Friday, September 16, 2011

Two "facts" about the diagnosis of ACS that you need to forget - Part I

Like the title says, I won't be discussing any new findings or developments in the recognition or diagnosis of ACS that you'll need to memorize, keep on your iPhone, or write on your hand. These are two "important facts" that are hammered into you during your training that are taking up valuable space in your brain, space that could be used for important stuff, like memorizing lines from The Simpsons.

The first thing you need to know is that asking crap about cardiac risk factors is a waste. Sure, it makes the chart look pretty and it gives the appearance of diligence. It also less-than-preferable when you are deep into presenting to an attending, having described in succinct but descriptive language the crushing pain the patient had, the sweating, the left arm heaviness, the relief with nitro, the concerning EKG findings... and all this becomes derailed when the attending asks about family history. "How can you possibly decide what to do with the patient," you are lectured, "if you don't know their risk factors!"

Don't fool yourself into believing that you are obtaining actionable intelligence, however. Knowing whether a patient smokes, takes Lipitor, or had a grampa that keeled over at 45 years of age should not be guiding your diagnostic reasoning.

Okay, so I can avoid sounding like some crazy heretic, let me start off the discussion with an appeal to authority. What do we find in the textbooks about this issue?

In the 2011 Tintinalli, we find that:
“All patients should also be questioned regarding the presence of cardiac risk factors, although these ‘risks’ are valid only for predicting the presence of coronary artery disease within a given population and are not predictive of the presence or absence of acute ischemia in an individual patient.”

Rosen's, as well, comments that:
“The presence of risk factors for a particular disease is primarily of value as an epidemiologic marker… In the ED, presence of risk factors in an individual patient without established disease has minimal or no effect on the clinical likelihood (pretest probability) of a specific disease process.”

Well, is this just the crazy, EM perspective? Nope, here's the American Heart Association in their 2007 NSTEMI guidelines, where they mention the use of risk factors in the initial history:

Even Lee Goldman, a cardiologist renowned for developing various eponymous criteria for determining cardiac risk, has not find the traditional risk factors to be useful, stating that "demographic and traditional risk factors (with the very notable exception of a history of MI or coronary disease) are of little importance in predicting the cause of acute chest pain.”

So, this isn't my wacky opinion, this is actually the conventional wisdom!

Now, perhaps you want some justification, a bit more meat in the argument. Let's check out the empirical evidence in two key studies. The first is the classic systematic review in JAMA, from their excellent Rational Clinical Exam series (results from these papers show up in the in-service!). Written by Panju et al., Is This Patient Having an MI? synthesizes a number of studies looking at the relative contribution of elements of the history, physical exam, and EKG.  There a few shortcomings in the methods, etc., but it is still one of the most oft-cited papers on the topic.
They find that:
Some of this was based on a nice study by Jayes in 1992. Multicenter, prospective collection of information of risk factors in adults being checked out for chest pain. Lotta patients, 1743 in fact. They ran the results through logistic regression and found:
Yeah, there seems to be a small bump in the RR for DM and family history, and barely for smoking. And only in men! These results are blown away by the RR for simply showing up with chest pain, in both men and women.

(Hey, wait a minute. Why is the risk of ACS higher in women with chest pain than for men? Aren't women supposed to present, like, all differently? I'll get to that "fact" in Part II.)

Well, that paper is a few years old now - anything newer out their? Why yes, in fact, there was an excellent single-center study done by Body et al. in Manchester, UK, in the ED. They prospectively obtained information on the standard risk factors (HTN, lipids, DM, smoking, and family history) on 804 adults coming in with "chest pain occurring within the past 24 h that the treating physician suspected to be cardiac in origin." Then they looked at who ruled in/out, and found that knowing the risk factors helped...
... not so much.

Of course, why did we ever expect that these classic risk factors would help us in the ED? This is epidemiological data, used in the office with the asymptomatic patient, to determine the risk of having an event in the next 10 years. Hey, if you're going to bust out your Framingham Risk calculator, knock yourself out, but realize you may be are answering a different question that the one you were aiming for.

You can also listen to this great podcast at ERCAST.

So, when it comes to the utility of cardiac risk factors in the ED, I'm actually not much of a heretic, and I'm not really espousing anything very controversial. I am going out on a bit of a limb, however, in Part II, where I deconstruct the "fact" that women with ACS present with significantly different symptoms and histories than men. Gonna have to show you some evidence there!

No comments:

Post a Comment