Mechanical Ventilation Management During Extracorporeal Membrane Oxygenation for Acute Respiratory Distress Syndrome: A Retrospective International Multicenter Study Matthieu Schmidt, MD1,2; Claire Stewart, MBBS (Hons), BSc (Adv)3,4; Michael Bailey, PhD, MSc (Statistics), BSc (Hons)1,5; Ania Nieszkowska, MD2; Joshua Kelly, BBiomedSc/BE1; Lorna Murphy, MBBS3,4; David Pilcher, MRCP, FRACP, FCICM;1,5 D. James Cooper, BMBS, MD1,5; Carlos Scheinkestel, MBBS5; Vincent Pellegrino, MBBS1,5; Paul Forrest, MBChB, FANZCA3,4; Alain Combes, MD, PhD2; Carol Hodgson, PhD, FACP1,5

Department of Epidemiology and Preventive Medicine, Australian and New Zealand Intensive Care Research Centre, School of Public Health, Monash University, Melbourne, VIC, Australia. 2 Medical-Surgical Intensive Care Unit, iCAN, Institute of Cardiometabolism and Nutrition, Hôpital de la Pitié–Salpêtrière, Assistance Publique– Hôpitaux de Paris, Paris, France. 3 Department of Anaesthetics, Royal Prince Alfred Hospital, Sydney, NSW, Australia. 4 Sydney University Medical School, The University of Sydney, Sydney, NSW, Australia. 5 Intensive Care Department, Alfred Hospital, Melbourne, VIC, Australia. Supplemental digital content is available for this article. Direct URL citations appear in the printed text and are provided in the HTML and PDF versions of this article on the journal’s website (http://journals.lww.com/ccmjournal). Dr. Schmidt was supported by the French Intensive Care Society (SRLF), the “Fonds de dotation Recherche en Santé Respiratoire, 2012”, the “Collège des Enseignants de Réanimation Médicale”, and the “Fonds d’Etude et de Recherche du Corps Médical des Hôpitaux de Paris”. Dr. Hodgson was supported by an early career fellowship from the National Medical and Research Council of Australia (2012–2015). Dr. Cooper was supported by a National Medical and Research Council of Australia Practitioner Fellowship (2013–2017). Dr. Pellegrino received support for travel from Maquet (Travel and accommodation provided to attend the 4th International Pediatric Perfusion Symposium [Singapore]). His institution received grant support from the Extracorporeal Life Support Organization (research grant for studies into the predictors of mortality for patients receiving venoarterial and venovenous extracorporeal membrane oxygenation [VV ECMO]) and from the Intensive Care Foundation (research grant for studies into the long term health effects of VV ECMO). Dr. Combes consulted for Maquet. Dr. Hodgson is employed by Monash University and received other support from the National Health and Medical Research Council Fellowship. The remaining authors have disclosed that they do not have any potential conflicts of interest. Address requests for reprints to: Matthieu Schmidt, MD, Department of Epidemiology and Preventive Medicine, The Australian & New Zealand Intensive Care Research Centre, School of Public Health and Preventive Medicine, Level 6. The Alfred Centre, Commercial Road, Melbourne, VIC 3004, Australia. E-mail: [email protected] 1

Copyright © 2014 by the Society of Critical Care Medicine and Lippincott Williams & Wilkins DOI: 10.1097/CCM.0000000000000753

Critical Care Medicine

Objective: To describe mechanical ventilation settings in adult patients treated for an acute respiratory distress syndrome with extracorporeal membrane oxygenation and assess the potential impact of mechanical ventilation settings on ICU mortality. Design: Retrospective observational study. Setting: Three international high-volume extracorporeal membrane oxygenation centers. Patients: A total of 168 patients treated with extracorporeal membrane oxygenation for severe acute respiratory distress syndrome from January 2007 to January 2013. Interventions: We analyzed the association between mechanical ventilation settings (i.e. plateau pressure, tidal volume, and positive end-expiratory pressure) on ICU mortality using multivariable logistic regression model and Cox-proportional hazards model. Measurement and Main Results: We obtained detailed demographic, clinical, daily mechanical ventilation settings and ICU outcome data. One hundred sixty-eight patients (41 ± 14 years old; Pao2/Fio2 67 ± 19 mm Hg) fulfilled our inclusion criteria. Median duration of extracorporeal membrane oxygenation and ICU stay were 10 days (6–18 d) and 28 days (16–42 d), respectively. Lower positive end-expiratory pressure levels and significantly lower plateau pressures during extracorporeal membrane oxygenation were used in the French center than in both Australian centers (23.9 ± 1.4 vs 27.6 ± 3.7 and 27.8 ± 3.6; p < 0.0001). Overall ICU mortality was 29%. Lower positive end-expiratory pressure levels (until day 7) and lower delivered tidal volume after 3 days on extracorporeal membrane oxygenation were associated with significantly higher mortality (p < 0.05). In multivariate analysis, higher positive end-expiratory pressure levels during the first 3 days of extracorporeal membrane oxygenation support were associated with lower mortality (odds ratio, 0.75; 95% CI, 0.64–0.88; p = 0.0006). Other independent predictors of ICU mortality included time between ICU admission and extracorporeal membrane oxywww.ccmjournal.org

1

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Schmidt et al genation initiation, plateau pressure greater than 30 cm H2O before extracorporeal membrane oxygenation initiation, and lactate level on day 3 of extracorporeal membrane oxygenation support. Conclusions: Protective mechanical ventilation strategies were routinely used in high-volume extracorporeal membrane oxygenation centers. However, higher positive end-expiratory pressure levels during the first 3 days on extracorporeal membrane oxygenation support were independently associated with improved survival. Further prospective trials on the optimal mechanical ventilation strategy during extracorporeal membrane oxygenation support are warranted. (Crit Care Med 2014; XX:00–00) Key Words: acute respiratory distress syndrome; adult; extracorporeal membrane oxygenation; mechanical ventilation; outcome

