Advertisement

The role of ECMO in COVID-19 acute respiratory failure: Defining risk factors for mortality

Published:December 23, 2022DOI:https://doi.org/10.1016/j.amjsurg.2022.12.017

      Highlights

      • ECMO utilization was greatest during the Pre-Delta and Delta periods of the pandemic.
      • Frequency of ECMO duration >3 weeks was much higher during Delta period.
      • Mortality 72 h post-decannulation was 51.9% and crude in-hospital mortality 60.5%.
      • Age ≥50 years increased risk of death by 30%.
      • Respiratory support ≥7 days and pre-ECMO dialysis increased risk of death by 40%.

      Abstract

      Background

      Veno-venous extracorporeal membrane oxygenation (VV ECMO) utilization increased substantially during the COVID-19 pandemic, but without patient selection criteria.

      Methods

      We conducted a retrospective review of all adult patients with COVID-19-associated ARDS placed on VV ECMO at our institution from April 2020 through June 2022.

      Results

      162 patients were included (n = 95 Pre-Delta; n = 58 Delta; n = 9 Omicron). The frequency of ECMO duration greater than three weeks was variable by pandemic period (17% pre-Delta, 41% Delta, 22% Omicron, p = 0.003). In-hospital mortality was 60.5%. Age ≥50 years (RR 1.28, 95% CI 1.01, 1.62), ≥7 days of respiratory support (1.39, 95% CI 1.05, 1.83) and pre-cannulation renal failure requiring dialysis (RR 1.42, 95% CI 1.13, 1.78) were associated with mortality.

      Conclusions

      In this cohort of VV ECMO patients with COVID-19, older age, a longer duration of pre-ECMO respiratory support, and pre-ECMO renal failure all increased the risk of mortality by approximately 30%.

      Keywords

      1. Introduction

      The novel SARS-CoV-2 disease (COVID-19), in the most severe cases, can lead to acute respiratory distress syndrome (ARDS). After the failure of maximal ventilatory therapy, extracorporeal membrane oxygenation (ECMO) may be considered as a rescue therapy.
      • Zangrillo A.
      • Landoni G.
      • Biondi-Zoccai G.
      • et al.
      A meta-analysis of complications and mortality of extracorporeal membrane oxygenation.
      While simple in concept, ECMO is complex in its execution, requiring trained staff and specialist equipment, and has a high burden of complications. ECMO is not a therapeutic modality; it simply supports patients with failing lungs, providing time for lung recovery or—exceptionally— serving as a bridge to lung transplantation.
      • Bharat A.
      • Machuca T.N.
      • Querrey M.
      • et al.
      Early outcomes after lung transplantation for severe COVID-19: a series of the first consecutive cases from four countries.
      Early in the COVID pandemic, concerns about the high mortality rate observed in COVID-19 patients requiring ECMO support were raised.
      • Henry B.M.
      COVID-19, ECMO, and lymphopenia: a word of caution.
      However, subsequent observational cohort studies, including one from the Extracorporeal Life Support Organization (ELSO) registry, reported outcomes for patients with COVID-19 receiving ECMO that were comparable to ECMO-supported patients with non-COVID-19-related ARDS.
      • Combes A.
      • Hajage D.
      • Capellier G.
      • et al.
      Extracorporeal membrane oxygenation for severe acute respiratory distress syndrome.
      • Barbaro R.P.
      • MacLaren G.
      • Boonstra P.S.
      • et al.
      Extracorporeal membrane oxygenation support in COVID-19: an international cohort study of the Extracorporeal Life Support Organization registry.
      • Shaefi S.
      • Brenner S.K.
      • Gupta S.
      • et al.
      Extracorporeal membrane oxygenation in patients with severe respiratory failure from COVID-19.
      However, there has been a significant variance in reported outcomes from various ECMO centers in the United States (US) and Europe.
      Prior to the pandemic, the role of careful patient selection on ECMO outcomes was relatively well established. However, with changes in COVID-19 management and virulence, patient selection for ECMO has become more challenging.
      • Horby P.
      • Lim W.S.
      • Emberson J.R.
      • et al.
      Dexamethasone in hospitalized patients with covid-19.
      • Beigel J.H.
      • Tomashek K.M.
      • Dodd L.E.
      • et al.
      Remdesivir for the treatment of covid-19 - final report.

      Raff L.A., Gallaher J.R., Johnson D., Raff E.J., Charles A.G., Reid T.S. Time to Cannulation after ICU Admission Increases Mortality for Patients Requiring Veno-venous ECMO for COVID-19 Associated Acute Respiratory Distress Syndrome. Ann Surg. 2022 Dec 1;276(6):e659-e663. doi: 10.1097/SLA.0000000000004683. Epub 2020 Dec 22. PMID: 33630477.

      In addition, the duration of ECMO support in the absence of clear improvement in lung recovery has been controversial. We, therefore, sought to evaluate all COVID-19 patients presenting to our high-volume quaternary care institution that progressed to ECMO support. We examined patient characteristics and changes in ECMO-related COVID-19 mortality over the pandemic period while accounting for differing pandemic periods based on COVID-19 variants.

      2. Methods

      We performed a retrospective study of all COVID-19 patients cannulated for veno-venous (VV) ECMO at the University of North Carolina (UNC) Medical Center (UNCMC). UNCMC is a quaternary academic medical center located in Chapel Hill, NC, with approximately 900 inpatient beds. The UNC ECMO program includes neonatal, pediatric, and adult cardiac and respiratory failure patients. Adult patients placed on VV ECMO for respiratory failure are cannulated and managed by trauma/critical care surgeons in the Surgical Intensive Care Unit (SICU). We previously published a description of our VV ECMO management protocol.

