A PROTOCOL FOR - Home - Critical Care Londoncriticalcarelondon.ca/wp-content/uploads/2014/07/12...n engl j med 368;23 nejm.org june 6, 2013 2159 The new england journal of medicine

Martin Cieslak A PROTOCOL FOR

PRONING

¡  Background ¡  Rationale ¡  Update

OUTLINE

n engl j med 368;23 nejm.org june 6, 2013 2159

The new england journal of medicineestablished in 1812 june 6, 2013 vol. 368 no. 23

Prone Positioning in Severe Acute Respiratory Distress Syndrome

Claude Guérin, M.D., Ph.D., Jean Reignier, M.D., Ph.D., Jean-Christophe Richard, M.D., Ph.D., Pascal Beuret, M.D., Arnaud Gacouin, M.D., Thierry Boulain, M.D., Emmanuelle Mercier, M.D., Michel Badet, M.D.,

Alain Mercat, M.D., Ph.D., Olivier Baudin, M.D., Marc Clavel, M.D., Delphine Chatellier, M.D., Samir Jaber, M.D., Ph.D., Sylvène Rosselli, M.D., Jordi Mancebo, M.D., Ph.D., Michel Sirodot, M.D., Gilles Hilbert, M.D., Ph.D., Christian Bengler, M.D., Jack Richecoeur, M.D., Marc Gainnier, M.D., Ph.D., Frédérique Bayle, M.D.,

Gael Bourdin, M.D., Véronique Leray, M.D., Raphaele Girard, M.D., Loredana Baboi, Ph.D., and Louis Ayzac, M.D., for the PROSEVA Study Group*

A bs tr ac t

The authors’ affiliations are listed in the Appendix. Address reprint requests to Dr. Guérin at Service de Réanimation Médi-cale, Hôpital de la Croix-Rousse, 103 Grande Rue de la Croix-Rousse, 69004 Lyon, France, or at [email protected].

* The Proning Severe ARDS Patients (PROSEVA) study investigators are listed in the Supplementary Appendix, available at NEJM.org.

This article was published on May 20, 2013, at NEJM.org.

N Engl J Med 2013;368:2159-68. DOI: 10.1056/NEJMoa1214103Copyright © 2013 Massachusetts Medical Society.

BackgroundPrevious trials involving patients with the acute respiratory distress syndrome (ARDS) have failed to show a beneficial effect of prone positioning during mechanical ven-tilatory support on outcomes. We evaluated the effect of early application of prone positioning on outcomes in patients with severe ARDS.

MethodsIn this multicenter, prospective, randomized, controlled trial, we randomly as-signed 466 patients with severe ARDS to undergo prone-positioning sessions of at least 16 hours or to be left in the supine position. Severe ARDS was defined as a ratio of the partial pressure of arterial oxygen to the fraction of inspired oxygen (Fio2) of less than 150 mm Hg, with an Fio2 of at least 0.6, a positive end-expira-tory pressure of at least 5 cm of water, and a tidal volume close to 6 ml per kilogram of predicted body weight. The primary outcome was the proportion of patients who died from any cause within 28 days after inclusion.

ResultsA total of 237 patients were assigned to the prone group, and 229 patients were as-signed to the supine group. The 28-day mortality was 16.0% in the prone group and 32.8% in the supine group (P<0.001). The hazard ratio for death with prone position-ing was 0.39 (95% confidence interval [CI], 0.25 to 0.63). Unadjusted 90-day mortal-ity was 23.6% in the prone group versus 41.0% in the supine group (P<0.001), with a hazard ratio of 0.44 (95% CI, 0.29 to 0.67). The incidence of complications did not differ significantly between the groups, except for the incidence of cardiac arrests, which was higher in the supine group.

ConclusionsIn patients with severe ARDS, early application of prolonged prone-positioning ses-sions significantly decreased 28-day and 90-day mortality. (Funded by the Programme Hospitalier de Recherche Clinique National 2006 and 2010 of the French Ministry of Health; PROSEVA ClinicalTrials.gov number, NCT00527813.)

The New England Journal of Medicine Downloaded from nejm.org at UNIVERSITY OF WESTERN ONTARIO on September 30, 2013. For personal use only. No other uses without permission.

Copyright © 2013 Massachusetts Medical Society. All rights reserved.

BACKGROUND

Multicenter RCT 466 patients Early and prolonged proning with lung protective ventilation in severe ARDS Absolute 16.8% 28day mortality benefit NNT=6!

2013 PROSEVA TRIAL T h e n e w e ngl a nd j o u r na l o f m e dic i n e

n engl j med 368;23 nejm.org june 6, 20132166

of invasive mechanical ventilation, length of stay in the ICU, incidence of pneumothorax, rate of use of noninvasive ventilation after extubation, and tracheotomy rate did not differ significantly be-tween the two groups (Table 3).

ComplicationsA total of 31 cardiac arrests occurred in the su-pine group, and 16 in the prone group (P = 0.02). There were no significant differences between the groups with respect to other adverse effects (Table S6 in the Supplementary Appendix).

Discussion

Survival after severe ARDS was significantly high-er in the prone group than in the supine group. Furthermore, the effect size was large despite the fact that mortality in the supine group was lower than anticipated.

Our results are consistent with findings from previous meta-analyses2,11 and an observational study,18 even though prior randomized trials have failed to show a survival benefit with prone po-sitioning. Meta-analyses of ARDS studies have suggested that the outcomes with prone posi-tioning are better in the subgroup of patients with severe hypoxemia.2,11 However, when we stratified our analysis according to quartile of Pao2:Fio2

ratio at enrollment, we found no significant dif-ferences in outcomes (Table S8 in the Supple-mentary Appendix).

Several factors may explain our results. First, patients with severe ARDS were selected on the basis of oxygenation together with PEEP and Fio2 levels. Second, patients were included after a 12-to-24-hour period during which the ARDS criteria were confirmed. This period may have contributed to the selection of patients with more severe ARDS19 who could benefit from the ad-vantages of the prone positioning, such as relief of severe hypoxemia and prevention of ventilator-induced lung injury. A previous study has shown that prone positioning, as compared with supine positioning, markedly reduces the overinflated lung areas while promoting alveolar recruitment.5 These effects (reduction in overdistention and recruitment enhancement) may help prevent ven-tilator-induced lung injury by homogenizing the distribution of stress and strain within the lungs. In our trial, alveolar recruitment was not directly assessed. However, studies have shown that lung recruitability correlates with the extent of hypoxemia20,21 and that the transpulmonary pressure along the ventral-to-dorsal axis is more homogeneously distributed in the prone position than in the supine position.22 We therefore sug-gest that prone positioning in our patients induced a decrease in lung stress and strain.

Third, as in previous investigations,9,10 we used long prone-positioning sessions. Fourth, the prone position was applied for 73% of the time ascribed to the intervention and was concentrated over a period of a few days. Fifth, in our trial, the tidal volume was lower than in previous tri-als,9,10 and the PplatRS was kept below 30 cm of water. However, because all patients were returned to the supine position at least once a day, the effect of the prone position itself cannot be dis-tinguished from the effects of being moved from the supine to the prone position over the course of a day.

We should acknowledge that the technical aspects of prone positioning are not simple and that a coordinated team effort is required (see Videos 1 and 2, available at NEJM.org). All cen-ters participating in this study were skilled in the process of turning patients from the supine to the prone position, as shown by the absence of adverse events directly related to repositioning. Because the experience of the units may explain

Cum

ulat

ive

Prob

abili

ty o

f Sur

viva

l

1.0

0.6

0.8

0.4

0.2

0.00 10 20 30 40 80 90

Days

P<0.001

No. at RiskProne groupSupine group

237229

202163

191150

182136

7060

186139

50

Prone group

Supine group

4 col22p3

AUTHOR:

FIGURE:

SIZE

4-C H/TLine Combo

Revised

AUTHOR, PLEASE NOTE: Figure has been redrawn and type has been reset.

Please check carefully.

Guérin

2 of 2

ARTIST:

TYPE:

ts

06-06-13JOB: 36823 ISSUE:

Figure 2. Kaplan–Meier Plot of the Probability of Survival from Randomiza-tion to Day 90.

Videos showing prone positioning

of patients with ARDS are

available at NEJM.org

The New England Journal of Medicine Downloaded from nejm.org at UNIVERSITY OF WESTERN ONTARIO on September 30, 2013. For personal use only. No other uses without permission.

Copyright © 2013 Massachusetts Medical Society. All rights reserved.

