Advanced airway placement (ETT vs SGA)

Transcription

1 Advanced airway placement (ETT vs SGA) Among adults who are in cardiac arrest in any setting (P), does tracheal tube insertion as first advanced airway (I), compared with insertion of a supraglottic airway as first advanced airway (C), change ROSC, CPR parameters, development of aspiration pneumonia, Survival with Favorable neurological/functional outcome at discharge, 30 days, 60 days, 180 days AND/OR 1 year, Survival only at discharge, 30 days, 60 days, 180 days AND/OR 1 year (O)? EGTA (I) versus tracheal intubation (C) For the critical outcome of survival to hospital discharge we have identified very low quality evidence (downgraded for very serious concerns about risk of bias and imprecision) from one RCT enrolling 175 OHCAs show no difference between EGTA and tracheal intubation (OR % CI ) [Goldenberg ] Combitube (I) versus tracheal intubation (C) For the critical outcome of survival to hospital discharge we have identified very low quality evidence (downgraded for very serious concerns about risk of bias and imprecision) from one RCT enrolling 173 OHCAs that showed no difference between Combitube and tracheal intubation (OR % CI ) [Rabitsch ] and very low quality evidence from one observational study of 5822 OHCAs that showed no difference between tracheal intubation by paramedics and Combitube insertion by emergency medical technicians (EMTs) (adjusted OR 1.02; 95% CI ) [Cady ]. For the important outcome of ROSC we have identified very low quality evidence from one observational study of 5822 OHCAs that showed no difference between tracheal intubation by paramedics and Combitube insertion by emergency medical technicians (EMTs) (adjusted OR 0.93; 95% CI ). [Cady ]. LMA (I) versus tracheal intubation (C) For the critical outcome of survival to hospital discharge we have identified very low quality evidence from one observational study of 641 OHCAs that showed lower rates of survival to hospital discharge with insertion of an LMA compared with tracheal tube (OR 0.69; 95% CI ) [Shin ]. Supraglottic airways (SGAs: Combitube, LMA, laryngeal tube) versus tracheal intubation For the critical outcome of favourable neurological survival we have identified low quality evidence from one observational study of 5377 OHCAs showing no difference between tracheal intubation and insertion of a SGA (adjusted OR 0.71; 95% CI ) [Kajino 2011 R236], from one observational study of 281,522 OHCAs showing higher rates of favourable neurological outcome between insertion of a SGA and tracheal intubation (OR 1.11; 95% CI ) [Hasegawa ] and from two studies showing higher rates of favourable neurological outcome between tracheal intubation and insertion of a SGA (8701 OHCAs adjusted OR 1.44; 95% CI [McMullan ]) and (10,455 OHCAs adjusted OR 1.40; 95% CI [Wang ]). Supraglottic airways (SGAs: Combitube and LMA) versus tracheal intubation For the important outcome of ROSC we have identified very low quality evidence from one observational study of 713 OHCAs that showed no difference between tracheal intubation and Combitube or LMA insertion by EMTs or emergency lifesaving technicians (ELTs) (OR 0.65; 95% CI ). [Yanagawa ]. Supraglottic airways (SGAs: Esophageal obturator airway and LMA) versus tracheal intubation For the critical outcome of neurologically favourable one-month survival we have identified very low quality evidence from one observational study of 138,248 OHCAs that showed higher rates of neurologically favourable one-month survival with tracheal intubation compared with insertion of an esophageal obturator airway or LMA (OR 0.89; 95% CI ). [Tanabe ] For the critical outcome of one-year survival we have identified very low quality evidence from one observational study of 923 OHCAs that showed no difference in one-year survival with tracheal intubation compared with insertion of an esophageal obturator airway or LMA (OR 0.89; 95% CI ). [Takei ]. For the critical outcome of one-month survival we have identified very low quality evidence from one observation study that showed no difference in one-month survival between tracheal intubation and insertion of an esophageal obturator airway of an LMA (OR 0.75; 95% CI ) [Takei ] and very low quality evidence from another observation study that showed higher one-month survival with tracheal intubation compared with insertion of an esophageal obturator airway of an LMA (OR 1.03; 95% CI ) [Tanabe ] For the important outcome of ROSC we have identified very low quality evidence from one observational study of 923 OHCAs that showed a higher rate of ROSC with tracheal intubation compared with insertion of an esophageal obturator airway or LMA (OR 0.71; 95% CI ). [Takei ]. We suggest using either a supraglottic airway or tracheal tube as the initial advanced airway management during CPR (weak recommendation, very low quality evidence) for out of hospital cardiac arrest. We suggest using either a supraglottic airway or tracheal tube as the initial advanced airway management during CPR 1

2 (weak recommendation, very low quality evidence) for in hospital cardiac arrest. Airway placement (Basic vs Advanced) Among adults who are in cardiac arrest in any setting (P), does insertion of an advanced airway (ETT or supraglottic airway) (I), compared with basic airway (bag mask +/- oropharyngeal airway) (C), change Survival with Favorable neurological/functional outcome at discharge, 30 days, 60 days, 180 days AND/OR 1 year, Survival only at discharge, 30 days, 60 days, 180 days AND/OR 1 year, ROSC, CPR parameters, development of aspiration pneumonia (O)? All advanced airways (I) versus bag-mask (C) For the critical outcome of favourable neurological survival at one month we have identified very low quality evidence (downgraded for very serious concerns about risk of bias and indirectness, and serious concerns about inconsistency) from one observational study of OHCAs showing a lower unadjusted rate of survival with insertion of an advanced airway (tracheal tube, LMA, LT or Combitube) compared with management with a bag-mask (1.1% vs 2.9%; odds ratio [OR], 0.38; 95% CI, )) [Hasegawa ]. When adjusted for all known variables the odds ratio was 0.32 ( ). For the critical outcome of favourable neurological survival to hospital discharge we have identified very low quality evidence (downgraded for very serious concerns about risk of bias and indirectness, and serious concerns about inconsistency) from one observational study of OHCAs showing a lower unadjusted rate of survival with insertion of an advanced airway (tracheal tube, LMA, LT or Combitube) compared