      Raff L.A., Gallaher J.R., Johnson D., Raff E.J., Charles A.G., Reid T.S. Time to Cannulation after ICU Admission Increases Mortality for Patients Requiring Veno-venous ECMO for COVID-19 Associated Acute Respiratory Distress Syndrome. Ann Surg. 2022 Dec 1;276(6):e659-e663. doi: 10.1097/SLA.0000000000004683. Epub 2020 Dec 22. PMID: 33630477.

      In general, patient management was consistent throughout the pandemic and the decision to cannulate a patient was made by the on-call ECMO surgeon within the guidelines and recommendation of ELSO. Data was collected using chart reviews by a research assistant and reviewed by an ECMO surgeon.
      We included patients from April 1, 2020, through June 30, 2022. All patients with COVID-19-associated ARDS cannulated for VV ECMO were included in the analysis, including patients <18 years old who were greater than 50 kg in weight and managed by the adult team. We recorded each patient's clinical history from the electronic medical record. The Charlson Comorbidity Index (CCI) was used to compare medical comorbidities.
      • D'Hoore W.
      • Bouckaert A.
      • Tilquin C.
      Practical considerations on the use of the Charlson comorbidity index with administrative data bases.
      The primary outcome of this study was to identify pre-cannulation risk factors associated with in-hospital mortality after VV ECMO for COVID-19 ARDS. We initially examined patients based on the time of presentation relative to the dominant circulating variant during the pandemic. We categorized patients into the Pre-Delta Period, Delta Period, and Omicron Period. The transition from the Pre-Delta to the Delta period was defined as a diagnosis on June 10, 2021, when the Delta variant rapidly increased in its proportion of new COVID-19 cases at UNCMC.
      • Hospitals U.
      UNC Surveillance Report.
      The Omicron period was similarly defined, beginning on December 10, 2021. During December 2021 and January 2022, there was substantial overlap in of patients with either Delta or Omicron infections. Starting in December 2021, all ECMO patients who had routine admission COVID-19 testing also had SARS-CoV-2 genomic sequencing performed if the viral load was sufficient. Consequently, to better examine the potential effects of different variant infections, any patients with confirmed Omicron infection or those who did not have variant data were included in the Omicron Period for analysis, while patients with confirmed Delta infection during the Omicron Period were included in the Delta Period.
      We used bivariate analysis to compare patients based on period presentation (three categories). We utilized Chi-squared tests to compare categorical variables and ANOVA (analysis of variance) for continuous variables with a normal distribution. For continuous variables with a non-normal distribution, we used a Kruskal-Wallis test to compare medians. Means are reported with standard deviations (SD) and medians with interquartile ranges (IQR). We then used a similar strategy to compare patients based on whether they survived to hospital discharge.
      Using variables identified on bivariate analysis, we modeled pre-cannulation predictors of in-hospital mortality after VV ECMO for COVID-19 using a modified Poisson regression model.
      • Zou G.
      A modified Poisson regression approach to prospective studies with binary data.
      ,
      • Chen W.
      • Qian L.
      • Shi J.
      • Franklin M.
      Comparing performance between log-binomial and robust Poisson regression models for estimating risk ratios under model misspecification.
      Initially, the model was fit with statistically significant variables from the bivariate analysis. Variables were removed stepwise if they did not significantly contribute to mortality in our multivariable model. We also tested whether there was a better model fit after adding a polynomial term. The final model is reported with adjusted risk ratios (RR) and a 95% confidence interval.
      We performed all statistical analyses with Stata/SE 17.0 (Stata- Corp LP, College Station, TX). The University of North Carolina Institutional Review Board approved this study (#19–3513 and #20–2448) and waived the need for informed consent.