M

echanical ventilation (MV) remains the cornerstone of management in patients with acute respiratory distress syndrome (ARDS) care, and numerous large randomized controlled trials have investigated different ventilation strategies (1–4). Despite recent trials on rescue therapies (5–8), the management of severe ARDS remains challenging. Extracorporeal membrane oxygenation (ECMO) was initially derived from conventional cardiopulmonary bypass circuits used in operating theatres, and previous efforts to expand its application to ARDS were marred by high complication rates (9, 10). Technological advances in centrifugal pumps, polymethylpentene (diffusion) membrane oxygenators, and heparin-bonded circuits have largely overcome these early technical problems (11). In addition, the promising outcomes during the 2009 influenza A (H1N1) pandemic (12) and the conventional ventilatory support versus extracorporeal membrane oxygenation for severe adult respiratory failure trial (13) have renewed interest in venovenous ECMO (VV-ECMO) as a salvage therapy for patients with severe ARDS. By replacing the usual functions of ventilation, ECMO facilitates ultraprotective ventilation strategies, potentially minimizing ventilation-induced lung injury (14). Although there is high level evidence that low-tidal volume (VT) ventilation improves survival (1, 15) in ARDS, ventilation strategies during ECMO have received little attention (16). Ultraprotective ventilation strategies are supported in a few experimental studies (17, 18), but there are no evidence into a clinical benefit of ventilation strategies during ECMO support (19) and therefore little consensus regarding optimal ventilator management. Practices are guided by clinician preference, experience of high-volume centers and local resource availability (16). The purpose of this study was therefore to describe MV settings in adult patients treated for ARDS with ECMO in three international high-volume ECMO centers and assess the potential impact of MV settings on ICU mortality

PATIENTS AND METHODS Setting This study was conducted in three university hospital centers with medical-surgical ICUs: a 45-bed ICU in Melbourne, 2

www.ccmjournal.org

Australia (The Alfred Hospital), a 49-bed ICU in Sydney, Australia (Royal Prince Alfred Hospital), and a 26-bed ICU in Paris, France (La Pitié-Salpêtrière Hospital). Each center supports more than 30 ECMO cases annually (20, 21). In addition, each hospital provides a referral and retrieval service for patients requiring ECMO support in regional centers. Ethics approval was obtained in each participating institution, and the need for informed consent was waived. Patients We retrospectively analyzed the hospital records of 60 consecutive patients per center (n = 180), who received ECMO for refractory ARDS from January 2007 to January 2013 (Melbourne and Sydney) or from January 2011 to January 2013 (Paris). Study patients were included if they received ECMO for more than 24 hours, and data on MV settings were available. Two patients were excluded being enrolled in the Extracorporeal Membrane Oxygenation for Severe Acute Respiratory Distress Syndrome trial (22). One hundred sixty-eight patients were eligible for enrolment in the analysis (Fig. 1). Hospital Characteristics and ECMO Management Detailed center-specific characteristics and criteria for assessing ECMO for ARDS are provided in eTable 1 (Supplemental Digital Content 1, http://links.lww.com/CCM/B140). Percutaneous VV-ECMO was used in all cases. In the Australian centers, the commonest configuration was femoro-femoral (with additional jugular access if required), whereas the French center mostly used a femoro-jugular configuration. Although MV on ECMO was not standardized between study centers, each center used lung protective strategies to minimize plateau pressures. Data Collection Demographic data, Acute Physiology and Chronic Health Evaluation (APACHE II) (23) and Sequential Organ Failure Assessment (SOFA) (24) were collected at the time of ECMO initiation. Date of hospital and ICU admission, date of initiation of MV and ECMO, primary indication for ECMO, preexisting immunodeficiency according to APACHE II criteria (23), and surgical intervention within 7 days before ECMO initiation were also recorded. Adjunctive rescue therapies prior to ECMO support were recorded (nitric oxide, neuromuscular blocker agents, prone positioning, or high-frequency oscillatory ventilation), along with ECMO cannulation configuration and retrieval cases. Once established on ECMO, the following patient variables were collected once daily at 10 am (or nearest available result), from 24 hours before ECMO support until support withdrawal, death, or day 28 (whichever occurred first): 1) level of sedation using the Richmond Agitation–Sedation Scale (RASS) (25); 2) ventilator settings (mode, VT, respiratory rate, positive end-expiratory pressure [PEEP], plateau pressure, driving pressure [plateau pressure—PEEP], Fio2; 3) ECMO settings (blood flow, sweep gas flow, and Fio2); 4) blood gas; 5) hemoglobin; 6) complications (death, cardiac arrest, pneumothorax, major bleeding, and renal replacement therapy); and 7) tracheotomy. Major bleeding event was defined XX 2014 • Volume XX • Number XXX