with management with a bag-mask (5.3% vs 18.6%; odds ratio [OR], 0.25; 95% CI, )) [McMullan ]. In an analysis of 3398 propensity-matched patients from the same study, the odds ratio for favourable neurological survival at hospital discharge (bag-mask versus advanced airway) adjusted for all variables was 4.19 ( ). Tracheal intubation (I) versus bag-mask (C) For the critical outcome of favourable neurological survival at one month we have identified very low quality evidence (downgraded for very serious concerns about risk of bias and indirectness, and serious concerns about inconsistency) from one observational study of OHCAs showing a lower unadjusted rate of survival with tracheal intubation compared with management with a bag-mask (1.0% vs 2.9%; OR 0.35 ( )) [Hasegawa ]. In an analysis of propensity-matched patients from the same study, the odds ratio for favourable neurological survival at one month (tracheal intubation versus bag-mask) adjusted for all variables was 0.42 ( ). For the critical outcome of favourable neurological survival to hospital discharge we have identified very low quality evidence (downgraded for very serious concerns about risk of bias and indirectness, and serious concerns about inconsistency) from one observational study of 7520 OHCAs showing a lower unadjusted rate of survival with tracheal intubation compared with management with a bag-mask (5.4% vs 18.6%; odds ratio [OR], 0.25; 95% CI, )) [McMullan ]. Supraglottic airways (I) versus bag-mask (C) For the critical outcome of favourable neurological survival at one month we have identified very low quality evidence (downgraded for very serious concerns about risk of bias and indirectness, and serious concerns about inconsistency) from one observational study of OHCAs showing a lower unadjusted rate of survival with insertion of a supraglottic airway (LMA, LT or Combitube) compared with management with a bag-mask (1.1% vs 2.9%; OR 0.38 ( )) [Hasegawa 2013]. In an analysis of propensity-matched patients from the same study, the odds ratio for favourable neurological survival at one month (supraglottic airway versus bag-mask) adjusted for all variables was 0.36 ( ). We suggest using either an advanced airway or a bag mask for airway management during CPR (weak recommendation, very low quality evidence) for out of hospital cardiac arrest. We suggest using either an advanced airway or a bag mask for airway management during CPR (weak recommendation, very low quality evidence) for in hospital cardiac arrest. ETCO2 to predict outcome of cardiac arrest Prognostic Among adults who are in cardiac arrest in any setting (P), does any end-tidal CO2 level value, when present (I), compared 2

3 with any end-tidal CO2 level below that value (C), change (O)? For the important outcome of ROSC we have identified low quality evidence from 3 observational studies enrolling 302 patients (Ahrens 2001, 391; Callaham 1990, 358; Cantineau 1996, 791) showing a correlation with Initial ETCO2 10 mmhg when compared to < 10 mmhg (OR % CI ). For the important outcome of ROSC we have identified low quality evidence from 3 observational studies enrolling 367 patients (Ahrens 2001, 391; Levine 1997, 301; Wayne 1995, 762) showing correlation with 20 min ETCO2 10 mmhg when compared to < 10 mmhg (OR % CI ). For the critical outcome of survival at discharge we have identified low quality evidence from 1 observational study enrolling 127 patients (Ahrens 2001, 391) showing a correlation with Initial ETCO2 10 mmhg when compared to < 10 mmhg (OR % CI ). For the critical outcome of survival at discharge we have identified low quality evidence from 1 observational study enrolling 127 patients (Ahrens 2001, 391) showing a correlation with 20 min ETCO2 20 mmhg when compared to < 20 mmhg (OR % CI ). We did not identify any evidence to address the critical outcome of neurologically intact survival. We suggest that ETCO2 10 mmhg, measured after the intubation or at 20 min of resuscitation, may be a predictor of ROSC (weak recommendation, low quality of evidence). We suggest that ETCO2 10 mmhg, measured after the intubation, or ETCO2 20 mmhg, measured at 20 min of resuscitation, may be a predictor of survival at discharge (weak recommendation, low quality of evidence). Although certain ETCO2 cutoff values appear to be a strong predictor of ROSC and mortality, their utility in accurately predict outcome during CPR is not established. Thus, we recommend against using ETCO2 cutoff values alone as a mortality predictor or on the decision to stop the resuscitation attempt (weak recommendation, low quality of evidence). Fever after cardiac arrest Adults with ROSC after cardiac arrest in any setting (P), does Prevention of fever to maintain strict normothermia (I), compared with No fever control (C), change Survival with Favorable neurological/functional outcome at discharge, 30 days, 60 days, 180 days AND/OR 1 year, Survival only at discharge, 30 days, 60 days, 180 days AND/OR 1 year (O)? No interventional or observational studies were identified addressing whether fever prevention (or treatment) after cardiac arrest improves outcome. Fever after ROSC: For the critical outcomes of survival with good neurologic/functional outcome and/or survival only, we found very low quality evidence from 5 observational studies (downgraded for risk of bias and indirectness) that fever after ROSC is associated with poor outcome when TTM is not used. Fever after TTM: For the critical outcomes of survival with good neurologic/functional outcome and/or survival only, we found very low quality evidence from 6 observational studies (n =856) (downgraded for risk of bias and indirectness) that fever after TTM is not associated with outcome For the same critical outcomes we also found very low quality evidence from 2 observational studies (n = 411) (downgraded for risk of bias, inconsistency and indirectness) that fever after TTM is associated with poor outcome We suggest prevention and treatment of fever in persistently comatose adults after completion of targeted temperature management between 32 and 36 degrees Celsius (weak recommendation, very low quality evidence) 3

