To the Editor,
The Acute Respiratory Distress Syndrome (ARDS) is still associated with high mortality [1], despite application of recent guidelines [2, 3]. The EOLIA study suggested that Extra- Corporeal Membrane Oxygenation (ECMO) could be effective in some of the most severe patients, but failed to demonstrate a 20% increase in survival [4]. One reason could be that criteria for selecting patients were only based on blood gas analysis. Our hypothesis is that adding other factors could allow a better selection of patients who could benefit from ECMO.
We took advantages to have a large multicentric cohort of patients under protective ventilation for moderate-to-severe ARDS [5] to determine the incidence, characteristics and outcome of patients eligible for ECMO according to EOLIA-based criteria and to identify patients who would benefit the most of the technique. ECMO was only used in these centers as a recue therapy.
Anzeige
Mechanical ventilation was applied in the volume-assist control mode, with a target tidal volume (TV) of 6–8 mL/kg (predicted body weight) and a plateau pressure < 30 cmH2O. Respiratory rate could be increased in case of high arterial carbon dioxide partial pressure (PaCO2). Prone positioning was left to the discretion of the attending physician, but was typically performed in patients with a PaO2/FiO2 < 100 mmHg and/or an acute core pulmonale (ACP) [6]. Patients eligible for ECMO according to EOLIA-based criteria were identified as follows: PaO2/FiO2 < 80 mmHg with optimal PEEP, or a pH < 7.25 and PaCO2 > 60 mmHg with a respiratory rate ≥ 35 cycles/min, despite the use of prone positioning or nitric oxide inhalation.
Statistical analysis was performed with R.4.0.4. Patients eligible for ECMO were compared to the rest of the cohort. Continuous data, expressed as medians (interquartile ranges), were compared with Mann–Whitney test. Categorical variables, expressed as numbers and percentages, were compared using the chi-square test or Fisher exact test. To evaluate independent factors associated with ICU mortality in this identified subgroup of patients, significant or marginally significant (p < 0.10) bivariate risk factors were examined using univariate and multivariable backward stepwise mixed logistic regression stratified on the center. SAPS II was forced in the model.
752 patients were studied. Characteristics and outcome are given in the Table 1. 67 (9%) patients were potentially eligible for ECMO. They had lower PaO2/FiO2 (62 [55–72] versus 114 [90–120] mmHg: p < 0.01) and higher incidence of ACP (42% versus 20%, p < 0.001). Only 8 of them underwent the procedure. In-ICU mortality in the whole cohort was 36%. Causes of death in patients eligible for ECMO was multi-organ failure in 21 (68%), neurologic in 4 (13%) and ECMO complication in 3 (10%). Only 3 patients (10%) died from hypoxic cardiac arrest.
Table 1
Clinical characteristics and outcome of the entire cohort according to EOLIA criteria
Characteristics, outcomes and complications | ARDS patients without EOLIA criteria (n = 685) | ARDS patients with EOLIA criteria (n = 67) | p-value |
|---|---|---|---|
Age (years) | 59 (47–72) | 56 (47–70) | 0.43 |
Male sex, n (%) | 465 (68) | 42 (63) | 0.47 |
SAPS II | 51 (38–65) | 47 (33–64) | 0.27 |
Weight (kg) | 77 (65–84) | 70 (59–86) | 0.43 |
Cause of ARDS, n (%) | 0.20 | ||
Pneumonia | 83 (12) | 8 (12) | |
Aspiration | 259 (39) | 34 (51)) | |
Non-pulmonary sepsis | 245 (37) | 17 (26) | |
Other causes | 83 (12) | 7 (11) | |
Respiratory setting at inclusion | |||
Tidal volume (ml/kg) | 6.7 (6.0–8.0) | 6.02 (5.4–6.9) | < 0.01 |
Respiratory rate (cycle/min) | 22 (16–27) | 26 (22–30) | < 0.01 |
PEEP (cmH2O) | 8 (5–10) | 10 (7–12) | < 0.01 |
Plateau pressure (cmH2O) | 24 (21–28) | 27 (25–29) | < 0.01 |
Compliance (ml/cmH2O) | 30.7 (24–39.1) | 25.9 (19.3–33.8) | < 0.01 |
Driving pressure (cmH2O) | 15 (13–19) | 17 (14–19.8) | 0.04 |
Arterial blood gases | |||
PaO2/FiO2 ratio (mmHg) | 114 (90–120) | 62 (55–72) | < 0.01 |
PaCO2 (mmHg) | 44 (38–52) | 48 (41–60) | < 0.01 |
Shock, n (%) | 449 (66) | 53 (79) | 0.04 |
Prone positioning, n (%) | 163 (24) | 55 (82) | < 0.01 |
VV ECMO in rescue during ARDS course, n (%) | 0 (0) | 8 (12) | < 0.01 |
RRT during ARDS course, n (%) | 126 (30) | 15 (40) | 0.27 |
Echocardiographic findings | |||
RVEDA/LVEDA | 0.68 (0.57–0.81) | 0.83 (0.64–1.04) | < 0.01 |
Systolic pulmonary artery pressure (mmHg) | 35 (20–48) | 52 (35–59) | < 0.01 |
Severe acute cor pulmonale | 43 (6) | 11 (16) | < 0.01 |
Outcome, n (%) | |||
ICU mortality | 243 (36) | 31 (46) | 0.10 |
ICU stay (days) | 16 (8–30) | 15 (6–31) | 0.43 |
Characteristics and outcome of patients potentially eligible for ECMO according to ICU mortality are given in the Table 2. In multivariable analysis, severe right ventricular dilatation (right-to-left ventricle end-diastolic area ratio > 1) and driving pressure were the only factors associated with in-ICU mortality (OR [95% CI]: 5.62 [1.44–27.39], p = 0.02 and 1.14 [1.01–1.31], p = 0.04, respectively).
