Supplementary Appendix

Transcription

1 Supplementary Appendix This appendix has been provided by the authors to give readers additional information about their work. Supplement to: Perkins GD, Ji C, Deakin CD, et al. A randomized trial of epinephrine in out-of-hospital cardiac arrest. N Engl J Med. DOI: /NEJMoa

2 Electronic supplementary material Table of contents PARAMEDIC2 collaborators... 2 System description (based on Utstein 2014 recommendations) Resuscitation Guidelines... 5 Ethical and legal considerations Patient and public involvement Sample size and other statistical considerations Figure S1: Patient flow of the trial from recruitment to hospital discharge Figure S2: Kaplan Meier survival curve for patients 30 days survival status by treatment arm Figure S3: Bayesian analysis for 30 day survival Figure S4: Bayesian analysis for survival with favorable neurological outcome at hospital discharge 28 Figure S5: ICU free stay (days) in 30 days post randomisation by treatment arm Figure S6: Hospital free stay (days) in 30 days post randomisation by treatment arm Figure S7 Sub-group analyses Table S1: Concurrent pre-hospital treatments Table S2: Distribution of mrs at discharge Table S3: Distribution of mrs at 3 months Table S4: Sensitivity analyses analysis of survival at 30 days, survival to hospital discharge and survival to discharge with good neurological outcome Page 1 of 39

3 PARAMEDIC2 collaborators PARAMEDIC 2 collaborators: Matthew Cooke, Sarah Lamb, Andrew Carson (RIP), Ian Jacobs (RIP), Ed England, Nicola Brock, Claire Godfrey, Sarah Taylor, Michelle Thomson, Isabel Rodriguez-Bachiller, Claire King,, Johanna Lazarus, Helen Werts, Joshua Golding, Alex Boda, Richard Whitfield, Laura Galligan, Rob Lovett, Jennifer Bradley, Gill Price, Andy Rosser, Garry Parcell, Mindy Jhamat, Josh Miller, Jenny Sears Brown, Alice Pretty, Emma Harris, Jenny Lumley-Holmes, Rhiannon Boldy, Prudence Horwood, Sonia Byers, Gary Shaw, Matt Limmer, Craig Wynne, Michelle Jackson, Emma Bell, Oliver Gupta, Rima Gupta, Susie Hennings, Jessica Horton, James Buck, Sarah Rumble, Hayley Johnson, Eva Kritzer, Chockalingham Muthiah, Adrian Willis, Claire Daffern, Louise Clarkson, Felix Achana, Nicola Cashin, Emma Skilton, Malvenia Richmond, Martin Underwood, Natalie Strickland, Sarah Duggan, Mike Smyth, Marie Stevens; Trial Steering Committee (Jon Nicholl, Neil Bayliss, Helen Snooks, Jonathan Benger, Robert Andrews, David Pitcher); Data Monitoring Committee (Marion Campbell, Jasmeet Soar, Kathy Rowan, Sue Mason). We would also like to thank collaborators at all receiving hospitals, all staff involved at participating ambulance services and our patient and public partners. Page 2 of 39

4 System description (based on Utstein 2014 recommendations). The trial was conducted amongst 5 National Health Service Ambulance Services. North East Ambulance Service NHS Foundation Trust South Central Ambulance Service NHS Foundation Trust London Ambulance Service NHS Trust Welsh Ambulance Service NHS Trust West Midlands Ambulance Service NHS Foundation Trust These Ambulance Services serve a mix of urban and rural locations in England and Wales covering a population of 24 million people. They attend approximately 32,000 cases of cardiac arrest each year of which resuscitation is attempted or continued by ambulance staff in approximately 45% of cases. The Ambulance Services are activated through a central emergency call number (999, 112 or 111) which directs the caller to the geographically relevant emergency operations and dispatch centre. Calls are received and processed by trained NHS Ambulance Dispatch staff. Call handlers use one of two dispatch support systems NHS Pathways or Advanced Medical Priority Dispatch System. Cases identified as a cardiac arrest are assigned the highest priority response. At the time of the study, Ambulance Services are commissioned to reach 75% of these cases with a defibrillator capable response within 8 min and to have an ambulance on scene within 19 min in 95% of cases. The nearest available resource(s) which may be a community responder, car, ambulance, motorbike or helicopter is dispatched to the scene. Paramedics or emergency medical technicians, either or both of which may be assigned to a cardiac arrest case, staff ambulance vehicles. Paramedics can deliver advanced life support (ALS) interventions (including advanced airway management and intravenous drugs) and after training were able to recruit patients to the trial. Technicians, and many community responders dispatched by the NHS ambulance service, can deliver CPR and defibrillation, and some use supraglottic airways. Upon arrival at a cardiac arrest, EMS personnel assess the appropriateness of a full resuscitation attempt. Resuscitation may be with-held if there is Page 3 of 39

5 unequivocal evidence of death (major traumatic injuries, putrefaction, rigor mortis, post mortem staining etc.), a Do not attempt resuscitation (DNAR) decision is documented or if there is asystole, no bystander CPR with more than 15 min having elapsed from the time of collapse. When resuscitation is attempted, EMS follow national Resuscitation Council (UK) guidelines or regional clinical guidelines derived from these national guidelines, which in turn, are derived from the European Resuscitation Council guidelines. The overall UK survival to discharge rates for all cases of out of hospital cardiac arrest where resuscitation is attempted is 7.9%. 1 Page 4 of 39

6 Resuscitation Guidelines Resuscitation algorithms were based on the Resuscitation Council (UK) guidelines, which are informed by the recommendations of the European Resuscitation Council and International Committee on Resuscitation. Reproduced with permission from the Resuscitation Council (UK). Page 5 of 39

7 Treatment of shockable rhythms (VF/VT) [Resuscitation Council UK] 1. Confirm cardiac arrest check for signs of life and normal breathing, and if trained to do so check for breathing and a pulse simultaneously. 2. Call resuscitation team. 3. Perform uninterrupted chest compressions while applying self-adhesive defibrillation/monitoring pads one below the right clavicle and the other in the V6 position in the midaxillary line. 4. Plan actions before pausing CPR for rhythm analysis and communicate these to the team. 5. Stop chest compressions; confirm VF/pVT from the ECG. This pause in chest compressions should be brief and no longer than 5 seconds. 6. Resume chest compressions immediately; warn all rescuers other than the individual performing the chest compressions to stand clear and remove any oxygen delivery device as appropriate. 7. The designated person selects the appropriate energy on the defibrillator and presses the charge button. Choose an energy setting of at least 150 J for the first shock, the same or a higher energy for subsequent shocks, or follow the manufacturer s guidance for the particular defibrillator. If unsure of the correct energy level for a defibrillator choose the highest available energy. 8. Ensure that the rescuer giving the compressions is the only person touching the patient. 9. Once the defibrillator is charged and the safety check is complete, tell the rescuer doing the chest compressions to stand clear ; when clear, give the shock. 10. After shock delivery immediately restart CPR using a ratio of 30:2, starting with chest compressions. Do not pause to reassess the rhythm or feel for a pulse. The total pause in chest compressions should be brief and no longer than 5 seconds. 11. Continue CPR for 2 min; the team leader prepares the team for the next pause in CPR. 12. Pause briefly to check the monitor. 13. If VF/pVT, repeat steps 6 12 above and deliver a second shock. 14. If VF/pVT persists, repeat steps 6 8 above and deliver a third shock. Resume chest compressions immediately. Give epinephrine 1 mg IV and amiodarone 300 mg IV while performing a further 2 min CPR. Withhold epinephrine if there are signs of return of spontaneous circulation (ROSC) during CPR. 15. Repeat this 2 min CPR rhythm/pulse check defibrillation sequence if VF/pVT persists. 16. Give further epinephrine 1 mg IV after alternate shocks (i.e. approximately every 3 5 min). 17. If organised electrical activity compatible with a cardiac output is seen during a rhythm check, seek evidence of ROSC (check for signs of life, a central pulse and end-tidal CO 2 if available). 1. If there is ROSC, start post-resuscitation care. 2. If there are no signs of ROSC, continue CPR and switch to the non-shockable algorithm. 18. If asystole is seen, continue CPR and switch to the non-shockable algorithm. The interval between stopping compressions and delivering a shock must be minimised. Longer interruptions to chest compressions reduce the chance of a shock restoring a spontaneous Page 6 of 39

8 circulation. Chest compressions are resumed immediately after delivering a shock (without checking the rhythm or a pulse) because even if the defibrillation attempt is successful in restoring a perfusing rhythm, it is very rare for a pulse to be palpable immediately after defibrillation. The duration of asystole before ROSC can be longer than 2 min in as many as 25% of successful shocks. If a shock has been successful immediate resumption of chest compressions does not increase the risk of VF recurrence. 12 Furthermore, the delay in trying to palpate a pulse will further compromise the myocardium if a perfusing rhythm has not been restored. The use of waveform capnography can enable ROSC to be detected without pausing chest compressions and may be used as a way of avoiding a bolus injection of epinephrine after ROSC has been achieved. Several human studies have shown that there is a significant increase in end-tidal CO 2when ROSC occurs. If ROSC is suspected during CPR withhold epinephrine. Give epinephrine if cardiac arrest is confirmed at the next rhythm check. Regardless of the arrest rhythm, after the initial epinephrine dose has been given, give further doses of epinephrine 1 mg every 3 5 min until ROSC is achieved; in practice, this will be about once every two cycles of the algorithm. If signs of life return during CPR (e.g. purposeful movement, normal breathing or coughing), or there is an increase in end-tidal CO 2, check the monitor; if an organised rhythm is present, check for a pulse. If a pulse is palpable, start post-resuscitation care. If no pulse is present, continue CPR. Give amiodarone 300 mg IV after three defibrillation attempts irrespective of whether they are consecutive shocks, or interrupted by CPR, or for recurrent VF/pVT during cardiac arrest. Consider a further dose of amiodarone 150 mg IV after a total of five defibrillation attempts. Lidocaine 1 mg kg - 1 may be used as an alternative if amiodarone is not available but do not give lidocaine if amiodarone has been given already. Treatment of PEA and asystole 1. Start CPR 30:2 2. Give epinephrine 1 mg IV as soon as intravascular access is achieved 3. Continue CPR 30:2 until the airway is secured then continue chest compressions without pausing during ventilation 4. Recheck the rhythm after 2 min: a. If electrical activity compatible with a pulse is seen, check for a pulse and/or signs of life i. If a pulse and/or signs of life are present, start post resuscitation care ii. If no pulse and/or no signs of life are present (PEA OR asystole): 1. Continue CPR 2. Recheck the rhythm after 2 min and proceed accordingly 3. Give further epinephrine 1 mg IV every 3 5 min (during alternate 2-min loops of CPR) b. If VF/pVT at rhythm check, change to shockable side of algorithm. Whenever a diagnosis of asystole is made, check the ECG carefully for the presence of P waves because the patient may respond to cardiac pacing when there is ventricular standstill with continuing P waves. There is no value in attempting to pace true asystole. Page 7 of 39

9 Discontinuing Resuscitation When to stop CPR In most patients where ROSC is not achieved on scene, despite appropriate ALS and treatment of any potentially reversible causes, little is to be gained from transferring these patients to hospital. Conveying a patient in cardiac arrest to hospital is not easy, through the logistics of having to move a patient down stairs, off the floor or into an ambulance, each of which may cause an interruption in chest compressions as well as the risks associated with manual handling. Conveying a patient in a vehicle traveling under emergency conditions is not without risk to the patient and the clinical team. Asystole for more than 20 min The UK Ambulance Service Clinical Practice Guidelines (2013) state that where a patient has had persistent asystole for more than 20 min, despite ALS, and where drowning, hypothermia, poisoning or overdose, and pregnancy have been excluded, it is appropriate for the resuscitation attempt to be stopped. Pulseless electrical activity The decision about when to stop a cardiac arrest where pulseless electrical activity (PEA) persists is less clear and is not currently within the UK Ambulance Service Clinical Practice Guidelines (2013). The reported survival to discharge rate for PEA is very low (one UK registry reported this to be 4.2 %). There is limited evidence to support when one should terminate a PEA cardiac arrest, but the length of time in arrest without life support, the absence of reversible causes and co-morbidities are important factors to consider when making this decision. Ventricular fibrillation Where practical, transport patients with persistent VF or pvt to a cardiac arrest centre with ongoing CPR, because further in-hospital treatment may occasionally be successful. Page 8 of 39

10 Post resuscitation care Page 9 of 39

11 Ethical and legal considerations Context What happens to someone when they sustain a cardiac arrest? A cardiac arrest is the medical term used to describe sudden cessation of the heartbeat. Outside of a hospital cardiac arrest is usually a sudden, unexpected event. When cardiac arrest occurs blood stops circulating around the body and consciousness and mental capacity are lost within seconds. Treatment must be started urgently every second that treatment is delayed is associated with less chance of survival. Treatment comprises combinations of CPR (chest compressions and ventilation), electric shocks and other advanced treatments. Treatment is usually continued for up to 20 minutes. After this time window the chances of survival are very small. If the heart is re-started the person is taken to hospital. Most people that survive initially are unconscious when they arrive in hospital and are admitted to intensive care. Unfortunately many people die within the first 24 hours of admission to hospital. People surviving more than 24 hours are treated in intensive care for around 4-6 days. Most remain unconscious for the majority of the time in intensive care. People that recover (about 1 in 20 of those initially treated by the ambulance service) are discharged to the hospital ward. Most people will have regained capacity by this point but some will have sustained brain damage and may never regain capacity. Footnote: Cardiac arrest is different from the term heart attack. A heart attack refers to the blockage of one of the blood vessels supplying the heart with blood. Although some heart attacks may progress to cardiac arrest most do not. Most people survive a heart attack, whilst few survive a cardiac arrest. Page 10 of 39

12 Legal considerations The trial was run in accordance with the Directive 2001/20/EC of the European Parliament and of the Council of 4 April 2001, The Medicines for Human Use Act (Clinical Trial) Regulations, Statutory Instrument 2004 No 1031 and Amendment (No.2) Statutory Instrument 2006 No The Regulations allow trials involving incapacitated adults in emergency situations: Where the treatment to be given to an incapacitated adult as part of the trial needs to be given urgently, time may not allow for the written consent of a legal representative to be obtained first. The regulations allows incapacitated adults to be entered into a trial prior to consent being obtained from a legal representative provided that: Having regard to the nature of the trial and the particular circumstances of the case, it is necessary to take action for the purpose of the trial as a matter of urgency but It is not reasonably practicable to obtain informed consent prior to entering the subject, and The action to be taken is carried out in accordance with a procedure approved by the ethics committee. Ethical considerations Ethical considerations for the trial were framed around the guidance from the Health Research Authority Workshop 2012 on conducting emergency research in patients who lack capacity. Is this research needed and is there uncertainty about treatment? The Resuscitation Council (UK), Association of Ambulance Medical Directors and International Liaison Committee on Resuscitation identified the need to conduct a placebo controlled trial of adrenaline use in cardiac arrest. The trial protocol summarises the current evidence about the use of adrenaline as a treatment for cardiac arrest. In brief there was some evidence that adrenaline leads to better short-term survival. Whether it improves long-term survival and how it affects brain function after cardiac arrest was uncertain. Whilst some studies suggested it may be beneficial, others suggested it may be Page 11 of 39

13 ineffective or potentially harmful. Our survey of UK clinician s views about the effectiveness of adrenaline showed substantial uncertainty amongst the clinical community (see protocol figure 2a). Is there a need to recruit participants who lack capacity? The clinical trial related directly to the treatment of cardiac arrest, which is a life-threatening emergency. All patients that suffer a cardiac arrest will lack capacity. There are no alternative groups of patients amongst whom this research could have been be conducted. In the context of the research is consent or consultation feasible? The occurrence of an out of hospital cardiac arrest is unpredictable. Within seconds of cardiac arrest a person becomes unconscious and thus incapacitated. It was therefore possible to obtain prospective consent directly from the research participant. It was not practical to consult with a personal legal representative without delaying immediately necessary. The focus of the paramedics attending the incident must be on the patient. Does treatment need to be given quickly and might delay change the effect of treatment or the results? Treatment (in the form of CPR) must be started immediately in an attempt to save the person s life. Delay in the initiation of CPR and other emergency treatment (e.g. defibrillation) is associated with worse outcomes. Observational studies suggest that if adrenaline is effective, the earlier it is given the more likely it will be beneficial. Will procedures accommodate variations in capacity? All patients will lack capacity throughout the intervention period of the trial due to the nature of the underlying medical condition (cardiac arrest). Is it practical to consult a professional legal representative unconnected to the research? Page 12 of 39

14 In this setting it is practical to consult a carer or independent registered medical practitioner without placing the potential participant at risk of harm from delaying emergency treatment. What should the patient or legal representative be asked later? We sought consent to continue in the trial. Our rationale was that the patient would be enrolled in the trial at the time of their cardiac arrest if they met the eligibility criteria. The duration of the research intervention (adrenaline or placebo) would typically last for 5-20 minutes and rarely more than 45 minutes. Adrenaline is rapidly metabolised by the body after administration with a half-life duration of 5-10 minutes (time it takes for the body to remove half the drug). The intervention phase of the trial was therefore be completed within an hour of the patient sustaining a cardiac arrest. We approached patients or their legal representative, as soon as practicable after the initial emergency had passed to inform them of their participation and request consent to continue using the steps outlined below. Provision of general information about the trial General information about the trial and contact details for further information was made freely available throughout the trial. Information about the trial was placed on Ambulance Trust and University websites, Ambulance Service Public Newsletters, Posters / information leaflets were shared with GP practices, pharmacies, Hospital A&E departments and waiting areas, totaling 6,500 mail-outs, including 2,356 GP Practices and 3,181 pharmacies. Prior to the start of the trial a press release was issued providing information about the trial. This was followed by periodic regional press releases as the trial progressed. The trial website was updated with information throughout the trial ( which was accessed 178,000 times during the trial. Although not required by the relevant regulations, the trial team developed a system to allow a patient to decline participation in the trial in the event that they sustain a cardiac arrest. Page 13 of 39

15 Requests not to participate were sent to and managed by the WCTU Trial Team. An online form could be completed on the website or the team could be contacted by phone or . A stainless steel No Study bracelet was issued to the person s home address and with the person s permission their home address was passed to the ambulance service to register the person s wishes. Those requesting a no study bracelet were also told to tell those close to them their wishes and that those wishes will be respected by the treating paramedics. Paramedics were trained to look for the bracelet. Informing the patient about participation in the trial: The first attempt to contact the patient and inform them of their enrolment into the trial was during their stay in hospital before hospital discharge. The clinical team made contact as soon as practically possible after the initial emergency had passed taking the utmost care and sensitivity in doing so. Following our experience from a 4,400 patient study in out of hospital cardiac arrest (PARAMEDIC trial) and talking to fellow researchers from the REVIVE AIRWAYS cardiac arrest study and discussions with patient and public representatives, we believed the earliest practicable time to approach patients and relatives was once the patient was discharged from ICU and was on a hospital ward. This allowed sufficient time for the research team to be made aware of enrolment, identify who the patient is, check which hospital the patient was transferred to, whether they were still alive and to verify with the hospital team where the patient was within the hospital. Transfer to a ward indicated that the initial emergency had passed and the patient s condition had stabilised. It was also considered more likely that the patient had regained consciousness and it would avoid any confusion or additional distress of making an approach while the patient remained critically ill in intensive care. Procedure The Research Paramedic, or hospital team assessed if the patient had capacity to consent. If the patient had capacity, they were provided with the information sheet explaining the trial and the options for their involvement. The patient was allowed time to consider the information provided, have the opportunity to ask questions and discuss with others. The Research Paramedic Page 14 of 39

16 or hospital team then asked when the patient would like someone to come back to discuss participation further and potentially take consent. The patient could decide it was not an appropriate time to discuss the trial or they could decide that they did not want to be involved in which case their feelings were respected and their decision about continuing in the trial was recorded. Participants who lack capacity In the event that a patient lacked capacity to consent, the Research Paramedic worked with the hospital team to identify a legal representative as defined below: Personal Legal Representative: A person independent of the trial, who by virtue of their relationship with the potential study participant is suitable to act as their legal representative for the purposes of that trial, and who is available and willing to so act for those purposes. Or if there is no such person: Professional Legal Representative: A person independent of the trial, who is the doctor primarily responsible for the medical treatment provided to that adult. Or a person nominated by the relevant healthcare provider The legal representative was approached and provided with the information sheet explaining the trial and the options for theirs and the patient s involvement, including the need for them to give consent on behalf of the patient and complete questionnaires on behalf of the patient. The legal representative was given time to consider the information provided. The Research Paramedic or hospital team then asked when the legal representative would like someone to come back to discuss participation further and potentially take consent. The legal representative could decide it was not an appropriate time to discuss the trial or they could decide that the patient would not want to take part in which case their feelings were respected and their decision about taking part was recorded. It is possible that the patient could have regained capacity by the time the 3 month visit was due. When contacting the legal representative to arrange the 3 month visit, we asked if we could speak Page 15 of 39

17 with the patient. If on assessment of the patient either on the phone or at the visit was found that the patient still lacked capacity the legal representative was asked to complete the questionnaires on behalf of the patient. If the patient had regained capacity then information was provided about the trial and consent sought. Contact Patient at Home Address: If a patient was discharged from hospital before contact was established an invitation letter and information sheet were posted to their discharge address as soon as possible. If there was no response after 2 weeks the WCTU team would try and contact the patient by phone (if phone number was known) or by a second letter. If the patient wished to be contacted about the trial they could contact the WCTU by phone, or by , or by returning the reply slip. For a small number of patients, if there was a delay between discharge and contacting the patient, to be sure the patient was still alive before writing, the ambulance service would conduct its own checks on patients survival using its own data systems, which differed between services. Where possible, they consulted the NHS Patient Demographics Service, but access may not have been set up in all areas. Other checks carried out by either the ambulance service and or the WCTU trial team included contacts with GPs (where known), hospitals or Health and Social Care Information Service (HSCIC). To ensure we wrote to the correct address, we confirmed the patient address with hospitals, GPs or public access online systems such as 192 before writing. After these checks, if someone was still believed to be alive the WCTU would contact them at their home address by letter, as detailed above. Page 16 of 39

18 Obtaining Consent The research team, or hospital staff where specific R&D approval had been obtained, were fully trained regarding informed consent and assessing capacity, GCP guidelines, relevant legislation and the trial related procedures around consent. Informed consent was taken with the patient or their legal representative on the hospital ward. In exceptional circumstances if consent was not obtained during the hospital stay, and the patient responded to the invitation letter sent by post and agreed to have a home visit, written consent was taken at the 3 month visit. The consent form listed the different sorts of information that were collected. If the patient or legal representative did not want the trial team to continue to collect data about the patient s survival, or to access the patient s health records then they indicated this on the consent form or could tell the trial team verbally at which point the patient was withdrawn. Staff confirmed the patient s willingness to continue with the trial at each contact point. People that do not survive: Key issues: The sad reality of an out of hospital cardiac arrest is that less than 1 in 10 survive. Suffering the sudden unexpected loss of a loved one due to cardiac arrest is a traumatic event that frequently leads to symptoms of anxiety, depression and post-traumatic distress. Careful consideration was given to how, when and if the relatives of non-survivors should be informed about participation in the trial. By the time the patient s death had occurred the trial intervention had been implemented and no further follow up would occur. Thus there was no requirement to or utility in seeking consent to continue. The purpose of any communication with the family/next of kin of the deceased was Page 17 of 39

19 therefore to inform them about the patient s involvement in the trial. Informing the family about the trial ensures that the process of trial recruitment was open and transparent. It would reduce the likelihood that family members would discover at a later date that their relative was involved in a trial without their knowledge. However, knowledge of the trial participation after the event may also place a significant burden on the next of kin at a time of heightened emotional distress due to the loss of their relative or friend. Any strategy to inform family or next of kin following a patient s death needed to carefully balance the need for transparency with the need to minimise their distress. Strategies for informing relatives There are a number of ways in which we could have approached informing the relatives of those that do not survive. These can be broadly categorised as passive or active methods. Passive methods include placing information about the trial in publically accessible places (e.g. websites, newsletters) and targeted sites likely to be attended by relatives of the deceased (e.g. hospitals, GP surgeries, Registrar of Births and Deaths offices, libraries, council websites). Such information would contain brief details about the study and a contact telephone number and address for further information. An advantage of passive methods is that they allow people to make a choice about whether they wish to seek further information and the timing of that approach. The disadvantage of passive methods is that one cannot be certain that relatives of all participants will see them. Discussion with investigators of previous UK trials (e.g. CRASH trials, Brain injury trials) indicates that passive strategies, although not formally evaluated, have been used successfully. Active strategies involve making direct contact with relatives (e.g. posting or hand-delivering a participant information leaflet, organising for a face to face meeting or telephone call). Concerns about the potential burdens to inform recipients and the practicalities of this approach mean that it has not been used in previous UK out of hospital cardiac arrest trials (e.g. LINC [mechanical CPR], Page 18 of 39

20 PARAMEDIC-1 [mechanical CPR], REVIVE [airway device]) [personal communication with Chief Investigators]. We are therefore unable to draw on relatives or researchers experience of this process. There are practical barriers to providing information actively. The sudden and unpredicted nature of cardiac arrest mean that the relatives / next of kin are neither universally present nor identifiable at the time of the cardiac arrest. Information on the identity of the relatives / next of kin are also not held by ambulance services. For people where resuscitation efforts are terminated in the home (approximately 40% of total cases) it is not possible for the paramedic who attends the cardiac arrest to spend the necessary time to explain about the study and answer questions due to the high likelihood that they will be tasked to attend another life threatening incident before the informing process is complete. This approach would also require that every paramedic was trained to discuss the trial with the family in this setting. For patients that die early in hospital (before consent is obtained), there is likely to be a delay in notification to the ambulance service making an in-hospital visit impractical. Given these difficulties with face-to-face consultation an alternative is to send written information by post. Outwith the difficulties described above about the accuracy and completeness of contact details, we had concerns that this un-solicited approach and absence of an opportunity to ask questions immediately upon receipt could exacerbate an already traumatic and stressful experience. Page 19 of 39

21 Advantages of active approaches are that the process of information giving can be more actively monitored. Whether it leads to greater dissemination of information, given the practical difficulties described is unclear. Disadvantages are active approaches remove the relatives choice about whether they wish to receive information about the trial or be reminded about the final stages of the deceased life and the risk that the receipt of such information, without the appropriate support could cause significant additional distress. We carefully considered the benefits and burdens of different approaches to informing the relatives of the deceased, about the trial. Our assessment of the balance of benefits and burdens for relatives was that the burden of actively informing them would outweigh the potential benefit. After consultation with the Research Ethics Committee we decided to inform relatives through passive communication processes described above. We discussed this in detail with our clinical ethicist and patient representatives and have their support for this approach. Enquiries regarding trial participation: If a member of the public wanted to enquire whether their next of kin was enrolled into the PARAMEDIC2 Trial they could complete a Trial Participation Enquiry Form. On receipt of a completed Trial Participation Enquiry Form the PARAMEDIC2 Trial Office worked with the appropriate site to check whether the next of kin of the enquirer was enrolled into the PARAMEDIC2 Trial. Once it was confirmed a letter was sent from the appropriate ambulance service to confirm that the enquirer s next of kin was enrolled into the trial along with an information sheet about the trial, which included contact details for the PARAMEDIC2 Trial Office. Page 20 of 39

22 Patient and public involvement During the planning and development phase we have worked with the PPI members of the PARAMEDIC trial who contributed to the trial design and proposed follow-up systems. Specific contributions related to the selection of outcome measures, summary / presentation of research in plain English. We held a community engagement event (supported by West Midlands South CLRN) in late November 2012 where we presented the scientific rationale behind this trial to a group of 280 lay-people who were interested in first aid. After preparing the talk in collaboration with one of our PPI representatives, to ensure concepts were presented in plain, understandable English, we delivered the presentation and addressed questions / queries from the group. We explained the concept of short term and longer-term outcomes and briefly sought community views about priorities for outcomes and their views on a trial of adrenaline for out of hospital cardiac arrest. Ninety-five percent of respondents prioritized long-term survival with favourable neurological outcome over short-term survival (hours to days). Participants broadly agreed there was a need for further research about adrenaline as a treatment for cardiac arrest (86% agreed, 8% neither agree nor disagreed, 6% disagreed). Page 21 of 39

23 Our PPI partners played key roles in the management of the research. Our PPI co-applicant has extensive experience of working with charities dedicated to reducing death from cardiac arrest was a member of the trial management group and the lead PPI representative. They liaised with our other PPI partners. There were also two PPI representatives as independent members of the Trial Steering Committee. We provided updates on progress to user representative groups during the conduct of the study. In addition the a PPI advisory group comprising of 8 members of the public, gave advice on the content and distribution of patient and public facing documents. The group was put together through the University of Warwick s UNTRAP group and includes a cross section of age, gender, religion and experience. The trial PPI group met face to face at the University of Warwick for the 1 st meeting before the start of the pilot (August 2014) where we sought their views on the general concept of the research in this area and to ensure the patient facing documents were sensitive, clear and understandable. We also sought their advice on the trial communication strategy (ways in which information about the trial can be shared within relevant communities) and wording of the press release to reflect the key messages are understandable and clear. Documents were sent out to the group members by , before the meeting, for them to review. Several of the comments made were taken forwards and were fed back to the group at the second meeting in May 2015, towards the end of the pilot. This second meeting was to discuss the revised communication strategy, the new posters and leaflets and the website in order for us to ensure the key messages are clear. We continued to meet at least annually thereafter and on an ad hoc basis when a review or consultation is required. Page 22 of 39

24 Sample size and other statistical considerations In the original protocol we reported that estimates of long-term survival of patients who receive epinephrine during a resuscitation attempt vary between about 3.5% and 12%. 2-6 The trial s primary aim was to estimate the treatment effect of epinephrine and the uncertainty around this; we therefore based the target sample size primarily on the precision of the estimate of the risk ratio. The target sample size was set as 8,000, which was expected to give a width of the 95% CI of approximately 0.4 or slightly less. With this target sample size, if the risk ratio for the Epinephrine group was estimated to be 1.25, the corresponding 95% confidence interval would range from 1.07 to A risk ratio of 1.25 corresponds to a rate of 30-day survival of 6.0% in the placebo group and 7.5% in the epinephrine group. In August 2016, the DMC reported that the overall survival rate was lower than anticipated in the original sample size estimate. The implications of this observation on the sample size estimates was considered. It was determined that to maintain a risk ratio of 1.25 would require a sample size of 24,000. This was considered impractical and too costly. It was noted that if the trial continued without any sample size adjustment (i.e. sample of 8000 patients), the trial would have 90% to detect a RR of 1.6% and absolute risk reduction of 1.1%. A threshold of 1% in absolute risk reduction for outcomes has been used widely in resuscitation science as the threshold that defines the minimally important clinical difference. 7,8 The TSC, DMC and funder endorsed continuation of the trial with the sample size of The Data Monitoring Committee performed interim reviews every 3 months. Lan-DeMets O Brien Fleming and Pocock alpha spending methods were used to form the upper and lower stopping boundaries for the primary outcome with the final analysis unadjusted. Page 23 of 39

25 Bayesian analysis We performed unadjusted Bayesian analyses for two outcomes, 30-day survival and survival with favourable neurological outcome, to enable us to quantify the probability of treatment effects of different sizes. We modelled the epinephrine and placebo arms separately, with the outcome for each participant being a Bernoulli (0/1) variable, with treatment arm-specific probabilities of success. Non-informative beta (1, 1) priors were used for both groups, initially. We also performed a sensitivity analysis assuming more realistic, informative, priors. These were beta (5, 150) for the epinephrine group, reflecting previous knowledge that survival in the epinephrine group was very unlikely to exceed 10%, and was most likely to be around 3%, and beta (2,20) for the placebo group, because a survival rate exceeding 20% in the placebo group was very unlikely. Beta (1,1) Beta(2,20) Beta(5,150) We calculated 95% and 80% highest density intervals, and, for risk differences, the probability that the treatment effect exceeded 0, 1% and 2% (which have been suggested as representing clinically important differences). We present the results for the analyses of risk differences for both outcomes with the noninformative prior. The informative priors made only very minor differences to the results. Page 24 of 39

26 Figure S1: Patient flow of the trial from recruitment to hospital discharge Note: Percentages of transportation to in-hospital status are based on enrolled patients in each arm. *: 213 and 391 patients are included in Placebo and Epinephrine boxes, respectively, due to unknown ICU/ward admission information. Page 25 of 39

27 Figure S2: Kaplan Meier survival curve for patients 30 days survival status by treatment arm Page 26 of 39

28 Figure S3: Bayesian analysis for 30 day survival Page 27 of 39

29 Figure S4: Bayesian analysis for survival with favorable neurological outcome at hospital discharge Page 28 of 39

30 Figure S5: ICU free stay (days) in 30 days post randomisation by treatment arm Note: *: Mann-Whitney test is used Page 29 of 39

31 Figure S6: Hospital free stay (days) in 30 days post randomisation by treatment arm Note: *: Mann-Whitney test is used Page 30 of 39

32 Figure S7 Sub-group analyses (1) Unadjusted analyses (categorical variables) Analysis of bystander CPR was carried out in unwitnessed and bystander witnessed cases only. There were no interactions for continuous variables age, ambulance response time, time to drug administration. Page 31 of 39

33 (2) Adjusted analyses (categorical variables) Analysis of bystander CPR was carried out in unwitnessed and bystander witnessed cases only. There were no interactions for continuous variables age, ambulance response time, time to drug administration. None of the overall interaction with treatment in each subgroup analysis was statistically significant. The analyses were for exploratory purpose and the p values were not adjusted for multiplicity. Hence, the results should be interpreted with caution. Page 32 of 39

34 Table S1: Concurrent pre-hospital treatments Epinephrine (N = 4015) Placebo (N = 3999) Intra Venous access 2739 (68.2%) 2763 (69.1%) Intra Venous access Unobtainable 5 (0.1%) 3 (0.1%) Unknown 76 (1.9%) 66 (1.7%) Intraosseous access 1340 (33.4%) 1319 (33.0%) Intraosseous access Unobtainable 5 (0.1%) 4 (0.1%) Unknown 76 (1.9%) 64 (1.6%) Number of trial drug doses Mean (standard deviation) 4.9 (2.5) 5.1 (2.3) Missing 7 9 Administration of Amiodarone 584 (14.5%) 368 (9.2%) Administration of Amiodarone Unobtainable 11 (0.3%) 14 (0.4%) Unknown 47 (1.2%) 39 (1.0%) Supraglottic airway used 2844 (70.8%) 2847 (71.2%) Supraglottic airway used Unobtainable 5 (0.1%) 3 (0.1%) Unknown 176 (4.4%) 164 (4.1%) Tracheal tube used 1206 (30.0%) 1125 (28.1%) Tracheal tube used Unobtainable 5 (0.1%) 4 (0.1%) Unknown 167 (4.2%) 157 (3.9%) Number of shocks after randomisation N Mean (standard deviation) 1.4 (3.0) 0.9 (2.4) Page 33 of 39

35 Compression rate in the first 5 minutes (per minute) N Mean (standard deviation) (14.4) (13.3) Compression fraction in the first 5 minutes (%) N Mean (standard deviation) 76.2 (11.2) 78.4 (13.0) : Compression rate is derived by only including compressions in the periods of uninterrupted compression (no pause or less than 3 seconds). : Compression ratio is defined as the proportion of time of uninterrupted chest compression. Page 34 of 39

36 Table S2: Distribution of mrs at discharge Epinephrine Placebo Odds ratio (95% Confidence Interval) Unadjusted Adjusted 0 - No symptoms 12 (0.3%) 15 (0.4%) (0.373, 1.704) (0.358, 1.769) 1 - No significant disability 17 (0.4%) 10 (0.3%) (0.677, 1.980) (0.658, 2.029) 2 - Slight disability 23 (0.6%) 29 (0.7%) (0.654, 1.407) (0.651, 1.476) 3 - Moderate disability 35 (0.9%) 20 (0.5%) (0.860, 1.608) (0.878, 1.727) 4 - Moderately severe 12 (0.3%) 8 (0.2%) (0.899, 1.625) (0.914, 1.745) disability 5 - Severe disability 27 (0.7%) 8 (0.2%) (1.071, 1.852) (1.115, 2.039) 6 - Dead 3881 (96.9%) 3904 (97.7%) - - Page 35 of 39

37 Table S3: Distribution of mrs at 3 months Epinephrine Placebo Odds ratio (95% Confidence Interval) Unadjusted Adjusted 0 - No symptoms 20 (0.5%) 20 (0.5%) (0.623, 2.325) 1 - No significant disability 30 (0.8%) 20 (0.5%) (0.889, 2.184) 2 - Slight disability 10 (0.3%) 11 (0.3%) 3 - Moderate disability 22 (0.6%) 12 (0.3%) (0.979, 1.801) (0.845, 1.900) (0.986, 2.025) 4 - Moderately severe disability 6 (0.2%) 5 (0.1%) (0.992, 1.990) 5 - Severe disability 10 (0.3%) 6 (0.2%) (1.038, 2.020) 6 - Dead 3888 (97.5%) 3905 (98.1%) - - Page 36 of 39

38 Table S4: Sensitivity analyses analysis of survival at 30 days, survival to hospital discharge and survival to discharge with good neurological outcome 30 days survival status Survived to hospital discharge 90 days survival status 180 days survival status Survival with good neurological outcome at discharge (mrs<=3) Survival with good neurological outcome at 3 months (mrs<=3) Scenario* Epinephrine Placebo Difference (95% Confidence Interval) Unadjusted odds ratio (95% Confidence Interval) A 133/4015 (3.3%) 98/3999 (2.5%) 0.9% (0.1%, 1.6%) (1.047, 1.777) B 130/4015 (3.2%) 94/3999 (2.4%) 0.9% (0.2%, 1.6%) (1.062, 1.819) C 130/4015 (3.2%) 98/3999 (2.5%) 0.8% (0.1%, 1.5%) (1.021, 1.738) D 133/4015 (3.3%) 94/3999 (2.4%) 1.0% (0.2%, 1.7%) (1.089, 1.860) E (1.055, 1.807) A 134/4015 (3.3%) 95/3999 (2.4%) 1.0% (0.2%, 1.7%) (1.087, 1.852) B 128/4015 (3.2%) 91/3999 (2.3%) 0.9% (0.2%, 1.6%) (1.077, 1.857) C 128/4015 (3.2%) 95/3999 (2.4%) 0.8% (0.1%, 1.5%) (1.034, 1.771) D 134/4015 (3.3%) 91/3999 (2.3%) 1.1% (0.3%, 1.8%) (1.132, 1.942) E (1.075, 1.853) A 127/4015 (3.2%) 94/3999 (2.4%) 0.8% (0.1%, 1.5%) (1.036, 1.778) B 121/4015 (3.0%) 86/3999 (2.2%) 0.9% (0.2%, 1.6%) (1.069, 1.871) C 121/4015 (3.0%) 94/3999 (2.4%) 0.7% (-0.0%, 1.4%) (0.982, 1.696) D 127/4015 (3.2%) 86/3999 (2.2%) 1.0% (0.3%, 1.7%) (1.126, 1.961) E (1.068, 1.870) A 126/4015 (3.1%) 93/3999 (2.3%) 0.8% (0.1%, 1.5%) (1.037, 1.785) B 117/4015 (2.9%) 85/3999 (2.1%) 0.8% (0.1%, 1.5%) (1.042, 1.834) C 117/4015 (2.9%) 93/3999 (2.3%) 0.6% (-0.1%, 1.3%) (0.957, 1.661) D 126/4015 (3.1%) 85/3999 (2.1%) 1.0% (0.3%, 1.7%) (1.129, 1.971) E (1.041, 1.833) A 95/4015 (2.4%) 79/3999 (2.0%) 0.4% (-0.2%, 1.0%) (0.889, 1.626) B 87/4015 (2.2%) 74/3999 (1.9%) 0.3% (-0.3%, 0.9%) (0.859, 1.607) C 87/4015 (2.2%) 79/3999 (2.0%) 0.2% (-0.4%, 0.8%) (0.808, 1.495) D 95/4015 (2.4%) 74/3999 (1.9%) 0.5% (-0.1%, 1.1%) (0.946, 1.747) E (0.861, 1.610) A 111/4015 (2.8%) 83/3999 (2.1%) 0.7% (0.0%, 1.4%) (1.006, 1.789) B 82/4015 (2.0%) 63/3999 (1.6%) 0.5% (-0.1%, 1.1%) (0.935, 1.814) C 82/4015 (2.0%) 83/3999 (2.1%) -0.0% (-0.7%, 0.6%) (0.723, 1.339) D 111/4015 (2.8%) 63/3999 (1.6%) 1.2% (0.6%, 1.8%) (1.300, 2.428) E (0.935, 1.814) Note: *: A: Best case scenario in both arms best value, alive ( good ), was assigned to all missing survival (mrs) status; B: Worst case scenario in both arms worst value, deceased ( poor ) was assigned to all missing survival (mrs) status; C: Best case scenario in placebo and worst case scenario in epinephrine best value was assigned to all missing data in placebo and worst value assigned in epinephrine; D: Best case scenario in Page 37 of 39

39 epinephrine and worst case scenario in placebo best value was assigned to all missing data in epinephrine and worst value assigned in placebo; E: Multiple imputation assuming missing at random. Page 38 of 39

40 1. Hawkes C, Booth S, Ji C, et al. Epidemiology and outcomes from out-of-hospital cardiac arrests in England. Resuscitation 2017;110: Sherman BW, Munger MA, Foulke GE, Rutherford WF, Panacek EA. High-dose versus standard-dose epinephrine treatment of cardiac arrest after failure of standard therapy. Pharmacotherapy 1997;17: Gueugniaud PY, David JS, Chanzy E, et al. Vasopressin and epinephrine vs. epinephrine alone in cardiopulmonary resuscitation. N Engl J Med 2008;359: Ong ME, Tiah L, Leong BS, et al. A randomised, double-blind, multi-centre trial comparing vasopressin and adrenaline in patients with cardiac arrest presenting to or in the Emergency Department. Resuscitation Bergner M, Bobbitt RA, Carter WB, Gilson BS. The Sickness Impact Profile: development and final revision of a health status measure. Med Care 1981;19: Stiell IG, Hebert PC, Weitzman BN, et al. High-dose epinephrine in adult cardiac arrest. N Engl J Med 1992;327: Schneiderman LJ. Defining Medical Futility and Improving Medical Care. J Bioeth Inq 2011;8: Sasson C, Hegg AJ, Macy M, Park A, Kellermann A, McNally B. Prehospital termination of resuscitation in cases of refractory out-of-hospital cardiac arrest. JAMA 2008;300: Page 39 of 39