4 Hemodynamic support post resuscitation Among adults with ROSC after cardiac arrest in any setting (P), does titration of therapy to achieve a specific hemodynamic goal goal (e.g. mean arterial pressure > 65 mmhg) (I), compared with no hemodynamic goal (C), change Survival with Favorable neurological/functional outcome at discharge, 30 days, 60 days, 180 days AND/OR 1 year, Survival only at discharge, 30 days, 60 days, 180 days AND/OR 1 year (O)? There are no RCTs addressing hemodynamic goals post resuscitation. 1. Titration of therapy to achieve a specific hemodynamic goal (e.g. mean arterial pressure > 65 mmhg) compared to no hemodynamic goal For the critical outcome of survival with favorable neurological/functional outcome we have identified very low quality evidence (downgraded for risks of bias and publication bias) from one multicentre retrospective non-intervention study including 8736 subjects that found post return of spontaneous circulation (ROSC) systolic blood pressure (SBP) <90 mm Hg was associated with higher mortality (65% vs 37%) and diminished discharge functional status in survivors (49% vs 38%). Trzeciak (2009, 2895) For the critical outcome of survival we have identified very low quality evidence (downgraded for risks of bias and publication bias) from two retrospective single centre studies including 2282 patients that showed reduced survival for patients with post ROSC SBP <90 mm Hg (Bray 2014, 509) and <100 mm Hg (Kilgannon 2008, 410). 2. Does a bundle of therapies including a specific blood pressure target compared to no bundle For the critical outcome of survival with favorable neurological/functional outcome we have identified very low quality evidence (downgraded for risks of bias and publication bias) from seven studies including 813 patients: One before and after study of EGDT of 36 patients including a mean arterial pressure (MAP) target >80 mm Hg showed no difference in mortality or neurological outcome at hospital discharge (Gaieski 2009, 418). One prospective observational study of 118 patients using historical controls and a aiming for MAP >65 mm Hg showed survival to hospital discharge with a favourable neurological outcome in 34/61 (56%) up to one year, versus 15/58 (26%) in the control period (OR 3.61, CI , p=0.001) (Sunde 2007, 29). One cohort study of 148 patients when a MAP <75 mm Hg was a threshold for intervention showed no difference in neurological outcome at hospital discharge (Laurent 2002, 2110). One retrospective study of 136 patients identified groups with MAP >100 mm Hg or < 100 mm Hg after ROSC. Good neurological recovery was independently and directly related to MAP measured during 2 hours after ROSC (rs=.26; P<.01) (Mullner 1996, 59). One before and after observational study of a care bundle including 55 patients aiming for a MAP > 65 mmhg within 6 hours, resulted in an in-hospital mortality of 55.2% (bundle) vs. 69.2% (prebundle) (P = 0.29). Bundle patients achieved good neurologic outcome compared patients, CPC 1 or 2 in 31 vs. 12% patients, respectively (P = 0.08) (Walters 2011, 360). One prospective single centre observational study assessed the association between increasing time-weighted average mean arterial pressures and good neurologic outcome. Among 151 patients receiving a bundle of therapies, 44 (29%) experienced good neurologic outcome and a time-weighted average MAP threshold > 70 mm Hg had the strongest association with good neurologic outcome (odds ratio, 4.11; 95% CI, ; p = 0.014) (Kilgannon 2014, 2083). One retrospective study of bundle therapy targeting a MAP >80 mm Hg in 168 patients showed survivors had higher MAPs at 1 h (96vs. 84 mmhg, p\0.0001), 6 h (96 vs. 90 mmhg, p = 0.014), and 24 h (86 vs. 78 mmhg, p = 0.15) than nonsurvivors. Increased requirement for vasoactive agents was associated with mortality at all time points. Among those requiring vasoactive agents, survivors had higher MAPs than non-survivors at 1 h (97 vs. 82 mmhg, p =\0.0001) and 6 h (94 vs 87 mmhg, p = 0.05) (Beylin 2013, 1982). For the critical outcome of survival we have identified very low quality evidence (downgraded for risks of bias and publication bias) from two studies including 91 patients assessed the impact of post-resuscitation goal directed/bundles of care (including blood pressure targets) on survival: One before and after study of EGDT of 36 patients including a mean arterial pressure (MAP) target >80 mm Hg showed no difference in mortality at hospital discharge (Gaieski 2009, 418). One before and after observational study of a care bundle including 55 patients aiming for a MAP > 65 mmhg within 6 hours, resulted in an in-hospital mortality of 55.2% (bundle) vs. 69.2% (prebundle) (P = 0.29). (Walters 2011, 360). 4

5 We suggest hemodynamic goals (e.g. MAP, SBP) are considered during post resuscitation care and as part of any bundle of post-resuscitation interventions. (weak recommendation, low quality evidence). There is insufficient evidence to recommend specific hemodynamic goals, such goals should be considered on an individual patient basis and are likely to be influenced by post resuscitation status and pre-existing morbidities that are considered during post resuscitation care. (weak recommendation, low quality evidence). Impedance threshold device Among adults who are in cardiac arrest in any setting (P), does use of an inspiratory impedance threshold device (ITD) during CPR (I), compared with no ITD (C), change Survival with Favorable neurological/functional outcome at discharge, 30 days, 60 days, 180 days AND/OR 1 year, Survival only at discharge, 30 days, 60 days, 180 days AND/OR 1 year, ROSC (O)? Impedance Threshold Device + Standard CPR (I) vs Standard CPR (C): For the critical outcome of neurologically intact survival at hospital discharge (assessed with Modified Rankin 3), there was one RCT (Aufderheide 2011, 798) of high quality in 8718 out-of-hospital cardiac arrests that was unable to demonstrate a clinically significant benefit from the addition of the ITD to standard CPR: RR 0.97 (95% CI 0.82 to 1.15). For the critical outcome of survival to hospital discharge, there was one RCT (Aufderheide 2011, 798) of high quality in 8718 out-of-hospital cardiac arrests that was unable to demonstrate a clinically significant benefit from the addition of the ITD to standard CPR: RR 1 (95% CI 0.87 to 1.15). Impedance Threshold Device + Active Compression Decompression CPR (I) vs Active Compression Decompression CPR (C): For the critical outcome of neurologically intact survival there were no studies identified that compared the use of Impedance Threshold Device with Active Compression Decompression CPR with Active Compression Decompression CPR in cardiac arrests. For the critical outcome of survival to hospital discharge there were two RCTs (Plaisance 2000, 989, Plaisance 2004, 265) of very low quality (downgraded for imprecision and indirectness) that that were unable to demonstrate a clinically significant benefit from the addition of the Impedance Threshold Device to Active