Review Articles

www.ccmjournal.org 1261

main bronchus differed between prone and supine position-ing. However, the prevalence of endotracheal tube obstruction significantly increased with prone positioning; this result was similar to that of a study by Sud et al (8). Contrary to the find-ings of the Prone-Supine II trial, which reported a higher prev-alence of adverse events in the prone position group (including increased sedation, transient desaturation or hypotension, and loss of venous access or thoracostomy tube), the more recent PROSEVA trial showed no significant difference between the two groups with regard to transient desaturation or hypoten-sion. These discrepancies in adverse events might be due to dif-ferent protocols for prone positioning among the various ICUs (8, 27). Together with the previous RCTs and meta-analyses, our findings suggest that prone positioning during ventilation can be harmful and is a complicated procedure that requires a coordinated, highly skilled team effort (28). Although our meta-analysis supports a significant reduction in overall mor-tality in patients with ARDS, the risk of adverse events should be carefully considered during the decision-making process, especially in ICUs with less experience in prone positioning.

LimitationsThis meta-analysis included clinically and methodologically diverse studies. Although we included only RCTs for the final

analysis and statistical heterogeneity was insignificant, there were some differences in the enrollment criteria and the tar-get population of each study. Furthermore, different endpoint times were used for mortality evaluation (up to 180 d after admission). Because this study was a study-level meta-analysis, individual patient data were not included in the analysis; there-fore, we could not adjust for patient-level confounders.

CONCLUSIONSThe use of prone positioning during mechanical ventilation for patients with acute hypoxemic respiratory failure resulted in a sig-nificant reduction of overall mortality. The effect of prone posi-tioning was more obvious in patients with ARDS who underwent a longer duration of prone ventilation and lung protective venti-lation strategy. The prevalence of pressure ulcers and major air-way problems was significantly higher in the prone positioning group. Before using prone positioning, the risks and benefits for that particular patient with ARDS should be carefully weighed.

ACKNOWLEDGMENTSWe thank Dr. Soyeon Ahn, PhD, Medical Research Collaborating Center, Seoul National University Bundang Hospital, for dedicated statistical review of the study as an independent statistician.

TABLE 2. Adverse Events Related to Prone Positioning

Adverse Events

No. of Trials Reporting the

Outcome Events/ProneEvents/ Supine

Treatment Effect (Random-Effect Model) Heterogeneity

OR (95% CI) p

Number Needed to Treat/Number Needed to Harm I2 (%) p

Ventilator- associated pneumonia

6 120/567 128/513 0.76 (0.44–1.33) 0.343 26 34.4 0.192

Pressure ulcers 6 294/698 218/646 1.49 (1.18–1.89) 0.001 12 0.0 0.617

Major airway problema

9 255/1,104 180/1,063 1.55 (1.10–2.17) 0.012 16 32.7 0.167

Unplanned extubation

7 113/1,091 98/1,050 1.17 (0.80–1.73) 0.421 98 25.5 0.234

Selective intubation

2 12/642 5/615 2.73 (0.29–25.46) 0.378 95 55.9 0.132

Endotracheal tube obstruction

4 130/823 77/802 2.16 (1.53–3.05) < 0.001 16 0.0 0.580

Loss of venous or arterial access

4 36/407 22/397 1.34 (0.29–6.26) 0.712 30 75.5 0.007

Thoracostomy tube dislodgement or kinking

4 14/407 14/397 1.14 (0.35–3.75) 0.827 1,154 42.6 0.175

Pneumothorax 4 29/513 33/462 0.77 (0.46–1.30) 0.333 67 0.0 0.528

Cardiac arrest 3 104/718 119/675 0.74 (0.47–1.17) 0.197 32 30.3 0.238

Tachyarrhythmia or bradyarrhythmia

3 115/663 102/634 1.08 (0.78–1.50) 0.643 80 8.8 0.334

a

Critical Care Medicine May 2014 Vol 42-5

Check! ROYAL COLLEGE

META ANALYSIS

Acute respiratory distress syndrome(ARDS) is a fulminant disease charac-terized by impaired oxygenation, pul-

monary congestion and decreased lung com -pliance following a direct pulmonary insult,such as aspiration or pneumonia, or a systemicinjury, such as sepsis or trauma. In many pa -tients with ARDS, mechanical ventilation canfurther injure damaged lungs,1–3 leading to long-term morbidity and mortality. Although mini-mizing tidal volumes4 and optimizing positiveend-expiratory pressure (PEEP)5 may reducelung injury, mortality associated with ARDSremains high.6,7

Placing patients in the prone position for aportion of time each day during mechanical ven-tilation, first suggested in 1974,8 is sometimesused as a protective lung strategy in patients withARDS.9,10 Alveolar distension varies regionallybecause of gravity and anatomic relationshipswith the chest wall and heart. In the prone posi-

tion, the volume of lung collapsed under its ownweight and that of the heart is decreased relativeto the supine position.10,11 Because pulmonaryperfusion is preserved in both the ventral anddorsal lung regions, ventilation–perfusion match-ing is improved in the prone position.12 Morehomogenous dispersion of tidal volume in theprone position may minimize alveolar stretchand strain.10 Improvements in oxygenation mayreduce the risk of death from hypoxia.13,14

Early randomized controlled trials (RCTs) ofprone positioning did not show reductions inmortality.15,16 However, these trials includedpatients with mild ARDS, the duration of pronepositioning each day was short, and protectivelung ventilation (low tidal volumes) was notused. A subsequent meta-analysis suggested thatprone positioning reduces mortality amongpatients with severe hypoxemia.13 Because of theavailability of new data,17 we undertook a sys-tematic review and meta- analysis, in collabora-

Competing interests: JordiMancebo received grantsfrom Covidien and GeneralElectric for research onproportional assistventilation and functionalresidual capacity; he servedon the data monitoring andsafety board for a trialsponsored by Air Liquide,and on the steeringcommittees for FaronPharmaceuticals and ALungTechnologies. LucianoGattinoni was a member ofan advisory board for KCIMedical. No othercompeting interests weredeclared.

This article has been peerreviewed.

Correspondence to:Sachin Sud, [email protected]

CMAJ 2014. DOI:10.1503/cmaj.140081

ResearchCMAJ

Background: Mechanical ventilation in theprone position is used to improve oxygena-tion and to mitigate the harmful effects ofmechanical ventilation in patients with acuterespiratory distress syndrome (ARDS). Wesought to determine the effect of prone posi-tioning on mortality among patients withARDS receiving protective lung ventilation.

Methods: We searched electronic databasesand conference proceedings to identify rele-vant randomized controlled trials (RCTs) pub-lished through August 2013. We included RCTsthat compared prone and supine positioningduring mechanical ventilation in patients withARDS. We assessed risk of bias and obtaineddata on all-cause mortality (determined at hos-pital discharge or, if unavailable, after longestfollow-up period). We used random-effectsmodels for the pooled analyses.

Results: We identified 11 RCTs (n = 2341) thatmet our inclusion criteria. In the 6 trials (n =1016) that used a protective ventilation strat-egy with reduced tidal volumes, prone posi-tioning significantly reduced mortality (riskratio 0.74, 95% confidence interval 0.59–0.95; I2 = 29%) compared with supine posi-tioning. The mortality benefit remained inseveral sensitivity analyses. The overall qual-ity of evidence was high. The risk of biaswas low in all of the trials except one, whichwas small. Statistical heterogeneity was low(I2 < 50%) for most of the clinical and physio-logic outcomes.

Interpretation: Our analysis of high-quality evi-dence showed that use of the prone positionduring mechanical ventilation improved survivalamong patients with ARDS who received protec-tive lung ventilation.

