Extracorporeal membrane oxygenation in covid 19 related acute respiratory distress syndrome a euro elso international survey

Lepper1,2 for the COVEC-Study 14 2 Department of Internal Medicine V - Pneumology, Allergology and Critical Care Medicine, 15 University Medical Centre, Saarland University, Homburg/Saa

2 Respiratory Distress Syndrome – a EuroELSO international survey

3

4 Sebastian Mang1,2, Armin Kalenka3, Lars Mikael Broman4, Alexander Supady5,

5 Justyna Swol6, Guy Danziger1,2, André Becker1,2, Sabrina I Hörsch1,7, Thilo

6 Mertke1,7, Ralf Kaiser1,2, Hendrik Bracht8, Viviane Zotzmann5, Frederik Seiler1,2,

7 Robert Bals1,2, Fabio Silvio Taccone9, Onnen Moerer10, Roberto Lorusso11, Jan

8 Bělohlávek12, Ralf M Muellenbach13, and Philipp M Lepper1,2 for the COVEC-Study

14 2 Department of Internal Medicine V - Pneumology, Allergology and Critical Care Medicine,

15 University Medical Centre, Saarland University, Homburg/Saar, Germany

16 3 Department of Anaesthesiology and Intensive Care Medicine, District Hospital Bergstrasse,

17 University Hospital Heidelberg, Heppenheim, Germany

18 4 ECMO Centre Karolinska, Department of Pediatric Perioperative Medicine and Intensive

19 Care, Karolinska University Hospital, Stockholm, Sweden

20 5 Department of Medicine III (Interdisciplinary Medical Intensive Care), Medical Center –

21 University of Freiburg, Faculty of Medicine, University of Freiburg, Germany

22 6 Department of Pneumology, Allergology and Sleep Medicine, and Intensive Care Medicine,

23 Paracelsus Medical University, General Hospital Nuremberg, Nuremberg, Germany

24 7 Department of Anaesthesiology, Critical Care Medicine and Pain Medicine, University

25 Medical Centre, Saarland University, Homburg/Saar, Germany

26 8 Department of Anaesthesiology and Critical Care Medicine, University Hospital of Ulm,

27 Ulm, Germany

28 9 Department of Intensive Care, Erasme University Hospital, Université Libre de Bruxelles,

29 Brussels, Belgium

30 10 Department of Anaesthesiology, University Hospital of Göttingen, Göttingen, Germany

31 11 Cardio-Thoracic Surgery Department - Heart & Vascular Centre - Maastricht University

32 Medical Centre, Maastricht, Netherlands

33 12 2 nd Department of Internal Cardiovascular Medicine, General University Hospital, Prague,

34 Czech Republic

35 13 Department of Anaesthesiology and Critical Care Medicine, Campus Kassel of the

36 University of Southampton, Kassel, Germany

49 Correspondence to:

50 Philipp M Lepper, MD

51 Department of Internal Medicine V – Pneumology, Allergology and Intensive Care

52 Medicine and ECLS Center Saar

53 University Hospital of Saarland

60 Background: Extracorporeal membrane oxygenation (ECMO) is a means to

61 support patients with acute respiratory failure Initially, recommendations to treat

62 severe cases of pandemic Coronavirus Disease 2019 (COVID-19) with ECMO have

63 been restrained In the meantime, ECMO has been shown to produce similar

64 outcomes in patients with severe COVID-19 compared to existing data on ARDS

65 mortality

66 Objective: We performed an international email survey to assess how ECMO

67 providers worldwide have previously used ECMO during the treatment of critically ill

68 patients with COVID-19

69 Methods: A questionnaire with 45 questions (covering e.g indication,

70 technical aspects, benefit and reasons for treatment discontinuation), mostly

71 multiple-choice, was distributed by email to ECMO centers The survey was

72 approved by the European branch of the Extracorporeal Life Support Organization

74 Results: 276 centers worldwide responded that they employed ECMO for

75 very severe COVID-19 cases, mostly in veno-venous configuration (87%) The most

