When approaching an undifferentiated patient, it is important to create a differential of critical or can’t miss conditions off of the chief complaint. The critical differential guides what questions to ask in your HPI, what physical exam findings to focus on, what labs/imaging to order in your workup, and skeletonizes the medical decision-making portion of your note. This organized approach is especially helpful initially when there is diagnostic uncertainty or when patients present with vague or multiple complaints. In the undifferentiated patient it is often easier to figure out what is NOT going on, than it is to figure out the final diagnosis. Think of it as trying to solve a case with limited clues or leads, it is hard initially but once you obtain more data you begin to get a better idea of what is going on. Sometimes the uncertainty can be paralyzing and you will not know where to start your workup. This is why we advocate focusing on the acute chief complaint and using that to build a critical differential of diseases you can’t miss (patients with multiple chief complaints may have many chronic complaints and only one acute complaint that caused them to come to the ER). Once you have initiated the workup, you then methodically rule out life-threatening conditions using an evidence-based approach. After you have convinced yourself that you have ruled out a life-threatening process, you can focus on building a differential of common diseases from the chief complaint, focusing now on what most likely is going on.
For undifferentiated chest pain we routinely talk about the 6 chest pain killers. These are the can’t miss diagnoses we must always consider:
- Acute Coronary Syndrome (ACS)
- Pulmonary Embolism (PE)
- Aortic Dissection (AD)
- Tension Pneumothorax
- Esophageal Rupture
Most of the time the approach boils down to:
“Could the patient’s chest pain be due to ACS, PE, or AD?” This is because half of our critical differential can usually be ruled out using history, vitals, and exam.
For example, tamponade and tension pneumothorax are both obstructive shock states. This means that for someone to have tamponade physiology, they must be hypotensive. If they are not hypotensive then this is a patient with a simple pericardial effusion. The same goes for tension pneumothorax, for the patient to have tension physiology they must have hypotension. There are also classic exam findings for both of these, but ultimately both of these conditions refer to an obstructive shock state. Therefore, if your patient with chest pain is not hypotensive then it is unlikely either of these two things are going on. Emergency Medicine is a dynamic field and it is possible for a patient with a simple pneumothorax or pericardial effusion to develop tension or tamponade physiology but don’t worry about those nuances now. Focus on simplifying your approach to patients first and then you can add layers of complexity once you gain more experience.
The reason esophageal rupture is a life-threatening diagnosis is because it can lead to mediastinitis and overwhelming sepsis. These patients usually have a historical trigger that will help you consider esophageal rupture or they will present as very sick septic patients with chest pain. These patients typically report persistent retching/vomiting or will report recent upper GI instrumentation. We can usually rule out esophageal rupture using history alone.
After cutting your differential in half, you can focus on the three remaining conditions because they usually require labs, imaging, or specialist consultation.
For AD, it is important to determine at least 3 things:
- Does the patient have aortic pain?
- Pain that is ripping, tearing, or radiates to the back.
- Is the mediastinum widened on CXR?
- >8 cm is considered widened.
- Are the pulses symmetric throughout?
- If there is asymmetry, you should measure blood pressure on both sides.
If your patient denies aortic pain, has a normal mediastinum, has symmetric pulses throughout, AND has an alternative diagnosis that is more likely, then you have effectively ruled out AD. Patients coming in with one or more of the aforementioned 3 things or those that are critically ill without a clear diagnosis may warrant further workup for AD.
We use the Pulmonary Embolism Rule-out Criteria (PERC) to rule out the diagnosis in patients under 50 years of age where PE is suspected but unlikely (i.e. low risk). In patients over 50 years of age, we use either the Wells’ or Geneva score (in conjunction with a D-dimer) to rule out PE. For patients over 50 years of age, many providers have adopted using an age adjusted D-dimer cutoff, where you multiply the patient’s Age (in years) x 10 ug/L = age adjusted cutoff. Please ensure you are using the appropriate d-dimer assay when adjusting for age. You will often notice that some providers use these scoring systems without obtaining a D-dimer, just be aware that they were developed to be used in conjunction with a D-dimer to rule out PE in low risk patients. Studies have shown that Wells’ score is a bit more effective than Geneva score in discriminating PE in suspected patients. The purpose of using these scoring systems is to save you from having to do a CTA chest on patients you consider low to low-intermediate risk for PE. Patients that you consider intermediate to high risk for PE should have a CTA chest performed to definitively rule out PE (D-dimer alone may not be enough to rule out PE).