Compression Decompression CPR in a total of 421 out-of-hospital cardiac arrests: RR 0.91 (95% CI 0.07 to 12.7) (Plaisance 2000, 989) and RR 1.25 (0.5 to 3.1) (Plaisance 2004, 265). Impedance Threshold Device + Active Compression Decompression CPR (I) vs Standard CPR (C): For the critical outcome of neurologically intact survival at 12 months (assessed with CPC 2), there was one RCT (Frascone 2013, 1214) of very low quality (downgraded for risk of bias and suspected publication bias) in 2738 out-ofhospital cardiac arrests that was unable to demonstrate a clinically significant benefit from the addition of the ITD to ACD CPR (when compared with standard CPR): RR 1.34 (95% CI 0.97 to 1.85). For the critical outcome of neurologically intact survival at hospital discharge (assessed with CPC 2), there was one RCT (Frascone 2013, 1214, which incorporated data published in Aufderheide 2011, 301) of very low quality (downgraded for risk of bias, inconsistency, and suspected publication bias) in 2738 out-of-hospital cardiac arrests that was unable to demonstrate a clinically significant benefit from the addition of the ITD to ACD CPR (when compared with standard CPR): RR 1.28 (95% CI 0.98 to 1.69). For the critical outcome of survival to 12 months, there was one RCT (Frascone 2013, 1214) of very low quality (downgraded for risk of bias, and suspected publication bias) in 2738 out-of-hospital cardiac arrests that was unable to demonstrate a clinically significant benefit from the addition of the ITD to ACD CPR (when compared with standard CPR): RR 1.39 (95% CI 1.04 to 1.85, or from 2 more to 47 more per 1000). For the critical outcome of survival to hospital discharge, there were two RCTs (Wolcke 2003, 2201 and Frascone 2013, 1214 (which incorporated data published in Aufderheide 2011, 301)) of very low quality (downgraded for risk of bias, indirectness and suspected publication bias) in a total of 2948 out-of-hospital cardiac arrests that were unable to demonstrate a clinically significant benefit from the addition of the ITD to ACD CPR (when compared with standard CPR): RR 1.17 (95% CI 0.94 to 1.45)(Frascone 2013) and RR 1.41 (95% CI 0.75 to 2.66)(Wolcke 2003, 2201). Impedance Threshold Device + Standard CPR (I) vs Standard CPR (C): We recommned against routine use of ITD in addition to standard CPR (strong recommendation, high quality of evidence). Values and preferences statement: In making this recommendation we place a higher value on not allocating resources to an ineffective intervention over any yet to be proven benefit for critical or important outcomes. Impedance Threshold Device + Active Compression Decompression CPR (I) vs Active Compression Decompression CPR (C): We suggest against the routine use of ITD in addition to Active Compression-Decompression CPR (weak recommendation, very low quality of evidence). Values and preferences statement: In making this recommendation we place a higher value 5

6 on not allocating resources to an ineffective intervention over any yet to be proven benefit for critical or important outcomes. Impedance Threshold Device + Active Compression Decompression CPR (I) vs Standard CPR (C): We suggest against the routine use of ITD with Active Compression-Decompression CPR as an alternative to standard CPR (weak recommendation, very low quality of evidence). Values and preferences statement: In making this recommendation we place a higher value on not allocating resources to an intervention with equivocal benefit for critical or important outcomes. Induced Hypothermia Among patients with ROSC after cardiac arrest in any setting (P), does inducing mild hypothermia (target temperature 32-34ºC) (I), compared with normothermia (36-37ºC) (C), change Survival with Favorable neurological/functional outcome at discharge, 30 days, 60 days, 180 days AND/OR 1 year, Survival only at discharge, 30 days, 60 days, 180 days AND/OR 1 year (O)? Targeted temperature management (I) vs no targeted temperature management (C): Out-of-hospital cardiac arrest with shockable rhythm: For the critical outcome of survival with good neurological outcome, we have identified moderate quality evidence from one randomized control trial (RCT) [1] and one pseudorandomized trial [2] enrolling 275 and 77 patients, demonstrating a benefit in patients with out of hospital cardiac arrest (OHCA) with ventricular fibrillation (VF) or pulseless ventricular tachycardia (VT) as an initial rhythm. In these studies, cooling to degrees Celsius compared to no temperature management was associated with a risk ratio (RR) and odds ratio (OR) for good neurologic outcome at 6 months and hospital discharge of 1.4 (95% CI ) and 2.65 (95% CI ), respectively. For the critical outcome of survival, moderate quality evidence in the larger study demonstrates benefit in patients treated with induced hypothermia (RR for 180-day mortality 0.74, 95% CI ),1 while the other study found no significant difference (51% vs 68% hospital mortality, p=0.145).2 Out-of-hospital cardiac arrest with nonshockable rhythm: We found no randomized controlled trials comparing mild induced hypothermia (32-34 degrees Celsius) to no temperature management in patients with OHCA with pulseless electrical activity or asystole (i.e. nonshockable) as the initial rhythm.for the critical outcome of survival with good neurologic outcome, we found three cohort studies including a total of 1,034 patients, providing overall very low-quality evidence for no difference in poor neurologic outcome in patients with nonshockable OHCA (adjusted pooled OR, 0.90 [95% CI, ].[3-5] One additional retrospective study utilizing a large registry, including 1830 patients, provided very low quality evidence for an increase in poor neurologic outcome in patients with nonshockable OHCA (adjusted OR 1.44 [95% CI, ].[6] This data was not pooled with the above studies due to the lack of temperature data and limited patient information available. For the critical outcome of survival, we found very low quality evidence of a benefit in mortality at 6 months (OR 0.56, 95% CI ) from one of these studes.[4]inhospital cardiac arrest: We found no RCT s comparing mild induced hypothermia (32-34 degrees Celsius) to no temperature management in patients with in-hospital cardiac arrest (IHCA). For the critical outcome of survival to hospital discharge, we found very low quality evidence in one retrospective cohort study including 8316 patients that showed no benefit in patients with IHCA of any initial rhythm who were treated with targeted temperature management vs. no active temperature management (OR 0.9, 95% CI ).