Prediction of readiness to decannulation from venovenous extracorporeal membrane oxygenation

Abstract

Objective:

The criteria and process for liberation from extracorporeal membrane oxygenation in patients with severe acute respiratory distress syndrome are not standardized. The predictive accuracy of the oxygen challenge test as a diagnostic test in determining weaning and decannulation from venovenous extracorporeal membrane oxygenation was tested.

Design:

A single-centre, retrospective, observational cohort study.

Setting:

Tertiary referral severe respiratory failure centre in a university hospital in the United Kingdom.

Patients:

253 adults with severe acute respiratory distress syndrome requiring extracorporeal membrane oxygenation.

Intervention:

None

Measurements and Main Results:

Patients had median age: 43 years (interquartile range: 32-52) years, extracorporeal membrane oxygenation days: 9 (interquartile range: 6-14) and acute physiology and chronic health evaluation II score 17.5 (interquartile range: 15-20). Oxygen challenge test value (PaO_2-OCT) with best prediction was 31 kPa (232 mmHg; sensitivity 0.74; specificity 0.70; area under curve 0.77 (confidence interval: 0.73-0.81)). PaO_2-OCT did not perform well as a prospective test to identify readiness to decannulation. Only 24 patients (10%) were decannulated 48 hours after their first positive oxygen challenge test (true positive) and 73.4% patients were false positives (positive oxygen challenge test but not decannulated). True positives had higher tidal volume (541 ± 218 vs 368 mL ± 210; p < 0.05) and minute ventilation (9.34 ± 5.36 vs 6.33 L/min ± 4.43; p < 0.05). Blood flow (3.17 ± 0.23 vs 3.53 L/min ± 0.56; p < 0.05), sweep gas flow (1.42 ±1.83 vs 3.74 L/min ± 2.43; p < 0.05) and extracorporeal membrane oxygenation minute volume at time of first positive oxygen challenge test was lower in true positives (1.66 ± 2.26 vs 4.82 ± 3.43 L/min). This was a strong predictor for decannulation within 48 hours (area under curve: 0.88, confidence interval: 0.88-0.89).

Conclusions:

In severe acute respiratory distress syndrome requiring venovenous extracorporeal membrane oxygenation, the PaO_2-OCT is a poor predictor of readiness to decannulate from extracorporeal membrane oxygenation. Additional factors involved in the control of respiratory drive and carbon dioxide clearance, particularly native lung dead space and total minute ventilation, should be assessed.

Keywords

adult acute respiratory distress syndrome mechanical ventilation recruitment extracorporeal membrane oxygenation weaning

Introduction

The use of venovenous extracorporeal membrane oxygenation (VV ECMO) for severe respiratory failure is increasing year on year, with improved survival rates.¹ A recent randomized trial and its subsequent post hoc Bayesian analysis have shown that ECMO improves survival when initiated early and particularly in patients with less severe disease.^2,3 These benefits have been further confirmed in meta-analyses and propensity-matched studies.^4–8 The effect of ECMO on survival is mainly due to two main factors: (1) the improvement in technology that allows lower levels of anticoagulation and less need for transfusion of blood products and (2) the possibility to achieve lung rest through a reduction of mechanical power both in terms of tidal volume (and therefore driving pressure) and respiratory rate (RR), allowing the lungs time to recover.^9,10

While initiation criteria for ECMO are well defined by national and international societies as well as clinical trials,² the criteria and the process for weaning and subsequent decannulation are not standardized.^2,6,11 In general terms, as the native lung function improves extracorporeal blood flow (Q_EC) and sweep gas flow (Q_G) are gradually reduced. Sweep gas flow is then ‘trialled off’ and if lung function is sufficient to maintain an acceptable gas exchange and work of breathing, decannulation can occur after >1-24 hours, depending on local protocols and clinical judgement.¹² Testing for readiness to wean is important to identify the patient who may no longer benefit from ECMO and reduce the exposure to the adverse effects of ECMO.

