0.05 seconds of occlusion was 0.0 ± 1.05 cm H2O (e.g., see Paw in Figures 1 and 6). It follows that there should have been no systematic difference in the driving pressure at the onset of expiratory flow (i.e., Palv) in occluded and unoccluded breaths, and any differences that may have existed in individual patients were small and bidirectional.

In the presence of flow limitation (FL), recruitment of expiratory muscles will not result in an increase in expiratory flow. It may be argued, therefore, that expiratory recruitment could have been missed because of the presence of expiratory FL. We believe that this is very unlikely for the following reasons: (1) The difference between Vexpocc and Vexp@iso-time was minimal in all patients (Figure 3). For expiratory FL to explain this finding, all patients must have had FL, a possibility that is extremely unlikely in a general medical intensive care unit (ICU) and which was discounted by the {Delta}R results (Figure 2), or expiratory recruitment occurred exclusively in patients with FL. There is no functional or known neurophysiologic reason to expect this selectivity in response. (2) The indexes used (Vexpocc and Vexp@iso-time) represent the integration of flow over a substantial period (TEM of occluded breaths), beginning with the onset of expiratory flow where lung volume is highest and FL is least likely. For there to be no difference, despite expiratory recruitment, FL must have been present throughout expiration. This occurs only in patients with very severe COPD. Only four patients (8%) could be placed in this category. Although we cannot exclude expiratory recruitment in these patients, it would be extremely unlikely that such recruitment would occur exclusively in patients with very severe COPD. In fact, given that these patients demonstrated a smaller timing response, one would not expect them to display a stronger neuromuscular response; at least in laboratory animals, timing and neuromuscular responses appear to be linked (22).

It is difficult to envision a situation, in the absence of FL, where an important increase (e.g., > 2 cm H2O) in expiratory muscle pressure related to the occlusion could have been missed. In the patient who had the largest difference between Vexpocc and Vexp@iso-time (0.049 L, Figure 3), the increase in mean expiratory flow in the postocclusion period over the corresponding duration in control breaths was 0.08 L/second. Given this patient's resistance (8.0 cm H2O/L/second, including tubing), the increase in mean Vexp could be accounted for by a difference in net forces of no more than 1 cm H2O. Given that the differences between Vexpocc and Vexp@iso-time were smaller in all patients (mean ± SD = 0.005 ± 0.023 L), we believe we are justified in concluding that the patients did not mount a functionally important expiratory muscle response.

The aforementioned conclusion does not entirely rule out an increase in the electrical activity of expiratory muscles. Although demonstration of such a response, if any, may be of interest from a neurophysiologic perspective, it would not be interpretable in functional terms without the demonstration of a change in expiratory muscle pressure; the units by which electrical activity of respiratory muscles is measured (volts) are, per se, meaningless in terms of mechanical action. We elected to proceed directly to a mechanical index because we were primarily interested in functionally important responses.

Neural timing was also estimated indirectly from mechanical variables. The end of neural expiration was identified as the point at which expiratory flow and Paw deviated, in an inspiratory direction, from their earlier trajectories. According to the equation of motion (Pmus = V· · R + V · E - Paw) such a change could only occur if Pmus changed in an inspiratory direction (see section B in the online data supplement). Although this could occur either because of an increase in inspiratory Pmus or, in the presence of phasic expiratory activity, decline in expiratory Pmus, in practice either event defines the end of neural expiration.

The point at which inspiratory flow ended was used to define the onset of TEN. In PAV, this point invariably occurs during the declining phase of inspiratory Pmus (24, 32). The validity of this approach depends on how one defines TIN and TEN. If, as is customary, the transition from TIN to TEN is defined from the point at which inspiratory activity begins declining rapidly (33), then our reported absolute values of TIN slightly overestimate TIN, and the reported TEN slightly underestimates TEN. Such discrepancies between reported and actual absolute values of TIN and TEN should, however, have no impact on the variable of interest (namely, change in TEN in response to occlusion). Because the occlusion began at the end of inspiratory flow, and there was no intervention before this point, any difference between reported and actual absolute TEN was the same for "control" and occluded breaths. Thus, any real change in TEN in occluded breaths was faithfully expressed in our reported {Delta}TEN.