Abstract

Effect of prone positioning during mechanical ventilationon mortality among patients with acute respiratorydistress syndrome: a systematic review and meta-analysis

Sachin Sud MD MSc, Jan O. Friedrich MD DPhil, Neill K. J. Adhikari MDCM MSc, Paolo Taccone MD,Jordi Mancebo MD, Federico Polli MD, Roberto Latini MD, Antonio Pesenti MD, Martha A.Q. Curley RN PhD,Rafael Fernandez MD, Ming-Cheng Chan MD, Pascal Beuret MD, Gregor Voggenreiter MD,Maneesh Sud MD, Gianni Tognoni MD, Luciano Gattinoni MD, Claude Guérin MD PhD

© 2014 Canadian Medical Association or its licensors CMAJ, July 8, 2014, 186(10) E381

Acute respiratory distress syndrome(ARDS) is a fulminant disease charac-terized by impaired oxygenation, pul-

monary congestion and decreased lung com -pliance following a direct pulmonary insult,such as aspiration or pneumonia, or a systemicinjury, such as sepsis or trauma. In many pa -tients with ARDS, mechanical ventilation canfurther injure damaged lungs,1–3 leading to long-term morbidity and mortality. Although mini-mizing tidal volumes4 and optimizing positiveend-expiratory pressure (PEEP)5 may reducelung injury, mortality associated with ARDSremains high.6,7

Placing patients in the prone position for aportion of time each day during mechanical ven-tilation, first suggested in 1974,8 is sometimesused as a protective lung strategy in patients withARDS.9,10 Alveolar distension varies regionallybecause of gravity and anatomic relationshipswith the chest wall and heart. In the prone posi-

tion, the volume of lung collapsed under its ownweight and that of the heart is decreased relativeto the supine position.10,11 Because pulmonaryperfusion is preserved in both the ventral anddorsal lung regions, ventilation–perfusion match-ing is improved in the prone position.12 Morehomogenous dispersion of tidal volume in theprone position may minimize alveolar stretchand strain.10 Improvements in oxygenation mayreduce the risk of death from hypoxia.13,14

Early randomized controlled trials (RCTs) ofprone positioning did not show reductions inmortality.15,16 However, these trials includedpatients with mild ARDS, the duration of pronepositioning each day was short, and protectivelung ventilation (low tidal volumes) was notused. A subsequent meta-analysis suggested thatprone positioning reduces mortality amongpatients with severe hypoxemia.13 Because of theavailability of new data,17 we undertook a sys-tematic review and meta- analysis, in collabora-

Competing interests: JordiMancebo received grantsfrom Covidien and GeneralElectric for research onproportional assistventilation and functionalresidual capacity; he servedon the data monitoring andsafety board for a trialsponsored by Air Liquide,and on the steeringcommittees for FaronPharmaceuticals and ALungTechnologies. LucianoGattinoni was a member ofan advisory board for KCIMedical. No othercompeting interests weredeclared.

This article has been peerreviewed.

Correspondence to:Sachin Sud, [email protected]

CMAJ 2014. DOI:10.1503/cmaj.140081

ResearchCMAJ

Background: Mechanical ventilation in theprone position is used to improve oxygena-tion and to mitigate the harmful effects ofmechanical ventilation in patients with acuterespiratory distress syndrome (ARDS). Wesought to determine the effect of prone posi-tioning on mortality among patients withARDS receiving protective lung ventilation.

Methods: We searched electronic databasesand conference proceedings to identify rele-vant randomized controlled trials (RCTs) pub-lished through August 2013. We included RCTsthat compared prone and supine positioningduring mechanical ventilation in patients withARDS. We assessed risk of bias and obtaineddata on all-cause mortality (determined at hos-pital discharge or, if unavailable, after longestfollow-up period). We used random-effectsmodels for the pooled analyses.

Results: We identified 11 RCTs (n = 2341) thatmet our inclusion criteria. In the 6 trials (n =1016) that used a protective ventilation strat-egy with reduced tidal volumes, prone posi-tioning significantly reduced mortality (riskratio 0.74, 95% confidence interval 0.59–0.95; I2 = 29%) compared with supine posi-tioning. The mortality benefit remained inseveral sensitivity analyses. The overall qual-ity of evidence was high. The risk of biaswas low in all of the trials except one, whichwas small. Statistical heterogeneity was low(I2 < 50%) for most of the clinical and physio-logic outcomes.

Interpretation: Our analysis of high-quality evi-dence showed that use of the prone positionduring mechanical ventilation improved survivalamong patients with ARDS who received protec-tive lung ventilation.

Abstract

Effect of prone positioning during mechanical ventilationon mortality among patients with acute respiratorydistress syndrome: a systematic review and meta-analysis

Sachin Sud MD MSc, Jan O. Friedrich MD DPhil, Neill K. J. Adhikari MDCM MSc, Paolo Taccone MD,Jordi Mancebo MD, Federico Polli MD, Roberto Latini MD, Antonio Pesenti MD, Martha A.Q. Curley RN PhD,Rafael Fernandez MD, Ming-Cheng Chan MD, Pascal Beuret MD, Gregor Voggenreiter MD,Maneesh Sud MD, Gianni Tognoni MD, Luciano Gattinoni MD, Claude Guérin MD PhD

© 2014 Canadian Medical Association or its licensors CMAJ, July 8, 2014, 186(10) E381

Interpretation

Our analysis of high-quality evidence showedthat prone positioning during mechanical venti-lation reduces mortality among patients withARDS receiving protective lung ventilation. Thequality of evidence was high, and the numberneeded to treat to save one life was 11 (95% CI6–50). Our findings complement those of arecent positive RCT17 and showed consistency ofeffect across previous RCTs and in the sensitiv-ity analyses.

Most RCTs of prone positioning duringmechanical ventilation in patients with ARDSfailed on their own to show statistically significantreductions in mortality despite improvements inoxygenation.15,16,36,39 Previous systematic reviewswere similarly unable to show reductions in mor-tality,41−43 although some suggested a mortalitybenefit among sicker patients.13,42 Limitations ofearlier trials, including use of injurious tidal vol-umes (> 8 mL/kg of predicted body weight),enrolment of patients with mild ARDS,15,16,35,37−39 andinadequate duration of prone positioning,15,16,38,39

Research

CMAJ, July 8, 2014, 186(10) E387

Table 3: Results of primary and sensitivity analyses for the effect of prone positioning during mechanical ventilation on mortality among patients with acute respiratory distress syndrome (ARDS)

Analysis* No. of trials

No. of deaths, n/N

Risk ratio (95% CI)

I2 value, %

Primary

Trials mandating protective ventilation† 6 363/1016 0.74 (0.59–0.95) 29

Sensitivity

Included all trials‡ 10 797/1869 0.86 (0.73–1.00) 42

Assumed patients lost to follow-up lived 6 363/1020 0.74 (0.59–0.95) 28

Assumed patients lost to follow-up died 6 366/1020 0.74 (0.59–0.94) 26

Excluded trial in which allocation was not concealed35

5 352/994 0.73 (0.55–0.98) 43

Excluded trial with pediatric population37 5 355/914 0.73 (0.56–0.96) 42

Included trial that used moderate tidal volume (< 10 mL/kg)36

7 438/1152 0.77 (0.65–0.91) 16

Fixed-effects model 6 363/1016 0.74 (0.63–0.87) 29

Note: CI = confidence interval. *Random-effects models were used for all analyses except in the final sensitivity analysis. †Tidal volume < 8 mL/kg of predicted or actual body weight. ‡For the 2 trials that enrolled some patients without ARDS,16,39 we included only patients whose condition met the authors’ definition of ARDS; when the analysis was redone to include all patients in these trials, the risk ratio changed minimally (0.87, 95% CI 0.74–1.02; I2 = 48%).

0.1 1 10

MandatedNot mandated

≥ 16 h/d< 16 h/d

SevereModerateMild

0.74 (CI 0.59–0.95)0.98 (CI 0.86–1.12)

0.77(CI 0.64–0.92)1.02 (CI 0.88–1.17)

0.76 (CI 0.61–0.94)0.74 (CI 0.48–1.16)0.98 (CI 0.18–5.24)

RR (95% CI)

p = 0.05

p = 0.02

p > 0.9

VariableNo. oftrials RR (95% CI)

I2 value,%

64

64

664

290

210

0420

Protective lung ventilation

Duration of prone positioning

Level of hypoxemia*

Deaths, n/N

Prone SupineFavours

proneFavourssupine

154/510229/458

191/565192/403

75/21075/2743/22

209/506205/395

243/547171/354

102/209102/268

3/23

Figure 3: Effect of prone positioning during mechanical ventilation on all-cause mortality according to prespecified patient-level andtrial-level subgroups. Risk ratios less than 1.0 indicate a decreased risk of death with prone positioning. *Severe hypoxemia = ratio ofpartial pressure of arterial oxygen to fraction of inspired oxygen (PaO2/FIO2) < 100 mm Hg; moderate = PaO2/FIO2 ratio 100–199 mm Hg;mild = PaO2/FIO2 ratio 200–299 mm Hg. CI = 95% confidence interval, RR = risk ratio. Baseline PaO2/FIO2 ratios were unavailable for 10patients in 3 trials.17,34,35

Research

E384C

MA

J, July 8, 2014, 186(10)

Table 1: Characteristics of randomized controlled trials included in the systematic review*