The prediction of readiness to wean – similarly to weaning from mechanical ventilation – requires criteria and tests to ascertain whether native lung function has recovered enough to support gas exchange necessary to meet metabolic demands. One of the tests that is often employed as a marker of lung recovery is the oxygen challenge test (OCT), where the arterial partial pressure of oxygen (P_aO₂) is measured 10-15 minutes after increasing the inspired oxygen fraction delivered by the ventilator (F_iO₂) to 1.0 (Cilley Test). Although this test is often reported in the ECMO literature and in the Extracorporeal Life Support Organization (ELSO), Ann Arbor, MI, USA guidelines,¹³ there is little data to support its use in clinical practice and its predictive value as indicator of readiness to decannulation. The aim of this study was to test the predictive accuracy of the OCT as a diagnostic test in determining weaning and decannulation from VV ECMO.

Methods

Patients

This was a single-centre, retrospective observational cohort study of adult patients admitted to the severe respiratory failure and ECMO service at Guy’s and St Thomas’ Hospital (GSTT), London (United Kingdom), who required VV ECMO between January 2014 and December 2017 and were successfully decannulated. The study had institutional approval (institutional research governance reference number 7820). The need for individual informed consent was waived as this was a retrospective analysis of data collected prospectively for usual clinical care, with no breach of privacy or anonymity. Data were reviewed from our electronic patient record (ICIP; Philips, Eindhoven, The Netherlands). The study qualified as a service evaluation as defined by the United Kingdom National Health Service (UK NHS) Health Research Authority and therefore did not require review by the Research Ethics Committee (http://www.hra.nhs.uk).

ECMO management

The criteria used for VV ECMO initiation is as per the conventional ventilation or ECMO for severe adult respiratory failure (CESAR) trial.⁶ Our standard approach of bifemoral percutaneous cannulation (25 F multi-stage access cannula and 23 F return cannula (BioMedicus, Medtronic Inc, Minneapolis, USA) with ultrasound and fluoroscopy use has previously been described.^14,15 Once stability is achieved, rest ventilator settings are employed (F_iO₂: 0.3, positive end-expiratory pressure (PEEP): 10 cmH₂O, plateau pressure 20 cmH₂O, RR 10) while keeping the membrane oxygenator F_SO₂ at 1.0. Daily assessment for weaning and titration of Q_G and Q_EC is managed by the intensive care team. Once tidal volume improves to >2-3 mL kg⁻¹ predicted body weight (PBW), with the ventilation settings as decided by the treating clinical team, OCTs are performed daily. Patients are usually trialled off sweep gas flow for 12-24 hours prior to decannulation.

Data collection

Collected data included demographics, duration of ECMO, all documented arterial blood gas results including OCTs, ventilatory parameters (pulmonary minute volume (V_e), tidal volume (V_T), RR), and ECMO settings (Q_EC and Q_G)). Patients included in the study had OCTs performed prior to decannulation.

We defined and calculated the following:

Corrected pulmonary minute volume (V_eCORR)

V_{e C O R R} = V_{e} \cdot \frac{P_{a} C O_{2}}{5.5}

(1)

where V_e is the pulmonary minute volume (L min⁻¹) and P_aCO₂ is the arterial partial pressure of CO₂ (kPa).

ECMO minute volume (V_ECMO)

V_{E C M O} = l o g (Q_{E C}) \times Q_{G}

(2)

ECMO corrected minute volume (V_ECMO-CORR)

V_{E C M O - C O R R} = V_{E C M O} . \frac{P_{a} C O_{2}}{5.5}

(3)

where V_ECMO is the ECMO minute volume (L min⁻¹; equation (2)) and P_aCO₂ the arterial partial pressure of CO₂ in kPa.

Total corrected minute volume (V_TOTAL-CORR)

V_{T O T A L - C O R R} = (V_{e} + l o g (Q_{E C}) . Q_{G}) . \frac{P_{a} C O_{2}}{5.5}

(4)

where V_e is the pulmonary minute volume (L min⁻¹), Q_EC is the ECMO blood flow, and Q_G is the ECMO sweep gas flow in L min⁻¹.

