Sunday, March 16, 2008

Is a Radiologist Doing CPR Like a Pig with a Wristwatch?

ResearchBlogging.orgI've been taking cardiopulmonary resuscitation (CPR) classes periodically since medical school many years ago. For me, one of the hardest part of these classes has been keeping track of the ever-changing Official Ratio™ of chest compressions to breaths. Is it 5:1? 15:2? That's one reason why the following recent study in the Journal of the American Medical Association (JAMA) caught my eye:
Bobrow, B.J. (2008). Minimally Interrupted Cardiac Resuscitation by Emergency Medical Services for Out-of-Hospital Cardiac Arrest. JAMA, 299(10), 1158-1165.
This study compared standard advanced life support with a new protocol known as minimally interrupted cardiac resuscitation (MICR). The results: patients treated with MICR were at least 3 times more likely to survive their cardiac arrest than patients treated with the standard life support protocol.

Just what is MICR and why should it make such a difference?

The main idea behind MICR is that the "C" in CPR is a whole lot more important than the "P" part. Unlike the mixture of chest compressions and assisted breaths given in standard CPR, the MICR protocol starts off with 200 uninterrupted chest compressions. Electrical defibrillation was not performed until after these initial 200 chest compressions. Why do it this way? As the authors state:
During resuscitation efforts, the forward blood flow produced by chest compressions is so marginal that any interruption of chest compressions is extremely deleterious, especially for favorable neurological outcomes. Excessive interruptions of chest compressions by prehospital personnel are common. Therefore, MICR emphasizes uninterrupted chest compressions.
In other words: CPR is a very poor substitute for an actual heart. Interrupting chest compression downgrades this "poor" to "piss poor".

But what's the benefit of not ventilating a patient?  How does a patient get any oxygen if one omits the "pulmonary" part of CPR?
...positive pressure ventilations during cardiac arrest may be harmful because they increase intrathoracic pressure, thereby decreasing venous return and subsequent myocardial and cerebral blood flow. Probably due to the excitement and stress of resuscitation efforts, excessive ventilations by both physicians and EMS personnel are common.
Immediately after a sudden VF cardiac arrest, aortic oxygen and carbon dioxide concentrations do not vary from the prearrest state because there is no blood flow and oxygen consumption is minimal. Therefore, when chest compressions are initiated, the blood flowing from the aorta to the coronary and cerebral circulations provides adequate oxygenation at an acceptable pH. At that time, myocardial oxygen delivery is limited more by blood flow than oxygen content. Adequate oxygenation and ventilation can continue without rescue breathing because the lungs serve as a reservoir of oxygen that allows adequate oxygen exchange with the limited pulmonary blood flow during cardiopulmonary resuscitation...
In addition, substantial ventilation occurs from chest compression–induced gas exchange (ie, small volumes exhaled with each compression and inhaled with chest recoil) and spontaneous gasping by the patient in cardiac arrest during cardiopulmonary resuscitation.
Is it time to change the way we do CPR? Maybe. Maybe not.

This is a nicely designed prospective study with a large sample size (n = 883), and published in an English-language major medical journal. I commend the authors for this excellent work and am impressed by their results. However, this JAMA article has a major flaw: the CPR protocols used were not randomized. Lack of randomization can allow all sorts of bias to creep into a study. For example, the effects seen in this study could even represent an example of the Hawthorne effect. I don't think I can improve on the authors' own conclusion:
These results need to be confirmed in a randomized trial.
I agree. If a randomized controlled trial corroborates the findings in this study, I expect to see a lot of changes in official CPR protocols. For an excellent further commentary on this study, please see this JAMA editorial.

At this point, you might be wondering: "How can CPR possibly matter to a radiologist?"

If so, 5 points for Gryffindor for raising that question. But just 5.

A fair number of radiology procedures involve injecting some contrast agent (lay term = "dye") into a patient. Cardiac arrythmias are an uncommon but known complication of these agents. Thus, these studies are only performed when a radiologist is around to monitor the exam and treat any adverse reactions, including cardiac arrest.

I haven't had to do CPR on a patient since I entered radiology. However, a friend had a cardiac arrest before my eyes at a dance last year. Fortunately, CPR and a nearby automatic external defibrillator (AED) were enough for me, my spouse and several other pals to restart my friend's heart, which is still beating strong at local dances 15 months later.

I guess you could say this event made CPR really, really relevant to me -- relevant not only as a physician, but also as a precocious geezer and potential CPR customer myself. These days I travel to local dances with a personal AED just in case cardiac lightning strikes again in my vicinity. If it does, I'm also going to pay a lot more attention to my chest compressions, and keep them as uninterrupted as possible.


Oldfart said...

Well. I'm confused. 200 chest compressions is roughly 3 minutes and 20 seconds without oxygen. If you stay under water that long and the water is not very cold and you are not a kid, you are generally dead or brain damaged. Or so I have understood.

The Samurai Radiologist said...

@ oldfart:

Yes, the idea of not ventilating for several minutes does seem a bit nonintuitive.

I highly recommend reading the whole article and some of their references for this bold idea. In summary though, here are three main points of their rationale:

1. Blood in the aorta still contains lots of oxygen -- enough, apparently to last for the first few minutes of CPR.

2. Air trapped in the lungs still contains lots of oxygen to exchange, and substantial ventilation occurs just with chest compression alone.

3. In the first few minutes before defibrillation, we may be doing more harm than good by over-ventilating.

Good point about drowning. Seems like water in the lungs could make these patients a different kettle of fish with regard to the need for ventilation. In any event, the authors of the JAMA study focused on cardiogenic arrests, and specifically excluded non-cardiac causes, such as trauma and drowning. Therefore, we can't automatically extrapolate the findings from this study to non-cardiogenic arrests.

steppen wolf said...


one must also consider that, when one is drowning, the heartbeat goes up, people tend to get really scared, and therefore oxygen consumption is increased. After a cardiac arrest, this problem is not there.

Also, normal air contains about 20% oxygen; expired air contains about 15%: which means, we are really only using 5% under normal conditions. This percent probably goes down during cardiac arrest. Did they measure the oxygen content in the expired air of these patients at all?

The Samurai Radiologist said...

@ steppen wolf:

Good points about O2 consumption in drowning and O2 content in expired air.

The authors did not measure O2 concentration in the expired air.

I probably should have mentioned in my original post that this was a study of out-of-hospital arrests in which CPR was done by bystanders in about 35 - 40%, and by EMT's in 100%.

Anonymous said...


Oldfart said...

" Well. I'm confused. 200 chest compressions is roughly 3 minutes and 20 seconds without oxygen. If you stay under water that long and the water is not very cold and you are not a kid, you are generally dead or brain damaged. Or so I have understood."

Where did you get 3 minutes 20 seconds from? Chest compressions should be given at around 100 compressions per minute. That is, a typical 30 compression cycle should take about 18 seconds. I assume this new study uses the same 100beat/s rate. "Hard and Fast!" has been the CPR mantra for quite some time regarding ways to improve bystander CPR. The 200 cycles should take 2 minutes or less, which isn't enough time to cause much damage even to a panicking drowning victim, if breathing can be restored immediately afterward.