Neuromuscular Response
In anesthetized animals, occlusion during expiration is associated with expiratory muscle activation (21, 22). Such recruitment may, therefore, have been expected in the current study. The lack of a functionally important increase in expiratory pressure is probably related to the fact that the increase in activity observed in animals is strictly related to the vagally mediated prolongation of TEN, which permits the ramping phasic expiratory activity to progress to a higher level (22). In our patients, the prolongation of TEN was modest. Furthermore, inflation during expiration actually delays the onset of phasic expiratory activity when the latter is present in unobstructed breaths (22). Thus, there are two opposing responses to lung inflation during expiration, one (delayed onset) that tends to decrease expiratory activity at comparable times during TEN, and another that tends to increase it through TEN prolongation. The likely explanation for our finding is that the two responses canceled each other out. Our findings do not preclude recruitment of expiratory muscles in the steady state. Delaying onset of expiratory flow results in respiratory slowing (Figure 7) and, unless VT is concomitantly increased, reduction in V·E. Expiratory muscles may be recruited later as a result of the increase in chemical drive. Our results indicate that if delayed expiration, when sustained breath after breath, is associated with expiratory muscle recruitment, the recruitment is chemically, and not reflexly, mediated.

The absence of a functionally important neuromuscular response was evident in all patients, even though in many patients elastic recoil pressure during the occlusion was more than 20 cm H2O, and in some it was greater than 30 cm H2O (Paw during occlusion was 19.1 ± 5.9 cm H2O). These pressures are associated with volumes near the patient's pathophysiologic total lung capacity. That a response was absent in all these patients, and at the largest volumes in 12 patients (Figure 4), indicates that the negative results were not due to failure to reach a high threshold for recruitment of expiratory muscles.

Timing Responses
In the absence of significant expiratory muscle recruitment, compensation for the delayed onset of mechanical expiration is entirely dependent on prolongation of TEN. Although statistically significant prolongation occurred in 45 of 50 patients, the response was incomplete, and fairly weak in most patients (Figure 5A). On average TEN prolongation countered less than 40% of the initial delay. As a result, the duration of mechanical expiration decreased in nearly all patients (Figure 5B).

To our knowledge, there has been only one study that assessed the response of TEN to changes in timing of ventilator cycling off, relative to end of neural TI, in mechanically ventilated patients. Kondili and coworkers measured the change in TEN when VT was increased at constant flow (i.e., mechanical TI was increased [23]). They found that the increase in TEN was, on average, equal to the increase in delay between ventilator cycling off and end of neural TI. On the basis of their findings, one may conclude that the timing response is powerful and results in complete compensation. This is clearly different from what was seen here ({Delta}TEN/Tocc = 0.36 ± 0.2). Several factors can account for this difference. First, they only studied patients with acute respiratory distress syndrome. We studied a larger population with assorted clinical disorders, representative of the general ICU population. Second, the range of delays examined by Kondili and coworkers (23) was very limited. With the longest delay, the interval between end of ventilator TI and end of neural TI (defined as onset of decline in inspiratory Pmus) was only 0.44 seconds. In our study, delays ranged from 0.4 to 1.6 seconds, measured from zero flow crossing. When it is considered that flow crossing in PAV occurs on the declining phase of Pmus, sometime beyond the onset of decline in inspiratory Pmus (32), the range of delays examined here is well beyond that examined in their study. Third, baseline TEN, with which the TEN of the delayed breaths was compared, was associated with a smaller VT. It is, therefore, not possible to ascertain how much of the increase in TEN was related to increased delay and how much to the increased VT, relative to baseline; an increase in VT, even without lengthening ventilator TI, also lengthens TEN (19, 34, 35).

The timing response was highly variable among patients. Correlation analysis with many potentially relevant variables revealed a significant effect for only three: elastance, {Delta}R, and the rapid shallow breathing index. Collectively, however, these variables accounted for only 28% of the scatter. It follows that the strength of this compensatory response is largely unpredictable.