V_eCORR (equation (1)) is a recognized proxy of physiological dead space of the native lung, calculated as the pulmonary minute volume (V_e) adjusted to achieve a P_aCO₂ of 5.5 kPa (41.3 mmHg).^16,17 Further variables were calculated to adjust for the relative contribution of ECMO and native lung on the readiness to wean. We reasoned that the CO₂ removal (VCO₂) of the membrane lung, for any given pre-membrane PCO₂ (P_preCO₂), can be roughly determined as proportional to the product of Q_G and the logarithm of Q_EC. This is because the effective ventilation (V_EFF) of the membrane lung (in L min⁻¹) is proportional to VCO₂ (i.e. V_EFF = VCO₂/PCO₂ × K; or if we consider the fraction and PCO₂ constant, V_EFF is directly proportional to VCO₂). The product of Q_G and the logarithm of Q_EC may therefore represent a measure of ECMO minute volume (V_ECMO; equation (2)). We also estimated the corrected ECMO minute volume (V_ECMO-CORR) required to obtain a P_aCO₂ of 5.5 kPa (41.3 mmHg; equation (3)) in a fashion analogous to the corrected pulmonary minute volume (equation (1)). It follows that the corrected total minute volume $(V_{TOTAL-CORR})$ is the sum of the ECMO and pulmonary minute volume, required to achieve P_aCO₂ = 5.5 kPa (41.3 mmHg; equation (4)).

Statistical analysis

Diagnostic test evaluations (sensitivity, specificity, area under the curve (AUC), and Youden J statistic) were performed by the construction and evaluation of receiver operating characteristic (ROC) curves. Ventilation and ECMO parameters were compared between differing OCT outcome groups. The assumption of homogeneous variance between groups was assessed and one-way analysis of variance (ANOVA) performed if homogeneous and Welch’s ANOVA if not. If statistically significant between-group differences were found, subsequent post hoc comparisons were made with Tukey’s honest significance test and pairwise comparisons or Games-Howell post hoc comparisons in the case of heterogeneous variance between groups. Two group comparisons employed Welch’s T-test to assess for statistically significant differences. Ventilation and ECMO parameter means are reported with standard deviations for OCT outcome groups. Multivariable logistic regression models were constructed using a pragmatic forward variable selection strategy (see Online Additional Information) to examine the predictive ability of multiple combinations of ventilation and ECMO parameters. All data extraction, cleaning and analysis routines were performed using the R project for Statistical Computing.¹⁸

Results

Patient characteristics

ECMO was performed in 310 patients, of whom 253 (81.6%) survived to decannulation. An OCT was not performed in 12 patients (4.7%), and therefore, these patients were excluded from the study (See Online Additional Information). Baseline characteristics are shown in Table 1.

Table 1.

Baseline characteristics.

		n (%)	Median	Interquartile range (IQR)
Female		101 (42)
Male		140 (58)
Age (years)			43	32-52
ECMO (days)			9	6-14
PaO_2-OCT (kPa (mmHg)) 24 h pre decannulation			41.7 (312)	32.3-53.8 (242-403)
APACHE II			17.5	15-20
Parameter	First 24 hours		Last 24 hours
	Mean (SD)	Median (IQR)	Mean (SD)	Median (IQR)
ECMO blood flow (L min⁻¹)	4.2 (0.8)	4.2 (3.6, 4.8)	3.1 (0.4)	3.1 (3.0, 3.2)
ECMO sweep gas flow (L min⁻¹)	5 (2.2)	5 (3, 6)	0.3 (0.9)	0 (0, 0)
ECMO minute volume (L min⁻¹)	7.2 (3.8)	6.4 (4.4, 19.3)	0.3 (1.0)	0 (0, 0)
Respiratory rate (min⁻¹)	12 (6)	10 (10, 12)	22 (8)	21 (16, 26)
PEEP (cmH₂O)	10.1 (3.6)	10 (10, 10)	10.1 (3.7)	10 (8, 12)
Expiratory tidal volume (mL)	189 (148)	160 (70, 280)	573 (202)	540 (430, 690)
Pulmonary minute volume (L min⁻¹)	2.3 (2.2)	1.8 (0.8, 3.3)	11.4 (4.4)	10.7 (8.4, 13.7)
Corrected minute volume (L min⁻¹)	2.4 (2.3)	1.9 (0.8, 3.5)	12.3 (4.5)	11.9 (9.3, 14.5)
Ventilator F_iO₂	0.37 (0.14)	0.3 (0.3, 0.4)	0.34 (0.09)	0.35 (0.3, 0.4)

