Clinical Scenario and PICO question
You have finally gotten away from work for a week to take that ski trip you have been planning. Waiting for the chairlift you see a bearded skier, with equipment and an outfit straight out of 1972, collapse. You look around desperately for Ski Patrol in hopes of avoiding the taint of work on your hard-earned vacation, but to no avail.
The teenage snowboarder with the pierced nose uses his iPhone 4 to call 911 and tells you an ambulance is “totally” en route, ETA 10 minutes. What do you do? Do you really want to start mouth to mouth on this bearded antiquity? When EMS arrives, should they defibrillate immediately?
PICO question
P: Patient with V-fib cardiac arrest, prolonged ambulance response time
I: Compression only CPR; Delayed defibrillation after CPR by EMS
C: Compression and ventilation CPR; Immediate defibrillation by EMS
O: Return of spontaneous circulation, Survival to hospital discharge, Neurologic outcome
Synopsis
A big thank you to hosts Mike Lambert and Hannah Watts. Shout outs to Tyler, Vijay, Jesse, Matt, CDO and Mark for their thoughtful and entertaining presentations.
Article #1
Rea TD, Fahrenbruch C, Culley L, et al. CPR with Chest Compression Alone or With Rescue Breathing. NEJM 2010;363:423-33.
This is the second large RCT published in the same issue of NEJM on this topic (both with same conclusions). In this multicenter, RCT of dispatcher instructions to bystanders performing CPR on patients after cardiac arrest, 981 patients were randomly assigned to receive chest compressions alone, and 960 to receive chest compressions plus rescue breathing. The entire resuscitation was taped, with the dispatcher coaching the bystander through the resuscitation until EMS arrival. There was no significant difference between the two groups in the primary outcome, survival to hospital discharge (12.5% with compressions alone, 11.0% with compressions plus rescue breathing, P=0.31). There was also no difference between the 2 groups for the secondary outcome of favorable neurologic outcome at discharge (14.4% and 11.5%, respectively; P=0.13). In other words, in this study, compression only CPR was as effective as compressions plus rescue breathing, and didn’t save lives at the expense of neurologic outcome (in fact trend towards improved neurologic outcome with compressions alone).
Consensus in the room was that this was a methodologically sound study. Majority of patients were enrolled in Seattle, where your chance of survival after cardiac arrest is near miraculous anyway. Approximately 34% of patients were enrolled in London (London patients had a 4% survival rate). There is a threat to external validity because of the variation in CPR training/post-arrest survival in different areas of the country and the world, but conceptually, compression only CPR should be even less intimidating to bystanders in areas of low-community CPR training rates (everywhere outside Seattle). As an illustration of this, in the study, patients randomly assigned to receive chest compressions alone were more likely to actually receive those compressions than patients randomized to compressions + rescue breathing (80.5% vs. 72.7%, P<0.001).
A few small methodology issues: CPR guidelines changed in 2006 (during this study), and differences between the 2 groups might be smaller with the newer recommended 30:2 rather than 15:2 ratio of compressions/ventilations. Also, there was no controlling for differences in ACLS interventions, therapeutic hypothermia, etc. Also, nice point Harwood, some patients were excluded because when EMS arrived, they were found not be in arrest. Maybe some of those who received rescue breathing came out of arrest, and then therefore were excluded (bias against rescue breathing).
A bone for the Bayesians in the audience: Erik pointed out that the conclusion paragraph of the abstract contained the statement that there was a trend towards better outcomes in certain subgroups. The idea of stressing point estimates and giving value to trends rather than defining meaningful conclusions only as those with p values less than 0.05 represents an important and recent change in editorial philosophy.
This study did contain 4 pre-specified subgroup analyses. There was a trend towards improved outcomes with chest compressions alone in patients with arrest from a cardiac cause, and a trend towards improved outcomes with chest compressions + rescue breathing for patients with non-cardiac causes of arrest. While it might be tempting to start giving different CPR instructions to bystanders treating patients with different etiologies of arrest, subgroup analyses should not be viewed as reliable conclusions, but only as ways to generate future hypotheses. An important exception was discussed-the pediatric patient in arrest. All 3 JC studies excluded children, and it’s well known that children have much higher rates of respiratory arrest, and therefore should receive both compressions and ventilations.
This study highlights the difference between efficacy and effectiveness. The efficacy of a drug is the measure of a drug’s ability to treat a certain condition (ideal, or study world). The effectiveness of a drug takes into account real world factors such as tolerability and ease of use. We anticipate that regardless of the efficacy of compression only CPR, the effectiveness should be significant, as bystanders reluctant to perform mouth-to mouth rescue breathing may be more willing to jump in and perform CPR if compressions are all that is required. Interestingly in this study, they performed efficacy analyses, and the magnitude of differences favoring chest compression alone was larger than that observed in the effectiveness analyses. In other words, something about compression only CPR (e.g. fewer interruptionsàimproved circulation) appears to be beneficial, beyond the effectiveness issue of reluctance to perform rescue breathing.
In 2008, the AHA published a “Call to Action”, emphasizing the importance of high-quality chest compressions. Unless bystanders are trained and comfortable with rescue breathing, it is recommended to call 911 and then simply provide hard and fast compression only CPR with minimal interruptions until EMS arrival (Circulation 2008;117:2162-2167). This recommendation is limited to arrest in adults presumed to be of cardiac origin. A current billboard and public service announcement campaign is promoting this recommendation.
