A quick journal club this week in what’s been a very busy few days. As you know we were at the BMJ Awards this week. We had a fabulous time but our section was won by a great team teaching CPR to students in the hope that rates of bystander CPR will increase and thus patients will benefit. Interestingly our paper this week addresses exactly this issue, does it really make a difference and if so by how much.
This paper from Denmark1 is an observational study of outcomes in cardiac arrest that specifically seeks to quantify the effects of bystander CPR and defibrillation on outcomes amongst those patients who survive for at least 30 days. The abstract is below, but as we always say, do go read the paper yourself.
What kind of paper is this?
It’s an observational study based on a prospectively captured database of all cardiac arrests in Denmark between 2001 and 2012. This is clearly the only way that you can do a study like this as there is no ethical way in which we could withold bystander CPR from one arm of a trial. Systematically acquired data is a good way of getting information together which then may reveal associations between what happened during a cardiac arrest and subsequent outcome.
Which patients are we talking about here?
Importantly the cohort examined here is those who survived to 30 days post arrest. It’s not all comers and thus the outcome data here relates to those who have already survived to 30 days. That’s a reasonably well defined group and that’s fine but it’s important that we don’t confuse these figures with all cardiac arrest patients, or all ROSC patients.
What did they look for?
From the registry they determined what happened to patients during their arrest, and also folowed the patients through to a year post cardiac arrest to determine a meaningful outcome. They had good follow up data on survival, nursing home admission and anoxic brain injury. This is important as we want to know whether we get more survivors, but also that the survivors have a good functional outcome. In that latter respect the authors are using the coding of nursing home admission or anoxic brain injury as a marker of adverse outcome. That’s OK, but it’s a proxy marker of outcome as compared to something like a functional outcome score (e.g. Glasgow Outcome Score).
What did they find.
Over the 12 year period the incidence of cardiac arrest remained pretty much the same but survival increased from 3.9 to 12.4% which is a great result in itself reflecting improvements in cardiac arrest care detailed in the paper. Of note the proportion of patients getting bystander CPR or defibrillation increased significantly. For the 2855 who survived 30 days then 9.7 % died within a year and 10.5% had an anoxic brain injury or were admitted to a nursing home. However, the proportion of patients with anoxic brain injury or nursing home admission decreased during the study period.
Sounds good, can we believe the results?
We always have to be cautious with database studies, the results here show an association, but that’s not the same as causality. Or at least that’s the critical appraisal answer. In reality this type of study is probably the best way of getting an answer and the association of bystander CPR/defibrillation has a biological plausibility supported by lots of other evidence. In fact in terms of initial survival we would be surprised not to find an association with a positive outcome. As for longer term survival, nursing home admission and anoxic brain injury the link is still strong but not to the same degree as there are other aspects of intra and post arrest care that might affect outcome. The authors have controlled for this through statistical analysis and the link remains so it may be as strong an evidence as we will get. We should also note that questions around admission criteria to nursing homes may vary and the authors acknowledge this.
What does this mean.
This study is reassuring that we should continue to promote early and effective CPR and defibrillation without the increased risk of anoxic brain injury patients in nursing homes. As an emergency physician we should continue to promote interventions that take place before our prehospital teams arrive.