Study

Patient characteristics at baseline Details of prone positioning Protective lung

ventilation† mandated

No. of patients Enrolment criteria

Mean PaO2/FIO2 ratio, mm Hg

Mean PEEP, cm H2O

Mean FIO2, %

Duration of ARDS before enrolment Mean duration

Criteria for discontinuation

Guérin et al.,17 2013

474 ARDS‡ with PaO2/FIO2 ratio < 150 mm Hg, PEEP ≥ 5 cm H2O and FIO2 ≥ 60% after stabilization period of 12–24 h

100 10 79 < 36 h 17 h daily for 4 d PaO2/FIO2 ratio ≥ 150 mm Hg, PEEP ≤ 10 cm H2O and FIO2 ≤ 0.6

Yes

Taccone et al.,14 2009

344 ARDS‡ with PEEP ≥ 5 cm H2O 113 10 72 < 72 h 18 h daily for 8.3 d FIO2 ≤ 40% and PEEP ≤ 10 cm H2O

Yes

Fernandez et al.,34 2008

42 ARDS‡ 118 11 85 < 48 h 18 h daily§ PaO2/FIO2 ratio > 250 mm Hg and PEEP ≤ 8 cm H2O for 12 h

Yes

Chan et al.,35 2007 22 ARDS‡ secondary to community-acquired pneumonia

109 13 88 < 72 h 24 h daily for 4.4 d SpO2 > 90%, FIO2 < 60% for > 24 h (after 72 h)

Yes

Mancebo et al.,36 2006

142 ARDS‡ with infiltrates in 4 quadrants on chest radiograph

105 7 82 < 48 h 17 h daily for 10.1 d FIO2 ≤ 45% and PEEP ≤ 5 cm H2O

No

Curley et al.,37 2005

102 children (age 2 wk–

18 yr)

Acute lung injury or ARDS‡ 100 9 60 < 48 h 18 h daily for 4 d Spontaneous breathing and O2 index < 6

Yes

Voggenreiter et al.,38 2005

40 Acute lung injury (duration ≥ 24 h) or ARDS‡ (duration ≥ 8 h) with PEEP ≥ 5 cm H2O

221 11.5 49 < 48 h 11 h daily for 7 d PaO2/FIO2 ratio > 300 mm Hg for 48 h

Yes


802 Hypoxemic acute respiratory failure (n = 413 with acute lung injury or ARDS‡)

152 8 66 > 12–24 h 9 h daily for 4.1 d Clinical improvement¶ No

Beuret et al.,39 2002

53 Coma, intubation required (n = 7 with acute lung injury or ARDS‡)

326 (238 in pts with

acute lung injury or ARDS)

NR NR < 24 h 4 h daily for 6.0 d Able to sit in chair No

Watanabe et al.,40 2002

16 PaO2/FIO2 ratio < 200 mm Hg, PEEP > 5 cm H2O 5 d after esophagectomy

166** NR NR Not prespecified 6 h daily for 4 d Not applicable No

Gattinoni et al.,15 2001

304 Acute lung injury or ARDS‡ with PEEP ≥ 5 cm H2O

127 10 73 Not prespecified 7 h daily for 4.7 d None No

Note: ARDS = acute respiratory distress syndrome, FIO2 = fractional concentration of inspired oxygen, NR = not reported, PaO2 = partial pressure of arterial oxygen, PEEP = positive end-expiratory pressure. *Additional details about the study characteristics are available in Appendix 3 (www.cmaj.ca/lookup/suppl/doi:10.1503/cmaj.140081/-/DC1). †Tidal volume < 8 mL/kg of predicted body weight. ‡Defined according to the criteria of the American–European Consensus Conference.18 §Average over first 2 d only; average number of days that prone positioning was used was not available. ¶Defined by 1 major criterion (improvement in PaO2/FIO2 ratio ≥ 30% relative to value at randomization, with FIO2 ≤ 60%) and at least 1 minor criterion (PEEP ≤ 8 cm H2O, no sepsis, cause of acute respiratory failure under control [stable or signs of improvement on chest radiograph, and < 3 organ dysfunctions, including lung dysfunction]). **Prone group only (baseline ratio in supine group not reported).

¡  Since summer of 2013 proning has become more common in our ICUs

¡  Although best clinical judgment to follow the PROSEVA trial’s protocol as a guideline are used, there remains inconsistency in the timing and duration of proning in our ARDS patients.

RATIONALE

¡  Introduce and implement an evidence-based clinical decision tool for implementing proning in ARDS §  Including rationale for each decision made

¡  Criteria § What are the indications and contraindications to Proning?

¡  Timing § When to start and stop?

¡  Implementation

PROJECT GOAL

UPDATE

¡ Big Project

¡ KISS!!

FEEDBACK

¡  Criteria § What are the indications and contraindications to Proning?

¡  Timing § When to start and stop?

¡  Implementation

Indications (need all!) 1.  Confirmed ARDS ≤36hrs 2.  12-24hrs of Ventilator Optimization 3.  PaO2/FiO2 ≤150 4.  PEEP ≥ 10cmH2O or ≥14cmH2O 5.  Central Line Access 6.  Functioning Arterial Line (if not, consider brachial)

¡  FiO2 Criteria? §  PROSEVA added FiO2 ≥ 0.6…. More on this later.

CRITERIA

¡  PaO2/FiO2 ≤ 150 ¡  PEEP ≥ 10cmH2O

CRITERIA

Research

E384C

MA

J, July 8, 2014, 186(10)

Table 1: Characteristics of randomized controlled trials included in the systematic review*

Study

Patient characteristics at baseline Details of prone positioning Protective lung

ventilation† mandated

No. of patients Enrolment criteria

Mean PaO2/FIO2 ratio, mm Hg

Mean PEEP, cm H2O

Mean FIO2, %

Duration of ARDS before enrolment Mean duration

Criteria for discontinuation


474 ARDS‡ with PaO2/FIO2 ratio < 150 mm Hg, PEEP ≥ 5 cm H2O and FIO2 ≥ 60% after stabilization period of 12–24 h

100 10 79 < 36 h 17 h daily for 4 d PaO2/FIO2 ratio ≥ 150 mm Hg, PEEP ≤ 10 cm H2O and FIO2 ≤ 0.6

Yes

Taccone et al.,14 2009

344 ARDS‡ with PEEP ≥ 5 cm H2O 113 10 72 < 72 h 18 h daily for 8.3 d FIO2 ≤ 40% and PEEP ≤ 10 cm H2O

Yes

Fernandez et al.,34 2008

42 ARDS‡ 118 11 85 < 48 h 18 h daily§ PaO2/FIO2 ratio > 250 mm Hg and PEEP ≤ 8 cm H2O for 12 h

Yes

Chan et al.,35 2007 22 ARDS‡ secondary to community-acquired pneumonia

109 13 88 < 72 h 24 h daily for 4.4 d SpO2 > 90%, FIO2 < 60% for > 24 h (after 72 h)

Yes

Mancebo et al.,36 2006

142 ARDS‡ with infiltrates in 4 quadrants on chest radiograph

105 7 82 < 48 h 17 h daily for 10.1 d FIO2 ≤ 45% and PEEP ≤ 5 cm H2O

No

Curley et al.,37 2005

102 children (age 2 wk–

18 yr)

Acute lung injury or ARDS‡ 100 9 60 < 48 h 18 h daily for 4 d Spontaneous breathing and O2 index < 6

Yes

Voggenreiter et al.,38 2005

40 Acute lung injury (duration ≥ 24 h) or ARDS‡ (duration ≥ 8 h) with PEEP ≥ 5 cm H2O

221 11.5 49 < 48 h 11 h daily for 7 d PaO2/FIO2 ratio > 300 mm Hg for 48 h

Yes


802 Hypoxemic acute respiratory failure (n = 413 with acute lung injury or ARDS‡)

152 8 66 > 12–24 h 9 h daily for 4.1 d Clinical improvement¶ No

Beuret et al.,39 2002

53 Coma, intubation required (n = 7 with acute lung injury or ARDS‡)

326 (238 in pts with

acute lung injury or ARDS)

NR NR < 24 h 4 h daily for 6.0 d Able to sit in chair No

Watanabe et al.,40 2002

16 PaO2/FIO2 ratio < 200 mm Hg, PEEP > 5 cm H2O 5 d after esophagectomy

166** NR NR Not prespecified 6 h daily for 4 d Not applicable No

Gattinoni et al.,15 2001

304 Acute lung injury or ARDS‡ with PEEP ≥ 5 cm H2O

127 10 73 Not prespecified 7 h daily for 4.7 d None No

Note: ARDS = acute respiratory distress syndrome, FIO2 = fractional concentration of inspired oxygen, NR = not reported, PaO2 = partial pressure of arterial oxygen, PEEP = positive end-expiratory pressure. *Additional details about the study characteristics are available in Appendix 3 (www.cmaj.ca/lookup/suppl/doi:10.1503/cmaj.140081/-/DC1). †Tidal volume < 8 mL/kg of predicted body weight. ‡Defined according to the criteria of the American–European Consensus Conference.18 §Average over first 2 d only; average number of days that prone positioning was used was not available. ¶Defined by 1 major criterion (improvement in PaO2/FIO2 ratio ≥ 30% relative to value at randomization, with FIO2 ≤ 60%) and at least 1 minor criterion (PEEP ≤ 8 cm H2O, no sepsis, cause of acute respiratory failure under control [stable or signs of improvement on chest radiograph, and < 3 organ dysfunctions, including lung dysfunction]). **Prone group only (baseline ratio in supine group not reported).