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Clinical Investigation

(interquartile range). Categorical variables were compared using chi-square test for equal proportion or Fisher exact tests and have been reported as numbers (percentages). Changes over time were analyzed using repeated-measures analysis of variance fitting main effects for group, time, and an interaction between group and time with results presented as least square means (95% CI). Multivariate logistic regression was used to identify early independent risk factors for ICU mortality in patients receiving ECMO, with results presented as odds ratios (95% CI). Multivariate models were constructed using both stepwise selection and backward elimination techniques before undergoing a final assessment for clinical and biological plausibility. A cox proportional hazard modeling Figure 1. Flow chart of the study. ARDS = acute respiratory distress syndrome, ECMO = extracorporeal membrane oxygenation, EOLIA = Extracorporeal Membrane Oxygenation for Severe Acute Respiratory Distress was also applied to data analySyndrome, MV = mechanical ventilation. sis. Multivariate modeling was done including patient inforas bleeding, which lead to a hemostatic treatment (pleural mation until day 3. Thus, three patients with ICU length of drainage, surgical procedure, arterial embolization, and gasstay (LOS) less than 3 days (i.e. 48 hr) were not included in trointestinal endoscopy), a transfusion greater than or equal the model. With the exception of the “country “(i.e. Australia to five blood units of packed RBCs during 24 hours, in cases of or France)” (automatically forced into the model), all other intracerebral hemorrhage or death. Maximal ionotrope dose variables before day 3 with a univariate value of p less than per day was calculated as the addition of the daily maximal 0.10 were considered for model inclusion. Possible interactions epinephrine and maximal norepinephrine doses, respectively. between independent risk factors (including “country”) were Status at ICU discharge and duration of MV and ECMO sup- tested by including proper cross-product terms in the regresport were recorded. sion models, and likelihood ratio tests comparing models with and without the interaction term were used to estimate the sigOutcome Measures nificance of the interaction. The primary endpoint was to examine the management of Analysis were performed using SAS version 9.3 (SAS MV during ECMO, including mode, VT, plateau pressure, and Institute, Cary, NC), and a two-sided value of p equal to 0.05 PEEP. These data were collated for six intervals: “pre-ECMO,” was considered to be statistically significant. “day 1–3,” “day 4–7,” “day 8–14,” “day 15–21,” and “day 22–28.” Secondary endpoints were to examine whether any MV variables either before or during early ECMO support were associ- RESULTS ated with ICU mortality. Characteristics of Included Patients A total of 168 severe patients with ARDS (age, 41 ± 14 years; Statistical Analyses 62% men) were enrolled in the study, of whom 98% received Data were initially assessed for normality. Normally distributed VV-ECMO (Fig. 1). Details of patient characteristics are procontinuous variables were compared using Student t tests or vided in Table 1. Bacterial pneumonia and influenza infection analysis of variance and presented as mean±sd, whereas nonwere the main risk factors for ARDS requiring ECMO support. normally distributed variables were compared using Wilcoxon Median time to initiation of ECMO was 2 days (1–6 d) after rank sum tests or Kruskal–Wallis tests and presented as median commencement of MV. Of note, French patients were older, Critical Care Medicine

www.ccmjournal.org

3

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Schmidt et al

Table 1. Demographics Characteristics and Mechanical Ventilation Data According to the Center All Patients (n = 168)

Melbourne, Australia (n = 52)

Paris, France (n = 57)

Sydney, Australia (n = 59)

pa

Age (yr)

41 ± 14

37 ± 14

46 ± 12

40 ± 13

0.002

Men

105 (62)

31 (60)

43 (75)

31 (52)

0.03

Body mass index (kg/m2)

28.6 ± 8.4

28.6 ± 9.2

28.5 ± 7.2

28.8 ± 8.9

0.99

Immunocompromisedb

29 (17)

7 (13)

11 (19)

11 (19)

0.68

Acute Physiology and Chronic Health Evaluation II

20 ± 8

19 ± 7

25 ± 8

16 ± 6

< 0.0001

Sequential Organ Failure Assessment score at ECMO instauration

10.9 ± 3.3

10.8 ± 2.9

11.0 ± 3.7

9.9 ± 2.9

0.007

Influenza

41 (24)

14 (27)

7 (12)

20 (34)

0.02

Bacterial pneumonia

92 (55)

21 (40)

37 (65)

34 (58)

0.03

Patients retrieval

120 (71)

37 (71)

42 (74)

41 (69)

0.88

Interval ICU-ECMO (d)

2 (1–7)

1 (0–3.5)

5 (1–9)

2 (1–7)

0.0002

Interval MV-ECMO (d)

2 (1–6)

1 (0–3.5)

4 (1–8)

2 (1–5)

0.004

   Prone positioning

44 (26)

1 (2)

31 (54)

12 (20)

< 0.0001

   Nitric oxide

74 (44)

21 (40)

48 (84)

5 (8)

< 0.0001

   Neuromuscular blockers

94 (56)

27 (52)

52 (91)

15 (25)

0.002

6 (4)

3 (6)

0 (0)

3 (5)

0.04

   Pao2/ Fio2 (mm Hg)

67 ± 19

75 ± 20

61 ± 15

66 ± 20

0.002

  PEEP (cm H2O)

13 ± 4

14 ± 3

10 ± 4

15 ± 4

< 0.0001

  Tidal volume (mL/kg)

6.3 ± 1.5

6.3 ± 1.5

6.4 ± 1.3

6.3 ± 1.7

0.94

  Plateau pressure (cm H2O)

32.2 ± 4.7

31.6 ± 3.7

32.3 ± 5.9

32.8 ± 4.2

0.48

  Driving pressure (cm H2O)c

19.0 ± 5.8

17.1 ± 4.1

21.9 ± 6.7

17.9 ± 5.3

0.0001

  Compliance (mL/cm H2O)

23.2 ± 9.9

25.0 ± 9.0

20.3 ± 10.2

24.4 ± 9.9

0.04

66 ± 32

60 ± 21

61 ± 22

76 ± 46

0.03

–4.3 ± 0.9

–4.7 ± 0.4

–3.4 ± 1

–4.7 ± 0.5

< 0.0001

12 ± 3

14 ± 3

10 ± 2

12 ± 3

< 0.0001

  Tidal volume (mL/kg)

3.9 ± 1.6

4.4 ± 1.6

3.5 ± 1.7

3.7 ± 1.5

0.008

  Plateau pressure(cm H2O)

26.4 ± 3.5

27.6 ± 3.7

23.9 ± 1.4

27.8 ± 3.6

< 0.0001

13.7 (12.0–15.3)

13.4 (11.4–15.6)

13.0 (12.0– 15.0)

4.5 ± 0.9

4.7 ± 0.9

5.0 ± 0.7

4.0 ± 0.9

< 0.0001

1.8 (1.4–2.3)

1.7 (1.3–2.7)

1.9 (1.6–2.3)

1.5 (1.3–2.2)

0.02

Before ECMO  Rescue therapy

   High-frequency oscillation ventilation  Ventilation parameters

   PaCo2 (mm Hg) From day 1 to day 3 on ECMO   Richmond Agitation Sedation Scale   PEEP (cm H2O)

  Driving pressure (cm H2O)   ECMO flow (L/min)    Lactate (mmol/L)

14.3 (12.0–18.7)

0.006

(Continued) 4

www.ccmjournal.org

XX 2014 • Volume XX • Number XXX

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Clinical Investigation

(Continued). Demographics Characteristics and Mechanical Ventilation Data According to the Center

Table 1.