ECMO: extracorporeal membrane oxygenation; PaO_2-OCT: arterial PO₂ during an oxygen challenge test; APACHE II: Acute Physiology And Chronic Health Evaluation II score; SD: standard deviation; PEEP: positive end-expiratory pressure.

OCT as a decannulation predictor

We reviewed all OCT results, considering whether they were within the 48 hours prior to decannulation, to identify the cut-off value of OCT (P_aO_2-OCT) with the best prediction for successful decannulation. Using ROC curve analysis, we identified a threshold P_aO_2-OCT value of 31 kPa (232 mmHg; sensitivity: 0.74; specificity: 0.70; AUC: 0.77 (CI: 0.73-0.81)).

Prospective evaluation of P_aO_2-OCT as a predictor of ECMO decannulation

We defined positive OCT as an P_aO_2-OCT > 30 kPa (225 mmHg). We then tested the P_aO_2-OCT prospectively to understand the power to predict decannulation after the first positive P_aO_2-OCT. OCT results were defined according to subgroups. A true positive was defined as a positive OCT achieved within the 48-hour period prior to decannulation; a false positive when a positive OCT occurred more than 48 hours prior to decannulation and a negative OCT where a positive result was never achieved, but the patient was successfully decannulated from ECMO. A period of 48 hours prior to decannulation was utilized because at our institution, patients are off sweep gas flow for 12-24 hours before proceeding to decannulation. True positives were found in 24 (10%) patients, false positives in 177 (73.4%) and negative in 40 (16.6%).

Comparisons of observations were reviewed from the 12-hour period before and 12-hour period after the first positive OCT. Comparison between the true positive and the false positive patient groups showed higher mean tidal volume (541 ± 218 vs 368 ± 210 mL; p < 0.05) and higher mean pulmonary minute volume (9.34 ± 5.36 vs 6.33 ± 4.43 L min⁻¹; p < 0.05) in the true positive group. When comparing V_eCORR, however, we found that it was also higher in the true positive compared to false-positive group (9.95 ± 5.46 vs 6.97 ± 4.8 L min⁻¹; p < 0.05; Figure 1). This result was associated with a statistically lower Q_EC (3.17 ± 0.23 vs 3.53 ± 0.56 L min⁻¹; p < 0.05) and Q_G (1.42 ± 1.83 vs 3.74 ± 2.43 L min⁻¹; p < 0.05) in true positives (i.e. patients who were decannulated 48 hours after the first positive OCT).

Figure 1.

Pulmonary, ECMO and total corrected minute volumes at the time of first positive OCT.

Pre-decannulation data

Ventilator and ECMO settings were reviewed for the 24 hours prior to decannulation. Mean tidal volumes of true-positive to false-positive group (597 ± 209 vs 576 ± 201 mL; p = 0.05), mean minute volume (11.23 ± 4.84 vs 11.68 ± 4.4 L min⁻¹; p = 0.05) and mean corrected minute volume (11.82 ± 5.11 vs 12.66 ± 4.45 L min⁻¹; p < 0.05) were all clinically similar. ECMO settings were clinically similar for true-positive to false-positive groups: mean Q_EC (3.1 ± 0.16 vs 3.06 ± 0.38 L min⁻¹; p = 0.032) and Q_G (0.26 ± 0.85 vs 0.22 ± 0.84 L min⁻¹; p = 0.609).

ECMO minute volume as a decannulation predictor

Given the poor predictive power of the OCT when applied prospectively, we used similar ROC analysis to identify single parameters which might better predict decannulation. As predicted, the mean ECMO minute volume, within 12 hours of first positive OCT, was higher in patients with a false-positive OCT compared to true positive (4.82 ± 3.43 vs 1.66 ± 2.26 L min⁻¹; Figure 1) and was a strong predictor for decannulation within 48 hours (AUC: 0.88, CI: 0.88-0.89).