Table 2
Clinical characteristics and echocardiographic findings of ARDS patients eligible for ECMO
Characteristics, outcomes and complications | Survivors (n = 36) | Non survivors (n = 31) | p-value |
|---|---|---|---|
Age (years) | 56 (45–70) | 57 (48–70) | 0.44 |
Male sex, n (%) | 23 (64) | 19 (61) | 0.47 |
SAPS II | 44 (32–58) | 53 (36–75) | 0.15 |
Weight (kg) | 74 (63–97) | 67 (56–80) | 0.22 |
Cause of ARDS, n (%) | 0.63 | ||
Pneumonia | 6 (17) | 2 (7) | |
Aspiration | 18 (50)) | 16 (53) | |
Non-pulmonary sepsis | 8 (22) | 9 (30) | |
Other causes | 4 (11) | 3 (10) | |
Respiratory setting at inclusion | |||
Tidal volume (ml/kg) | 6.3 (5.5–7.6) | 5.9 (5.3–6.6) | 0.13 |
Respiratory rate (cycle/min) | 25 (20–27) | 30 (25–30) | < 0.01 |
PEEP (cmH2O) | 10 (8–12) | 10 (7–12) | 0.72 |
Plateau pressure (cmH2O) | 26 (24–29) | 28 (25–38) | 0.07 |
Compliance (ml/cmH2O) | 30 (2338) | 23 (16–28) | < 0.01 |
Driving pressure (cmH2O) | 16 (14–19) | 19 (16–22) | 0.05 |
Arterial blood gases | |||
PaO2/FiO2 ratio (mmHg) | 69 (58–74) | 60 (55–67) | 0.08 |
PaCO2 (mmHg) | 48 (40–52) | 51 (42–70) | 0.17 |
Shock, n (%) | 25 (70) | 28 (90) | 0.07 |
Prone positioning, n (%) | 31 (86) | 24 (77) | 0.52 |
VV ECMO in rescue during ARDS course, n (%) | 2 (1) | 6 (19) | 0.13 |
RRT during ARDS course, n (%) | 9 (33) | 6 (55) | 0.28 |
Echocardiographic findings | |||
RVEDA/LVEDA | 0.71 (0.57–0.93) | 0.98 (0.71–1.10) | 0.10 |
Pulmonary hypertension (mmHg) | 51 (44–55) | 52 (34–63) | 0.85 |
Severe acute cor pulmonale | 3 (8) | 8 (26) | 0.09 |
ICU stay (days) | 19 (14–34) | 12 (3–20) | 0.16 |
Anzeige
A limitation of our study is that eight patients of the cohort received ECMO as a rescue therapy, which may have influenced our results, especially since the technique is now safer when performed in expert centers. However, six of these eight patients died.
In conclusion, we report a 9% incidence of patients who reach the EOLIA-based criteria for ECMO in a large non-selected cohort of ARDS patients ventilated with moderate-to-severe ARDS. These patients exhibited higher driving pressure and more frequent right ventricle failure, both being independently associated with ICU mortality. How this subgroup of patients could be considered as the ideal target for ECMO selection strategy should better be evaluated in the future.
Acknowledgements
Not applicable.
Declaration
Ethics approval and consent to participate
The study was approved by the Institutional Review Boards of participating centers as a component of standard of care and requirement for patient’s consent was waived.
Consent for publication
Not applicable.
Competing interests
MP, PM, AL, PV declare no competing interest. AMD reports grants from Fischer Paykel, Baxter, and Ferring, and personal fees from Air Liquide, Amomed, and Addmedica, all outside the submitted work. AVB reports research Grant from GSK, outside the submitted work.
Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.