The correlation with the rapid shallow breathing index (RR/VT) was negative. Thus, patients with rapid shallow breathing are, all else being the same, less equipped to deal with delays in onset of expiratory flow. The multiple regression coefficient was -0.003. Quantitatively, this translates into an average difference in {Delta}TEN/Tocc of 0.3 across the range of RR/VT observed (RR/VT averaged 55 ± 30 per minute per liter). Thus, whereas in an average patient with RR/VT of 55, {Delta}TEN is 36% of Tocc, a patient with RR/VT of 125 would have, on average, a {Delta}TEN of only 15% of Tocc. The mechanism of this association between RR/VT and {Delta}TEN/Tocc is not clear.

The most obvious mechanism to explain TEN prolongation is the Hering–Breuer (HB) reflex (36). In awake (37), sleeping (37), and anesthetized (22) animals, maintaining the lung inflated, even at very modest volumes, throughout expiration invariably causes a vagally mediated prolongation, often marked, of TEN. When inflations are maintained for only a fraction of TEN, and withdrawn before the onset of the next inspiration (as in the present study), there is still (in animals) a prolongation of TEN, evidencing residual after effect (memory) from the earlier inflation (14, 38). The main difficulty with advocating this mechanism is that, unlike sleeping animals (37), in sleeping humans the lungs must be maintained inflated at volumes in excess of half inspiratory capacity for any TEN prolongation to be observed (39, 40). Evidently, then, this expiratory prolonging HB reflex has a high volume threshold in humans. Because the HB reflex is mediated by pulmonary stretch receptors that respond to transpulmonary pressure (41), and in view of the increased lung stiffness in most of our patients, it is possible that the threshold was reached even though VT (0.49 ± 0.14 L) was considerably less than one-half of a normal inspiratory capacity. The strength of the timing response correlated positively with elastance and negatively with {Delta}R, an index of dynamic collapse of the airways during expiration, and by extension, of loss of elastic support of airways. These observations are in keeping with a contribution from airway stretch receptors to the response. However, the variability in response could only be explained to a small extent by these variables. Two explanations may be entertained: (1) There is considerable variability in the strength of this reflex even among healthy animals (42). Thus, innate variability may have contributed to the considerable variability in response among patients (Figure 5A). (2) In conscious normal humans, delaying lung emptying using an expiratory resistance is associated with considerable prolongation of TEN (15–17). Likewise, in mechanically ventilated awake normal subjects, introduction of a pause at the end of inflation markedly lengthens cycle duration (18). These responses are to be contrasted with the absence of timing responses to complete expiratory obstructions at moderate tidal volumes in sleeping normal humans (39, 40). The different responses in awake and sleeping humans cannot be explained by a sleep effect on HB expiratory prolonging reflex; this response remains very strong during sleep (37), and even under deep anesthesia (22), in animals. It is, therefore, likely that other mechanisms exist that cause expiratory lengthening when expiratory flow is retarded and that are operative only in the conscious state. These mechanisms, whose nature is unclear, may have been operative in our patients, many of whom were conscious.

On average, {Delta}TEN in the current study was considerably less than the prolongation observed in awake normal subjects. For example, in the study of Laghi and colleagues (18) in mechanically ventilated awake normal subjects, the addition of a 2-second pause lengthened cycle duration by 1.7 seconds, or 85% of pause duration. By contrast, in the current study, prolongation of cycle duration, as expressed in {Delta}TEN, was only 36% of occlusion (pause) duration. The weaker response in the patients may reflect their generally less alert state, when the mechanism involved is sensitive to level of consciousness (LOC). We did not formally document LOC. However, because patients were not selected on the basis of LOC, we expect that LOC covered the usual range encountered in a general ICU (alert to comatose). Whether this accounts for the variability in response remains to be determined.

We do not believe the TEN prolongation was a nonautomatic behavioral response to the occlusion. In a separate report on the same patients (32), we carefully looked for evidence of new respiratory muscle action in the course of these occlusions. Less than 1% of occlusions displayed motor acts that could possibly be considered nonautomatic (32). These rare occlusions were not considered in the present study. It would be extremely unlikely for TEN prolongation to represent a nonautomatic behavioral response after the occlusion when there were no nonautomatic motor responses during the occlusion.