Composite criteria were examined by construction of multivariable logistic regression models and ROC analysis of their respective outcome predictions. In particular, the performance of the models that included the ECMO contribution (as ECMO minute volume) and the contribution of native lung either as P_aO_2-OCT (AUC: 0.89 (95% CI: 0.888-0.897)) or as expiratory minute volume (AUC: 0.89 (95% CI: 0.887-0.896)) (Figure 2), were statistically nearly identical (Table 2), demonstrating that P_aO_2-OCT does not add to predictive models that include minute volumes alone (see Online Additional Information).

Figure 2.

Receiver operating characteristics (ROC) curves for prospective P_aO_2-OCT values and logistic regression models of corrected and non-corrected pulmonary and ECMO minute volumes.

Table 2.

ROC AUCs for prospective P_aO_2-OCT values and logistic regression models of corrected and non-corrected pulmonary and ECMO minute volumes, with odds ratios for logistic regression model parameters.

Model	AUC (95% CI)	Variable	Odds ratio (95% CI)
P_aO_2-OCT	0.77 (0.73-0.81)	–	–
P_aO_2-OCT + V_ECMO	0.89 (0.89-0.90)	P_aO_2-OCT	1.006 (1.004-1.009)
		V_ECMO	0.52 (0.51-0.53)
V_e + V_ECMO	0.89 (0.89-0.90)	V_e	1.02 (1.01-1.03)
		V_ECMO	0.53 (0.52-0.54)
V_eCORR + V_ECMO-CORR	0.90 (0.89-0.90)	V_eCORR	0.99 (0.99-1.0)
		V_ECMO-CORR	1.24 (1.22-1.38)

ROC: receiver operator characteristic; AUC: area under the curve; P_aO_2-OCT, arterial oxygen partial pressure measured during oxygen challenge test; ECMO: extracorporeal membrane oxygenation; CI: confidence interval; V_ECMO: ECMO minute volume; Ve: minute volume; V_eCORR + V_ECMO-CORR: corrected minute volume and corrected ECMO minute volume.

Discussion

Testing for readiness to wean is important to identify the patient who may no longer benefit from ECMO and reduce the exposure to the adverse effects of ECMO. In this study, we show that in patients with severe acute respiratory distress syndrome (ARDS) requiring VV ECMO, the P_aO_2-OCT is a poor predictor of readiness to wean from ECMO. The performance of P_aO_2-OCT may have also unintended consequences of increasing the risks for lung tissue and arterial hyperoxia, particularly in patients with false-positive OCT test, as they may be exposed to more prolonged cumulative doses of hyperoxia.^19–21 The poor positive predictive value highlights three points: (1) ECMO dependency is not necessarily related to oxygenation, (2) testing oxygenation on ECMO may not predict sufficient P_aO₂ once the patient is trialled off ECMO and (3) prediction of readiness to weaning and decannulation from ECMO requires the assessment of additional factors involved in the control of respiratory drive and carbon dioxide clearance, particularly native lung dead space and total minute volume.

Regarding the first point, P_aO_2-OCT does not predict decannulation because oxygenation may not be the primary reason for ECMO dependency. This is consistent with data from ECMO to rescue lung injury in severe ARDS trial (EOLIA) that shows the conventional and ECMO groups have similar oxygenation but different carbon dioxide post randomization.² Extracorporeal membrane lung may affect oxygenation through various mechanisms. One is the effect of ECMO on the hypoxic pulmonary vasoconstriction (HPV). HPV physiologically aims to reduce blood flow to poorly ventilated lung preserving ventilation perfusion (V/Q) ratio. However, HPV is impaired in ECMO because of the increased mixed venous oxygen (P_mixO₂).^22,23 The abolition of this physiological response results in increased effective shunt fraction and can result in arterial desaturation and increased venous admixture if V/Q matching is profound and there is no time for restoring HPV.¹¹ Another factor is the effect of CO₂ removed extracorporeally on the native lung function and reduction of the alveolar PO₂ based on modification of the respiratory quotient and change in the pulmonary hypoxic vasocontriction.^24,25 Finally, in a subgroup of patients, ECMO may cause re-opening of a foramen ovale which can cause a positive OCT but severe arterial desaturation during trialling off ECMO.²⁶

