Documente Academic
Documente Profesional
Documente Cultură
Intensive Care Medicine, University Hospital of Heraklion, Medical School, University of Crete, Heraklion, Crete, Greece
Pulmonary Department, University Hospital of Heraklion, Medical School, University of Crete, Heraklion, Crete, Greece
Sleep Disorders Centre, Misericordia Health Centre, Winnipeg, Manitoba, Canada
a r t i c l e
i n f o
Article history:
Received 17 January 2016
Received in revised form 3 March 2016
Accepted 15 March 2016
Available online 17 March 2016
Keywords:
Proportional assist ventilation
Reex feedback
Chemical feedback
Respiratory system compliance
Tidal volume
a b s t r a c t
Tidal volume (VT ) is the controlled variable during passive mechanical ventilation (CMV) in order to
avoid ventilator-induced-lung-injury. However, recent data indicate that the driving pressure [P; VT
to respiratory system compliance (Crs) ratio] is the parameter that best straties the risk of death. In
order to study which variable (VT or P) is controlled by critically ill patients, 108 previously studied
patients were assigned to receive PAV+ (a mode that estimates Crs and permits the patients to select
their own breathing pattern) after CMV, were re-analyzed. When patients were switched from CMV to
PAV+ they controlled P without constraining VT to narrow limits. VT was increased when the resumption of spontaneous breathing was associated with an increase in Crs. When P was high during CMV,
the patients (n = 12) decreased it in 58 out of 67 measurements. We conclude that critically ill patients
control the driving pressure by sizing the tidal volume to individual respiratory system compliance using
appropriate feedback mechanisms aimed at limiting the degree of lung stress.
2016 Elsevier B.V. All rights reserved.
1. Introduction
T
idal volume (VT ) has been recognized as the variable that
should be controlled during mechanical ventilation in order to
avoid ventilator-induced lung injury (VILI) (Amato et al., 1998; The
ARDSnet, 2000; Malhotra, 2007; Mascia et al., 2010; Serpa Neto
et al., 2012; Futier et al., 2013). However, in a recent study on 3562
patients with acute respiratory distress syndrome (ARDS) enrolled
in previously reported randomized trials and ventilated passively,
Amato et al. (2015) showed that reductions in VT or increases in
external positive end-expiratory pressure (PEEP) increased survival only if associated with decreases in driving pressure [P;
static end-inspiratory plateau pressure (Pplat ) minus PEEP, or VT
to respiratory system compliance (Crs) ratio]. Increases in P were
strongly associated with decreased survival, particularly at P values above 15 cmH2 O.
P essentially reects the extent of lung stretch during tidal
breathing. The respiratory system is endowed with mechanoreceptors that sense the degree of lung stretch. These tend to naturally
Corresponding author at: Intensive Care Medicine, University Hospital of Heraklion, Medical School University of Crete, Heraklion, Crete, Greece.
E-mail address: georgop@med.uoc.gr (D. Georgopoulos).
http://dx.doi.org/10.1016/j.resp.2016.03.009
1569-9048/ 2016 Elsevier B.V. All rights reserved.
70
We previously reported (Xirouchaki et al., 2008) on breathing pattern and respiratory mechanics in 108 critically ill patients
ventilated with PAV+ for 48 h. The majority of patients (59.3%)
met ARDS denition (Force et al., 2012) at study entry. Pplat averaged 17 cmH2 O while PEEP averaged 7 cmH2 O, giving an average
P of 10 cmH2 O, well below the values found by Amato et al.
(2015) to adversely affect survival. Although these results indicate
that, on average, the protective reexes are effective, we did not
examine the range of P in individual patients. This is of importance since the potency of these reexes is highly variable among
mechanically-ventilated patients (Kondili et al., 2001; Younes et al.,
2002). Accordingly, in the current study, we determined individual P in these patients and examined how it related to P when
the same patients were ventilated with CMV using the currently
accepted lung-protective strategy (The ARDSnet, 2000; Malhotra,
2007). We hypothesized that the patients control of breathing
system is adept at protecting the lungs by preventing high P
using appropriate feedback mechanisms, while not unnecessarily
restricting tidal volume when this has no protective value.