[7] For the critical outcome of neurologically favorable survival, we found very low quality evidence from the same observational study showing no benefit (OR 0.93, 95% CI ). This evidence was downgraded for high probability of selection bias, residual confounding, and a low prevalence of patients (2.6%) receiving the intervention. Note: Although we found numerous observational studies on the implementation of temperature management, these data are extremely difficult to interpret in light of other changes in post-cardiac arrest care that accompanied implementation, making it impossible to isolate the effect of temperature on outcomes after cardiac arrest. For this reason, we excluded all before and after studies. Other observational studies with concurrent controls also represent low quality evidence due to residual confounding and other factors. We therefore did not include these in the consensus on science, except for specific patient populations lacking higher quality (i.e. RCT) evidence. Is there evidence for an ideal temperature (I) when using targeted temperature management? For the critical outcomes of survival and survival with good neurologic outcome, we found high quality evidence from one RCT including 939 patients. This study compared cooling to 33 degrees Celsius compared to tight temperature control at 36 degrees Celsius in adult patients with OHCA of any initial rhythm except unwitnessed asystole, and found no benefit (HR for mortality at end of trial 1.06, 95% CI ; RR for death or poor neurologic outcome at 6 months 1.02, 95% CI ).[8] For the critical outcome of survival with good neurologic outcome we found low quality evidence in one additional small pilot RCT enrolling 36 patients comparing 32 vs 34 degrees Celsius in patients with OHCA VT/VF or asystole.[9] This study found no benefit (neurologically intact survival 44.4% vs 11.1%, p=0.12), although with only 36 patients it was underpowered. [1] The Hypothermia After Cardiac Arrest Study Group,2002;549. [2] Bernard;2002; 557.[3] Dumas; 2011; 877.[4] Testori;2011;1162.[5]Vaahersalo;826; [6] Mader;2014;21.[7] Nichol;2013;620. [8] Nielsen;2013;2197. [9] Lopez-de-sa;2012;2826. We recommend targeted temperature management as opposed to no targeted temperature management for adults with OHCA with an initial shockable rhythm who remain unresponsive after ROSC (strong recommendation, low-quality evidence). 6

7 We suggest targeted temperature management as opposed to no targeted temperature management for adults with OHCA with an initial nonshockable rhythm (weak recommendation, very low-quality evidence) who remain unresponsive after ROSC. We suggest targeted temperature management as opposed to no targeted temperature management for adults with IHCA (weak recommendation, very low-quality evidence) with any initial rhythm who remain unresponsive after ROSC. We recommend selecting and maintaining a constant, target temperature between 32 C and 36 C for those patients in whom temperature control is used (strong recommendation, moderate-quality evidence). Whether certain subpopulations of cardiac arrest patients may benefit from lower (32-34oC) or higher (36oC) temperatures remains unknown, and further research may help elucidate this. Mechanical CPR Devices Among adults who are in cardiac arrest in any setting (P), does Automated mechanical CPR devices (I), compared with Standard Manual CPR (C), change Survival with Favorable neurological/functional outcome at discharge, 30 days, 60 days, 180 days AND/OR 1 year, Survival only at discharge, 30 days, 60 days, 180 days AND/OR 1 year, ROSC (O)? For the critical outcome of survival to hospital discharge with good neurologic outcome (defined as CPC 1-2 or mrs 0-3), we have identified moderate quality evidence from 3 RCTs enrolling 7582 OHCA patients showing variable results (Hallstrom , Wik , Rubertsson ). One study (Hallstrom ) (n=767) demonstrated harm with the use of a load-distributing band mechanical chest compression device compared to manual chest compressions (7.5% of patients in the control group versus 3.1% in the intervention group, p=0.006). Two other RCTs (Wik , Rubertsson ) (n=6820), one using a load-distributing band and the other a LUCAS (Lund University Cardiac Arrest System), did not demonstrate benefit or harm when compared with manual chest compressions. For the critical outcomes of survival at 30 days with good neurologic outcome and survival at 180 days with good neurologic outcome, we identified moderate quality evidence from 1 RCT (Rubertsson ) using a LUCAS device and enrolling 2589 OHCA patients that did not demonstrate benefit or harm when compared with manual chest compressions. For the critical outcome of survival to hospital discharge, we identified very low quality evidence from 5 RCTs (Hallstrom , Lu , Rubertsson , Smekal , Wik ) enrolling 7734 OHCA patients and 150 in-hospital cardiac arrest patients showing heterogeneous results. One study of patients with in-hospital cardiac arrests (Lu ) (n=150) demonstrated benefit with use of a piston device compared with manual chest compressions (32.9% versus 14.7%, p=0.02). Two other RCTs (Rubertsson , Smekal ) did not demonstrate benefit or harm. One large RCT (Wik ) (n=4231) using a load-distributing band device demonstrated equivalence when compared with high-quality manual chest compressions. For the critical outcome of survival to 30 days we identified moderate quality evidence from two RCTs (Rubertsson , Perkins , n=7060) using the LUCAS device showing no benefit or harm when compared with manual chest compressions where quality of compressions in the manual arm was not measured. For the critical outcome of survival to 180 days, we identified moderate quality evidence from one RCT (Rubertsson ) using a LUCAS device enrolling 2589 OHCA patients showing no benefit or harm when compared with manual chest compressions where quality of chest compressions in the manual arm was not measured. For the critical outcome of survival to 1 year, we identified moderate quality evidence from one cluster RCT (Perkins ) using the LUCAS device showing no benefit or harm when compared with manual chest compressions. For the important outcome of return of spontaneous circulation, we identified low quality evidence from 7 RCTs enrolling 11,638 cardiac arrest patients (in-hospital and OHCA) (Dickinson , Halperin , Lu , Perkins , Rubertsson , Smekal , Wik ). Four studies (Dickinson , Halperin , Lu , Wik ) (n=4,432) demonstrated benefit with mechanical chest compression devices, however methodology of the Wik study did not allow adjustment for interim analyses. Three studies (Perkins , Rubertsson , Smekal )(n=7206) did not demonstrate benefit or harm when compared to manual chest compressions. We suggest mechanical chest compression devices should not be considered the standard of care for cardiac arrest patients, but can be considered a reasonable alternative to high quality manual chest compressions in some settings (weak recommendation, moderate quality of evidence). Values and Preferences Statement: 7