The second point is that P_aO_2-OCT does not reflect just the function of the native lung, but it results from the interaction between ECMO and native lung. Therefore, a positive OCT may not guarantee sufficient oxygenation to allow decannulation from ECMO. In particular, ECMO increases S_preO₂, and therefore, for the same native lung shunt fraction and Q_EC to cardiac output ratio, P_aO_2-OCT will be higher while the patient is on ECMO. When the ECMO is trialled off, the reduction in S_preO₂ plays a role in increasing venous admixture with consequent decrease in P_aO₂. If we use an example of a patient with a cardiac output of 5 L min⁻¹, an oxygen consumption of 300 mL min⁻¹, a native lung shunt fraction of 30% and a Q_EC of 3 L min⁻¹ (assuming a recirculation fraction (R_f) of 20% and an intramembrane shunt fraction of 10%), the P_aO_2-OCT would be 37.9 kPa (284 mmHg; S_aO₂: 98.1%) with a S_mixO₂ of 83% and a PO₂ of final mixture (P_mixO₂) of 6.6 kPa (49 mmHg). This patient would have a positive OCT. However, when the F_SO₂ of the membrane lung is reduced to 0.21, but the F_iO₂ of the ventilator left at 1, the P_aO_2-OCT will decrease to 25.7 kPa (193 mmHg; S_aO₂: 97.7%) with a S_mixO₂ of 75.2% and a P_mixO₂ of 5.6 kPa (42 mmHg). If this patient is decannulated from ECMO, the P_aO₂ at F_iO₂ of 1 would further drop to 9.6 kPa (72 mmHg; S_aO₂ 93%) with a S_preO₂ of 50.8% and a P_vO₂ of 3.7 kPa (28 mmHg). If the F_iO₂ on the ventilator is reduced to 0.6, the P_aO₂ would be 7.3 kPa (55 mmHg) in the absence of a change in hypoxic vasoconstriction. If, as a consequence of better V/Q matching and HPV, the functional shunt fraction decreases to 20%, P_aO₂ may improve overtime to 10.7 kPa (80 mmHg). Still this is far from the original 37.9 kPa (284 mmHg; calculations using http://www.ecmomodel.unimi.it/p/index.php).

In addition, given the lower S_mixO₂ off ECMO, the same amount of the native lung oxygen consumption (VO₂) will have first to be spent in increasing oxygen saturation by a quantity that is equal to (1.39 × (100-S_preO₂) × haemoglobin (Hb)). This condition, by definition, will reduce the amount of dissolved oxygen (P_aO₂).

The final point is that ECMO CO₂ removal depends on sweep gas flow and blood flow for a given membrane size and function and a P_preCO₂.²⁷ Clearance of CO₂ by the native lung is essential for weaning and decannulation from ECMO; however, we have no data on the precise artificial and native lung CO₂ transfer (VCO₂). Therefore, we used a proxy estimate of the contribution of ECMO on the CO₂ elimination (ECMO minute volume) and as a surrogate for physiological dead space and impaired ability for native lung to eliminate CO₂. We found that ECMO minute volume was higher in false-positive OCTs. This result indicates, as evident in daily clinical practice, a greater contribution from the membrane lung and indicates that the majority of patients require ECMO not for oxygenation but mainly for CO₂ removal.

This study has the strengths of including a large cohort of patients, protocolised care and a well-documented electronic physiological and clinical data. The study combines a physiological hypothesis with a ‘big data/data puddle approach’ (see online Additional Information).