2. Methods
One hundred and eight patients who were randomized in our
previous study (Xirouchaki et al., 2008) to receive PAV+ [PuritanBennett 840 ventilator (Covidien, Boulder, CO)] were re-analyzed.
Before switching to PAV+, the patients were on mechanical ventilation for at least 36 h and ventilated passively with a CMV mode
(volume or pressure control) [see Ref. Georgopoulos et al., 2016].
Criteria for switching to PAV+ were (Xirouchaki et al., 2008): ability to trigger the ventilator at >10 breaths/min; PaO2 > 60 mmHg,
with fractional concentration of inspired O2 (FIO2 ) <65%; total [PEEP
and intrinsic PEEP (PEEPi)] positive end-expiratory airway pressure <15 cmH2 O; pH > 7.30; no severe hemodynamic instability;
no severe bronchospasm; and a stable neurological status. Initially,
PEEP and FI O2 were set to values similar to those during CMV immediately before the randomization. The PAV+ assist was set relatively
high [70% (7070), median (interquartile range)]. Subsequently,
specic pre-dened written algorithms were used to adjust the
ventilator settings (Xirouchaki et al., 2008). PAV+ was continued for
48 h unless the patient met pre-dened criteria either for switching
to CMV, or for breathing without ventilator assistance (Xirouchaki
et al., 2008).
During CMV respiratory system mechanics were assessed
within 8 h before switching to PAV+ (passive mechanical ventilation), using the 3-s occlusion technique on volume control mode
with square-wave inspiratory ow-time prole (Bates et al., 1985;
Gottfried et al., 1985). P1 (the pressure level at which airway
pressure drops immediately after end-inspiratory occlusion), Pplat
(PplatCMV ) and PEEPi were measured. During PAV+ VT , Pplat , PEEP
and PEEPi were measured approximately 510 min after switching
to PAV+ and at 1, 4, 8, 12, 24, 28, 32, 36 and 48 h later (if the patient
continued to be on PAV+). Pplat during PAV+ (PplatPAV+ ) was measured using a modication of end-inspiratory occlusion technique
(Younes et al., 2001), and PEEPi by calculating the time course of
elastic recoil pressure during expiration, assuming that expiration
is passive [see Ref. Georgopoulos et al., 2016].
During CMV, the primary physician, who was not involved in the
study, was responsible for ventilator settings based on the principles of lung protective strategy (VT 47 ml/Kg, Pplat < 30 cmH2 O)
(The ARDSnet, 2000). To follow the methods used in patients
of the Amato et al. (2015) study, respiratory system compliance
(CrsCMV ) was measured as VT /(PplatCMV PEEP) and driving pressure (PCMV ) was calculated as VT /CrsCMV ratio. Crs and P during
PAV+ (CrsPAV+ , PPAV+ ) were similarly calculated from average VT
(VTPAV+ ), PEEP and average PplatPAV+ in the occluded breaths. Arte-
Table 1
P1 and PplatCMV during CMV and PplatPAV+ measured during the rst 8 h on PAV+.
P1 (cmH2 O)
PplatCMV (cmH2 O)
PplatPAV+ (cmH2 O)
0h
1h
4h
8h
20.5 (18.023.0)
17.8 (15.320.1)
17.2 (14.720.5)
16.5 (14.720.3)
16.4 (14.519.1)
16.5 (14.520.0)
Values are medians (interquartile range). P1 ; the pressure level at which airway pressure drops immediately after end-inspiratory occlusion during controlled
mechanical ventilation (CMV). PplatCMV ; airway pressure measured at the end of
the 3 s end-inspiratory occlusion maneuver during CMV. PplatPAV+ ; airway pressure
measured 0.3 s after the end-inspiratory pause maneuver at the end of selected
inspirations at 0, 1, 4 and 8 h on PAV+. For clarity of presentation, PplatPAV+ values
after 8 h on PAV+ are not shown.
rial blood gasses were measured during CMV and after switching
to PAV+ at 1, 4, 8, 12, 24, 28, 32, 36 and 48 h later.