CONTRAINDICATIONS

¡  ABSOLUTE Contraindications §  Ongoing hemodynamic instability despite vasopressors (5 Trials) §  Elevated Intracranial Pressure (6 Trials) §  Unstable Fractures (Spine, long-bone, Pelvis) (6 Trials) §  Pregnancy (2nd or 3rd Trimester) (2 Trials)

¡  Relative Contraindications § Massive haemoptysis needing urgent treatment (2 Trials) §  Tracheal or thoracic surgery in the last 15 days (2 Trials) §  Facial trauma or surgery in the last 15 days (1 Trial) §  DVT or PE treated in the last 2 days (2 Trials) §  Cardiac pace maker implantation in the last 2 days (2 Trials) §  Bronchopleural fistula treated with a single anterior chest tube (risk

of kinking in prone position) (2 Trials)

1.  Key Differences in ARDS protocols in studies and current LHSC routine practice

1.  Criteria to Start and Stop – PEEP Tables 2.  May require a general consensus on how we treat ARDS in

LHSC 3.  Streamline timing of when patients are Supine

TIMING

¡  ALL major trials investigating proning to date have used the “classic ARDSnet” PEEP table

¡  LHSC has adopted high-PEEP LOVS-protocol ventilation strategy for managing ARDS patients

¡  Is there a role for high-PEEP ventilation in the prone patient?

HIGH PEEP VERSUS LOW PEEP

FiO2 0.3 0.4 0.4 0.5 0.5 0.6 0.7 0.7 0.7 0.8 0.9 0.9 0.9 1.0 1.0 1.0 1.0

PEEP 5 5 8 8 10 10 10 12 14 14 14 16 18 18 20 22 24

FiO2 0.3 0.3 0.3 0.4 0.4 0.4 0.4 0.4 0.5 0.5 0.6 7.0 0.8 0.8 0.9 1.0 1.0

PEEP 5 8 10 10 12 14 16 18 18 20 20 20 20 22 22 22 24

LOVS TRIAL

mation to the exact binomial method ofCassagrande et al).14 An independentdata monitoring committee conducted2 interim analyses using a nominalP!.001 as a threshold to consider earlystopping.

The primary analysis was a Mantel-Haenszel analysis of hospital mortal-ity, using center as the single stratifi-cation variable.

In a planned secondary analysis ofhospital mortality, we adjusted for 4baseline variables: age, the Acute Physi-

ology Score component of the AcutePhysiology and Chronic Health Evalu-ation (APACHE) II score,15 sepsis, andduration of hospitalization. To pre-sent study results as relative risks, weplanned to use the exact log–binomialapproach. With failure to convergeusing this method, we used an indi-rect logistic regression analysis, usingthe bootstrap method to derive confi-dence intervals.16 We also conducted asubgroup analysis to investigate an in-teraction between severity of lung in-

jury at baseline, defined by quartiles ofPaO2/FIO2, and treatment effect.

Four sensitivity analyses address-ing the outcome of hospital mortalityexamined potential bias introduced bythe blocked randomization program-ming error; these results did not differfrom our primary analysis. Formal com-parisons of 25 baseline characteristicsusing the Bonferroni correction re-vealed no statistically significant im-balances. Analyses of the duration ofmechanical ventilation and hospital-

Table 4. Respiratory Dataa

Variables

Day 1 Day 3 Day 7

Lung OpenVentilation Control

PValue


PValue


PValue

Tidal volume, mean (SD),mL/kg predictedbody weight

6.8 (1.4) 6.8 (1.3) .76 6.9 (1.5) 6.7 (1.5) .02 6.9 (1.3) 7.0 (1.6) .53

No. of patients 436 469 337 395 177 243Total respiratory rate,

mean (SD), /min25.2 (6.6) 26.0 (6.5) .08 25.1 (6.6) 27.1 (8.0) !.001 25.5 (8.0) 26.1 (7.6) .26

No. of patients 471 507 447 479 316 351Plateau pressure, mean (SD),

cm H2O30.2 (6.3) 24.9 (5.1) !.001 28.6 (6.0) 24.7 (5.7) !.001 28.8 (6.3) 25.1 (6.8) !.001

No. of patients 435 424 334 380 174 23230.1-35.0 112 33 76 38 37 2735.1-40.0 88 4 41 12 27 13"40.0 8 1 8 3 4 4

FIO2, mean (SD) 0.50 (0.16) 0.58 (0.17) !.001 0.41 (0.12) 0.52 (0.16) !.001 0.39 (0.12) 0.48 (0.17) !.001No. of patients 471 507 447 482 319 356

Set PEEP, mean (SD), cm H2OAll patients 15.6 (3.9) 10.1 (3.0) !.001 11.8 (4.1) 8.8 (3.0) !.001 10.3 (4.3) 8.0 (3.1) !.001

No. of patients 471 507 447 479 316 348First 161 patients 15.3 (3.6) 10.6 (2.9) !.001 12.1 (4.1) 9.3 (3.0) !.001 10.4 (4.3) 8.2 (3.1) .005

No. of patients 77 82 72 79 47 63Subsequent 822 patients 15.7 (4.0) 10.0 (3.0) !.001 11.8 (4.1) 8.7 (3.0) !.001 10.3 (4.3) 8.0 (3.1) !.001

No. of patients 394 425 375 400 269 285I:E ratio, mean (SD) 0.62 (0.19) 0.56 (0.19) !.001 0.64 (0.21) 0.56 (0.21) !.001 0.64 (0.19) 0.59 (0.22) .02

No. of patients 410 420 329 373 170 212PaO2/FIO2, mean (SD) 187.4 (68.8) 149.1 (60.6) !.001 196.8 (60.6) 164.1 (63.5) !.001 212.7 (70.5) 180.8 (73.0) !.001

No. of patients 464 498 444 472 314 342PaO2, mean (SD), mm Hg 88.1 (32.0) 80.1 (25.2) !.001 75.3 (14.8) 76.4 (16.2) .30 76.3 (15.6) 77.0 (17.1) .56

No. of patients 464 498 444 472 314 342PaCO2, mean (SD), mm Hg 45.5 (12.0) 44.6 (10.9) .22 44.8 (9.5) 45.3 (9.8) .41 44.8 (10.3) 46.6 (11.7) .04

No. of patients 464 498 444 472 314 342pH, mean (SD) 7.33 (0.10) 7.35 (0.09) .17 7.38 (0.07) 7.37 (0.08) .11 7.40 (0.07) 7.39 (0.08) .03

No. of patients 464 498 444 472 314 34224-h fluid balance, mean (SD), mL 2131.4

(2506.6)2110.6(2641.7)

.90 1029.0(2222.9)

722.9(2201.4)

.04 270.6(2078.2)

102.4(1808.4)

.26

No. of patients 465 500 445 473 326 363Abbreviations: FIO2, fraction of inspired oxygen; I:E, inspiration:expiration; PEEP, positive end-expiratory pressure; PaO2, partial pressure of arterial oxygen; PaCO2, partial pressure of

arterial carbon dioxide.aData shown were derived from the average value obtained for each patient over 3 measurements each day. Values were recorded on days 1, 3, and 7 after enrollment. For tidal volume

and plateau airway pressure measurements, data exclude patients weaning in pressure support mode, with FIO2 less than or equal to 0.40 and PEEP less than or equal to 10 cm H2O.