Melbourne, Australia (n = 52)

Paris, France (n = 57)

Sydney, Australia (n = 59)

pa

12 (7)

2 (4)

0 (0)

10 (17)

0.001

Tracheotomy on ECMO

30 (22)

2 (4)

26 (46)

9 (15)

0.0003

Renal replacement therapy during ICU stay

75 (45)

35 (67)

23 (41)

17 (29)

0.0002

 ECMO

10 (6–18)

10 (7–16)

10 (6–29)

9 (6–14)

0.30

 Mechanical ventilation

23 (13–37)

20 (14–33)

28 (15–41)

22 (10–34)

0.12

28 (16–42)

21 (15–35)

33 (20–62)

27 (15–37)

0.03

34 (23–54)

30 (19–52)

39 (26–70)

34 (26–47)

0.21

44 (28)

17 (33)

14 (25)

13 (25)

0.59

ECMO-free days at day 28

11.6 ± 9.9

12.6 ± 9.7

9.1 ± 10.2

13.0 ± 9.4

0.07

MV-free days at day 60

23.1 ± 20.3

26.8 ± 20.5

18.1 ± 20.4

24.8 ± 19.4

0.06

48 (29)

13 (25)

20 (35)

15 (25)

0.41

All Patients (n = 168)

Pneumothorax (day 1–7)

Duration or length of stay (d)

 ICU  Hospital Major bleeding events

d

ICU mortality

ECMO = extracorporeal membrane oxygenation, MV = mechanical ventilation, PEEP = positive end-expiratory pressure. a Overall difference between the three centers. b Immunocompromised status included hematological malignancies, solid tumors, solid-organ transplantation, high-dose or long-term corticosteroid and/ or immunosuppressant use, or human immunodeficiency virus infection. c Driving pressure is calculated as plateau pressure—positive end-expiratory pressure. d Major bleeding event was defined as bleeding, which lead to a hemostatic treatment (pleural drainage, surgical procedure, arterial embolization, and gastrointestinal endoscopy), a transfusion ≥ 5 blood units of packed red blood cells during 24 hours, in cases of intracerebral hemorrhage or death.

had higher severity scores at ICU admission and ECMO cannulation, and had less influenza as the reason for ARDS (Table 1). They also had a longer period of MV prior to initiation of ECMO, while they received more rescue therapies (p < 0.05). MV data in the French patients revealed higher median driving pressures and lower median compliance before ECMO initiation in more severe patients with mainly bacterial pneumonia (i.e. less patients with influenza). Clinician-determined pre-ECMO VT and subsequent pre-ECMO plateau pressures were similar within the three ICUs; however, PEEP before ECMO was lower in Paris (p < 0.0001; Table 1). MV Management During ECMO This study reported daily MV management during a total of 2514 ECMO days. Fifty-five percent of patients received pressure targeted modes (i.e. airway press release ventilation/bilevel positive airway pressure/synchronized intermittent mechanical ventilation/pressure control and pressure-­ control ventilation [PCV]) before ECMO initiation, whereas it increased up to 90% and 85% at day 1 and day 7 during ECMO (eFig. 1, Supplemental Digital Content 1, http://links.lww.com/CCM/B140). ECMO initiation facilitated significant reduction in the proportion of patients with a plateau pressure greater than 30 cm H2O, falling from 58% to 13% by ECMO day 3 (Fig. 2A). Once established on ECMO support, a median plateau pressure of 26 cm H2O was maintained throughout the ICU course in both survivors and nonsurvivors (Fig. 2B). Similarly, VT was Critical Care Medicine

reduced from 6.3 ± 1.5 to 3.9 ± 1.6 after ECMO initiation (Fig. 3). After 4 days on ECMO, ICU survivors exhibited significantly higher VT than nonsurvivors, despite both receiving targets of VT less than 6 mL/kg with plateau pressure less than 28 cm H2O. VT in 29% patients was less than 2 and greater than 6 mL/kg in 15% (Fig. 3A). Consistent with previously described protective strategies, almost 80% of patients were ventilated with a PEEP greater than or equal to 10 cm H2O; however, patients who died in ICU received significant lower PEEP during the first week of ECMO support. Of note, a lower PEEP from day 1 to day 7 was associated with a higher prevalence of pneumothorax preECMO or during the first ECMO week (p = 0.047). MV management during the first 3 days on ECMO varied between centers. Lower PEEP (10 ± 2 vs 14 ± 3 and 12 ± 3; p < 0.0001) and lower plateau pressures (23.9 ± 1.4 vs 27.6 ± 3.7 and 27.8 ± 3.6; p < 0.0001) were used in the French center than in both Australian centers. In addition, patients in Australian sites received more sedation, as suggested by a significantly more negative RASS during the first days of the ECMO course (Table 1). Outcomes Nearly half of the patients received renal replacement therapy during their ICU stay and 22% underwent tracheotomy on ECMO. Duration of ECMO and MV was 10 days (6–18 d) and 23 days (13–37 d), respectively, with no significant difference between centers. However, we observed a longer ICU www.ccmjournal.org