Patients with severe COPD had a significantly depressed timing response. In a recent study on ambulatory stable patients with COPD connected to a ventilator via a mouthpiece, Laghi and colleagues found that delaying the opening of the exhalation valve by 2 seconds was associated with a 1.34-second (i.e., 67% of pause duration) increase in cycle time (20). Although somewhat less than the response observed in normal subjects under similar conditions (85%, [18]), it was still a substantial response in comparison with our patients (9% of occlusion time). The differences may reflect differences in LOC (the patients in Laghi and coworkers' study were fully alert). Alternatively, the patients in Laghi and coworkers' study (20) may not have been as severely affected as our four patients. Regardless of the reason, patients with severe COPD on mechanical ventilation in the ICU are very poorly equipped to deal with delays in onset of expiration.

The increase in timing response with occlusion volume in the same patients (Figure 4) is to be expected, given that these expiratory timing responses are well known to be volume dependent both in animals (14, 22, 37, 38) and humans (39, 40). It is noteworthy that even in the highest volume range, the response remained rather weak ({Delta}TEN/Tocc = 0.5), and highly variable (SD ± 0.28). In fact, there were four patients with VT greater than 0.8 L in this subset (0.83, 0.89, 0.92, and 1.04 L). Their timing ratios were 0.51, 0.38, 0.15, and 0.86, respectively. Thus, the main conclusion regarding the timing response, namely that it is variable but generally weak, is not related to a generally lower (than used in practice with other modes) VT with PAV (43).

In the volume-cycled and pressure support modes, volume usually increases in the interval between end of inspiratory effort and end of ventilator cycle. In the current study, volume was kept constant in this interval. It may be argued that the responses may be different if volume increases during the period of delay in ventilator cycling off. This possibility is difficult to address in practice, as well as being very unlikely given the totality of our results. The practical difficulty of addressing this issue stems from the fact that, in practice, there is a wide variety of patterns and rates of increase in volume during the period of delay.§ It would be impractical to design experiments to address all these patterns and also not to cause concurrent changes in ventilator output during inspiration, which would greatly confound interpretation. In the current study, we examined the responses when volume was kept constant at a low value (0.44 ± 0.17 L) during the delay and when volume was higher by, on average, 0.3 L at the same delay (Figure 4). These two conditions essentially cover the extremes of possible changes in volume, encountered in practice, during the delay. That the same range of responses was seen in the high and low volume ranges, as well as the fact that patients who had weak responses at low volumes continued to have weak responses at the high volume (and vice versa), strongly suggest that our current conclusions apply equally to situations in which volume increases during the delay. The timing response may be expected to be somewhat higher if volume increased during the delay, simply because volume at the end of the delay would be higher than otherwise. However, there is little reason to believe that the act of volume increasing specifically during the delay will convert a weak responder to a strong responder, or vice versa.

DH and Trigger Delay
The changes in end-expiratory lung volume ({Delta}VDH) and elastic recoil pressure ({Delta}PDH) in response to delaying expiration were highly variable (Figure 8). Delays of as little as 0.4 seconds could result in an increase in end-expiratory volume (EEV) greater than 0.2 L and in end-expiratory pressure (i.e., PEEPi) greater than 4.0 cm H2O. Conversely, the changes in these variables with much longer delays (e.g., > 1.0 seconds) could be considerably less (Figure 8). There are two major factors that account for this variable response to a given Tocc. First, the actual reduction in the duration of expiratory flow (TEM), at a given Tocc, was quite variable depending on the strength of the TEN response (Figure 5). Second, for a given reduction in TEM, the extent of increase in EEV is very dependent on the magnitude of expiratory flow rate in the terminal part of expiration (V·expterminal), and this was highly variable among patients (-0.23 ± 0.14 L/second). Thus, if expiratory flow in the interval before the onset of TI, and corresponding to {Delta}TEM, is close to zero (e.g., Figure 6B), the reduction in TEM will have little consequence on EEV and, hence, PEEPi. Conversely, if V·expterminal is high, even a small reduction in TEM can result in a large increase in EEV. The importance of these two variables (i.e., strength of TEN response [{Delta}TEN/Tocc] and V·expterminal) is emphasized by the fact that, for a given Tocc, {Delta}VDH can be precisely predicted from the equation: pred. {Delta}VDH = Tocc(1 - {Delta}TEN/Tocc) · V·expterminal. This prediction equation had a correlation coefficient (with actual {Delta}VDH) of 0.96 with a slope of 0.98 and intercept of 0. This agreement further emphasizes the lack of any significant other responses to defend against DH.

