Tuesday, March 20, 2012

Dextrose for cardiac arrest

Meir and I were talking during a recent shift, and he asked the perceptive question "What's the deal with giving dextrose in a cardiac arrest?"

Good question!

You know that there's a lot of controversy about how to run a cardiac arrest - intubating or not, how often to ventilate, or doing a short trial of CPR before defibrillation. This is especially true regarding the "code drugs," like epinephrine. 
Well, okay. Epi works.
But at least when it comes to epinephrine and amiodarone, there are some studies out there, some base of evidence to start the discussion from. This is not true for D50.
If you look in the 2000 ACLS guidelines, you'll see the list of the "reversible causes" of cardiac arrest. It doesn't include hypoglycemia.
If you don't believe me, look at the fine print at the bottom.
Now skip ahead 5 years, and we now see that hypoglycemia has been added (2005 ACLS):
Down in the green box, at the bottom.
Read through the guidelines, though, and you'll see that not a word is uttered about why this was added.

Now, fast-forward to 2010, and they've taken it out! And, just like they added it without comment, it's gone without justification or evidence, nor even a mere "clinical evidence suggests."
Since the AHA elected not to review any relevant evidence about the topic, I decided to answer some questions about hypoglycemia, cardiac arrest, and the relative benefit of trying to squeeze that huge syringe of syrup into an IO.

1. Does hypoglycemia cause cardiac arrest? 

You figure this would be easy enough to answer, but there is almost no direct data that suggests hypoglycemia is a cause of cardiac arrest, let alone a treatable cause.

One case series in 1995 reviewed 3 arrests that the authors thought were associated with hypoglycemia. These patients all had significant primary problems (active CAD, cerebral hemorrhage, and severe pancreatitis), so it's hard to assign blame to the blood sugar.

Another case series, looking at patients with severe heart failure, concluded that one cardiac arrest was due to hypoglycemia (oddly enough, she didn't have diabetes). And in one last example, a patient in the ICU became asystolic at the same time her blood sugar was plummeting, although she also was developing a severe hyperkalemia at the same time.

The problem with this handful of case reports is that, given the uncontrolled nature of the situations, it's hard to point out cause and effect. Just because one thing occurred at the same time as something else, or even right after, doesn't mean they're related.

Well, look at this from a different point of view. Is there a proposed "mechanism," some physiological theory, that suggests that hypoglycemia could cause an arrest?

There is the phenomenon of the "dead in bed" syndrome, where a relatively healthy diabetic is found deceased in the morning. A number of researchers think they've found a link - the usual dip in blood sugar levels at night can cause a prolongation of the QT interval. And long QT intervals can sometimes cause problems! (See these examples at Dr. Smith's ECG Blog.)

A long QTc. (source)
They've been able to show this effect both in the lab (giving insulin to healthy people) and at home (people on continuous ECG and glucose monitoring).

But, just because you can show a longer QT, doesn't mean you have a smoking gun! Others have pointed out that there are probably a number of other factors involved. For instance, it may not be the hypoglycemia that triggers the QT changes, but in fact may be the body's own epinephrine that kicks off the arrythmias!
Now epinephrine is bad for the heart?!?

So, we just don't know.

2. Is the finger stick accurate in cardiac arrest?

Usually, the capillary blood glucose is pretty close to the venous level, close enough that we all trust it. However, in the critically ill patient, the capillary level becomes less accurate, as a number of studies have shown.

Only one study has looked at patients getting CPR, though.  This was a pretty big study by cardiac arrest standards - they checked the venous and capillary glucose levels in 50 cardiac arrest patient. It wasn't encouraging.

There were 4 patients with "true" hypoglycemia, found on the venous samples sent to the lab. The fingerstick missed 1 of those, and also managed to misdiagnose 5 patients as having hypoglycemia, when they really didn't. (That works out to 75% sensitive, and 38% specific).

Not so useful. So, perhaps we should just skip testing, and treat empirically. What could be the harm in that?

 3. What's the harm in empirically treating for hypoglycemia?

Glad you asked.

It's the same reason we're trying to cool people down after we get a pulse back  - neurologic outcomes. In some studies, they gave dextrose to some cats before they put 'em into cardiac arrest, while other cats they didn't. The cats who didn't get sugar beforehand had better brains afterwards.

You can't really do this same kind of study in humans (for example). One group in Helsinki, though, checked the blood sugar on VF cardiac arrests, and looked at how well they recovered. Patients who had increases in their glucose after resuscitation didn't survive to hospital discharge as often. This is just the latest evidence - see this review article (50% dextrose: antidote or toxin?) for plenty of other examples.

4. Any evidence giving sugar helps? 

Well, yes and no. There is no evidence that pumping liquid rock candy into someone's tibia helps in cardiac arrest.

Now, there are a lot of other sick people out there, people teetering on the edge, critically ill, septic, metabolically deranged - with a blood sugar headed south, and fast. You have to find those folks and treat them quick. Some of these are kids, with weird metabolic problems, or with sepsis.  But the key is to get to them before they crash, while they still have a pulse.

The authors of the only review I found on this topic concluded that (my emphasis):
"This is obviously a controversial issue and raises the point of whether we should still be teaching that hypoglycaemia is a reversible cause of cardiac arrest when there seems to be not enough evidence to support this. 
Current evidence would suggest that patients may suffer cardio-respiratory arrest with hypoglycaemia, but not because of it." 
The Bottom Line

Despite the huge crowd in the room whenever a patient is "coding," there are only a limited number of spaces around the patient. Likewise, you have the contents of the whole code cart available to you, but can only push each drug one at a time. You have to pick your priorities in a code, and it doesn't appear as though pushing sugar water ought to one of those.

No comments:

Post a Comment