      3. Results

      A total of 162 patients were cannulated for VV ECMO during the study period, with 95 patients cannulated during the Pre-Delta period, 58 patients during the Delta period, and 9 during the Omicron period. A total of 18 patients were cannulated during the Omicron period, but 9 were confirmed to be Delta by viral genomic sequencing and were included in the Delta group. (Fig. 1).
      Fig. 1
      Fig. 1Changes in COVID-19 ECMO patient volume by pandemic period.
      A comparison of patients, stratified by period, is shown in Table 1. Notable differences included a much higher median age in the Pre-Delta period at 50 years (IQR 42–58) compared to 42 years (IQR 36–48) and 49 years (IQR 43–51, p < 0.001) in the Delta and Omicron periods, respectively. Hispanic whites comprised 37% of all patients in the pre-Delta period compared to only 21% in the Delta period (p = 0.006). African Americans also represented a higher proportion of patients early in the pandemic at 28% Pre-Delta compared to 12% during the Delta period. The median BMI was consistently high at over 36 kg/m2 in each period (p = 0.4), as shown in Fig. 2. The prevalence of diabetes mellitus was variable at over 25% (p = 0.031) for all patients but highest during the Pre-Delta (n = 38, 40%) and Omicron periods (n = 6, 67%). The Charlson Comorbidity Index was low across all cohorts. Fewer patients were insured in the Pre-Delta period at 66% compared to 83% during Delta (p = 0.050). Vaccination was rare, with only 2 patients fully vaccinated at the time of cannulation. Crude in-hospital mortality was statistically similar between the cohorts at 67% (n = 64/95) during the Pre-Delta period, 50% (n = 29/58) in the Delta period, and 56% (n = 5/9, p = 0.1) during the Omicron period.
      Table 1A comparison of COVID-19 patients placed on VV ECMO, stratified by time period of presentation.
      Pre-Delta (n = 95)Delta (n = 58)Omicron (n = 9)p value
      Patient Age (years)
      Median (IQR)50.0 (42.0–58.0)42.0 (36.0–48.0)49.0 (43.0–51.0)<0.001
      Gender: N (%)
      Female31 (33)18 (31)2 (22)0.8
      Male64 (67)40 (69)7 (78)
      Race: N (%)
      African American27 (28)7 (12)1 (11)0.006
      Hispanic White35 (37)12 (21)4 (44)
      Non-Hispanic White30 (32)36 (62)4 (44)
      Other3 (3)3 (5)0 (0)
      BMI at hospital admission (kg/m2)
      Median (IQR)36.2 (10.3)36.8 (9.1)41.0 (9.3)0.4
      Diabetes?
      Yes: N (%)38 (40)15 (26)6 (67)0.031
      Charlson Comorbidity Score
      029 (31)35 (60)0 (0)<0.001
      135 (37)20 (34)7 (78)
      220 (21)3 (5)2 (22)
      35 (5)0 (0)0 (0)
      46 (6)0 (0)0 (0)
      Insurance status
      Self-Pay32 (34)10 (17)4 (44)0.050
      Insured63 (66)48 (83)5 (56)
      Vaccinated against COVID-19 prior to admission?
      Yes: N % ()0 (0)1 (2)1 (11)0.043
      Transferred to UNCH?
      Yes: N (%)82 (86)57 (98)9 (100)0.027
      Crude In-Hospital Mortality: N (%)64 (67)29 (50)5 (56)0.10
      Fig. 2
      Fig. 2Distribution of patient body mass index (BMI) at the time of ECMO cannulation.
      Table 2 compares patients based on ECMO-related outcomes, stratified by the presentation period. During the Pre-Delta period, the median duration was 14 days (IQR 7–18) compared to 17 days (IQR 11–28) during Delta and 9 days (IQR 7–19, p = 0.008) during Omicron. The proportion of patients who were on ECMO for three weeks or greater increased significantly throughout the pandemic, with only 17% (n = 16) during the Pre-Delta period compared to 41% (n = 24) in the Delta period and 22% (n = 2, p = 0.003) in the Omicron cohort. The need for renal replacement therapy (CRRT) during ECMO was common in over 35% of patients but similar across the periods (p = 0.2). Other complications, including cerebrovascular accidents, liver failure, and thromboembolic events, were not common in any group. Bleeding complications were common but were primarily bleeding at cannula sites or tracheostomy sites with no differences between the cohorts except for gastrointestinal bleeding. Median total blood transfusions while on ECMO for packed red blood cells, platelets, and fresh frozen plasma were similar.
      Table 2A comparison of ECMO related outcomes, stratified by time period of presentation.
      Pre-Delta (n = 95)Delta (n = 58)Omicron (n = 9)p value
      Time on ECMO (days)
      Median (IQR)14 (7-18)17 (11-28)9 (7-19)0.008
      On ECMO for Greater than 3 weeks?
      Yes: N (%)16 (17)24 (41)2 (22)0.003
      Tracheostomy Performed on ECMO?
      Yes: N (%)52 (55)48 (83)6 (67)0.001
      Number of ECMO circuit changes: N (%)
      062 (65)31 (53)6 (67)0.3
      127 (28)16 (28)2 (22)
      26 (6)7 (12)1 (11)
      30 (0)2 (3)0 (0)
      40 (0)2 (3)0 (0)
      CRRT while on ECMO?
      Yes: N (%)34 (36)26 (45)5 (62)0.2