Article #2:
Wik L, Hansen TB, Fylling F et al. Delaying Defibrillation to Give Basic Cardiopulmonary Resuscitation to Patients with Out-of-Hospital Ventricular Fibrillation. JAMA 2003;289:1389-1395.
Switching gears, the second topic discussed was the concept of delaying defibrillation for several minutes to first perform high-quality CPR. The concept here is that ventricular fibrillation is a huge energy suck (resulting in a low ATP, severe acidosis state). If tissue perfusion can be improved by performing high quality CPR for several minutes, defibrillation may have a higher chance of success.
In this RCT of 200 patients with out-of-hospital ventricular fibrillation arrest in Oslo, Norway, patients either received standard care with immediate defibrillation, or “CPR first” with 3 minutes of CPR by EMS personnel prior to defibrillation. The CPR first group also received more prolonged CPR prior to subsequent defibrillations. Their primary end point was survival to hospital discharge. Survival for the standard group was 15% versus 22% in the CPR first group (P=0.17, not significant, but trend favoring CPR first). The secondary outcome of good neurologic outcome at hospital discharge also demonstrated no significant difference between the two groups. A pre-specified subgroup analysis evaluated outcomes in patients with ambulance response times of up to or longer than 5 minutes. For patients with ambulance response times greater than 5 minutes, hospital and one year survival were statistically better in the group who received CPR first prior to defibrillation. Again, this was sub-group analysis (hypothesis generating) with small subgroups, and as Erik mentioned, small sub-groups lead to an even higher chance of not seeing these results replicated in future trials.
Other methodologic issues: as Christine pointed out, this study was performed back in the day of stacked shocks. Initial CPR wasn’t just priming the pump: multiple defibrillation attempts in the standard group meant significantly less CPR, and might have affected the study outcomes (favoring CPR first). Harwood also pointed out his favorite pet peeve; in this article published in JAMA, “the numbers don’t add up”. Probably not a fatal flaw, but come on…get the tables right!
There were a similar number of witnessed arrests (>90%) and bystander CPR being performed prior to EMS arrival (about half) in both groups. While bystander CPR is often ineffective, this study didn’t (couldn’t) measure the quality of CPR performed before EMS arrival, and therefore these results might be quite different if high-quality CPR was provided by bystanders.
In the future, whether defibrillation should be postponed in favor of immediate CPR may depend on the frequency spectrum of the electrocardiogram as this can predict the probability of ROSC after defibrillation-stay tuned.
So, no difference overall, but results favoring CPR first in situations with prolonged ambulance response times, based on sub-group analysis. Whether these results hold up in other studies, and how long CPR should be performed prior to defibrillation all need to be investigated. There was disagreement in the room about whether or not to defibrillate the scenario patient immediately, but it was agreed that aggressive CPR should be performed while the defibrillator is being prepared.
A final important point about witnessed VT/VF in-hospital arrests or witnessed out-of-hospital arrests when an AED is immediately available. Shock them. Shock them right away. Do not intubate them first. Do not delay defibrillation to perform CPR. Shock them immediately. Shock them now.
Article #3:
Baker PW, Conway J, Cotton C et al. Defibrillation or CPR first for patients with out-of-hospital cardiac arrests found by paramedics to be in ventricular fibrillation? Resuscitation 2008;79:424-431.
Last and least, the most recent RCT on the topic of CPR before defibrillation in VF cardiac arrest. In this Australian study, a total of 202 patients with out-of-hospital VF arrest were randomized to receive either 3 minutes of EMS CPR prior to defibrillation, or immediate defibrillation. Primary outcome was survival to hospital discharge, and secondary outcomes included neurologic status at discharge. The study was powered to detect the outcome difference found in the Wik study for patients with an ambulance response time of >5 minutes. In this study, for all response times combined, as well as for the group with ambulance response times >5 minutes, there was a tendency for reduced survival to hospital discharge in the CPR before defibrillation group (lower by 6.8% and 4.7%, respectively), although these were not statistically significant.
There were significant problems with this study. First, it was stopped early, but you had to discover this in the discussion section, which is weird and suspicious. Second, they performed at per-protocol, rather than an intention to treat (ITT) analysis. ITT analysis is based on initial treatment intent, not treatment eventually received. For the purposes of an ITT analysis, everyone who begins the treatment in a trial is considered to be part of the trial, regardless of whether or not they finish the trial or crossover to another treatment. This helps to prevent erroneous interpretation of the results. In addition, 139 patients were “not randomized by mistake”; a significant number of patients excluded and no further explanation provided.
In sum, a negative study (against CPR prior to defibrillation) with significant methodologic flaws. See Article #2 for our JC group’s bottom line.
The one interesting and unforeseen conclusion from this study: it took place between 2005 and 2007, and the protocol changed mid-study to reflect the new international CPR guidelines published in 2006. Overall survival in this study increased more than two-fold following introduction of the 2006 guidelines (8.8% to 18%). Although this was not statistically significant, it is a positive trend in survival and indirectly supports the concept of aggressive CPR with minimal interruptions for patients with cardiac arrest.
The most recent AHA guidelines (2006) give very wishy-washy recommendations on CPR first before defibrillation for out-of-hospital arrest not witnessed by EMS (“may consider”, “may give” CPR before defibrillation). This reflects the conflicting evidence on the topic. The new AHA CPR and Emergency Cardiovascular Care guidelines are due out in February 2011.