There are clear limitations to this study particularly the retrospective design, and the lack of clear documentation on whether patients who were not decannulated after the first positive P_aO_2-OCT (false positive) remained on ECMO because of desaturation during a trial off test or because of hypercapnia. In addition, we have no routine data on the VCO₂ of the membranes and native lungs. However, we believe that the results of this large data approach show the importance of the interaction between artificial and native lung and the role of functional dead space or CO₂ removal in ECMO dependence. In addition, repeated OCT may expose patients, particularly the group we defined as ‘false positive’ to additional cumulative hours of hyperoxaemia with possible, although unquantified, effects of short-term outcomes.

Conclusion

In patients with severe ARDS requiring VV ECMO, the P_aO_2-OCT is a poor predictor of readiness to decannulate from ECMO. Additional factors involved in the control of respiratory drive and carbon dioxide clearance, particularly native lung dead space and total minute volume should be routinely assessed.

Supplemental Material

Sup_Material_908115_-_Prediction_of_readiness_to_decannulation_from_venovenous_extracorporeal_membrane_oxygenation – Supplemental material for Prediction of readiness to decannulation from venovenous extracorporeal membrane oxygenation

Supplemental material, Sup_Material_908115_-_Prediction_of_readiness_to_decannulation_from_venovenous_extracorporeal_membrane_oxygenation for Prediction of readiness to decannulation from venovenous extracorporeal membrane oxygenation by Emma L Hartley, Barnaby Sanderson, Francesco Vasques, Kathleen JR Daly, Maria Lozinski, Nicholas A Barrett and Luigi Camporota in Perfusion

Footnotes

Acknowledgements

We are grateful to all the members of the Severe Respiratory Failure team at Guy’s and St Thomas’ Trust for the collection of data, retrieval and subsequent care of these very sick patients.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

ORCID iDs

Emma L Hartley

Nicholas A Barrett

Supplemental Material

Supplemental material for this article is available online.

References

Extracorporeal Life Support Organization. ECLS registry report. Ann Arbor, MI: Extracorporeal Life Support Organisation, https://www.elso.org/Registry/Statistics/InternationalSummary.aspx (2019, accessed 30 November 2019).

Combes

Hajage

Capellier

, et al. Extracorporeal membrane oxygenation for severe acute respiratory distress syndrome. N Engl J Med 2018; 378: 1965–1975.

Goligher

Tomlinson

Hajage

, et al. Extracorporeal membrane oxygenation for severe acute respiratory distress syndrome and posterior probability of mortality benefit in a post hoc Bayesian analysis of a randomized clinical trial. JAMA 2018; 320: 2251–2259.

Munshi

Walkey

Goligher

, et al. Venovenous extracorporeal membrane oxygenation for acute respiratory distress syndrome: a systematic review and meta-analysis. Lancet Respir Med 2019; 7: 163–172.

The Australia and New Zealand Extracorporeal Membrane Oxygenation (ANZ ECMO) Influenza Investigators. Extracorporeal membrane oxygenation for 2009 influenza A (H1N1) acute respiratory distress syndrome. JAMA 2009; 302: 1888–1895.

Peek

Mugford

Tiruvoipati

, et al. 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.

Noah

Peek

Finney

, et al. Referral to an extracorporeal membrane oxygenation center and mortality among patients with severe 2009 influenza A (H1N1). JAMA 2011; 306: 1659–1668.

Pham

Combes

Roze

, et al. 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.

Brodie

Slutsky

Combes

Extracorporeal life support for adults with respiratory failure and related indications: a review. JAMA 2019; 322: 557–568.

10.

Schmidt

Pham

Arcadipane

, et al. Mechanical ventilation management during extracorporeal membrane oxygenation for acute respiratory distress syndrome. An international multicenter prospective cohort. Am J Respir Crit Care Med 2019; 200: 1002–1012.

11.

Vasques

Romitti

Gattinoni

, et al. How I wean patients from veno-venous extra-corporeal membrane oxygenation. Crit Care 2019; 23: 316.

12.

Bartlett

. Physiology of extracorporeal life support. In: Annich

GMLW

MacLaren

Wilson

, et al. (eds) ECMO extracorporeal cardiopulmonary support in critical care. 4th ed. Ann Arbor, MI: Extracorporeal Life Support Organisation, 2012, p. 30.