      Other Complications: N (%)
      Cerebrovascular accident4 (4)2 (3)1 (11)0.4
      Liver Failure3 (3)6 (10)1 (11)0.1
      Deep Vein Thrombosis or Pulmonary Embolism2 (2)1 (2)0 (0)1.00
      Bleeding Complications: N (%)
      Cannulation site18 (19)8 (14)1 (11)0.7
      Trachesostomy19 (20)17 (29)3 (33)0.3
      Oropharynx29 (31)16 (28)1 (11)0.6
      Gastrointestinal16 (17)2 (3)0 (0)0.026
      Number of Blood Product Units Transfused on ECMO: Median (IQR)
      Packed Red Blood Cells6 (3-12)8 (2-14)3 (0–7)0.1
      Platelets1 (0–4)0 (0–4)0 (0–3)0.8
      Fresh Frozen Plasma0 (0–0)0 (0–0)0 (0–0)0.2
      Total crude in-hospital mortality was 60.5% (n = 98/162). Table 3 compares differences in patient characteristics stratified by hospital survival. The mean age was significantly higher among patients who died at 48.5 years (SD 10.9) compared to 41.9 years (SD 10.4, p < 0.001) in survivors. For patients <50 years old, crude in-hospital mortality was 52.0% (n = 52/100) and for patients ≥50 years, it was 74.2% (n = 46/62). Notably, there were no differences in gender composition, race, BMI, diabetes, Charlson Comorbidity Index, respiratory comorbidities, insurance status, or COVID-19 vaccination status related to mortality.
      Table 3A comparison of characteristics between patients who survived to hospital discharge and those who did not. NIPPV = Noninvasive positive-pressure ventilation.
      Survived to Discharge (n = 64)Died in Hospital (n = 98)p-value
      Patient Age (years)
      Mean (SD)41.9 (10.4)48.5 (10.9)<0.001
      Gender: N (%)
      Male42 (66)69 (70)0.6
      Female22 (34)29 (30)
      Race: N (%)
      Black13 (20)22 (22)0.2
      Hispanic White21 (33)30 (31)
      Non-Hispanic White30 (47)40 (41)
      Other0 (0)6 (6)
      BMI at hospital admission
      Median (IQR)36.0 (30.4–44.7)34.3 (30.2–39.9)0.1
      Diabetes?
      Yes: N (%)27 (42)32 (33)0.2
      Charleson Comorbidity Index: N (%)
      028 (44)36 (37)0.5
      126 (41)36 (37)
      28 (12)17 (17)
      31 (2)4 (4)
      41 (2)5 (5)
      Patient Respiratory Comorbidities: N (%)
      Asthma6 (9)6 (6)0.5
      COPD2 (3)3 (3)1.00
      OSA6 (9)9 (9)1.00
      Smoker4 (6)10 (10)0.6
      Insurance status: N (%)
      Self-Pay14 (22)32 (33)0.1
      Insured50 (78)66 (67)
      Vaccinated against COVID-19 prior to admission?
      Yes: N (%)0 (0)2 (2)0.5
      Transferred to UNCH?
      Yes: N (%)64 (100)84 (86)<0.001
      Time from Diagnosis to ECMO (days)
      Median (IQR)10.0 (6.0–17.0)13.5 (10.0–17.0)0.025
      Time from Hospital Admission to ECMO (days)
      Median (IQR)7.0 (3.0–11.5)10.0 (7.0–14.0)0.001
      Organ Failure Prior to Cannulation: N (%)
      Renal5 (8)25 (26)0.006
      Cardiac arrest with ROSC2 (3)8 (8)0.3
      P:F ratio at cannulation
      Median (IQR)67 (56–80)71 (60–80)0.4
      Evidence of barotrauma at cannulation?
      Yes: N (%)11 (17)26 (27)0.2
      Received NIPPV?
      Yes: N (%)54 (87)83 (90)0.6
      Days of NIPPV prior to intubation
      Median (IQR)2.5 (0.5–6.0)4.0 (1.0–8.0)0.1
      Time from Intubation to ECMO (days)
      Median (IQR)2.0 (0.5–4.0)3.0 (1.0–5.0)0.043
      Total Days of NIPPV and Intubation Prior to ECMO (days)
      Median (IQR)6.0 (2.0–8.5)8.0 (4.0–12.0)0.005
      Time from diagnosis to ECMO (10 days vs. 13.5 days, p = 0.025) and time from hospital admission to ECMO (7.0 days vs. 10.0 days, p = 0.001) were significantly higher in non-survivors. Pre-cannulation renal failure was less common among those who survived (8% vs. 26%, p = 0.006). The median Pao2:Fio2 ratio was approximately 70 in both groups (p = 0.4). The median number of days of NIPPV before intubation was statistically similar between the two groups (2.5 vs. 4.0, p = 0.1) but the median number of days from intubation to ECMO was significantly lower among those who survived (2.0 days vs. 3.0 days, p = 0.043). When the total number of days requiring NIPPV and intubation were combined, the median total days of respiratory support were significantly lower in the survival group (6.0 vs. 8.0 days, p = 0.005).
      Upon modeling, three factors were associated with an increased risk of in-hospital mortality. Age ≥50 years had a RR of death of 1.28 (95% CI 1.01, 1.62). Total days (≥7 days) of respiratory support (NIPPV + mechanical ventilation with intubation) had a RR of 1.39 (95% CI 1.05, 1.83) and pre-cannulation renal failure requiring dialysis had a RR of 1.42 (95% CI 1.13, 1.78). A total of 45 patients (27.8%) had at least two of these factors, while 11 (6.8%) had all three. The RR of death if a patient had any two factors was 1.54 (95% CI 1.22, 1.95). Among the 11 patients with all three factors, 9 died (82%). Notably, the pandemic period was not a significant contributor to the adjusted mortality.