VENTILATION STRATEGY FOR ACUTE LUNG INJURY

©2008 American Medical Association. All rights reserved. (Reprinted) JAMA, February 13, 2008—Vol 299, No. 6 641

Downloaded From: http://jama.jamanetwork.com/ by a Western University User on 01/17/2014








Variables

Day 1 Day 3 Day 7


PValue


PValue


PValue


6.8 (1.4) 6.8 (1.3) .76 6.9 (1.5) 6.7 (1.5) .02 6.9 (1.3) 7.0 (1.6) .53


mean (SD), /min25.2 (6.6) 26.0 (6.5) .08 25.1 (6.6) 27.1 (8.0) !.001 25.5 (8.0) 26.1 (7.6) .26


cm H2O30.2 (6.3) 24.9 (5.1) !.001 28.6 (6.0) 24.7 (5.7) !.001 28.8 (6.3) 25.1 (6.8) !.001

No. of patients 435 424 334 380 174 23230.1-35.0 112 33 76 38 37 2735.1-40.0 88 4 41 12 27 13"40.0 8 1 8 3 4 4











(2506.6)2110.6(2641.7)

.90 1029.0(2222.9)

722.9(2201.4)

.04 270.6(2078.2)

102.4(1808.4)

.26














Variables

Day 1 Day 3 Day 7


PValue


PValue


PValue


6.8 (1.4) 6.8 (1.3) .76 6.9 (1.5) 6.7 (1.5) .02 6.9 (1.3) 7.0 (1.6) .53


mean (SD), /min25.2 (6.6) 26.0 (6.5) .08 25.1 (6.6) 27.1 (8.0) !.001 25.5 (8.0) 26.1 (7.6) .26


cm H2O30.2 (6.3) 24.9 (5.1) !.001 28.6 (6.0) 24.7 (5.7) !.001 28.8 (6.3) 25.1 (6.8) !.001

No. of patients 435 424 334 380 174 23230.1-35.0 112 33 76 38 37 2735.1-40.0 88 4 41 12 27 13"40.0 8 1 8 3 4 4











(2506.6)2110.6(2641.7)

.90 1029.0(2222.9)

722.9(2201.4)

.04 270.6(2078.2)

102.4(1808.4)

.26







with a recruitment maneuver: 61(4.5%) resulted in a mean arterialpressure of less than 60 mm Hg, 58(4.2%) were associated with a decreasein oxygen saturation to less than 85%,24 (1.8%) were associated with brady-cardia or tachycardia, 4 (0.3%) wereassociated with cardiac arrhythmia,and 4 (0.3%) were associated with anew air leak through an existing tho-racostomy tube. In 3 patients, clini-cians detected new barotrauma imme-diately following a recruitmentmaneuver. TABLE 5 summarizes theuse of selected intensive care unitinterventions, which clinicians admin-istered similarly in both groups.

There were 173 hospital deaths(36.4%) in the experimental group and205 (40.4%) in the control group. Therelative risk of death in the hospital was0.90 (95% confidence interval, 0.77-1.05; P=.19) (FIGURE 2 and TABLE 6).The secondary adjusted analysis of hos-pital mortality showed a relative risk of0.97 (95% confidence interval, 0.84-1.12; P= .74). We found no interac-tion between severity of baseline lunginjury and response to treatment(TABLE 7).

There were 53 experimental pa-tients vs 47 controls who developed anepisode of barotrauma, for an abso-lute difference of 6 events. There wasa lower incidence of refractory hypox-emia as a cause for deviation from theassigned ventilation settings, and alower rate of associated deaths, amongpatients in the experimental group(Table 6). The median duration of me-chanical ventilation among survivors ofmechanical ventilation was 10 days (in-terquartile range, 6-17 days) in the ex-perimental group and 10 days (inter-quartile range, 6-16 days) in the controlgroup (P=.92). The median durationof hospitalization among survivors was28 days (interquartile range, 17-48days) vs 29 days (interquartile range,16-51 days) (P=.96).

COMMENTThis trial comparing 2 lung-protec-tive ventilation strategies, an estab-lished low-tidal-volume strategy and an

experimental lung open ventilationstrategy that includes low tidal vol-umes, recruitment maneuvers, andhigher levels of PEEP, resulted in no sta-tistically significant difference in ratesof all-cause hospital mortality. Thelower mortality rate observed in the ex-perimental group was not statisticallysignificant and became negligible in asecondary adjusted analysis. The 2 strat-egies resulted in similar rates of baro-trauma and similar duration of me-chanical ventilation. The experimentalstrategy was associated with less use of

rescue therapies and fewer deaths as-sociated with refractory hypoxemia.

A number of hypotheses could ex-plain the similar mortality rates we ob-served. First, our experimental strat-egy may have no appreciable impact onsurvival beyond that achieved with lowtidal volumes and standard PEEP lev-els alone. Alternatively, the experimen-tal strategy may reduce deaths amongpatients similar to those studied; how-ever, our trial did not have sufficientpower to detect a relatively small mor-tality reduction. Finally, benefits to the

Table 6. Outcomesa

Outcomes

No. (%)

Relative Risk(95% Confidence

Interval)P

Value

Lung OpenVentilation(n = 475)

ControlVentilation(n = 508)

Death in hospital 173 (36.4) 205 (40.4) 0.90 (0.77-1.05) .19Death in intensive care unit 145 (30.5) 178 (35.0) 0.87 (0.73-1.04) .13Death during mechanical

ventilation136 (28.6) 168 (33.1) 0.87 (0.72-1.04) .13

Death during first 28 d 135 (28.4) 164 (32.3) 0.88 (0.73-1.07) .20Barotraumab 53 (11.2) 47 (9.1) 1.21 (0.83-1.75) .33Refractory hypoxemia 22 (4.6) 52 (10.2) 0.54 (0.34-0.86) .01

Death with refractoryhypoxemia

20 (4.2) 45 (8.9) 0.56 (0.34-0.93) .03

Refractory acidosis 29 (6.1) 42 (8.3) 0.81 (0.51-1.31) .39Death with refractory

acidosis27 (5.7) 38 (7.5) 0.85 (0.51-1.40) .52

Refractory barotrauma 14 (3.0) 12 (2.4) 1.10 (0.54-2.26) .80Death with refractory

barotrauma8 (1.7) 8 (1.6) 1.00 (0.41-2.40) .99

Eligible use of rescue therapiesc 24 (5.1) 47 (9.3) 0.61 (0.38-0.99) .045Total use of rescue therapiesc 37 (7.8) 61 (12.0) 0.68 (0.46-1.00) .05Days of mechanical ventilationd 10 (6-17) 10 (6-16) .92Days of intensive cared 13 (8-23) 13 (9-23) .98Days of hospitalizationd 28 (17-48) 29 (16-51) .96aAll analyses of relative risk are stratified by hospital.bBarotrauma includes study onset of pneumothorax, pneumomediastinum, pneumoperitoneum, subcutaneous emphy-

sema, and chest tubes inserted for spontaneous pneumothorax.cEligible use of rescue therapies refers to use among patients who met a priori criteria. Total use of rescue therapies refers

to use among all patients whether or not they met criteria. Rescue therapies included inhaled nitric oxide, prone venti-lation, high-frequency oscillation, high-frequency jet ventilation, and extracorporeal membrane oxygenation.

dContinuous data are presented as median (interquartile range) among survivors of mechanical ventilation, intensive care,and hospitalization, respectively.

Table 7. Hospital Mortality Based on Severity of Lung Injury at Baseline

PaO2/FIO2

No. (%)Relative Risk

(95% ConfidenceInterval)

PValuea


Quartile 1: 41-106 57 (50) 77 (58) 0.86 (0.68-1.09)Quartile 2: !106-142 46 (39) 55 (43) 0.92 (0.68-1.24)

.94Quartile 3: !142-180 43 (33) 40 (33) 0.99 (0.69-1.41)Quartile 4: !180-250 27 (25) 33 (26) 0.90 (0.58-1.40)Abbreviations: FIO2, fraction of inspired oxygen; PaO2, partial pressure of arterial oxygen.aP value for homogeneity among the quartiles.