13.

Extracorporeal Life Support Organization. ELSO guidelines for adult respiratory failure (version 1.4). Ann Arbor, MI: Extracorporeal Life Support Organization, https://www.elso.org/Portals/0/ELSO%20Guidelines%20For%20Adult%20Respiratory%20Failure%201_4.pdf (2017, accessed 30 November 2019).

14.

Camporota

Caricola

Bartolomeo

, et al. Lung recruitability in severe acute respiratory distress syndrome requiring extracorporeal membrane oxygenation. Crit Care Med 2019; 47: 1177–1183.

15.

Sherren

Shepherd

Glover

, et al. Capabilities of a mobile extracorporeal membrane oxygenation service for severe respiratory failure delivered by intensive care specialists. Anaesthesia 2015; 70: 707–714.

16.

Wexler

Lok

A simple formula for adjusting arterial carbon dioxide tension. Can Anaesth Soc J 1981; 28: 370–372.

17.

Nuckton

Alonso

Kallet

, et al. Pulmonary dead-space fraction as a risk factor for death in the acute respiratory distress syndrome. N Engl J Med 2002; 346: 1281–1286.

18.

Team

RC.

R: A language and environment for statistical computing. Vienna: R Foundation for Statistical Computing, 2019.

19.

Chu

Kim

Young

, et al. Mortality and morbidity in acutely ill adults treated with liberal versus conservative oxygen therapy (IOTA): a systematic review and meta-analysis. Lancet 2018; 391: 1693–1705.

20.

Palmer

Post

Klapaukh

, et al. The association between supraphysiologic arterial oxygen levels and mortality in critically ill patients. A multicenter observational cohort study. Am J Respir Crit Care Med 2019; 200: 1373–1380.

21.

Girardis

Busani

Damiani

, et al. Effect of conservative vs conventional oxygen therapy on mortality among patients in an intensive care unit: the oxygen-ICU randomized clinical trial. JAMA 2016; 316: 1583–1589.

22.

Marshall

BE.

Influence of perfusate PO₂ on hypoxic pulmonary vasoconstriction in rats. Circ Res 1983; 52: 691–696.

23.

Marshall

Site and sensitivity for stimulation of hypoxic pulmonary vasoconstriction. J Appl Physiol Respir Environ Exerc Physiol 1983; 55: 711–716.

24.

Gattinoni

Ultra-protective ventilation and hypoxemia. Crit Care 2016; 20: 130.

25.

Diehl

Mercat

Pesenti

Understanding hypoxemia on ECCO2R: back to the alveolar gas equation. Intensive Care Med 2019; 45: 255–256.

26.

Gillon

Barrett

Ioannou

, et al. Intracardiac right-to-left shunt impeding liberation from veno-venous extracorporeal membrane oxygenation: two case studies. Crit Care Med 2016; 44: e583–e586.

27.

Kolobow

Gattinoni

Tomlinson

, et al. Control of breathing using an extracorporeal membrane lung. Anesthesiology 1977; 46: 138–141.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

1.16 MB

Prediction of readiness to decannulation from venovenous extracorporeal membrane oxygenation

Abstract

Objective:

Design:

Setting:

Patients:

Intervention:

Measurements and Main Results:

Conclusions:

Keywords

Introduction

Methods

Patients

ECMO management

Data collection

Statistical analysis

Results

Patient characteristics

OCT as a decannulation predictor

Prospective evaluation of PaO2-OCT as a predictor of ECMO decannulation

Pre-decannulation data

ECMO minute volume as a decannulation predictor

Discussion

Conclusion

Supplemental Material

Sup_Material_908115_-_Prediction_of_readiness_to_decannulation_from_venovenous_extracorporeal_membrane_oxygenation – Supplemental material for Prediction of readiness to decannulation from venovenous extracorporeal membrane oxygenation

Footnotes

Acknowledgements

Declaration of Conflicting Interests

Funding

ORCID iDs

Supplemental Material

References

Supplementary Material

Prospective evaluation of P_aO_2-OCT as a predictor of ECMO decannulation