      4. Discussion

      In this study, we report one of the largest single-institution experiences with VV ECMO for patients with severe ARDS secondary to COVID-19 infection. Throughout the pandemic, certain patient factors were predictive of mortality after VV ECMO, including older age, length of respiratory support, and the need for dialysis before cannulation. Independently, each increased the risk of death by over 30%. Patients with at least two factors had a more than 50% increased risk of death. Notably, gender, race, insurance status, medical comorbidity, and BMI were not associated with increased mortality. We also found significant differences in patient characteristics across different pandemic periods, especially in age, with a higher proportion of older patients placed on VV ECMO early in the pandemic. ECMO duration also increased in the Delta period, with more than 40% of patients on ECMO for greater than three weeks compared to only 17% pre-Delta. The number of patients requiring ECMO during the Omicron period decreased dramatically.
      Globally, respiratory ECMO use increased significantly during the COVID-19 pandemic. As of July 2022, almost 15,000 COVID-19 cases are reported on the ELSO Dashboard, eclipsing the total number of pre-pandemic ECMO cannulations.
      The widespread use of this expensive resource, especially during a pandemic, has led to important questions about appropriate indications and the ideal patient population when there is high demand.
      • Karagiannidis C.
      • Bein T.
      • Welte T.
      ECMO during the COVID-19 pandemic: moving from rescue therapy to more reasonable indications.
      ,
      • Supady A.
      • Biever P.M.
      • Staudacher D.L.
      • Wengenmayer T.
      Choosing the right reference cohort for assessing outcome of venovenous ECMO.
      One emerging theme of published data has been the effect of age on survival. In our patient cohort, age was a significant predictor of mortality, with an age of 50 years or older increasing the adjusted risk of death by 30%. Crude mortality was close to 50% for patients under the age of 50 but nearly 75% for those over the age of 50. Hall et al. found similar findings in their multi-institutional US study of 505 COVID-19 ECMO patients, showed that the median age was significantly lower in survivors, 44 vs. 51 years (p < 0.001).
      • Hall C.A.
      • Jacobs J.P.
      • Stammers A.H.
      • et al.
      Multi-institutional analysis of 505 patients with coronavirus disease-2019 supported with extracorporeal membrane oxygenation: predictors of survival.
      A similar multi-institutional study from Paris of 302 COVID-19 patients also demonstrated that approximately 50 years of age was a meaningful age threshold for significantly increased mortality.
      • Lebreton G.
      • Schmidt M.
      • Ponnaiah M.
      • et al.
      Extracorporeal membrane oxygenation network organisation and clinical outcomes during the COVID-19 pandemic in Greater Paris, France: a multicentre cohort study.
      Lastly, an early analysis of the first 1035 patients in the ELSO registry provided evidence that older age was associated with higher mortality, with every age group over 50 having a higher hazard ratio for mortality compared to patients under the age of 40 years.
      • Barbaro R.P.
      • MacLaren G.
      • Boonstra P.S.
      • et al.
      Extracorporeal membrane oxygenation support in COVID-19: an international cohort study of the Extracorporeal Life Support Organization registry.
      Current ELSO guidelines for adult respiratory ECMO do not include age as a contraindication for cannulation but acknowledge that older age increases the risk of death.
      • Tonna J.E.
      • Abrams D.
      • Brodie D.
      • et al.
      Management of adult patients supported with venovenous extracorporeal membrane oxygenation (VV ECMO): guideline from the extracorporeal Life support organization (ELSO).
      With the currently available evidence, ECMO clinicians must give substantial weight to a patient's age starting at 50 years, especially in the presence of other risk factors.
      In addition to age, seven or more total days of respiratory support had a similar association with mortality in our study. Patients who died also had a longer time from diagnosis to cannulation and from hospitalization to cannulation, although this was not significant in our model. While there is limited published data on this relationship for COVID-19 patients placed on VV ECMO, another US institutional study showed that mortality was 100% in patients that were mechanically ventilated for greater than 7 days and who did not receive a lung transplant.
      • Kurihara C.
      • Manerikar A.
      • Gao C.A.
      • et al.
      Outcomes after extracorporeal membrane oxygenation support in COVID-19 and non-COVID-19 patients.
      Lebreton et al. also demonstrated a significant relationship between the duration of intubation and mortality, finding that each decrease in the number of ventilation days had an odds ratio of 0.91 (95% CI 0.84, 0.99).
      • Lebreton G.
      • Schmidt M.
      • Ponnaiah M.
      • et al.
      Extracorporeal membrane oxygenation network organisation and clinical outcomes during the COVID-19 pandemic in Greater Paris, France: a multicentre cohort study.
      Unfortunately, during the COVID-19 pandemic, calculating ventilation days is often more complicated due to a very high proportion of patients managed with prolonged NIPPV prior to intubation. Our study found that the duration of respiratory support, including NIPPV and mechanical ventilation, significantly predicted mortality over seven days. Current ELSO guidelines recommend using seven days of high-setting ventilatory support as a relative contraindication for ECMO support but do not include NIPPV.
      • Tonna J.E.
      • Abrams D.
      • Brodie D.
      • et al.
      Management of adult patients supported with venovenous extracorporeal membrane oxygenation (VV ECMO): guideline from the extracorporeal Life support organization (ELSO).
      However, our data suggest that the duration of NIPPV should also be considered, especially in higher risk patients.
      Pre-cannulation renal failure requiring dialysis is not a contraindication to ECMO support and was not considered in decision-making by our center during the pandemic. However, it was a significant predictor of mortality in our cohort, independent of age or respiratory support duration. A recent German study of COVID-19 ECMO patients found that renal failure was much more common among non-survivors compared to survivors at 73% versus 33% (p < 0.001).
      • Natanov R.
      • Kunkel E.R.
      • Wiesner O.
      • et al.
      Determinants of survival in patients on extracorporeal membrane oxygenation therapy due to severe covid-19.
      In addition, data from France also found that patients with a lower renal component of the Sequential Organ Failure Assessment (SOFA) score pre-ECMO had a lower risk of mortality.18 Although dialysis can easily be provided to patients while on ECMO, pre-cannulation renal failure may signify greater systemic disease severity and affect prognosis. There is insufficient evidence to include renal failure as a contraindication to ECMO, but like older age, it should be considered in the presence of other adverse prognostic factors, especially if resources are limited. Consistent with other recent reports, BMI was not associated with mortality in our cohort and should not be considered a contraindication for ECMO in COVID-19 patients.

      Powell E.K., Haase D.J., Lankford A., Boswell K., Esposito E., Hamera J., Dahi S., Krause E., Bittle G., Deatrick K.B., Young B.A.C., Galvagno S.M. Jr, Tabatabai A. Body mass index does not impact survival in COVID-19 patients requiring veno-venous extracorporeal membrane oxygenation. Perfusion. 2022 Apr 25:2676591221097642. doi: 10.1177/02676591221097642. Epub ahead of print. PMID: 35467981.