As may be expected (as a consequence of the increase in PEEPi), trigger delay was lengthened in nearly all patients (Figure 9). The magnitude of this response was quite variable, reflecting the wide range of {Delta}PDH (Figure 8B) and the expected wide range in {Delta}Pmus /{Delta}t in the early phase of inspiratory effort.

On average, 6% of postocclusion inspiratory efforts failed to trigger the ventilator. Leung and associates reported that approximately 28% of efforts are ineffective in patients placed in the PSV or volume-cycled modes at a VT of 10 ml/kg (9). This much higher incidence during routine ventilator care suggests that the degree of nonsynchrony between the end of the ventilator's and the patient's inspiratory cycles under these conditions is considerably more pronounced than was deliberately induced here (average Tocc = 0.78 ± 0.34 seconds). In other words, we believe that the delays imposed in the present study are likely to be well within the range of delays that occur during routine ventilator care.

Clinical Implications
The current study shows that, all else being the same, delays in onset of mechanical expiration relative to neural expiration will tend to: (1) Increase DH with, consequently, an increase in trigger delay and in frequency of ineffective efforts. This is primarily due to lack of a reflex (i.e., immediate) neuromuscular response to effectively expedite expiratory flow in the abbreviated TEM, and secondarily, to incomplete timing response in most patients. (2) Decrease ventilation. This is in part due to slowing of the patient's RR and in part to promotion of ineffective efforts.

In practice, these primary effects of delayed cycling-off may be enhanced or mitigated by other ventilator settings that affect overall level of ventilation and respiratory timing via chemical and other neural mechanisms (see reference 44 for review). Nonetheless, this study shows that delayed cycling off, per se, has adverse consequences that, if not offset by other ventilator settings, will inevitably lead to an increase in chemical drive and respiratory efforts.

In the current study, the delay in onset of expiration was applied for only one breath. The results, therefore, need not reflect what might happen in the steady state. The latter, it may be argued, is what is clinically relevant. Although steady state responses need to be documented, we believe that identifying the immediate responses is valuable in that these responses point out the primary changes that will challenge the respiratory control system. In fact, interpretation of steady state changes may be difficult without knowledge of the immediate responses. For example, an immediate response that causes a reduction in V·E will, in due time, evoke secondary responses (e.g., via chemical feedback) that may restore V·E toward the pre-intervention value. Without knowledge of the immediate responses, an unchanged steady state V·E may lead to the false impression that the intervention is inconsequential. With knowledge of the immediate responses, it will be realized that the maintenance of V·E was at the expense of increased respiratory drive and efforts.

Patients vary considerably in their susceptibility to DH in response to cycling delays. This susceptibility can, to a considerable extent, be assessed at the bedside by noting the magnitude of expiratory flow just before the onset of inspiratory effort (defined here as the point at which flow and Paw deviate, in an inspiratory direction, from the trajectory established earlier in expiration, Figure 1). The higher this terminal flow value, the more susceptible the patient is, and the more careful one needs to be to avoid cycling delays. Given the relatively weak timing responses observed here ({Delta}TE / Tocc = 0.37 ± 0.2) and the resulting magnitude of DH (Equation 1), the increase in DH as a result of delays is, on average, two-thirds (i.e., 1.00-0.37) of the value of terminal flow per second of delay. As a corollary, a patient with a terminal expiratory flow of, for example, 0.4 L/second is likely to benefit substantially (in terms of reduced DH and, by extension, reduced inspiratory effort) from measures that would reduce or eliminate the cycling delay, whereas in a patient with a near zero terminal flow such measures would have little consequence in terms of DH.