5

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Schmidt et al

nonsurvivors were more likely men, immunocompromised, with lower body mass index and SOFA scores at ECMO instauration and longer intervals between MV and ECMO initiation. Lower PEEP and plateau pressure greater than 30 cm H2O prior to ECMO were also associated with increased ICU mortality. PEEP during the first week of ECMO support was also higher in ICU survivors, whereas plateau pressure was the same for both groups (Figs. 3B and 4B). After adjusting for other significant confounding variables using logistic regression analysis, independent predictors of death in ICU included lower PEEP during the first 3 days on ECMO support, duration between ICU admission and ECMO initiation, plateau pressure greater than 30 cm H2O before ECMO and elevated lactate on ECMO day 3 (Table 3). Consistently, a higher PEEP for MV while on ECMO was independently associated with a longer “time to death in ICU” using Cox proportional hazards modeling (Table 3). In addition, higher PEEP was still independently associated with reduced death in ICU when the multivariate logistic regression was performed on the Australian population only (eTable 3, Supplemental Digital Content 1, http:// Figure 2. A, Mean plateau pressure targeted during the extra-corporeal membrane oxygenation (ECMO) links.lww.com/CCM/B140). course and (B) mean plateau pressure according to time and ICU outcome. *A value of p less than 0.05 with ICU death. Results are presented as least square means (95% CI). No significant interactions between “country,” “interval LOS in Paris, whereas there was no difference between the between ICU admission and ECMO initiation,” “plateau prescenters for ECMO-free days at day 28 and MV-free days at day sure before ECMO less than 30 cm H2O,” “PEEP from day 1 to 3 on ECMO,” and “lactate at day 3” were found. 60 (Table 1). Analysis of Mortality in ICU ICU mortality was not significantly different within all centers with an overall rate of 48 of 168 death (29%) in ICU. Univariate analysis of factors associated with ICU mortality is provided in Table 2 and in eTable 2 (Supplemental Digital Content 1, http://links.lww.com/CCM/B140). Briefly, 6

www.ccmjournal.org

DISCUSSION Although previous studies have examined outcomes of patients with ARDS receiving ECMO support (12, 26, 27), there are little data on MV during ECMO support (28). Although ECMO support facilitates the provision of “ultraprotective” MV, the ideal MV strategy on ECMO is unknown. XX 2014 • Volume XX • Number XXX

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Clinical Investigation

we found that PCV was used in nearly all cases during the first week of ECMO support. Limiting pressure during MV provided lung protection by minimizing barotrauma in both assist-volume control ventilation and PCV, whereas recovery of lung compliance is easily monitored in PCV by increasing VT for the set pressure (16). Our finding that elevated MV plateau prior to ECMO was associated with increased mortality is consistent with a reported relationship between decreased lung compliance and mortality (27). Higher preECMO plateau pressure was associated with shorter time to death. We observed a major reduction in plateau pressure (< 30 cm H2O) after ECMO initiation in all centers (Fig. 2B). Limiting MV plateau pressure after ECMO was established by decreasing VT on day 1 to 3 to less than 4  mL/kg ideal body weight (Fig. 3B), which is consistent with “ultraprotective ventilation” (13, 30). Although animal and human experimental studies report less pulmonary edema and lung injury with VT below 4 mL/ kg (17, 31), a recent multicenter randomized controlled trial failed to demonstrate 28 and 60 days ventilator-free Figure 3. A, Mean tidal volume set during the extra-corporeal membrane oxygenation (ECMO) course and days benefit from ultra ver(B) mean tidal volume according to time and ICU outcome. *A value of p less than 0.05 with ICU death. Results sus conventionally protective are presented as least square means (95% CI). MV (19). However, it is worth noting that a post hoc analysis demonstrated that survivWe found that in three experienced ECMO centers, MV settings during ECMO varied significantly. We also found that ing patients with greater hypoxemia, who would likely be higher plateau pressures prior to ECMO and lower PEEP levels treated with ECMO, were treated with Co2 removal devices and had a significantly shorter period of ventilation comduring the first 3 days on ECMO support were independently pared to control (19). associated with increased ICU mortality. The importance of PEEP in ARDS is well described, but its MV Settings and Outcomes use remains controversial (3, 4, 32). The Extracorporeal Life Assist-volume control ventilation is the most common Support Organization recommends a PEEP of 10 cm H2O durMV mode worldwide for patients with ARDS (29). Less ing ECMO support (28). In our study, the median PEEP during than 40% of patients with conventionally managed ARDS the first 3 days of ECMO support was 12 ± 3 cm H2O; however, we receive PCV in large observational studies (29). By contrast, noted significant differences in PEEP values between countries. Critical Care Medicine

www.ccmjournal.org

7

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Schmidt et al

Table 2.

Univariate Analysis of Factors Associated With ICU Mortality Status at ICU Discharge

Characteristics

Age (yr) Men

Alive (n = 120)

Dead (n = 48)

p

40 ± 13

44 ± 14

0.11

69 (58)

Body mass index (kg/m )

29.5 ± 9.0

2

Immunocompromised

16 (13)

a

36 (75) 26.3 ± 6.2 13 (27)

0.03 0.02 0.03

APACHE II

19.6 ± 7.9

20.9 ± 7.9

0.34

APACHE II (age removed)

18.6 ± 7.7

19.5 ± 7.5

0.50

APACHE II (pH removed)

17.9 ± 7.7

18.7 ± 7.8

0.58

SOFA score at ECMO initiation

10.6 ± 3.0

11.6 ± 3.7

0.05

6.6 ± 3.0

7.8 ± 3.6

0.02

Nonrespiratory SOFA score at ECMO instauration Influenza

36 (30)

5 (10)

Bacterial pneumonia

59 (49)

33 (69)

0.008 0.02

Duration of mechanical ventilation prior to ECMO initiation (d)

2 (0–6)

4 (1–10)

0.006

Duration between ICU admission and ECMO initiation (d)

2 (0–5)

3 (1–7.5)

0.07

Before ECMO  Pneumothorax  Any rescue therapy  Pao2/Fio2 (mm Hg)

9 (8)

10 (21)

0.01

80 (67)