8 In making this recommendation we place value on data from a large, high-quality RCT demonstrating equivalence between high quality manual chest compressions and manual chest compressions. Local considerations such as relative costs and resource availability for maintenance of high quality manual chest compressions and mechanical chest compression device implementation should guide decisions around which mode of chest compression delivery is most appropriate. Also, there may be scenarios not directly addressed in the literature reviewed to support this treatment recommendation such as CPR in a moving ambulance, in the angiography suite or during preparation for ECLS, where mechanical chest compressions are more practical. Organ donation Among adults and children who are receiving an organ transplantation in any setting (P), does an organ retrieved from a donor who has had CPR (e.g. donor dies on ICU after intial successful CPR, or donation after unsuccessful CPR) (I), compared with an organ retrieved from a donor who did not have CPR (C), change increase survival rates, Complication Rate (O)? Consensus on Science Donors with Prior CPR Two papers reported the mean yield of organs procured from donors who had been resuscitated by CPR prior to donation was 3.9 (Faucher 2014) or 2.9 (Orioles 2013). For the critical outcome of immediate graft survival, low quality evidence from non-randomized studies did not detect any worse outcome when donors have had CPR and resuscitation for adult hearts (3239 organs, 7 studies), pediatric hearts (557 organs, 4 studies), adult lungs (1031 organs, 4 studies), pediatric lungs (105 organs, 1 study), adult kidneys (5,000 organs, 2 studies), pediatric kidneys (1122 organs, 2 studies), adult livers (2,911 organs, 3 studies), pediatric livers (689 organs, 2 studies), adult intestines (25 organs, 2 studies), and pediatric intestines (79 organs, 1 study) For the important outcome of graft survival for 1 year, low quality evidence from non-randomized studies did not detect any worse outcome when donors have had CPR and resuscitation for adult hearts (3230 organs, 8 studies), pediatric hearts (1605 organs, 8 studies), adult lungs (1031 organs, 4 studies), pediatric lungs (105 organs, 1 study), adult kidneys (5,000 organs, 2 studies), pediatric kidneys (1122 organs, 2 studies), adult livers (2,911 organs, 3 studies), pediatric livers (689 organs, 2 studies), adult intestines (25 organs, 2 studies), and pediatric intestines (79 organs, 1 study) For the important outcome of graft survival for 5 years, low quality evidence from non-randomized studies did not detect any worse outcome when donors have had CPR and resuscitation for adult hearts (3230 organs, 8 studies), pediatric hearts (1537 organs, 8 studies), adult lungs (1031 organs, 4 studies), pediatric lungs (105 organs, 1 study), adult kidneys (5,000 organs, 2 studies), pediatric kidneys (1122 organs, 2 studies), adult livers (2,911 organs, 3 studies), pediatric livers (689 organs, 2 studies), adult intestines (25 organs, 2 studies), and pediatric intestines (79 organs, 1 study) Donors with ongoing CPR (uncontrolled on-heart beating donors or uncontrolled donation after circulatory death) Two papers reported the mean number of organs procured from donors with ongoing CPR was 1.5 (Fondevilla 2012) and 3.2 (Mateos-Rodriguez 2012) For the critical outcome of immediate graft survival, low quality evidence from non-randomized studies did not detect any worse outcome when organs were recovered from non-heart beating donors with ongoing CPR compared to other types of donors for adult kidneys (203 organs, 4 studies) or adult livers (64 organs, 4 studies). For the important outcome of graft survival for 1 year, low quality evidence from non-randomized studies did not detect any worse outcome when organs were recovered from non-heart beating donors with ongoing CPR compared to other types of donors for adult kidneys (199 organs, 3 studies) or adult livers (60 organs, 3 studies). For the important outcome of graft survival for 5 years, low quality evidence from non-randomized studies did not detect any worse outcome when organs were recovered from non-heart beating donors with ongoing CPR compared to other types of donors for adult kidneys (177 organs, 2 studies) or adult livers (34 organs, 1 study). 1. We recommend that all patients who have restoration of circulation after CPR and who subsequently progress to death be evaluated for organ donation (strong recommendation, low quality of evidence). 8

9 In making this recommendation, we consider the absence of any evidence of worse graft function from donors with antecedent CPR, the desirability of providing more organs to waiting recipients, and the absence of any risk to the donor. As in all organ donation, the function of the donated organ determines whether procurement and transplantation proceed. Therefore, there is also precaution to ensure the safety of the recipient. 2. We suggest that patients who fail to have restoration of circulation after CPR and who would otherwise have termination of efforts be considered candidates for kidney or liver donation in settings where programs exist (weak recommendation, low quality of evidence). In making this recommendation, we consider the evidence that kidney grafts obtained from donors with ongoing CPR can function at rates comparable to kidneys obtained from other donors, and that recipients can safely tolerate delayed graft function that is common with kidneys obtained in this manner. We also consider the immediate life-saving potential of liver grafts, which offsets the potentially greater rate of long-term graft failure in livers obtained from donors with ongoing CPR. Oxygen dose after ROSC in adults Among adults who have ROSC after cardiac arrest in any setting (P), does an inspired oxygen concentration titrated to oxygenation (normal oxygen saturation or partial pressure of oxygen) (I), compared with the use of 100% inspired oxygen concentration (C), change survival to 30 days with good neurological outcome, survival to hospital discharge with good neurological outcome, improve survival, survival to 30 days, survival to hospital discharge (O)? 