Effects of Prone Positioning on Lung Protection inPatients with Acute Respiratory Distress SyndromeRodrigo A. Cornejo1, Juan C. Dıaz2, Eduardo A. Tobar1, Alejandro R. Bruhn3, Cristobal A. Ramos2,Roberto A. Gonzalez1, Claudia A. Repetto1, Carlos M. Romero1, Luis R. Galvez1, Osvaldo Llanos1,Daniel H. Arellano1, Wilson R. Neira1, Gonzalo A. Dıaz1, Anıbal J. Zamorano1, and Gonzalo L. Pereira2

1Unidad de Pacientes Crıticos, Departamento de Medicina, Hospital Clınico Universidad de Chile; 2Departamento de Radiologıa, Hospital ClınicoUniversidad de Chile, Santiago, Chile; and 3Departamento de Medicina Intensiva, Facultad de Medicina, Pontificia Universidad Catolica de Chile,Santiago, Chile

Rationale: Positive end-expiratory pressure (PEEP) and prone posi-tioningmay induce lung recruitment and affect alveolar dynamics inacute respiratory distress syndrome (ARDS). Whether there is inter-dependence between the effects of PEEP and prone positioning onthese variables is unknown.Objectives: To determine the effects of high PEEP and prone posi-tioning on lung recruitment, cyclic recruitment/derecruitment,and tidal hyperinflation and how these effects are influenced bylung recruitability.Methods: Mechanically ventilated patients (VT 6 ml/kg ideal bodyweight) underwent whole-lung computed tomography (CT) duringbreath-holding sessions at airway pressures of 5, 15, and 45 cmH2OandCine-CTs on a fixed thoracic transverse slice at PEEP 5 and 15 cmH2O. CT images were repeated in supine and prone positioning. Arecruitment maneuver at 45 cm H2O was performed before eachPEEP change. Lung recruitability was defined as the difference inpercentage of nonaerated tissue between 5 and 45 cm H2O. Cyclicrecruitment/derecruitment and tidal hyperinflation were deter-mined as tidal changes in percentage of nonaerated and hyperin-flated tissue, respectivelyMeasurements and Main Results: Twenty-four patients with ARDSwere included. Increasing PEEP from 5 to 15 cmH2O decreased non-aerated tissue (501 6 201 to 322 6 132 grams; P , 0.001) and in-creased tidal-hyperinflation (0.416 0.26 to 0.576 0.30%; P¼ 0.004)in supine.Pronepositioning furtherdecreasednonaerated tissue (3226132 to2906141grams;P¼0.028) andreduced tidal hyperinflationobserved at PEEP 15 in supine patients (0.576 0.30 to 0.416 0.22%).Cyclic recruitment/derecruitmentonlydecreasedwhenhighPEEPandpronepositioningwere applied together (4.16 1.9 to2.96 0.9%; P¼0.003), particularly in patients with high lung recruitability.Conclusions: Prone positioning enhances lung recruitment anddecreases alveolar instability and hyperinflation observed at highPEEP in patients with ARDS.

Ventilator-induced lung injury (VILI) seems to play an impor-tant role in patients with acute respiratory distress syndrome

(ARDS) (1). The mechanisms by which mechanical ventilationexerts its detrimental effect are not completely understood, butit appears that hyperinflation of lung units and shear forcesgenerated during cyclic recruitment/derecruitment of unstablealveoli exacerbate, or even initiate, lung injury (1).

High levels of positive end-expiratory pressure (PEEP) andprone positioning have been demonstrated to reduce VILI in ex-perimental models of acute lung injury (ALI) (2–5). However,analysis of several large clinical trials in patients with ALI orARDS suggests that these interventions may be effective onlyin patients with severe ARDS (6–9). Patients who seem to benefitfrom prone positioning are frequently subjected to higher levelsof PEEP. Thus, there may be a potential interaction between theeffects of both interventions on the mechanisms of VILI.

Regarding VILI, PEEP may have a protective effect by fa-voring lung recruitment and by reducing cyclic recruitment/derecruitment (1, 6), but other mechanisms (e.g., redistributionof extravascular lung water, redistribution of pulmonary bloodflow to better aerated units, or preservation of surfactant activ-ity) may be involved. However, PEEP may induce hyperin-flation and increase the risk of VILI, especially in patientswith low recruitability or lobar ARDS (10–12). In fact, somepatients exhibit tidal hyperinflation despite using low VT andmoderate PEEP levels according to the ARDS-Net strategy(13, 14).

Prone positioning may influence mechanisms of VILI. Byrecruiting nonaerated tissue and by reducing the vertical pleuralpressure gradient, prone positioning may provide a more uniformdistribution of transpulmonary pressures during mechanical

(Received in original form July 23, 2012; accepted in final form January 12, 2013)

This work was supported by FONDECYT grant 11070156, Chile.

Author Contributions: Substantial contributions to conception and design: R.A.C.Acquisition of data: R.A.C., J.C.D., C.M.R., R.A.G., C. A. Repetto, D.H.A., W.R.N.,G.A.D., A.J.Z., G.L.P. Analysis and interpretation of data: R.A.C., E.A.T., A.R.B.,C.M.R., L.R.G., O.L. Drafting the article or revising it critically for importantintellectual content: R.A.C., J.C.D., E.A.T., A.R.B., C.M.R., R.A.G., C. A. Repetto,C. A. Ramos, L.R.G., O.L., D.H.A., W.R.N., G.A.D., A.J.Z., G.L.P. Final approval ofthe version to be published: R.A.C., J.C.D., E.A.T., A.R.B., C.M.R., R.A.G., C. A.Repetto, C. A. Ramos, L.R.G., O.L., D.H.A., W.R.N., G.A.D., A.J.Z., G.L.P.

Correspondence and requests for reprints should be addressed to RodrigoCornejo, M.D., Universidad de Chile, Santos Dumont 999, Independencia, San-tiago, Chile. E-mail: [email protected]

This article has an online supplement, which is accessible from this issue’s table ofcontents at www.atsjournals.org

Am J Respir Crit Care Med Vol 188, Iss. 4, pp 440–448, Aug 15, 2013Copyright ª 2013 by the American Thoracic SocietyOriginally Published in Press as DOI: 10.1164/rccm.201207-1279OC on January 24, 2013Internet address: www.atsjournals.org

AT A GLANCE COMMENTARY

Scientific Knowledge on the Subject

Experimental and clinical studies suggest that high levels ofpositive end-expiratory pressure (PEEP) and prone posi-tioning may favor protective mechanical ventilation inpatients with acute respiratory distress syndrome. HighPEEP may induce lung recruitment and decrease cyclicrecruitment/derecruitment; however, increasing PEEP mayincrease hyperinflation. Prone positioning could have syn-ergistic effects with high PEEP by providing a more uniformrecruitment and better distribution of lung stress.

What This Study Adds to the Field

In ventilated patients with acute respiratory distress syn-drome, prone positioning enhances the effects of high PEEPin terms of lung recruitment and reduction of cyclicrecruitment/derecruitment and prevents the negative im-pact of PEEP on tidal hyperinflation.

HIGH PEEP VS LOW PEEP

induced by high PEEP was lower in prone positioning. Thesefindings may be related to the effects of prone positioning in de-creasing pleural pressure gradients and homogenizing transpul-monary pressures in the lungs of patients with ARDS (15, 16, 18,20). A regional analysis of CT images, as performed by Grassoand colleagues (40), would be a valuable complement to thepresent study to assess the effects of prone positioning on inho-mogeneity and recruitment of individual lung regions.

The results of the present study suggest that a high PEEPstrategy applied in prone positioning, instead of supine position-ing, could have more beneficial and less adverse effects in termsof respiratory mechanics and determinants of VILI. These find-ings are consistent with the observations of ametaanalysis, whichindicate that patients with severe forms of ARDS, who are usu-ally ventilated with high levels of PEEP,may have a survival ben-efit when treated in prone positioning (7–9).

Recruitment and Cyclic Recruitment/Derecruitment:Methodological Issues

CT has been the gold standard to assess lung recruitment, al-though different definitions have been applied (10, 41). We chosethe original definition of recruitment based on the decrease ofnonaerated tissue expressed in lung weight as we have used inthe past (22, 42) because it may be applied to analyze cyclicrecruitment/derecruitment in dynamic CT. Other authors havedefined recruitment as the reaeration of the nonaerated andpoorly aerated compartment (10). Because we applied a definitionlimited to the nonaerated compartment, we acknowledge that

our results for PEEP-induced recruitment may appear as subes-timated compared with studies that include the poorly aeratedcompartment, as shown in a recent study using transthoracic ul-trasound (43).

The threshold of 13.9% used in the present study to classifypatients as having high or low lung recruitability in supine posi-tion was predefined arbitrarily based on the median value of thesubgroup of 49 patients with ARDS from Gattinoni’s study. Theoriginal lung recruitability threshold of Gattinoni’s study was9%, which corresponded to the median value of the wholeALI/ARDS population (68 patients). By applying a differentthreshold, our subgroups of higher and lower lung recruitabilityare not comparable to the subgroups defined in the originalstudy by Gattinoni (22).