      • Ramanathan K.
      • Shekar K.
      • Ling R.R.
      • et al.
      Extracorporeal membrane oxygenation for COVID-19: a systematic review and meta-analysis.
      • Mustafa A.K.
      • Joshi D.J.
      • Alexander P.J.
      • et al.
      Comparative propensity matched outcomes in severe COVID-19 respiratory failure-extracorporeal membrane oxygenation or maximum ventilation alone.
      Reported outcomes after ECMO for COVID-19 have been variable. While the ELSO registry reported mortality of approximately 40% for VV ECMO patients in 2020, recent large single-institution studies have shown much greater variability. These reports have ranged from 54% mortality in France18, 50% in Germany,
      • Natanov R.
      • Kunkel E.R.
      • Wiesner O.
      • et al.
      Determinants of survival in patients on extracorporeal membrane oxygenation therapy due to severe covid-19.
      46% in Italy,
      • Fanelli V.
      • Giani M.
      • Grasselli G.
      • et al.
      Extracorporeal membrane oxygenation for COVID-19 and influenza H1N1 associated acute respiratory distress syndrome: a multicenter retrospective cohort study.
      74% in Poland,
      • Trejnowska E.
      • Drobiński D.
      • Knapik P.
      • et al.
      Extracorporeal membrane oxygenation for severe COVID-19-associated acute respiratory distress syndrome in Poland: a multicenter cohort study.
      and 60% at our center. Differences in patient characteristics during various pandemic periods may explain the heterogeneity of published ECMO-associated outcomes. For example, early in the pandemic, our patient cohort's median age was significantly higher than in the Delta period, consistent with multi-center data from France.
      • Schmidt M.
      • Langouet E.
      • Hajage D.
      • et al.
      Evolving outcomes of extracorporeal membrane oxygenation support for severe COVID-19 ARDS in Sorbonne hospitals, Paris.
      Pre-vaccination, older patients comprised a larger proportion of patients with COVID-19-associated ARDS, which may have biased some centers towards placing relatively older patients on ECMO before evidence on clinical outcomes was available. In contrast, other centers may have been more cautious due to limited resources.
      • Banerjee A.
      • Pasea L.
      • Harris S.
      • et al.
      Estimating excess 1-year mortality associated with the COVID-19 pandemic according to underlying conditions and age: a population-based cohort study.
      In addition to the strong correlation between age and survival, we also found that older patients did not tolerate a longer ECMO duration due to multi-organ failure and sepsis. Our data suggest that a longer duration of ECMO may have benefited some patients later in the pandemic and that firm cut-offs for ECMO duration for COVID-19 should be re-examined.
      • Dreier E.
      • Malfertheiner M.V.
      • Dienemann T.
      • et al.
      ECMO in COVID-19—prolonged therapy needed? A retrospective analysis of outcome and prognostic factors.
      Differences in COVID-19 variant characteristics, including clinical severity and rapidly changing population immunity due to vaccination and infection, make comparisons across different pandemic periods difficult.
      • Lauring A.S.
      • Tenforde M.W.
      • Chappell J.D.
      • et al.
      Clinical severity of, and effectiveness of mRNA vaccines against, covid-19 from omicron, delta, and alpha SARS-CoV-2 variants in the United States: prospective observational study.
      It also made patient selection more challenging due to changing demographics of patients affected by the virus. Outside of age and ECMO duration, the primary differences in patient characteristics in our cohort were race and insurance status, but neither contributed to mortality in our analysis. Early in the pandemic, African Americans and Hispanic Whites comprised a much higher proportion of our ECMO cohort compared to later periods which shifted towards insured, non-Hispanic Whites. This finding is consistent with published US data that minority racial and ethnic groups were disproportionally affected by COVID-19, especially in 2020, with a higher risk of infection, hospitalization, and death.
      • Romano S.D.
      • Blackstock A.J.
      • Taylor E.V.
      • et al.
      Trends in racial and ethnic disparities in COVID-19 hospitalizations, by region - United States, march-december 2020.
      The other notable difference was the dramatic decrease in ECMO volume during the Omicron period. Despite rapidly increasing case numbers, the number of patients cannulated for ECMO during the Omicron period was relatively low, with several patients confirmed to be infected with a Delta variant. While our Omicron sample is small, our findings are consistent with evidence that Omicron has a higher case-to-hospitalization ratio than previous variants and may lead to substantially lower rates of ARDS.
      • Madhi S.A.
      • Kwatra G.
      • Myers J.E.
      • et al.
      Population immunity and covid-19 severity with omicron variant in South Africa.
      ,
      • Ulloa A.C.
      • Buchan S.A.
      • Daneman N.
      • Brown K.A.
      Estimates of SARS-CoV-2 omicron variant severity in ontario, Canada.
      At high volume ECMO centers like ours, ECMO utilization may be a helpful indicator of disease severity as the pandemic evolves.
      Our study is limited by its retrospective, single-institution design, which may limit generalizability to other centers. Patient selection criteria were not formally changed during the pandemic, but our results are biased by the patients referred to our center, which trended younger during the Delta period. However, this is one of the largest single-institution cohorts published and restricting our data to our center allowed for a controlled sample of patients with management consistency throughout the pandemic with a complete data set. Working at a single institution also allowed us to partner with the clinical laboratory to perform viral genomic sequencing, which was especially important during the Delta to Omicron transition. Our sample size may limit the ability to identify less common risk factors for mortality.

      5. Conclusion

      In this large cohort of VV ECMO patients with ARDS secondary to COVID-19, age ≥50 years, a duration of 7 or more days of total respiratory support, and pre-ECMO renal failure requiring dialysis all increased the risk of mortality by approximately 30%. More data is needed on the appropriate duration of ECMO support for COVID-19, especially for younger patients, as some patients appeared to benefit from an extended duration later in the pandemic. Careful consideration of ECMO candidates older than 50 years or with a prolonged course of critical illness is imperative.

      Funding

      SARS-CoV-2 genomic sequencing was funded by the NC Policy Collaboratory.

      Declaration of competing interest

      The authors have no conflict of interest to disclose. The authors have no financial relationships to disclose.

      References

        • Zangrillo A.
        • Landoni G.
        • Biondi-Zoccai G.
        • et al.
        A meta-analysis of complications and mortality of extracorporeal membrane oxygenation.
        Crit Care Resusc. 2013 Sep; 15 (PubMed PMID: 23944202. eng): 172-178
        • Bharat A.
        • Machuca T.N.
        • Querrey M.
        • et al.
        Early outcomes after lung transplantation for severe COVID-19: a series of the first consecutive cases from four countries.
        Lancet Respir Med. 2021 May; 9 (PubMed PMID: 33811829. Pubmed Central PMCID: PMC8012035. Epub 20210331. eng): 487-497
        • Henry B.M.
        COVID-19, ECMO, and lymphopenia: a word of caution.
        Lancet Respir Med. 2020 Apr; 8 (PubMed PMID: 32178774. Pubmed Central PMCID: PMC7118650. Epub 20200313. eng): e24
        • Combes A.
        • Hajage D.
        • Capellier G.
        • et al.
        Extracorporeal membrane oxygenation for severe acute respiratory distress syndrome.
        N Engl J Med. 2018 May 24; 378 (PubMed PMID: 29791822. eng): 1965-1975
        • Barbaro R.P.
        • MacLaren G.
        • Boonstra P.S.
        • et al.
        Extracorporeal membrane oxygenation support in COVID-19: an international cohort study of the Extracorporeal Life Support Organization registry.
        Lancet. 2020 Oct 10; 396 (PubMed PMID: 32987008. Pubmed Central PMCID: PMC7518880. Epub 20200925. eng): 1071-1078
        • Shaefi S.
        • Brenner S.K.
        • Gupta S.
        • et al.
        Extracorporeal membrane oxygenation in patients with severe respiratory failure from COVID-19.
        Intensive Care Med. 2021 Feb; 47 (PubMed PMID: 33528595. Pubmed Central PMCID: PMC7851810. Epub 20210202. eng): 208-221
        • Horby P.
        • Lim W.S.
        • Emberson J.R.
        • et al.
        Dexamethasone in hospitalized patients with covid-19.
        N Engl J Med. 2021 Feb 25; 384 (PubMed PMID: 32678530. Pubmed Central PMCID: PMC7383595. Epub 20200717. eng): 693-704
        • Beigel J.H.
        • Tomashek K.M.
        • Dodd L.E.
        • et al.
        Remdesivir for the treatment of covid-19 - final report.
        N Engl J Med. 2020 Nov 5; 383 (PubMed PMID: 32445440. Pubmed Central PMCID: PMC7262788. Epub 20201008. eng): 1813-1826
      1. Raff L.A., Gallaher J.R., Johnson D., Raff E.J., Charles A.G., Reid T.S. Time to Cannulation after ICU Admission Increases Mortality for Patients Requiring Veno-venous ECMO for COVID-19 Associated Acute Respiratory Distress Syndrome. Ann Surg. 2022 Dec 1;276(6):e659-e663. doi: 10.1097/SLA.0000000000004683. Epub 2020 Dec 22. PMID: 33630477.

        • D'Hoore W.
        • Bouckaert A.
        • Tilquin C.
        Practical considerations on the use of the Charlson comorbidity index with administrative data bases.
        J Clin Epidemiol. 1996; 49: 1429-1433
        • Hospitals U.
        UNC Surveillance Report.
        2022 ([cited 2022 6 Aug 2022]; Available from:)
        • Zou G.
        A modified Poisson regression approach to prospective studies with binary data.
        Am J Epidemiol. 2004; 159: 702-706
        • Chen W.
        • Qian L.
        • Shi J.
        • Franklin M.
        Comparing performance between log-binomial and robust Poisson regression models for estimating risk ratios under model misspecification.
        BMC Med Res Methodol. 2018; 18: 63
      2. Registry Dashboard-COVID-19 Cases on ECMO in the ELSO Registry. 2022 ([database on the Internet]. Available from:)
        • Karagiannidis C.
        • Bein T.
        • Welte T.
        ECMO during the COVID-19 pandemic: moving from rescue therapy to more reasonable indications.
        Eur Respir J. 2022 Feb; 59 (PubMed PMID: 35115345. Pubmed Central PMCID: PMC8828992. Epub 20220210. eng)
        • Supady A.
        • Biever P.M.
        • Staudacher D.L.
        • Wengenmayer T.
        Choosing the right reference cohort for assessing outcome of venovenous ECMO.
        Crit Care. 2022 2022/01/10; 26: 17
        • Hall C.A.
        • Jacobs J.P.
        • Stammers A.H.
        • et al.
        Multi-institutional analysis of 505 patients with coronavirus disease-2019 supported with extracorporeal membrane oxygenation: predictors of survival.
        Ann Thorac Surg. 2022 Jul; 114 (PubMed PMID: 35189111. Pubmed Central PMCID: PMC8855605. Epub 20220218. eng): 61-68
        • Lebreton G.
        • Schmidt M.
        • Ponnaiah M.
        • et al.
        Extracorporeal membrane oxygenation network organisation and clinical outcomes during the COVID-19 pandemic in Greater Paris, France: a multicentre cohort study.
        Lancet Respir Med. 2021 Aug; 9 (PubMed PMID: 33887246. Pubmed Central PMCID: PMC8055207. Epub 20210419. eng): 851-862
        • Barbaro R.P.
        • MacLaren G.
        • Boonstra P.S.
        • et al.
        Extracorporeal membrane oxygenation support in COVID-19: an international cohort study of the Extracorporeal Life Support Organization registry.
        Lancet. 2020; 396 (2020/10/10/): 1071-1078
        • Tonna J.E.
        • Abrams D.
        • Brodie D.
        • et al.
        Management of adult patients supported with venovenous extracorporeal membrane oxygenation (VV ECMO): guideline from the extracorporeal Life support organization (ELSO).
        Am Soc Artif Intern Organs J. 2021; 67 (PubMed PMID: 00002480-202106000-00001): 601-610
        • Kurihara C.
        • Manerikar A.
        • Gao C.A.
        • et al.
        Outcomes after extracorporeal membrane oxygenation support in COVID-19 and non-COVID-19 patients.
        Artif Organs. 2022 Apr; 46 (PubMed PMID: 34694655. Pubmed Central PMCID: PMC8653196. Epub 20211104. eng): 688-696
        • Natanov R.
        • Kunkel E.R.
        • Wiesner O.
        • et al.
        Determinants of survival in patients on extracorporeal membrane oxygenation therapy due to severe covid-19.
        Perfusion. 2022; 02676591221113135
      3. Powell E.K., Haase D.J., Lankford A., Boswell K., Esposito E., Hamera J., Dahi S., Krause E., Bittle G., Deatrick K.B., Young B.A.C., Galvagno S.M. Jr, Tabatabai A. Body mass index does not impact survival in COVID-19 patients requiring veno-venous extracorporeal membrane oxygenation. Perfusion. 2022 Apr 25:2676591221097642. doi: 10.1177/02676591221097642. Epub ahead of print. PMID: 35467981.

        • Ramanathan K.
        • Shekar K.
        • Ling R.R.
        • et al.
        Extracorporeal membrane oxygenation for COVID-19: a systematic review and meta-analysis.
        Crit Care. 2021 2021/06/14; 25: 211
        • Mustafa A.K.
        • Joshi D.J.
        • Alexander P.J.
        • et al.
        Comparative propensity matched outcomes in severe COVID-19 respiratory failure-extracorporeal membrane oxygenation or maximum ventilation alone.
        Ann Surg. 2021 Nov 1; 274 (PubMed PMID: 34617934. Pubmed Central PMCID: PMC8500214. eng): e388-e394
        • Fanelli V.
        • Giani M.
        • Grasselli G.
        • et al.
        Extracorporeal membrane oxygenation for COVID-19 and influenza H1N1 associated acute respiratory distress syndrome: a multicenter retrospective cohort study.
        Crit Care. 2022 2022/02/05; 26: 34
        • Trejnowska E.
        • Drobiński D.
        • Knapik P.
        • et al.
        Extracorporeal membrane oxygenation for severe COVID-19-associated acute respiratory distress syndrome in Poland: a multicenter cohort study.
        Crit Care. 2022 Apr 7; 26 (PubMed PMID: 35392960. Pubmed Central PMCID: PMC8988534. Epub 20220407. eng): 97
        • Schmidt M.
        • Langouet E.
        • Hajage D.
        • et al.
        Evolving outcomes of extracorporeal membrane oxygenation support for severe COVID-19 ARDS in Sorbonne hospitals, Paris.
        Crit Care. 2021 Oct 9; 25 (PubMed PMID: 34627350. Pubmed Central PMCID: PMC8502094. Epub 20211009. eng): 355
        • Banerjee A.
        • Pasea L.
        • Harris S.
        • et al.
        Estimating excess 1-year mortality associated with the COVID-19 pandemic according to underlying conditions and age: a population-based cohort study.
        Lancet. 2020; 395 (2020/05/30/): 1715-1725
        • Dreier E.
        • Malfertheiner M.V.
        • Dienemann T.
        • et al.
        ECMO in COVID-19—prolonged therapy needed? A retrospective analysis of outcome and prognostic factors.
        Perfusion. 2021; 36: 582-591
        • Lauring A.S.
        • Tenforde M.W.
        • Chappell J.D.
        • et al.
        Clinical severity of, and effectiveness of mRNA vaccines against, covid-19 from omicron, delta, and alpha SARS-CoV-2 variants in the United States: prospective observational study.
        BMJ. 2022; 376e069761
        • Romano S.D.
        • Blackstock A.J.
        • Taylor E.V.
        • et al.
        Trends in racial and ethnic disparities in COVID-19 hospitalizations, by region - United States, march-december 2020.
        MMWR Morb Mortal Wkly Rep. 2021 Apr 16; 70 (PubMed PMID: 33857068. Pubmed Central PMCID: PMC8344991. Epub 20210416. eng): 560-565
        • Madhi S.A.
        • Kwatra G.
        • Myers J.E.
        • et al.
        Population immunity and covid-19 severity with omicron variant in South Africa.
        N Engl J Med. 2022; 386 (PubMed PMID: 35196424): 1314-1326
        • Ulloa A.C.
        • Buchan S.A.
        • Daneman N.
        • Brown K.A.
        Estimates of SARS-CoV-2 omicron variant severity in ontario, Canada.
        JAMA. 2022; 327: 1286-1288