Data are given as median (2575th interquartile range), unless
stated otherwise. Proportions were compared using the Fisher
exact test. Continuous variables were compared with Wilcoxon and
Man-Whitney tests, as appropriate. Regression analysis was performed using the least square method. Linear mixed effect models
on parameters of repeated measurements were used to investigate
changes in various variables over time during PAV+. The values
of the rst four serial measurements, corresponding to an 8-h
PAV+ period, were included in the model in order to compare with
the corresponding variables obtained within the 8-h CMV period.
P < 0.05 was considered as signicant.
3. Results
Since the results were similar when patients with (n = 64) and
without ARDS (n = 44) were analyzed separately [see Fig. 1, 2, 4, 5,
79 in Ref. Georgopoulos et al., 2016] we present data pertaining
to all patients (n = 108). Also, the results were not modied if PEEPi
was taken into consideration in calculating Crs (and thus P), due
to very low PEEPi in these patients (median PEEPi was <0.5 cmH2 O
both during CMV and PAV + ) (Xirouchaki et al., 2008).
Median and interquartile range of VTPAV+ , CrsPAV+ and PPAV+
are shown in Fig. 1 and P1 , PplatCMV and PplatPAV+ are reported in
Table 1. VTPAV+ , CrsPAV+ and PPAV+ during the initial 8-h PAV+
period did not differ as a function of time (linear effect model,
p > 0.05). Therefore, the values of each variable during this 8-h
period were averaged (VTPAV+aver , CrsPAV+aver and PPAV+aver ) and
compared to the corresponding values (VTCMV , CrsCMV , PCMV )
obtained during CMV within 8 h before switching to PAV+.
PPAV+aver did not differ from PCMV [10.2 cmH2 O (8.112.4)
vs. 10.7 cmH2 O (9.012.9), respectively], while VTPAV+aver and
CrsPAV+aver were signicantly higher than VTCMV and CrsCMV ,
respectively (Fig. 1). During PAV+ a total of 744 measurements
of VTPAV+ , CrsPAV+ and PPAV+ were performed [median 8 (410)
measurements per patient]. During CMV, VT was relatively tightly
controlled as dictated by the principle of lung protective strategy,
and as a result PCMV was less than 15 cmH2 O in the majority of
the patients (Fig. 2). When the patients were switched to PAV+,
PPAV+ (but not VT ) was tightly controlled by the patients (Fig. 2).
Compared to CMV, in 83 out of 108 patients (76.9%) VTPAV+aver
was higher, and in 70 of them (84.3%) this increase was associated with an increase in CrsPAV+aver (Fig. 3), [see also Fig. 3 in Ref.
Georgopoulos et al., 2016]. When CrsPAV+aver was lower than CrsCMV
(n = 10), the increase in VT was minimal [0.74 (0.620.79) ml/kg].
Fig. 4 shows individual differences between PPAV+ and PCMV
as a function of PCMV . Data are given for PPAV+ measured at
510 min (Fig. 4A) and at 1 h (Fig. 4B) after switching to PAV+.
At both times, a signicant negative relationship was observed
71
Fig. 1. Boxplot with whiskers from minimum to maximum of tidal volume (VT ), respiratory system compliance (Crs) and driving pressure (P). Left: values at various times
during the 48-h PAV+ period. The numbers in the top of the gure indicate median assist level (%) throughout the PAV+ period. Right: values of each variable during controlled
mechanical ventilation (CMV, grey box) obtained within 8 h before switching to PAV+ and average values (PAV + aver , open box) during the rst 8-h PAV+ period (time 0, 1, 4
and 8 h). *Signicantly different from CMV (p < 0.0001).
4. Discussion
The main ndings of our study are: (1) When critically ill
patients are switched from passive mechanical ventilation to an
assisted mode that permits them to choose their own breathing
pattern they control the driving pressure as well as it is controlled
with CMV using the lung protective strategy without constraining VT to narrow limits. (2) When the lung protective strategy
results in high driving pressure (15 cmH2 O), the patients control
of breathing system decreased the driving pressure in the majority of measurements. (3) Respiratory system compliance increased
on PAV+, which made it possible for VT to increase without compromising driving pressure. Overall, the patients increased the VT
when the resumption of spontaneous breathing was associated
with an increase in respiratory system compliance. These results
indicate that by using appropriate feedback systems, patients are
able to keep the driving pressure at a relatively safe range by sizing
the inspired volume to individual respiratory system compliance.