There is controversy regarding whether cyclic recruitment/derecruitment, as assessed by CT, corresponds to intratidal open-ing and closing of lung units or to flooded alveoli that becomepartially inflated during inspiration (44, 45). Whatever the un-derlying phenomenon (cyclic mechanical deformation or cyclicrecruitment/derecruitment of lung units), a reduction of insta-bility produced by prone positioning at high PEEP, as sup-ported by our study, appears as theoretically positive in termsof lung protection.

Several approaches to assess cyclic phenomena have beenused (46). A cine-CT analysis of a fixed transverse slice allowsdynamic imaging without mechanical ventilation interruption.This method has been recently validated by experimental andclinical studies to determine cyclic recruitment/derecruitment,tidal hyperinflation, and dynamic lung strain (18, 35, 36, 47, 48).

Figure 3. Effects of positive end-expiratory pressure (PEEP)and prone positioning on cyclic recruitment/derecruit-ment (R/D) and tidal hyperinflation (TH). Cyclic R/D andTH with PEEP 5 and 15 cm H2O, assessed in supine (supine5 and supine 15) and prone positioning (prone 5 andprone 15). Data are presented for the overall population(n ¼ 24) (A), for the subgroup of patients with high lungrecruitability (n ¼ 14) (B), and for the subgroup of patientswith low lung recruitability (n ¼ 10) (C). *P , 0.05 be-tween parameters at supine 5 and prone 15 cm H2O. yP,0.05 between parameters at supine 5 and supine 15 cmH2O. zP , 0.05 between parameters at supine 15 andprone 15 cm H2O.

Cornejo, Dıaz, Tobar, et al.: Prone Decreases Instability and Hyperinflation 445

The main limitations of this method are: 1) The more inhomo-geneous the lung impairment, the less representative the slicemay be (this handicap is particularly true in patients with lobarpattern, but only four of our patients had such pattern); 2) It isimpossible for dynamic CT to scan exactly the same anatomicalstructure in different settings, although the careful definition ofanatomical landmarks and fractional analysis used for dynamicCT may avoid artifacts created by the cranio-caudal motion;and 3) the absolute amount of grams of lung tissue undergoingrecruitment/derecruitment or tidal hyperinflation cannot be

determined. The alternative approach of using static CT imagesof the whole lung after end-expiratory and end-inspiratorybreath holds (15, 16, 18–20, 22, 34) also has several limitations,the most important being the time dependency of the recruit-ment and derecruitment phenomena (49, 50). Despite the lim-itations of both methods, a previous study comparing themshowed no major differences (48).

In conclusion, prone positioning induces lung recruitmenteven in patients classified as having low potential for lung recruit-ment. In addition, prone positioning applied together with high

Figure 4. Representative chest CT images obtained duringbreath-holding sessions. Representative CT slices of thelungs obtained 2 cm above the diaphragm dome at airwaypressures of 5 cm H2O (left) and 45 cm H2O (right) fromfour patients in supine (upper) and prone positioning(lower). The percentage of potentially recruitable lungwas defined as the proportion of nonaerated tissue inwhich aeration is restored when increasing airway pres-sures from 5 to 45 cm H2O. In patient A, the percentageof potentially recruitable lung was 24% in supine and 18%in prone positioning; in patient B, the percentage of po-tentially recruitable lung was 24% in supine and 15% inprone; in patient C, the percentage of potentially recruit-able lung was 14% in supine and 21% in prone; and inpatient D, the percentage of potentially recruitable lungwas 27% in supine and 24% in prone.

TABLE 4. COMPARISONS BETWEEN PATIENTS WITH HIGH AND LOW LUNG RECRUITABILITY

High Lung Recruitability Low Lung Recruitability

Supine Prone Supine Prone

PaO2/FIO2

at PEEP 5, mm Hg 128 6 61 156 6 62 164 6 49 191 6 80PaO2

/FIO2at PEEP 15, mm Hg 230 6 69* 272 6 70*† 244 6 70 241 6 49†

PaCO2at PEEP 5, mm Hg 49 6 7 47 6 7 45 6 6 45 6 6

PaCO2at PEEP 15, mm Hg 47 6 8 46 6 8 46 6 6 46 6 5

Compliance at 5 cmH2O, ml/cm H2O 34 6 6 33 6 8 35 6 6 37 6 9Compliance at 15 cmH2O, ml/cm H2O 40 6 7* 45 6 11*† 40 6 12 43 6 10†

Total lung weight at 5 cm H2O, ml 1,307 6 240 1,312 6 292 1,052 6 334‡ 1,083 6 376NAT at 5 cm H2O, g 576 (410–797) 448 (354–646)x 372 (305–472)‡ 325 (181–390)‡

NAT at 15 cm H2O, g 301 (230–400)* 296 (231–329)*† 290 (250–334)* 215 (163–330)*†jj

PAT at 5 cm H2O, g 454 (360–546) 443 (411–561) 387 (281–431) 457 (275–561)PAT at 15 cm H2O, g 486 (422–786) 459 (410–681) 378 (241–572) 407 (263–577)WAT at 5 cm H2O, g 208 (163–347) 255 (191–434)x 240 (210–351) 324 (239–343)x

WAT at 15 cm H2O, g 468 (388–593)* 534 (458–630)*†jj 461 (377–581)* 483 (336–593)*HIT at 5 cm H2O, g 3.6 (2.1–5.4) 2.7 (2.0–4.9) 5.8 (4.7–6.9) 6.0 (4.8–7.9)‡

HIT at 15 cm H2O, g 15 (7.3–18)* 12 (9.0–18)* 17 (14–26)* 15 (10–24)*

Definition of abbreviations: HIT ¼ hyperinflated tissue; NAT ¼ nonaerated tissue; PAT ¼ poorly aerated tissue; PEEP 5 ¼ positiveend-expiratory pressure 5 cm H2O; PEEP 15 ¼ positive end-expiratory pressure 15 cm H2O; WAT ¼ well aerated tissue.* P , 0.05 between parameters at 5 cm H2O (or PEEP 5) and 15 cm H2O (or PEEP 15).y P , 0.05 between parameters at 5 cm H2O (or PEEP 5) in supine positioning and 15 cm H2O (or PEEP 15) in prone

positioning.z P , 0.05 comparing patients with high lung recruitability versus low lung recruitability at the same level of PEEP and position.x P , 0.05 comparing parameters at 5 cm H2O (or PEEP 5) between supine and prone positioning.jj P , 0.05 comparing parameters at 15 cm H2O (or PEEP 15) between supine and prone positioning.

446 AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 188 2013

¡  To date, no clinical protocol has been used in any published trial combining HIGH PEEP and Prone Positioning

¡  Three possible solutions

HIGH PEEP VERSUS LOW PEEP

¡  LHSC stops using HIGH PEEP protocol in favor of “Classic” ARDSnet PEEP table

¡  Use PaO2/FiO2 ≤150 with PEEP ≥10 as targets for “flipping”

¡  Benefits §  Already using this PEEP table for routine ventilation § Maximize evidence-based benefits of Prone by using PROSEVA

protocol

¡  Risks §  Physician and RT education regarding change ventilation strategy §  Theoretical risk of increased refractory hypoxia and death secondary

to refractory hypoxia

SOLUTION 1

¡  Combine HIGH PEEP and PRONING Strategies ¡  Use PaO2/FiO2 ≤150 with PEEP ≥14 as targets for “flipping”

¡  BENEFITS §  Theoretical benefit combining strategies (small study) §  Can follow current LHSC best practice ARDS strategy

¡  RISKS §  Unproven – No direct evidence for combined strategy

SOLUTION 2

¡  Do BOTH!!

¡  Implement one strategy at each site as part of a PILOT study for a future RCT

SOLUTION 3

OUTSTANDING ISSUES

1.  Decision about which ventilation strategy LHSC will follow 2.  Implement protocol

1.  Laminated protocol sheets, HUGO check box 2.  Academic Half-day Fellow Education 3.  RT Education 4.  Nursing Education about protocol existence

3.  Test implementation 1.  Compare pre- and post- LHSC wide versus Site specific comparison

(CCTC vs MSICU) 1.  Outcome measures: Prone time, Sedation requirements, NMB

requirements, Ventilator time

THOUGHTS?? Thank You

Documents

A PROTOCOL FOR - Home - Critical Care Londoncriticalcarelondon.ca/wp-content/uploads/2014/07/12...n engl j med 368;23 nejm.org june 6, 2013 2159 The new england journal of medicine