Patients with severe COPD are notably lacking in timing responses. In them, mechanical TE decreases on a nearly 1:1 basis with increases in delay. This, and their need for a long exhalation time, makes them particularly vulnerable to developing more DH in the presence of delays.


   Acknowledgments
 
The authors wish to thank Warren Shewchuk for technical assistance and Kay Foster for typing the manuscript.

Supported by the Canadian Institutes of Health Research.


   FOOTNOTES
 
This article has an online data supplement, which is accessible from this issue's table of contents online at www.atsjournals.org

* The subtraction method for estimating {Delta}R is subject to errors due to: (1) Nonlinearities in the P–V relation within the tidal volume range. These can cause small bidirectional errors in estimated alveolar pressure during expiration. (2) Differences between the actual K2 value (of Rohrer's equation) of the ET tube and the assumed value of a clean tube used here. This can again cause small bidirectional errors if the flows present during the inspiratory and expiratory measurements are different (see online data supplement for more detail). Back

{dagger} The slope of expiratory flow in the last 0.5 seconds of TEN was estimated from: .Vexpslope = 2(.V0 - .V-0.5), where .V-0.5 is flow 0.5 seconds before T0. Back

{ddagger} During PAV, flow reversal (onset of TEM) occurs, as it does normally (31), during the declining phase of inspiratory Pmus (24). There is, therefore, some residual inspiratory Pmus at the beginning of TEM, which causes some braking of .Vexp. This inspiratory Pmus dissipates over a fraction of a second (32). To the extent that this early braking influence disappears later, the effect would be to cause .Vexp to be higher after the occlusion, even in the absence of occlusion-related expiratory muscle recruitment. The presence of phasic expiratory Pmus, unrelated to the occclusion, would also favor a higher .Vexp in the postocclusion breaths, relative to control breaths, because phasic expiratory activity, if present, is higher later in TEN (33). Back

§ In some ventilators, the last part of the inflation cycle may have no volume increase, or volume may increase initially during the delay but remain constant later on (e.g., Evita ventilator with the autoflow function switched off). Alternatively, with square-flow patterns, volume will increase at the same rate that applied throughout inflation, and this rate varies considerably from patient to patient. In the decelerating flow pattern of volume-cycled ventilation, volume increases at a progressively decreasing rate and this deceleration pattern varies among ventilators. Finally, in the PSV mode, volume also increases at a decelerating rate with the deceleration pattern being extremely variable depending on respiratory mechanics and pattern of inspiratory muscle pressure beyond the end of inspiratory effort. Back

Received in original form July 26, 2001; accepted in final form March 22, 2002


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 METHODS
 RESULTS
 DISCUSSION
 REFERENCES
 