72
Fig. 2. (A) Individual data of tidal volume (VT )driving pressure (P) relationship during PAV+ (open circles, 744 measurements) and controlled mechanical ventilation
(CMV, closed, red circles, 108 measurements). Continuous vertical and dotted and dashed horizontal lines show the thresholds of 15 cmH2 O of P, and 7 and 8 ml/kg of VT ,
respectively. (B) Percentage of measurements in which P was <15 cmH2O (open bars) and VT was <7 ml/kg (black bars) and <8 ml/kg (grey bars) during CMV and PAV+.
*Signicantly different from the corresponding value during CMV (p < 0.0001).
ance. At this time the lung has undergone stress relaxation (Bates
et al., 1985) and Pplat at 3 s underestimates elastic recoil at endinspiration of the un-occluded breaths. Accordingly, if all else is
the same, driving pressure during PAV+ should be higher than in
CMV by the change in Pplat between 0.3 s and 3.0 s after occlusion.
In our study, we found no difference in P between PAV+ and CMV
even though during CMV, early Pplat (i.e. close to P1 ) was higher
than 3-s Pplat by approximately 3 cmH2 O (Table 1). Thus, in reality,
end-inspiratory elastic recoil during PAV+ was not just comparable to that during CMV, but it was, on average, 3 cmH2 O lower
[P1 PplatCMV = 3 cmH2 O (24)]. Considering that driving pressure
includes the pressure dissipated against the elastic recoil of the
chest-wall and lung, dynamic lung stretch during PAV+ is very low
indeed.
In the current study Pplat measured at 510 min after switching to PAV+ and at high assist (70%) was 17 (1519) cmH2 O and
did not exceed 30 cmH2 O except in one patient, who immediately
failed on PAV+ (Georgopoulos et al., 2016). The lungs are rich with
stretch receptors that can limit the degree of lung stretch at endinspiration through the operation of the Hering-Breuer inspiratory
inhibitory reex (Clark and von Euler, 1972; Grunstein et al., 1973).
In animals, the volume threshold for activating the inspiratory inhibition is very low at FRC (Clark and von Euler, 1972; Grunstein
et al., 1973). In awake healthy humans VT -related reduction in
inspiratory time begins on average when VT exceeds 1.01.5 l,
corresponding to an end-inspiratory transpulmonary pressure of
73
Fig. 3. Relationships between VT (difference in VT between PAV+ and CMV, VTPAV+aver VTCMV ) and Crs (difference in Crs between PAV+ and CMV, CrsPPAV+aver CrsCMV ).
The change in VT and Crs were obtained by averaging these variables during the rst 8-h PAV+ period. Continuous line; regression line. Crs; respiratory system compliance.
VT ; tidal volume. CMV; controlled mechanical ventilation. During CMV, the measurements of VT and Crs were obtained within 8 h before switching to PAV+ when criteria
for passive mechanical ventilation were met.
strategy. The current study shows that when given the opportunity,
most patients increase VT over and above the recommended levels,
as long as this does not result in excessive lung stress.
We observed a signicant negative relationship between PaCO2
during CMV and the change in PaCO2 at 1 h on PAV+ (PaCO2 at
1 h on PAV+ minus PaCO2 on CMV). Overall, these results indicate
that when patients are switched to PAV+, PaCO2 values during CMV
largely determine the magnitude and the direction of the change
in PaCO2 as the proper operation of chemical feedback dictates. On
PAV+ in 86 out of 636 measurements, PaCO2 was <35 mmHg but
this was associated with P 15 cmH2 O only in 8 measurements.
It follows that in critically ill patients meeting certain criteria, permitting the chemical feedback to be expressed (since PAV+ does
not interfere with the operation of chemical feedback) does not
increase the driving pressure; even in patients with high drive, the
driving pressures was maintained less than 15 cmH2 O.