28 (58)

0.31

67 ± 20

65 ± 15

0.57

 PEEP (cm H2O)

13.6 ± 4.0

11.9 ± 4.9

0.04

 PEEP < 8 cm H2O

7 (6)

7 (15)

0.06

 8 < PEEP < 10 cm H2O

3 (3)

7 (15)

0.003

51 (43)

31 (65)

0.01

 Plateau pressure > 30 cm H2O  Driving pressure (cm H2O)

17.5 ± 5.2

22.8 ± 5.7

< 0.0001

 Compliance (mL/cm H2O)

23.3 (17.8–30.7)

16.7 (14.1–23.5)

0.003

 pH

7.28 (7.18–7.37)

7.17 (7.12–7.29)

0.005

b

 PaCo2 (mm Hg)

56 (45–71)

65 (58–82)

0.03

 Maximal inotrope dose (mg/hr)

0.5 (0–1.5)

1.2 (0.2–2.1)

0.02

ΔPEEP before ECMO: day 1

1.2 (3.87)

c

1.21 (4.67)

0.99

From day 1 to 3 on ECMO  PEEP (cm H2O)

12.7 ± 2.9

11.0 ± 2.7

0.0005

  10 < PEEP < 12 cm H2O

35 (29)

21 (44)

0.07

  PEEP > 12 cm H2O

67 (56)

14 (29)

0.002

 pH day 1

7.43 (7.37–7.48)

7.38 (7.34–7.47)

0.06

 pH day 2

7.42 ± 0.07

7.40 ± 0.10

0.07

 Tidal volume (mL/kg)

4 (2.8–5.1)

3.3 (2.3–4.6)

0.09

 Plateau pressure (cm H2O)

26.5 ± 3.4

26.3 ± 3.8

0.73

 Driving pressure (cm H2O)

13.0 (11.7–15.0)

15.0 (12.7–17.0)

0.008

 Compliance (mL/cm H2O)

9.45 (6.5–12.8)

7.95 (5.15–11.6)

0.07

b

(Continued) 8

www.ccmjournal.org

XX 2014 • Volume XX • Number XXX

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Clinical Investigation

Table 2.

(Continued). Univariate Analysis of Factors Associated With ICU Mortality Status at ICU Discharge p

Characteristics

Alive (n = 120)

Dead (n = 48)

 ECMO sweep gas flow (L/min)

5.3 (4.2–7.3)

6.5 (5.7–7.7)

0.01

 Hemoglobin (g/dL)

9.1 (8.3–10.0)

8.6 (8.0–9.5)

0.04

–4.3 ± 0.9

–4.2 ± 0.9

0.67

 Lactate at day 1 (mmol/L)

1.9 (1.3–2.6)

2.5 (1.7–2.9)

0.0002

 Lactate at day 3 (mmol/L)

1.4 (1.1–1.8)

1.8 (1.4–2.1)

0.0002

–0.3 (–1 to 0.1)

–0.7 (–1.5 to 0.3)

0.36

 Richmond Agitation Sedation Scale score

  ΔLactate day 3 to day 1 (mmol/L)  Maximal inotrope dose (mg/hr)

c

  At day 1

0.5 (0.1–1)

0.9 (0.2–1.4)

0.02

  At day 2

0.2 (0–0.6)

0.5 (0.05–1.15)

0.03

  At day 3

0.1 (0–0.4)

0.35 (0–1.1)

0.005

APACHE II = Acute Physiology and Chronic Health Evaluation, SOFA = Sequential Organ Failure Assessment, ECMO = extracorporeal membrane oxygenation, PEEP = positive end-expiratory pressure. a Immunocompromised status included hematological malignancies, solid tumors, solid-organ transplantation, high-dose or long-term corticosteroid and/or immunosuppressant use, or human immunodeficiency virus infection. b Driving pressure is calculated as plateau pressure—positive end-expiratory pressure. c Maximal ionotrope dose per day was calculated as the addition of the daily maximal epinephrine and maximal norepinephrine doses, respectively.

Table 3. Multivariate Logistic Regression and Cox Proportional Hazard Modeling With “ICU Death” as Outcome ICU Death

Time to ICU Death

OR (95% CI)

p

Hazard Ratio (95% CI)

p

Country (France vs Australia)

0.56 (0.22–1.42)

0.56

0.39 (0.19–0.81)

0.01

Duration between ICU admission and ECMO initiation (d)

1.15 (1.06–1.26)

0.001

1.02 (0.97–1.07)

0.56

Plateau pressure before ECMO > 30 cm H2O

5.18 (1.88–14.31)

0.02

3.31 (1.53–7.15)

0.002

Mean positive end-expiratory pressure from day 1 to 3 on ECMO

0.75 (0.64–0.88)

0.0006

0.78 (0.69–0.88)

< 0.0001

Lactate at day 3 (log transformed)

4.77 (2.12–10.73)

0.0002

3.64 (2.24–5.92)

< 0.0001

Variables

OR = odds ratio, ECMO = extracorporeal membrane oxygenation.

In addition, it is worth noting that a higher prevalence of pneumothorax may have led to lower PEEP levels in our population. Our observation of an independent association between early high PEEP level and reduced mortality in ECMO patients, regardless of management’s differences within countries, merits further investigation. Potential benefits of PEEP include reduced atelectasis (33, 34) and improved ventilation/perfusion matching, especially when VT is less than 4 mL/kg (33). In the cohort of patient with a very low VT (< 2 mL/kg; Fig. 3A), higher PEEP might also maintain convective ventilation to prevent atelectasis induced by alveolar nitrogen accumulation during apneic oxygenation on ECMO support (34) although it may also cause alveolar overdistension (35) and hemodynamic impairment. It might be worth considering the potential benefit of recruitment maneuvers in those patients (36, 37), especially at the early stage of the ECMO course. However, the combination of Critical Care Medicine

these two rescue therapies has not been investigated, and physiological studies are needed to support this hypothesis. Finally, the association between higher VT after day 4 (but maintained < 6 mL/kg with a constant plateau pressure) and ICU survival may support the great impact of the respiratory compliance on outcome in these patients. Other Prognostic Factors Several studies (27, 38–40) have demonstrated an adverse relationship between duration of MV prior to ECMO initiation and mortality, especially after 7 days (27, 41). In our study, almost all patients received ECMO within 7 days of MV although we were still able to identify duration between ICU admission and ECMO as a risk factor for mortality. Further trials are therefore needed on the use www.ccmjournal.org

9

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Schmidt et al

independent risk factors of the model, MV differences between France and Australia cannot explain the shorter time to ICU death in Australia. Some unknown confounding factors might have contributed to this result.