  1. Younes M. Proportional assist ventilation and pressure support ventilation: similarities and differences. Update Intensive Care Emerg Med 1991;15:361–380.
  2. Aslanian P, Brochard LJ. Partial ventilatory support. In: Marini JJ, Slutsky AS, editors. Physiological basis of ventilatory support. Lung biology in health and disease, Vol. 118. New York: Marcel Dekker; 1998. p. 817–846.
  3. Jubran A, Van de Graaff W, Tobin M. Variability of patient-ventilator interaction with pressure support ventilation in patients with chronic, obstructive pulmonary disease. Am J Respir Crit Care Med 1995;152:129–136.[Abstract]
  4. Nava S, Bruschi C, Rubini F, Palo A, Iotti G, Braschi A. Respiratory response and inspiratory effort during pressure support ventilation in COPD patients. Intensive Care Med 1995;21:871–879.[CrossRef][Medline]
  5. Nava S, Bruschi C, Fracchia C, Braschi A, Rubini F. Patient-ventilator interaction and inspiratory effort during pressure support ventilation in patients with different pathologies. Eur Respir J 1997;10:177–183.[Abstract]
  6. Cinnella G, Conti G, Lofaso F, Lorino H, Harf A, Lemaire F, Brochard L. Effects of assisted ventilation on the work of breathing: volume-controlled versus pressure-controlled ventilation. Am J Respir Crit Care Med 1996;153:1025–1033.[Abstract]
  7. Marini JJ. Dynamic hyperinflation. In: Marini JJ, Slutsky AS, editors. Physiological basis of ventilatory support. Lung biology in health and disease, Vol. 118. New York: Marcel Dekker; 1998. p. 453–490.
  8. Younes M. Patient-ventilator interaction with pressure-assisted modalities of ventilatory support. Semin Respir Med 1993;14:299–322.
  9. Leung P, Jubran A, Tobin MJ. Comparison of assisted ventilator modes on triggering, patients' effort, and dyspnea. Am J Respir Crit Care Med 1997;155:1940–1948.[Abstract]
  10. Giannouli E, Webster K, Roberts D, Younes M. Response of ventilator-dependent patients to different levels of pressure support and proportional assist. Am J Respir Crit Care Med 1999;159:1716–1725.[Abstract/Free Full Text]
  11. Tobin MJ, Jubran A, Laghi F. Patient-ventilator interaction. Am J Respir Crit Care Med 2001;163:1059–1063.[Free Full Text]
  12. Remmers JE, Bartlett D Jr. Reflex control of expiratory airflow and duration. J Appl Physiol 1977;42:80–87.[Abstract/Free Full Text]
  13. Zechman FW, Frazier DT, Lally DA. Respiratory volume-time relationships during resistive loading in the cat. J Appl Physiol 1976;40:177–183.[Abstract/Free Full Text]
  14. Knox CK. Characteristics of inflation and deflation reflexes during expiration in the cat. J Neurophysiol 1973;36:284–295.[Free Full Text]
  15. Gothe B, Cherniack NS. Effects of expiratory loading on respiration in humans. J Appl Physiol 1980;49:601–608.[Abstract/Free Full Text]
  16. Poon CS, Younes M, Gallagher C. Post-inspiratory muscle pressure, respiratory patterns and gas exchange: effects of expiratory resistive loading. J Appl Physiol 1987;63:1837–1845.[Abstract/Free Full Text]
  17. Hill AR, Kaiser DL, Lu JY, Rochester DF. Steady-state response of conscious man to small expiratory resistive loads. Respir Physiol 1985; 61:369–381.[CrossRef][Medline]
  18. Laghi F, Karamchandani K, Tobin MJ. Influence of ventilator settings in determining respiratory frequency during mechanical ventilation. Am J Respir Crit Care Med 1999;160:1766–1770.[Abstract/Free Full Text]
  19. Tolbert DG, Simon PM, Stroetz RW, Hubmayr RD. The determinants of respiratory rate during mechanical ventilation. Am J Respir Crit Care Med 1997;155:485–492.[Abstract]
  20. Laghi F, Segal J, Choe WK, Tobin MJ. Effect of imposed inflation time on respiratory frequency and hyperinflation in patients with chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2001;163:1365–1370.[Abstract/Free Full Text]
  21. Bishop B. Diaphragm and abdominal muscle responses to elevated airway pressures in the cat. J Appl Physiol 1967;22:959–965.[Free Full Text]
  22. Younes M, Vaillancourt P, Milic-Emili J. Interaction between chemical factors and duration of apnea following lung inflation. J Appl Physiol 1974;36:190–201.[Free Full Text]
  23. 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.[CrossRef][Medline]
  24. Younes M. Proportional assist ventilation. A new approach to ventilatory support: theory. Am Rev Respir Dis 1992;145:114–120.[Medline]
  25. Younes M, Webster K, Kun J, Roberts D. Gain of the expiratory prolonging Hering-Breuer (H-B) reflex in mechanically ventilated patients. Am J Respir Crit Care Med 1997;155:A685.
  26. Younes M, Puddy A, Roberts D, Light RB, Quesada A, Taylor K, Oppenheimer L, Cramp H. Proportional assist ventilation: results of an initial clinical trial. Am Rev Respir Dis 1992;145:121–129.[Medline]
  27. Rossi A, Polese G, Milic-Emili J. Monitoring respiratory mechanics in ventilator-dependent patients. In: Tobin MJ, editor. Principles and practice of intensive care monitoring. New York: McGraw-Hill; 1998. p. 553–597.
  28. Knudson RJ, Mead J, Knudson DE. Contribution of airway collapse to supramaximal expiratory flows. J Appl Physiol 1974;36:653–657.[Free Full Text]
  29. Gottfried SB, Rossi A, Higgs BD, Calverley PMA, Zochi L, Bozic C, Milic-Emili J. Noninvasive determination of respiratory system mechanics during mechanical ventilation for acute respiratory failure. Am Rev Respir Dis 1985;131:414–420.[Medline]
  30. Koutsoukou A, Armaganidis A, Stavrakaki-Kallergi C, Vassilakopoulos T, Lymberis A, Roussos C, Milic-Emili J. Expiratory flow limitation and intrinsic positive end-expiratory pressure at zero positive end-expiratory pressure in patients with adult respiratory distress syndrome. Am J Respir Crit Care Med 2000;161:1590–1596.[Abstract/Free Full Text]
  31. Mead J, Agostoni E. Dynamics of breathing. In: Handbook of physiology: respiration, Sect. 3, Vol. 1, chap. 14. Washington, DC: American Physiological Society; 1964. p. 411–427.
  32. Younes M, Webster K, Kun J, Roberts D, Masiowski B. A method for measuring passive elastance during proportional assist ventilation. Am J Respir Crit Care Med 2001;164:50–60.[Abstract/Free Full Text]
  33. von Euler C. Brain stem mechanisms for generation and control of breathing pattern. In: Cherniack NS, Widdicombe JG, editors. Handbook of physiology: the respiratory system, Vol. 2. Bethesda, MD: American Physiological Society; 1986. p. 1–68.
  34. Puddy A, Patrick W, Webster K, Younes M. Respiratory control during volume-cycled ventilation in normal humans. J Appl Physiol 1996;80: 1749–1758.[Abstract/Free Full Text]
  35. Xirouhaki N, Kondili E, Mitrouska I, Siafakas N, Georgopoulos D. Response of respiratory motor output to varying pressure in mechanically ventilated patients. Eur Respir J 1999;14:508–516.[Abstract]
  36. Breuer J. Die Selbststeuerung der Atmung durch den Nervus vagus. Sitzungber Akad Wiss Wien Abt 1868;258:909–937.
  37. Phillipson EA, Murphy E, Kozar LF. Regulation of respiration in sleeping dog. J Appl Physiol 1976;40:688–693.[Abstract/Free Full Text]
  38. Satoh M, Eastwood PR, Smith CA, Dempsey JA. Nonchemical elimination of inspiratory motor output via mechanical ventilation in sleep. Am J Respir Crit Care Med 2001;163:1356–1364.[Abstract/Free Full Text]
  39. Iber C, Simon PM, Skatrud JB, Mahowald MW, Dempsey JA. The Breuer-Hering reflex in humans: effects of pulmonary denervation and hypocapnia. Am J Respir Crit Care Med 1995;152:217–224.[Abstract]
  40. Hamilton RD, Winning AJ, Horner RL, Guz A. The effect of lung inflation on breathing in man during wakefulness and sleep. Respir Physiol 1988;73:145–154.[CrossRef][Medline]
  41. Sant'Ambrogio G. Information arising from the tracheobronchial tree of mammals. Physiol Rev 1982;62:531–569.[Free Full Text]
  42. Gautier H, Bonora M, Gaudy JH. Breuer-Hering inflation reflex and breathing pattern in anesthetized humans and cats. J Appl Physiol 1981;51:1162–1168.[Abstract/Free Full Text]
  43. Younes M. Proportional assist ventilation. In: Tobin M, editor. Principles and practice of mechanical ventilation, chap. 15. New York: McGraw- Hill; 1994. p. 349–370.
  44. Younes M, Georgopoulos D. Control of breathing relevant to mechanical ventilation. In: Marini J, Slutsky A, editors. Physiological basis of ventilatory support. Lung biology in health and disease, Vol. 118, chap. 1. New York: Marcel Dekker; 1998. p. 1–74.



This article has been cited by other articles:


Home page