Our study has some limitations: rstly, it is a post hoc observational analysis. Secondly, the ventilator settings during CMV were
titrated following the everyday clinical practice and did not follow a specic algorithm as they did during PAV+. However, during
CMV, the primary physicians followed the rules of lung protective strategy and VT , as well as end-inspiratory plateau pressure,
were maintained at recommended values. Thirdly, patients meeting certain criteria were studied and thus these results may not be
applicable to all critically ill patients. For example, high respiratory
drive due to extrapulmonary causes (i.e. severe metabolic acidosis) may override the protective mechanisms resulting in excessive
driving pressure and thus lung stress. Fourthly, due to the observational nature of the study, the effects of chemical and reex
feedback were not appropriately studied.
In conclusion, we demonstrated that during proportional assist
ventilation, critically ill patients are able to maintain the driving
pressure at an acceptable level by sizing the inspired volume to res-
74
Fig. 4. Individual changes in PPAV+ measured at 510 min (Fig. 4A) and 1 h (Fig. 4B) after switching to PAV+ from PCMV (PPAV+ PCMV ) as a function of PCMV . Curved
open arrow: Patient no. 162 of Ref. Xirouchaki et al., 2008 who failed immediately on PAV+ and was placed back on controlled mechanical ventilation (CMV). Black arrows:
Patient no. 130 of Ref. Xirouchaki et al., 2008 who presented with acute exacerbation of lung brosis and failed approximately 12 h after switching to PAV+. This patient
maintained tidal volume between 6 and 7 ml/kg throughout PAV+. Continuous lines; regression lines.
Fig. 5. Relationship between PaCO2 during CMV and the change in PaCO2 at 1 h on PAV+ (PaCO2 = PaCO2 at 1 h on PAV+ PaCO2 on CMV). Continuous line; regression line.
75
Jammes, Y., Auran, Y., Gouvernet, J., Delpierre, S., Grimaud, C., 1979. The
ventilatory pattern of conscious man according to age and morphology. Bull.
Eur. Physiopathol. Respir. 15, 527540.
Kao, K.C., Hu, H.C., Chang, C.H., Hung, C.Y., Chiu, L.C., Li, S.H., Lin, S.W., Chuang, L.P.,
Wang, C.W., Li, L.F., Chen, N.H., Yang, C.T., Huang, C.C., Tsai, Y.H., 2015. Diffuse
alveolar damage associated mortality in selected acute respiratory distress
syndrome patients with open lung biopsy. Crit. Care 19, 228.
Kondili, E., Prinianakis, G., Anastasaki, M., Georgopoulos, D., 2001. Acute effects of
ventilator settings on respiratory motor output in patients with acute lung
injury. Intensive Care Med. 27, 11471157.
Kondili, E., Prinianakis, G., Athanasakis, H., Georgopoulos, D., 2002. Lung emptying
in patients with acute respiratory distress syndrome: effects of positive
end-expiratory pressure. Eur. Respir. J. 19, 811819.
Lorente, J.A., Cardinal-Fernandez, P., Munoz, D., Frutos-Vivar, F., Thille, A.W.,
Jaramillo, C., Ballen-Barragan, A., Rodriguez, J.M., Penuelas, O., Ortiz, G., Blanco,
J., Pinheiro, B.V., Nin, N., Del Carmen Marin, M., Esteban, A., Thompson, T.B.,
2015. Acute respiratory distress syndrome in patients with and without diffuse
alveolar damage: an autopsy study. Intensive Care Med. 41, 19211930.
Malhotra, A., 2007. Low-tidal-volume ventilation in the acute respiratory distress
syndrome. N. Engl. J. Med. 357, 11131120.
Marantz, S., Patrick, W., Webster, K., Roberts, D., Oppenheimer, L., Younes, M.,
1996. Response of ventilator-dependent patients to different levels of
proportional assist. J. Appl. Physiol. (1985) 80, 397403.
Mascia, L., Pasero, D., Slutsky, A.S., Arguis, M.J., Berardino, M., Grasso, S., Munari,
M., Boifava, S., Cornara, G., Della Corte, F., Vivaldi, N., Malacarne, P., Del Gaudio,
P., Livigni, S., Zavala, E., Filippini, C., Martin, E.L., Donadio, P.P., Mastromauro, I.,
Ranieri, V.M., 2010. Effect of a lung protective strategy for organ donors on
eligibility and availability of lungs for transplantation: a randomized
controlled trial. JAMA 304, 26202627.