Figure 4. A, Mean positive end-expiratory pressure (PEEP) set during the extra-corporeal membrane oxygenation (ECMO) course and (B) mean PEEP according to time and ICU outcome. *A value of p less than 0.05 with ICU death. Results are presented as least square means (95% CI).

and optimal timing of ECMO support in ARDS. Moreover, it raises compelling questions about the value and timing of others rescues therapies before cannulation. Ongoing trials may answer this concern soon (22). It is worth noting that APACHE II was not retained in our multivariate models. Considering that this severity score is calculated at ICU admission, its performance may not likely be as good as the SOFA score, which is performed at ECMO initiation. Finally, in the absence of significant interaction between 10

www.ccmjournal.org

Strengths and Limitations Our study is the first detailed examination of MV settings and outcome in a large cohort of patients with ARDS from high-volume ECMO centers. Our study has several limitations, including its retrospective design. In addition, we cannot exclude that the result might have been biased by residual confounding not accounted for in this study. For logistic reasons, MV settings were collected as point data once a day. We are aware that MV is a continuous, dynamic function of our daily practice, and settings may have changed rapidly with a 24-hour period, especially during the start of the ECMO course. Data from only the first 3 days of ECMO support were included for analysis of mortality, because beyond this time, the size of our cohort was reduced to a number that prevented accurate multivariate analysis. Whether specific MV strategies after day 3 on ECMO would change patient outcomes is yet to be determined, and larger prospective study may overcome this difficulty.

CONCLUSIONS We found that early higher PEEP with MV during ECMO support was associated with decreased mortality. Further large, prospective trials on MV management during ECMO are warranted.

REFERENCES

  1. The Acute Respiratory Distress Syndrome Network: Ventilation with lower tidal volumes as compared with traditional tidal volumes for acute lung XX 2014 • Volume XX • Number XXX

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited

Clinical Investigation injury and the acute respiratory distress syndrome. The Acute Respiratory Distress Syndrome Network. N Engl J Med 2000; 342:1301–1308   2. Brower RG, Lanken PN, MacIntyre N, et al; National Heart, Lung, and Blood Institute ARDS Clinical Trials Network: Higher versus lower positive end-expiratory pressures in patients with the acute respiratory distress syndrome. N Engl J Med 2004; 351:327–336   3. Meade MO, Cook DJ, Guyatt GH, et al; Lung Open Ventilation Study Investigators: Ventilation strategy using low tidal volumes, recruitment maneuvers, and high positive end-expiratory pressure for acute lung injury and acute respiratory distress syndrome: A randomized controlled trial. JAMA 2008; 299:637–645 4. Mercat A, Richard JC, Vielle B, et al; Expiratory Pressure (Express) Study Group: Positive end-expiratory pressure setting in adults with acute lung injury and acute respiratory distress syndrome: A randomized controlled trial. JAMA 2008; 299:646–655 5. Ferguson ND, Cook DJ, Guyatt GH, et al; OSCILLATE Trial Investigators; Canadian Critical Care Trials Group: High-frequency oscillation in early acute respiratory distress syndrome. N Engl J Med 2013; 368:795–805 6. Guérin C, Reignier J, Richard JC, et al; PROSEVA Study Group: Prone positioning in severe acute respiratory distress syndrome. N Engl J Med 2013; 368:2159–2168 7. Papazian L, Forel JM, Gacouin A, et al; ACURASYS Study Investigators: Neuromuscular blockers in early acute respiratory distress syndrome. N Engl J Med 2010; 363:1107–1116 8. Taylor RW, Zimmerman JL, Dellinger RP, et al; Inhaled Nitric Oxide in ARDS Study Group: Low-dose inhaled nitric oxide in patients with acute lung injury: A randomized controlled trial. JAMA 2004; 291:1603–1609 9. Zapol WM, Snider MT, Hill JD, et al: Extracorporeal membrane oxygenation in severe acute respiratory failure. A randomized prospective study. JAMA 1979; 242:2193–2196 10. Morris AH, Wallace CJ, Menlove RL, et al: Randomized clinical trial of pressure-controlled inverse ratio ventilation and extracorporeal CO2 removal for adult respiratory distress syndrome. Am J Respir Crit Care Med 1994; 149:295–305 11. MacLaren G, Combes A, Bartlett RH: Contemporary extracorporeal membrane oxygenation for adult respiratory failure: Life support in the new era. Intensive Care Med 2012; 38:210–220 12. Davies A, Jones D, Bailey M, et al. Extracorporeal membrane oxygenation for 2009 Influenza A(H1N1) acute respiratory distress syndrome. JAMA 2009;302:1888–1895 13. Peek GJ, Mugford M, Tiruvoipati R, et al; CESAR Trial Collaboration: Efficacy and economic assessment of conventional ventilatory support versus extracorporeal membrane oxygenation for severe adult respiratory failure (CESAR): A multicentre randomised controlled trial. Lancet 2009; 374:1351–1363 14. Brodie D, Bacchetta M: Extracorporeal membrane oxygenation for ARDS in adults. N Engl J Med 2011; 365:1905–1914 15. Gajic O, Dara SI, Mendez JL, et al: Ventilator-associated lung injury in patients without acute lung injury at the onset of mechanical ventilation. Crit Care Med 2004; 32:1817–1824 16. Schmidt M, Pellegrino V, Combes A, et al: Mechanical ventilation during extracorporeal membrane oxygenation. Crit Care 2014; 18:203 17. Terragni PP, Del Sorbo L, Mascia L, et al: Tidal volume lower than 6 ml/kg enhances lung protection: Role of extracorporeal carbon dioxide removal. Anesthesiology 2009; 111:826–835 18. Terragni PP, Rosboch G, Tealdi A, et al: Tidal hyperinflation during low tidal volume ventilation in acute respiratory distress syndrome. Am J Respir Crit Care Med 2007; 175:160–166 19. Bein T, Weber-Carstens S, Goldmann A, et al: Lower tidal volume strategy (≈3 ml/kg) combined with extracorporeal CO2 removal versus ‘conventional’ protective ventilation (6 ml/kg) in severe ARDS: The prospective randomized Xtravent-study. Intensive Care Med 2013; 39:847–856 20. Freeman CL, Bennett TD, Casper TC, et al: Pediatric and neonatal extracorporeal membrane oxygenation: Does center volume impact mortality?*. Crit Care Med 2014; 42:512–519 21. Karamlou T, Vafaeezadeh M, Parrish AM, et al: Increased extracorporeal membrane oxygenation center case volume is associated with improved