1. 30% inspired oxygen for 60 minutes after ROSC vs. 100% inspired oxygen for 60 minutes after ROSC Survival to discharge with favourable neurological outcome (CPC 1 or 2): [1 study, observational, Kuisma ]. For the critical outcome of survival to hospital discharge with favourable neurological outcome (CPC 1 or 2) we have identified very low quality evidence (downgraded for small numbers, lack of blinding, indirectness, misallocation of patients) from one RCT [Kuisma ] enrolling 32 OHCA (of which 4 excluded) patients that showed no difference between 30% inspired oxygen for 60 minutes after ROSC vs.100% inspired oxygen for 60 minutes after ROSC (8/14 vs. 6/14, unadjusted RR for survival1.33 [95% CI 0.63 to 2.84] p=0.46). Survival to hospital discharge: [1 study, observational, Kuisma ]. For the critical outcome of survival to hospital discharge we have identified very low quality evidence (downgraded for small numbers, lack of blinding, indirectness, misallocation of patients) from one RCT [Kuisma ] enrolling 32 OHCA (of which 4 excluded) patients that showed no difference between 30% inspired oxygen for 60 minutes after ROSC and 100% inspired oxygen for 60 minutes of after ROSC (10/14 vs. 10/14, unadjusted RR for survival 1.0 [95% CI 0.63 to 1.60] p=1.00). 2. Hyperoxia vs. Normoxia Survival with favourable neurological outcome (CPC 1 or 2) at 12 months: [1 study, observational, Vaahersalo ] For the critical outcome of survival with favourable neurological outcome (CPC 1 or 2) at 12 months we have identified very low quality evidence from 1 study [Vaahersalo ] (downgraded for very serious risk of bias, and indirectness) that showed no harmful effect from hyperoxia during the first 24 hours of ICU care. Survival to discharge with favourable neurological outcome (CPC 1 or 2): [5 studies, observational, Kilgannon , Bellomo 2011 R90, Janz , Roberts , Elmer 2014] For the critical outcome of survival to hospital discharge with favourable neurological outcome (CPC 1 or 2) we have identified very low quality evidence (downgraded as very serious bias and serious inconsistency, indirectness, confounding)) from 5 observational studies with conflicting results [2 showed hyperoxia worse than normoxia]. Studies reported CPC 1 or 2 outcomes at discharge: Janz [ ] showed in a single centre study of 170 ICU patients treated with therapeutic hypothermia that the maximum PaO2 in the first 24 h after arrest was associated with a worse outcome (CPC 1 or 2) (poor neurological status at hospital discharge, adjusted OR [95% CI ] P =.033). Roberts [ ] showed in a single centre study of 193 ICU patients that the first PaO2 after ROSC was not associated with outcome (hyperoxia adjusted OR for poor neurological outcome 1.05 [95% CI ] P=0.911). Elmer [2014] showed in a single centre study of 184 ICU patients that oxygen exposure over first 24 h of ventilation was not associated with outcome with unadjusted and adjusted outcomes [effect size cannot be estimated from data]. Two studies did not report CPC outcomes, and used other measures as a surrogate marker: Kilgannon [ ] reported worse independent functional survival at hospital discharge (hyperoxia vs. normoxia 9

10 124/1156 vs. 245/1171 [29 vs. 38%], unadjusted OR 0.45 [95%CI ] P<0.0001). Bellomo [2011 R90] reported no difference in discharge to home [hyperoxia vs. normoxia (27 vs. 34%) - effect size cannot be estimated from data] Survival to hospital discharge (or survival to 30 days): [7 observational studies, Kilgannon , Kilgannon , Bellomo 2011 R90, Janz , Ihle , Nelskyla 2013, Elmer 2014] For the critical outcome of survival to discharge (or survival to 30 days) we have identified very low quality evidence (downgraded as very serious bias and serious inconsistency, indirectness, confounding) from 7 observational studies with conflicting results [4 showed hyperoxia worse than normoxia]. Kilgannon [ ] showed a worse outcome with hyperoxia vs. normoxia based on the first ICU PaO2 (in-hospital mortality 63 vs. 45%, adjusted OR hyperoxia exposure 1.8 [95% CI P = 0.001]). Kilgannon [ ] showed a 100 mm Hg increase in PaO2 was associated with a 24% increase in mortality risk (odds ratio 1.24 [95% confidence interval 1.18 to 1.31]). Bellomo [2011 R90] showed no association between hyperoxia vs. normoxia (based on the worse PaO2 in first 24 h on ICU) adjusted OR for hospital mortality of 1.2 (95% CI 1.0 to 1.5) P= 0.04). Janz [ ] showed in a single centre study of 170 ICU patients treated with therapeutic hypothermia that the maximum PaO2 in the first 24 h after arrest was associated with a worse outcome. Survivors had lower maximum PaO2 (198 mm Hg; interquartile range, ) vs. non survivors (254 mm Hg; interquartile range, ; p =0.022). Adjusted OR - higher PaO2 increased in-hospital mortality (odds ratio 1.439; 95% CI ; P =.034). Ihle [ ] showed in a data linkage study of worse PaO2 (highest/lowest) in first 24 on ICU hyperoxia was not associated with outcome (hospital mortality 47 vs. 41%, adjusted OR hyperoxia vs. normoxia 1.2 [95%CI ]). Nelskyla [2013] enrolling 122 ICU admissions patients that showed no difference between patients with hyperoxia (PaO2 > 300mmHg in 1st 24 h after arrest) and normoxia (22/49 vs. 25/70, unadjusted OR 0.68 [95% CI 0.32 to 1.44] P=0.31) for 30 day survival or survival to discharge (20/49 vs. 24/70, unadjusted OR 0.76 [95% CI 0.36 to 1.61] P=0.47). Elmer [2014] reported that of 184 ICU patients, 36 % were exposed to severe hyperoxia, there overall mortality was 54 %, and severe hyperoxia, was associated with decreased survival in both unadjusted and adjusted analysis [adjusted odds ratio (OR) for survival 0.83 per hour exposure (95% CI , P = 0.04]. Survival to ICU discharge: [2 observational studies, Ihle , Nelskyla 2013] For the important outcome of survival to ICU discharge we have identified very low quality evidence (downgraded as very serious bias, serious indirectness, confounding) from 2 observational studies that showed no harm from hyperoxia: Ihle [ ] showed in a data linkage study of worse PaO2 (highest/lowest) in first 24 on ICU hyperoxia was not associated with outcome (ICU mortality 35% vs. 32% for hyperoxia vs. normoxia, unadjusted OR 1.16 [95%CI ] P=0.68). One observational study [Nelskyla 2013, survival to 30 days] enrolling 122 ICU admissions patients that showed no difference between patients with hyperoxia (PaO2 > 300mmHg in 1st 24 h after arrest) and normoxia (ICU discharge 53% vs. 46%, adjusted OR 0.75 [95%CI ] P=0.43). 3. Hypoxia vs. normoxia Survival to hospital discharge (or survival to 30 days): [3 observational studies, Kilgannon , Bellomo 2011 R90, Ihle ] For the critical outcome of survival to discharge (or survival to 30 days) we have identified very low quality evidence (downgraded as very serious bias and serious indirectness, confounding) from 3 observational studies that showed : Kilgannon [ ] showed a worse outcome with hypoxia vs. normoxia based on the first ICU PaO2 (57 vs. 45%, adjusted OR hypoxia exposure 1.3 [95%CI ] P = 0.009). Bellomo [2011 R90] showed a hypoxia vs. normoxia (based on the worse PaO2 in first 24 h on ICU) hospital mortality of 60 vs. 47% (hypoxia/poor O2 exchange vs. normoxia, OR hospital mortality 1.2 (1.1 to 1.4) P= 0.002). This paper also reported discharge to home outcomes (hypoxia/poor O2 exchange vs. normoxia 26 vs. 24%). Ihle [ ] showed in a data linkage study of worse PaO2 (highest/lowest) in first 24 on ICU no difference in outcome between hypoxia and normoxia (for in hospital mortality, 51 vs. 41%, adjusted OR hypoxia vs. normoxia 0.93 [95%CI ]. Survival to ICU discharge: [1 observational study, Ihle ] For the important outcome of survival to ICU discharge we have identified very low quality evidence (downgraded as very serious bias and serious indirectness, confounding) from 1 observational study: Ihle [ ] showed in a data linkage study of worse PaO2 (highest/lowest) in first 24 on ICU hypoxia was associated with a worse unadjusted outcome (ICU mortality 49% vs. 32% for hypoxia vs. normoxia, unadjusted OR 2.15 [95% CI ] P=0.008). RR 0.74 (95% CI ) worse with hypoxia. We recommend the avoidance of hypoxia in adults with ROSC after cardiac arrest in any setting (strong recommendation, very low quality evidence). 10

11 We suggest the avoidance of hyperoxia in adults with ROSC after cardiac arrest in any setting (weak recommendation, very low quality evidence). We suggest the use of 100% inspired oxygen until the arterial oxygen saturation or the partial pressure of arterial oxygen can be measured reliably in adults with ROSC after cardiac arrest in any setting (weak recommendation, very low quality evidence). Prognostication in hypothermia Prognostic Among adults with ROSC who are treated with hypothermia, (P), does any clinical variable when normal (e.g. 1. Clinical Exam, 2. EEG, 3. SSEP, 4. Imaging, 5. Other) when present (I), compared with any clinical variable when abnormal (C), change Survival with Favorable neurological/functional outcome at discharge, 30 days, 60 days, 180 days AND/OR 1 year, Survival only at discharge, 30 days, 60 days, 180 days AND/OR 1 year (O)? Clinical Examination: In patients who are comatose after resuscitation from cardiac arrest and who are treated with targeted temperature management, bilaterally absent pupillary reflexes to light at 72 or later from ROSC predict a poor outcome (death, vegetative state or severe cerebral disability) with very high precision (FPR 0[0-3]%). Bilaterally absent corneal reflexes at 72h or later from ROSC is also an accurate predictor of poor outcome, but it is less specific (FPR 2[0-7]%). Both these predictors have a low sensitivity (. 25%). Bilaterally absent or extensor motor response to pain at 36h or later from ROSC has a relatively high sensitivity for prediction of poor outcome (70[65-74]%). However, its false positive rate is also high (10[7-15]%). There are little data at present on combination of clinical signs. Presence of myoclonus within 72h from cardiac arrest is inconsistently associated with poor outcome (FPR 6%; 95%CIs 3-9). A prolonged, continuous and generalised myoclonus (status myoclonus) within 72h from cardiac arrest predicts poor outcome with high precision (FPR 0 [0-4[%) but low sensitivity (16%; 95%CIs 11-22). Rare cases of good outcome in patients with an early-onset status myoclonus have been reported. Predictors based on clinical examination, especially corneal reflex and motor response to pain, may be affected by sedation or paralysis. Blinding of the treating team is very difficult to achieve for predictors based on clinical examination, which implies a risk of self-fulfilling prophecy. Electrophysiology: In patients who are comatose after resuscitation from cardiac arrest and who are treated with targeted temperature management a bilaterally absent N20 wave accurately predicts a poor neurological outcome after rewarming (at 72h from ROSC) (FPR 1[0-3]%) while prediction during TTM is less reliable (FPR 2[0-4]%). Absence of EEG background reactivity during TTM is inconsistently associated with a poor neurological outcome (FPR 2 [1-7]%) while after rewarming at.72 h from ROSC it predicts a poor outcome with 0[0-3]% FPR. Limitations of EEG reactivity include being operator dependent and non-quantitative, and lacking standardization. Presence of epileptiform discharges on EEG, both during TH or after rewarming is inconsistently associated with poor neurological outcome. Presence of electrographic seizures or status epilepticus, either during TTM or after rewarming is almost invariably associated with poor outcome. The definition of epileptiform activity and status epilepticus is inconsistent among studies. Limited evidence shows that prognosis is worse when status epilepticus is associated to a non-continuous background or absence of reactivity. A flat or low-voltage EEG was inconsistently associated with poor outcome. Timing and definitions of low voltage was inconsistent among studies. A lowest BIS value =0 during TTM, corresponding to a flat or very low voltage EEG during TH is inconsistently associated with poor outcome (FPR from 0% to 10%). BIS is affected by interference from muscular artefacts, which restricts its use in patients who are sedated and paralysed during TTM treatment. The amplitude of the EEG signal may be affected on a variety of technical conditions such as skin and scalp impedance, inter-electrode distances, size, type and placement of the exploring electrodes, and type of filters adopted.49 Use of EEG grading, brainstem auditory evoked potentials and pain-related somatosensory evoked potentials has been documented only in single studies and their predictive value needs confirmation from other studies. Biomarkers: In patients who are comatose after resuscitation from cardiac arrest and who are treated with targeted temperature management, high and increasing concentrations of biomarkers are associated with poor outcome. However, the biomarkers thresholds which predict a poor neurological outcome with 0% FPR vary between studies. Advantages of biomarkers over other predictors such as EEG and clinical examination include quantitative results and likely 11