Mirabella, L., Grasselli, G., Haitsma, J.J., Zhang, H., Slutsky, A.S., Sinderby, C., Beck, J.,
2014. Lung protection during non-invasive synchronized assist versus volume
control in rabbits. Crit. Care 18, R22.
Polacheck, J., Strong, R., Arens, J., Davies, C., Metcalf, I., Younes, M., 1980. Phasic
vagal inuence on inspiratory motor output in anesthetized human subjects. J.
Appl. Physiol.: Respir. Environ. Exerc. Physiol. 49, 609619.
Puddy, A., Patrick, W., Webster, K., Younes, M., 1996. Respiratory control during
volume-cycled ventilation in normal humans. J. Appl. Physiol. (1985) 80,
17491758.
Serpa Neto, A., Cardoso, S.O., Manetta, J.A., Pereira, V.G., Esposito, D.C., Pasqualucci
Mde, O., Damasceno, M.C., Schultz, M.J., 2012. Association between use of
lung-protective ventilation with lower tidal volumes and clinical outcomes
among patients without acute respiratory distress syndrome: a meta-analysis.
JAMA 308, 16511659.
Sinderby, C., Navalesi, P., Beck, J., Skrobik, Y., Comtois, N., Friberg, S., Gottfried, S.B.,
Lindstrom, L., 1999. Neural control of mechanical ventilation in respiratory
failure. Nat. Med. 5, 14331436.
Spieth, P.M., Guldner, A., Beda, A., Carvalho, N., Nowack, T., Krause, A., Rentzsch, I.,
Suchantke, S., Thal, S.C., Engelhard, K., Kasper, M., Koch, T., Pelosi, P., de Abreu,
M.G., 2012. Comparative effects of proportional assist and variable pressure
support ventilation on lung function and damage in experimental lung injury.
Crit. Care Med. 40, 26542661.
The Acute Respiratory Distress Syndrome Network, 2000. Ventilation with lower
tidal volumes as compared with traditional tidal volumes for acute lung injury
and the acute respiratory distress syndrome. The Acute Respiratory Distress
Syndrome Network. N. Engl. J. Med. 342, 13011308.
von Euler, C., 1986. Brainstem mechanisms for generation and control of breathing
pattern. In: Handbook of Physiology. The Respiratory System. American
Physiological Society, Bethesda, Maryland, pp. 168.
Widdicombe, J.G., 1961. Respiratory reexes in man and other mammalian species.
Clin. Sci. 21, 163170.
Xirouchaki, N., Kondili, E., Vaporidi, K., Xirouchakis, G., Klimathianaki, M.,
Gavriilidis, G., Alexandopoulou, E., Plataki, M., Alexopoulou, C., Georgopoulos,
D., 2008. Proportional assist ventilation with load-adjustable gain factors in
critically ill patients: comparison with pressure support. Intensive Care Med.
34, 20262034.
Younes, M., 1992. Proportional assist ventilation, a new approach to ventilatory
support. Theory. Am. Rev. Respir. Dis. 145, 114120.
Younes, M., Georgopoulos, D., 1998. Control of breathing relevant to mechanical
ventilation. In: Marini, J., Slutsky, A. (Eds.), Physiological Basis of Ventilatory
Support. Dekker, New York, pp. 174.
Younes, M., Remmers, J., 1981. Control of tidal volume and respiratory frequency.
In: Hornbein, T. (Ed.), Control of Breathing. Marcel Deckker, pp. 163170.
Younes, M., Youssef, M., 1978. Effect of ve human anesthetics on respiratory
control in cats. J. Appl. Physiol.: Respir. Environ. Exerc. Physiol. 44, 596606.
Younes, M., Webster, K., Kun, J., Roberts, D., Masiowski, B., 2001. A method for
measuring passive elastance during proportional assist ventilation. Am. J.
Respir. Crit. Care Med. 164, 5060.
Younes, M., Kun, J., Webster, K., Roberts, D., 2002. Response of
ventilator-dependent patients to delayed opening of exhalation valve. Am. J.
Respir. Crit. Care Med. 166, 2130.