Critical Care Medicine

extracorporeal membrane oxygenation survival among pediatric patients. J Thorac Cardiovasc Surg 2013; 145:470–475 22. EOLIA-TRIAL. Available at http://www.clinicaltrials.gov/ct2/show/ NCT01470703?term=EOLIA+ECMO&rank=1. 23. Knaus WA, Draper EA, Wagner DP, et al: APACHE II: A severity of disease classification system. Crit Care Med 1985; 13:818–829 24. Vincent JL, Moreno R, Takala J, et al: The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med 1996; 22:707–710 25. Sessler CN, Gosnell MS, Grap MJ, et al: The Richmond AgitationSedation Scale: Validity and reliability in adult intensive care unit patients. Am J Respir Crit Care Med 2002; 166:1338–1344 26. Pham T, Combes A, Rozé H, et al; REVA Research Network: Extracorporeal membrane oxygenation for pandemic influenza A(H1N1)-induced acute respiratory distress syndrome: A cohort study and propensity-matched analysis. Am J Respir Crit Care Med 2013; 187:276–285 27. Schmidt M, Zogheib E, Rozé H, et al: The PRESERVE mortality risk score and analysis of long-term outcomes after extracorporeal membrane oxygenation for severe acute respiratory distress syndrome. Intensive Care Med 2013; 39:1704–1713 28. ELSO Guidelines for Cardiopulmonary Extracorporeal Life Support and Patient Specific Supplements to the ELSO General Guidelines, Ann Arbor, MI. April 2009 29. Esteban A, Anzueto A, Frutos F, et al; Mechanical Ventilation International Study Group: Characteristics and outcomes in adult patients receiving mechanical ventilation: A 28-day international study. JAMA 2002; 287:345–355 30. Gattinoni L, Pesenti A, Mascheroni D, et al: Low-frequency positivepressure ventilation with extracorporeal CO2 removal in severe acute respiratory failure. JAMA 1986; 256:881–886 31. Frank JA, Gutierrez JA, Jones KD, et al: Low tidal volume reduces epithelial and endothelial injury in acid-injured rat lungs. Am J Respir Crit Care Med 2002; 165:242–249 32. Santa Cruz R, Rojas JI, Nervi R, et al: High versus low positive endexpiratory pressure (PEEP) levels for mechanically ventilated adult patients with acute lung injury and acute respiratory distress syndrome. Cochrane Database Syst Rev 2013; 6:CD009098 33. Dembinski R, Hochhausen N, Terbeck S, et al: Pumpless extracorporeal lung assist for protective mechanical ventilation in experimental lung injury. Crit Care Med 2007; 35:2359–2366 34. Nielsen ND, Kjaergaard B, Koefoed-Nielsen J, et al: Apneic oxygenation combined with extracorporeal arteriovenous carbon dioxide removal provides sufficient gas exchange in experimental lung injury. ASAIO J 2008; 54:401–405 35. Grasso S, Stripoli T, Sacchi M, et al: Inhomogeneity of lung parenchyma during the open lung strategy: A computed tomography scan study. Am J Respir Crit Care Med 2009; 180:415–423 36. Fan E, Wilcox ME, Brower RG, et al: Recruitment maneuvers for acute lung injury: A systematic review. Am J Respir Crit Care Med 2008; 178:1156–1163 37. Hodgson C, Keating JL, Holland AE, et al. Recruitment manoeuvres for adults with acute lung injury receiving mechanical ventilation. Cochrane Database Syst Rev 2009(2):CD006667 38. Beiderlinden M, Eikermann M, Boes T, et al: Treatment of severe acute respiratory distress syndrome: Role of extracorporeal gas exchange. Intensive Care Med 2006; 32:1627–1631 39. Brogan TV, Thiagarajan RR, Rycus PT, et al: Extracorporeal membrane oxygenation in adults with severe respiratory failure: A multicenter database. Intensive Care Med 2009; 35:2105–2114 40. Hemmila MR, Rowe SA, Boules TN, et al: Extracorporeal life support for severe acute respiratory distress syndrome in adults. Ann Surg 2004; 240:595–605; discussion 605 41. Schmidt M, Bailey M, Sheldrake J, et al. Predicting Survival after ECMO for Severe Acute Respiratory Failure: The Respiratory ECMO Survival Prediction (RESP)-Score. Am J Respir Crit Care Med 2014; 189:1374–1382 www.ccmjournal.org

11

Copyright (c) Society of Critical Care Medicine and Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited