Supplementary Appendix

This appendix has been provided by the authors to give readers additional information about their work.

Supplement to: The REMAP-CAP Investigators. Interleukin-6 receptor antagonists in critically ill patients with Covid-19. N Engl J Med. DOI: 10.1056/NEJMoa2100433

1

Interleukin-6 Receptor Antagonists in Critically Ill patients with Covid-19

The REMAP-CAP Investigators

1. REMAP-CAP investigators .................................................................................................. 2

REMAP-CAP Trial Investigators & Collaborators.................................................................... 2

2. Additional methods ......................................................................................................... 19

2.1 Additional information about REMAP-CAP general design ........................................... 19

2.2 REMAP-CAP Eligibility Criteria for suspected or proven Covid-19 ................................ 20

Platform inclusion criteria:........................................................................................... 20

Platform exclusion criteria: .......................................................................................... 20

Immune Modulation Therapy domain specific inclusion criteria: ............................... 20

Immune Modulation Therapy domain specific exclusion criteria: .............................. 21

Intervention specific exclusion criteria: ....................................................................... 21

2.2 Additional Statistical Methods: ...................................................................................... 22

Secondary outcomes:................................................................................................... 22

Analysis populations: ................................................................................................... 22

3. Additional Results ................................................................................................................ 25

Site Participation in the Immune Modulation Domain................................................ 25

Platform exclusions: ..................................................................................................... 25

Additional analyses: ............................................................................................................. 26

4. References: ...................................................................................................................... 38

5. Appendices ....................................................................................................................... 39

a. Statistical Analysis Committee Primary Analysis Report ............................................. 39

b. ITSC Secondary Analysis Report ................................................................................... 39

Table S1. Baseline Characteristics of Participants in the Immune Modulation Therapy domain * .................................................................................................................................. 27 Table S2. Randomization into other REMAP-CAP domains and other treatments given ....... 29 Table S3. Sensitivity analyses of the Primary Outcomes ......................................................... 30 Table S4. Exploratory post-hoc sensitivity analyses of the Primary Outcomes....................... 31 Table S5. Odds ratios for secondary analysis of organ support free days for the unblinded ITT population ................................................................................................................................ 32 Table S6. Odds ratios for secondary analysis of hospital survival for the unblinded ITT population ................................................................................................................................ 33 Table S7. Secondary Outcomes ............................................................................................... 34 Table S8. Subgroup analysis of the Primary Outcomes ........................................................... 36 Table S9. Exploratory post-hoc subgroup analyses of the Primary Outcomes ....................... 37

2

1. REMAP-CAP investigators

REMAP-CAP Trial Investigators & Collaborators

International Trial Steering Committee: Farah Al-Beidh, Derek Angus, Djillali Annane, Yaseen Arabi, Abi Beane, Wilma van Bentum-Puijk, Scott Berry, Zahra Bhimani, Marc Bonten, Charlotte Bradbury, Frank Brunkhorst, Meredith Buxton, Allen Cheng, Lennie Derde, Lise Estcourt, Herman Goossens, Anthony Gordon, Cameron Green, Rashan Haniffa, Francois Lamontagne, Patrick Lawler, Edward Litton, John Marshall, Colin McArthur, Daniel McAuley, Shay McGuinness, Bryan McVerry, Stephanie Montgomery, Paul Mouncey, Srinivas Murthy, Alistair Nichol, Rachael Parke, Jane Parker, Kathryn Rowan, Marlene Santos, Christopher Seymour, Alexis Turgeon, Anne Turner, Frank van de Veerdonk, Steve Webb (Chair), Ryan Zarychanski Regional Management Committees Australia and New Zealand Yaseen Arabi, Lewis Campbell, Allen Cheng, Lennie Derde, Andrew Forbes, David Gattas, Cameron Green, Stephane Heritier, Peter Kruger, Edward Litton, Colin McArthur (Deputy Executive Director), Shay McGuinness (Chair), Alistair Nichol, Rachael Parke, Jane Parker, Sandra Peake, Jeffrey Presneill, Ian Seppelt, Tony Trapani, Anne Turner, Steve Webb (Executive Director), Paul Young Canadian Regional Management Committee Zahra Bhimani, Brian Cuthbertson, Rob Fowler, Francois Lamontagne, John Marshall (Executive Director), Venika Manoharan, Srinivas Murthy (Deputy Executive Director), Marlene Santos, Alexis Turgeon, Ryan Zarychanski CRIT Care Asia (CCA) Regional Management Committee Diptesh Aryal, Abi Beane (Chair), Arjen M Dondrop, Cameron Green, Rashan Haniffa (Executive Director), Madiha Hashmi, Issrah Jawad, Deva Jayakumar, John Marshall, Colin McArthur, Srinivas Murthy, Timo Tolppa, Vanessa Singh, Steve Webb European Regional Management Committee Farah Al-Beidh, Derek Angus, Djillali Annane, Wilma van Bentum-Puijk, Scott Berry, Marc Bonten (Executive Director and Chair), Nicole Brillinger, Frank Brunkhorst, Maurizio Cecconi, Lennie Derde, Stephan Ermann, Bruno Francois, Herman Goossens, Anthony Gordon, Cameron Green, Sebastiaan Hullegie, Rene Markgraff, Colin McArthur, Paul Mouncey, Alistair Nichol, Mathias Pletz, Pedro Povoa, Gernot Rohde, Kathryn Rowan, Steve Webb United States Regional Management Committee Brian Alexander, Derek Angus (Executive Director), Kim Basile, Meredith Buxton (Chair), Timothy Girard, Christopher Horvat, David Huang, Kelsey Linstrum, Florian Mayr, Bryan McVerry, Stephanie Montgomery, Christopher Seymour Regional Coordinating Centers Australia, CCA region, and Saudi Arabia: The Australia and New Zealand Intensive Care Research Centre (ANZIC-RC), Monash University Canada: St. Michael’s Hospital, Unity Health Toronto

3

Europe: University Medical Center Utrecht (UMCU) Ireland: Irish Critical Care Clinical Trials Network, University College Dublin Research Centre, St. Vincent’s Hospital New Zealand: The Medical Research Institute of New Zealand (MRINZ) United Kingdom: Intensive Care National Audit and Research Centre (ICNARC), and Imperial College London United States: Global Coalition for Adaptive Research (GCAR), and University of Pittsburgh Medical Center CRIT Care Asia (CCA): NICS MORU. Domain-Specific Working Groups Antibiotic and Macrolide Duration Domain-Specific Working Group Richard Beasley, Marc Bonten, Allen Cheng (Chair), Nick Daneman, Lennie Derde, Robert Fowler, David Gattas, Anthony Gordon, Cameron Green, Peter Kruger, Colin McArthur, Steve McGloughlin, Susan Morpeth, Srinivas Murthy, Alistair Nichol, Mathias Pletz, David Paterson, Gernot Rohde, Steve Webb Corticosteroid Domain-Specific Working Group Derek Angus (Chair), Wilma van Bentum-Puijk, Lennie Derde, Anthony Gordon, Sebastiaan Hullegie, Peter Kruger, Edward Litton, John Marshall, Colin McArthur, Srinivas Murthy, Alistair Nichol, Bala Venkatesh, Steve Webb Influenza Antiviral Domain-Specific Working Group Derek Angus, Scott Berry, Marc Bonten, Allen Cheng, Lennie Derde, Herman Goossens, Sebastiaan Hullegie, Menno de Jong, John Marshall, Colin McArthur, Srinivas Murthy (Chair), Tim Uyeki, Steve Webb COVID-19 Antiviral Domain-Specific Working Group Derek Angus, Yaseen Arabi (Chair), Kenneth Baillie, Richard Beasley, Scott Berry, Marc Bonten, Allen Cheng, Menno de Jong, Lennie Derde, Eamon Duffy, Rob Fowler, Herman Goossens, Anthony Gordon, Cameron Green, Thomas Hills, Colin McArthur, Susan Morpeth, Srinivas Murthy, Alistair Nichol, Katrina Orr, Rachael Parke, Jane Parker, Asad Patanwala, Kathryn Rowan, Steve Tong, Tim Uyeki, Frank van de Veerdonk, Steve Webb COVID-19 Immune Modulation Domain-Specific Working Group Derek Angus, Yaseen Arabi, Kenneth Baillie, Richard Beasley, Scott Berry, Marc Bonten, Frank Brunkhorst, Allen Cheng, Nichola Cooper, Olaf Cremer, Menno de Jong, Lennie Derde (Chair), Eamon Duffy, James Galea, Herman Goossens, Anthony Gordon, Cameron Green, Thomas Hills, Andrew King, Helen Leavis, John Marshall, Florian Mayr, Colin McArthur, Bryan McVerry, Susan Morpeth, Srinivas Murthy, Mihai Netea, Alistair Nichol, Kayode Ogungbenro, Katrina Orr, Jane Parker, Asad Patawala, Ville Pettilä (Deputy Chair), Emma Rademaker, Kathryn Rowan, Manoj Saxena, Christopher Seymour, Wendy Sligl, Steven Tong, Tim Uyeki, Frank van de Veerdonk, Steve Webb, Taryn Youngstein COVID-19 Immune Modulation -2 Domain-Specific Working Group Derek Angus, Scott Berry, Lennie Derde, Cameron Green, David Huang, Florian Mayr, Bryan McVerry, Stephanie Montgomery, Christopher W. Seymour (Chair), Steve Webb

4

Therapeutic Anticoagulation Domain-Specific Working Group Derek Angus, Diptesh Aryal, Scott Berry, Shilesh Bihari, Charlotte Bradbury, Marc Carrier, Dean Fergusson, Robert Fowler, Ewan Goligher (Deputy Chair), Anthony Gordon, Christopher Horvat, David Huang, Beverley Hunt, Devachandran Jayakumar, Anand Kumar, Mike Laffan, Patrick Lawler, Sylvain Lother, Colin McArthur, Bryan McVerry, John Marshall, Saskia Middeldopr, Zoe McQuilten, Matthew Neal, Alistair Nichol, Christopher Seymour, Roger Schutgens, Simon Stanworth, Alexis Turgeon, Steve Webb, Ryan Zarychanski (Chair) Vitamin C Domain-Specific Working Group Neill Adhikari (Chair), Derek Angus, Djillali Annane, Matthew Anstey, Yaseen Arabi, Scott Berry, Emily Brant, Angelique de Man, Lennie Derde, Anthony Gordon, Cameron Green, David Huang, Francois Lamonagne (Chair), Edward Litton, John Marshall, Marie-Helene Masse, Colin McArthur, Shay McGuinness, Paul Mouncey, Srinivas Murthy, Rachael Parke, Alistair Nichol, Tony Trapani, Andrew Udy, Steve Webb COVID-19 Immunoglobulin Domain-Specific Working Group Derek Angus, Donald Arnold, Phillipe Begin, Scott Berry, Richard Charlewood, Michael Chasse, Mark Coyne, Jamie Cooper, James Daly, Lise Estcourt (Chair, UK lead), Dean Fergusson, Anthony Gordon, Iain Gosbell, Heli Harvala-Simmonds, Tom Hills (New Zealand lead), Christopher Horvat, David Huang, Sheila MacLennan, John Marshall, Colin McArthur (New Zealand lead), Bryan McVerry (USA lead), David Menon, Susan Morpeth, Paul Mouncey, Srinivas Murthy, John McDyer, Zoe McQuilten (Australia lead), Alistair Nichol (Ireland lead), Nicole Pridee, David Roberts, Kathryn Rowan, Christopher Seymour, Manu Shankar-Hari (UK lead), Helen Thomas, Alan Tinmouth, Darrell Triulzi, Alexis Turgeon (Canada lead), Tim Walsh, Steve Webb, Erica Wood, Ryan Zarychanski (Canada lead) Simvastatin Domain-Specific Working Group Derek Angus, Yaseen Arabi, Abi Beane, Carolyn Calfee, Anthony Gordon, Cameron Green, Rashan Haniffa, Deva Jayakumar, Peter Kruger, Patrick Lawler, Edward Litton, Colin McArthur, Daniel McAuley (Chair), Bryan McVerry, Matthew Neal, Alistair Nichol, Cecilia O’Kane, Murali Shyamsundar, Pratik Sinha, Taylor Thompson, Steve Webb, Ian Young Antiplatelet Domain-Specific Working Group Derek Angus, Scott Berry, Shailesh Bihari, Charlotte Bradbury (Chair), Marc Carrier, Timothy Girard, Ewan Goligher, Anthony Gordon, Ghady Haidar, Christopher Horvat, David Huang, Beverley Hunt, Anand Kumar, Patrick Lawler, Patrick Lawless, Colin McArthur, Bryan McVerry, John Marshall, Zoe McQuilten, Matthew Neal, Alistair Nichol, Christopher Seymour, Simon Stanworth, Steve Webb, Alexandra Weissman, Ryan Zarychanski Mechanical Ventilation Domain Derek Angus, Wilma van Bentum-Puijk, Lewis Campbell, Lennie Derde, Niall Ferguson, Timothy Girard, Ewan Goligher, Anthony Gordon, Cameron Green, Carol Hodgson, Peter Kruger, John Laffey, Edward Litton, John Marshall, Colin McArthur, Daniel McAuley, Shay McGuinness, Alistair Nichol (Chair) Neil Orford, Kathryn Rowan, Ary Neto, Steve Webb ACE-2 RAS Domain Rebecca Baron, Lennie Derde, Slava Epelman, Claudia Frankfurter, David Gattas, Frank Gommans, Anthony Gordon, Rashan Haniffa, David Huang, Edy Kim, Francois Lamontagne, Patrick Lawler (Chair), David Leaf, John Marshall, Colin McArthur, Bryan McVerry, Daniel McAuley, Muthiah Vaduganathan, Roland van Kimmenade, Frank van de Veerdonk, Steve Webb

5

Statistical Analysis Committee Michelle Detry, PhD, Mark Fitzgerald, PhD, Roger Lewis, MD, PhD (Chair), Anna McGlothlin, PhD, Ashish Sanil, PhD, Christina Saunders, PhD Statistical Design Team Lindsay Berry, PhD, Scott Berry, PhD, Elizabeth Lorenzi, PhD Project Management Australia and Saudi Arabia: Jane Parker, Vanessa Singh, Claire Zammit Canada: Zahra Bhimani, Marlene Santos CCA: Abi Beane, Rashan Haniffa, Timo Tolppa Europe: Wilma van Bentum Puijk, Wietske Bouwman, Radhika Ganpat, Erika Groenveld, Denise van Hout, Yara Mangindaan, Clementina Okundaye, Lorraine Parker, Svenja Peters, Ilse Rietveld, Linda Rikkert, Kik Raymakers, Irma Scheepstra-Beukers, Albertine Smit, Germany: Nicole Brillinger, Rene Markgraf Global: Cameron Green Ireland: Kate Ainscough, Kathy Brickell, Peter Doran New Zealand: Anne Turner United Kingdom: Farah Al-Beidh, Aisha Anjum, Janis-Best Lane, Elizbeth Fagbodun, Lorna Miller, Paul Mouncey, Karen Parry-Billings, Sam Peters, Alvin Richards-Belle, Michelle Saull, Stefan Sprinkmoller, Daisy Wiley United States of America: Kim Basile, Meredith Buxton, Kelsey Linstrum, Stephanie Montgomery Data and Safety Monitoring Board Julian Bion, Jason Connor, (Deputy Chair), Simon Gates, Victoria Manax (Chair), Tom van der Poll, John Reynolds Database Providers Research Online: Marloes van Beurden, Evelien Effelaar, Joost Schotsman, Spinnaker Software: Craig Boyd, Cain Harland, Audrey Shearer, Jess Wren University of Pittsburgh Medical Center Giles Clermont, William Garrard, Christopher Horvat, Kyle Kalchthaler, Andrew King, Daniel Ricketts, Salim Malakoutis, Oscar Marroquin, Edvin Music, Kevin Quinn NICS MORU: on behalf of CCA; Udara Attanayaka, Abi Beane, Sri Darshana, Rashan Haniffa, Pramodya Ishani, Issrah Jawad, Upulee Pabasara, Timo Tolppa, Ishara Udayanga. Clinical Trials Groups The REMAP-CAP platform is supported by the Australian and New Zealand Intensive Care Society Clinical Trials Group, the Canadian Critical Care Clinical Trials Group, the Irish Critical Care Clinical Trials Network, the UK Critical Care Research Group and the International Forum of Acute Care Trialists. REMAP-CAP was supported in the UK by the NIHR Clinical Research Network and we acknowledge the contribution of Kate Gilmour, BSc (Hons), Karen Pearson, MSc, Chris Siewerski, MSc, Sally-Anne

6

Hurford, MSc, Emma Marsh, FdSC, Debbie Campbell, Penny Williams, MSc, Kim Shirley, LLB Hons, Meg Logan, NVQ, Jane Hanson, Becky Dilley, BSc, Louise Phillips, CIM, Anne Oliver, MSc, Mihaela Sutu, MSc, Sheenagh Murphy, PGDip, Latha Aravindan, PhD, Joanne Collins, MRes, Holly Monaghan, Adam Unsworth, NVQ, Seonaid Beddows, MSc, Laura Ann Dawson, LLM, Sarah Dyas, Adeeba Asghar, MSc, Kate Donaldson, BA, Tabitha Skinner, BSc, Nhlanhla Mguni, BSc (Hons), Natasha Muzengi, BSc, Ji Luo, PhD, Joanna O’Reilly, BSc (Hons), Chris Levett, MSc, Alison Potter, David Porter, PhD, Teresa Lockett, MSc, Jazz Bartholomew, MSc, Clare Rook, MSc, Rebecca McKay, Hannah Williams, MSc, Alistair Hall, FRCP, Hilary Campbell, BSc (Hons), Holly Speight, BA, Sandra Halden, Susan Harrison, Mobeena Naz, Kaatje Lomme, MA, Paula Sharratt, MSc, Johnathan Sheffield, FRCP, William Van’t Hoff, FRCPCH, James D Williamson, PhD, Alex Barnard, BSc (Hons), Catherine Birch, BA, Morwenna Brend, PhD, Emma Chambers, PhD, Sarah Crawshaw, PhD, Chelsea Drake, BSc, Hayley Duckles-Leech, PhD, Justin Graham, BSc (Hons), Heather Harper, BSc (Hons), Stephen Lock, PGCert, Nicola McMillan, PhD, Cliodhna O’Flaherty, Eleanor OKell, PhD, Amber Hayes, Sally Sam, BSc, Heather Slade, MSc, Susan Walker, PhD, Karen Wilding, MRes, Jayne Goodwin, Helen Hodgson, Yvette Ellis, Dawn Williamson, Madeleine Bayne, MSc, Shane Jackson, Rahim Byrne, Sonia McKenna, Alsion Clinton, NIHR Urgent Public Health Group: https://www.nihr.ac.uk/documents/urgent-public-health-group-members/24638#Members REMAP-CAP was supported in France by the CRICS-TRIGGERSEP network REMAP-CAP was supported in Ireland by the Irish Critical Care Clinical Trials Network and we acknowledge the contribution of Kate Ainscough, Kathy Brickell and Peter Doran. REMAP-CAP was supported in the Netherlands by the Research Collaboration Critical Care the Netherlands (RCC-Net). Site Investigators and Research Coordinators Australia: The Alfred Hospital: Andrew Udy, PhD, Phoebe McCracken, MPH, Meredith Young, MPH, Jasmin

Board, MPH, Emma Martin, BPharmSc; Ballarat Health Services: Khaled El-Khawas, FCICM, Angus Richardson, FCICM, Dianne Hill, BN

GradCert(CritCare), Robert J Commons, PhD, Hussam Abdelkharim, FCICM; Bendigo Hospital: Cameron Knott, MBBS(Hons) GDipClinUS MClinEd FCICM FRACP, Julie Smith,

GDipN(CritCare), Catherine Boschert, PGDip; Caboolture Hospital: Julia Affleck, MBiotech, Yogesh Apte, FCICM, Umesh Subbanna, FCICM, Roland

Bartholdy, FCICM, Thuy Frakking, PhD; Campbelltown Hospital: Karuna Keat, MB BS(Hons), Deepak Bhonagiri, MD, Ritesh Sanghavi, MBBS,

Jodie Nema, B.App.Sc.(Ag), Megan Ford, BSc; Canberra Hospital: Harshel G. Parikh, FCICM, Bronwyn Avard, MLMEd, Mary Nourse,

PGCert(CritCare); Concord Repatriation General Hospital: Winston Cheung, MBChB, Mark Kol, MBBS, Helen Wong, RN,

Asim Shah, MBChB, Atul Wagh, MBBS; Eastern Health (Box Hill, Maroondah & Angliss Hospitals): Joanna Simpson, FCICM FANZCA, Graeme

Duke, FCICM, Peter Chan, FCICM, Brittney Carter, PGCert(CritCare), Stephanie Hunter, MN; Flinders Medical Centre: Shailesh Bihari, PhD, Russell D Laver, FCICM, Tapaswi Shrestha, MN, Xia Jin,

MN; Fiona Stanley Hospital: Edward Litton, FCICM, Adrian Regli, FCICM, Susan Pellicano, PGCert,

Annamaria Palermo, BA, Ege Eroglu, BSc(Hons);

7

Footscray Hospital: Craig French, MBBS, Samantha Bates, RN GDipCritCare, Miriam Towns, MPH, Yang Yang, MBBS, Forbes McGain, PhD;

Gold Coast University Hospital: James McCullough, FCICM, Mandy Tallott, MN; John Hunter Hospital: Nikhil Kumar, MBBS, Rakshit Panwar, FCICM, Gail Brinkerhoff, PGCert,

Cassandra Koppen, BPharm, Federica Cazzola, MBBS; Launceston General Hospital: Matthew Brain, FCICM, FRACP, DDU; Sarah Mineall, MN; Lyell McEwin Hospital: Roy Fischer, FCICM, Vishwanath Biradar, FCICM, Natalie Soar, BSN; Logan Hospital: Hayden White, PhD, Kristen Estensen, MBBS, Lynette Morrison, PGCert(CritCare),

Joanne Sutton, RN, Melanie Cooper, PGCert(CritCare); Monash Health (Monash Medical Centre, Dandenong Hospital & Casey Hospital): Yahya Shehabi,

PhD, Wisam Al-Bassam, MBChB, Amanda Hulley, BSc; Umesh Kadam, MSc, Kushaharan Sathianathan, MSc;

Nepean Hospital: Ian Seppelt, PhD, Christina Whitehead, MN, Julie Lowrey, BN, Rebecca Gresham, BN, Kristy Masters, BN;

Princess Alexandra Hospital: Peter Kruger, PhD, James Walsham, FCICM, Mr Jason Meyer, BN, Meg Harward, MN, Ellen Venz, PGCert(CritCare);

The Prince Charles Hospital: Kara Brady, MPharm, Cassandra Vale, GDipClinPharm, Kiran Shekar, PhD, Jayshree Lavana, FCICM, Dinesh Parmar, FCICM;

The Queen Elizabeth Hospital: Sandra Peake, PhD, Patricia Williams, BN, Catherine Kurenda, RN; Rockhampton Hospital: Helen Miles, MBBS MPH FCICM, Antony Attokaran, MBBS FCICM FRACP; Royal Adelaide Hospital: Samuel Gluck, MD, Stephanie O’Connor, MNSc, Marianne Chapman, PhD,

Kathleen Glasby, CCRN; Royal Darwin Hospital: Lewis Campbell, FCICM, Kirsty Smyth, PGCert(CritCare), Margaret Phillips,

MN; Royal Melbourne Hospital: Jeffrey Presneill, PhD, Deborah Barge, CCRN, Kathleen Byrne, MNSc,

Alana Driscoll GDipN(CritCare), Louise Fortune, MNSc; Royal North Shore Hospital: Pierre Janin, MD, Elizabeth Yarad, MN, Frances Bass, MSc, Naomi

Hammond, PhD, Anne O’Connor, RN; Royal Perth Hospital: Sharon Waterson, PGCert(CritCare), Steve Webb, PhD, Robert McNamara,

BMBS; Royal Prince Alfred Hospital: David Gattas, MMed(ClinEpid), Heidi Buhr, MScMed(ClinEpid), Jennifer

Coles, GDipN(CritCare); Sir Charles Gardiner Hospital: Sacha Schweikert, FCICM, Bradley Wibrow, FCICM, Matthew Anstey,

FCICM, Rashmi Rauniyar, MPH; St George Hospital: Kush Deshpande, FCICM, Pam Konecny, FRACP, Jennene Miller, BN, Adeline

Kintono, BN, Raymond Tung B.Pharm, MPS, MSHP St. John of God Midland Public and Private Hospitals: Ed Fysh, PhD, Ashlish Dawda, MD, Bhaumik

Mevavala, MHA; St. John of God Hospital, Murdoch: Annamaria Palermo, BA, Adrian Regli, FCICM, Bart De Keulenaer,

FCICM; St. John of God Hospital, Subiaco: Ed Litton, PhD, Janet Ferrier, BSc; St. Vincent’s Hospital (NSW): Priya Nair, PhD, Hergen Buscher, FCICM, Claire Reynolds, MClinNurse,

Sally Newman, PGCert(CritCare); St. Vincent’s Hospital (VIC): John Santamaria, MDBS, Leanne Barbazza, PGCert(CritCare), Jennifer

Homes, PGCert(CritCare), Roger Smith, MPH; Sunshine Coast University Hospital: Peter Garrett, FCICM, Lauren Murray, MSc, Jane Brailsford,

PGCert(CritCare), Loretta Forbes, PGCert(CritCare), Teena Maguire, BA;

8

Sunshine Hospital: Craig French, MBBS, Gerard Fennessy, MBBS, John Mulder, MBBS, Rebecca Morgan, RN PGCert(CritCare), Rebecca McEldrew, RN PGCert(CritCare);

The Sutherland Hospital: Anas Naeem, MClinEd, Laura Fagan, GDipN(CritCare), Emily Ryan, PGCert(CritCare);

Toowoomba Hospital: Vasanth Mariappa, FCICM, Judith Smith, PGCert; University Hospital Geelong: Scott Simpson, MBBS, Matthew Maiden, PhD, Allison Bone,

GDipN(CritCare), , Michelle Horton, MN, Tania Salerno, PGCert; Wollongong Hospital: Martin Sterba, PhD, Wenli Geng, MN;

Belgium: Ghent University Hospital: Pieter Depuydt, PhD, Jan De Waele, PhD, Liesbet De Bus, PhD, Jan Fierens,

MD, Stephanie Bracke, BSc, Joris Vermassen, MD, Daisy Vermeiren, BSc;

Canada: Brantford General Hospital: Brenda Reeve, MD, William Dechert, MSc; Centre de recherche de l’Institut universitaire de cardiologie et de pneumologie de Québec: Francois

Lellouche, Patricia Lizotte Centre Hospitalier de l’Universite de Montreal: Michaël Chassé, PhD, François Martin Carrier, MSc,

Dounia Boumahni, BSc, Fatna Benettaib, MSc., Ali Ghamraoui, BSc; CHU de Québec – Université Laval: Alexis Turgeon, MSc, David Bellemare, BSc, Ève Cloutier, Rana

Daher, François Lauzier, MSc, Charles Francoeur, MSc; Centre Hospitalier Universitaire de Sherbrooke: François Lamontagne, MD, Frédérick D’Aragon, MD,

Elaine Carbonneau, BACC, Julie Leblond, BACC; Grace Hospital: Gloria Vazquez-Grande, Nicole Marten Grand River Hospital (Kitchener): Theresa Liu, Atif Siddiqui Health Sciences Centre, Winnipeg: Ryan Zarychanski, MD, Gloria Vazquez-Grande, MD, Nicole

Marten, RN, Maggie Wilson, MSc; Hôpital du Sacré Coeur de Montréal: Martin Albert, MD, Karim Serri, MD, Alexandros Cavayas, MD,

Mathilde Duplaix, MSc, Virginie Williams, PhD; Juravinski Hospital: Bram Rochwerg, MD, Tim Karachi, MD, Simon Oczkowski, MD, John Centofanti,

MD, Tina Millen, RRT McGill University Health Centre: Kosar Khwaja, Josie Campisi Niagara Health (St. Catherine’s Hospital): Erick Duan, MD, Jennifer Tsang, MD, Lisa Patterson, BA; Regina General Hospital: Eric Sy, Chiraag Gupta, Sandy SHA Kassir Royal Alexandra Hospital: Demetrios Kutsogiannis, Patricia Thompson Sunnybrook Health Sciences Centre: Rob Fowler, Neill Adhikari, Maneesha Kamra, Nicole Marinoff St. Boniface General Hospital: Ryan Zarychanski, Nicole Marten St. Joseph’s Healthcare Hamilton: Deborah Cook, Frances Clarke St. Mary's General Hospital (Kitchener): Rebecca Kruisselbrink, Atif Siddiqui St. Michael’s Hospital: John Marshall, MD, Laurent Brochard, MD, Karen Burns, MD, Gyan Sandhu,

RN, Imrana Khalid, MD; The Ottawa Hospital: Shane English, MSc, Irene Watpool, BScN, Rebecca Porteous, BSN, Sydney

Miezitis, BSc, Lauralyn McIntyre, MSc; University Health Network: Elizabeth Wilcox, Lorenzo del Sorbo, Hesham Abdelhady, Tina

Romagnuolo University of Alberta: Wendy Sligl, Nadia Baig William Osler Health System: Alexandra Binnie, MD, Elizabeth Powell, MD, Alexandra McMillan, MD,

Tracy Luk, MD, Noah Aref, MSc

9

CCA: India Apollo Speciality Hospital - OMR, Chennai: Devachandran Jayakumar, MD, Suresh Babu BSc; Apollo Main Hospital, Chennai: C Vignesh, MD, Augustian James BSc. Apollo Speciality Vanagaram, Vanagaram, Chennai; R Ebenezer MD, S Krishnamurthy MD, Lakshmi

Ranganathan PhD, Manisha, MD Nepal Grande International Hospital : Sushil Khanal, MD, Sameena Amatya, RN; HAMS Hospital:

Hem Raj Paneru, MD, Sabin Koirala, MD, Pratibha Paudel ,RN; Nepal Mediciti Hospital: Diptesh Aryal,MD, Kanchan Koirala,RN, Namrata Rai,RN, Subekshya Luitel, BSc; Tribhuvan University Teaching Hospital : Hem Raj Paneru,MD, Binita Bhattarai, RN;

Pakistan (CCA): Ziauddin Hospital Clifton Campus: Madiha Hashmi, MD; Ashok Panjwani, MD; Zulfiqar Ali Umrani,

Shoaib Siddiq; Mohiuddin Shaikh; National Institute of Cardiovascular Diseases Pakistan: Nawal Salahuddin, MD, Sobia Masood;

Croatia: General Hospital Pozega: Zdravko Andric, Sabina Cviljevic, Renata Đimoti, Marija Zapalac, Gordan

Mirković; University Hospital of Infectious Diseases “Dr Fran Milhajevid”: Bruno Baršić, Marko Kutleša, Viktor

Kotarski; University Hospital of Zagreb: Ana Vujaklija Brajković, Jakša Babel, Helena Sever, Lidija Dragija, Ira

Kušan; Finland: Helsinki University Hospital: Suvi Vaara, PhD, Leena Pettilä, Jonna Heinonen, Ville Pettilä, PhD; Tampere University Hospital: Anne Kuitunen, PhD, Sari Karlsson, PhD, Annukka Vahtera, PhD, Heikki

Kiiski, PhD, Sanna Ristimäki; France: Ambroise Pare Hospital: Amine Azaiz, Cyril Charron, MD, Mathieu Godement, MD, Guillaume Geri,

Antoine Vieillard-Baron; Centre Hospitalier de Melun: Franck Pourcine, Mehran Monchi; Centre Hospitalier Simone Veil, Beauvais: David Luis, MD, Romain Mercier, MD, Anne Sagnier, MD,

Nathalie Verrier, MD, Cecile Caplin, MD, Jack Richecoeu, MD, Daniele Combaux, MD; Centre Hospitalier Sud Essonne: Shidasp Siami, PhD, Christelle Aparicio, Sarah Vautier, Asma

Jeblaoui, Delphine Lemaire-Brunel; Centre Hospitalier Tenon: Muriel Fartoukh, MD, Laura Courtin, Vincent Labbe, MD , Guillaume

Voiriot, MD, Sara Nesrine Salhi; Centre Hospitalier Victor Dupouy: Gaetan Plantefeve, MD, Cécile Leparco, RN, Damien Contou, MD ; CHR d'Orleans: Grégoire Muller, MD, Mai-Anh Nay, MD, Toufik Kamel, MD, Dalila Benzekri, MD,

Sophie Jacquier, MD, Isabelle Runge, MD, Armelle Mathonnet, MD, François Barbier, MD, Anne Bretagnol, MD;

CHRU Tours Hopital Bretonneau: Emmanuelle Mercier, MD, Delphine Chartier, Charlotte Salmon, MD, Pierre-François Dequin, PhD, Denis GAROT, MD;

CHU Dupuytren, Limoges; Hôpital Civil, Hôpitaux Universitaires de Strasbourg;

10

Hôpital de Hautepierre, Hôpitaux Universitaires de Strasbourg: Francis Schneider, PhD, Vincent Castelain, PhD, Guillaume Morel, MD, Sylvie L’Hotellier, MSc;

Hospital Nord Franche-Comté: Julio Badie, MD, Fernando Daniel Berdaguer, MD, Sylvain Malfroy, MD, Chaouki Mezher, MD, Charlotte Bourgoin, MSc, Guy Moneger, MD, Elodie Bouvier, MSc;

Lariboisière Hospital: Bruno Megarbane, PhD, Sebastian Voicu, PhD, Nicolas Deye, PhD, Isabelle Malissin, MD, Laetitia Sutterlin, MD, Aymen Mrad, MD, Adrien Pépin Lehalleur, MD, Giulia Naim, MD, Philippe Nguyen, MD, Jean-Michel Ekhérian, MD, Yvonnick Boué, MD, Georgios Sidéris, PhD, Dominique Vodovar, PhD, Emmanuelle Guérin, MD, Caroline Grant, MD;

Le Mans Hospital: Christophe Guitton, PhD, Cédric Darreau, MD, Mickaël Landais, MD, Nicolas Chudeau, MD, Alain Robert, PhD, Patrice Tirot, MD, Jean Christophe Callahan, MD, Marjorie Saint Martin, MD, Charlène Le Moal, MD, Rémy Marnai, MD, Marie Hélène Leroyer;

Raymond Poincaré Hospital: Djillali Annane, MD, Pierre Moine, MD, Nicholas Heming, MD, Virginie Maxime, MD, Isabelle Bossard, MSc, Tiphaine Barbarin Nicholier, MD, Bernard Clair, MD, David Orlikowski, MD, Rania Bounab, MD, Lilia Abdeladim, MD;

Vendee Hospital: Gwenhael Colin, MD, Vanessa Zinzoni, Natacha Maquigneau, Matthieu Henri-Lagarrigue, MD, Caroline Pouplet, MD;

Germany: Carl-Thiem-Klinikum Cottbus gGmbH: Jens Soukup, PD, Richard Wetzold, Madlen Löbel, Dr. Ing, Lisa

Starke, Patrick Grimm; Charité - Universitätsmedizin Berlin: André Finn, MD, Gabriele Kreß, Uwe Hoff, MD, Carl Friedrich

Hinrichs, MD, Jens Nee, MD; Jena University Hospital: Mathias W. Pletz, PhD, Stefan Hagel, PhD, Juliane Ankert, MSc, Steffi

Kolanos, BSc, Frank Bloos, PhD; Klinikum Dortmund gGmbH: Daniela Nickoleit-Bitzenberger, MD , Bernhard Schaaf, MD, Werner

Meermeier, MD, Katharina Prebeg, Harun Said Azzaui, Martin Hower, Klaus-Gerd Brieger, Corinna Elender, Timo Sabelhaus, Ansgar Riepe, Ceren Akamp, Julius Kremling, Daniela Klein, Elke Landsiedel-Mechenbier;

University Hospital of Leipzig: Sirak Petros, MD, Kevin Kunz, MD, Bianka Schütze, BSc; Universitätsklinikum Hamburg-Eppendorf: Stefan Kluge, MD, Axel Nierhaus, MD, Dominik Jarczak,

MD, Kevin Roedl, MD; University Hospital of Frankfurt: Gernot Gerhard Ulrich Rohde, MD, Achim Grünewaldt, MD, Jörg

Bojunga, MD; University Hospital of Wuerzburg: Dirk Weismann, MD, Anna Frey, MD; Maria Drayss, MD, M.E.

Goebeler, MEG, Thomas Flor, Gertrud Fragner, Nadine Wahl, Juliane Totzke, Cyrus Sayehli, MD;

Vivantes Klinikum Neukölln: Lorenz Reill, Michael Distler, MD, Astrid Maselli; Hungary: Almási Balogh Pál Hospital, Ózd: János Bélteczki, István Magyar, Ágnes Fazekas, Sándor Kovács,

Viktória Szőke; Jósa András County Hospital, Nyíregyháza: Gábor Szigligeti, János Leszkoven; Ireland: Beacon Hospital Dublin: Daniel Collins, MRCPI, Kathy Brickell, RGN, Liadain Reid, MPH, Michelle

Smyth, PgDip, Patrick Breen, MB FJFICMI, Sandra Spain, RGN; Beaumont Hospital: Gerard Curley; PhD, Natalie McEvoy, MSc, Pierce Geoghegan, MB, Jennifer

Clarke, MB;

11

Galway University Hospitals: John Laffey, MD, Bairbre McNicholas, PhD, Michael Scully, MD, Siobhan Casey, RN, Maeve Kernan, RN, Aoife Brennan, PhD, Ritika Rangan, PhD, Riona Tully, MB, Sarah Corbett, MB, Aine McCarthy, MB, Oscar Duffy, MB, David Burke, MB;

St Vincent’s University Hospital, Dublin: Alistair Nichol, PhD, Kathy Brickell, RGN, Michelle Smyth, PGDip, Leanne Hayes, PhD, Liadain Reid, MPH, Lorna Murphy, RGN, Andy Neill, MB, Bryan Reidy, MSc, Michael O’Dwyer, PhD, Donal Ryan, MD, Kate Ainscough, PhD;

Netherlands: Canisius Wilhelmina Ziekenhuis: Oscar Hoiting, MD, Marco Peters, MD, Els Rengers, MD, Mirjam

Evers, RN, Anton Prinssen, RN; Deventer Hospital: Huub L.A. van den Oever, MD, Arriette Kruisdijk-Gerritsen, CCRN; Jeroen Bosch Ziekenhuis: Koen Simons, PhD, Tamara van Zuylen, RN, Angela Bouman, RN; Meander Medisch Centrum: Laura van Gulik, PhD; Radboud University Medical Center Nijmegen: Jeroen Schouten, PhD, Peter Pickkers, PhD, Noortje

Roovers, BSc, Margreet Klop-Riehl, BSc, Hetty van der Eng, BSc; UMC Leiden: Evert de Jonge, PhD, Jeanette Wigbers, RN, Michael del Prado, RN; UMC Utrecht: Marc Bonten, PhD, Olaf Cremer, PhD, Lennie Derde, PhD, Diederik van Dijk, PhD,

Emma Rademaker, MD, Jelle Haitsma Mulier, MD, Anna Linda Peters, PhD, Birgit Romberg, MD;

Ziekenhuis Gelderse Vallei: Sjoerd van Bree, MD, Marianne Bouw-Ruiter, RN, Barbara Festen, MD, Fiona van Gelder, MD, Mark van Iperen, MD, Margreet Osinga, RN, Roel Schellaars, MD, Dave Tjan, MD, Ruben van der Wekken, MD, Max Melchers, MD, Arthur van Zanten, MD;

New Zealand: Auckland City Hospital, Cardiothoracic and Vascular ICU: Shay McGuinness, MBChB, Rachael Parke,

PhD, Eileen Guilder, MA, Magdalena Butler, RN, Keri-Anne Cowdrey, RN, Melissa Woollett, BHSc (Nurs);

Auckland City Hospital, DCCM: Colin McArthur, FJFICM, Thomas Hills, DPhil, Lynette Newby, MN, Yan Chen, MN, Catherine Simmonds, PGDipHSc, Rachael McConnochie, MN, Caroline O’Connor;

Christchurch Hospital: Jay Ritzema Carter, PhD, Seton Henderson, MD, Kymbalee Van Der Heyden, BSc, Jan Mehrtens, PGCert, Anna Morris, BN, Stacey Morgan, BN;

Middlemore Hospital: Tony Williams, BMedSc, Alex Kazemi, BMedSc, Susan Morpeth, PhD, Rima Song, PGDip, Vivian Lai, MHSc, Dinuraj Girijadevi, PGCert;

North Shore Hospital: Robert Everitt, FACEM, Robert Russell, BSc(Hons), Danielle Hacking, PGDipNurs;

Rotorua Hospital: Ulrike Buehner, FCARCSI, Erin Williams, MSc; Tauranga Hospital: Troy Browne, FCICM, Kate Grimwade, FRACP, Jennifer Goodson, RN, Owen Keet,

FANZCA, Owen Callender, FANZCA; Waikato Hospital: Robert Martynoga, FCICM, Kara Trask, PGCert, Amelia Butler, PGCert, Wellington Hospital: Paul Young, PhD, Chelsea Young, PGDip, Eden Lesona, MNSc, Shaanti Olatunji,

MClinIm, Leanlove Navarra, BSc (Nurs), Raulle Sol Cruz, BSc (Nurs) Whangarei Hospital: Katherine Perry, FJFICM, Ralph Fuchs, FANZCA, Bridget Lambert, PGDip; Taranaki Base Hospital: Jonathan Albrett, FCICM, Carolyn Jackson, RN, Simon Kirkham, PGDip; Portugal: Hospital de Abrantes: Nuno José Teodoro Amaro dos Santos Catorze, MD, Tiago Nuno Alfaro Lima

Pereira, MD, Ricardo Manuel Castro Ferreira, RN, Joana Margarida Pereira Sousa Bastos, PharmD, Teresa Margarida Oliveira Batista, RN;

12

Romania: "Dr. Victor Babes" Clinical Hospital of Infectious and Tropical Diseases Bucharest: Simin Aysel

Florescu, PhD, Delia Stanciu, MD, Mihaela Florentina Zaharia, MD, Alma Gabriela Kosa, MD, Daniel Codreanu;

Saudi Arabia: King Abdulaziz Medical City- Riyadh: Yaseen M Arabi, MD, Sheryl Ann Abdukahil, RN, Farhan Al Enezi,

MD, Lolowa Alswaidan, MSc, Samah Y. AlQahtani, MD Spain: Hospital del Mar: Rosana Muñoz-Bermúdez, MD, Judith Marin-Corral, PhD, Anna Salazar Degracia,

PhD, Francisco Parrilla Gómez, MD, Maria Isabel Mateo López; Reina Sofia University Hospital: Rafael León López, MD, Jorge Rodriguez, PhD, Sheila Cárcel, Rosario

Carmona, MD, Carmen de la Fuente, MD, Marina Rodriguez, MD; United Kingdom: Aberdeen Royal Infirmary: Callum Kaye, MBChB, Angela Allan, PGDip; Addenbrooke's Hospital: Charlotte Summers, PhD, Petra Polgarova MSc; Alder Hey Children’s NHS Foundation Trust: Stephen J McWilliam, PhD, Daniel B Hawcutt, MD, Laura

Rad, BSc(Hons), Laura O’Malley, BSc(Hons), Jennifer Whitbread, BSc(Hons); Alexandra Hospital Redditch: Olivia Kelsall, MBChB, Nicholas Cowley MD, Laura Wild, BSc(Hons),

Jessica Thrush, RGN, Hannah Wood, BSc(Hons), Karen Austin, RGN; Altnagelvin Hospital: Adrian Donnelly, FFICM, Martin Kelly, MD, Naoise Smyth MB ChB, Sinéad

O’Kane, BSc(Hons), Declan McClintock, MSc, Majella Warnock, MPharm, Ryan Campbell BSc, Edmund McCallion MPharm;

Antrim Area Hospital: Paul Johnson, FFARCSI, Shirley McKenna, MSc, Joanne Hanley, MSc, Andrew Currie, MSc, Barbara Allen, MPharm, Clare Mc Goldrick, MPhil, Moyra Mc Master, RGN, ;

Barnet Hospital: Rajeev Jha, MD, Michael Kalogirou, MD, Christine Ellis, PhD, Vinodh Krishnamurthy, PhD, Vashish Deelchand, MSc, Aibhilin O’Connor, MSc;

Basildon Universty Hospital: Dipak Mukherjee, MD, Agilan Kaliappan, MD, Anirudda Pai, MD,Mark Vertue, Anne Nicholson, Joanne Riches, Gracie Maloney, Lauren Kittridge, Amanda Solesbury, Kezia Allen

Belfast Health and Social Care Trust (Belfast City Hospital, Mater Infirmorium, Royal Victoria Hospital): Jon Silversides, PhD, Peter McGuigan, MBBCh, Kathryn Ward, BSc, Aisling O’Neill, BSc, Stephanie Finn, BSc;

Brighton and Sussex University Hospitals Trust: Barbara Phillips, Laura Oritz-Ruiz de Gordoa, BSc; Bristol Royal Infirmary: Jeremy Bewley, MBChB, Matthew Thomas, MBChB, Katie Sweet, BSc(Hons),

Lisa Grimmer, BSc(Hons), Rebekah Johnson, BSc(Hons); Calderdale and Huddersfield Foundation Trust: Jez Pinnell, MD, Matt Robinson, BSc(Hons), Lisa

Gledhill, MSc, Tracy Wood, BSc(Hons); Cardiff and Vale University Health Board: Matt Morgan, PhD, Jade Cole, BSc, Helen Hill, BSc, Michelle

Davies, BN, Angharad Williams, BSc, Emma Thomas, BSc, Rhys Davies, BSc, Matt Wise, DPhil; Charing Cross Hospital: David Antcliffe, PhD, Maie Templeton, MSc, Roceld Rojo, BSN, Phoebe

Coghlan, MA, Joanna Smee, BSc; Chesterfield Royal Hospital: Euan Mackay, MD, Jon Cort, MD, Amanda Whileman, BSc, Thomas

Spencer, Nick Spittle, Sarah Beavis, MD, Anand Padmakumar, MD, Katie Dale, BSc, Joanne Hawes, BSc, Emma Moakes, BSc, Rachel Gascoyne, BSc, Kelly Pritchard, BSc, Lesley Stevenson, BSc, Justin Cooke, MD, Karolina Nemeth-Roszpopa, MD;

13

The Christie NHS Foundation Trust: Vidya Kasipandian, FFICM, Amit Patel, Suzanne Allibone, Roman Mary-Genetu, BSc;

Colchester Hospital: Mohamed Ramali, FRCA, Ooi HC, MRCEM, Alison Ghosh, RN, Rawlings Osagie, PharmD, Malka Jayasinghe Arachchige, MBBS, Melissa Hartley, MBBS;

Countess of Chester Hospital: Peter Bamford, FFICM, Emily London, MBChB, Kathryn Cawley, MRes, Maria Faulkner, BSc, Helen Jeffrey, DipNS;

Croydon University Hospital: Ashok Sundar Raj, MD, Georgios Tsinaslanidis, MD, Reena Nair Khade, BSc, Gloria Nwajei Agha, BSc, Rose Nalumansi Sekiwala;

Cumberland Infirmary: Tim Smith, FRCA, Chris Brewer, BPharm(Hons), Jane Gregory, BSc(Hons); Darlington Memorial Hospital: James Limb, FRCA, Amanda Cowton, BSc(Hons), Julie O’Brien,

DipNurs, Kelly Postlethwaite, DipNurs; Derriford Hospital: Nikitas Nikitas, PhD, Colin Wells, MSc, Liana Lankester, PGCert, Helen McMillan,

MSc; Dorset County Hospital: Mark Pulletz, FFICM, Patricia Williams, AdDip, Jenny Birch, BA, Sophie

Wiseman, Mpharm, Sarah Horton, BA(Hons); East Kent Hospitals (Queen Elizabeth the Queen Mother Hospital): Ana Alegria, CCT, Salah Turki,

MBBch, Tarek Elsefi, MRCP, Nikki Crisp, BSc, Louise Allen, BSc; East Lancashire Hospitals NHS Trust (Royal Blackburn Hospital): Matthew Smith, MD, Sri

Chukkambotla, MD, Wendy Goddard, BSc, Stephen Duberley BSc; Freeman Hospital and Royal Victoria Infirmary, Newcastle upon Tyne: Iain J McCullagh, FRCA, Philip

Robinson, MSc, Bijal Patel, MSc, Sinead Kelly, PGDip; Frimley Health NHS Foundation Trust: Omar Touma, MD, Susan Holland, Christopher Hodge, Holly

Taylor, Meera Alderman, Nicky Barnes, Joana Da Rocha, BSc, Catherine Smith, BSc, Nicole Brooks, Thanuja Weerasinghe, BSc, Julie-Ann Sinclair, Yousuf Abusamra, MD, Ronan Doherty, MD, Joanna Cudlipp, MD, Rajeev Singh, MD, Haili Yu, MD, Admad Daebis, MD, Christopher NG, MD, Sara Kendrick, MD, Anita Saran, MD, Ahmed Makky, MD, Danni Greener, MD, Louise Rowe-Leete, Dip, Alexandra Edwards, Dip, Yvonne Bland, BSc, Rozzie Dolman, BSc, Tracy Foster, BSc;

Gateshead Health NHS Trust: Vanessa Linnett, MD, Amanda Sanderson, Jenny Ritzema, Helen Wild; George Eliot Hospital: Divya Khare, FRCA, Meredith Pinder, BSN, Selvin Selvamoni, MSc, Amitha

Gopinath, MBA; Glan Clwyd Hospital: Richard Pugh, FFICM, Daniel Menzies, FRCP, Richard Lean, MBChB, Xinyi Qiu,

MBChB, Jeremy James Scanlon, MBChB; Glasgow Royal Infirmary: Kathryn Puxty, MD, Susanne Cathcart, BSc, Chris Mc Govern, MBChB,

Samantha Carmichael, MRPharms, Dominic Rimmer, BSc; Glenfield Hospital Leicester: Hakeem Yusuff, FFICM, Graziella Isgro, FFICM, Chris Brightling, PhD,

Michelle Bourne, BSc(Hons), Michelle Craner, DipHE, Rebecca Boyles, BSc (Hons); Great Western Hospitals NHS Foundation Trust: Malcolm Watters, MBBCh, Rachel Prout, MBChB,

Louisa Davies, BSc, Suzannah Pegler, BSc(Hons), Lynsey Kyeremeh, BPharm, Aiman Mian, MBBS;

Guy’s & St Thomas’ NHS Foundation Trust: Manu Shankar-Hari, PhD, Marlies Ostermann, PhD, Marina Marotti, BSc, Neus Grau Novellas, BSc, Aneta Bociek, BSc;

Hammersmith Hospital: Stephen Brett, MD, Sonia Sousa Arias, BSc, Rebecca Elin Hall, BN; Homerton University Hospital NHS Foundation Trust: Susan Jain, MD, Abhinav Gupta MD, Catherine

Holbrook James Cook University Hospital: Jeremy Henning, MB, Stephen Bonner, BSc, Keith Hugill, BSc,

Emanuel Cirstea, MSc, Dean Wilkinson, BSc, Jessica Jones, BSc;

14

James Paget University Hospitals: Michal Karlikowski, MD, Helen Sutherland, BSc(Hons), Elva Wilhelmsen, DipHE, Jane Woods, BSc, Julie North, BSc(Hons);

Kettering General Hospital: Dhinesh Sundaran, FFICM, Laszlo Hollos, FFICM, Susan Coburn, PGCert, Anna Williams, BSc, Samantha Saunders, BTEC;

King’s College Hospital (Denmark Hill site): Phil Hopkins, MD, John Smith, RN, Harriet Noble, RN, Maria Theresa Depante, RN, Emma Clarey, RN;

Lancashire Teaching Hospitals NHS Foundation Trust: Shondipon Laha, FFICM, Mark Verlander, MBA, Alexandra Williams, MSc;

Leeds Teaching Hospitals Trust: Elankumaran Paramasivam, FRCP, Elizabeth Wilby, BSc (Hons), Bethan Ogg, BSc (Hons), Clare Howcroft, BSc (Hons), Angelique Aspinwall, BSc (Hons), Sam Charlton, BSc (Hons), Richard Gould, MBBS, Deena Mistry, MPharm, Sidra Awan, MPharm, Caroline Bedford, MPharm;

Leicester General Hospital: Andrew Hall, MRCP, Jill Cooke, RGN, Caroline Gardiner-Hill, RGN, Carolyn Maloney, Nigel Brunskill, PhD;

Leicester Royal Infirmary: Hafiz R Qureshi, MRCPI, Neil Flint, MBChB, Sarah Nicholson, Sara Southin, Andrew Nicholson, Amardeep Ghattaoraya, MSc;

Lewisham and Greenwich NHS Trust: Dr Daniel Harding, MD, Sinead O’Halloran , Amy Collins Emma Smith , Estefania Trues;

Liverpool Foundation Trust Aintree: Barbara Borgatta, PhD, Ian Turner-Bone, DipHE, Amie Reddy, Laura Wilding, DipHE;

Liverpool Heart and Chest Hospital: Craig Wilson, Zuhra Surti; Luton and Dunstable University Hospital: Loku Chamara Warnapura, FFICM, Ronan Agno, BSc,

Prasannakumari Sathianathan, MSc, Deborah Shaw, FFICM, Nazia Ijaz, FFICM, Dean Burns, MD, Mohammed Nisar, MD, Vanessa Quick, MD, Craig Alexander, BSc, Sanil Patel BSc, Nafisa Hussain, Yvynne Croucher, BSc, Eva-Maria Lang, MD, Banu Rudran, MD, Syed Gilani, MD, Talia Wieder, MD, Margaret Louise Tate, BSc;

Maidstone and Tunbridge Wells NHS Trust: David Golden, FFICM, Miriam Davey, PGDip, Rebecca Seaman BSc (Hons);

Manchester Royal Infirmary: Tim Felton, FFICM, Jonathan Bannard-Smith, FFICM, Joanne Henry, Richard Clark, DiPHE, Kathrine Birchall, BSc(Hons), Joanne Henry, MA, Fiona Pomeroy, BSc (Hons), Rachael Quayle, Dip.HE, Katharine Wylie, MSc, Anila Sukuraman, BSc, John McNamarra, MD;

Medway Maritime Hospital: Arystarch Makowski, PhD, Beata Misztal, PhD, Iram Ahmed, PhD, Kevin Neicker, MBA, Sam Millington, BMBS, Rebecca Squires, BSc, Masroor Phulpoto, MBBS;

Milton Keynes University Hospital: Richard Stewart, Esther Mwaura, BSc, Louise E Mew, BSc(Hons), Lynn Wren, BSc(Hons), Felicity Willams, PhD;

Mid & South Essex NHS Foundation Trust: Aneta Oborska, FRCA, Rino Maeda, MBBS, Selver Kalchko-Veyssal, MD, Raji Orat Prabakaran, BSc, Bernard Hadebe, MSc, Eric Makmur, MBBS, Guy Nicholls, MBBS;

Musgrove Park Hospital: Richard Innes, MBBCh, Patricia Doble, BSc(Hons), Libby Graham, RN, Charmaine Shovelton, RN;

Nevill Hall Hospital: Vincent Hamlyn, MBBCh, Nancy Hawkins, PhD, Anna Roynon-Reed, MSc, Sean Cutler, MSc, Sarah Lewis, MBChB;

Newham University Hospital: Juan Martin Lazaro, PhD, Tabitha Newman, MSc; Ninewells Hospital: Pauline Austin, MBChB, Susan Chapman, MBChB, Louise Cabrelli, BSc; Norfolk and Norwich University Hospital: Simon Fletcher, FFICM, Jurgens Nortje, FFICM, Deirdre

Fottrell-Gould, Dip, Georgina Randell, Dip, Katie Stammers, BSc;

15

Northampton General Hospital: Mohsin Zaman, MRCP, Einas Elmahi, MPhil, Andrea Jones, PhD, Kathryn Hall, Dip;

Northern General Hospital, Sheffield: Gary H Mills, PhD, Kim Ryalls, RegPharmTech, Kate Harrington RCN, Helen Bowler, BSc, Jas Sall, BSc, Richard Bourne, PhD;

North Manchester General Hospital: Zoe Borrill, MD, Tracy Duncan, MD, Thomas Lamb, MD, Joanne Shaw, BSc, Claire Fox, BSc, Kirstie Smith, BSc, Sarah Holland, Bethany Blackledge, BSc, Liam McMorrow, BSc, Laura Durrans, Jade Harris;

North Middlesex University Hospital: Jeronimo Moreno Cuesta, MD, Kugan Xavier, MD, Dharam Purohit, EDIC, Munzir Elhassan, MBBS, Anne Haldeos, BSc, Rachel Vincent, DipHE, Marwa Abdelrazik, MBBCH, Samuel Jenkins, BMBS, Arunkumar Ganesan, MD, Rohit Kumar, DA, David Carter, MBBS, Dhanalakshmi Bakthavatsalam, BSc;

Oxford University Hospitals: Matthew Rowland, FFICM, Paula Hutton, PGCert, Archana Bashyal, MSc, Neil Davidson, BSc, Clare Hird, MSc, Sally Beer, MSc;

Pilgrim Hospital Boston: Manish Chhablani, FFICM, Gunjan Phalod, MPharm, Amy Kirkby, BSc(Hons), Simon Archer, BSc(Hons), Kimberley Netherton, RGN;

Princess Royal Hospital: Barbara Philips, MD, Dee Mullan, BSc, Denise Skinner, BSc, Jane Gaylard, BSc, Julie Newman, BSc;

Princess of Wales Hospital: Sonia Arun Sathe, MD, Lisa Roche, BSc, Ellie Davies, BSc, Keri Turner; Poole Hospital: Henrik Reschreiter, FFICM, Julie Camsooksai, PGDE, Sarah Patch, BSc(Hons), Sarah

Jenkins, BSc(Hons), Charlotte Humphrey, BSc (Hons); Queen Alexandra Hospital Portsmouth: David Pogson, MSc, Steve Rose, BSc, Zoe Daly, BSc, Lutece

Brimfield, BN, Angie Nown; Queen Elizabeth Hospital, Birmingham: Dhruv Parekh, PhD, Colin Bergin, BSc, Michelle Bates, BSc,

Christopher McGhee, BSc, Daniella Lynch, BSc, Khushpreet Bhandal, Dip, Kyriaki Tsakiridou, MSc, Amy Bamford, BSc, Lauren Cooper, MSc, Tony Whitehouse, MD, Tonny Veenith, MD;

Queen Elizabeth University Hospital, Glasgow: Malcolm A.B. Sim, MD, Sophie Kennedy Hay, BN, Steven Henderson, MPH, Maria Nygren, MSc, Eliza Valentine, HNC;

Queen’s Hospital, Burton: Amro Katary, MD, Gill Bell, BSc, Louise Wilcox, BSc, Katy English BSc, Ann Adams;

Queen’s Hospital, Romford: Mandeep-Kaur Phull, MBBS, Abbas Zaidi, MBBS, Tatiana Pogreban, BN, Lace Paulyn Rosaroso, BN;

Queens Medical Centre and Nottingham City Hospital: Daniel Harvey, BMBS, Benjamin Lowe, BMBS, Megan Meredith, BSc(Hons), Lucy Ryan, MNSc, DREEAM Research Team;

The Rotherham NHS Foundation Trust: Anil Hormis, FRCA, Rachel Walker, BA, Dawn Collier, BSc, Sarah Kimpton, MSc, Susan Oakley;

Royal Alexandra Hospital: Kevin Rooney, MBChB, Natalie Rodden, BSc, Emma Hughes, BSc, Nicola Thomson, BSc(Hons), Deborah McGlynn, BSc, Charlotte Clark, Dip, Patricia Clark, BSc;

Royal Berkshire Hospital: Andrew Walden, FFICM, Liza Keating, MBChB, Matthew Frise, DPhil, Tolu Okeke, BSc, Nicola Jacques, MSc, Holly Coles, BSc, Emma Tilney, BSc, Emma Vowell, DipHE;

Royal Bournemouth and Christchurch Hospitals: Martin Schuster-Bruce, FRCA, Sally Pitts, BSc, Rebecca Miln, ADipHE, Laura Purandare, MBA, Luke Vamplew, BSc;

Royal Brompton Hospital: Brijesh Patel, FRCA, Debra Dempster, Mahitha Gummadi, Natalie Dormand, Shu Fang Wang;

Royal Cornwall NHS Trust: Michael Spivey, FFICM, Sarah Bean, RN, Karen Burt, RN, Lorraine Moore, MPharm;

Royal Devon and Exeter NHS Foundation Trust: Christopher Day, MD, Charly Gibson, MBChB, Elizabeth Gordon, BSc, Letizia Zitter, BSc, Samantha Keenan, BSc;

16

Royal Glamorgan Hospital: Jayaprakash Singh, MD, Ceri Lynch, MD, Lisa Roche, Justyna Mikusek, Bethan Deacon, Keri Turner;

Royal Gwent Hospital: Tamas Szakmany, PhD, Evelyn Baker, MSc, Shiney Cherian, BSc(Hons), John Hickey, MSc, Shreekant Champanerkar, MBBS ;

Royal Hallamshire Hospital, Sheffield: Gary H Mills, PhD, Ajay Raithatha, FFICM, Kris Bauchmuller, FFICM, Norfaizan Ahmad, FFICM, Matt Wiles FFICM, Jayne Willson, RN;

Royal Hampshire Hospitals: Irina Grecu, MD, Jane Martin, Caroline Wrey Brown, Ana-Marie Arias, Emily Bevan;

Royal Infirmary of Edinburgh: Thomas H Craven, PhD, David Hope, PGDip, Jo Singleton, BN, Sarah Clark, MNurs, Corrienne McCulloch, PhD;

Royal Liverpool University Hospital: Ingeborg D Welters, PhD, David Oliver Hamilton, BMBS, Karen Williams, RGN, Victoria Waugh, BA, David Shaw, DipHE, Suleman Mulla, MBChB, Alicia Waite, PhD, Jaime Fernandez Roman, BSc, Maria Lopez Martinez, BSc;

Royal London Hospital: Zudin Puthucheary, PhD, Timothy Martin, BA(Hons), Filipa Santos, RN, Ruzena Uddin, MSc(Hons), Maria Fernandez, MSc, Fatima Seidu, MSc, Alastair Somerville, MSc, Mari Lis Pakats, MSc, Priya Dias, PhD, Salam Begum, BSc, Tasnin Shahid, BSc;

The Royal Free Hospital: Sanjay Bhagani, FRCP, Mark De Neef, MSc, Helder Filipe, BSc, Sara Mingos, BSc, Amitaa Maharajh, BA, Glykeria Pakou, BA, Aarti Nandani, MPharm;

The Royal Marsden NHS Foundation Trust: Kate Colette Tatham, PhD, Shaman Jhanji, PhD, Ethel Black, BSNurs, Arnold Dela Rosa, BSNurs, Ryan Howle, FRCA, Ravishankar Rao Baikady, FRCA;

The Royal Oldham Hospital: Redmond P Tully, FFICM, Andrew Drummond, FFICM, Joy Dearden, BSc, Jennifer E Philbin, MSc, Sheila Munt, SRN;

The Royal Wolverhampton NHS Trust: Shameer Gopal, MBBCh, Jagtar- Singh Pooni, MBBS, Saibal Ganguly, MBBS, Andrew Smallwood, RGN, Stella Metherell, RGN;

Royal Papworth Hospital: Alain Vuylsteke, MD, Charles Chan, FRCA, Saji Victor, COVID Research Team, Papworth Hospital;

Royal Stoke Hospital: Ramprasad Matsa, FRCP, Minerva Gellamucho, BSN, Michelle Davies, NVQ; Royal Surrey County Hospital: Ben Creagh-Brown, PhD, Joe Tooley, MSc, Laura Montague, BSc, Fiona

De Beaux, BSc, Laetitia Bullman, MBChB; Royal United Hospital Bath: Ian Kerslake, FFICM, Carrie Demetriou, RN, Sarah Mitchard, MBBS, Lidia

Ramos, RN, Katie White, MSc; Russells Hall Hospital: Michael Reay, FFICM, Steve Jenkins, MD, Caroline Tuckwell, Angela Watts,

BSc, Eleanor Traverse, Stacey Jennings; Salisbury NHS Foundation Trust: Phil Donnison, FFICM, Maggie Johns, RGN, Ruth Casey, BSc,

Lehentha Mattocks, Dip, Sarah Salisbury; Salford Royal NHS Foundation Trust: Paul Dark, PhD, Alice Harvey, BSc, Reece, Doonan, BSc, Liam

McMorrow, BA (Hons), Karen Knowles, BA (Hons);Sandwell and West Birmingham NHS Trust: Jonathan Hulme, FFICM, Santhana Kannan, FFICM, Sibet Joseph, BSc, Fiona Kinney, RGN, Ho Jan Senya, BPharm;

Sherwood Forest Hospitals NHS Foundation Trust: Valli Ratnam, MD, Mandy Gill, Jill Kirk, Sarah Shelton

South Tyneside District Hospital: Christian Frey, MD, Riccardo Scano, MD, Madeleine McKee, BSc, Peter Murphy, BSc;

Southmead Hospital: Matt Thomas, FFICM, Ruth Worner, RGN, Beverley Faulkner, RGN, Emma Gendall, BSc, Kati Hayes, BSc, Hayley Blakemore, BSc, Borislava Borislavova, MSc;

St. Bartholomew’s Hospital: Colin Hamilton-Davies, MBBS, Carmen Chan, BSc, Celina Mfuko, BSc, Hakam Abbass, MSc, Vineela Mandadapu, MSc;

17

St. George’s Hospital: Susannah Leaver, MRCP, Kamal Patel, MRCP, Sarah Farnell-Ward, MSc, Romina Pepermans Saluzzio, BSc, John Rawlins, MBBS;

St. Mary’s Hospital: Anthony Gordon, MD; Dorota Banach, BSc, Ziortza Fernández de Pinedo Artaraz, BN, Leilani Cabreros, BSN;

St. Peter’s Hospital, Chertsey: Ian White, FFICM, Maria Croft, BSc(Hons), Nicky Holland, BN(Hons), Rita Pereira, MPharm;

Stepping Hill Hospital, Stockport: Ahmed Zaki, PhD, David Johnson, MPhil, Matthew Jackson, MBChB, Hywel Garrard, BMBS, Vera Juhaz, MD, Louise Brown BSc (Hons);

Sunderland Royal Hospital: Alistair Roy, MBChB, Anthony Rostron, PhD, Lindsey Woods, BSc, Sarah Cornell, BSc;

Swansea Bay University Health Board: Suresh Pillai, FFCIM, Rachel Harford, RN, Helen Ivatt, FRCA, Debra Evans, BN, Suzanne Richards, BN, Eilir Roberts, MBBCh, James Bowen, MBBCh James Ainsworth, MBBS;

Torbay and South Devon NHS Foundation Trust: Thomas Clark, MBChB, FRCA, FFICM, Angela Foulds, BSc, Stacey Atkins, Dip RN;

United Lincolnshire NHS Trust: Kelvin Lee, PhD, Russell Barber, FRCA, Anette Hilldrith, RGN, Claire Hewitt, RGN, Gunjan Phalod, MPharm;

University Hospitals Coventry & Warwickshire NHS Trust: Pamela Bremmer, BSc, Geraldine Ward, MA, Christopher Bassford, PhD;

University Hospital of North Tees: Farooq Brohi, FFARCSI, Vijay Jagannathan, FRCA, Michele Clark, MA, Sarah Purvis, Dip, Bill Wetherill, MSc;

University Hospital Southampton NHS Foundation Trust: Ahilanandan Dushianthan, PhD, Rebecca Cusack, MD, Kim de Courcy-Golder, PGDip, Karen Salmon, MSc, Rachel Burnish, Simon Smith, BN, Susan Jackson, BSc, Winningtom Ruiz, BSc (Hons), Zoe Duke BSc (Hons) Magaret Johns, BA, Michelle Male, Dip, Kirsty Gladas, BSc (Hons), Satwinder Virdee, MPharm, Jacqueline Swabe, MPharm, Helen Tomlinson;

Warwick Hospital: Ben Attwood, MBBCh, Penny Parsons, BSc, Bridget Campbell, BSc, Alex Smith, BSc;

Watford General Hospital: Valerie J Page, MBBCh, Xiao Bei Zhao, BSc (Hons), Deepali Oza, BPharm, Gail Abrahamson, DPhil, Ben Sheath, BSc (Hons), Chiara Ellis, BSc (Hons);

Western General Hospital, Edinburgh: Jonathan Rhodes, PhD, Thomas Anderson, MBChB, Sheila Morris;

Whipps Cross Hospital: Charlotte Xia Le Tai, MBChB, Amy Thomas, MSc, Alexandra Keen, MSc; Whiston Hospital: Dr Ascanio, Tridente, Karen Shuker, Jeanette Anders, Sandra Greer, Paula Scott,

Amy Millington, Philip Buchanan, Jodie Kirk Wirral University Teaching Hospital NHSFT: Craig Denmade, MBChB, Girendra Sadera, MBBS, Reni

Jacob, BSc, Cathy Jones, BSc, Debbie Hughes; Worcester Royal Hospital: Stephen Digby, MBBS, Nicholas Cowley, MD, Laura Wild, BSc(Hons),

Jessica Thrush, RGN, Hannah Wood, BSc(Hons), Karen Austin, RGN; Wrexham Maelor Betsi Cadwaladr University Hospital: David Southern, FFICM, Harsha Reddy, FFICM,

Sarah Hulse, BSc, Andy Campbell, FFICM, Mark Garton, Claire Watkins, PGDip, Sara Smuts, BN;

Wrightington, Wigan and Leigh Teaching Hospitals NHS Foundation Trust: Alison Quinn, MD, Benjamin Simpson, MD, Catherine McMillan, MD, Cheryl Finch, BSc, Claire Hill BSc, Josh Cooper;

Wye Valley NHS Trust: Joanna Budd, MBBS, Charlotte Small, PhD, Ryan O’Leary, MBBS, Janine Birch, RN, Emma Collins, BSc (Hons);

Wythenshawe Hospital: Peter D G Alexander, FFICM, Tim Felton, FFICM, Susan Ferguson, BSc, Katharine Sellers, BSc, Joanne Bradley-Potts, BSc;

18

York Teaching Hospital: David Yates, FCRA, Isobel Birkinshaw, BSc(Hons), Kay Kell, BSc(Hons), Zoe Scott, BN, Harriet Pearson, BSc;

United States of America University of Pittsburgh Medical Center: David Huang, MD, Tim Girard, MD, Kim Basile, BS, Erin

McCreary, PharmD, Kelsey Linstrum, MS;

19

2. Additional methods

2.1 Additional information about REMAP-CAP general design

REMAP-CAP is an international, multicenter, open-label adaptive platform trial to determine

best treatment strategies for patients with severe pneumonia in both pandemic and non-

pandemic settings. A full description of the trial design is provided elsewhere1 and below is

a short summary.

The trial uses a novel design entitled a randomized embedded multifactorial adaptive

platform (REMAP).2 The design has five key features: randomization, allowing causal

inference; embedding of study procedures into routine clinical care, facilitating enrollment

and generalizability; a multifactorial statistical model comparing multiple interventions

across multiple patient subgroups; response-adaptive randomization with preferential

assignment to those interventions that appear most favorable after interim analyses, and; a

platform structured to permit continuous, potentially perpetual, enrollment.

Once the Covid-19 outbreak began a new pandemic stratum was added within REMAP-CAP.

Specifically, all patients hospitalized with suspected or proven Covid-19 are assigned to the

pandemic stratum. They are further classified as being in either moderate or severe disease

state. Severe disease state is defined by receiving respiratory or cardiovascular organ failure

support in an intensive care unit (ICU). All other hospitalized patients are defined as being in

the moderate state.

This report relates to the analysis of the interleukin-6 receptor antagonists in the Immune

Modulation Therapy domain up to November 19, 2020. Other domains open to

randomisation within the pandemic stratum severe state up to this date were:

• Corticosteroid domain (closed June 17, 2020)

20

• Covid-19 Antiviral domain (closed November 19, 2020)

• Covid-19 Immunoglobulin domain

• Therapeutic Anticoagulation domain

• Macrolide Duration domain

• Vitamin C domain

• Anti-platelet domain

• Statin Therapy domain

2.2 REMAP-CAP Eligibility Criteria for suspected or proven Covid-19

Platform inclusion criteria:

• Adult patient admitted to hospital with acute illness due to suspected or proven

pandemic (Covid-19) infection

Platform exclusion criteria:

• Death is deemed to be imminent and inevitable during the next 24 hours AND one or

more of the patient, substitute decision maker or attending physician are not

committed to full active treatment

• Patient is expected to be discharged from hospital today or tomorrow

• More than 14 days have elapsed while admitted to hospital with symptoms of an

acute illness due to suspected or proven pandemic infection

• Previous participation in this REMAP within the last 90 days

Immune Modulation Therapy domain specific inclusion criteria:

• Severe disease state, defined by receiving respiratory or cardiovascular organ failure

support in an intensive care unit (ICU).

a. Respiratory organ support is defined as invasive or non-invasive mechanical

ventilation including via high flow nasal cannula if flow rate >30 L/min and

FIO2 >0.4. If non-invasive ventilation would normally be provided but is being

withheld, due to infection control concerns associated with aerosol

21

generating procedures, then the patient still meets the severe disease state

criteria.

b. Cardiovascular organ support was defined as the intravenous infusion of any

vasopressor or inotrope.

c. Pandemic surge capacity means that provision of advanced organ support

may need to occur in locations that do not usually provide ICU-level care.

Therefore, an ICU is defined as an area within the hospital that is repurposed

so as to be able to deliver one or more of the qualifying organ failure

supports (non-invasive ventilation, invasive ventilation, and vasopressor

therapy)

• Microbiological testing for SARS-CoV-2 of upper or lower respiratory tract secretions

or both has occurred or is intended to occur

Immune Modulation Therapy domain specific exclusion criteria:

• More than 24 hours has elapsed since ICU admission

• Patient has already received any dose of one or more of any form of interferon,

anakinra, tocilizumab, or sarilumab during this hospitalization or is on long-term

therapy with any of these agents prior to this hospital admission

• Known condition or treatment resulting in ongoing immune suppression including

neutropenia prior to this hospitalization

• Patient has been randomized in a trial evaluating an immune modulation agent for

proven or suspected Covid-19 infection, where the protocol of that trial requires

ongoing administration of study drug

• The treating clinician believes that participation in the domain would not be in the

best interests of the patient

Intervention specific exclusion criteria:

• Known hypersensitivity to an agent specified as an intervention in this domain will

exclude a patient from receiving that agent

• Intention to prescribe systemic corticosteroids for any reason, other than

participation in the Corticosteroid domain of this platform, will result in exclusion

from receiving IFN-β1a

22

• Known hypersensitivity to proteins produced by E. coli will result in exclusion from

receiving anakinra

• Known or suspected pregnancy will result in exclusion from the anakinra, IFN-β1a,

tocilizumab, and sarilumab interventions. It is normal clinical practice that women

admitted who are in an age group in which pregnancy is possible will have a

pregnancy test conducted. The results of such tests will be used to determine

interpretation of this exclusion criteria.

• A baseline alanine aminotransferase or an aspartate aminotransferase that is more

than five times the upper limit of normal will result in exclusion from receiving

tocilizumab or sarilumab

• A baseline platelet count < 50 x 109 / L will result in exclusion from receiving

tocilizumab or sarilumab

2.2 Additional Statistical Methods:

Secondary outcomes:

• Hospital mortality / survival (dichotomous)

• 90-day survival (time to event)

• Respiratory support-free days

• Cardiovascular support-free days

• Time to ICU discharge

• Time to hospital discharge

• WHO ordinal scale (range: 0-8, where 0 = no illness, 1-7 = increasing level of care,

and 8 = death) assessed at day 14.3 In this analysis categories 0,1,2 have been

condensed into one category for all patients discharge from hospital.

• Progression to invasive mechanical ventilation, extracorporeal membrane

oxygenation (ECMO), or death among those not ventilated at baseline

Analysis populations:

The primary analysis was conducted on all severe state patients with Covid-19 randomized

to any domain allowing maximal incorporation of all information, providing the most robust

estimation of the coefficients of all covariates. As the primary model included information

23

about assignment to interventions within domains whose evaluation is ongoing and blinded

to the investigators, the primary analysis was run by the fully unblinded Statistical Analysis

Committee (SAC), who conduct all protocol-specified trial update analyses and report those

results to the DSMB.

In addition to the primary analysis conducted by the Statistical Analysis Committee, analysis

of the primary outcome was then repeated in a second model using only data from those

patients enrolled in those domains that had ceased recruitment, Corticosteroid domain and

Covid-19 Antiviral domain, as well as those patients randomised in the Immune Modulation

Therapy domain; the Unblinded Intention To Treat (ITT) population. For this analysis there

was no adjustment for assignment to interventions in other ongoing domains. These

analyses were conducted by the investigator team who remain blinded to on-going domains

and interventions in the platform.

Further secondary analyses explored the effects of excluding patients who were ruled out

for Covid-19 (defined as documented negative tests for SARS-CoV-2 infection and no

positive tests). As described in the main text the analytical model includes a site covariate

term. The site covariate adjusts for site-level differences in the distribution of organ

support-free days that may be due to the patient characteristics or standard treatment

practices at the site. If a site was not taking part in the Immune Modulation Therapy

domain, these patients would not be included as controls in the analysis. If a site recruiting

in the Immune Modulation Therapy domain did not have tocilizumab or sarilumab available,

then those patients allocated to “no immune modulation” would be controls in the primary

model, adjusted for site, time and other effects. We therefore also carried out other

sensitivity analyses including only those patients randomised directly to tocilizumab or

sarilumab versus control and without any interactions from other domains (IL-6 RA specific

24

ITT population) and a subset of this population treated as per protocol (IL-6 RA per protocol

population)

One subgroup was defined in the statistical analysis plan, dividing patients into terciles

according to serum C-reactive protein levels at study inclusion.

Various sensitivity analyses were conducted including less informative standard normal

priors on pre-specified interactions excluding adjustment for site and time epoch, and

alternative coding of steroid interventions after closure of the corticosteroid domain.

Identical analyses were conducted to estimate the effect on mortality, except the outcome

was dichotomous (alive or dead at hospital discharge).

There was no imputation of missing outcomes in either primary or secondary analyses.

Cases were excluded on an analysis-by-analysis basis, i.e. patients missing outcome data

were included in treatment compliance and safety analyses. The secondary outcomes were

also analysed in the unblinded ITT model. The time-to-death and length of stay outcomes

were time-to-event analyses and used a piecewise exponential Bayesian model to estimate

hazard ratios. The primary safety analysis compared the proportion of patients who

developed one or more serious adverse events across groups.

25

3. Additional Results

Site Participation in the Immune Modulation Domain

During the study period, April 19 to November 19, 2020, 125 sites were open for enrollment

in the Immune Modulation Domain. There were 113 sites in six countries open to

randomisation for tocilizumab and / or sarilumab, 98 sites in the UK, 7 in The Netherlands, 3

in Australia, 2 in New Zealand, 2 in Ireland and 1 in the Kingdom of Saudi Arabia.

Platform exclusions:

From the patient flow figure 1

764 patients were excluded from the REMAP-CAP platform, severe state:

2 - Prior enrollment in REMAP-CAP within 90 days 5 - No lower respiratory signs or symptoms * 11 - No radiological infiltrate * 318 - No qualifying organ support in ICU 375 - Exceeded platform eligibility time windows 32 - Death deemed imminent and inevitable 136 - Other

* = exclusion criteria from the original core protocol for community acquired pneumonia,

but dropped once Covid-19 specific amendments described in the Pandemic Appendix to

Core protocol were implemented (see protocol). Patients could be excluded for more than

one reason.

In addition, 208 patients were potentially eligible for the platform, were screened for

inclusion in the Immune Modulation Therapy domain but were excluded and not included in

the platform in any other domain

2 - Samples for Covid-19 not taken or intended 95 - Admitted to ICU >24 hours earlier 2 - Already received immune modulation during hospital admission

26

19 - Known immune suppression 6 - Enrolled in another trial 22 - Contraindication to agents in domain ^ 15 - Not considered in patient’s best interests 107 - Prospective consent declined

^ Contraindications include hypersensitivity, raised ALT/AST, or thrombocytopenia, or

pregnancy

Additional analyses:

The additional analyses specified in the statistical analysis plan are provided in the “ITSC

Secondary Analysis Report” appendix to this supplement, with the exception of the anti-viral

interactions with immune modulation interventions. The Covid-19 antiviral primary analysis

has not yet been completed or reported, and these treatment interactions will be included in

that subsequent report. Provided below are the sensitivity analyses of the primary outcome

(organ support-free days and hospital survival), with patients with only PCR negative tests

removed and with no treatment interactions included (Table S3). There was one additional

post-hoc analysis undertaken and presented that does not use any borrowing between the

two IL-6 receptor antagonists, tocilizumab and sarilumab (Table S4). In this model they are no

longer “nested”.

Table S5 and S6 provide the odds ratios (OR) and 95% CrI for each parameter in the second

model using data from those patients enrolled in those domains that had ceased recruitment,

Corticosteroid domain and Covid-19 Antiviral domain, as well as those patients randomised

in the Immune Modulation Therapy domain; the Unblinded ITT population.

The pre-specified C-reactive protein (CRP) subgroup analysis is in Table S8. The cut points of

CRP terciles were 102 and 187 µg/ml. Table S9 provides post-hoc subgroup analyses

according to mechanical ventilation status at baseline.

27

Table S1. Baseline Characteristics of Participants in the Immune Modulation Therapy domain *

Characteristic

Tocilizumab (N=353)

Sarilumab (N=48)

Control a (N=402)

All participants in the Immune Modulation

Therapy domain b (N=865)

Age - mean (SD), years 61.5 (12.5) 63.4 (13.4) 61.1 (12.8) 61.4 (12.7)

Male Sex - n (%) 261 (73.9) 39 (81.3) 283 (70.4) 629 (72.7)

Race/Ethnicity c -

White - n/N (%) 160/228 (70.2) 29/39 (74.4) 206/279 (73.8) 420/580 (72.4)

Asian - n/N (%) 41/228 (18.0) 8/39 (20.5) 47/279 (16.9) 99/580 (17.1)

Black - n/N (%) 12/228 (5.3) 1/39 (2.6) 9/279 (3.2) 23/580 (4.0)

Mixed - n/N (%) 2/228 (0.9) 0/39 (0.0) 5/279 (1.8) 7/580 (1.2)

Other - n/N (%) 13/228 (5.7) 1/39 (2.6) 12/279 (4.3) 31/580 (5.3)

Body mass index d - median (IQR), kg/m2

30.5 (26.9-34.9) (n=342)

29.2 (26.0-33.8) (n=39)

30.9 (27.1-34.9) (n=377)

30.5 (26.8-34.9) (n=815)

APACHE II score e - median (IQR) 13 (8-19) (n=337) 10 (7-16) (n=42) 12 (8-18) (n=381) 12.5 (8-19) (n=820)

Confirmed SARS-CoV2 infectionf – n/N (%)

284/345 (82.3) 44/47 (93.6) 334/394 (84.8) 715/847 (84.4)

Pre-existing conditions – n/N (%)

Diabetes mellitus 123/349 (35.2) 13/48 (27.1) 150/401 (37.4) 304/860 (35.4)

Kidney disease 30/312 (9.6) 4/45(8.9) 43/372 (11.6) 81/789 (10.3)

Respiratory disease¤ 82/349 (23.5) 15/48 (31.3) 98/401 (24.4) 206/860 (24.0)

Immunosuppressive disease 8/348 (2.3) 0/48 (0.0) 14/401 (3.5) 25/859 (2.9)

Chronic immunosuppressive therapy 3/349 (0.9) 1/48 (2.1) 6/401 (1.5) 12/860 (1.4)

Severe cardiovascular disease 34/339 (10.0) 1/48 (2.1) 47/395 (11.9) 86/844 (10.2)

Liver cirrhosis/failure 2/339 (0.6) 0/48 (0.0) 1/395 (0.3) 5/844 (0.6)

Time to enrollment - median (IQR)

From hospital admission - days 1.2 (0.8-2.8) 1.4 (0.9-2.8) 1.2 (0.8-2.8) 1.2 (0.8-2.8)

From ICU admission - hours 13.1 (6.6-19.0) 16.0 (11.4-20.8) 14.0 (6.8-19.5) 13.6 (6.6-19.4)

Acute respiratory support

None/supplemental oxygen only 1/353 (0.3) 0/48 (0.0) 2/402 (0.5) 3/865 (0.4)

High flow nasal cannula 101/353 (28.6) 17/48 (35.4) 110/402 (27.4) 249/865 (28.8)

Non-invasive ventilation only 147/353 (41.6) 23/48 (47.9) 169/402 (42.0) 359/865 (41.5)

28

Invasive mechanical ventilation 104/353 (29.5) 8/48 (16.7) 121/402 (30.1) 254/865 (29.4)

Vasopressor support 63/353 (17.9) 4/48 (8.3) 79/402 (19.7) 163/865 (18.8)

PaO2/FIO2 – median (IQR) 115 (89-162)

(n=335) 126 (99-157)

(n=35) 118 (89-169)

(n=354) 116.5 (89-165)

(n=780)

Median laboratory values (IQR)g

C-reactive protein, µg/mL 150 (85-221)

(n=207) 136 (105-204)

(n=37) 130 (71-208)

(n=244) 136 (79-208)

(n=533)

D-dimer, ng/mL 832 (461-1763)

(n=159) 828 (355-1435)

(n=20) 1010 (500-2115)

(n=172) 910 (480-1916)

(n=385)

Ferritin, ng/mL 912 (525-1590)

(n=201) 1914 (696-2315)

(n=13) 887 (410-1573)

(n=199) 929 (472-1643)

(n=447)

Neutrophils, x109/L 8 (5.8-10.7)

(n=239) 7.7 (5.4-10.2)

(n=47) 7.9 (5.3-11.0)

(n=278) 7.9 (5.6-10.7)

(n=612)

Lymphocytes, x109/L 0.7 (0.5-1.0)

(n=239) 0.6 (0.4-0.9)

(n=47) 0.7 (0.5-1.0)

(n=279) 0.7 (0.5-1.0)

(n=613)

Platelets, x109/L 250 (183-323)

(n=346) 273 (205-319)

(n=48) 236 (177-297)

(n=398) 245 (182-311)

(n=854)

Lactate, mmol/L 1.3 (1.0-1.8)

(n=303) 1.5 (1.2-2.0)

(n=40) 1.4 (1.0-1.9)

(n=353) 1.4 (1.0-1.9)

(n=754)

Creatinine, mg/dL 0.8 (0.7-1.2)

(n=348) 1.0 (0.6-1.4)

(n=48) 0.9 (0.7-1.2)

(n=398) 0.9 (0.7-1.2)

(n=856)

Bilirubin, mg/dL 0.11 (0.08-0.15)

(n=327) 0.11 (0.09-0.18)

(n=46) 0.10 (0.08-0.15)

(n=382) 0.11 (0.08-0.15)

(n=817)

* Percentages may not sum to 100 because of rounding. SD denotes standard deviation; APACHE, Acute Physiology and Chronic Health Evaluation; COPD, chronic obstructive pulmonary disease; IQR, interquartile range; ECMO, extracorporeal membrane oxygenation. a Control patients include all patient randomized to control who were also eligible to be randomized to tocilizumab and/or sarilumab, i.e. direct concurrent controls. b All patients includes patient randomized in the Immune Modulation Therapy domain, including control (including where tocilizumab and sarilumab were not a randomization option, i.e. non-direct controls), tocilizumab, sarilumab, anakinra and interferon-beta-1a. c Data collection not approved in Canada and continental Europe. “Other” includes “declined” and “multiple”. d The body-mass index (BMI) is the weight in kilograms divided by the square of the height in meters. e Range 0 to 71, with higher scores indicating greater severity of illness. f SARS-CoV2 infection was confirmed by respiratory tract polymerase chain reaction test g Values were from the sample collected closest to randomization, up to 8 hours prior to randomization. If no samples were collected up to 8 hours prior to time of randomization, the sample collected closest to the time of randomization up to 2 hours after randomization was used (other than PaO2/FIO2 which was a pre-randomization value only). Laboratory values were only added to the case report form on August 6, 2020.

29

Table S2. Randomization into other REMAP-CAP domains and other treatments given

Domain

Tocilizumab (N=353)

Sarilumab (N=48)

Control (N=402)

All participants in the immune modulation

domain (N=865)

Corticosteroids, n (%) 50 (14.2) 0 (0.0) 52 (12.9) 107 (12.4)

Covid-19 Antiviral, n (%) 169 (47.9) 26 (54.2) 217 (54.0) 440 (50.9)

Covid-19 Immunoglobulin, n (%) 175 (49.6) 41 (85.4) 202 (50.3) 448 (51.8)

Therapeutic Anticoagulation, n (%)

119 (33.7) 31 (64.6) 146 (36.3) 331 (38.3)

Macrolide, n (%) 24 (6.8) 5 (10.4) 27 (6.7) 61 (7.1)

Vitamin C, n (%) 0 (0.0) 0 (0.0) 0 (0.0) 1 (0.1)

Antiplatelet, n (%) 7 (2.0) 3 (6.3) 4 (1.0) 14 (1.6)

Statins Therapy n (%) 13 (3.7) 8 (16.7) 13 (3.2) 41 (4.7)

Other treatments given as part of usual clinical care

Corticosteroids within 48 hours of randomization, n (%)

(patients enrolled post June17, 2020)

252/272 (92.7) 46/48 (95.8) 293/312 (93.9) 610/654 (93.3)

Remdesivir within 48 hours of randomization, n (%) 107/341 (31.4) 21/48 (43.8) 133/389 (34.2) 265/807 (32.8)

30

Table S3. Sensitivity analyses of the Primary Outcomes

Outcome/Analysis Tocilizumab (N=353) Sarilumab (N=48) Control (N=402)

Secondary Analysis of Primary Outcome, model restricted Immune Modulation Therapy domain patients and other closed domains, with PCR negative patients removed

Adjusted OR - mean (SD) 1.72 (0.26) 1.94 (0.50) 1

- median (95% CrI) 1.70 (1.26, 2.27) 1.86 (1.22, 3.19) 1

Probability of superiority to control, % >99.9 99.7 -

Secondary analysis of Hospital Survival, model restricted Immune Modulation Therapy domain patients and other closed domains, with PCR negative patients removed

Adjusted OR - mean (SD) 1.65 (0.32) 2.36 (1.10) 1

- median (95% CrI) 1.62 (1.12, 2.38) 2.07 (1.18, 5.24) 1

Probability of superiority to control, % 99.6 99.6 -

Secondary Analysis of Primary Outcome, model restricted to the IL-6 RA population only with no treatment interactions included

Adjusted OR - mean (SD) 1.69 (0.23) 1.91 (0.47) 1

- median (95% CrI) 1.67 (1.29 to 2.18) 1.83 (1.23 to 3.04) 1

Probability of superiority to control, % >99.9 99.9 -

Secondary analysis of Hospital Survival, model restricted to the IL-6 RA population only with no treatment interactions included

Adjusted OR - mean (SD) 1.72 (0.31) 2.31 (1.01) 1

- median (95% CrI) 1.69 (1.19 to 2.42) 2.05 (1.22 to 5.08) 1

Probability of superiority to control, % 99.8 99.7 -

All models are structured such that a higher OR or HR is favorable. The WHO scale ranges from 0 (no disease) to 8 (death).

SD - standard deviation; CrI - credible interval; OR - odds ratio; HR – Hazard ratio; ECMO – Extra-Corporeal Membrane Oxygenation.

31

Table S4. Exploratory post-hoc sensitivity analyses of the Primary Outcomes

Outcome/Analysis Tocilizumab (N=353) Sarilumab (N=48) Control (N=402)

Exploratory post-hoc Analysis of Primary Outcome, model restricted to Immune Modulation Therapy Domain participants, with no borrowing between IL-6 RA interventions

Adjusted OR - mean (SD) 1.64 (0.23) 1.98 (0.60) 1

- median (95% CrI) 1.62 (1.24 to 2.14) 1.90 (1.08 to 3.41) 1

Probability of superiority to control, % >99.9 98.7 -

Exploratory post-hoc Analysis of Hospital Survival, model restricted to Immune Modulation Therapy Domain participants, with no borrowing between IL-6 RA interventions

Adjusted OR - mean (SD) 1.58 (0.30) 2.73 (1.13) 1

- median (95% CrI) 1.56 (1.08 to 2.24) 2.54 (1.17 to 5.56) 1

Probability of superiority to control, % 99.1 99.2 -

All models are structured such that a higher OR or HR is favorable. The WHO scale ranges from 0 (no disease) to 8 (death).

SD - standard deviation; CrI - credible interval; OR - odds ratio; HR – Hazard ratio; ECMO – Extra-Corporeal Membrane Oxygenation.

32

Table S5. Odds ratios for secondary analysis of organ support free days for the unblinded ITT population Parameter Mean SD Median CrI Age<39 4.11 0.91 4.01 (2.61, 6.17) Age 40-49 2.11 0.36 2.08 (1.49, 2.88) Age 50-59 1.96 0.27 1.94 (1.48, 2.56) Age 70-79 0.52 0.08 0.51 (0.38, 0.69) Age 80+ 0.34 0.09 0.32 (0.19, 0.55) Female 1.17 0.13 1.16 (0.93, 1.44) Time epoch 1 0.95 0.08 0.95 (0.79, 1.11) Time epoch 2 0.91 0.13 0.90 (0.66, 1.17) Time epoch 3 0.96 0.17 0.95 (0.67, 1.33) Time epoch 4 1.06 0.22 1.04 (0.70, 1.56) Time epoch 5 1.16 0.27 1.13 (0.74, 1.78) Time epoch 6 1.22 0.29 1.18 (0.77, 1.88) Time epoch 7 1.26 0.29 1.22 (0.80, 1.92) Time epoch 8 1.29 0.30 1.25 (0.83, 1.98) Time epoch 9 1.23 0.27 1.20 (0.80, 1.88) Time epoch 10 1.13 0.24 1.10 (0.74, 1.69) Time epoch 11 0.97 0.20 0.95 (0.64, 1.41) Time epoch 12 0.86 0.19 0.84 (0.55, 1.27) Time epoch 13 0.85 0.19 0.83 (0.54, 1.26) Time epoch 14 0.94 0.22 0.92 (0.59, 1.43) Time epoch 15 1.11 0.32 1.06 (0.61, 1.86) Time epoch 16 1.41 0.66 1.26 (0.59, 3.08) Tocilizumab 1.68 0.24 1.66 (1.26, 2.18) Sarilumab 1.84 0.44 1.77 (1.18, 2.90) Fixed-dose corticosteroid 1.45 0.33 1.42 (0.92, 2.18) Shock-dependent corticosteroid 1.15 0.26 1.12 (0.72, 1.75) Tocilizumab*Fixed-dose interaction effect 1.00 0.05 1.00 (0.91, 1.10) Tocilizumab*Shock-dependent interaction effect 1.00 0.05 1.00 (0.91, 1.10) Sarilumab*Fixed-dose interaction effect 1.01 0.05 1.00 (0.91, 1.11) Sarilumab*Shock-dependent interaction effect 1.00 0.05 1.00 (0.91, 1.10) Tocilizumab*Fixed-dose combination 2.42 0.65 2.34 (1.39, 3.91) Tocilizumab*Shock-dependent combination 1.93 0.54 1.86 (1.08, 3.17) Sarilumab*Fixed-dose combination 2.68 0.90 2.53 (1.37, 4.86) Sarilumab*Shock-dependent combination 2.11 0.72 1.99 (1.08, 3.87)

The odds ratio (OR) effects for pre-specified combinations of interventions from the Immune Modulation Therapy and Corticosteroid domains are reported relative to control. These ORs incorporate the effect of each individual intervention and the interaction term for the combination of interventions. OR effects for pre-specified interaction effects are reported relative to and additive effect. For example, if an interaction effect OR is equal to 1, the combination of interventions is additive (equal to the sum of the effects of the two interventions taken separately). If an interaction OR is greater than 1, the effect of the combination of interventions is synergistic (greater than the sum of the effects of the two interventions taken separately). If the interaction OR is less than 1, the effect of the combination is sub-additive (less than the sum of the effects of the two interventions taken separately).

33

Table S6. Odds ratios for secondary analysis of hospital survival for the unblinded ITT population

Parameter Mean SD Median CrI Age<39 12.85 6.45 11.38 (5.00, 28.81) Age 40-49 3.33 0.89 3.21 (1.95, 5.45) Age 50-59 2.99 0.60 2.93 (2.00, 4.34) Age 70-79 0.41 0.07 0.41 (0.29, 0.58) Age 80+ 0.23 0.07 0.22 (0.12, 0.39) Female 1.19 0.19 1.18 (0.87, 1.59) Time epoch 1 0.94 0.08 0.94 (0.78, 1.10) Time epoch 2 0.87 0.14 0.87 (0.61, 1.14) Time epoch 3 0.86 0.17 0.85 (0.55, 1.22) Time epoch 4 0.87 0.20 0.86 (0.53, 1.33) Time epoch 5 0.90 0.23 0.87 (0.53, 1.41) Time epoch 6 0.91 0.24 0.88 (0.53, 1.48) Time epoch 7 0.92 0.25 0.89 (0.52, 1.48) Time epoch 8 0.95 0.25 0.92 (0.55, 1.53) Time epoch 9 0.99 0.26 0.95 (0.58, 1.59) Time epoch 10 1.02 0.27 0.98 (0.60, 1.66) Time epoch 11 1.02 0.27 0.97 (0.60, 1.65) Time epoch 12 0.99 0.26 0.96 (0.58, 1.58) Time epoch 13 1.00 0.28 0.96 (0.57, 1.63) Time epoch 14 1.08 0.33 1.03 (0.57, 1.86) Time epoch 15 1.23 0.50 1.14 (0.55, 2.43) Time epoch 16 1.52 1.03 1.27 (0.50, 3.95) Tocilizumab 1.67 0.31 1.65 (1.15, 2.34) Sarilumab 2.24 0.94 2.00 (1.17, 4.69) Fixed-dose corticosteroid 0.99 0.29 0.95 (0.54, 1.68) Shock-dependent corticosteroid 1.21 0.38 1.15 (0.63, 2.10) Tocilizumab*Fixed-dose interaction 0.99 0.05 0.99 (0.89, 1.09) Tocilizumab*Shock-dependent interaction 1.00 0.05 1.00 (0.91, 1.11) Sarilumab*Fixed-dose interaction 1.01 0.05 1.00 (0.91, 1.11) Sarilumab*Shock-dependent interaction 1.00 0.05 1.00 (0.91, 1.10) Tocilizumab*Fixed-dose combination 1.63 0.58 1.53 (0.77, 3.03) Tocilizumab*Shock-dependent combination 2.03 0.76 1.90 (0.92, 3.87) Sarilumab*Fixed-dose combination 2.21 1.14 1.94 (0.86, 5.21) Sarilumab*Shock-dependent combination 2.69 1.42 2.35 (1.02, 6.35)

The odds ratio (OR) effects for pre-specified combinations of interventions from the Immune Modulation Therapy and Corticosteroid domains are reported relative to control. These ORs incorporate the effect of each individual intervention and the interaction term for the combination of interventions. OR effects for pre-specified interaction effects are reported relative to and additive effect. For example, if an interaction effect OR is equal to 1, the combination of interventions is additive (equal to the sum of the effects of the two interventions taken separately). If an interaction OR is greater than 1, the effect of the combination of interventions is synergistic (greater than the sum of the effects of the two interventions taken separately). If the interaction OR is less than 1, the effect of the combination is sub-additive (less than the sum of the effects of the two interventions taken separately).

34

Table S7. Secondary Outcomes Outcome/Analysis Tocilizumab (N=353) Sarilumab (N=48) Control (N=402)

90-day Survival (time to event)

Adjusted HR - mean (SD) 1.60 (0.21) 1.94 (0.56) 1

- median (95% CrI) 1.59 (1.24 to 2.05) 1.82 (1.22 to 3.38) 1

Probability of superiority to control, % >99.9 99.8 -

Respiratory support-free days

Median (IQR) 9.5 (-1, 16) 11.5 (0, 16) 0 (-1, 14)

Adjusted OR - mean (SD) 1.74 (0.25) 2.04 (0.53) 1

- median (95% CrI) 1.73 (1.31 to 2.27) 1.94 (1.27 to 3.32) 1

Probability of superiority to control, % >99.9 99.9 -

Days free from respiratory support in survivors, median (IQR) 14 (6, 17) 15 (4.5, 17) 12 (2,17)

Cardiovascular support-free days

Median (IQR) 19 (-1, 21) 20.5 (1.5, 21) 15 (-1,21)

Adjusted OR - mean (SD) 1.70 (0.26) 1.95 (0.53) 1

- median (95% CrI) 1.68 (1.25 to 2.24) 1.85 (1.20 to 3.30) 1

Probability of superiority to control, % >99.9 99.5 -

Days free from cardiovascular support in survivors, median (IQR) 21 (17, 21) 21 (17.5, 21) 21 (16, 21)

Time to ICU discharge

Adjusted HR - mean (SD) 1.43 (0.13) 1.69 (0.32) 1

- median (95% CrI) 1.42 (1.18 to 1.70) 1.64 (1.21 to 2.45) 1

Probability of superiority to control, % >99.9 99.9 -

Time to hospital discharge

Adjusted HR - mean (SD) 1.42 (0.13) 1.65 (0.31) 1

- median (95% CrI) 1.41 (1.18 to 1.70) 1.60 (1.17 to 2.40) 1

Probability of superiority to control, % >99.9 99.8 -

WHO scale at day 14

Adjusted OR - mean (SD) 1.85 (0.26) 1.91 (0.43) 1

- median (95% CrI) 1.83 (1.40 to 2.41) 1.86 (1.22 to 2.91) 1

Probability of superiority to control, % >99.9 99.6 -

Progression to invasive mechanical ventilation, ECMO or death, restricted to those not intubated at baseline

Free of invasive mechanical ventilation at baseline, n 242 37 273

Progression to intubation, ECMO or death, n (%) 100/242 (41.3) 13/37 (35.1) 144/273 (52.7)

Adjusted OR - mean (SD) 1.72 (0.33) 1.82 (0.55) 1

- median (95% CrI) 1.69 (1.17 to 2.42) 1.74 (1.01 to 3.14) 1

Probability of superiority to control, % 99.8 97.7 -

Individual components of progression *

Intubation, n (%) 84/242 (34.7) 6/37 (16.2) 116/273 (42.5)

ECMO, n (%) 5/242 (5.0) 0/37 (0.0) 3/273 (1.1)

Death, n (%) 53/242 (21.9) 8/37 (21.6) 82/273 (30.0)

Serious Adverse Events

Patients with >1 serious adverse event, n (%) 9/353 (2.5) 0/48 (0.0) 11/402 (2.7)

Adjusted OR - mean (SD) 1.22 (0.55) 2.99 (2.95) 1

- median (95% CrI) 1.10 (0.48 to 2.58) 2.10 (0.51 to 10.77) 1

Probability of superiority to control, % 59.3 84.0 -

35

* Patients could be counted in more than one component

These secondary analyses were restricted to participants enrolled in the Immune Modulation domain and any domains that have ceased recruitment (Corticosteroid and Covid-19 Antiviral domains) (n=1293), adjusting for age, sex, time period, site, region, domain and intervention eligibility and intervention assignment.

Definitions of outcomes are provided in Methods and the study protocol.

All models are structured such that a higher OR or HR is favorable. The WHO scale ranges from 0 (no disease) to 8 (death).

SD - standard deviation; CrI - credible interval; OR - odds ratio; HR – Hazard ratio; ECMO – Extra-Corporeal Membrane Oxygenation.

36

Table S8. Subgroup analysis of the Primary Outcomes

Outcome/Analysis Tocilizumab (N=353) Sarilumab (N=48) Control (N=402)

Secondary Analysis of Primary Outcome, model restricted to Immune Modulation Therapy Domain participants, according to CRP tercile subgroups

CRP lowest tercile Adjusted OR - mean (SD) 1.51 (0.42) 1.86 (1.06) 1

- median (95% CrI) 1.45 (0.85 to 2.48) 1.62 (0.76 to 4.57) 1

Probability of superiority to control, % 91.3 90.3 -

CRP middle tercile Adjusted OR - mean (SD) 1.54 (0.41) 1.51 (0.54) 1

- median (95% CrI) 1.49 (0.89 to 2.49) 1.44 (0.73 to 2.75) 1

Probability of superiority to control, % 93.5 86.6 -

CRP highest tercile Adjusted OR - mean (SD) 2.00 (0.58) 2.60 (1.44) 1

- median (95% CrI) 1.92 (1.12 to 3.34) 2.25 (1.08 to 6.46) 1

Probability of superiority to control, % 99.1 98.6 -

Secondary Analysis of Hospital Survival, model restricted to Immune Modulation Therapy Domain participants, according to CRP tercile subgroups

CRP lowest tercile Adjusted OR - mean (SD) 1.60 (0.61) 2.05 (1.86) 1

- median (95% CrI) 1.49 (0.74 to 3.08) 1.66 (0.71 to 5.79) 1

Probability of superiority to control, % 86.8 87.5 -

CRP middle tercile Adjusted OR - mean (SD) 1.72 (0.68) 1.82 (1.06) 1

- median (95% CrI) 1.60 (0.77 to 3.40) 1.61 (0.66 to 4.26) 1

Probability of superiority to control, % 89.3 85.8 -

CRP highest tercile Adjusted OR - mean (SD) 1.62 (0.61) 2.43 (3.28) 1

- median (95% CrI) 1.51 (0.76 to 3.12) 1.80 (0.73 to 7.88) 1

Probability of superiority to control, % 87.4 90.0 -

All models are structured such that a higher OR or HR is favorable. The WHO scale ranges from 0 (no disease) to 8 (death).

SD - standard deviation; CrI - credible interval; OR - odds ratio; HR – Hazard ratio; ECMO – Extra-Corporeal Membrane Oxygenation.

37

Table S9. Exploratory post-hoc subgroup analyses of the Primary Outcomes

Outcome/Analysis Pooled Tocilizumab and Sarilumab (N=401) Control (N=402)

Exploratory post-hoc Analysis of Primary Outcome, model restricted to Immune Modulation Therapy Domain participants, according to ventilation subgroups

No mechanical ventilation at baseline Adjusted OR - mean (SD) 1.85 (0.27) 1

- median (95% CrI) 1.83 (1.37, 2.43) 1

Probability of superiority to control, % >99.9 -

Mechanical ventilation at baseline Adjusted OR - mean (SD) 1.30 (0.29) 1

- median (95% CrI) 1.27 (0.84, 1.94) 1

Probability of superiority to control, % 87.1 -

Exploratory post-hoc Analysis of Hospital Survival, model restricted to Immune Modulation Therapy Domain participants, according to ventilation subgroups

No mechanical ventilation at baseline Adjusted OR - mean (SD) 1.91 (0.40) 1

- median (95% CrI) 1.87 (1.26, 2.81) 1

Probability of superiority to control, % 99.9 -

Mechanical ventilation at baseline Adjusted OR - mean (SD) 1.57 (0.45) 1

- median (95% CrI) 1.51 (0.88, 2.63) 1

Probability of superiority to control, % 93.4 -

All models are structured such that a higher OR or HR is favorable. The WHO scale ranges from 0 (no disease) to 8 (death).

SD - standard deviation; CrI - credible interval; OR - odds ratio; HR – Hazard ratio; ECMO – Extra-Corporeal Membrane Oxygenation.

38

4. References:

1. Angus DC, Berry S, Lewis RJ, et al. The REMAP-CAP (Randomized Embedded

Multifactorial Adaptive Platform for Community-acquired Pneumonia) Study. Rationale and

Design. Ann Am Thorac Soc 2020;17:879-91.

2. Angus DC. Fusing Randomized Trials With Big Data: The Key to Self-learning Health

Care Systems? JAMA 2015;314:767-8.

3. WHO Working Group on the Clinical Characterisation and Management of COVID-19

infection. A minimal common outcome measure set for COVID-19 clinical research. Lancet

Infect Dis 2020;20:e192-e7.

39

5. Appendices

a. SAC Primary Analysis Report

b. ITSC Secondary Analysis Report

REMAP-CAP (REMAP-COVID) January 5, 2021

REMAP-CAP (REMAP-COVID)

Analysis of Tocilizumab and Sarilumab

January 5, 2021

Contents

1 Introduction 11.1 Overview of the Adaptive Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Purpose of this Report . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.3 Endpoints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.3.1 Primary Endpoint: Organ-Support Free-Days (OSFD) . . . . . . . . . . . . . . . . . . 21.3.2 Secondary Endpoint: In-Hospital Mortality . . . . . . . . . . . . . . . . . . . . . . . . 2

1.4 Vocabulary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.5 Current Trial Status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31.6 Analysis Population . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 Data Summaries 62.1 Overview of Descriptive Data Summaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62.2 Overall Severe State Summaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72.3 COVID-19 Immune Modulation Domain . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

2.3.1 Description of the COVID-19 Immune Modulation domain . . . . . . . . . . . . . . . 72.3.2 Observed data within the COVID-19 Immune Modulation domain . . . . . . . . . . . 82.3.3 Data summaries for pooled interventions in the COVID-19 Immune Modulation domain 10

3 Analysis Results and Conclusions 113.1 Definition of Statistical Triggers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113.2 Overview of the model results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 113.3 Analyses with separate estimates for tocilizumab and sarilumab . . . . . . . . . . . . . . . . . 12

3.3.1 Primary analysis for OSFD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123.3.2 Primary analysis for in-hospital mortality . . . . . . . . . . . . . . . . . . . . . . . . . 133.3.3 Sensitivity analysis of OSFD with less informative priors on interaction e↵ects . . . . 143.3.4 Sensitivity analysis of the proportional odds assumption . . . . . . . . . . . . . . . . . 16

3.4 Analyses that pool tocilizumab and sarilumab . . . . . . . . . . . . . . . . . . . . . . . . . . . 163.4.1 Primary analysis for OSFD (pooled tocilizumab/sarilumab) . . . . . . . . . . . . . . . 163.4.2 Primary analysis for in-hospital mortality (pooled tocilizumab/sarilumab) . . . . . . . 173.4.3 Sensitivity analysis of OSFD with less informative priors on interaction e↵ects (pooled

tocilizumab/sarilumab) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

4 Other Data Summaries 19

5 Analysis Conventions 23

6 Model Stability 23

7 Report Production 23

Analysis of Tocilizumab and Sarilumab Page 1

REMAP-CAP (REMAP-COVID) January 5, 2021

1 Introduction

1.1 Overview of the Adaptive Design

This trial is a Randomized, Embedded, Multifactorial Adaptive Platform (REMAP) trial that was originallydesigned to investigate treatments for Community-Acquired Pneumonia (CAP). The platform trial has theability to investigate multiple interventions within multiple domains, across di↵erent patient strata. Thenumber of interventions, domains, and strata may increase or decrease as the trial progresses. The platformtrial includes a pandemic stratum that was activated when COVID-19 emerged. The pandemic stratum-specific protocol details are provided in a Pandemic Appendix to the Core (PAtC) protocol. The PAtCinvestigates therapies for patients with pandemic infection that are classified as suspected or proven (PISOP).This report focuses on the COVID-19 PISOP stratum.

For the PISOP stratum, patients may be randomized to interventions while they are in a Severe disease stateor a Moderate disease state. State definitions are in the PAtC. Patients initially randomized in a Moderatestate may progress in their disease severity, and subsequently meet the criteria for Severe state, and haveadditional randomization and reveal of interventions for Severe state domains.

1.2 Purpose of this Report

The international trial steering committee (ITSC) halted randomization to the standard of care (control)intervention in the COVID-19 Immune Modulation domain in the Severe PISOP stratum following thedisclosure from the Data and Safety Monitoring Board (DSMB) that the tocilizumab intervention, an IL-6 receptor antagonist, had met the pre-specified threshold for e�cacy. The REMAP-CAP ITSC publiclydisclosed the e�cacy of Tocilizumab on November 19, 2020 and started a process for reporting results. TheITSC prepared a statistical analysis plan (SAP) for the tocilizumab analysis and provided this plan to theStatistical Analysis Committee (SAC). When the full analysis was performed, the sarilumab intervention,also an IL-6 receptor antagonist, met the pre-specified threshold for e�cacy. Thus, as set forth in the SAP,both sarilumab and tocilizumab are included in the analysis and reporting.

Although the ITSC will be unblinded to the tocilizumab, sarilumab, and control interventions within theCOVID-19 Immune Modulation domain, they will not be unblinded to the other interventions within thedomain, nor to the other domains to which patients have been randomized (except the corticosteroid andCOVID-19 antiviral domains, which were previously closed). The fully unblinded SAC performed the set ofanalyses that use the full statistical model including data from all domains in the Severe PISOP stratum.This report summarizes the data and the results for tocilizumab and sarilumab resulting from the analysesusing the full statistical model. This report is restricted to summaries and results pertaining to the unblindedinterventions. Summaries and results for other ongoing interventions and other domains are contained in aseparate unblinded report only viewed by the SAC and DSMB.

1.3 Endpoints

1.3.1 Primary Endpoint: Organ-Support Free-Days (OSFD)

The primary endpoint is organ support-free days (OSFD) (days alive and free of ICU-based respiratory orcardiovascular support) within 21 days, where patients who die before discharge from the index hospitaliza-tion, and before day 90, were assigned a �1 day, even if the death occurred after day 21. The cumulativehours of organ support are computed and then rounded to the nearest day. Patients who receive no organsupport in an ICU will be coded as 22 days. An outcome of 22 days is not possible for patients in Severestate. An outcome of 21 organ-support free-days is only possible in the Severe state if the patient had lessthan 12 hours of organ support.

Analysis of Tocilizumab and Sarilumab Page 2

REMAP-CAP (REMAP-COVID) January 5, 2021

1.3.2 Secondary Endpoint: In-Hospital Mortality

The secondary endpoint is a dichotomous endpoint of in-hospital mortality, i.e. those patients with a �1 forthe OSFD endpoint.

1.4 Vocabulary

• Domain: a specific set of competing alternative interventions within a common clinical mode

• Intervention: is a treatment option that is subject to variation in clinical practice (comparativee↵ectiveness intervention) or has been proposed for introduction into clinical practice (experimentalintervention) and also is being subjected to experimental manipulation within the design of a REMAP.

• Regimen: Each patient is assigned a single intervention from each domain. The regimen is thecombination of assigned interventions across the domains.

• Immediate Reveal Domain: is one for which all participants are eligible, the allocation status ismade known, and the intervention is initiated at the time of randomization.

• Delayed Reveal Domain: is one for which all participants received a randomization assignment, butthe allocation status is only made known and the intervention initiated if and when eligibility occurs.This occurs for example, when a domain is appropriate only for patients in a certain disease state andthe patient transitions to that disease state.

• Deferred Reveal Domain: is one for which patients receive a randomization assignment and theallocation status is made known based on eligibility criterion known at the time of randomization,but additional information to assess that eligibility becomes known after some time. This occurs forexample, when a test results confirming an eligibiltiy criterion are returned after some time.

• Nest: A grouping of interventions within a domain that are modeled hierarchically in order to allowfor borrowing among the interventions e↵ect estimates.

• State: Defined by the disease characteristics of the patient and may change over time as the diseaseprogresses. States are used to define eligibility for certain domains.

1.5 Current Trial Status

The data transfers provided to the SAC (see Section 7 for details) include all patients randomized to the trial(including both the pandemic and non-pandemic strata) through December 17, 2020 from the combined Spiraland Research Online databases. The UPMC transfer was received December 13, 2020, but only includespatients randomized through November 19, 2020. These data include:

• 3744 total randomized, consented patients in the REMAP-CAP/REMAP-COVID trial, with:

– 3036 randomized, consented patients with pandemic infection suspected or proven (PISOP),

‘ 2696 PISOP patients initially randomized while in Severe disease state.

‘ 340 PISOP patients initially randomized while in Moderate disease state,

· 14 PISOP patients initially randomized while in Moderate disease state that later metthe criteria for Severe disease state and had additional randomization for Severe statedomains,

– 708 randomized, consented patients with pandemic infection neither suspected nor proven (PIN-SNP).

Analysis of Tocilizumab and Sarilumab Page 3

REMAP-CAP (REMAP-COVID) January 5, 2021

These counts exclude patients that withdrew consent for the use of their data. The patients who withdrewconsent do not appear in the SAC data export, so no information is available to the SAC regarding when inthe process (e.g. before or after eligibility assessment) consent was withdrawn.

Figure 1.1 gives an overview of the interventions, domains, and strata currently being investigated in theCOVID-19 pandemic portion of the trial. Each intervention is represented by a colored box, with similarcolors used for interventions within the same domain. The figure also indicates features of the statisticalmodel. For example, interactions are represented with an arrow and star (H). Within a domain, interventionsthat are nested within a hierarchical model are grouped within a curly bracket. Interventions that are closedto enrollment are indicated by an “X”. Table 1.1 is a companion to the current state figure, and providesthe mapping of intervention codes to the actual intervention names (e.g. X2 = Lopinavir/ritonavir).

M1

M2

All Severe Pandemic Infection Suspected or Proven (PISOP) Patients

X3

X4

X1

X2

C4

C1

C2

C3

Y4Y5

Y1Y2

Y3

L1

L2

H1

H2

Corticosteroid

COVID-19 Antiviral

Immune Modulation – 1Macrolide Duration

Vitamin C

Anticoagulation

Nest

Nest

Nest

Antibiotic

Nest

Influenza Antiviral

Suspected or Confirmed Influenza

A4A5

A1A2

A3

I3

I2

I1

Domains Not Included in the Pandemic Model *

* RAR probabilities for these domains are determined from the inter-pandemic model

Pool

S1

S2

Statin†

P1

P2

P3

Immunoglobulin

B1

B3

B2

Antiplatelet†

= Closed

= Effective

or Superior

Figure 1.1: Current state of the Severe State pandemic domains and interventions. Each colored boxrepresents an intervention, grouped by domain, with similar colors used for interventions within the samedomain. Domains connected with an arrow and indicated with a star (H) will have interaction terms fit betweenthe interventions in those domains. Within a domain, interventions that are grouped with a curly bracket arepart of a nest whose main e↵ects are estimated with a hierarchical model. Interventions that are closed toenrollment are indicated by a grey “X”. Interventions that have met an E�cacy or Superiority trigger areindicated by a green checkmark. As indicated by the dagger (†), the Statin and Antiplatelet domains are openfor randomization but do not contribute data the model for the current analysis.

Analysis of Tocilizumab and Sarilumab Page 4

REMAP-CAP (REMAP-COVID) January 5, 2021

Table 1.1: List of all interventions to which a patient may be allocated in Severe State.

Code Intervention Status

Antibiotic

A1 Ceftriaxone + MacrolideA2 Moxifloxacin or LevofloxacinA3 Piperacillin-Tazobactam + MacrolideA4 Ceftaroline + MacrolideA5 Amoxicillin-Clavulanate + Macrolide

Macrolide Duration

M1 Standard course (3 to 5 days)M2 Extended course (14 days)

Corticosteroid

C1 No corticosteroids ClosedC2 Hydrocortisone (50mg) ClosedC3 Shock dependent hydrocortisone ClosedC4 High-dose hydrocortisone (100mg) Closed

Influenza Antiviral

I1 No antiviralI2 Oseltamivir 5 daysI3 Oseltamivir 10 days

COVID-19 Antiviral

X1 No antiviral for COVID-19 ClosedX2 Lopinavir/ritonavir ClosedX3 Hydroxychloroquine ClosedX4 Hydroxychloroquine + lopinavir/ritonavir Closed

COVID-19 Immune Modulation

Y1 No immune modulation for COVID-19 ClosedY2 Interferon-Beta-1aY3 AnakinraY4 Tocilizumab E↵ectiveY5 Sarilumab E↵ective

COVID-19 Immunoglobulin

P1 No Immunoglobulin against COVID-19P2 Convalescent plasmaP3 Delayed convalescent plasma Closed

COVID-19 Therapeutic Anticoagulation

H1 Standard practice thromboprophylaxis ClosedH2 Therapeutic anticoagulation Closed

Vitamin C

L1 No vitamin CL2 Vitamin C

Statin Therapy†

S1 No simvastatinS2 Simvastatin

COVID-19 Antiplatelet†

B1 No antiplatelet therapyB2 AspirinB3 P2Y12 inhibitor (clopidogrel, prasugrel, or ticagrelor)

† Domain open for randomization but data not yet available for inclusion in the model

1.6 Analysis Population

This report restricts the analysis population to consented patients with pandemic infection suspected orproven (PISOP) that were randomized for the first time in Severe disease state (excluding patients firstrandomized in Moderate state that progressed to Severe state) on or before 12:00 UTC on November 19,2020. This population is defined as the REMAP-CAP COVID-19 severe state ITT population. Thepatient population breakdown is as follows:

Analysis of Tocilizumab and Sarilumab Page 5

REMAP-CAP (REMAP-COVID) January 5, 2021

• 2328 PISOP consented patients randomized to at least one domain in any disease state (Moderate orSevere) on or before 12:00 UTC on November 19, 2020.

– 1991 PISOP consented patients randomized initially in Severe disease state (never randomized inModerate state) to at least one domain on or before 12:00 UTC on November 19, 2020

‘ 1928 PISOP consented patients randomized initially in Severe disease state (never randomizedin Moderate state) to at least one domain on or before 12:00 UTC on November 19, 2020 forwhom 21 days have elapsed since randomization and there is a known outcome on the 21-dayorgan-support free-days endpoint

– 865 PISOP consented patients randomized initially in Severe disease state (never randomizedin Moderate state) to the COVID-19 Immune Modulation domain on or before 12:00 UTC onNovember 19, 2020

‘ 853 PISOP consented patients randomized initially in Severe disease state (never randomizedin Moderate state) to the COVID-19 Immune Modulation domain on or before 12:00 UTC onNovember 19, 2020 for whom 21 days have elapsed since randomization and there is a knownoutcome on the 21-day organ-support free-days endpoint

Thus the analysis model is run on the 1928 PISOP Severe patients for whom 21 days have elapsed sincerandomization and there is a known outcome on the 21-day OSFD endpoint. This count excludes 14 patientswho were initially randomized while in Moderate State and later progressed to Severe State with additionalrandomized assignments for Severe state domains.

2 Data Summaries

2.1 Overview of Descriptive Data Summaries

Data for the Severe PISOP patient population will be summarized, both across all Severe state PISOPpatients (without respect to intervention assignments), and at the intervention level for the unblinded inter-ventions in the COVID-19 Immune Modulation domain. A description of each of the summary tables andfigures is provided here.

Summary of the availability of data:

• Number Eligible: Eligibility is assessed both at the domain level and the intervention level. Wetabulate the number of patients eligible for the domain, and within each category of domain eligibility,the number of patients eligible for each intervention. Eligibility captures both the patient meeting theinclusion criteria, and the domain or intervention being available and active at their site.

• Number Assigned: We tabulate the number of patients assigned to each intervention, by eligibilitycategory. No randomized assignment can be given when a patient is ineligible for a domain, or whena patient is eligible for only one intervention within a domain. A patient must be eligible for at leasttwo interventions within a domain to receive a randomized assignment.

• Number Revealed: Among the patients eligible and assigned to each intervention, we tabulate thenumber of patients whose assignment was revealed. Reveal means that the randomization assignmentwas made known and the patient then commences treatment according to their assigned intervention.

• Number Past 21 Days: Among the patients eligible and assigned to each intervention, we tabulatethe number of patients who have had the opportunity to complete the 21 days of follow-up for theprimary endpoint. A patient must have been in the trial at least 21 days to be included in the analysis.

• Number Missing: Among the patients eligible and assigned to each intervention, we tabulate thenumber of patients who have completed 21 days of follow-up but do not have an outcome available onthe primary endpoint.

Analysis of Tocilizumab and Sarilumab Page 6

REMAP-CAP (REMAP-COVID) January 5, 2021

• Number Known: Among the patients eligible and assigned to each intervention, we tabulate thenumber of patients who have completed 21 days of follow-up and have a known outcome on theprimary endpoint.

• For the subjects that are eligible for the domain, a bar chart summarizes the number and percent ofpatients assigned to each intervention.

Summary of the observed outcomes data:

For patients that are eligible for the domain and assigned to an intervention, we repeat the tabulation ofthe number of patients assigned to an intervention and with a known outcome on the 21-day endpoint.Additionally, we provide summaries of the following:

• Number Deaths: The number of in-hospital deaths, where the death corresponds to �1 on the OSFDendpoint.

• Mortality Rate: We calculate the observed in-hospital mortality rate as the number of in-hospitaldeaths out of the total number of patients with a known 21-day outcome.

• OSFD median (IQR): Among the patients with a known 21-day outcome, we compute the 25th,50th, and 75th percentiles of the Organ-Support Free-Days endpoint. The interquartile range (IQR)is shown in parentheses as the range between the 25th and 75th percentiles.

• Conditional OSFD: Among the patients with a known 21-day outcome that were not deceased,we compute the 25th, 50th, and 75th percentiles of the Organ-Support Free-Days endpoint. Theinterquartile range (IQR) is shown in parentheses as the range between the 25th and 75th percentiles.

• For the subjects that are eligible for the domain, we show a plot of the cumulative distribution functionfor the OSFD endpoint for each intervention within the domain.

• For the subjects that are eligible for the domain, we show a stacked bar plot for the OSFD outcomesfor each intervention within the domain. Red represents worse outcomes and blue represents betteroutcomes.

2.2 Overall Severe State Summaries

Figure 2.1 displays the distribution of outcomes on the primary endpoint for all patients in the analysispopulation (across all domains), without respect to treatment assignments. Table 2.1 provides descriptivesummaries of the OSFD and in-hospital mortality outcomes for all patients in the analysis population.

Table 2.1: Overall summary of the OSFD and In-Hospital mortality data

Participant GroupNumberAssigned

(N)

NumberPast

Day 21

NumberKnown(n)

NumberDeaths(y)

MortalityRate(y/n)

OSFDmedian (IQR)

Conditional*OSFD

median (IQR)

COVID Severe State 2696 1991 1928 642 0.333 3.00 (�1.00� 15.00) 13.00 (3.00� 17.00)* Conditional OSFD reports the median and IQR for subjects that did not die.

2.3 COVID-19 Immune Modulation Domain

2.3.1 Description of the COVID-19 Immune Modulation domain

The COVID-19 Immune Modulation domain includes a total of 5 interventions. This domain:

• has closed randomization to the control arm, Y1 (no immune modulation). Randomization continuesto the remaining active interventions in the domain. Because 2 interventions (Anakinra and Interferon-Beta-1a) are still actively being investigated, they are collapsed into an “other immune modulation”category in this report to preserve the blind;

Analysis of Tocilizumab and Sarilumab Page 7

REMAP-CAP (REMAP-COVID) January 5, 2021

Death

0

200

400

600

−1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21Days free from organ support

Num

ber o

f Pat

ient

s(n = 1928 with outcome)OSFD Distribution for Severe State

Figure 2.1: Overall distribution of the primary organ support free days endpoint.

• is an immediate reveal domain with immediate initiation of the randomized assignment, unless prospec-tive agreement to participate is required, in which case it is deferred reveal domain;

• has no strata identified as being of interest. Analyses and response adaptive randomization are appliedto all randomized patients in Severe State;

• has possible interactions modeled with the corticosteroid domain; A previous study suggested thatthe interaction of interferon-� and corticosteroids may be harmful; therefore an informative prior isused to reflect a harmful interaction. Furthermore, initial (burn-in) randomization probabilities wereconstructed to limit the number of patients randomized to the combination of corticosteroids andinterfereon-�;

• has one nest, comprised of tocilizumab and sarilumab, which are both interleukin-6 inhibitors;

• enforces a minimum marginal RAR probability of 0.05 for each active intervention within the domain.

2.3.2 Observed data within the COVID-19 Immune Modulation domain

Table 2.2: Summary of the availability of data (COVID-19 Immune Modulation domain)

InterventionNumberEligible

NumberAssigned

NumberRevealed

NumberPast

Day 21

NumberMissing

NumberKnown

Eligible for domain: N=871No assignment 6 3 6 0 6No immune modulation for COVID-19 871 406 406 406 5 401Tocilizumab 765 353 353 353 3 350Sarilumab 147 48 48 48 3 45Other immune modulation 58 58 58 1 57

Not eligible for domain: N=470No assignment 470 0 470 17 453

Domain not active/not available: N=650No assignment 650 0 650 34 616

Analysis of Tocilizumab and Sarilumab Page 8

REMAP-CAP (REMAP-COVID) January 5, 2021

0.0

0.2

0.4

0.6

0.8

1.0

Death 0 2 4 6 8 10 12 14 16 18 20 21Days free from organ support

Cum

ulat

ive P

ropo

rtion

of P

atie

nts

No immunemodulation forCOVID−19(n=401)

Tocilizumab(n=350)

Sarilumab(n=45)

COVID−19 Immune Modulation Domain; Severe StateCumulative Distribution of Organ−Support Free−Days

Figure 2.2: Empirical cumulative distribution of organ support free days for the IL-6ra and control interventionsin the COVID-19 Immune Modulation domain. This plot is restricted to patients who were eligible for thedomain.

Sarilumab(n=45)

Tocilizumab(n=350)

No immunemodulation for

COVID−19(n=401)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0Proportion

Death 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 2.3: Stacked proportion of organ support free days for the IL-6ra and control interventions in theCOVID-19 Immune Modulation domain. Red represents worse outcomes and blue represents better out-comes. This plot is restricted to include only patients who were eligible for the domain.

Analysis of Tocilizumab and Sarilumab Page 9

REMAP-CAP (REMAP-COVID) January 5, 2021

Table 2.3: Summary of the OSFD and In-Hospital mortality data for patients that were eligible for theCOVID-19 Immune Modulation domain

InterventionNumberAssigned

(N)

NumberKnown(n)

Numberof Deaths

(y)

MortalityRate(y/n)

OSFDmedian (IQR)

Conditional*OSFD

median (IQR)

No immune modulation for COVID-19 406 401 142 0.354 0.00 (�1.00� 15.00) 13.00 (3.50� 17.00)Tocilizumab 353 350 98 0.280 10.00 (�1.00� 16.00) 14.00 (7.00� 17.00)Sarilumab 48 45 10 0.222 11.00 (0.00� 16.00) 15.00 (6.50� 17.00)* Conditional OSFD reports the median and IQR for subjects that did not die.

2.3.3 Data summaries for pooled interventions in the COVID-19 Immune Modulation domain

For the primary analysis model, treatment e↵ects are estimated separately for the two IL-6 antagoniststocilizumab and sarilumab. A sensitivity analysis is performed that pools the tocilizumab and sarilumabinterventions. This section summarizes the observed data on the OSFD and in-hospital mortality endpointsfor the combined arms for the IL-6 antagonists as they will be coded in the sensitivity analysis model.

Table 2.4: Summary of the OSFD and In-Hospital mortality data for patients that were eligible for theCOVID-19 Immune Modulation domain

InterventionNumberAssigned

(N)

NumberKnown(n)

Numberof Deaths

(y)

MortalityRate(y/n)

OSFDmedian (IQR)

Conditional*OSFD

median (IQR)

No immune modulation for COVID-19 406 401 142 0.354 0.00 (�1.00� 15.00) 13.00 (3.50� 17.00)Tocilizumab/Sarilumab 401 395 108 0.273 10.00 (�1.00� 16.00) 14.00 (7.00� 17.00)* Conditional OSFD reports the median and IQR for subjects that did not die.

0.0

0.2

0.4

0.6

0.8

1.0

Death 0 2 4 6 8 10 12 14 16 18 20 21Days free from organ support

Cum

ulat

ive P

ropo

rtion

of P

atie

nts

No immunemodulation forCOVID−19(n=401)

Tocilizumab/Sarilumab(n=395)

COVID−19 Immune Modulation Domain; Severe StateCumulative Distribution of Organ−Support Free−Days

Figure 2.4: Empirical cumulative distribution of organ support free days for the IL-6ra and control interventionsin the COVID-19 Immune Modulation domain. This plot is restricted to patients who were eligible for thedomain. IL-6ra intervention arms tocilizumab and sarilumab are pooled.

Analysis of Tocilizumab and Sarilumab Page 10

REMAP-CAP (REMAP-COVID) January 5, 2021

Tocilizumab/Sarilumab(n=395)

No immunemodulation for

COVID−19(n=401)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0Proportion

Death 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 2.5: Stacked proportion of organ support free days for the IL-6ra and control interventions in theCOVID-19 Immune Modulation domain. Red represents worse outcomes and blue represents better out-comes. This plot is restricted to include only patients who were eligible for the domain. IL-6ra interventionarms tocilizumab and sarilumab are pooled.

3 Analysis Results and Conclusions

3.1 Definition of Statistical Triggers

The adpative design defines several statistical triggers within the trial, that at any analysis of the trial wouldresult in public disclosure and declaration of a platform conclusion. The following statistical triggers weredefined for the COVID-19 Immune Modulation domain:

1. Domain Superiority. If a single intervention within a domain has at least a 99% posterior probabilityof being in the best regimen for patients in the severe state of the PISOP stratum, this would triggersuperiority of that intervention.

2. Intervention E�cacy. If an intervention is deemed to have at least a 99% posterior probability ofbeing superior to the control, then a declaration of e�cacy of that intervention would be declared.This statistical trigger is active for each of the non-control arms in the domain.

3. Intervention Equivalence. If two non-control interventions have a 90% probability of equivalence— that is, that the odds ratio comparing the two interventions is between 0.83 (inverse of 1/1.2 and1.2) /textemdash , then a declaration of intervention equivalence would be made.

4. Intervention Futility. If an intervention is deemed to have less than 5% posterior probability ofat least a 20% odds ratio improvement compared to the control, then a declaration of futility of thatintervention would be declared. This statistical trigger is active for each of the non-control arms in thedomain.

5. Intervention Inferiority. If an intervention has low posterior probability of being the optimalintervention within a state, then that intervention will be deemed inferior. Specifically, an interventionis considered inferior if the probability of being the optimal intervention is less than 0.01/(Jd � 1),where Jd is the number of interventions within the domain. For the COVID-19 Immune Modulationdomain, there were initially J = 5 interventions. Since the closure of the control arm, there are nowJ = 4 interventions in the domain.

3.2 Overview of the model results

The OSFD endpoint is an ordered categorical endpoint that is modeled with a cumulative logistic model.The median and 95% Bayesian credible intervals for the odds-ratios are presented for each intervention,

Analysis of Tocilizumab and Sarilumab Page 11

REMAP-CAP (REMAP-COVID) January 5, 2021

relative to the control intervention in the domain. The model is structured so that an odds-ratio greaterthan 1 implies patient benefit. We also present the posterior mean and standard deviation, but caution thatthe mean tends to be higher than the median due to the skewed nature of the posterior distribution.

3.3 Analyses with separate estimates for tocilizumab and sarilumab

3.3.1 Primary analysis for OSFD

Table 3.1 summarizes the model-estimated odds-ratios for the covariates in the model, including age cate-gories, sex at birth, and time e↵ects.

Table 3.1: Model-estimated Odds-Ratios for the OSFD endpoint; REMAP-CAP COVID-19 severe state ITTpopulation; model with Tocilizumab and Sarilumab estimated separately

Odds-RatioParameter

Median95% Credible

IntervalMean (SD)

39 4.06 2.87� 5.80 4.13 (0.74)40-49 2.40 1.82� 3.17 2.42 (0.34)50-59 1.86 1.49� 2.32 1.88 (0.21)70-79 0.53 0.41� 0.67 0.53 (0.07)80+ 0.36 0.24� 0.54 0.37 (0.08)Female 1.12 0.94� 1.34 1.12 (0.10)Time-1 0.99 0.85� 1.14 0.99 (0.07)Time-2 1.03 0.81� 1.27 1.03 (0.12)Time-3 1.15 0.88� 1.53 1.17 (0.17)Time-4 1.27 0.93� 1.76 1.29 (0.21)Time-5 1.35 0.98� 1.90 1.37 (0.24)Time-6 1.40 1.00� 1.97 1.42 (0.25)Time-7 1.46 1.06� 2.05 1.49 (0.26)Time-8 1.52 1.08� 2.17 1.55 (0.28)Time-9 1.49 1.05� 2.16 1.52 (0.28)Time-10 1.34 0.92� 1.95 1.37 (0.26)Time-11 1.21 0.78� 1.81 1.23 (0.26)Time-12 1.10 0.69� 1.71 1.13 (0.26)Time-13 1.07 0.66� 1.72 1.10 (0.28)Time-14 1.09 0.64� 1.89 1.13 (0.32)Time-15 1.11 0.58� 2.18 1.18 (0.42)Tocilizumab 1.64 1.25� 2.14 1.65 (0.23)Sarilumab 1.76 1.17� 2.91 1.83 (0.44)Tocilizumab * Fixed-Dose Corticosteroid 1.00 0.91� 1.10 1.00 (0.05)Sarilumab * Fixed-Dose Corticosteroid 1.00 0.91� 1.11 1.00 (0.05)Tocilizumab * Shock-based Corticosteroid 1.00 0.90� 1.10 1.00 (0.05)Sarilumab * Shock-based Corticosteroid 1.00 0.91� 1.10 1.00 (0.05)Tocilizumab * HCQ 1.00 0.91� 1.10 1.00 (0.05)Sarilumab * HCQ 1.00 0.91� 1.11 1.00 (0.05)Tocilizumab * Kaletra 1.01 0.92� 1.11 1.01 (0.05)Sarilumab * Kaletra 1.00 0.91� 1.10 1.00 (0.05)Tocilizumab * (HCQ + Kaletra) 1.00 0.91� 1.11 1.00 (0.05)Sarilumab * (HCQ + Kaletra) 1.00 0.91� 1.10 1.00 (0.05)

Note: For Age, Odds-Ratio is relative to the Age 60-69 category. Time bucketX is the Xth 2-week interval prior to the most recent month, and Odds-Ratios arerelative to the most recent month.

Analysis of Tocilizumab and Sarilumab Page 12

REMAP-CAP (REMAP-COVID) January 5, 2021

Table 3.2: Posterior Probabilities for the OSFD endpoint; REMAP-CAP COVID-19 severe state ITT popu-lation; model with Tocilizumab and Sarilumab estimated separately

Posterior Probability

Quantity of Interest Tocilizumab Sarilumab

Prob(OR > 1) 0.9999 0.9951Prob(OR < 1.2) 0.0119 0.0306Prob(Y4 equiv Y5) 0.6795Prob(OR > 1) interaction with Fixed-dose corticosteroid 0.4880 0.5287Prob(OR > 1) interaction with Shock-based corticosteroid 0.4905 0.5034Prob(OR > 1) interaction with HCQ 0.5071 0.5136Prob(OR > 1) interaction with Kaletra 0.5691 0.4978Prob(OR > 1) interaction with HCQ + Kaletra 0.5165 0.5030

3.3.2 Primary analysis for in-hospital mortality

Table 3.3: Model-estimated Odds-Ratios for the Mortality endpoint; REMAP-CAP COVID-19 severe stateITT population; model with Tocilizumab and Sarilumab estimated separately

Odds-RatioParameter

Median95% Credible

IntervalMean (SD)

39 11.19 5.82� 23.31 12.09 (4.57)40-49 4.20 2.80� 6.52 4.32 (0.96)50-59 2.98 2.23� 4.13 3.04 (0.49)70-79 0.45 0.34� 0.59 0.45 (0.06)80+ 0.29 0.18� 0.45 0.30 (0.07)Female 1.14 0.89� 1.45 1.15 (0.14)Time-1 0.99 0.83� 1.15 0.98 (0.08)Time-2 0.99 0.72� 1.26 0.99 (0.13)Time-3 1.05 0.75� 1.46 1.06 (0.18)Time-4 1.14 0.78� 1.68 1.16 (0.23)Time-5 1.23 0.82� 1.84 1.25 (0.26)Time-6 1.32 0.88� 1.99 1.34 (0.28)Time-7 1.40 0.93� 2.13 1.44 (0.31)Time-8 1.51 1.00� 2.33 1.55 (0.34)Time-9 1.62 1.04� 2.75 1.69 (0.44)Time-10 1.61 1.00� 2.85 1.70 (0.51)Time-11 1.52 0.90� 2.68 1.59 (0.44)Time-12 1.43 0.80� 2.54 1.49 (0.44)Time-13 1.40 0.74� 2.61 1.47 (0.48)Time-14 1.43 0.69� 2.94 1.53 (0.59)Time-15 1.50 0.60� 3.67 1.68 (0.83)Tocilizumab 1.64 1.14� 2.35 1.66 (0.31)Sarilumab 2.01 1.18� 4.71 2.25 (0.96)Tocilizumab * Fixed-Dose Corticosteroid 0.99 0.90� 1.09 0.99 (0.05)Sarilumab * Fixed-Dose Corticosteroid 1.01 0.91� 1.11 1.01 (0.05)Tocilizumab * Shock-based Corticosteroid 1.00 0.91� 1.10 1.00 (0.05)Sarilumab * Shock-based Corticosteroid 1.00 0.91� 1.11 1.00 (0.05)Tocilizumab * HCQ 1.00 0.91� 1.11 1.01 (0.05)Sarilumab * HCQ 1.00 0.90� 1.10 1.00 (0.05)Tocilizumab * Kaletra 1.00 0.91� 1.11 1.00 (0.05)Sarilumab * Kaletra 1.00 0.91� 1.10 1.00 (0.05)Tocilizumab * (HCQ + Kaletra) 1.00 0.91� 1.10 1.00 (0.05)Sarilumab * (HCQ + Kaletra) 1.00 0.91� 1.10 1.00 (0.05)

Note: For Age, Odds-Ratio is relative to the Age 60-69 category. Time bucketX is the Xth 2-week interval prior to the most recent month, and Odds-Ratios arerelative to the most recent month.

Analysis of Tocilizumab and Sarilumab Page 13

REMAP-CAP (REMAP-COVID) January 5, 2021

Table 3.4: Posterior Probabilities for the Mortality endpoint; REMAP-CAP COVID-19 severe state ITTpopulation; model with Tocilizumab and Sarilumab estimated separately

Posterior Probability

Quantity of Interest Tocilizumab Sarilumab

Prob(OR > 1) 0.9963 0.9948Prob(OR < 1.2) 0.0489 0.0288Prob(Y4 equiv Y5) 0.4782Prob(OR > 1) interaction with Fixed-dose corticosteroid 0.4038 0.5471Prob(OR > 1) interaction with Shock-based corticosteroid 0.5151 0.5003Prob(OR > 1) interaction with HCQ 0.5388 0.5046Prob(OR > 1) interaction with Kaletra 0.5179 0.5054Prob(OR > 1) interaction with HCQ + Kaletra 0.5191 0.5011

3.3.3 Sensitivity analysis of OSFD with less informative priors on interaction e↵ects

To assess whether results are strongly influenced by the informative priors on the interaction terms betweenthe IL-6ra interventions and corticosteroids and between the IL-6ra interventions and antivirals, the modelwas evaluated with less informative priors on those terms.

Analysis of Tocilizumab and Sarilumab Page 14

REMAP-CAP (REMAP-COVID) January 5, 2021

Table 3.5: Model-estimated Odds-Ratios for the OSFD endpoint; REMAP-CAP COVID-19 severe stateITT population; model with Tocilizumab and Sarilumab estimated separately; sensitivity analysis with lessinformative priors on interaction e↵ects

Odds-RatioParameter

Median95% Credible

IntervalMean (SD)

39 4.05 2.88� 5.73 4.12 (0.73)40-49 2.40 1.83� 3.16 2.43 (0.34)50-59 1.87 1.48� 2.35 1.88 (0.22)70-79 0.52 0.41� 0.67 0.53 (0.07)80+ 0.36 0.23� 0.54 0.36 (0.08)Female 1.12 0.94� 1.35 1.13 (0.10)Time-1 1.00 0.85� 1.15 1.00 (0.07)Time-2 1.04 0.82� 1.28 1.04 (0.12)Time-3 1.16 0.89� 1.54 1.18 (0.17)Time-4 1.28 0.94� 1.80 1.30 (0.22)Time-5 1.37 0.99� 1.97 1.40 (0.26)Time-6 1.42 1.01� 2.02 1.45 (0.26)Time-7 1.49 1.06� 2.12 1.51 (0.28)Time-8 1.54 1.09� 2.20 1.57 (0.30)Time-9 1.52 1.06� 2.19 1.54 (0.29)Time-10 1.37 0.93� 1.97 1.39 (0.27)Time-11 1.22 0.80� 1.82 1.25 (0.26)Time-12 1.12 0.71� 1.72 1.14 (0.26)Time-13 1.08 0.67� 1.72 1.11 (0.27)Time-14 1.10 0.65� 1.87 1.14 (0.31)Time-15 1.11 0.58� 2.17 1.18 (0.41)Tocilizumab 1.52 0.84� 2.71 1.59 (0.49)Sarilumab 1.54 0.56� 4.35 1.78 (1.42)Tocilizumab * Fixed-Dose Corticosteroid 1.02 0.56� 1.87 1.07 (0.35)Sarilumab * Fixed-Dose Corticosteroid 1.30 0.43� 3.75 1.48 (0.90)Tocilizumab * Shock-based Corticosteroid 0.89 0.36� 2.17 0.98 (0.48)Sarilumab * Shock-based Corticosteroid 0.97 0.14� 6.95 1.62 (2.21)Tocilizumab * HCQ 1.04 0.39� 2.81 1.19 (0.64)Sarilumab * HCQ 0.98 0.15� 6.96 1.63 (2.13)Tocilizumab * Kaletra 1.28 0.74� 2.21 1.33 (0.38)Sarilumab * Kaletra 0.99 0.32� 3.03 1.17 (0.73)Tocilizumab * (HCQ + Kaletra) 1.57 0.37� 6.27 2.00 (1.62)Sarilumab * (HCQ + Kaletra) 1.00 0.14� 6.82 1.63 (2.05)

Note: For Age, Odds-Ratio is relative to the Age 60-69 category. Time bucketX is the Xth 2-week interval prior to the most recent month, and Odds-Ratios arerelative to the most recent month.

Table 3.6: Posterior Probabilities for the OSFD endpoint; REMAP-CAP COVID-19 severe state ITT popu-lation; model with Tocilizumab and Sarilumab estimated separately; sensitivity analysis with less informativepriors on interaction e↵ects

Posterior Probability

Quantity of Interest Tocilizumab Sarilumab

Prob(OR > 1) 0.9138 0.8371Prob(OR < 1.2) 0.2220 0.2794Prob(Y4 equiv Y5) 0.5296Prob(OR > 1) interaction with Fixed-dose corticosteroid 0.5208 0.6927Prob(OR > 1) interaction with Shock-based corticosteroid 0.3909 0.4872Prob(OR > 1) interaction with HCQ 0.5238 0.4913Prob(OR > 1) interaction with Kaletra 0.8122 0.4932Prob(OR > 1) interaction with HCQ + Kaletra 0.7250 0.4993

Analysis of Tocilizumab and Sarilumab Page 15

REMAP-CAP (REMAP-COVID) January 5, 2021

3.3.4 Sensitivity analysis of the proportional odds assumption

An assumption of the ordinal logistic regression model being used to analyze OSFD is that e↵ects havea proportional e↵ect on log-odds. That is, a treatment e↵ect that increases the log-odds of OSFD beinggreater than X is the same for all values of X. In order to assess this modeling assumption, a logisticregression model is fit to dichotomized versions of the OSFD values ( X versus > X) across the possiblerange of OSFD values, to see if the estimated treatment e↵ect is nearly constant. The prior distribution forthe dichotomized outcomes is constructed in like manner, converting the Dirichlet distribution across OSFDvalues to a Beta distribution by summing across the parameter values for the corresponding OSFD ranges.

The logistic regression model is subject to poor estimation when categories in the model contain only asingle outcome type (e.g. all observations are X). With the large number of covariate e↵ects being usedin the current model, some categories of these covariate crossings may contain single outcomes, particularlyas the dichotomization goes to the higher end of the OSFD values with low frequencies. The Bayesian modeluses informative priors and thus a model fit can always be constructed. However, because many of the priordistributions in the model are relatively non-informative, the MCMC fitting algorithms can perform poorlyin the more extreme dichotomizations. Some poor MCMC behavior was observed for dichotomizations at� 15 OSFD and higher.

Table 3.7: Sensitivity analysis of proportional odds assumption. Values are Odds-Ratio estimates for theOSFD

endpoint for di↵erent dichotomizations of OSFD; REMAP-CAP COVID-19 severe state ITT population; modelwith Tocilizumab and Sarilumab estimated separately

Tocilizumab Sarilumab

OSFDDichotomization

Median95% Credible

IntervalMean (SD) Median

95% CredibleInterval

Mean (SD)

-1 vs �0 1.64 1.14� 2.35 1.66 (0.31) 2.01 1.18� 4.71 2.25 (0.96)0 vs �1 1.68 1.22� 2.32 1.70 (0.28) 1.79 1.09� 3.20 1.88 (0.54)1 vs �2 1.72 1.26� 2.37 1.74 (0.28) 1.82 1.13� 3.21 1.90 (0.52)2 vs �3 1.71 1.24� 2.39 1.74 (0.29) 1.83 1.13� 3.22 1.92 (0.54)3 vs �4 1.61 1.18� 2.26 1.64 (0.27) 1.74 1.09� 3.18 1.84 (0.54)4 vs �5 1.63 1.19� 2.22 1.65 (0.27) 1.73 1.10� 3.05 1.81 (0.49)5 vs �6 1.64 1.17� 2.25 1.66 (0.28) 1.76 1.12� 3.21 1.86 (0.54)6 vs �7 1.65 1.20� 2.20 1.65 (0.24) 1.63 1.08� 2.83 1.73 (0.44)7 vs �8 1.66 1.22� 2.30 1.69 (0.28) 1.74 1.07� 2.93 1.79 (0.47)8 vs �9 1.68 1.22� 2.33 1.70 (0.28) 1.75 1.10� 3.07 1.84 (0.51)9 vs �10 1.73 1.27� 2.39 1.76 (0.28) 1.81 1.13� 3.06 1.88 (0.49)10 vs �11 1.71 1.25� 2.34 1.73 (0.28) 1.70 1.04� 2.77 1.75 (0.45)11 vs �12 1.64 1.18� 2.26 1.66 (0.27) 1.59 0.97� 2.58 1.64 (0.41)12 vs �13 1.66 1.21� 2.33 1.68 (0.28) 1.63 0.99� 2.70 1.68 (0.43)13 vs �14 1.70 1.22� 2.35 1.72 (0.29) 1.68 1.01� 2.78 1.74 (0.46)14 vs �15 1.61 1.16� 2.27 1.64 (0.29) 1.64 1.00� 2.71 1.69 (0.44)15 vs �16 1.45 1.03� 2.06 1.48 (0.27) 1.50 0.91� 2.61 1.56 (0.43)16 vs �17 1.27 0.87� 1.83 1.29 (0.24) 1.26 0.73� 2.18 1.31 (0.37)17 vs �18 1.37 0.92� 2.03 1.40 (0.29) 1.27 0.62� 2.22 1.31 (0.41)18 vs �19 1.44 0.90� 2.34 1.49 (0.37) 1.42 0.71� 2.75 1.50 (0.53)19 vs �20 0.67 0.30� 1.42 0.73 (0.29) 0.82 0.32� 2.62 0.98 (0.62)20 vs �21 0.68 0.15� 2.70 0.87 (0.73) 0.60 0.01� 2.76 0.80 (0.80)

3.4 Analyses that pool tocilizumab and sarilumab

3.4.1 Primary analysis for OSFD (pooled tocilizumab/sarilumab)

Analysis of Tocilizumab and Sarilumab Page 16

REMAP-CAP (REMAP-COVID) January 5, 2021

Table 3.8: Model-estimated Odds-Ratios for the OSFD endpoint; REMAP-CAP COVID-19 severe state ITTpopulation; model with Tocilizumab and Sarilumab pooled

Odds-RatioParameter

Median95% Credible

IntervalMean (SD)

39 4.06 2.86� 5.71 4.12 (0.73)40-49 2.39 1.82� 3.14 2.41 (0.34)50-59 1.87 1.49� 2.34 1.88 (0.21)70-79 0.52 0.41� 0.67 0.53 (0.06)80+ 0.36 0.24� 0.54 0.37 (0.08)Female 1.12 0.94� 1.34 1.12 (0.10)Time-1 0.99 0.85� 1.14 0.99 (0.07)Time-2 1.03 0.81� 1.28 1.04 (0.12)Time-3 1.16 0.87� 1.55 1.17 (0.17)Time-4 1.27 0.93� 1.80 1.30 (0.22)Time-5 1.36 0.98� 1.92 1.38 (0.24)Time-6 1.41 1.02� 1.98 1.44 (0.25)Time-7 1.47 1.07� 2.08 1.50 (0.26)Time-8 1.54 1.11� 2.21 1.57 (0.28)Time-9 1.51 1.07� 2.22 1.55 (0.29)Time-10 1.36 0.94� 1.97 1.38 (0.27)Time-11 1.22 0.80� 1.81 1.24 (0.26)Time-12 1.12 0.71� 1.74 1.14 (0.27)Time-13 1.08 0.67� 1.76 1.12 (0.28)Time-14 1.11 0.65� 1.91 1.15 (0.32)Time-15 1.14 0.58� 2.22 1.21 (0.43)Tocilizumab/Sarilumab 1.65 1.27� 2.14 1.66 (0.22)Tocilizumab/Sarilumab * Fixed-Dose Corticosteroid 1.00 0.91� 1.11 1.00 (0.05)Tocilizumab/Sarilumab * Shock-based Corticosteroid 1.00 0.91� 1.10 1.00 (0.05)Tocilizumab/Sarilumab * HCQ 1.00 0.91� 1.10 1.00 (0.05)Tocilizumab/Sarilumab * Kaletra 1.01 0.92� 1.11 1.01 (0.05)Tocilizumab/Sarilumab * (HCQ + Kaletra) 1.00 0.91� 1.11 1.00 (0.05)

Note: For Age, Odds-Ratio is relative to the Age 60-69 category. Time bucket X is theXth 2-week interval prior to the most recent month, and Odds-Ratios are relative to the mostrecent month.

Table 3.9: Posterior Probabilities for the OSFD endpoint; REMAP-CAP COVID-19 severe state ITT popu-lation; model with Tocilizumab and Sarilumab pooled

Posterior Probability

Quantity of Interest Tocilizumab/Sarilumab

Prob(OR > 1) 1.0000Prob(OR < 1.2) 0.0096Prob(OR > 1) interaction with Fixed-dose corticosteroid 0.5212Prob(OR > 1) interaction with Shock-based corticosteroid 0.4792Prob(OR > 1) interaction with HCQ 0.5125Prob(OR > 1) interaction with Kaletra 0.5835Prob(OR > 1) interaction with HCQ + Kaletra 0.5170

3.4.2 Primary analysis for in-hospital mortality (pooled tocilizumab/sarilumab)

Analysis of Tocilizumab and Sarilumab Page 17

REMAP-CAP (REMAP-COVID) January 5, 2021

Table 3.10: Model-estimated Odds-Ratios for the Mortality endpoint; REMAP-CAP COVID-19 severe stateITT population; model with Tocilizumab and Sarilumab pooled

Odds-RatioParameter

Median95% Credible

IntervalMean (SD)

39 11.19 5.81� 23.35 12.03 (4.60)40-49 4.26 2.78� 6.54 4.35 (0.96)50-59 3.01 2.20� 4.14 3.05 (0.49)70-79 0.45 0.34� 0.59 0.45 (0.06)80+ 0.30 0.19� 0.46 0.30 (0.07)Female 1.14 0.89� 1.45 1.15 (0.14)Time-1 0.98 0.81� 1.15 0.98 (0.08)Time-2 0.98 0.71� 1.26 0.98 (0.14)Time-3 1.05 0.73� 1.45 1.06 (0.19)Time-4 1.13 0.76� 1.77 1.16 (0.26)Time-5 1.22 0.79� 1.96 1.25 (0.29)Time-6 1.29 0.84� 2.08 1.33 (0.31)Time-7 1.37 0.88� 2.19 1.42 (0.33)Time-8 1.48 0.97� 2.36 1.53 (0.37)Time-9 1.58 0.98� 2.63 1.64 (0.43)Time-10 1.57 0.92� 2.71 1.64 (0.46)Time-11 1.48 0.80� 2.50 1.53 (0.43)Time-12 1.38 0.71� 2.44 1.43 (0.44)Time-13 1.35 0.65� 2.53 1.42 (0.47)Time-14 1.40 0.60� 2.89 1.49 (0.58)Time-15 1.48 0.54� 3.77 1.66 (0.86)Tocilizumab/Sarilumab 1.69 1.18� 2.40 1.71 (0.31)Tocilizumab/Sarilumab * Fixed-Dose Corticosteroid 1.00 0.90� 1.09 1.00 (0.05)Tocilizumab/Sarilumab * Shock-based Corticosteroid 1.00 0.91� 1.10 1.00 (0.05)Tocilizumab/Sarilumab * HCQ 1.00 0.91� 1.11 1.00 (0.05)Tocilizumab/Sarilumab * Kaletra 1.00 0.91� 1.10 1.00 (0.05)Tocilizumab/Sarilumab * (HCQ + Kaletra) 1.00 0.91� 1.10 1.00 (0.05)

Note: For Age, Odds-Ratio is relative to the Age 60-69 category. Time bucket X is theXth 2-week interval prior to the most recent month, and Odds-Ratios are relative to the mostrecent month.

Table 3.11: Posterior Probabilities for the Mortality endpoint; REMAP-CAP COVID-19 severe state ITTpopulation; model with Tocilizumab and Sarilumab pooled

Posterior Probability

Quantity of Interest Tocilizumab/Sarilumab

Prob(OR > 1) 0.9974Prob(OR < 1.2) 0.0311Prob(OR > 1) interaction with Fixed-dose corticosteroid 0.4630Prob(OR > 1) interaction with Shock-based corticosteroid 0.5188Prob(OR > 1) interaction with HCQ 0.5181Prob(OR > 1) interaction with Kaletra 0.4924Prob(OR > 1) interaction with HCQ + Kaletra 0.5033

3.4.3 Sensitivity analysis of OSFD with less informative priors on interaction e↵ects (pooledtocilizumab/sarilumab)

Analysis of Tocilizumab and Sarilumab Page 18

REMAP-CAP (REMAP-COVID) January 5, 2021

Table 3.12: Model-estimated Odds-Ratios for the OSFD endpoint; REMAP-CAP COVID-19 severe state ITTpopulation; model with Tocilizumab and Sarilumab pooled; sensitivity analysis with less informative priors oninteraction e↵ects

Odds-RatioParameter

Median95% Credible

IntervalMean (SD)

39 4.04 2.85� 5.69 4.10 (0.72)40-49 2.39 1.80� 3.14 2.41 (0.34)50-59 1.86 1.48� 2.34 1.87 (0.22)70-79 0.52 0.40� 0.67 0.53 (0.07)80+ 0.36 0.24� 0.53 0.36 (0.08)Female 1.12 0.94� 1.34 1.13 (0.10)Time-1 0.99 0.84� 1.14 0.99 (0.07)Time-2 1.03 0.81� 1.28 1.04 (0.12)Time-3 1.15 0.88� 1.54 1.17 (0.17)Time-4 1.27 0.93� 1.81 1.30 (0.22)Time-5 1.36 0.98� 1.95 1.39 (0.25)Time-6 1.41 1.01� 2.00 1.44 (0.25)Time-7 1.48 1.06� 2.09 1.50 (0.26)Time-8 1.53 1.10� 2.19 1.56 (0.28)Time-9 1.51 1.06� 2.22 1.54 (0.29)Time-10 1.35 0.92� 2.01 1.38 (0.28)Time-11 1.21 0.79� 1.85 1.24 (0.27)Time-12 1.11 0.69� 1.75 1.14 (0.27)Time-13 1.08 0.66� 1.77 1.11 (0.28)Time-14 1.09 0.64� 1.90 1.14 (0.32)Time-15 1.11 0.58� 2.21 1.18 (0.42)Tocilizumab/Sarilumab 1.48 0.81� 2.74 1.55 (0.49)Tocilizumab/Sarilumab * Fixed-Dose Corticosteroid 1.09 0.59� 2.00 1.14 (0.37)Tocilizumab/Sarilumab * Shock-based Corticosteroid 0.90 0.38� 2.21 1.00 (0.48)Tocilizumab/Sarilumab * HCQ 1.06 0.40� 2.84 1.20 (0.65)Tocilizumab/Sarilumab * Kaletra 1.25 0.74� 2.11 1.30 (0.35)Tocilizumab/Sarilumab * (HCQ + Kaletra) 1.55 0.38� 6.00 1.97 (1.59)

Note: For Age, Odds-Ratio is relative to the Age 60-69 category. Time bucket X is theXth 2-week interval prior to the most recent month, and Odds-Ratios are relative to the mostrecent month.

Table 3.13: Posterior Probabilities for the OSFD endpoint; REMAP-CAP COVID-19 severe state ITT pop-ulation; model with Tocilizumab and Sarilumab pooled; sensitivity analysis with less informative priors oninteraction e↵ects

Posterior Probability

Quantity of Interest Tocilizumab/Sarilumab

Prob(OR > 1) 0.8952Prob(OR < 1.2) 0.2480Prob(OR > 1) interaction with Fixed-dose corticosteroid 0.6069Prob(OR > 1) interaction with Shock-based corticosteroid 0.4001Prob(OR > 1) interaction with HCQ 0.5467Prob(OR > 1) interaction with Kaletra 0.7979Prob(OR > 1) interaction with HCQ + Kaletra 0.7231

4 Other Data Summaries

This section provides summary tables and graphics for variables that are included as covariates in the model,including age, sex at birth, sites within country, and time e↵ects.

Analysis of Tocilizumab and Sarilumab Page 19

REMAP-CAP (REMAP-COVID) January 5, 2021

8 (7)8 (8)7 (7)6 (5)5 (5)5 (5)5 (5)4 (4)3 (3)3 (3)3 (3)3 (3)2 (1)2 (2)2 (2)1 (1)1 (1)1 (0)1 (1)1 (1)1 (1)1 (0)

22 (16)20 (18)17 (13)14 (10)10 (10)

9 (6)8 (8)8 (8)7 (7)6 (6)4 (4)4 (3)4 (1)4 (0)4 (4)2 (1)1 (0)1 (1)1 (1)1 (1)1 (1)

7 (7)3 (3)1 (1)

2 (2)2 (2)

20 (20)14 (14)

44 (44)17 (16)11 (11)

9 (9)7 (6)6 (6)2 (2)

5 (5)1 (1)1 (1)1 (1)1 (1)

3 (3) 114 (114) 69 (66)60 (60)48 (48)38 (37)37 (36)32 (32)31 (31)29 (29)29 (29)28 (28)27 (27)26 (25)26 (26)24 (24)23 (23)23 (23)23 (23)22 (22)22 (22)21 (21)21 (20)19 (19)18 (18)18 (18)18 (17)17 (17)17 (17)16 (16)16 (15)15 (15)15 (13)15 (15)15 (15)15 (15)14 (14)14 (14)14 (14)13 (13)13 (12)13 (13)13 (13)13 (13)12 (12)12 (12)12 (10)12 (12)12 (10)11 (11)11 (11)11 (11)11 (10)11 (11)10 (10)10 (9)10 (10)10 (10)9 (9)9 (9)9 (9)8 (8)8 (8)8 (8)8 (8)8 (6)8 (8)7 (7)7 (6)7 (7)7 (7)7 (5)6 (6)6 (6)6 (6)6 (6)6 (6)6 (6)6 (6)6 (6)6 (6)6 (6)5 (5)5 (5)5 (5)4 (4)4 (4)4 (4)4 (4)3 (3)3 (3)3 (3)3 (3)3 (3)3 (3)3 (2)3 (3)3 (2)3 (3)3 (3)3 (3)3 (3)2 (2)2 (2)2 (2)2 (2)2 (2)2 (2)2 (0)2 (2)2 (2)1 (1)1 (1)1 (1)1 (1)1 (1)1 (1)1 (1)1 (1)1 (1)1 (1)1 (1)1 (1)

92 (92)2 (2)

121120119118117116115114113112111110109108107106105104103102101100999897969594939291908988878685848382818079787776757473727170696867666564636261605958575655545352515049484746454443424140393837363534333231302928272625242322212019181716151413121110

987654321

Australia Canada France Germany Ireland Netherlands NewZealand

Portugal SaudiArabia

UnitedKingdom

UnitedStates ofAmerica

Site

with

in c

ount

ry

Figure 4.1: Sample size at each site within each country. The values in each cell represent the number ofpatients randomized to any domain at that site and, in parentheses, the number of patients for whom theoutcome on the 21-day endpoint is known. Within each country, all sites having fewer than 5 randomizedpatients are combined into a single site for the statistical model.

Analysis of Tocilizumab and Sarilumab Page 20

REMAP-CAP (REMAP-COVID) January 5, 2021

Table 4.1: Summary of the number of sites and patients randomized within each country

All Domains Immune Modulation Domain

Region CountryNumber of

Sites

Number ofPatients

Randomized

Number ofPatients w/Outcomes

Number ofSites

Number ofPatients

Randomized

Number ofPatients w/Outcomes

Canada 21 148 119 1 4 4AmericasUnited States of America 2 94 94France 3 11 11Germany 2 4 4Ireland 2 34 34 2 9 9Netherlands 7 96 94 7 68 66Portugal 1 3 3

Europe

United Kingdom 121 1405 1378 88 698 688Middle East Saudi Arabia 1 114 114 1 83 83

Australia 22 73 68 1 1 1OceanaNew Zealand 5 9 9 2 2 2

Table 4.2: Summary of age groups by sex at birth

Age Group (years)

39 40� 49 50� 59 60� 69 70� 79 � 80 Total

Male 76 155 324 416 313 85 1369Female 52 79 143 166 136 46 622Total 128 234 467 582 449 131 1991

0

200

400

600

<=39 40−49 50−59 60−69 70−79 >=80Age (years)

Num

ber o

f Sub

ject

s

Severe StateAge Groups

0

500

1000

Male FemaleSex at Birth

Num

ber o

f Sub

ject

s

Severe StateSex at Birth

Figure 4.2: Distribution of age groups and sex at birth. The total height of each bar represents the numberof patients in each category. The darker shaded area indicates the number of patients for whom 21 days haveelapsed since randomization and have a known OSFD outcome. The lighter shaded area indicates the numberof patients who do not have a known OSFD outcome.

Analysis of Tocilizumab and Sarilumab Page 21

REMAP-CAP (REMAP-COVID) January 5, 2021

500

1000

1500

2000

Mar 20

20

Apr 20

20

May 20

20

Jun 2

020

Jul 2

020

Aug 20

20

Sep 20

20

Oct 20

20

Nov 20

20

Dec 20

20

Num

ber o

f Sub

ject

s

Accrual over time (PISOP Severe)

0

200

400

600

Feb 2

5 − M

ar 10

Mar 11

− Mar

25

Mar 26

− Apr

09

Apr 10

− Apr

24

Apr 25

− May

09

May 10

− May

24

May 25

− Ju

n 08

Jun 0

9 − Ju

n 23

Jun 2

4 − Ju

l 08

Jul 0

9 − Ju

l 23

Jul 2

4 − Aug

07

Aug 08

− Aug

22

Aug 23

− Sep

06

Sep 07

− Sep

21

Sep 22

− Oct

06

Oct 07

− Oct

21

Oct 22

− Nov

19

Num

ber o

f Sub

ject

sTime Machine Intervals

Figure 4.3: Accrual over time and distribution of patients within each of the time buckets used to estimatetime trends in the analysis model. The time buckets are derived so that the first bucket is the most recentmonth going backwards in time from the most recently randomized patient in the dataset that has an outcome.Thereafter, each bucket is defined as the next two-week interval backwards in time. The total height of each barrepresents the number of patients in each category. The darker shaded area indicates the number of patientsfor whom 21 days have elapsed since randomization and have a known OSFD outcome. The lighter shaded areaindicates the number of patients who do not have a known OSFD outcome. The vertical dashed line indicatesthe randomization date for the last patient who has past 21 days and has a known outcome on the primaryendpoint at the time of this analysis.

Analysis of Tocilizumab and Sarilumab Page 22

REMAP-CAP (REMAP-COVID) January 5, 2021

5 Analysis Conventions

The following conventions were applied to the analyses contained in this report:

• The ITSC closed randomization to the Corticosteroid domain within the PISOP stratum on June 17,2020. This decision was made following the release of the RECOVERY trial results on June 16, 2020which reported strong positive e↵ects of dexamethasone. Following this decision, the results from theCorticosteroid domain in REMAP-CAP were publicly disclosed. Patients who are randomized withinthe PISOP stratum after June 17 receive no randomized assignment within the Corticosteroid domain;however it is asssumed they likely receive fixed duration steroid. Thus, for the statistical model, patientsrandomized after June 17 are coded identically to patients randomized to fixed duration steroid.

• All sites within a country that have < 5 patients randomized in the analysis population will have theirresults combined into a single site within that country.

• For the estimation of time trends in the model, time buckets with < 5 patients randomized within thebucket were combined with a neighboring bucket.

• Patients with no randomized assignment in any domain were removed from the analysis population.

• Data provided to the SAC only include patients who consented for use of their data.

• Some patients for whom 21 days have elapsed since randomization have missing 21-day OSFD outcomesin the data export. A supplemental file was provided to the SAC in which some additional outcomedata was obtained based on a manual review.

• For unique patient identifiers that exist in both the Research Online and Spiral databases, we generallypull the eligibility and randomization information from the Spiral database and the outcomes from theResearch Online database. If outcomes were reported in both places, the reported outcome in Spiralwas selected per instructions from the global project manager for the trial (email dated August 6,2020).

6 Model Stability

The Bayesian model was computed in R version 4.0.3 (2020-10-10), using the rstan package version 2.21.0.This package computes the Markov Chain Monte Carlo (MCMC) using the highly e�cient HamiltonianMonte Carlo method. The MCMC used 5 separate chains, with each chain using a burnin of 500 samples,followed by 2000 samples, for a total of 10000 samples. Convergence diagnostics were assessed, and noconcerns regarding mixing or convergence were identified. All R̂ values were less than 1.05. All model runsused a random number seed of 10252020 for the MCMC initialization.

7 Report Production

All analyses in this report are based on the following documents:

• Statistical Analysis Appendix for REMAP-COVID, version 1, dated August 18, 2020;

• Statistical Analysis Appendix for the Tocilizumab Analysis of the Immune Modulation Therapy Do-main for Patients with COVID-19 Pandemic Infection Suspected or Proven (PISOP), version 1, datedNovember 25, 2020;

• Current State of the Statistical Model: Pandemic Model, version 2.2, dated October 1, 2020;

• Instructions (“Single Source of Truth”), dated July 21, 2020.

Analysis of Tocilizumab and Sarilumab Page 23

REMAP-CAP (REMAP-COVID) January 5, 2021

Berry Consultants performed the analysis using data received from multiple sources. Table 7.1 shows thefile names for the data exports from each database along with the names of supplemental files received bythe SAC, and the dates on which each file was received by the SAC.

Table 7.1: Summary of data sources.File Name Date Received DescriptionUPMC SACDataExport 12132020 1035.csv December 13, 2020 UPMC dataremapcap spiral interimexport 2020-12-18 091609 v10.1.csv December 18, 2020 Spiral dataRAR Unscrambled RO 20201214 v3.csv December 15, 2020 Research Online datamissingOSFD PISOPSevereModeling RandDomainH 2020-12-21 CG.csv December 22, 2020 Supplemental OSFD

outcome datamissingOSFD PISOPSevereModeling ExcludingRandDomainH 2020-12-21 CG.csv December 22, 2020 Supplemental OSFD

outcome data

All data summaries were completed using the R1 statistical computing environment R version 3.5.2 (2018-12-20).

1R Development Core Team (2005). R: A language and environment for statistical computing. R Foundation for StatisticalComputing, Vienna, Austria. ISBN 3-900051-07-0. URL http://www.R-project.org.

Analysis of Tocilizumab and Sarilumab Page 24

REMAP-CAP Unblinded Analysis Report for IL-6ra

Interven ons in the Immune Modula on Therapy Domain

Prepared by the ITSC Analysis Commi ee

January 22, 2021

Contents

1 Introduc on 2

1.1 Analysis Popula ons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.2 Modeling Conven ons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

1.3 Overall summaries of OSFD in Unblinded ITT popula on . . . . . . . . . . . . . . . . . . . . . . 6

2 Secondary analyses of OSFD endpoint 8

2.1 Empirical distribu on of OSFD in IL-6ra specific ITT popula on . . . . . . . . . . . . . . . . . . . 8

2.2 Secondary analysis of OSFD endpoint for Unblinded ITT popula on . . . . . . . . . . . . . . . . . 11

2.3 Secondary analysis of OSFD endpoint for Unblinded ITT popula on with pooled IL-6ra interven ons 13

2.4 Secondary analysis of OSFD endpoint for Unblinded ITT popula on restricted to non-nega ve COVID-

19 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2.5 Secondary analysis of OSFD endpoint for IL-6ra specific ITT popula on . . . . . . . . . . . . . . . 17

3 Secondary analyses of in-hospital mortality 18

3.1 Empirical distribu on of in-hospital mortality in IL-6ra specific ITT popula on . . . . . . . . . . . 18

3.2 Secondary analysis of in-hospital mortality for Unblinded ITT popula on . . . . . . . . . . . . . . 18

3.3 Secondary analysis of in-hospital mortality for Unblinded ITT popula on with pooled IL-6ra inter-

ven ons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.4 Secondary analysis of in-hospital mortality for Unblinded ITT popula on restricted to non-nega ve

COVID-19 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

3.5 Secondary analysis of in-hospital mortality for IL-6ra specific ITT popula on . . . . . . . . . . . . 24

4 Primary safety analysis 25

1

4.1 Empirical distribu on of SAE in IL-6ra specific ITT popula on . . . . . . . . . . . . . . . . . . . . 26

4.2 Primary safety analysis in IL-6ra specific ITT popula on . . . . . . . . . . . . . . . . . . . . . . . 26

5 Analyses of secondary endpoints 26

5.1 Secondary analysis of mortality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

5.2 Secondary analysis of mortality with pooled IL-6ra interven ons . . . . . . . . . . . . . . . . . . 29

5.3 Secondary analysis of progression to intuba on, ECMO, or death . . . . . . . . . . . . . . . . . . 31

5.4 Secondary analysis of days free of vasopressors/inotropes . . . . . . . . . . . . . . . . . . . . . 34

5.5 Secondary analysis of days free of ven la on . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

5.6 Secondary analysis of length of ICU stay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

5.7 Secondary analysis of length of hospital stay . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43

5.8 Secondary analysis of WHO scale at 14 days . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46

6 Sensi vity Analyses 48

6.1 Sensi vity analyses of OSFD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

6.2 Sensi vity analyses of in-hospital mortality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52

7 Subgroup Analyses 56

7.1 Subgroup analyses of OSFD endpoint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

7.2 Subgroup analyses of in-hospital mortality . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62

8 Post Hoc Analyses 67

8.1 Post hoc analysis of OSFD without borrowing between IL-6ra interven ons . . . . . . . . . . . . . 68

8.2 Post hoc analysis of in-hospital mortality without borrowing between IL-6ra interven ons . . . . . 70

8.3 Data summaries of baseline mechanical ven la on status in the IL-6ra specific ITT popula on . . . 72

8.4 Post hoc analysis with differen al effect on OSFD endpoint by baselinemechanical ven la on status

with pooled IL-6ra interven ons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75

8.5 Post hoc analysis with differen al effect on in-hospital mortality by baseline mechanical ven la on

status with pooled IL-6ra interven ons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77

1 Introduc on

This report summarizes the data and the results for the IL-6ra interven on analyses run by the ITSC analysis com-

mi ee. The ITSC analysis commi ee is blinded to all ongoing domains and interven ons in REMAP-CAP.

2

1.1 Analysis Popula ons

The SAP for the IL-6ra interven ons outlines five popula ons in which analyses will be performed. This report in-

cludes analysis results for four of the five analysis popula ons. The analysis results for the REMAP-CAP COVID-19 se-

vere state intent-to-treat (ITT) popula on are not included in this report since the ITSC analysis commi ee remains

blinded to ongoing domains and interven ons. Results for the REMAP-CAP COVID-19 severe state intent-to-treat

(ITT) popula on are included in a separate report prepared by the fully unblinded Sta s cal Analysis Commi ee

(SAC). In this report, tables and figures summarize data in the IL-6ra specific ITT popula on. Thus, the figures/tables

in this report may differ from those in the SAC report based on the difference in popula on defini on. In this report,

we summarize analysis results from the following four analysis popula ons:

1. The Unblinded ITT popula on is defined as all pa ents randomized to tocilizumab, sarilumab or no immune

modula on (control) in the Immune Modula on Therapy Domain, the Cor costeroid Domain, or the An -

Viral Domain within the pandemic stratum. The unblinded ITT popula on consists of 1293 pa ents. There

are 22 pa ents within this popula on that are missing values of OSFD and in-hospital mortality.

• 406 pa ents randomized to no immune modula on (control) of which 401 have known OSFD outcomes

• 353 pa ents randomized to tocilizumab of which 350 have known OSFD outcomes

• 48 pa ents randomized to sarilumab of which 45 have known OSFD outcomes

• 381 pa ents randomized to the cor costeroid domain of which 380 have known OSFD outcomes

• 694 pa ents randomized to the an viral domain of which 677 have known OSFD outcomes.

2. The Unblinded non-nega ve COVID-19 popula on is defined as all pa ents in the Unblinded ITT popula on

a er removing those with >1 nega ve test for COVID-19 and no posi ve tests. The unblinded ITT popula on

restricted to non-nega ve COVID-19 consists of 1136 pa ents. There are 21 pa ents within this popula on

that are missing values of OSFD and in-hospital mortality.

• 365 pa ents randomized to no immune modula on (control) of which 360 have known OSFD outcomes

• 314 pa ents randomized to tocilizumab of which 311 have known OSFD outcomes

• 47 pa ents randomized to sarilumab of which 44 have known OSFD outcomes

• 315 pa ents randomized to the cor costeroid domain of which 314 have known OSFD outcomes

• 619 pa ents randomized to the an viral domain of which 603 have known OSFD outcomes

3. The IL-6ra specific ITT popula on consists of only pa ents eligible to be randomized to an IL-6ra that were

randomized to tocilizumab, sarilumab or the no immunemodula on interven on in the ImmuneModula on

3

Therapy Domain within the pandemic stratum. The IL-6ra specific ITT popula on consists of 803 pa ents.

There are 11 pa ents within this popula on that are missing values of OSFD and in-hospital mortality.

• 402 pa ents randomized to no immune modula on (control) of which 397 have known OSFD outcomes

• 353 pa ents randomized to tocilizumab of which 350 have known OSFD outcomes

• 48 pa ents randomized to sarilumab of which 45 have known OSFD outcomes

4. The IL-6ra specific PP popula on consists of the pa ents in the IL-6ra specific ITT popula on who have been

treated as per protocol. In this analysis that is defined as pa ents randomized to tocilizumab or sarilumab

that received at least one dose of IL-6ra or pa ents randomized to no immunemodula on that did not receive

any interven ons in the Immune Modula on Therapy Domain. The IL-6ra specific per protocol popula on

consists of 740 pa ents. There are 6 pa ents within this popula on that are missing values of OSFD and

in-hospital mortality.

• 382 pa ents randomized to no immune modula on (control) of which 379 have known OSFD outcomes

• 315 pa ents randomized to tocilizumab of which 315 have known OSFD outcomes

• 43 pa ents randomized to sarilumab of which 40 have known OSFD outcomes

1.2 Modeling Conven ons

• All reported credible intervals (CrIs) are 95% equal-tailed intervals.

• Results from models of ordinal and dichotomous endpoints are reported as odds ra os (ORs). Results from

models of me to event endpoints are reported as hazard ra os (HRs). For consistency of interpreta on, all

models are parameterized so that anOR/HR greater than 1 indicates pa ent benefit rela ve to the reference

group and an OR/HR less than 1 indicates pa ent harm rela ve to the reference group.

• The reference group for the age category OR/HRs is the age category from 60-69 years old.

• The reference group for the me epochs OR/HRs is the most recent me epoch consis ng of the four-week

period preceding Nov 19, 2020. Time epoch 0 is the most recent epoch and the epochs move backwards

in me in two-week periods from epoch 1 to 16. The Unblinded ITT and Unblinded non-nega ve COVID-19

popula ons include pa ents randomized in epochs 0 to 16. The IL-6ra specific ITT and PP popula ons include

pa ents randomized in epochs 0 to 13.

• The reference group for the sex at birth OR/HRs is the male category.

• The reference group for the CRP tercile subgroup OR/HRs is the largest subgroup of pa ents with unknown

CRP.

4

• Tocilizumab and sarilumab are compared to the control arm. A posterior probability of superiority of 99% is

used as a sta s cal trigger for efficacy.

• Tocilizumab and sarilumab are compared to control for fu lity. A 95% probability of a smaller than 1.2 odds

ra o for Tocilizumab or sarilumab rela ve to control is used as a sta s cal trigger for fu lity.

• Tocilizumab and sarilumab are compared for equivalence. A 90% probability of equivalence (defined as an

odds ra o of tocilizumab rela ve to sarilumab between 1/1.2 and 1.2) is used as a sta s cal trigger for inter-

ven on equivalence.

• OR/HR effects for pre-specified combina ons of interven ons from the Immune Modula on Therapy and

Cor costeroid domain are reported rela ve to control. These OR/HRs incorporate the effect of each individ-

ual interven on and the interac on term for the combina on of interven ons.

• OR/HR effects for pre-specified interac on effects are reported rela ve to an addi ve effect. For example, if

an interac on effect OR/HR is equal to 1, the combina on of interven ons is addi ve (equal to the sum of

the effects of the two interven ons taken separately). If an interac on OR/HR is greater than 1, the effect of

the combina on of interven ons is synergis c (greater than the sum of the effects of the two interven ons

taken separately). If the interac on OR/HR is less than 1, the effect of the combina on is sub-addi ve (less

than the sum of the effects of the two interven ons taken separately).

• Interac on effects are reported for pre-specified interven ons from the Immune Modula on Therapy and

Cor costeroid domain.

5

1.3 Overall summaries of OSFD in Unblinded ITT popula on

38413277312623212641234861951261004611

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Mar 12−M

ar 25

(16)

Mar 26−A

pr 08

(15)

Apr 09−A

pr 22

(14)

Apr 23−M

ay 06

(13)

May 07−M

ay 20

(12)

May 21−J

un 03

(11)

Jun 0

4−Ju

n 17 (

10)

Jun 1

8−Ju

l 01 (

9)

Jul 0

2−Ju

l 15 (

8)

Jul 1

6−Ju

l 29 (

7)

Jul 3

0−Aug

12 (6

)

Aug 13−A

ug 26

(5)

Aug 27−S

ep 09

(4)

Sep 10−S

ep 23

(3)

Sep 24−O

ct 07

(2)

Oct 08−O

ct 21

(1)

Oct 22−N

ov 19

(0)

Randomization Date

Prop

ortio

n

OSFD−1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 1: Empirical distribu on of organ support free days (OSFD) by me epoch. The Oct 22-Nov 19 category

(epoch 0) is the reference me epoch in all models. This plot shows the me epochs defined for the Unblinded

ITT popula on, with the labeled number shown in parantheses. Other analysis popula ons may have fewer me

epochs based on the randomiza on dates of included pa ents.

6

78 154 308 380 282 69

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

<39 40−49 50−59 60−69 70−79 80+Age Category

Prop

ortio

n

OSFD−1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 2: Empirical distribu on of organ support free days (OSFD) by age category for the Unblinded ITT popula on.

The 60−69 year age category is the reference group in all models.

899

372

Male

Female

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0Proportion

Sex

at B

irth

OSFD−1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 3: Empirical distribu on of organ support free days (OSFD) by sex at birth for the Unblinded ITT popula on.

The male category is the reference group in all models.

7

85 1118945 25 459 13111

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

United

Kingdo

m

Saudi

Arabia

Netherl

ands

Canad

a

Austra

liaIre

land

United

States

France

New Zea

land

Prop

ortio

n

OSFD−1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 4: Empirical distribu on of organ support free days (OSFD) by country for the Unblinded ITT popula on.

2 Secondary analyses of OSFD endpoint

2.1 Empirical distribu on of OSFD in IL-6ra specific ITT popula on

Table 1: Summary of OSFD for the IL-6ra specific ITT popula on

Interven on # Pa ents # Known OSFD median (IQR) Days Free of Organ Support in Survivors* median (IQR)Tocilizumab 353 350 10 (-1, 16) 14 (7, 17)Sarilumab 48 45 11 (0, 16) 15 (6.5, 17)Control 402 397 0 (-1, 15) 13 (4, 17)

Pooled IL-6ra 401 395 10 (-1, 16) 14 (7, 17)* Days Free of Organ Support in Survivors within 21 days

8

Control (397)

Sarilumab (45)

Tocilizumab (350)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0Proportion

OSFD−1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 5: Empirical distribu on of organ support free days (OSFD) for tocilizumab, sarilumab, and control. This plot

is restricted to the IL-6ra ITT popula on.

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 5 10 15 20Organ support free days

Cum

ulat

ive P

ropo

rtion

of P

atie

nts

Tocilizumab (350)

Sarilumab (45)

Control (397)

Figure 6: Empirical cumula ve distribu on of organ support free days (OSFD) for tocilizumab, sarilumab, and control.

This plot is restricted to the IL-6ra ITT popula on.

9

Control (397)

Pooled IL−6ra (395)

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0Proportion

OSFD−1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 7: Empirical cumula ve distribu on of organ support free days (OSFD) for IL-6ra interven ons and control.

This plot is restricted to the IL-6ra ITT popula on.

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 5 10 15 20Organ support free days

Cum

ulat

ive P

ropo

rtion

of P

atie

nts

Pooled IL−6ra (395)

Control (397)

Figure 8: Empirical cumula ve distribu on of organ support free days (OSFD) for IL-6ra and control interven ons.

This plot is restricted to the IL-6ra ITT popula on.

10

361

358

36

37

No corticosteroids

Pooled corticosteroids

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Control

Pooled IL−6ra

Control

Pooled IL−6ra

Proportion

OSFD−1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 9: Empirical cumula ve distribu on of organ support free days (OSFD) for the IL-6ra, separated by pa ents

who received no cor costeroid interven ons and pa ents who received a cor costeroid interven ons. This plot is

restricted to the IL-6ra ITT popula on.

2.2 Secondary analysis of OSFD endpoint for Unblinded ITT popula on

• Model: Primary analysis ordinal model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, fixed-dura on, shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ),

and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

11

Table 2: Odds ra o parameters for secondary analysis of OSFD endpoint for Unblinded ITT popula on

Mean SD Median CrIAge<39 4.11 0.91 4.01 (2.61, 6.17)Age 40-49 2.11 0.36 2.08 (1.49, 2.88)Age 50-59 1.96 0.27 1.94 (1.48, 2.56)Age 70-79 0.52 0.08 0.51 (0.38, 0.69)Age 80+ 0.34 0.09 0.32 (0.19, 0.55)Female 1.17 0.13 1.16 (0.93, 1.44)Time epoch 1 0.95 0.08 0.95 (0.79, 1.11)Time epoch 2 0.91 0.13 0.90 (0.66, 1.17)Time epoch 3 0.96 0.17 0.95 (0.67, 1.33)Time epoch 4 1.06 0.22 1.04 (0.70, 1.56)Time epoch 5 1.16 0.27 1.13 (0.74, 1.78)Time epoch 6 1.22 0.29 1.18 (0.77, 1.88)Time epoch 7 1.26 0.29 1.22 (0.80, 1.92)Time epoch 8 1.29 0.30 1.25 (0.83, 1.98)Time epoch 9 1.23 0.27 1.20 (0.80, 1.88)Time epoch 10 1.13 0.24 1.10 (0.74, 1.69)Time epoch 11 0.97 0.20 0.95 (0.64, 1.41)Time epoch 12 0.86 0.19 0.84 (0.55, 1.27)Time epoch 13 0.85 0.19 0.83 (0.54, 1.26)Time epoch 14 0.94 0.22 0.92 (0.59, 1.43)Time epoch 15 1.11 0.32 1.06 (0.61, 1.86)Time epoch 16 1.41 0.66 1.26 (0.59, 3.08)Tocilizumab 1.68 0.24 1.66 (1.26, 2.18)Sarilumab 1.84 0.44 1.77 (1.18, 2.90)Fixed-dose cor costeroid 1.45 0.33 1.42 (0.92, 2.18)Shock-dependent cor costeroid 1.15 0.26 1.12 (0.72, 1.75)Tocilizumab*Fixed-dose interac on effect 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Shock-dependent interac on effect 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Fixed-dose interac on effect 1.01 0.05 1.00 (0.91, 1.11)Sarilumab*Shock-dependent interac on effect 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose combina on 2.42 0.65 2.34 (1.39, 3.91)Tocilizumab*Shock-dependent combina on 1.93 0.54 1.86 (1.08, 3.17)Sarilumab*Fixed-dose combina on 2.68 0.90 2.53 (1.37, 4.86)Sarilumab*Shock-dependent combina on 2.11 0.72 1.99 (1.08, 3.87)

12

Table 3: Posterior probabili es for secondary analysis of OSFD endpoint for Unblinded ITT popula on

Posterior ProbabilityTocilizumab is superior to control 1.000Tocilizumab is fu le (OR < 1.2) 0.010Sarilumab is superior to control 0.996Sarilumab is fu le (OR < 1.2) 0.030Tocilizumab*Fixed dose combina on OR > 1 0.999Tocilizumab*Shock-dependent combina on OR > 1 0.988Sarilumab*Fixed dose combina on OR > 1 0.998Sarilumab*Shock-dependent combina on OR > 1 0.985Tocilizumab and Sarilumab are equivalent 0.676

2.3 Secondary analysis ofOSFDendpoint forUnblinded ITT popula onwith pooled IL-6ra interven ons

• Model: Primary analysis ordinal model

• Factors: Age, sex, site, me, pooled tocilizumab and sarilumab, and control interven ons, cor costeroid

interven ons (control, fixed-dura on, shock-dependent) and an viral interven ons (control, HCQ, Kaletra,

Kaletra + HCQ), and interac ons between pooled tocilizumab and sarilumab, cor costeroids, and an virals

13

Table 4: Odds ra o parameters for secondary analysis of OSFD endpoint for Unblinded ITT popula on with pooled

IL-6ra interven ons

Mean SD Median CrIAge<39 4.11 0.89 4.01 (2.64, 6.10)Age 40-49 2.11 0.36 2.08 (1.50, 2.92)Age 50-59 1.96 0.27 1.95 (1.48, 2.54)Age 70-79 0.52 0.08 0.51 (0.38, 0.69)Age 80+ 0.34 0.09 0.33 (0.19, 0.55)Female 1.17 0.13 1.16 (0.93, 1.44)Time epoch 1 0.95 0.08 0.95 (0.79, 1.11)Time epoch 2 0.90 0.13 0.90 (0.66, 1.17)Time epoch 3 0.96 0.17 0.95 (0.67, 1.33)Time epoch 4 1.06 0.22 1.03 (0.70, 1.56)Time epoch 5 1.16 0.27 1.12 (0.73, 1.81)Time epoch 6 1.21 0.29 1.17 (0.76, 1.91)Time epoch 7 1.25 0.29 1.21 (0.78, 1.93)Time epoch 8 1.28 0.30 1.24 (0.81, 1.96)Time epoch 9 1.22 0.27 1.19 (0.79, 1.84)Time epoch 10 1.12 0.24 1.09 (0.73, 1.68)Time epoch 11 0.97 0.20 0.95 (0.63, 1.41)Time epoch 12 0.86 0.19 0.84 (0.54, 1.27)Time epoch 13 0.86 0.19 0.84 (0.54, 1.28)Time epoch 14 0.94 0.22 0.92 (0.59, 1.44)Time epoch 15 1.10 0.32 1.06 (0.62, 1.87)Time epoch 16 1.39 0.63 1.26 (0.59, 2.98)Pooled IL-6ra 1.68 0.23 1.67 (1.27, 2.18)Fixed-dose cor costeroid 1.45 0.33 1.41 (0.92, 2.20)Shock-dependent cor costeroid 1.15 0.27 1.11 (0.71, 1.75)Pooled IL-6ra*Fixed-dose interac on 1.00 0.05 1.00 (0.91, 1.10)Pooled IL-6ra*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Pooled IL-6ra*Fixed-dose combina on 2.45 0.66 2.36 (1.40, 3.99)Pooled IL-6ra*Shock-dependent combina on 1.93 0.53 1.86 (1.10, 3.15)

Table 5: Posterior probabili es for secondary analysis of OSFD endpoint for Unblinded ITT popula on with pooled

IL-6ra interven ons

Posterior ProbabilityPooled IL-6ra is superior to control 1.000Pooled IL-6ra is fu le (OR < 1.2) 0.008Pooled IL-6ra*Fixed-dose combina on OR > 1 1.000Pooled IL-6ra*Shock-dependent combina on OR > 1 0.990

14

2.4 Secondary analysis of OSFD endpoint for Unblinded ITT popula on restricted to non-nega ve

COVID-19

• Model: Primary analysis ordinal model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, fixed-dura on, shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ),

and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

15

Table 6: Odds ra o parameters for secondary analysis of OSFD endpoint for Unblinded ITT popula on restricted to

non-nega ve COVID-19

Mean SD Median CrIAge<39 4.48 1.09 4.35 (2.73, 7.00)Age 40-49 2.10 0.38 2.07 (1.45, 2.93)Age 50-59 1.96 0.29 1.94 (1.46, 2.56)Age 70-79 0.51 0.08 0.50 (0.36, 0.68)Age 80+ 0.27 0.08 0.26 (0.15, 0.45)Female 1.12 0.14 1.11 (0.87, 1.41)Time epoch 1 0.96 0.08 0.96 (0.80, 1.13)Time epoch 2 0.94 0.13 0.93 (0.69, 1.21)Time epoch 3 0.98 0.18 0.97 (0.67, 1.37)Time epoch 4 1.07 0.24 1.04 (0.69, 1.64)Time epoch 5 1.15 0.30 1.10 (0.70, 1.87)Time epoch 6 1.17 0.32 1.12 (0.70, 1.91)Time epoch 7 1.15 0.31 1.11 (0.68, 1.87)Time epoch 8 1.12 0.29 1.08 (0.67, 1.79)Time epoch 9 1.05 0.26 1.01 (0.65, 1.64)Time epoch 10 0.96 0.22 0.93 (0.60, 1.45)Time epoch 11 0.85 0.19 0.83 (0.53, 1.28)Time epoch 12 0.80 0.19 0.78 (0.48, 1.21)Time epoch 13 0.83 0.19 0.81 (0.51, 1.27)Time epoch 14 0.92 0.23 0.89 (0.55, 1.45)Time epoch 15 1.05 0.32 1.00 (0.56, 1.82)Time epoch 16 1.26 0.57 1.14 (0.52, 2.67)Tocilizumab 1.72 0.26 1.70 (1.26, 2.27)Sarilumab 1.94 0.50 1.86 (1.22, 3.19)Fixed-dose cor costeroid 1.29 0.31 1.25 (0.78, 2.01)Shock-dependent cor costeroid 0.97 0.24 0.94 (0.58, 1.53)Tocilizumab*Fixed-dose interac on 1.00 0.05 0.99 (0.91, 1.10)Tocilizumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Fixed-dose interac on 1.01 0.05 1.00 (0.91, 1.11)Sarilumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose combina on 2.21 0.65 2.11 (1.22, 3.72)Tocilizumab*Shock-dependent combina on 1.66 0.50 1.59 (0.89, 2.87)Sarilumab*Fixed-dose combina on 2.52 0.91 2.36 (1.23, 4.78)Sarilumab*Shock-dependent combina on 1.89 0.69 1.76 (0.91, 3.61)

16

Table 7: Posterior probabili es for secondary analysis of OSFD endpoint for Unblinded ITT popula on restricted to

non-nega ve COVID-19

Posterior ProbabilityTocilizumab is superior to control 1.000Tocilizumab is fu le (OR < 1.2) 0.011Sarilumab is superior to control 0.997Sarilumab is fu le (OR < 1.2) 0.021Tocilizumab*Fixed-dose combina on OR > 1 0.996Tocilizumab*Shock-dependent combina on OR > 1 0.945Sarilumab*Fixed-dose combina on OR > 1 0.995Sarilumab*Shock-dependent combina on OR > 1 0.954Tocilizumab and Sarilumab are equivalent 0.642

2.5 Secondary analysis of OSFD endpoint for IL-6ra specific ITT popula on

• Model: Primary analysis ordinal model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons

Table 8: Odds ra o parameters for sensi vity analysis of OSFD in IL-6ra specific ITT popula on

Mean SD Median CrIAge<39 4.28 1.18 4.14 (2.44, 7.04)Age 40-49 2.19 0.47 2.13 (1.41, 3.25)Age 50-59 2.50 0.44 2.45 (1.76, 3.46)Age 70-79 0.59 0.11 0.58 (0.40, 0.85)Age 80+ 0.33 0.10 0.32 (0.17, 0.58)Female 1.22 0.18 1.21 (0.91, 1.61)Time epoch 1 0.98 0.08 0.98 (0.83, 1.15)Time epoch 2 0.98 0.14 0.97 (0.73, 1.26)Time epoch 3 1.03 0.19 1.01 (0.70, 1.44)Time epoch 4 1.10 0.24 1.07 (0.71, 1.67)Time epoch 5 1.17 0.30 1.13 (0.72, 1.90)Time epoch 6 1.18 0.30 1.14 (0.72, 1.92)Time epoch 7 1.18 0.30 1.14 (0.72, 1.88)Time epoch 8 1.15 0.28 1.12 (0.72, 1.80)Time epoch 9 1.07 0.24 1.04 (0.69, 1.62)Time epoch 10 0.98 0.20 0.96 (0.65, 1.44)Time epoch 11 0.87 0.16 0.86 (0.59, 1.24)Time epoch 12 0.76 0.15 0.75 (0.51, 1.09)Time epoch 13 0.69 0.17 0.67 (0.41, 1.07)Tocilizumab 1.69 0.23 1.67 (1.29, 2.18)Sarilumab 1.91 0.47 1.83 (1.23, 3.04)

17

Table 9: Posterior probabili es for sensi vity analysis of OSFD in IL-6ra specific ITT popula on

Posterior ProbabilityTocilizumab is superior to control 1.000Tocilizumab is fu le (OR < 1.2) 0.006Sarilumab is superior to control 0.999Sarilumab is fu le (OR < 1.2) 0.019Tocilizumab and Sarilumab are equivalent 0.658

3 Secondary analyses of in-hospital mortality

For consistency of interpreta on, all models are parameterized so that an OR/HR greater than 1 indicates pa ent

benefit rela ve to the reference group and an OR/HR less than 1 indicates pa ent harm rela ve to the reference

group. In this sec on, the ORs can be interpreted for the outcome of in-hospital survival.

3.1 Empirical distribu on of in-hospital mortality in IL-6ra specific ITT popula on

Table 10: Summary of in-hospital mortality for the IL-6ra specific ITT popula on

Interven on # Pa ents (N) # Known (n) Deaths (y) Mortality Rate (y/n)Tocilizumab 353 350 98 0.280Sarilumab 48 45 10 0.222Control 402 397 142 0.358

Pooled IL-6ra 401 395 108 0.273

3.2 Secondary analysis of in-hospital mortality for Unblinded ITT popula on

• Model: Primary dichotomous model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, fixed-dura on, shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ),

and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

18

Table 11: Odds ra o parameters for secondary analysis of in-hospital mortality for Unblinded ITT popula on

Mean SD Median CrIAge<39 12.85 6.45 11.38 (5.00, 28.81)Age 40-49 3.33 0.89 3.21 (1.95, 5.45)Age 50-59 2.99 0.60 2.93 (2.00, 4.34)Age 70-79 0.41 0.07 0.41 (0.29, 0.58)Age 80+ 0.23 0.07 0.22 (0.12, 0.39)Female 1.19 0.19 1.18 (0.87, 1.59)Time epoch 1 0.94 0.08 0.94 (0.78, 1.10)Time epoch 2 0.87 0.14 0.87 (0.61, 1.14)Time epoch 3 0.86 0.17 0.85 (0.55, 1.22)Time epoch 4 0.87 0.20 0.86 (0.53, 1.33)Time epoch 5 0.90 0.23 0.87 (0.53, 1.41)Time epoch 6 0.91 0.24 0.88 (0.53, 1.48)Time epoch 7 0.92 0.25 0.89 (0.52, 1.48)Time epoch 8 0.95 0.25 0.92 (0.55, 1.53)Time epoch 9 0.99 0.26 0.95 (0.58, 1.59)Time epoch 10 1.02 0.27 0.98 (0.60, 1.66)Time epoch 11 1.02 0.27 0.97 (0.60, 1.65)Time epoch 12 0.99 0.26 0.96 (0.58, 1.58)Time epoch 13 1.00 0.28 0.96 (0.57, 1.63)Time epoch 14 1.08 0.33 1.03 (0.57, 1.86)Time epoch 15 1.23 0.50 1.14 (0.55, 2.43)Time epoch 16 1.52 1.03 1.27 (0.50, 3.95)Tocilizumab 1.67 0.31 1.65 (1.15, 2.34)Sarilumab 2.24 0.94 2.00 (1.17, 4.69)Fixed-dose cor costeroid 0.99 0.29 0.95 (0.54, 1.68)Shock-dependent cor costeroid 1.21 0.38 1.15 (0.63, 2.10)Tocilizumab*Fixed-dose interac on 0.99 0.05 0.99 (0.89, 1.09)Tocilizumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.11)Sarilumab*Fixed-dose interac on 1.01 0.05 1.00 (0.91, 1.11)Sarilumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose combina on 1.63 0.58 1.53 (0.77, 3.03)Tocilizumab*Shock-dependent combina on 2.03 0.76 1.90 (0.92, 3.87)Sarilumab*Fixed-dose combina on 2.21 1.14 1.94 (0.86, 5.21)Sarilumab*Shock-dependent combina on 2.69 1.42 2.35 (1.02, 6.35)

19

Table 12: Posterior probabili es for secondary analysis of in-hospital mortality for Unblinded ITT popula on

Posterior ProbabilityTocilizumab is superior to control 0.996Tocilizumab is fu le (OR < 1.2) 0.045Sarilumab is superior to control 0.994Sarilumab is fu le (OR < 1.2) 0.030Tocilizumab*Fixed-dose combina on OR > 1 0.893Tocilizumab*Shock-dependent combina on OR > 1 0.962Sarilumab*Fixed-dose combina on OR > 1 0.944Sarilumab*Shock-dependent combina on OR > 1 0.977Tocilizumab and Sarilumab are equivalent 0.495

3.3 Secondary analysis of in-hospital mortality for Unblinded ITT popula on with pooled IL-6ra inter-

ven ons

• Model: Primary analysis ordinal model

• Factors: Age, sex, site, me, pooled tocilizumab and sarilumab, and control interven ons, cor costeroid

interven ons (control, fixed-dura on, shock-dependent) and an viral interven ons (control, HCQ, Kaletra,

Kaletra + HCQ), and interac ons between pooled tocilizumab and sarilumab, cor costeroids, and an virals

20

Table 13: Odds ra o parameters for secondary analysis of in-hospital mortality for Unblinded ITT popula on with

pooled IL-6ra interven ons

Mean SD Median CrIAge<39 13.06 6.69 11.58 (5.07, 30.45)Age 40-49 3.32 0.88 3.20 (1.97, 5.40)Age 50-59 2.97 0.60 2.91 (1.99, 4.33)Age 70-79 0.41 0.07 0.41 (0.29, 0.58)Age 80+ 0.23 0.07 0.22 (0.12, 0.39)Female 1.20 0.19 1.18 (0.87, 1.61)Time epoch 1 0.93 0.08 0.93 (0.77, 1.10)Time epoch 2 0.86 0.14 0.85 (0.60, 1.14)Time epoch 3 0.84 0.18 0.83 (0.53, 1.22)Time epoch 4 0.86 0.21 0.84 (0.52, 1.34)Time epoch 5 0.89 0.24 0.86 (0.52, 1.45)Time epoch 6 0.90 0.25 0.86 (0.51, 1.47)Time epoch 7 0.90 0.24 0.87 (0.52, 1.47)Time epoch 8 0.93 0.25 0.90 (0.54, 1.50)Time epoch 9 0.97 0.26 0.94 (0.58, 1.57)Time epoch 10 1.02 0.28 0.98 (0.60, 1.67)Time epoch 11 1.01 0.27 0.97 (0.59, 1.65)Time epoch 12 0.98 0.26 0.94 (0.57, 1.59)Time epoch 13 0.98 0.28 0.94 (0.55, 1.65)Time epoch 14 1.07 0.34 1.01 (0.56, 1.86)Time epoch 15 1.23 0.51 1.14 (0.55, 2.44)Time epoch 16 1.55 1.10 1.30 (0.50, 4.06)Pooled IL-6ra 1.72 0.32 1.68 (1.18, 2.43)Fixed-dose cor costeroid 0.99 0.30 0.95 (0.53, 1.71)Shock-dependent cor costeroid 1.21 0.39 1.15 (0.63, 2.16)Pooled IL-6ra*Fixed-dose interac on 0.99 0.05 0.99 (0.90, 1.09)Pooled IL-6ra*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.11)Pooled IL-6ra*Fixed-dose combina on 1.69 0.61 1.59 (0.80, 3.18)Pooled IL-6ra*Shock-dependent combina on 2.09 0.81 1.94 (0.95, 4.12)

Table 14: Posterior probabili es for secondary analysis of in-hospital mortality for Unblinded ITT popula on with

pooled IL-6ra interven ons

Posterior ProbabilityPooled IL-6ra is superior to control 0.997Pooled IL-6ra is fu le (OR < 1.2) 0.031Pooled IL-6ra*Fixed-dose combina on OR > 1 0.902Pooled IL-6ra*Shock-dependent combina on OR > 1 0.964

21

3.4 Secondary analysis of in-hospital mortality for Unblinded ITT popula on restricted to non-nega ve

COVID-19

• Model: Primary dichotomous model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, fixed-dura on, shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ),

and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

22

Table 15: Odds ra o parameters for secondary analysis of in-hospital mortality for Unblinded ITT popula on re-

stricted to non-nega ve COVID-19

Mean SD Median CrIAge<39 12.94 7.47 11.13 (4.40, 32.35)Age 40-49 3.10 0.85 2.97 (1.80, 5.09)Age 50-59 2.84 0.60 2.78 (1.87, 4.17)Age 70-79 0.43 0.08 0.42 (0.29, 0.60)Age 80+ 0.21 0.07 0.20 (0.11, 0.37)Female 1.18 0.20 1.16 (0.84, 1.62)Time epoch 1 0.95 0.08 0.94 (0.79, 1.12)Time epoch 2 0.89 0.14 0.88 (0.63, 1.17)Time epoch 3 0.87 0.18 0.85 (0.55, 1.26)Time epoch 4 0.87 0.22 0.85 (0.52, 1.37)Time epoch 5 0.87 0.25 0.84 (0.49, 1.44)Time epoch 6 0.87 0.26 0.83 (0.47, 1.47)Time epoch 7 0.86 0.26 0.82 (0.46, 1.46)Time epoch 8 0.87 0.25 0.84 (0.48, 1.47)Time epoch 9 0.90 0.26 0.86 (0.50, 1.51)Time epoch 10 0.93 0.27 0.89 (0.52, 1.58)Time epoch 11 0.94 0.27 0.90 (0.53, 1.56)Time epoch 12 0.95 0.27 0.91 (0.53, 1.57)Time epoch 13 0.99 0.30 0.96 (0.54, 1.69)Time epoch 14 1.10 0.37 1.05 (0.56, 1.98)Time epoch 15 1.28 0.55 1.18 (0.55, 2.64)Time epoch 16 1.58 1.10 1.33 (0.49, 4.18)Tocilizumab 1.65 0.32 1.62 (1.12, 2.38)Sarilumab 2.36 1.10 2.07 (1.18, 5.24)Fixed-dose cor costeroid 0.92 0.30 0.87 (0.46, 1.62)Shock-dependent cor costeroid 1.07 0.37 1.01 (0.52, 1.93)Tocilizumab*Fixed-dose interac on 0.99 0.05 0.99 (0.90, 1.09)Tocilizumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Fixed-dose interac on 1.01 0.05 1.01 (0.91, 1.11)Sarilumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose combina on 1.49 0.57 1.39 (0.67, 2.90)Tocilizumab*Shock-dependent combina on 1.77 0.71 1.64 (0.79, 3.54)Sarilumab*Fixed-dose combina on 2.18 1.26 1.86 (0.79, 5.53)Sarilumab*Shock-dependent combina on 2.53 1.52 2.14 (0.89, 6.40)

23

Table 16: Posterior probabili es for secondary analysis of in-hospital mortality for Unblinded ITT popula on re-

stricted to non-nega ve COVID-19

Posterior ProbabilityTocilizumab is superior to control 0.996Tocilizumab is fu le (OR < 1.2) 0.056Sarilumab is superior to control 0.996Sarilumab is fu le (OR < 1.2) 0.032Tocilizumab*Fixed-dose combina on OR > 1 0.816Tocilizumab*Shock-dependent combina on OR > 1 0.903Sarilumab*Fixed-dose combina on OR > 1 0.915Sarilumab*Shock-dependent combina on OR > 1 0.956Tocilizumab and Sarilumab are equivalent 0.446

3.5 Secondary analysis of in-hospital mortality for IL-6ra specific ITT popula on

• Model: Primary analysis ordinal model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, fixed-dura on, shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ)

24

Table 17: Odds ra o parameters for sensi vity analysis of in-hospital mortality in IL-6ra specific ITT popula on

Mean SD Median CrIAge<39 8.93 4.97 7.81 (3.25, 21.48)Age 40-49 3.30 1.06 3.13 (1.74, 5.81)Age 50-59 4.22 1.12 4.07 (2.45, 6.82)Age 70-79 0.56 0.13 0.55 (0.36, 0.84)Age 80+ 0.26 0.09 0.25 (0.12, 0.48)Female 1.35 0.27 1.32 (0.91, 1.95)Time epoch 1 0.94 0.09 0.94 (0.78, 1.12)Time epoch 2 0.86 0.15 0.86 (0.58, 1.17)Time epoch 3 0.84 0.19 0.83 (0.51, 1.25)Time epoch 4 0.85 0.22 0.82 (0.48, 1.36)Time epoch 5 0.86 0.26 0.83 (0.46, 1.47)Time epoch 6 0.87 0.27 0.83 (0.45, 1.49)Time epoch 7 0.88 0.27 0.84 (0.45, 1.51)Time epoch 8 0.92 0.27 0.89 (0.50, 1.55)Time epoch 9 0.99 0.28 0.96 (0.56, 1.65)Time epoch 10 1.06 0.30 1.02 (0.61, 1.79)Time epoch 11 1.07 0.30 1.03 (0.63, 1.76)Time epoch 12 1.00 0.27 0.97 (0.57, 1.62)Time epoch 13 0.95 0.33 0.90 (0.45, 1.72)Tocilizumab 1.72 0.31 1.69 (1.19, 2.42)Sarilumab 2.31 1.01 2.05 (1.22, 5.08)

Table 18: Posterior probabili es for sensi vity analysis of in-hospital mortality in IL-6ra specific ITT popula on

Posterior ProbabilityTocilizumab is superior to control 0.998Tocilizumab is fu le (OR < 1.2) 0.027Sarilumab is superior to control 0.997Sarilumab is fu le (OR < 1.2) 0.022Tocilizumab and Sarilumab are equivalent 0.494

4 Primary safety analysis

25

4.1 Empirical distribu on of SAE in IL-6ra specific ITT popula on

Table 19: Summary of Serious Adverse Events (SAE) for the IL-6ra specific ITT popula on

Interven on # Pa ents (N) # Known (n) # SAE (y) SAE Rate (y/n)Tocilizumab 353 353 9 0.025Sarilumab 48 48 0 0.000Control 402 402 11 0.027

4.2 Primary safety analysis in IL-6ra specific ITT popula on

• Model: Primary dichotomous model

• Factors: Age, sex, site, tocilizumab, sarilumab, and control interven ons

Table 20: Odds ra o parameters for primary safety analysis in IL-6ra specific ITT popula on

Mean SD Median CrIAge<39 9.89 12.04 6.41 (1.22, 40.63)Age 40-49 1.97 1.54 1.58 (0.49, 5.80)Age 50-59 2.53 1.59 2.13 (0.77, 6.67)Age 70-79 1.84 1.13 1.56 (0.58, 4.74)Age 80+ 1.47 1.23 1.13 (0.33, 4.72)Female 1.27 0.66 1.12 (0.46, 2.97)Tocilizumab 1.22 0.55 1.10 (0.48, 2.58)Sarilumab 2.99 2.95 2.10 (0.51, 10.77)

Table 21: Posterior probabili es for primary safety analysis in IL-6ra specific ITT popula on

Posterior ProbabilityTocilizumab is superior to control 0.593Sarilumab is superior to control 0.840

5 Analyses of secondary endpoints

5.1 Secondary analysis of mortality

• Model: Primary TTE model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, fixed-dura on, shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ),

and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

26

• Popula on: Unblinded ITT

+ ++++ ++++++++ + +++++ + + + + ++++++ + +

+ +++ +++++ ++ ++++ ++ + ++ + +

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0 15 30 45 60 75 90Days

Surv

ival p

roba

bilit

y

+ + +Tocilizumab Sarilumab Control

353 300 266 242 230 226 22448 42 37 31 31 31 31402 323 268 242 231 226 225Control

Sarilumab

Tocilizumab

0 15 30 45 60 75 90Days

Number at risk

Figure 10: Empirical distribu on of mortality for tocilizumab, sarilumab, and control. This plot is restricted to the

IL-6ra ITT popula on.

Table 22: Summary of 2.5th, 10th, 25th, 50th, 75th, 90th, and 97.5th percen les from the Kaplan-Meier es mates

for mortality (in days). Displaying the observed percen les for this outcome.

2.5 10.0 25.0 50.0 75.0 90.0 97.5Tocilizumab 4.88 10.72 33.71 - - - -Sarilumab 6.22 7.85 - - - - -Control 2.71 8.48 19.53 - - - -

27

Table 23: Hazard ra o parameters for secondary analysis of mortality

Mean SD Median CrIAge<39 5.85 2.33 5.39 (2.80, 11.80)Age 40-49 2.35 0.51 2.29 (1.52, 3.51)Age 50-59 2.28 0.38 2.25 (1.64, 3.10)Age 70-79 0.49 0.06 0.49 (0.38, 0.62)Age 80+ 0.32 0.06 0.31 (0.22, 0.45)Female 1.08 0.12 1.07 (0.86, 1.34)Time epoch 1 0.96 0.07 0.96 (0.83, 1.12)Time epoch 2 0.92 0.12 0.92 (0.70, 1.16)Time epoch 3 0.93 0.15 0.91 (0.66, 1.25)Time epoch 4 0.94 0.18 0.93 (0.65, 1.34)Time epoch 5 0.95 0.20 0.93 (0.63, 1.40)Time epoch 6 0.93 0.20 0.91 (0.61, 1.38)Time epoch 7 0.90 0.19 0.88 (0.59, 1.32)Time epoch 8 0.91 0.18 0.89 (0.61, 1.31)Time epoch 9 0.93 0.18 0.91 (0.63, 1.33)Time epoch 10 0.96 0.19 0.94 (0.66, 1.38)Time epoch 11 0.95 0.18 0.93 (0.65, 1.34)Time epoch 12 0.91 0.17 0.90 (0.62, 1.28)Time epoch 13 0.91 0.17 0.89 (0.61, 1.28)Time epoch 14 0.98 0.20 0.96 (0.65, 1.42)Time epoch 15 1.12 0.32 1.07 (0.65, 1.87)Time epoch 16 1.36 0.69 1.21 (0.61, 3.05)Tocilizumab 1.60 0.21 1.59 (1.24, 2.05)Sarilumab 1.94 0.56 1.82 (1.22, 3.38)Fixed-dose cor costeroid 0.87 0.15 0.86 (0.61, 1.18)Shock-dependent cor costeroid 1.03 0.23 1.00 (0.67, 1.55)Tocilizumab*Fixed-dose interac on 0.98 0.05 0.98 (0.89, 1.08)Tocilizumab*Shock-dependent interac on 1.01 0.05 1.01 (0.91, 1.11)Sarilumab*Fixed-dose interac on 1.01 0.05 1.00 (0.91, 1.11)Sarilumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.11)Tocilizumab*Fixed-dose combina on 1.36 0.28 1.33 (0.89, 1.98)Tocilizumab*Shock-dependent combina on 1.66 0.43 1.60 (0.98, 2.67)Sarilumab*Fixed-dose combina on 1.69 0.57 1.58 (0.93, 3.09)Sarilumab*Shock-dependent combina on 2.00 0.74 1.86 (1.00, 3.83)

28

Table 24: Posterior probabili es for secondary analysis of mortality

Posterior ProbabilityTocilizumab is superior to control 1.000Tocilizumab is fu le (HR<1.2) 0.013Sarilumab is superior to control 0.998Sarilumab is fu le (HR<1.2) 0.021Tocilizumab*Fixed-dose combina on HR > 1 0.916Tocilizumab*Shock-dependent combina on HR > 1 0.969Sarilumab*Fixed-dose combina on HR > 1 0.954Sarilumab*Shock-dependent combina on HR > 1 0.976Tocilizumab and Sarilumab are equivalent 0.567

5.2 Secondary analysis of mortality with pooled IL-6ra interven ons

• Model: Primary TTE model

• Factors: Age, sex, site, me, pooled tocilizumab and sarilumab, and control interven ons, cor costeroid

interven ons (control, fixed-dura on, shock-dependent) and an viral interven ons (control, HCQ, Kaletra,

Kaletra + HCQ), and interac ons between pooled tocilizumab and sarilumab, cor costeroids, and an virals

• Popula on: Unblinded ITT

29

+++++++ ++++++++ + +++++ + + + + ++ +++ +++++ ++ ++++ ++ + ++ + +

0.00

0.25

0.50

0.75

1.00

0 15 30 45 60 75 90Days

Surv

ival p

roba

bilit

y+ +Pooled IL−6ra Control

401 342 303 273 261 257 255

402 323 268 242 231 226 225Control

Pooled IL−6ra

0 15 30 45 60 75 90Days

Number at risk

Figure 11: Empirical distribu on of mortality for pooled IL-6ra interven ons and control. This plot is restricted to

the IL-6ra ITT popula on.

Table 25: Summary of 2.5th, 10th, 25th, 50th, 75th, 90th, and 97.5th percen les from the Kaplan-Meier es mates

for mortality (in days). Displaying the observed percen les for this outcome.

2.5 10.0 25.0 50.0 75.0 90.0 97.5Pooled IL-6ra 5.2 10.56 35.36 - - - -Control 2.71 8.48 19.53 - - - -

30

Table 26: Hazard ra o parameters for secondary analysis of mortality with pooled IL-6ra interven ons

Mean SD Median CrIAge<39 5.83 2.35 5.32 (2.77, 11.77)Age 40-49 2.34 0.50 2.29 (1.54, 3.52)Age 50-59 2.28 0.38 2.23 (1.64, 3.15)Age 70-79 0.49 0.06 0.49 (0.38, 0.61)Age 80+ 0.32 0.06 0.31 (0.21, 0.46)Female 1.08 0.12 1.07 (0.86, 1.34)Time epoch 1 0.96 0.07 0.96 (0.82, 1.11)Time epoch 2 0.92 0.11 0.91 (0.70, 1.15)Time epoch 3 0.92 0.15 0.90 (0.66, 1.23)Time epoch 4 0.93 0.18 0.91 (0.64, 1.32)Time epoch 5 0.94 0.19 0.92 (0.63, 1.39)Time epoch 6 0.92 0.19 0.90 (0.60, 1.37)Time epoch 7 0.89 0.18 0.87 (0.58, 1.30)Time epoch 8 0.89 0.18 0.88 (0.59, 1.28)Time epoch 9 0.92 0.18 0.90 (0.63, 1.32)Time epoch 10 0.95 0.19 0.93 (0.65, 1.38)Time epoch 11 0.94 0.18 0.92 (0.65, 1.33)Time epoch 12 0.90 0.16 0.89 (0.62, 1.27)Time epoch 13 0.90 0.17 0.88 (0.61, 1.27)Time epoch 14 0.97 0.20 0.95 (0.65, 1.40)Time epoch 15 1.11 0.31 1.06 (0.65, 1.86)Time epoch 16 1.36 0.74 1.20 (0.61, 3.11)Pooled IL-6ra 1.62 0.21 1.61 (1.25, 2.08)Fixed-dose cor costeroid 0.87 0.15 0.86 (0.61, 1.20)Shock-dependent cor costeroid 1.04 0.23 1.01 (0.67, 1.54)Pooled IL-6ra*Fixed-dose interac on 0.99 0.05 0.98 (0.90, 1.08)Pooled IL-6ra*Shock-dependent interac on 1.01 0.05 1.01 (0.91, 1.11)Pooled IL-6ra*Fixed-dose combina on 1.39 0.29 1.36 (0.91, 2.04)Pooled IL-6ra*Shock-dependent combina on 1.69 0.43 1.64 (1.00, 2.70)

Table 27: Posterior probabili es for secondary analysis of mortality with pooled IL-6ra interven ons

Posterior ProbabilityPooled IL-6ra is superior to control 1.000Pooled IL-6ra is fu le (HR<1.2) 0.011Pooled IL-6ra*Fixed-dose combina on HR > 1 0.929Pooled IL-6ra*Shock-dependent combina on HR > 1 0.974

5.3 Secondary analysis of progression to intuba on, ECMO, or death

• Model: Primary dichotomous model

31

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, fixed-dura on, shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ),

and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

• Popula on: Unblinded ITT not on ven la on or ECMO at baseline. There are 552 pa ents that meet this

criterion.

Table 28: Summary of progression to intuba on, ECMO, or death for the Unblinded ITT popula on restricted to

pa ents not on mechanical ven la on or ECMO at baseline

Interven on # Pa ents (N) # Progressors (y) Progression Rate (y/N)Tocilizumab 242 100 0.413Sarilumab 37 13 0.351Control 273 144 0.527

Table 29: Summary of progression to intuba on, ECMO, or death by component for the Unblinded ITT popula on

restricted to pa ents not onmechanical ven la on or ECMO at baseline. Note that a pa ent may progress onmore

than one component of the composite endpoint, so the sum of events in this table will not match the total number

of progressors.

Interven on # Pa ents (N) Death, n (%) Intuba on, n (%) ECMO, n (%)Tocilizumab 242 53 (21.9) 84 (34.7) 5 (2.1)Sarilumab 37 8 (21.6) 6 (16.2) 0 (0)Control 273 82 (30) 116 (42.5) 3 (1.1)

32

Table 30: Odds ra o parameters for secondary analysis of progression to intuba on, ECMO, or death

Mean SD Median CrIAge<39 5.38 2.13 4.96 (2.45, 10.63)Age 40-49 1.88 0.46 1.83 (1.14, 2.92)Age 50-59 1.70 0.35 1.66 (1.12, 2.50)Age 70-79 0.79 0.16 0.78 (0.52, 1.16)Age 80+ 0.51 0.16 0.48 (0.26, 0.88)Female 1.10 0.19 1.08 (0.78, 1.52)Time epoch 1 0.93 0.08 0.93 (0.78, 1.10)Time epoch 2 0.84 0.13 0.84 (0.59, 1.11)Time epoch 3 0.80 0.16 0.79 (0.50, 1.14)Time epoch 4 0.78 0.19 0.76 (0.46, 1.19)Time epoch 5 0.77 0.21 0.75 (0.44, 1.24)Time epoch 6 0.77 0.21 0.74 (0.43, 1.26)Time epoch 7 0.76 0.21 0.74 (0.43, 1.26)Time epoch 8 0.78 0.21 0.75 (0.45, 1.27)Time epoch 9 0.80 0.22 0.77 (0.45, 1.30)Time epoch 10 0.82 0.24 0.79 (0.45, 1.37)Time epoch 11 0.83 0.25 0.79 (0.45, 1.43)Time epoch 12 0.85 0.27 0.81 (0.44, 1.51)Time epoch 13 0.89 0.31 0.84 (0.44, 1.63)Time epoch 14 0.99 0.40 0.92 (0.43, 1.97)Time epoch 15 1.16 0.62 1.02 (0.41, 2.68)Tocilizumab 1.72 0.33 1.69 (1.17, 2.42)Sarilumab 1.82 0.55 1.74 (1.01, 3.14)Fixed-dose cor costeroid 2.92 1.18 2.71 (1.24, 5.84)Shock-dependent cor costeroid 1.34 0.54 1.24 (0.59, 2.67)Tocilizumab*Fixed-dose interac on 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Fixed-dose interac on 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose combina on 5.04 2.34 4.57 (1.92, 10.87)Tocilizumab*Shock-dependent combina on 2.30 1.03 2.11 (0.92, 4.91)Sarilumab*Fixed-dose combina on 5.34 2.81 4.72 (1.82, 12.54)Sarilumab*Shock-dependent combina on 2.43 1.25 2.18 (0.86, 5.58)

33

Table 31: Posterior probabili es for secondary analysis of progression to intuba on, ECMO, or death

Posterior ProbabilityTocilizumab is superior to control 0.998Tocilizumab is fu le (OR < 1.2) 0.035Sarilumab is superior to control 0.977Sarilumab is fu le (OR < 1.2) 0.078Tocilizumab*Fixed-dose combina on OR > 1 1.000Tocilizumab*Shock-dependent combina on OR > 1 0.960Sarilumab*Fixed-dose combina on OR > 1 1.000Sarilumab*Shock-dependent combina on OR > 1 0.951Tocilizumab and Sarilumab are equivalent 0.658

5.4 Secondary analysis of days free of vasopressors/inotropes

• Model: Primary analysis ordinal model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, fixed-dura on, shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ),

and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

• Popula on: Unblinded ITT

• There are 25 missing values of days free of vasopressors/inotropes in the Unblinded ITT popula on due to

missing daily data. These pa ents are removed from this analysis.

Control

Sarilumab

Tocilizumab

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0Proportion

Days Free ofVasopressors/Inotropes −1 0 1 2 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 12: Empirical distribu on of days-free of vasopressors and inotropes for tocilizumab, sarilumab, and control.

This plot is restricted to the IL-6ra ITT popula on.

34

Table 32: Summary of Days Free of Vasopressors/Inotropes for the IL-6ra specific ITT popula on

Interven on # Pa ents # Known Days Free

Vasopressors/Inotropes

median (IQR)

Days Free of

Vasopressors/Inotropes

in Survivors* median

(IQR)

Tocilizumab 353 348 19 (-1, 21) 21 (17, 21)

Sarilumab 48 48 20.5 (1.5, 21) 21 (17.5, 21)

Control 402 391 15 (-1, 21) 21 (16, 21)

* Days Free of vasopressors/inotropes in survivors within 21 days

35

Table 33: Odds ra o parameters for secondary analysis of days free of vasopressors/inotropes

Mean SD Median CrIAge<39 3.96 0.95 3.85 (2.44, 6.15)Age 40-49 2.49 0.45 2.45 (1.73, 3.50)Age 50-59 2.06 0.29 2.05 (1.56, 2.70)Age 70-79 0.50 0.08 0.50 (0.37, 0.68)Age 80+ 0.30 0.08 0.29 (0.17, 0.49)Female 1.18 0.14 1.18 (0.93, 1.48)Time epoch 1 0.95 0.08 0.95 (0.81, 1.10)Time epoch 2 0.91 0.13 0.90 (0.66, 1.16)Time epoch 3 0.94 0.16 0.92 (0.65, 1.29)Time epoch 4 1.00 0.21 0.97 (0.66, 1.47)Time epoch 5 1.04 0.25 1.01 (0.67, 1.62)Time epoch 6 1.05 0.25 1.01 (0.66, 1.62)Time epoch 7 1.04 0.23 1.01 (0.65, 1.57)Time epoch 8 1.03 0.22 1.01 (0.66, 1.55)Time epoch 9 1.01 0.22 0.99 (0.65, 1.50)Time epoch 10 0.97 0.21 0.94 (0.63, 1.43)Time epoch 11 0.88 0.18 0.86 (0.57, 1.29)Time epoch 12 0.80 0.17 0.79 (0.52, 1.19)Time epoch 13 0.77 0.17 0.75 (0.49, 1.15)Time epoch 14 0.79 0.18 0.76 (0.49, 1.20)Time epoch 15 0.85 0.24 0.81 (0.48, 1.40)Time epoch 16 0.96 0.38 0.89 (0.45, 1.90)Tocilizumab 1.70 0.26 1.68 (1.25, 2.24)Sarilumab 1.95 0.53 1.85 (1.20, 3.30)Fixed-dose cor costeroid 1.44 0.33 1.40 (0.90, 2.20)Shock-dependent cor costeroid 1.13 0.27 1.10 (0.70, 1.74)Tocilizumab*Fixed-dose interac on 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Shock-dependent interac on 1.00 0.05 1.00 (0.90, 1.10)Sarilumab*Fixed-dose interac on 1.01 0.05 1.00 (0.91, 1.11)Sarilumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose combina on 2.45 0.70 2.35 (1.37, 4.07)Tocilizumab*Shock-dependent combina on 1.91 0.55 1.84 (1.06, 3.22)Sarilumab*Fixed-dose combina on 2.82 1.03 2.62 (1.39, 5.31)Sarilumab*Shock-dependent combina on 2.20 0.83 2.05 (1.06, 4.21)

36

Table 34: Posterior probabili es for secondary analysis of days free of vasopressors/inotropes

Posterior ProbabilityTocilizumab is superior to control 1.000Tocilizumab is fu le (OR < 1.2) 0.013Sarilumab is superior to control 0.995Sarilumab is fu le (OR < 1.2) 0.024Tocilizumab*Fixed-dose combina on OR > 1 0.999Tocilizumab*Shock-dependent combina on OR > 1 0.986Sarilumab*Fixed-dose combina on OR > 1 0.998Sarilumab*Shock-dependent combina on OR > 1 0.984Tocilizumab and Sarilumab are equivalent 0.634

5.5 Secondary analysis of days free of ven la on

• Model: Primary analysis ordinal model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, fixed-dura on, shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ),

and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

• Popula on: Unblinded ITT

• There are 25 missing values of days free of ven la on in the Unblinded ITT popula on due to missing daily

data. These pa ents are removed from this analysis.

Control

Sarilumab

Tocilizumab

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0Proportion

Days Free ofVentilation −1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 13: Empirical distribu on of days-free of ven la on for tocilizumab, sarilumab, and control. This plot is

restricted to the IL-6ra ITT popula on.

37

Table 35: Summary of Days Free of Ven la on for the IL-6ra specific ITT popula on

Interven on # Pa ents # Known Days Free of Ven la on

median (IQR)

Days Free of Ven la on

in Survivors* median

(IQR)

Tocilizumab 353 348 9.5 (-1, 16) 14 (6, 17)

Sarilumab 48 48 11.5 (0, 16) 15 (4.5, 17)

Control 402 391 0 (-1, 14) 12 (2, 17)

* Days Free of ven la on in survivors within 21 days

38

Table 36: Odds ra o parameters for secondary analysis of days free of ven la on

Mean SD Median CrIAge<39 4.16 0.90 4.07 (2.68, 6.18)Age 40-49 2.20 0.38 2.17 (1.55, 3.04)Age 50-59 1.96 0.27 1.94 (1.49, 2.54)Age 70-79 0.52 0.08 0.51 (0.38, 0.69)Age 80+ 0.35 0.10 0.34 (0.19, 0.57)Female 1.20 0.14 1.19 (0.96, 1.49)Time epoch 1 0.97 0.08 0.97 (0.81, 1.13)Time epoch 2 0.93 0.13 0.93 (0.68, 1.21)Time epoch 3 0.99 0.17 0.98 (0.68, 1.37)Time epoch 4 1.10 0.22 1.08 (0.73, 1.61)Time epoch 5 1.21 0.28 1.17 (0.78, 1.87)Time epoch 6 1.27 0.30 1.23 (0.80, 1.98)Time epoch 7 1.32 0.31 1.28 (0.82, 2.04)Time epoch 8 1.36 0.32 1.31 (0.86, 2.13)Time epoch 9 1.29 0.29 1.25 (0.83, 1.97)Time epoch 10 1.17 0.25 1.14 (0.77, 1.75)Time epoch 11 1.01 0.21 0.99 (0.66, 1.48)Time epoch 12 0.90 0.20 0.88 (0.58, 1.34)Time epoch 13 0.88 0.19 0.86 (0.56, 1.32)Time epoch 14 0.96 0.22 0.93 (0.60, 1.45)Time epoch 15 1.09 0.31 1.05 (0.62, 1.81)Time epoch 16 1.33 0.58 1.22 (0.59, 2.80)Tocilizumab 1.74 0.25 1.73 (1.31, 2.27)Sarilumab 2.04 0.53 1.94 (1.27, 3.32)Fixed-dose cor costeroid 1.47 0.33 1.44 (0.93, 2.21)Shock-dependent cor costeroid 1.22 0.28 1.19 (0.76, 1.87)Tocilizumab*Fixed-dose interac on 1.00 0.05 1.00 (0.90, 1.10)Tocilizumab*Shock-dependent interac on 1.00 0.05 1.00 (0.90, 1.10)Sarilumab*Fixed-dose interac on 1.01 0.05 1.01 (0.91, 1.11)Sarilumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose combina on 2.55 0.69 2.47 (1.49, 4.15)Tocilizumab*Shock-dependent combina on 2.13 0.59 2.05 (1.21, 3.51)Sarilumab*Fixed-dose combina on 3.01 1.04 2.83 (1.52, 5.49)Sarilumab*Shock-dependent combina on 2.49 0.88 2.33 (1.25, 4.70)

39

Table 37: Posterior probabili es for secondary analysis of days free of ven la on

Posterior ProbabilityTocilizumab is superior to control 1.000Tocilizumab is fu le (OR < 1.2) 0.005Sarilumab is superior to control 0.999Sarilumab is fu le (OR < 1.2) 0.013Tocilizumab*Fixed-dose combina on OR > 1 1.000Tocilizumab*Shock-dependent combina on OR > 1 0.996Sarilumab*Fixed-dose combina on OR > 1 1.000Sarilumab*Shock-dependent combina on OR > 1 0.995Tocilizumab and Sarilumab are equivalent 0.601

5.6 Secondary analysis of length of ICU stay

• Model: Primary TTE model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, fixed-dura on, shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ),

and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

• Popula on: Unblinded ITT

40

++ ++++ ++++++ + ++++ + + + +++++ ++ +

+++ + + +++++ +++ +++ ++ + ++ + +

1

0.75

0.5

0.25

0

0 15 30 45 60 75 90Days

ICU

dis

char

ge p

roba

bilit

y+ + +Tocilizumab Sarilumab Control

353 162 125 99 91 90 8948 18 14 7 7 7 7402 236 184 157 140 134 132Control

Sarilumab

Tocilizumab

0 15 30 45 60 75 90Days

Number at risk

Figure 14: Empirical distribu on of length of ICU stay for tocilizumab, sarilumab, and control. This plot is restricted

to the IL-6ra ITT popula on.

Table 38: Summary of 2.5th, 10th, 25th, 50th, 75th, 90th, and 97.5th percen les from the Kaplan-Meier es mates

for length of ICU stay (in days). Displaying the observed percen les for this outcome.

2.5 10.0 25.0 50.0 75.0 90.0 97.5Tocilizumab 2.55 3.69 6.53 13.31 - - -Sarilumab 1.8 3.58 5.75 11.18 - - -Control 2.64 4.58 8.52 23.79 - - -

41

Table 39: Hazard ra o parameters for secondary analysis of length of ICU stay

Mean SD Median CrIAge<39 2.54 0.35 2.52 (1.91, 3.30)Age 40-49 1.70 0.19 1.69 (1.36, 2.11)Age 50-59 1.62 0.15 1.61 (1.35, 1.93)Age 70-79 0.68 0.07 0.67 (0.54, 0.83)Age 80+ 0.71 0.13 0.70 (0.48, 0.98)Female 1.18 0.09 1.18 (1.01, 1.36)Time epoch 1 1.01 0.07 1.01 (0.88, 1.14)Time epoch 2 1.03 0.10 1.02 (0.83, 1.24)Time epoch 3 1.10 0.14 1.09 (0.85, 1.38)Time epoch 4 1.18 0.18 1.17 (0.88, 1.58)Time epoch 5 1.22 0.20 1.20 (0.89, 1.67)Time epoch 6 1.19 0.19 1.18 (0.88, 1.60)Time epoch 7 1.16 0.17 1.15 (0.87, 1.53)Time epoch 8 1.16 0.16 1.15 (0.87, 1.50)Time epoch 9 1.15 0.16 1.14 (0.87, 1.49)Time epoch 10 1.13 0.16 1.12 (0.86, 1.46)Time epoch 11 1.08 0.15 1.07 (0.82, 1.39)Time epoch 12 1.04 0.14 1.03 (0.79, 1.33)Time epoch 13 1.05 0.14 1.04 (0.79, 1.35)Time epoch 14 1.14 0.16 1.13 (0.85, 1.48)Time epoch 15 1.28 0.23 1.26 (0.89, 1.77)Time epoch 16 1.51 0.42 1.45 (0.89, 2.51)Tocilizumab 1.43 0.13 1.42 (1.18, 1.70)Sarilumab 1.69 0.32 1.64 (1.21, 2.45)Fixed-dose cor costeroid 0.92 0.12 0.91 (0.71, 1.17)Shock-dependent cor costeroid 0.96 0.13 0.95 (0.73, 1.23)Tocilizumab*Fixed-dose interac on 1.00 0.05 0.99 (0.90, 1.09)Tocilizumab*Shock-dependent interac on 1.00 0.05 1.00 (0.90, 1.10)Sarilumab*Fixed-dose interac on 1.01 0.05 1.01 (0.91, 1.11)Sarilumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose combina on 1.30 0.20 1.29 (0.96, 1.73)Tocilizumab*Shock-dependent combina on 1.37 0.23 1.35 (0.98, 1.87)Sarilumab*Fixed-dose combina on 1.56 0.34 1.52 (1.01, 2.35)Sarilumab*Shock-dependent combina on 1.62 0.39 1.56 (1.02, 2.52)

42

Table 40: Posterior probabili es for secondary analysis of length of ICU stay

Posterior ProbabilityTocilizumab is superior to control 1.000Tocilizumab is fu le (HR<1.2) 0.035Sarilumab is superior to control 0.999Sarilumab is fu le (HR<1.2) 0.023Tocilizumab*Fixed-dose combina on HR > 1 0.949Tocilizumab*Shock-dependent combina on HR > 1 0.966Sarilumab*Fixed-dose combina on HR > 1 0.978Sarilumab*Shock-dependent combina on HR > 1 0.980Tocilizumab and Sarilumab are equivalent 0.598

5.7 Secondary analysis of length of hospital stay

• Model: Primary TTE model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, fixed-dura on, shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ),

and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

• Popula on: Unblinded ITT

43

+ ++++ ++++++++ + +++++ + + + + ++++++ + +

++++ +++++ ++ ++++ ++ + ++ + +

1

0.75

0.5

0.25

0

0 15 30 45 60 75 90Days

Hos

pita

l dis

char

ge p

roba

bilit

y+ + +Tocilizumab Sarilumab Control

353 234 163 118 106 103 10148 26 19 10 10 10 10402 297 218 182 159 148 145Control

Sarilumab

Tocilizumab

0 15 30 45 60 75 90Days

Number at risk

Figure 15: Empirical distribu on of length of hospital stay for tocilizumab, sarilumab, and control. This plot is re-

stricted to the IL-6ra ITT popula on.

Table 41: Summary of 2.5th, 10th, 25th, 50th, 75th, 90th, and 97.5th percen les from the Kaplan-Meier es mates

for length of hospital stay (in days). Displaying the observed percen les for this outcome.

2.5 10.0 25.0 50.0 75.0 90.0 97.5Tocilizumab 5 8 12 27 - - -Sarilumab 5 7 10 19 - - -Control 5 8 13 44 - - -

44

Table 42: Hazard ra o parameters for secondary analysis of length of hospital stay

Mean SD Median CrIAge<39 2.84 0.40 2.82 (2.13, 3.69)Age 40-49 1.82 0.21 1.81 (1.44, 2.25)Age 50-59 1.73 0.17 1.72 (1.42, 2.08)Age 70-79 0.62 0.07 0.62 (0.49, 0.77)Age 80+ 0.46 0.10 0.45 (0.29, 0.66)Female 1.13 0.09 1.12 (0.96, 1.31)Time epoch 1 1.00 0.07 1.00 (0.87, 1.13)Time epoch 2 1.00 0.11 1.00 (0.80, 1.22)Time epoch 3 1.06 0.14 1.05 (0.81, 1.35)Time epoch 4 1.14 0.17 1.12 (0.84, 1.52)Time epoch 5 1.18 0.19 1.16 (0.85, 1.61)Time epoch 6 1.19 0.20 1.17 (0.86, 1.63)Time epoch 7 1.20 0.19 1.18 (0.87, 1.60)Time epoch 8 1.22 0.18 1.20 (0.90, 1.61)Time epoch 9 1.23 0.18 1.21 (0.91, 1.62)Time epoch 10 1.21 0.18 1.20 (0.90, 1.60)Time epoch 11 1.14 0.16 1.13 (0.86, 1.50)Time epoch 12 1.08 0.15 1.07 (0.81, 1.40)Time epoch 13 1.06 0.15 1.06 (0.80, 1.38)Time epoch 14 1.13 0.17 1.12 (0.83, 1.48)Time epoch 15 1.27 0.24 1.25 (0.87, 1.80)Time epoch 16 1.53 0.47 1.46 (0.86, 2.67)Tocilizumab 1.42 0.13 1.41 (1.18, 1.70)Sarilumab 1.65 0.31 1.60 (1.17, 2.40)Fixed-dose cor costeroid 0.92 0.12 0.91 (0.70, 1.17)Shock-dependent cor costeroid 0.94 0.13 0.93 (0.71, 1.21)Tocilizumab*Fixed-dose interac on 0.99 0.05 0.99 (0.90, 1.09)Tocilizumab*Shock-dependent interac on 1.01 0.05 1.01 (0.91, 1.11)Sarilumab*Fixed-dose interac on 1.01 0.05 1.01 (0.92, 1.11)Sarilumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose combina on 1.29 0.20 1.27 (0.93, 1.73)Tocilizumab*Shock-dependent combina on 1.34 0.23 1.32 (0.94, 1.86)Sarilumab*Fixed-dose combina on 1.52 0.34 1.47 (0.98, 2.32)Sarilumab*Shock-dependent combina on 1.55 0.37 1.50 (0.97, 2.45)

45

Table 43: Posterior probabili es for secondary analysis of length of hospital stay

Posterior ProbabilityTocilizumab is superior to control 1.000Tocilizumab is fu le (HR<1.2) 0.039Sarilumab is superior to control 0.998Sarilumab is fu le (HR<1.2) 0.036Tocilizumab*Fixed-dose combina on HR > 1 0.940Tocilizumab*Shock-dependent combina on HR > 1 0.949Sarilumab*Fixed-dose combina on HR > 1 0.969Sarilumab*Shock-dependent combina on HR > 1 0.965Tocilizumab and Sarilumab are equivalent 0.628

5.8 Secondary analysis of WHO scale at 14 days

• Model: Primary analysis ordinal model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, fixed-dura on, shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ),

and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

• Popula on: Unblinded ITT

• There are 19 missing values of WHO scale at day 14 due to missing daily data. These pa ents are removed

from this analysis.

Control

Sarilumab

Tocilizumab

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0Proportion

WHO scale atday 14 8 7 6 5 4 3 0+1+2

Figure 16: Empirical distribu on of modified WHO scale at day 14 for tocilizumab, sarilumab, and control. This plot

is restricted to the IL-6ra ITT popula on.

46

Table 44: Odds ra o parameters for secondary analysis of WHO scale

Mean SD Median CrIAge<39 4.20 1.00 4.08 (2.62, 6.53)Age 40-49 2.02 0.35 1.99 (1.43, 2.78)Age 50-59 1.76 0.24 1.74 (1.34, 2.29)Age 70-79 0.59 0.08 0.58 (0.44, 0.78)Age 80+ 0.33 0.08 0.32 (0.19, 0.53)Female 1.06 0.12 1.05 (0.84, 1.32)Time epoch 1 1.04 0.09 1.04 (0.88, 1.22)Time epoch 2 1.11 0.15 1.10 (0.85, 1.43)Time epoch 3 1.20 0.21 1.18 (0.85, 1.68)Time epoch 4 1.34 0.30 1.29 (0.87, 2.02)Time epoch 5 1.41 0.36 1.36 (0.87, 2.25)Time epoch 6 1.35 0.33 1.31 (0.84, 2.13)Time epoch 7 1.27 0.30 1.22 (0.80, 1.97)Time epoch 8 1.18 0.27 1.14 (0.76, 1.78)Time epoch 9 1.06 0.22 1.03 (0.70, 1.57)Time epoch 10 0.95 0.19 0.93 (0.62, 1.39)Time epoch 11 0.85 0.17 0.83 (0.55, 1.23)Time epoch 12 0.80 0.17 0.78 (0.52, 1.18)Time epoch 13 0.82 0.18 0.80 (0.52, 1.21)Time epoch 14 0.94 0.21 0.92 (0.60, 1.42)Time epoch 15 1.18 0.35 1.13 (0.66, 2.01)Time epoch 16 1.61 0.75 1.44 (0.67, 3.57)Tocilizumab 1.85 0.26 1.83 (1.40, 2.41)Sarilumab 1.91 0.43 1.86 (1.22, 2.91)Fixed-dose cor costeroid 1.35 0.31 1.32 (0.84, 2.04)Shock-dependent cor costeroid 0.99 0.23 0.96 (0.62, 1.52)Tocilizumab*Fixed-dose interac on 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Fixed-dose interac on 1.00 0.05 1.00 (0.91, 1.11)Sarilumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose combina on 2.50 0.68 2.41 (1.44, 4.08)Tocilizumab*Shock-dependent combina on 1.83 0.51 1.76 (1.04, 3.03)Sarilumab*Fixed-dose combina on 2.59 0.86 2.47 (1.33, 4.66)Sarilumab*Shock-dependent combina on 1.89 0.63 1.79 (0.96, 3.39)

47

Table 45: Posterior probabili es for secondary analysis of WHO scale

Posterior ProbabilityTocilizumab is superior to control 1.000Tocilizumab is fu le (OR < 1.2) 0.001Sarilumab is superior to control 0.996Sarilumab is fu le (OR < 1.2) 0.022Tocilizumab*Fixed-dose combina on OR > 1 1.000Tocilizumab*Shock-dependent combina on OR > 1 0.983Sarilumab*Fixed-dose combina on OR > 1 0.998Sarilumab*Shock-dependent combina on OR > 1 0.968Tocilizumab and Sarilumab are equivalent 0.727

6 Sensi vity Analyses

6.1 Sensi vity analyses of OSFD

6.1.1 Sensi vity analysis of OSFD endpoint for IL-6ra specific per protocol popula on

• Model: Primary analysis ordinal model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons

48

Table 46: Odds ra o parameters for sensi vity analysis of OSFD in IL-6ra specific PP popula on

Mean SD Median CrIAge<39 4.02 1.17 3.86 (2.21, 6.76)Age 40-49 2.08 0.46 2.03 (1.32, 3.12)Age 50-59 2.50 0.45 2.46 (1.73, 3.48)Age 70-79 0.58 0.11 0.57 (0.39, 0.83)Age 80+ 0.30 0.10 0.28 (0.14, 0.54)Female 1.25 0.19 1.23 (0.92, 1.65)Time epoch 1 0.98 0.08 0.98 (0.83, 1.14)Time epoch 2 0.96 0.13 0.95 (0.73, 1.23)Time epoch 3 0.97 0.17 0.96 (0.67, 1.34)Time epoch 4 0.99 0.21 0.97 (0.64, 1.47)Time epoch 5 1.02 0.24 0.98 (0.63, 1.57)Time epoch 6 1.01 0.25 0.98 (0.62, 1.61)Time epoch 7 1.00 0.24 0.97 (0.61, 1.56)Time epoch 8 0.98 0.23 0.95 (0.62, 1.51)Time epoch 9 0.94 0.21 0.92 (0.61, 1.43)Time epoch 10 0.89 0.18 0.87 (0.59, 1.32)Time epoch 11 0.82 0.16 0.81 (0.56, 1.17)Time epoch 12 0.75 0.15 0.73 (0.50, 1.08)Time epoch 13 0.69 0.18 0.67 (0.41, 1.10)Tocilizumab 1.68 0.23 1.66 (1.28, 2.19)Sarilumab 1.87 0.47 1.79 (1.17, 3.05)

Table 47: Posterior probabili es for sensi vity analysis of OSFD in IL-6ra specific PP popula on

Posterior ProbabilityTocilizumab is superior to control 1.000Tocilizumab is fu le (OR < 1.2) 0.008Sarilumab is superior to control 0.997Sarilumab is fu le (OR < 1.2) 0.030Tocilizumab and Sarilumab are equivalent 0.655

6.1.2 Sensi vity analysis of OSFD endpoint for Unblinded ITT popula on with site and me factors removed

• Model: Primary analysis ordinal model

• Factors: Age, sex, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons (control,

fixed-dura on, Shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ), and in-

terac ons between tocilizumab, sarilumab, cor costeroids, and an virals

49

Table 48: Odds ra o parameters for sensi vity analysis of OSFD with site and me factors removed

Mean SD Median CrIAge<39 3.59 0.76 3.51 (2.34, 5.32)Age 40-49 2.04 0.34 2.01 (1.45, 2.77)Age 50-59 1.87 0.25 1.85 (1.43, 2.41)Age 70-79 0.57 0.08 0.56 (0.42, 0.75)Age 80+ 0.40 0.11 0.39 (0.23, 0.64)Female 1.14 0.12 1.13 (0.92, 1.40)Tocilizumab 1.58 0.21 1.56 (1.20, 2.04)Sarilumab 1.78 0.41 1.70 (1.16, 2.77)Fixed-dose cor costeroids 1.61 0.31 1.58 (1.09, 2.30)Shock-dependent steroids 1.29 0.29 1.26 (0.82, 1.95)Tocilizumab*Fixed-dose interac on 1.00 0.05 1.00 (0.90, 1.09)Tocilizumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Fixed-dose interac on 1.01 0.05 1.00 (0.91, 1.11)Sarilumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose combina on 2.53 0.61 2.46 (1.54, 3.91)Tocilizumab*Shock-dependent combina on 2.04 0.55 1.98 (1.17, 3.29)Sarilumab*Fixed-dose combina on 2.87 0.86 2.74 (1.58, 4.94)Sarilumab*Shock-dependent combina on 2.30 0.75 2.19 (1.19, 4.08)

Table 49: Posterior probabili es for sensi vity analysis of OSFD with site and me factors removed

Posterior ProbabilityTocilizumab is superior to control 0.999Tocilizumab is fu le (OR < 1.2) 0.025Sarilumab is superior to control 0.997Sarilumab is fu le (OR < 1.2) 0.036Tocilizumab*Fixed-dose combina on OR > 1 1.000Tocilizumab*Shock-dependent combina on OR > 1 0.995Sarilumab*Fixed-dose combina on OR > 1 1.000Sarilumab*Shock-dependent combina on OR > 1 0.995Tocilizumab and Sarilumab are equivalent 0.663

6.1.3 Sensi vity analysis of OSFD endpoint for Unblinded ITT popula onwith alterna ve coding of cor costeroid

interven ons

• Model: Primary analysis ordinal model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, pooled fixed-dura on and Shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kale-

tra + HCQ), and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

50

• Pa ents randomized a er the closure of the cor costeroid domain (June 17, 2020) will be coded as receiving

steroids if they received steroids prior to randomiza on or within the first two study days. There are 803

pa ents randomized a er the closure of the cor costeroid domain. Of those pa ents, 55 pa ents did not

receive steroids prior to randomiza on or within the first two study days. In the model, the pa ents that

received steroids post-closure of the cor costeroid domain will be pooled with the pa ents randomized to

a cor costeroid interven on. The pa ents that did not receive steroids post-closure of the cor costeroid

domain will be pooled with pa ents randomized to no cor costeroid interven on.

Table 50: Odds ra o parameters for sensi vity analysis of OSFD endpoint for Unblinded ITT popula on with alter-

na ve coding of cor costeroid interven ons

Mean SD Median CrIAge<39 3.94 0.87 3.84 (2.52, 5.92)Age 40-49 2.02 0.34 2.00 (1.43, 2.75)Age 50-59 1.90 0.26 1.89 (1.43, 2.45)Age 70-79 0.50 0.08 0.50 (0.37, 0.67)Age 80+ 0.32 0.09 0.31 (0.17, 0.53)Female 1.16 0.13 1.15 (0.92, 1.43)Time epoch 1 0.94 0.08 0.94 (0.78, 1.10)Time epoch 2 0.88 0.13 0.88 (0.63, 1.15)Time epoch 3 0.93 0.17 0.92 (0.63, 1.29)Time epoch 4 1.04 0.22 1.01 (0.68, 1.55)Time epoch 5 1.15 0.28 1.11 (0.72, 1.79)Time epoch 6 1.20 0.30 1.16 (0.74, 1.89)Time epoch 7 1.23 0.30 1.19 (0.76, 1.90)Time epoch 8 1.24 0.29 1.20 (0.78, 1.91)Time epoch 9 1.15 0.25 1.12 (0.74, 1.73)Time epoch 10 1.01 0.20 0.99 (0.68, 1.47)Time epoch 11 0.84 0.16 0.82 (0.57, 1.17)Time epoch 12 0.73 0.14 0.71 (0.49, 1.03)Time epoch 13 0.71 0.14 0.70 (0.48, 1.02)Time epoch 14 0.79 0.16 0.78 (0.52, 1.15)Time epoch 15 0.94 0.25 0.91 (0.54, 1.53)Time epoch 16 1.20 0.52 1.09 (0.52, 2.50)Tocilizumab 1.64 0.23 1.62 (1.24, 2.13)Sarilumab 1.79 0.43 1.72 (1.14, 2.85)Steroids 1.32 0.22 1.30 (0.93, 1.81)Tocilizumab*cor costeroid interac on 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*cor costeroid interac on 1.01 0.05 1.01 (0.91, 1.11)Tocilizumab*cor costeroid combina on 2.15 0.47 2.09 (1.38, 3.24)Sarilumab*cor costeroid combina on 2.36 0.70 2.25 (1.32, 4.09)

51

Table 51: Posterior probabili es for sensi vity analysis of OSFD endpoint for Unblinded ITT popula on with alterna-

ve coding of cor costeroid interven ons

Posterior ProbabilityTocilizumab is superior to control 1.000Tocilizumab is fu le (OR < 1.2) 0.014Sarilumab is superior to control 0.993Sarilumab is fu le (OR < 1.2) 0.040Tocilizumab*steroids OR > 1 1.000Sarilumab*steroids OR > 1 0.997Tocilizumab and Sarilumab are equivalent 0.687

6.2 Sensi vity analyses of in-hospital mortality

6.2.1 Sensi vity analysis of in-hospital mortality for IL-6ra specific per protocol popula on

• Model: Primary dichotomous model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons

Table 52: Odds ra o parameters for sensi vity analysis of in-hospital mortality in IL-6ra specific PP popula on

Mean SD Median CrIAge<39 9.67 6.01 8.26 (3.15, 24.80)Age 40-49 3.16 1.06 2.97 (1.61, 5.71)Age 50-59 4.20 1.13 4.04 (2.42, 6.87)Age 70-79 0.55 0.13 0.54 (0.34, 0.84)Age 80+ 0.24 0.09 0.23 (0.11, 0.44)Female 1.41 0.30 1.37 (0.92, 2.08)Time epoch 1 0.94 0.09 0.94 (0.77, 1.14)Time epoch 2 0.84 0.15 0.83 (0.56, 1.15)Time epoch 3 0.79 0.19 0.78 (0.46, 1.20)Time epoch 4 0.76 0.22 0.74 (0.39, 1.25)Time epoch 5 0.75 0.25 0.72 (0.36, 1.30)Time epoch 6 0.74 0.26 0.71 (0.32, 1.33)Time epoch 7 0.76 0.27 0.73 (0.33, 1.36)Time epoch 8 0.82 0.27 0.78 (0.40, 1.43)Time epoch 9 0.93 0.29 0.88 (0.49, 1.62)Time epoch 10 1.06 0.36 1.00 (0.57, 1.98)Time epoch 11 1.11 0.38 1.04 (0.61, 2.02)Time epoch 12 1.02 0.30 0.97 (0.57, 1.72)Time epoch 13 0.93 0.34 0.88 (0.42, 1.75)Tocilizumab 1.75 0.33 1.72 (1.19, 2.49)Sarilumab 2.14 0.83 1.96 (1.13, 4.36)

52

Table 53: Posterior probabili es for sensi vity analysis of in-hospital mortality in IL-6ra specific PP popula on

Posterior ProbabilityTocilizumab is superior to control 0.999Tocilizumab is fu le (OR < 1.2) 0.027Sarilumab is superior to control 0.991Sarilumab is fu le (OR < 1.2) 0.038Tocilizumab and Sarilumab are equivalent 0.563

6.2.2 Sensi vity analysis of in-hospitalmortality for Unblinded ITT popula onwith site and me factors removed

• Model: Primary dichotomous model

• Factors: Age, sex, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons (control,

fixed-dura on, Shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ), and in-

terac ons between tocilizumab, sarilumab, cor costeroids, and an virals

Table 54: Odds ra o parameters for sensi vity analysis of in-hospital mortality with site and me factors removed

Mean SD Median CrIAge<39 10.37 5.33 9.13 (4.15, 24.04)Age 40-49 3.17 0.80 3.06 (1.92, 4.98)Age 50-59 2.75 0.52 2.70 (1.87, 3.91)Age 70-79 0.51 0.08 0.51 (0.37, 0.69)Age 80+ 0.32 0.09 0.31 (0.18, 0.52)Female 1.17 0.17 1.15 (0.87, 1.55)Tocilizumab 1.57 0.28 1.55 (1.10, 2.17)Sarilumab 2.07 0.80 1.88 (1.13, 4.17)Fixed-dose cor costeroids 1.05 0.26 1.02 (0.63, 1.65)Shock-dependent steroids 1.43 0.42 1.38 (0.77, 2.41)Tocilizumab*Fixed-dose interac on 0.99 0.05 0.99 (0.89, 1.09)Tocilizumab*Shock-dependent interac on 1.01 0.05 1.01 (0.91, 1.11)Sarilumab*Fixed-dose interac on 1.01 0.05 1.00 (0.91, 1.11)Sarilumab*Shock-dependent interac on 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose combina on 1.63 0.50 1.57 (0.88, 2.83)Tocilizumab*Shock-dependent combina on 2.26 0.80 2.14 (1.08, 4.13)Sarilumab*Fixed-dose combina on 2.19 1.01 1.97 (0.96, 4.79)Sarilumab*Shock-dependent combina on 2.98 1.49 2.64 (1.18, 6.82)

53

Table 55: Posterior probabili es for sensi vity analysis of in-hospital mortality with site and me factors removed

Posterior ProbabilityTocilizumab is superior to control 0.994Tocilizumab is fu le (OR < 1.2) 0.074Sarilumab is superior to control 0.994Sarilumab is fu le (OR < 1.2) 0.044Tocilizumab*Fixed-dose combina on OR > 1 0.931Tocilizumab*Shock-dependent combina on OR > 1 0.986Sarilumab*Fixed-dose combina on OR > 1 0.967Sarilumab*Shock-dependent combina on OR > 1 0.993Tocilizumab and Sarilumab are equivalent 0.496

6.2.3 Sensi vity analysis of in-hospital mortality for Unblinded ITT popula on with alterna ve coding of cor -

costeroid interven ons

• Model: Primary analysis dichotomous model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, pooled fixed-dura on and Shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kale-

tra + HCQ), and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

• Pa ents randomized a er the closure of the cor costeroid domain (June 17, 2020) will be coded as receiving

steroids if they received steroids prior to randomiza on or within the first two study days. There are 803

pa ents randomized a er the closure of the cor costeroid domain. Of those pa ents, 55 pa ents did not

receive steroids prior to randomiza on or within the first two study days. In the model, the pa ents that

received steroids post-closure of the cor costeroid domain will be pooled with the pa ents randomized to

a cor costeroid interven on. The pa ents that did not receive steroids post-closure of the cor costeroid

domain will be pooled with pa ents randomized to no cor costeroid interven on.

54

Table 56: Odds ra o parameters for sensi vity analysis of in-hospital mortality for Unblinded ITT popula on with

alterna ve coding of cor costeroid interven ons

Mean SD Median CrIAge<39 12.36 6.47 10.78 (4.75, 29.17)Age 40-49 3.15 0.82 3.04 (1.89, 5.07)Age 50-59 2.86 0.59 2.80 (1.89, 4.19)Age 70-79 0.40 0.07 0.39 (0.28, 0.56)Age 80+ 0.22 0.07 0.21 (0.12, 0.38)Female 1.21 0.19 1.20 (0.89, 1.64)Time epoch 1 0.94 0.08 0.94 (0.78, 1.11)Time epoch 2 0.87 0.14 0.87 (0.61, 1.15)Time epoch 3 0.86 0.17 0.85 (0.56, 1.23)Time epoch 4 0.88 0.21 0.86 (0.54, 1.34)Time epoch 5 0.90 0.24 0.88 (0.52, 1.45)Time epoch 6 0.92 0.25 0.89 (0.51, 1.48)Time epoch 7 0.94 0.25 0.91 (0.53, 1.51)Time epoch 8 0.98 0.26 0.95 (0.57, 1.56)Time epoch 9 1.04 0.26 1.01 (0.62, 1.64)Time epoch 10 1.10 0.28 1.06 (0.67, 1.73)Time epoch 11 1.10 0.26 1.07 (0.69, 1.71)Time epoch 12 1.08 0.25 1.06 (0.68, 1.64)Time epoch 13 1.10 0.27 1.08 (0.67, 1.71)Time epoch 14 1.20 0.33 1.15 (0.68, 1.95)Time epoch 15 1.38 0.51 1.29 (0.68, 2.60)Time epoch 16 1.71 1.10 1.47 (0.61, 4.28)Tocilizumab 1.65 0.30 1.63 (1.12, 2.33)Sarilumab 2.17 0.87 1.95 (1.15, 4.48)Steroids 1.21 0.27 1.18 (0.76, 1.82)Tocilizumab*cor costeroid interac on 0.99 0.05 0.99 (0.90, 1.09)Sarilumab*cor costeroid interac on 1.01 0.05 1.01 (0.91, 1.11)Tocilizumab*cor costeroid combina on 1.97 0.56 1.90 (1.07, 3.26)Sarilumab*cor costeroid combina on 2.64 1.26 2.33 (1.17, 5.95)

55

Table 57: Posterior probabili es for sensi vity analysis of in-hospital mortality for Unblinded ITT popula on with

alterna ve coding of cor costeroid interven ons

Posterior ProbabilityTocilizumab is superior to control 0.996Tocilizumab is fu le (OR < 1.2) 0.054Sarilumab is superior to control 0.993Sarilumab is fu le (OR < 1.2) 0.035Tocilizumab*steroids OR > 1 0.988Sarilumab*steroids OR > 1 0.991Tocilizumab and Sarilumab are equivalent 0.502

7 Subgroup Analyses

Subgroups are defined based on terciles of C-reac ve protein (CRP) levels for pa ents in the Unblinded ITT popu-

la on. Pa ents are included in the “lower” CRP subgroup if their CRP level is lower than 102 µg/mL. Pa ents are

included in the “middle” CRP subgroup if their CRP level is greater than or equal to 102 µg/mL and lower than 187

µg/mL. Pa ents are included in the “upper” CRP subgroup if their CRP level is greater than or equal to 187 µg/mL.

Pa ents without a measured CRP level are included in the “unknown” subgroup.

56

698 194 193 186

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Unknown Lower Middle UpperCRP Tercile

Prop

ortio

n

OSFD−1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 17: Empirical distribu on of organ support free days (OSFD) by C-reac ve protein (CRP) tercile subgroup for

the Unblinded ITT popula on.

7.1 Subgroup analyses of OSFD endpoint

7.1.1 Subgroup analysis with differen al effect on OSFD endpoint by CRP terciles

• Model: Primary analysis ordinal model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, fixed-dura on, Shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ),

and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

• Differen al treatment effect for Tocilizumab and Sarilumab by terciles of CRP will be compared to the control

arm. A posterior probability of superiority of 99%will be used as a sta s cal trigger for efficacy for each level

of CRP.

• Popula on: Unblinded ITT

57

Table 58: Odds ra o parameters for subgroup analysis with differen aleffect on OSFD endpoint by CRP tercile

Mean SD Median CrIAge<39 4.18 0.90 4.09 (2.70, 6.21)Age 40-49 2.16 0.37 2.13 (1.54, 2.96)Age 50-59 1.98 0.27 1.96 (1.50, 2.58)Age 70-79 0.52 0.08 0.52 (0.38, 0.70)Age 80+ 0.33 0.09 0.32 (0.19, 0.54)Female 1.16 0.13 1.15 (0.92, 1.44)Time epoch 1 0.94 0.08 0.94 (0.79, 1.11)Time epoch 2 0.90 0.13 0.89 (0.65, 1.17)Time epoch 3 0.97 0.18 0.96 (0.66, 1.35)Time epoch 4 1.11 0.24 1.08 (0.72, 1.66)Time epoch 5 1.25 0.31 1.21 (0.78, 1.96)Time epoch 6 1.35 0.34 1.30 (0.82, 2.13)Time epoch 7 1.43 0.36 1.38 (0.88, 2.25)Time epoch 8 1.51 0.39 1.45 (0.92, 2.42)Time epoch 9 1.47 0.38 1.41 (0.89, 2.35)Time epoch 10 1.36 0.34 1.31 (0.83, 2.17)Time epoch 11 1.15 0.28 1.12 (0.70, 1.79)Time epoch 12 1.02 0.25 0.99 (0.61, 1.58)Time epoch 13 1.02 0.26 0.98 (0.60, 1.60)Time epoch 14 1.15 0.30 1.11 (0.67, 1.85)Time epoch 15 1.38 0.45 1.31 (0.71, 2.46)Time epoch 16 1.81 0.89 1.61 (0.71, 4.12)CRP lower tercile 1.53 0.36 1.49 (0.95, 2.33)CRP middle tercile 1.63 0.37 1.59 (1.03, 2.48)CRP upper tercile 0.97 0.22 0.94 (0.60, 1.46)Tocilizumab, CRP unknown 1.68 0.33 1.64 (1.11, 2.41)Tocilizumab, CRP lower tercile 1.51 0.42 1.45 (0.85, 2.48)Tocilizumab, CRP middle tercile 1.54 0.41 1.49 (0.89, 2.49)Tocilizumab, CRP upper tercile 2.00 0.58 1.92 (1.12, 3.34)Sarilumab, CRP unknown 1.63 0.59 1.56 (0.72, 3.00)Sarilumab, CRP lower tercile 1.86 1.06 1.62 (0.76, 4.57)Sarilumab, CRP middle tercile 1.51 0.54 1.44 (0.73, 2.75)Sarilumab, CRP upper tercile 2.60 1.44 2.25 (1.08, 6.46)Fixed-dose cor costeroid 1.44 0.33 1.41 (0.91, 2.18)Shock-dependent cor costeroid 1.14 0.26 1.11 (0.71, 1.73)Tocilizumab*Fixed-dose interac on, CRP unknown 1.00 0.05 1.00 (0.90, 1.10)Tocilizumab*Fixed-dose interac on, CRP lower tercile 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose interac on, CRP middle tercile 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose interac on, CRP upper tercile 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Shock-dependent interac on, CRP unknown 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Shock-dependent interac on, CRP lower tercile 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Shock-dependent interac on, CRP middle tercile 1.00 0.05 1.00 (0.90, 1.10)

58

Table 58: Odds ra o parameters for subgroup analysis with differen aleffect on OSFD endpoint by CRP tercile (con nued)

Mean SD Median CrITocilizumab*Shock-dependent interac on, CRP upper tercile 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Fixed-dose interac on, CRP unknown 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Fixed-dose interac on, CRP lower tercile 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Fixed-dose interac on, CRP middle tercile 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Fixed-dose interac on, CRP upper tercile 1.01 0.05 1.00 (0.91, 1.11)Sarilumab*Shock-dependent interac on, CRP unknown 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Shock-dependent interac on, CRP lower tercile 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Shock-dependent interac on, CRP middle tercile 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Shock-dependent interac on, CRP upper tercile 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose combina on, CRP unknown 2.42 0.75 2.32 (1.26, 4.17)Tocilizumab*Fixed-dose combina on, CRP lower tercile 2.17 0.77 2.04 (1.04, 4.05)Tocilizumab*Fixed-dose combina on, CRP middle tercile 2.24 0.80 2.10 (1.06, 4.14)Tocilizumab*Fixed-dose combina on, CRP upper tercile 2.89 1.08 2.70 (1.34, 5.52)Tocilizumab*Shock-dependent combina on, CRP unknown 1.92 0.60 1.83 (0.99, 3.35)Tocilizumab*Shock-dependent combina on, CRP lower tercile 1.71 0.62 1.61 (0.80, 3.19)Tocilizumab*Shock-dependent combina on, CRP middletercile

1.76 0.64 1.65 (0.82, 3.35)

Tocilizumab*Shock-dependent combina on, CRP uppertercile

2.28 0.87 2.12 (1.04, 4.35)

Sarilumab*Fixed-dose combina on, CRP unknown 2.37 1.05 2.19 (0.91, 4.85)Sarilumab*Fixed-dose combina on, CRP lower tercile 2.69 1.66 2.30 (0.97, 6.96)Sarilumab*Fixed-dose combina on, CRP middle tercile 2.19 0.94 2.03 (0.90, 4.48)Sarilumab*Fixed-dose combina on, CRP upper tercile 3.77 2.30 3.19 (1.35, 9.81)Sarilumab*Shock-dependent combina on, CRP unknown 1.86 0.82 1.73 (0.70, 3.89)Sarilumab*Shock-dependent combina on, CRP lower tercile 2.12 1.32 1.81 (0.75, 5.51)Sarilumab*Shock-dependent combina on, CRP middle tercile 1.73 0.76 1.59 (0.70, 3.60)Sarilumab*Shock-dependent combina on, CRP upper tercile 2.96 1.81 2.50 (1.05, 7.84)

Table 59: Posterior probabili es for subgroup analysis with differen aleffect on OSFD endpoint by CRP tercile

Posterior ProbabilityTocilizumab is superior to control with CRP unknown 0.993Tocilizumab is fu le with CRP unknown (OR < 1.2) 0.057Tocilizumab is superior to control in CRP lower tercile 0.913Tocilizumab is fu le in CRP lower tercile (OR < 1.2) 0.243Tocilizumab is superior to control in CRP middle tercile 0.935Tocilizumab is fu le in CRP middle tercile (OR < 1.2) 0.204Tocilizumab is superior to control in CRP upper tercile 0.991Tocilizumab is fu le in CRP upper tercile (OR < 1.2) 0.044Sarilumab is superior to control with CRP unknown 0.900

59

Table 59: Posterior probabili es for subgroup analysis with differen aleffect on OSFD endpoint by CRP tercile (con nued)

Posterior ProbabilitySarilumab is fu le with CRP unknown (OR < 1.2) 0.198Sarilumab is superior to control in CRP lower tercile 0.903Sarilumab is fu le in CRP lower tercile (OR < 1.2) 0.208Sarilumab is superior to control in CRP middle tercile 0.866Sarilumab is fu le in CRP middle tercile (OR < 1.2) 0.283Sarilumab is superior to control in CRP upper tercile 0.986Sarilumab is fu le in CRP upper tercile (OR < 1.2) 0.046Tocilizumab*Fixed-dose combina on, CRP unknown OR > 1 0.996Tocilizumab*Fixed-dose combina on, CRP lower tercile OR > 1 0.981Tocilizumab*Fixed-dose combina on, CRP middle tercile OR > 1 0.984Tocilizumab*Fixed-dose combina on, CRP upper tercile OR > 1 0.997Tocilizumab*Shock-dependent combina on, CRP unknown OR > 1 0.973Tocilizumab*Shock-dependent combina on, CRP lower tercile OR > 1 0.911Tocilizumab*Shock-dependent combina on, CRP middle tercile OR > 1 0.923Tocilizumab*Shock-dependent combina on, CRP upper tercile OR > 1 0.981Sarilumab*Fixed-dose combina on, CRP unknown OR > 1 0.962Sarilumab*Fixed-dose combina on, CRP lower tercile OR > 1 0.971Sarilumab*Fixed-dose combina on, CRP middle tercile OR > 1 0.957Sarilumab*Fixed-dose combina on, CRP upper tercile OR > 1 0.996Sarilumab*Shock-dependent combina on, CRP unknown OR > 1 0.898Sarilumab*Shock-dependent combina on, CRP lower tercile OR > 1 0.911Sarilumab*Shock-dependent combina on, CRP middle tercile OR > 1 0.876Sarilumab*Shock-dependent combina on, CRP upper tercile OR > 1 0.981

7.1.2 Subgroup analysis with differen al effect on OSFD endpoint by CRP tercile with pooled IL-6ra interven ons

• Model: Primary analysis ordinal model

• Factors: Age, sex, site, me, control and pooled IL-6ra interven ons, cor costeroid interven ons (control,

fixed-dura on, Shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ), and in-

terac ons between pooled IL-6ra interven ons, cor costeroids, and an virals

• The treatment effect for pooled IL-6ra interven ons by tercile of CRP will be compared to the control arm. A

posterior probability of superiority of 99% will be used as a sta s cal trigger for efficacy for each level of CRP.

• Popula on: Unblinded ITT

60

Table 60: Odds ra o parameters for subgroup analysis with differen aleffect on OSFD endpoint by CRP tercile with pooled IL-6ra interven ons

Mean SD Median CrIAge<39 4.18 0.91 4.07 (2.66, 6.21)Age 40-49 2.16 0.37 2.13 (1.53, 2.97)Age 50-59 1.98 0.27 1.96 (1.50, 2.57)Age 70-79 0.52 0.08 0.52 (0.39, 0.70)Age 80+ 0.33 0.09 0.32 (0.19, 0.54)Female 1.16 0.13 1.15 (0.93, 1.43)Time epoch 1 0.94 0.08 0.94 (0.79, 1.11)Time epoch 2 0.90 0.13 0.90 (0.65, 1.17)Time epoch 3 0.97 0.17 0.96 (0.66, 1.35)Time epoch 4 1.11 0.23 1.08 (0.72, 1.62)Time epoch 5 1.25 0.30 1.21 (0.79, 1.92)Time epoch 6 1.35 0.33 1.30 (0.83, 2.11)Time epoch 7 1.43 0.35 1.39 (0.88, 2.23)Time epoch 8 1.52 0.38 1.46 (0.92, 2.40)Time epoch 9 1.47 0.36 1.42 (0.91, 2.31)Time epoch 10 1.36 0.34 1.31 (0.84, 2.12)Time epoch 11 1.16 0.28 1.13 (0.70, 1.79)Time epoch 12 1.02 0.25 1.00 (0.61, 1.59)Time epoch 13 1.02 0.26 0.99 (0.61, 1.61)Time epoch 14 1.15 0.30 1.11 (0.67, 1.85)Time epoch 15 1.37 0.44 1.31 (0.71, 2.45)Time epoch 16 1.79 0.88 1.60 (0.70, 4.08)CRP lower tercile 1.53 0.36 1.49 (0.94, 2.36)CRP middle tercile 1.64 0.38 1.61 (1.02, 2.51)CRP upper tercile 0.97 0.22 0.94 (0.62, 1.47)Pooled IL-6ra, CRP unknown 1.67 0.33 1.63 (1.11, 2.41)Pooled IL-6ra, CRP lower tercile 1.52 0.42 1.46 (0.87, 2.48)Pooled IL-6ra, CRP middle tercile 1.52 0.40 1.47 (0.90, 2.44)Pooled IL-6ra, CRP upper tercile 2.05 0.58 1.98 (1.15, 3.39)Fixed-dose cor costeroid 1.44 0.33 1.41 (0.91, 2.18)Shock-dependent cor costeroid 1.14 0.27 1.11 (0.70, 1.75)Pooled IL-6ra*Fixed-dose interac on, CRP unknown 1.00 0.05 1.00 (0.91, 1.10)Pooled IL-6ra*Fixed-dose interac on, CRP lower tercile 1.00 0.05 1.00 (0.91, 1.11)Pooled IL-6ra*Fixed-dose interac on, CRP middle tercile 1.00 0.05 1.00 (0.91, 1.11)Pooled IL-6ra*Fixed-dose interac on, CRP upper tercile 1.00 0.05 1.00 (0.91, 1.10)Pooled IL-6ra*Shock-dependent interac on, CRP unknown 1.00 0.05 1.00 (0.91, 1.11)Pooled IL-6ra*Shock-dependent interac on, CRP lower tercile 1.00 0.05 1.00 (0.91, 1.10)Pooled IL-6ra*Shock-dependent interac on, CRP middletercile

1.00 0.05 1.00 (0.91, 1.10)

Pooled IL-6ra*Shock-dependent interac on, CRP upper tercile 1.00 0.05 1.00 (0.90, 1.10)Pooled IL-6ra*Fixed-dose combina on, CRP unknown 2.40 0.74 2.29 (1.27, 4.17)Pooled IL-6ra*Fixed-dose combina on, CRP lower tercile 2.19 0.80 2.06 (1.05, 4.08)

61

Table 60: Odds ra o parameters for subgroup analysis with differen aleffect on OSFD endpoint by CRP tercile with pooled IL-6ra interven ons(con nued)

Mean SD Median CrIPooled IL-6ra*Fixed-dose combina on, CRP middle tercile 2.20 0.77 2.07 (1.07, 4.01)Pooled IL-6ra*Fixed-dose combina on, CRP upper tercile 2.96 1.09 2.78 (1.39, 5.64)Pooled IL-6ra*Shock-dependent combina on, CRP unknown 1.91 0.60 1.82 (0.99, 3.33)Pooled IL-6ra*Shock-dependent combina on, CRP lowertercile

1.73 0.65 1.63 (0.80, 3.29)

Pooled IL-6ra*Shock-dependent combina on, CRP middletercile

1.74 0.62 1.63 (0.83, 3.23)

Pooled IL-6ra*Shock-dependent combina on, CRP uppertercile

2.34 0.88 2.20 (1.08, 4.49)

Table 61: Posterior probabili es for subgroup analysis with differen aleffect on OSFD endpoint by CRP tercile with pooled IL-6ra interven ons

Posterior ProbabilityPooled IL-6ra is superior to control with CRP unknown 0.994Pooled IL-6ra is fu le with CRP unknown (OR < 1.2) 0.059Pooled IL-6ra is superior to control in CRP lower tercile 0.923Pooled IL-6ra is fu le in CRP lower tercile (OR < 1.2) 0.232Pooled IL-6ra is superior to control in CRP middle tercile 0.934Pooled IL-6ra is fu le in CRP middle tercile (OR < 1.2) 0.212Pooled IL-6ra is superior to control in CRP upper tercile 0.994Pooled IL-6ra is fu le in CRP upper tercile (OR < 1.2) 0.036Pooled IL-6ra*Fixed-dose combina on, CRP unknown OR > 1 0.997Pooled IL-6ra*Fixed-dose combina on, CRP lower tercile OR > 1 0.981Pooled IL-6ra*Fixed-dose combina on, CRP middle tercile OR > 1 0.984Pooled IL-6ra*Fixed-dose combina on, CRP upper tercile OR > 1 0.998Pooled IL-6ra*Shock-dependent combina on, CRP unknown OR > 1 0.973Pooled IL-6ra*Shock-dependent combina on, CRP lower tercile OR > 1 0.910Pooled IL-6ra*Shock-dependent combina on, CRP middle tercile OR > 1 0.921Pooled IL-6ra*Shock-dependent combina on, CRP upper tercile OR > 1 0.985

7.2 Subgroup analyses of in-hospital mortality

7.2.1 Subgroup analysis with differen al effect on in-hospital mortality by CRP tercile

• Model: Primary dichotomous model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, fixed-dura on, Shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ),

62

and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

• Differen al treatment effect for Tocilizumab and Sarilumab by terciles of CRP will be compared to the control

arm. A posterior probability of superiority of 99%will be used as a sta s cal trigger for efficacy for each level

of CRP.

• Popula on: Unblinded ITT

Table 62: Odds ra o parameters for subgroup analysis with differen aleffect on in-hospital mortality by CRP tercile

Mean SD Median CrIAge<39 13.44 7.00 11.78 (5.10, 31.44)Age 40-49 3.43 0.91 3.31 (2.02, 5.55)Age 50-59 3.06 0.62 3.01 (2.02, 4.45)Age 70-79 0.42 0.08 0.42 (0.29, 0.59)Age 80+ 0.22 0.07 0.21 (0.12, 0.38)Female 1.18 0.19 1.16 (0.86, 1.59)Time epoch 1 0.94 0.08 0.93 (0.78, 1.10)Time epoch 2 0.86 0.14 0.86 (0.59, 1.14)Time epoch 3 0.87 0.18 0.86 (0.55, 1.26)Time epoch 4 0.91 0.22 0.88 (0.54, 1.41)Time epoch 5 0.95 0.26 0.92 (0.53, 1.57)Time epoch 6 0.98 0.29 0.95 (0.54, 1.65)Time epoch 7 1.02 0.30 0.98 (0.55, 1.70)Time epoch 8 1.08 0.31 1.03 (0.59, 1.80)Time epoch 9 1.16 0.34 1.11 (0.64, 1.96)Time epoch 10 1.23 0.38 1.16 (0.66, 2.13)Time epoch 11 1.22 0.38 1.16 (0.66, 2.12)Time epoch 12 1.18 0.36 1.13 (0.63, 2.03)Time epoch 13 1.20 0.38 1.14 (0.63, 2.08)Time epoch 14 1.30 0.44 1.24 (0.65, 2.36)Time epoch 15 1.53 0.65 1.40 (0.65, 3.17)Time epoch 16 1.97 1.52 1.60 (0.60, 5.45)CRP lower tercile 1.26 0.37 1.20 (0.70, 2.11)CRP middle tercile 1.93 0.59 1.84 (1.04, 3.29)CRP upper tercile 1.15 0.34 1.10 (0.64, 1.94)Tocilizumab, CRP unknown 1.76 0.48 1.69 (1.03, 2.87)Tocilizumab, CRP lower tercile 1.60 0.61 1.49 (0.74, 3.08)Tocilizumab, CRP middle tercile 1.72 0.68 1.60 (0.77, 3.40)Tocilizumab, CRP upper tercile 1.62 0.61 1.51 (0.76, 3.12)Sarilumab, CRP unknown 2.20 1.53 1.84 (0.86, 5.88)Sarilumab, CRP lower tercile 2.05 1.86 1.66 (0.71, 5.79)Sarilumab, CRP middle tercile 1.82 1.06 1.61 (0.66, 4.26)Sarilumab, CRP upper tercile 2.43 3.28 1.80 (0.73, 7.88)Fixed-dose cor costeroid 0.98 0.30 0.94 (0.52, 1.69)

63

Shock-dependent cor costeroid 1.21 0.39 1.16 (0.62, 2.14)Tocilizumab*Fixed-dose interac on, CRP unknown 0.99 0.05 0.99 (0.89, 1.09)Tocilizumab*Fixed-dose interac on, CRP lower tercile 1.00 0.05 1.00 (0.90, 1.10)Tocilizumab*Fixed-dose interac on, CRP middle tercile 1.00 0.05 1.00 (0.90, 1.10)Tocilizumab*Fixed-dose interac on, CRP upper tercile 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Shock-dependent interac on, CRP unknown 1.01 0.05 1.01 (0.91, 1.11)Tocilizumab*Shock-dependent interac on, CRP lower tercile 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Shock-dependent interac on, CRP middle tercile 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Shock-dependent interac on, CRP upper tercile 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Fixed-dose interac on, CRP unknown 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Fixed-dose interac on, CRP lower tercile 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Fixed-dose interac on, CRP middle tercile 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Fixed-dose interac on, CRP upper tercile 1.00 0.05 1.00 (0.91, 1.11)Sarilumab*Shock-dependent interac on, CRP unknown 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Shock-dependent interac on, CRP lower tercile 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Shock-dependent interac on, CRP middle tercile 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Shock-dependent interac on, CRP upper tercile 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose combina on, CRP unknown 1.72 0.74 1.57 (0.73, 3.54)Tocilizumab*Fixed-dose combina on, CRP lower tercile 1.57 0.79 1.41 (0.55, 3.60)Tocilizumab*Fixed-dose combina on, CRP middle tercile 1.70 0.89 1.50 (0.58, 3.94)Tocilizumab*Fixed-dose combina on, CRP upper tercile 1.60 0.80 1.43 (0.58, 3.57)Tocilizumab*Shock-dependent combina on, CRP unknown 2.15 0.96 1.95 (0.88, 4.50)Tocilizumab*Shock-dependent combina on, CRP lower tercile 1.94 1.00 1.73 (0.67, 4.45)Tocilizumab*Shock-dependent combina on, CRP middletercile

2.10 1.13 1.84 (0.69, 4.91)

Tocilizumab*Shock-dependent combina on, CRP uppertercile

1.97 1.03 1.75 (0.68, 4.64)

Sarilumab*Fixed-dose combina on, CRP unknown 2.17 1.65 1.76 (0.66, 6.30)Sarilumab*Fixed-dose combina on, CRP lower tercile 2.02 1.99 1.59 (0.54, 6.20)Sarilumab*Fixed-dose combina on, CRP middle tercile 1.80 1.24 1.52 (0.50, 4.69)Sarilumab*Fixed-dose combina on, CRP upper tercile 2.42 3.64 1.73 (0.59, 8.45)Sarilumab*Shock-dependent combina on, CRP unknown 2.67 2.09 2.16 (0.79, 7.78)Sarilumab*Shock-dependent combina on, CRP lower tercile 2.48 2.57 1.95 (0.65, 7.53)Sarilumab*Shock-dependent combina on, CRP middle tercile 2.21 1.56 1.86 (0.61, 5.89)Sarilumab*Shock-dependent combina on, CRP upper tercile 2.96 3.80 2.10 (0.68, 10.41)

Table 63: Posterior probabili es for subgroup analysis with differen aleffect on in-hospital mortality by CRP tercile

Posterior ProbabilityTocilizumab is superior to control with CRP unknown 0.980Tocilizumab is fu le with CRP unknown (OR < 1.2) 0.091Tocilizumab is superior to control in CRP lower tercile 0.868Tocilizumab is fu le in CRP lower tercile (OR < 1.2) 0.269

64

Table 63: Posterior probabili es for subgroup analysis with differen aleffect on in-hospital mortality by CRP tercile (con nued)

Posterior ProbabilityTocilizumab is superior to control in CRP middle tercile 0.893Tocilizumab is fu le in CRP middle tercile (OR < 1.2) 0.221Tocilizumab is superior to control in CRP upper tercile 0.874Tocilizumab is fu le in CRP upper tercile (OR < 1.2) 0.261Sarilumab is superior to control with CRP unknown 0.950Sarilumab is fu le with CRP unknown (OR < 1.2) 0.119Sarilumab is superior to control in CRP lower tercile 0.875Sarilumab is fu le in CRP lower tercile (OR < 1.2) 0.232Sarilumab is superior to control in CRP middle tercile 0.858Sarilumab is fu le in CRP middle tercile (OR < 1.2) 0.252Sarilumab is superior to control in CRP upper tercile 0.900Sarilumab is fu le in CRP upper tercile (OR < 1.2) 0.189Tocilizumab*Fixed-dose combina on, CRP unknown OR > 1 0.871Tocilizumab*Fixed-dose combina on, CRP lower tercile OR > 1 0.771Tocilizumab*Fixed-dose combina on, CRP middle tercile OR > 1 0.800Tocilizumab*Fixed-dose combina on, CRP upper tercile OR > 1 0.775Tocilizumab*Shock-dependent combina on, CRP unknown OR > 1 0.946Tocilizumab*Shock-dependent combina on, CRP lower tercile OR > 1 0.870Tocilizumab*Shock-dependent combina on, CRP middle tercile OR > 1 0.890Tocilizumab*Shock-dependent combina on, CRP upper tercile OR > 1 0.874Sarilumab*Fixed-dose combina on, CRP unknown OR > 1 0.872Sarilumab*Fixed-dose combina on, CRP lower tercile OR > 1 0.800Sarilumab*Fixed-dose combina on, CRP middle tercile OR > 1 0.776Sarilumab*Fixed-dose combina on, CRP upper tercile OR > 1 0.830Sarilumab*Shock-dependent combina on, CRP unknown OR > 1 0.932Sarilumab*Shock-dependent combina on, CRP lower tercile OR > 1 0.884Sarilumab*Shock-dependent combina on, CRP middle tercile OR > 1 0.866Sarilumab*Shock-dependent combina on, CRP upper tercile OR > 1 0.900

7.2.2 Subgroup analysis with differen al effect on in-hospital mortality by CRP tercile with pooled IL-6ra inter-

ven ons

• Model: Primary dichotomous model

• Factors: Age, sex, site, me, control and pooled tocilizumab and sarilumab interven ons, cor costeroid inter-

ven ons (control, fixed-dura on, Shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra

+ HCQ), and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

• Differen al treatment effect for pooled Tocilizumab/Sarilumab by terciles of CRP will be compared to the

control arm. A posterior probability of superiority of 99% will be used as a sta s cal trigger for efficacy for

65

each level of CRP.

• Popula on: Unblinded ITT

Table 64: Odds ra o parameters for subgroup analysis with differen aleffect on in-hospital mortality by CRP tercile with pooled IL-6ra inter-ven ons

Mean SD Median CrIAge<39 13.39 7.05 11.81 (5.17, 31.13)Age 40-49 3.42 0.90 3.29 (2.01, 5.50)Age 50-59 3.06 0.64 3.00 (2.02, 4.52)Age 70-79 0.42 0.08 0.42 (0.29, 0.59)Age 80+ 0.22 0.07 0.21 (0.12, 0.38)Female 1.18 0.19 1.16 (0.86, 1.58)Time epoch 1 0.93 0.08 0.93 (0.77, 1.10)Time epoch 2 0.87 0.14 0.86 (0.61, 1.15)Time epoch 3 0.86 0.18 0.85 (0.55, 1.24)Time epoch 4 0.89 0.21 0.87 (0.54, 1.38)Time epoch 5 0.94 0.25 0.90 (0.54, 1.52)Time epoch 6 0.97 0.27 0.94 (0.55, 1.59)Time epoch 7 1.00 0.28 0.97 (0.55, 1.66)Time epoch 8 1.06 0.30 1.02 (0.59, 1.75)Time epoch 9 1.13 0.33 1.08 (0.63, 1.90)Time epoch 10 1.19 0.36 1.14 (0.66, 2.04)Time epoch 11 1.19 0.36 1.14 (0.65, 2.05)Time epoch 12 1.17 0.35 1.12 (0.63, 1.99)Time epoch 13 1.18 0.37 1.13 (0.62, 2.06)Time epoch 14 1.29 0.44 1.22 (0.64, 2.34)Time epoch 15 1.49 0.64 1.36 (0.63, 3.12)Time epoch 16 1.85 1.22 1.55 (0.58, 5.01)CRP lower tercile 1.24 0.35 1.19 (0.69, 2.06)CRP middle tercile 1.91 0.58 1.82 (1.03, 3.29)CRP upper tercile 1.13 0.33 1.09 (0.63, 1.91)Pooled IL-6ra, CRP unknown 1.75 0.46 1.69 (1.03, 2.79)Pooled IL-6ra, CRP lower tercile 1.63 0.59 1.53 (0.78, 3.10)Pooled IL-6ra, CRP middle tercile 1.69 0.64 1.58 (0.78, 3.24)Pooled IL-6ra, CRP upper tercile 1.67 0.61 1.57 (0.80, 3.15)Fixed-dose cor costeroid 0.98 0.29 0.94 (0.53, 1.66)Shock-dependent cor costeroid 1.21 0.39 1.15 (0.63, 2.13)Pooled IL-6ra*Fixed-dose interac on, CRP unknown 0.99 0.05 0.99 (0.90, 1.09)Pooled IL-6ra*Fixed-dose interac on, CRP lower tercile 1.00 0.05 1.00 (0.91, 1.10)Pooled IL-6ra*Fixed-dose interac on, CRP middle tercile 1.00 0.05 1.00 (0.91, 1.10)Pooled IL-6ra*Fixed-dose interac on, CRP upper tercile 1.00 0.05 1.00 (0.91, 1.11)Pooled IL-6ra*Shock-dependent interac on, CRP unknown 1.01 0.05 1.01 (0.91, 1.11)Pooled IL-6ra*Shock-dependent interac on, CRP lower tercile 1.00 0.05 1.00 (0.90, 1.11)

66

Table 64: Odds ra o parameters for subgroup analysis with differen aleffect on in-hospital mortality by CRP tercile with pooled IL-6ra inter-ven ons (con nued)

Mean SD Median CrIPooled IL-6ra*Shock-dependent interac on, CRP middletercile

1.00 0.05 1.00 (0.91, 1.11)

Pooled IL-6ra*Shock-dependent interac on, CRP upper tercile 1.00 0.05 1.00 (0.90, 1.10)Pooled IL-6ra*Fixed-dose combina on, CRP unknown 1.70 0.69 1.58 (0.73, 3.39)Pooled IL-6ra*Fixed-dose combina on, CRP lower tercile 1.60 0.78 1.45 (0.59, 3.53)Pooled IL-6ra*Fixed-dose combina on, CRP middle tercile 1.66 0.83 1.47 (0.59, 3.77)Pooled IL-6ra*Fixed-dose combina on, CRP upper tercile 1.64 0.79 1.47 (0.61, 3.67)Pooled IL-6ra*Shock-dependent combina on, CRP unknown 2.14 0.92 1.97 (0.88, 4.44)Pooled IL-6ra*Shock-dependent combina on, CRP lowertercile

1.99 1.00 1.78 (0.70, 4.54)

Pooled IL-6ra*Shock-dependent combina on, CRP middletercile

2.06 1.07 1.81 (0.71, 4.78)

Pooled IL-6ra*Shock-dependent combina on, CRP uppertercile

2.02 1.00 1.81 (0.73, 4.57)

Table 65: Posterior probabili es for subgroup analysis with differen aleffect on in-hospital mortality by CRP tercile with pooled IL-6ra inter-ven ons

Posterior ProbabilityPooled IL-6ra is superior to control with CRP unknown 0.981Pooled IL-6ra is fu le with CRP unknown (OR < 1.2) 0.092Pooled IL-6ra is superior to control in CRP lower tercile 0.887Pooled IL-6ra is fu le in CRP lower tercile (OR < 1.2) 0.242Pooled IL-6ra is superior to control in CRP middle tercile 0.897Pooled IL-6ra is fu le in CRP middle tercile (OR < 1.2) 0.230Pooled IL-6ra is superior to control in CRP upper tercile 0.903Pooled IL-6ra is fu le in CRP upper tercile (OR < 1.2) 0.221Pooled IL-6ra*Fixed-dose combina on, CRP unknown OR > 1 0.876Pooled IL-6ra*Fixed-dose combina on, CRP lower tercile OR > 1 0.790Pooled IL-6ra*Fixed-dose combina on, CRP middle tercile OR > 1 0.802Pooled IL-6ra*Fixed-dose combina on, CRP upper tercile OR > 1 0.807Pooled IL-6ra*Shock-dependent combina on, CRP unknown OR > 1 0.949Pooled IL-6ra*Shock-dependent combina on, CRP lower tercile OR > 1 0.890Pooled IL-6ra*Shock-dependent combina on, CRP middle tercile OR > 1 0.895Pooled IL-6ra*Shock-dependent combina on, CRP upper tercile OR > 1 0.901

8 Post Hoc Analyses

67

8.1 Post hoc analysis of OSFD without borrowing between IL-6ra interven ons

• Model: Primary analysis ordinal model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, fixed-dura on, Shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ),

and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

• Popula on: Unblinded ITT

• The nes ng of tocilizumab and sarilumab has been removed from this model. Each interven on log odds

ra o is modeled with independent standard normal prior distribu ons.

68

Table 66: Odds ra o parameters for exploratory analysis of OSFD endpoint without borrowing between IL-6ra inter-

ven ons

Mean SD Median CrIAge<39 4.09 0.88 3.99 (2.65, 6.08)Age 40-49 2.11 0.36 2.08 (1.50, 2.90)Age 50-59 1.96 0.27 1.94 (1.50, 2.54)Age 70-79 0.52 0.08 0.51 (0.38, 0.69)Age 80+ 0.33 0.09 0.32 (0.19, 0.54)Female 1.17 0.13 1.16 (0.93, 1.45)Time epoch 1 0.95 0.08 0.95 (0.80, 1.11)Time epoch 2 0.91 0.13 0.91 (0.67, 1.19)Time epoch 3 0.96 0.17 0.95 (0.66, 1.34)Time epoch 4 1.07 0.22 1.04 (0.70, 1.57)Time epoch 5 1.17 0.28 1.13 (0.74, 1.81)Time epoch 6 1.23 0.30 1.19 (0.77, 1.92)Time epoch 7 1.26 0.30 1.23 (0.79, 1.96)Time epoch 8 1.30 0.30 1.26 (0.82, 2.01)Time epoch 9 1.23 0.27 1.20 (0.80, 1.87)Time epoch 10 1.13 0.24 1.10 (0.74, 1.68)Time epoch 11 0.97 0.20 0.95 (0.63, 1.42)Time epoch 12 0.86 0.19 0.85 (0.55, 1.28)Time epoch 13 0.86 0.19 0.84 (0.54, 1.28)Time epoch 14 0.95 0.22 0.92 (0.58, 1.45)Time epoch 15 1.11 0.32 1.06 (0.61, 1.86)Time epoch 16 1.40 0.62 1.27 (0.59, 2.96)Tocilizumab 1.64 0.23 1.62 (1.24, 2.14)Sarilumab 1.98 0.60 1.90 (1.08, 3.41)Fixed-dose cor costeroid 1.44 0.32 1.40 (0.91, 2.16)Shock-dependent cor costeroid 1.14 0.26 1.11 (0.71, 1.72)Tocilizumab*Fixed-dose interac on effect 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Shock-dependent interac on effect 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Fixed-dose interac on effect 1.00 0.05 1.00 (0.91, 1.10)Sarilumab*Shock-dependent interac on effect 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose combina on 2.35 0.64 2.27 (1.36, 3.83)Tocilizumab*Shock-dependent combina on 1.87 0.52 1.80 (1.05, 3.11)Sarilumab*Fixed-dose combina on 2.86 1.10 2.68 (1.28, 5.56)Sarilumab*Shock-dependent combina on 2.26 0.90 2.10 (1.00, 4.43)

69

Table 67: Posterior probabili es for exploratory analysis of OSFD endpoint without borrowing between IL-6ra inter-

ven ons

Posterior ProbabilityTocilizumab is superior to control 1.000Tocilizumab is fu le (OR < 1.2) 0.015Sarilumab is superior to control 0.987Sarilumab is fu le (OR < 1.2) 0.057Tocilizumab*Fixed-dose combina on OR > 1 0.999Tocilizumab*Shock-dependent combina on OR > 1 0.984Sarilumab*Fixed-dose combina on OR > 1 0.996Sarilumab*Shock-dependent combina on OR > 1 0.975Tocilizumab and Sarilumab are equivalent 0.401

8.2 Post hoc analysis of in-hospital mortality without borrowing between IL-6ra interven ons

• Model: Primary dichotomous model

• Factors: Age, sex, site, me, tocilizumab, sarilumab, and control interven ons, cor costeroid interven ons

(control, fixed-dura on, Shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ),

and interac ons between tocilizumab, sarilumab, cor costeroids, and an virals

• Popula on: Unblinded ITT

• The nes ng of tocilizumab and sarilumab has been removed from this model. Each interven on log odds

ra o is modeled with independent standard normal prior distribu ons.

70

Table 68: Odds ra o parameters for exploratory analysis of in-hospital mortality without borrowing between IL-6ra

interven ons

Mean SD Median CrIAge<39 12.91 6.62 11.36 (4.94, 29.87)Age 40-49 3.32 0.87 3.22 (1.96, 5.35)Age 50-59 2.99 0.60 2.93 (2.01, 4.34)Age 70-79 0.41 0.07 0.41 (0.29, 0.58)Age 80+ 0.23 0.07 0.22 (0.12, 0.39)Female 1.19 0.19 1.18 (0.86, 1.62)Time epoch 1 0.94 0.08 0.94 (0.78, 1.11)Time epoch 2 0.88 0.14 0.88 (0.62, 1.16)Time epoch 3 0.87 0.17 0.86 (0.56, 1.25)Time epoch 4 0.89 0.21 0.86 (0.54, 1.35)Time epoch 5 0.91 0.24 0.88 (0.53, 1.45)Time epoch 6 0.92 0.25 0.89 (0.53, 1.49)Time epoch 7 0.93 0.25 0.90 (0.53, 1.52)Time epoch 8 0.97 0.26 0.93 (0.57, 1.56)Time epoch 9 1.01 0.26 0.97 (0.60, 1.63)Time epoch 10 1.04 0.27 1.00 (0.62, 1.68)Time epoch 11 1.03 0.27 0.99 (0.61, 1.64)Time epoch 12 1.00 0.26 0.97 (0.58, 1.60)Time epoch 13 1.01 0.28 0.97 (0.56, 1.66)Time epoch 14 1.08 0.33 1.04 (0.58, 1.87)Time epoch 15 1.23 0.48 1.15 (0.56, 2.43)Time epoch 16 1.50 0.94 1.28 (0.52, 3.90)Tocilizumab 1.58 0.30 1.56 (1.08, 2.24)Sarilumab 2.73 1.13 2.54 (1.17, 5.56)Fixed-dose cor costeroid 0.99 0.30 0.95 (0.53, 1.68)Shock-dependent cor costeroid 1.21 0.39 1.15 (0.63, 2.14)Tocilizumab*Fixed-dose interac on effect 0.99 0.05 0.99 (0.90, 1.09)Tocilizumab*Shock-dependent interac on effect 1.01 0.05 1.00 (0.91, 1.11)Sarilumab*Fixed-dose interac on effect 1.00 0.05 1.00 (0.91, 1.11)Sarilumab*Shock-dependent interac on effect 1.00 0.05 1.00 (0.91, 1.10)Tocilizumab*Fixed-dose combina on 1.55 0.55 1.46 (0.73, 2.88)Tocilizumab*Shock-dependent combina on 1.93 0.74 1.80 (0.87, 3.73)Sarilumab*Fixed-dose combina on 2.71 1.41 2.42 (0.91, 6.32)Sarilumab*Shock-dependent combina on 3.31 1.77 2.93 (1.11, 7.71)

71

Table 69: Posterior probabili es for exploratory analysis of in-hospital mortality endpoint without borrowing be-

tween IL-6ra interven ons

Posterior ProbabilityTocilizumab is superior to control 0.991Tocilizumab is fu le (OR < 1.2) 0.084Sarilumab is superior to control 0.992Sarilumab is fu le (OR < 1.2) 0.029Tocilizumab*Fixed-dose combina on OR > 1 0.864Tocilizumab*Shock-dependent combina on OR > 1 0.946Sarilumab*Fixed-dose combina on OR > 1 0.962Sarilumab*Shock-dependent combina on OR > 1 0.984Tocilizumab and Sarilumab are equivalent 0.185

8.3 Data summaries of baseline mechanical ven la on status in the IL-6ra specific ITT popula on

Table 70: Summary of baseline mechanical ven la on by interven on

Interven on No mechanical ven la on at baseline Mechanical ven la on at baselineTocilizumab 249 (70.5%) 104 (29.5%)Sarilumab 40 (83.3%) 8 (16.7%)Control 281 (69.9%) 121 (30.1%)

Pooled IL-6ra 289 (72.1%) 112 (27.9%)Pooled IL-6ra and control 570 (71%) 233 (29%)

Table 71: Summary of subjects on mechanical ven la on at baseline from the IL-6ra specific ITT popula on

Interven on # Pa ents # Known # Deaths (%) OSFD median (IQR) OSFD in Survivors* median (IQR)Tocilizumab 104 102 42 (41.2%) 0 (-1, 10.75) 10 (0, 14)Sarilumab 8 7 1 (14.3%) 0 (0, 5) 2 (0, 5.5)Control 121 119 58 (48.7%) 0 (-1, 8.5) 8 (0, 14)

Pooled IL-6ra 112 109 43 (39.4%) 0 (-1, 10) 8 (0, 14)* Days Free of Organ Support in Survivors measured within 21 days.

Table 72: Summary of subjects on no mechanical ven la on at baseline from the IL-6ra specific ITT popula on

Interven on # Pa ents # Known # Deaths (%) OSFD median (IQR) OSFD in Survivors* median (IQR)Tocilizumab 249 248 56 (22.6%) 13 (0, 17) 15 (11, 18)Sarilumab 40 38 9 (23.7%) 12.5 (0, 16.75) 15 (11, 17)Control 281 278 84 (30.2%) 7 (-1, 16) 14 (5, 17.75)

Pooled IL-6ra 289 286 65 (22.7%) 13 (0, 17) 15 (11, 18)* Days Free of Organ Support in Survivors measured within 21 days.

72

564

228Mechanical ventilation at baseline

No mechanical ventilation at baseline

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0Proportion

OSFD−1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 18: Empirical distribu on of organ support free days (OSFD) by baseline mechanical ven la on status. This

plot is restricted to the IL-6ra ITT popula on.

278

248

38

119

102

7

Mechanical ventilation at baseline

No mechanical ventilation at baseline

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Control

Sarilumab

Tocilizumab

Control

Sarilumab

Tocilizumab

Proportion

OSFD−1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 19: Empirical distribu on of organ support free days (OSFD) for Tocilizumab, Sarilumab, and control by base-

line mechanical ven la on status. This plot is restricted to the IL-6ra ITT popula on.

73

286

278

109

119

Mechanical ventilation at baseline

No mechanical ventilation at baseline

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Control

Pooled IL−6ra

Control

Pooled IL−6ra

Proportion

OSFD−1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

Figure 20: Empirical distribu on of organ support free days (OSFD) for pooled IL-6ra and control interven ons by

baseline mechanical ven la on status. This plot is restricted to the IL-6ra ITT popula on.

8.4 Post hoc analysis with differen al effect on OSFD endpoint by baseline mechanical ven la on sta-

tus with pooled IL-6ra interven ons

• Model: Primary analysis ordinal model

• Factors: Age, sex, site, me, control and pooled IL-6ra interven ons, cor costeroid interven ons (control,

fixed-dura on, Shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ), and in-

terac ons between pooled IL-6ra interven ons, cor costeroids, and an virals

• The treatment effect for pooled IL-6ra interven ons by baseline mechanical ven la on status will be com-

pared to the control arm. A posterior probability of superiority of 99% will be used as a sta s cal trigger for

efficacy for each mechanical ven la on subgroup.

• Popula on: Unblinded ITT

74

Table 73: Odds ra o parameters for subgroup analysis with differen-al effect by baseline mechanical ven la on status with pooled IL-6ra

interven ons

Mean SD Median CrIAge<39 4.34 0.93 4.24 (2.81, 6.49)Age 40-49 1.98 0.33 1.95 (1.41, 2.71)Age 50-59 1.94 0.27 1.92 (1.46, 2.51)Age 70-79 0.50 0.08 0.49 (0.36, 0.66)Age 80+ 0.30 0.08 0.29 (0.17, 0.50)Female 1.16 0.13 1.16 (0.93, 1.44)Time epoch 1 0.95 0.08 0.95 (0.79, 1.10)Time epoch 2 0.90 0.13 0.90 (0.66, 1.17)Time epoch 3 0.96 0.17 0.95 (0.67, 1.32)Time epoch 4 1.07 0.22 1.04 (0.71, 1.57)Time epoch 5 1.18 0.27 1.14 (0.76, 1.82)Time epoch 6 1.25 0.30 1.21 (0.80, 1.94)Time epoch 7 1.30 0.30 1.26 (0.84, 2.02)Time epoch 8 1.36 0.31 1.32 (0.87, 2.06)Time epoch 9 1.35 0.29 1.31 (0.88, 2.03)Time epoch 10 1.31 0.28 1.28 (0.86, 1.95)Time epoch 11 1.22 0.25 1.19 (0.79, 1.77)Time epoch 12 1.17 0.25 1.14 (0.75, 1.72)Time epoch 13 1.21 0.26 1.19 (0.77, 1.81)Time epoch 14 1.42 0.32 1.38 (0.88, 2.13)Time epoch 15 1.75 0.49 1.68 (0.98, 2.91)Time epoch 16 2.32 1.01 2.12 (1.00, 4.87)Mechanical ven la on at baseline 0.39 0.05 0.38 (0.29, 0.50)Pooled IL-6ra, No mechanical ven la on at baseline 1.85 0.27 1.83 (1.37, 2.43)Pooled IL-6ra, Mechanical ven la on at baseline 1.30 0.29 1.27 (0.84, 1.94)Fixed-dose cor costeroid 1.54 0.35 1.50 (0.98, 2.34)Shock-dependent cor costeroid 1.07 0.25 1.04 (0.67, 1.65)Pooled IL-6ra*Fixed-dose interac on, No mechanicalven la on at baseline

1.01 0.05 1.01 (0.91, 1.11)

Pooled IL-6ra*Fixed-dose interac on, Mechanical ven la onat baseline

0.99 0.05 0.99 (0.90, 1.09)

Pooled IL-6ra*Shock-dependent interac on, No mechanicalven la on at baseline

1.00 0.05 1.00 (0.91, 1.10)

Pooled IL-6ra*Shock-dependent interac on, Mechanicalven la on at baseline

1.00 0.05 1.00 (0.91, 1.10)

Pooled IL-6ra*Fixed-dose combina on, No mechanicalven la on at baseline

2.86 0.78 2.75 (1.66, 4.68)

Pooled IL-6ra*Fixed-dose combina on, Mechanicalven la on at baseline

1.99 0.65 1.89 (1.02, 3.50)

Pooled IL-6ra*Shock-dependent combina on, No mechanicalven la on at baseline

1.98 0.57 1.91 (1.11, 3.30)

75

Table 73: Odds ra o parameters for subgroup analysis with differen-al effect by baseline mechanical ven la on status with pooled IL-6ra

interven ons (con nued)

Mean SD Median CrIPooled IL-6ra*Shock-dependent combina on, Mechanicalven la on at baseline

1.41 0.47 1.34 (0.70, 2.50)

Table 74: Posterior probabili es for subgroup analysis with differen aleffect on OSFD endpoint by baseline mechanical ven la on status withpooled IL-6ra interven ons

Posterior ProbabilityPooled IL-6ra is superior to control with no mechanical ven la on at baseline 1.000Pooled IL-6ra is fu le with no mechanical ven la on at baseline (OR < 1.2) 0.002Pooled IL-6ra is superior to control with mechanical ven la on at baseline 0.871Pooled IL-6ra is fu le with mechanical ven la on at baseline (OR < 1.2) 0.392Pooled IL-6ra*Fixed-dose combina on, No mechanical ven la on at baseline OR> 1

1.000

Pooled IL-6ra*Fixed-dose combina on, Mechanical ven la on at baseline OR > 1 0.979Pooled IL-6ra*Shock-dependent combina on, No mechanical ven la on atbaseline OR > 1

0.990

Pooled IL-6ra*Shock-dependent combina on, Mechanical ven la on at baselineOR > 1

0.816

8.5 Post hoc analysiswith differen al effect on in-hospitalmortality by baselinemechanical ven la on

status with pooled IL-6ra interven ons

• Model: Primary analysis dichotomous model

• Factors: Age, sex, site, me, control and pooled IL-6ra interven ons, cor costeroid interven ons (control,

fixed-dura on, Shock-dependent) and an viral interven ons (control, HCQ, Kaletra, Kaletra + HCQ), and in-

terac ons between pooled IL-6ra interven ons, cor costeroids, and an virals

• The treatment effect for pooled IL-6ra interven ons by baseline mechanical ven la on status will be com-

pared to the control arm. A posterior probability of superiority of 99% will be used as a sta s cal trigger for

efficacy for each mechanical ven la on subgroup.

• Popula on: Unblinded ITT

76

Table 75: Odds ra o parameters for subgroup analysis with differen aleffect on in-hospital mortality by baselinemechanical ven la on statuspooled IL-6ra interven ons

Mean SD Median CrIAge<39 14.10 7.16 12.53 (5.44, 32.37)Age 40-49 3.18 0.85 3.07 (1.86, 5.19)Age 50-59 3.04 0.62 2.98 (2.03, 4.41)Age 70-79 0.41 0.07 0.40 (0.28, 0.57)Age 80+ 0.21 0.06 0.20 (0.11, 0.36)Female 1.19 0.19 1.18 (0.87, 1.60)Time epoch 1 0.94 0.08 0.93 (0.78, 1.10)Time epoch 2 0.86 0.14 0.86 (0.60, 1.15)Time epoch 3 0.85 0.17 0.84 (0.54, 1.22)Time epoch 4 0.88 0.21 0.85 (0.53, 1.34)Time epoch 5 0.91 0.24 0.88 (0.53, 1.48)Time epoch 6 0.93 0.26 0.90 (0.53, 1.55)Time epoch 7 0.95 0.26 0.92 (0.53, 1.56)Time epoch 8 0.99 0.27 0.96 (0.57, 1.63)Time epoch 9 1.06 0.28 1.02 (0.62, 1.72)Time epoch 10 1.14 0.31 1.09 (0.67, 1.86)Time epoch 11 1.17 0.31 1.13 (0.69, 1.91)Time epoch 12 1.18 0.31 1.14 (0.70, 1.88)Time epoch 13 1.22 0.33 1.18 (0.70, 1.98)Time epoch 14 1.36 0.41 1.31 (0.74, 2.32)Time epoch 15 1.63 0.64 1.50 (0.74, 3.20)Time epoch 16 2.12 1.45 1.76 (0.70, 5.53)Mechanical ven la on at baseline 0.43 0.09 0.42 (0.28, 0.62)Pooled IL-6ra, No mechanical ven la on at baseline 1.91 0.40 1.87 (1.26, 2.81)Pooled IL-6ra, Mechanical ven la on at baseline 1.57 0.45 1.51 (0.88, 2.63)Fixed-dose cor costeroid 1.10 0.33 1.05 (0.59, 1.89)Shock-dependent cor costeroid 1.25 0.40 1.19 (0.64, 2.20)Pooled IL-6ra*Fixed-dose interac on, No mechanicalven la on at baseline

1.00 0.05 1.00 (0.91, 1.10)

Pooled IL-6ra*Fixed-dose interac on, Mechanical ven la onat baseline

0.99 0.05 0.99 (0.90, 1.09)

Pooled IL-6ra*Shock-dependent interac on, No mechanicalven la on at baseline

1.00 0.05 1.00 (0.91, 1.11)

Pooled IL-6ra*Shock-dependent interac on, Mechanicalven la on at baseline

1.00 0.05 1.00 (0.91, 1.11)

Pooled IL-6ra*Fixed-dose combina on, No mechanicalven la on at baseline

2.09 0.76 1.97 (0.99, 3.91)

Pooled IL-6ra*Fixed-dose combina on, Mechanicalven la on at baseline

1.71 0.73 1.58 (0.71, 3.56)

Pooled IL-6ra*Shock-dependent combina on, No mechanicalven la on at baseline

2.39 0.93 2.21 (1.07, 4.64)

77

Table 75: Odds ra o parameters for subgroup analysis with differen aleffect on in-hospital mortality by baselinemechanical ven la on statuspooled IL-6ra interven ons (con nued)

Mean SD Median CrIPooled IL-6ra*Shock-dependent combina on, Mechanicalven la on at baseline

1.98 0.91 1.79 (0.77, 4.25)

Table 76: Posterior probabili es for subgroup analysis with differen aleffect on in-hospital mortality by baselinemechanical ven la on statuswith pooled IL-6ra interven ons

Posterior ProbabilityPooled IL-6ra is superior to control with no mechanical ven la on at baseline 0.999Pooled IL-6ra is fu le with no mechanical ven la on at baseline (OR < 1.2) 0.014Pooled IL-6ra is superior to control with mechanical ven la on at baseline 0.934Pooled IL-6ra is fu le with mechanical ven la on at baseline (OR < 1.2) 0.204Pooled IL-6ra*Fixed-dose combina on, No mechanical ven la on at baseline OR> 1

0.973

Pooled IL-6ra*Fixed-dose combina on, Mechanical ven la on at baseline OR > 1 0.872Pooled IL-6ra*Shock-dependent combina on, No mechanical ven la on atbaseline OR > 1

0.985

Pooled IL-6ra*Shock-dependent combina on, Mechanical ven la on at baselineOR > 1

0.911

78