76 common reason to establish ECMO was isolated hypoxemic respiratory failure

77 (50%), followed by a combination of hypoxemia and hypercapnia (39%) Only a

78 small fraction of patients required veno-arterial cannulation due to heart failure (3%)

79 Time on ECMO varied between less than two and more than four weeks The main

80 reason to discontinue ECMO treatment prior to patient’s recovery was lack of clinical

81 improvement (53%), followed by major bleeding, mostly intracranially (13%) Only

82 4% of respondents reported that triage situations, lack of staff or lack of oxygenators

83 were responsible for discontinuation of ECMO support Most ECMO physicians (66%

84  26%) agreed that patients with COVID-19 induced ARDS (CARDS) benefitted from

85 ECMO Overall mortality of COVID-19 patients on ECMO was estimated to be about

86 55%, scoring higher than what has previously been reported for Influenza patients on

87 ECMO (29 – 36%)

88 Conclusion: ECMO has been utilized successfully during the COVID-19

89 pandemic to stabilize CARDS patients in hypoxemic or hypercapnic lung failure Age

90 and multimorbidity limited the use of ECMO Triage situations were rarely a concern

91 ECMO providers stated that patients with severe COVID-19 benefitted from ECMO

Preprint not peer reviewed

92 An increasing use in patients with respiratory failure in a future stage of the

93 pandemic may be expected

Preprint not peer reviewed

94 Introduction

95

96 Early in 2020, countries worldwide have been facing a surge of patients with

97 acute respiratory distress syndrome (ARDS) due to pandemic Severe Acute

98 Respiratory Syndrome Coronavirus-2 (SARS-CoV-2) disease 2019 (COVID-19)

99 Survival of those most severely affected by COVID-19-related ARDS (CARDS) might

100 depend on extracorporeal membrane oxygenation (ECMO) as bridge to recovery1

101 In this global pandemic, hospitals and healthcare systems have been pushed

102 to the verge of collapse During the first phase of the COVID-19 pandemic, the

103 number of critically ill patients requiring invasive ventilation often exceeded ventilator

104 capacities, creating a need for ICU triage2 In this scenario, it was highly unlikely that

105 ECMO would be broadly recommended to critical care providers to treat COVID-19,

106 given its high demands on personnel and resources3 In its initial guidance

107 document, ELSO considered to offer ECMO only to specific patients not responding

108 to maximal conventional therapy, including proning and neuromuscular blockade4

109 Additionally, early reports suggested mortality rates could be higher than 90% in

110 COVID-19 patients supported with ECMO5

111 A recent trial reported that ECMO reduced 60-day mortality in non-COVID-19

112 related ARDS to 35% in the ECMO group versus 46% in the conventional

113 management group (relative risk 0.76, 95% CI 0.55–1.04; p=0.09)6 The study

114 highlighted that ECMO can facilitate protective ventilation of ARDS patients with

115 reduced tidal volumes, plateau and driving pressures, mostly due to effective

116 extracorporeal CO2 removal CARDS might not differ as much from non-COVID

117 ARDS as was previously expected7 Physiological considerations make it thus

118 reasonable to think about ECMO as a bail-out strategy in critically ill patients with

119 CARDS A recently published retrospective data suggested that mortality of patients

120 with CARDS receiving ECMO might be comparable to past ARDS cohorts8 Given

121 that COVID-19 pathophysiology is still poorly understood, little is currently known

122 about how to tailor ECMO treatment to meet COVID-19 specific challenges, e.g

123 hypercoagulable state9, or how long ECMO should be continued when patients fail to

124 improve

125 We therefore designed an online survey to elicit how ECMO providers

126 worldwide have previously employed ECMO to treat critically ill COVID-19 patients Preprint not peer reviewed

127 Our survey was approved by the European branch of the Extracorporeal Life Support

128 Organization (EuroELSO)

Preprint not peer reviewed

129 Methods

130

131 We created a questionnaire consisting of 45 questions and distributed it to

132 4,193 physicians that had published on an ECMO-related topic since the year 2000

133 in a PubMed-listed journal with an available E-Mail using a commercially available

134 internet survey platform (SurveyMonkey Inc., San Mateo, California)

135 The ethical committee (Ärztekammer des Saarlandes) waived the need for a

136 formal approval since the questionnaire did not retrieve actual patient data

137

139

140 The questionnaire was composed of two sections: the first dealt with general

141 questions regarding contact information, details on hospital and ICU capacity as well

142 as years of ECMO experience The second part was designed to elicit most common

143 indications for ECMO use in COVID-19, details about ECMO circuit configuration as

144 well as complications and reasons for possible treatment discontinuation We did not

145 ask for any patient-specific data For conformity reasons and to facilitate participation

146 in the survey, most of the questions were multiple choice with two to nine possible

147 answers per question The last eight items requested the participant to express his

148 extent of agreement with a specific statement about ECMO therapy in the context of

149 COVID-19 on a visual analogue scale The survey is partly available in the

150 Supplementary Material

151 The survey questions and multiple-choice responses with their respective

152 organization in the different sections were circulated and consented between a group

153 of 23 very experienced physicians in this field When consensus of all questions and

154 answers was reached, the survey was transferred to an online platform

155 (SurveyMonkey Inc., San Mateo, California) Automatic data retrieval and descriptive

156 statistics were retrieved through this platform More than one answer from centers

157 were possible This was allowed, as many centers comprise several departments

158 with physicians from different backgrounds (e.g anesthesiology and surgery)

159 Results from multiple-choice questions are expressed in median, participants’

160 extent of agreement or disagreement in mean and standard deviation in percent The

161 survey was launched on 8Preprint not peer reviewedth, deadline for return was 20th of June, 2020 Final

162 analysis of results was performed using an extrapolation tool provided by

163 SurveyMonkey as well as SPSS

164 Participants were given the opportunity to be listed as collaborators Those

165 participants who did not supply hospital or contact information or who did not

166 complete the survey could not be included in the list of collaborators

167

169

170 General data on ECMO centers and treatment capacities

171 276 ECMO professionals from 98 centers in 30 different countries on four

172 continents (North America, South America, Europe, Asia) responded to the survey,

173 yielding a response rate of 6.6% Sixty-four percent of responding centers were

174 ELSO members As it was possible to skip questions, sometimes the denominator is

175 less than 276 In this case, the number of respondents is given in brackets

176 Centers’ median number of years with ECMO experience was 14, mainly in ECMO

177 treatment of adults or adult and pediatric patients (85%) Only 1.3% of participants

178 were exclusively specialized in neonatal ECMO Most common numbers of patients

179 supported with ECMO per year prior to COVID-19 in the participants’ centers ranged

180 from 21 to 50, 13% of centers having even supported more the 100 patients on

181 veno-venous ECMO (VV-ECMO) per year prior to the pandemic

182

183 Numbers of patients with COVID-19 with or without ECMO

184 The majority of ECMO providers (30%) stated that 2 to 6 patients with

COVID-185 19 had received ECMO in their center, 85% of all centers having supported a

186 maximum of 15 patients on ECMO by survey deadline ECMO treatment had mostly

187 been initiated in the participants’ hospitals (63%), only a minority of patients was

188 retrieved on ECMO by mobile ECMO retrieval teams from other hospitals

189

190 Indication for ECMO and circuit configuration

191 The most common reason to initiate ECMO for COVID-19 was isolated

192 hypoxemia (50%), followed by a combination of hypoxemia and hypercapnia (39%)

193 Isolated hypercapnia was rarely a reason to cannulate a patient (3%) Only 6%

194 stated that ECMO was started to facilitate lung-protective ventilation (n = 105) The

195 majority of ECMO cannulations (88%) were performed in veno-venous configuration Preprint not peer reviewed

196 Eight percent of centers used veno-arterial configuration (VA-ECMO) in one or more

197 patients and 3% had to extend to a V-AV-circuit in at least one case (one venous

198 draining cannula, one arterial returning cannula, one venous returning cannula) In

199 those cases, where an arterial cannulation was required, the indications were

200 specified as biventricular failure (n = 2) and, in one case, right heart failure due to

201 pulmonary embolism (n = 1) See also Figure 1, panel A and B

202

203 Anticoagulation management

204 Targeting anticoagulation therapy in patients with COVID-19 on ECMO, 60%

205 of participants (n = 110) stated that they did not change their standard

206 anticoagulation strategy compared to cases of ARDS due to other causes Forty

207 percent used higher doses of anticoagulants than usual, monitored by higher

208 prothrombin time or higher activated clotting time Only one of 110 ECMO providers

209 stated that they deliberately used lower doses of anticoagulants than usual for

210 ECMO in COVID-19 Antiplatelet therapy was also rarely used (1%) to prevent

211 clotting The details of the anticoagulants or antiplatelet agents administered were

212 not part of the survey See also Figure 1 C.

213

214 Reasons to abstain from ECMO

215 The two main reasons to refrain from ECMO initiation were patient age (74%)

216 and comorbidities (85%, not further specified) 28% of participants stated that ECMO

217 was withdrawn due to a patient’s known or suspected wishes Nine percent decided

218 against ECMO because it was not actively recommended for COVID-19 induced

219 ARDS by responsible scientific societies at that time Seven percent reported that a

220 surge of COVID-19 patients and overwhelming workload made ECMO impracticable

221 Only 5% of participants reported that they had to abstain from ECMO initiation due to

222 a shortage of oxygenators, machines or ECMO cannulas

Preprint not peer reviewed

223 Duration of ECMO support

224 Most patients were supported with ECMO for less than two weeks However,

225 50% of all participants stated that they had also treated patients with ECMO for more

226 than four weeks (Figure 1 D).

227

228 Reasons for ECMO discontinuation

229 Seventy-two percent of participants confirmed that their center would

230 withdraw ECMO if there was no perspective for a COVID-19 patient to recover If

231 ECMO treatment was discontinued prior to recovery, futility was mostly stated as the

232 reason (n = 50 from 94 responses, 53%) ECMO-related complications were the

233 second most important reason for treatment discontinuation Fourteen percent of

234 ECMO providers stated that they had terminated ECMO due to major bleeding (n =

235 15), mainly intracranial hemorrhage (n = 13) and, less frequently, extracranially In

236 1% of cases, not further specified technical issues led to ECMO withdrawal The

237 question also offered ‘lack of staff’ as a possible answer, which was not chosen

238 However, 2% of participants (n = 2, Germany and France) stated that a triage

239 situation forced physicians to discontinue ECMO prior to the patient’s possible

240 recovery Two percent of respondents named lack of ECMO oxygenators, ECMO

241 machines or consumables as the reason for ECMO discontinuation See also Figure

242 1 E.

243

244 Estimation of patients’ outcome

245 When asked to estimate the percentage of patients who died while on ECMO

246 due to COVID-19, average mortality was estimated to be 55%, meaning that 45% of

247 patients had survived on ECMO at least until the end of the survey

248

249 ECMO providers’ opinions on COVID-19 and ECMO

250 The last eight questions were designed to investigate a participant’s opinion

251 on certain statements about ECMO and COVID-19, measured in percentage of

252 agreement Participants agreed to 63%  24% on average that patients were longer

253 on ECMO due to COVID-19 compared to other causes of ARDS The claim that

254 CARDS patients on ECMO required more sweep gas flow than what the individual

255 ECMO physician was used to was accepted by 58% Preprint not peer reviewed 25% The statement that

256 oxygenator change was more frequently required in CARDS patients on ECMO had

257 an acceptance rate of 54% on average  29% The claim that disturbed coagulation

258 in COVID-19 patients would make ECMO impossible was mostly rejected (24%

259 agreement  26%) The assumption that ECMO offers patients with COVID-19

260 induced ARDS a chance to recover found relatively strong acceptance (agreement

261 extent of 76%  25%), with 66%  26% agreement with the claim that CARDS

262 patients benefitted from ECMO therapy Box plots are displayed in Figure 2.

263

264 For results not outlined here, see Supplementary Material.

Preprint not peer reviewed

265 Discussion

266

267 Survey results prove that critical care providers worldwide have repeatedly

268 successfully used extracorporeal membrane oxygenation to support patients in

269 COVID-19 related respiratory failuredespite official recommendations were limited to

270 carefully singled-out patients At the time the survey was launched, data regarding

271 the outcome of COVID-19 patients treated with ECMO was limited to case reports

272 and small case series Recent retrospective cohort studies on COVID-19 patients

273 receiving ECMO have produced encouraging results8 but were unpublished at that

274 time Critical care providers hence had to rely on what was already known about

275 ARDS from different causes Both clinical experience and pathophysiology of

276 CARDS made ECMO appealing to treat COVID-19, even when data on mortality or

277 complication rate was missing

278 Several groups have reported evidence that COVID-19 might be associated

279 with a hypercoagulable state, resulting in an increased risk of thromboembolic

280 complications11 Exposing hypercoagulable blood to the artificial surface of an ECMO

281 circuit could therefore come with a considerable risk of thromboembolism

282 Interestingly, when asked about ECMO-related complications in COVID-19, 13% of

283 ECMO providers stated that ECMO had to be terminated due to major bleeding, but

284 only one case of relevant pulmonary embolism on VV-ECMO was reported It should

285 be noted though that minor thrombosis or thromboembolism could have remained

286 undetected during ECMO therapy, hence a reliable incidence of all thrombotic events

287 cannot be provided or was underestimated by the ECMO providers Nevertheless,

288 given that 50% of all participants had treated patients on ECMO for more than four

289 weeks, the overall occurrence of lethal thromboembolism on ECMO in COVID-19

290 was surprisingly low Only a minority of ECMO physicians said that they increased

291 the dosage of anticoagulants to prevent clotting on ECMO There was also no

292 general agreement that oxygenator change was required more frequently compared

293 to patients with ARDS from other causes Although increased rate of oxygenator

294 pump head thrombosis in COVID-19 has been reported12, suspected

295 hypercoagulation in COVID-19 did not seem to translate to a higher incidence of

life-296 threatening pulmonary or systemic thromboembolism in patients with COVID-19 on

297 ECMO Prospective studies should address how much anticoagulation for ECMO in Preprint not peer reviewed

298 COVID-19 is needed so that both the risk of thrombosis and hemorrhage can be

299 minimized

300 Although predominantly affecting the lungs, multiple organ involvement of

301 COVID-19 has been reported Cardiac complications in COVID-19 have recently

302 gathered broad attention Acute myocardial injury in the absence of macroscopic

303 coronary artery disease is suspected to occur in up to 10% of patients especially in

304 the critically ill13,14,15,16 While 11% of our participants reported that they employed

305 also veno-arterial ECMO for COVID-19, only 3% of participants specified that arterial

306 cannulation was required because of acute heart or circulatory failure, in one case

307 caused by fulminant pulmonary embolism So far, our survey results do not support

308 the hypothesis that severe cardiac involvement in COVID-19 translated to an

309 increased need for cardiac support (e.g with VA-ECMO) in patients with CARDS

310 A recently published study investigating outcomes of all adult patients with

311 CARDS treated with ECMO using a EuroELSO registry estimated overall 90-day

312 mortality to be less than 40%8 When asked about the mortality of COVID-19

313 patients on ECMO, the mean estimation in our survey was 55%, meaning that

314 roughly 45% could be weaned off extracorporeal support at survey deadline If this

315 data translates to real-life mortality, survival of COVID-19 patients on ECMO could

316 be comparable to non-COVID-19 induced ARDS At the time of report, no

317 prospective data on the outcome nor complication rate of ECMO therapy in

COVID-318 19 is available For the time being, our survey suggests that ECMO for COVID-19 is

319 practicable, effective and does not lead to higher complication rates in COVID-19 if it

320 is utilized in experienced ECMO centers Most participants agreed that their patients

321 benefitted from ECMO therapy and stated that, in a hypothetical scenario of a future

322 respiratory pandemic, they would use it more readily to treat acute respiratory failure

323

324 Our survey has several limitations Firstly, while it reflects opinions of a fairly

325 large cohort of physicians providing ECMO services, we nevertheless did not ask for

326 patient-specific data outcome Hence, we cannot provide statistical evidence on

327 ECMO-related end points, e.g overall survival, 28-day mortality or discharge from

328 ICU Secondly, the survey was addressed mainly to tertiary care centers who have

329 repeatedly published on ECMO, leading to a skewness of representation, possibly to

330 the disadvantage of primary and secondary care hospitals with less ECMO

333 Critical care providers need to know not only whether to use ECMO, but also

334 how to manage ECMO therapy in a disease as poorly understood as COVID-19

335 Currently, this knowledge is expanding8 Across Europe, more than 2,100 patients in

336 180 centers were already treated on ECMO for respiratory and/or circulatory failure

337 due to COVID-19 (https://www.euroelso.net/covid-19/covid-19-survey/; last accessed

338 on November 23rd) The data from this survey (theoretical considerations early in the

339 pandemia) and the EuroELSO (real world data with weekly updates) data are in line,

340 suggesting that ECMO is a useful adjunctive tool in COVID-19-related respiratory

341 failure

Preprint not peer reviewed

342 Declarations

343

344 Conflict of interest disclosure

345

346 Robert Bals declares funding from AstraZeneca, Boehringer Ingelheim,

347 GlaxoSmithKline, Grifols, Novartis, CLS Behring, the German Federal Ministry of

348 Education and Research (BMBF) Competence Network Asthma and COPD

349 (ASCONET), Sander-Stiftung, Schwiete-Stiftung, Krebshilfe and Mukoviszidose eV

350 Conflicts that the editors consider relevant to the content of the manuscript have

351 been disclosed All other authors: No potential conflicts of interest

352

353 Financial funding

354

355 COVID-19 research was funded by the Federal state of Saarland, Saarland

356 University and Dr Rolf M Schwiete Foundation

357

359

360 The authors are grateful to all colleagues worldwide for their generous

361 participation in our survey We thank the EuroELSO steering committee for

362 supporting the survey

363

364 Consent to be listed as collaborator

365

366 All colleagues who participated in the survey and agreed to be named publicly

367 were included as collaborators

368

369 Authorship Contribution

370

371 S.M., A.K., A.B., G.D., S.H., T.M and P.L drafted the survey L.M.B, J.S.,

372 H.B., F.S.T., O.M., R.L., J.B and R.M revised the survey for important content P.L

373 launched the survey and oversaw collection of data R.K provided IT-tools for survey

374 preparation S.M drafted manuscript A.K., L.M.B., A.S., J.S., G.D., S.H., F.S., R.B., Preprint not peer reviewed

375 F.T., O.M., R.L, J.B., R.M and P.L revised the manuscript for important intellectual

376 content All authors have seen and approved the final version of the manuscript

Preprint not peer reviewed

377 *The COVID-19 ECMO (COVEC) study group

378

379 Nicholas Barrett, Guy’s and St Thomas’ NHS Foundation Trust, London, UK;

380 Daniel Duerschmied, University Hospital of Freiburg, Freiburg, Germany;

381 Eddy Fan, Toronto General Hospital, Toronto, Canada;

382 Falk Fichtner, Universitätsklinikum Leipzig, Leipzig, Germany;

383 Hendrik Haake, Krankenhaus Maria Hilf, Mönchengladbach, Germany;

384 Frank Langer, University Medical Centre, Saarland University, Homburg/Saar,

386 Haitham Mutlak, Sana Klinikum Offenbach, Offenbach, Germany;

387 Markus Kredel, University Hospital of Würzburg, Würzburg, Germany;

388 Thomas Müller, University Hospital of Regensburg, Regensburg, Germany;

389 Alessandro Protti, Humanitas Research and Clinical Centre-IRCCS, Milan, Italy

390 Alexander Raddatz, University Medical Centre, Saarland University, Homburg/Saar,

392 Tobias Spangenberg, Marienkrankenhaus Hamburg, Hamburg, Germany;

393 Dawid Staudacher, University Hospital of Freiburg, Freiburg, Germany;

394 Holger Wehrfritz, University Medical Centre, Saarland University, Homburg/Saar,

396 Tobias Wengenmayer, University Hospital of Freiburg, Freiburg, Germany;

397 Arne Westheider, University Medical Centre, Saarland University, Homburg/Saar,

399

Preprint not peer reviewed