Is There Anything Else Going On?
You will notice that after a long discussion and workup, most patients presenting to the ER with chest pain get placed into the ACS pathway. Before you put your patient into the ACS pathway, this is where you want to take a step back and think about what most likely is going on. At this point in the encounter you must really try to determine if you can build a story for this chest pain being caused by anything other than ACS. Try to find alternative explanation for the chest pain. For example:
- Does this patient’s chest pain have a clear cut MSK component to the pain?
- Is there a zoster rash on the chest?
You must be careful here and ensure that you can build a logical alternative explanation to the patient’s chest pain that is supported by your HPI, vitals, physical exam, labs, etc. If you cannot convince yourself of an alternative explanation then you place your patient into the ACS pathway.
Acute Coronary Syndrome
ACS is divided into two main categories:
- ST Elevation ACS (STE-ACS)
- Non-ST Elevation ACS (NSTE-ACS)
STE-ACS only requires a “positive” ECG to diagnose, meaning STE on ECG. This can also get nuanced as not all STE are due to a STE myocardial infarction, but for now just keep things simple and treat all STEs as potential STEMIs. The troponin doesn’t matter here because the acute coronary occlusion could be so fresh that it has not resulted in cardiac cell death.
NSTE-ACS is further subdivided into:
- NSTE myocardial infarction (NSTEMI)
- Unstable Angina (UA)
To diagnose NSTEMI you must have a “negative” ECG. This means that the ECG can be normal, have T-wave inversions (TWI), or ST Depressions (STD). Each one of these ECG findings is more concerning than the one before. The only thing you cannot have on ECG is STEs because then it would be called a STEMI. NSTEMI patients are diagnosed with a positive troponin. There are different types of NSTEMIs and you will primarily deal with type I (ACS) and type II (so called demand ischemia) NSTEMIs.
In UA both the ECG and troponin are negative. Similarly, to NSTEMI, the ECG can show anything but STEs. UA is a clinical diagnosis (no lab test or ECG will help you diagnose it). UA is considered in patients presenting with a history of worsening anginal pain and is a specific syndrome of chest pain. The classical definition of UA is categorized as chest pain that presents with 1 or more of the following:
- Chest pain that occurs at rest or with minimal exertion and usually lasts more than 20 minutes.
- New onset severe chest pain.
- Chest pain that occurs with a crescendo pattern (brought on by less activity, more severe, more prolonged or increased frequency than previously).
You will often hear providers say that patient has a “good” chest pain story. What they mean is that this patient has chest pain that is likely cardiac or due to ACS. Below is a list of clinical factors that increase likelihood of ACS/AMI (obtained from a paper published in Resuscitation 2010 on the value of symptoms and signs in the emergent diagnosis of ACS). The presence of each one in isolation does not automatically mean that you have ACS, but if your patient reports multiple items from this list you should strongly consider ACS as a diagnosis:
- Chest pain radiating to both arms > R arm > L arm.
- Chest pain that is exertional.
- Chest pain associated with nausea or vomiting.
- Chest pain associated with diaphoresis.
Another term you will hear providers say is that a patient has “typical” angina . Typical angina is categorized by the presence of all 3 of the following criteria:
- Chest pain that is substernal.
- Chest pain that is exertional.
- Chest pain that is relieved by nitroglycerin or rest.
Patients with 2 of the aforementioned items will be said to have “atypical” angina. Whereas those with 0-1 of the items on the list may be said to have “non-anginal” chest pain.
The terms atypical angina, atypical chest pain, non-angina, and non-cardiac chest pain will often be used interchangeably by providers and that is very confusing. They do not mean the same thing! It is important that you understand that atypical-angina and non-angina refer to chest pain that could still be due to ACS versus atypical chest pain and non-cardiac chest pain refer to pain that is not caused by ACS. It is important to know this because the definition of typical chest pain was determined based off of symptoms that suggested ACS in a predominantly middle-aged male cohort. We know that certain patient populations with ACS, tend to present less with chest pain and more with what we term anginal equivalents. For example, men are more likely to report central chest pain in ACS while women are more likely to report fatigue/weakness, dyspnea, diaphoresis, indigestion (burning or heartburn), nausea/vomiting, or palpitations. Populations that present “atypically” or with anginal equivalents rather than chest pain:
- Immunocompromised (HIV, Autoimmune disease)
- Patients with ESRD
- Patients with CHF
The definition of noncardiac pain is in the name itself. When you categorize something as noncardiac chest pain, you are stating that this pain is not due to the heart and therefore not due to ACS.
Conversely, you will hear providers refer to a patient having a “bad” chest pain story. What they mean is that this patient has chest pain that sounds noncardiac. Below is a list of clinical factors that decrease the likelihood of ACS/AMI (this list is from that previously mentioned Resuscitation article from 2010). The presence of each one of these in isolation does not automatically mean that you have ruled out ACS, but if your patient reports multiple items from this list, the likelihood of ACS decreases:
- Chest pain that lasts seconds
- Chest pain that is pleuritic
- Chest pain that is sharp or stabbing
- Chest pain that is positional,
- Chest pain that is reproducible with palpation
Sometimes we get lucky and patients present with a clearly “good” chest pain story that includes almost all of the items on the high likelihood of ACS list or they present with a really “bad” chest pain story with almost all of the items on the low likelihood of ACS list. However, you will see that most of the time patients provide vague symptoms and report a mix of items from both lists.
Risk Factors Put Framingham On The Map
In the Emergency Department (ED), traditional cardiac risk factors are poor predictors of determining if the current chest pain your patient is having is due to ACS. These traditional cardiac risk factors and their associated scoring systems (GRACE, TIMI) are great at predicting your patient’s likelihood of having coronary artery disease (CAD) they aren’t good at predicting ACS in the ED. We will go over how to determine if your patient has ACS when we discuss the HEART score below. It is important to obtain these historical risk factors from your patient because they are adjuvants in your final decision making, but in isolation, they will not help you determine if the chest pain is due to ACS. The reason the scoring systems based off risk factors alone are poor predictors of ACS in the ED is because not all acute MIs result from large plaques (those with a critical degree of luminal stenosis >50-75% occlusion). There is data to suggest that plaque size is not the key determinant of MI. We know that vessels with immature plaques (<50% luminal stenosis) can also result in acute MI. Size does matter somewhat (which is why knowing risk factors for coronary artery disease is helpful) because larger plaques are more likely to cause ACS, but it is important to keep in mind that small plaques that are not detected by provocative testing (treadmills, P-mibis, cardiac CTs) also play a role. This stuff is complicated!
What Is Happening At The Coronaries?
What we ultimately care about in the ED is whether or not the patient with undifferentiated chest pain is presenting with an acute coronary syndrome. That is our goal in the moment. It is important to identify cardiac risk factors and the likelihood of CAD in our patients so that we can help reduce them, but they are more helpful to primary care physicians. To definitively know if your patient’s chest pain is due to ACS you must look at what is going on at the coronary arteries and the only currently available test for that is a cardiac catheterization. This is an invasive test that we do not perform on everyone. We have relatively non-invasive tests (exercise treadmills, P-mibis, cardiac CTs, etc) that help us determine what is going on at the coronaries but they are less sensitive and time/resource intensive. In the ED, we only have 3 ways to non-invasively look at what is going on at the coronaries upon patient presentation:
The ECG helps us diagnose STEMIs, the troponin helps us diagnose NSTEMIs. In UA the ECG and troponin are both negative, therefore only the history helps us in the diagnosis.
What Is The Risk of Unstable Angina?
Since the diagnosis of STE-ACS and NSTEMIs is clear-cut, we spend most of our brainpower trying to figuring out if our patient with chest pain has UA. Historically we have been horrible at it. We still aren’t that good and there are multiple papers that point out how difficult it is for us to determine if our patient with undifferentiated chest pain has ACS. Diagnosing UA is easy when the patient presents with all of the typical symptoms that define UA. But what do we do with that elderly patient with risk factors for CAD but a story that has items from BOTH the high likelihood of ACS and low likelihood of ACS list? Or that elderly patient with risk factors with a very bad story? Or that young patient with no risk factors and a good story? The more patients with undifferentiated patients you see the more nuanced these questions become. Luckily, smart people decided to study this and came up with the HEART score.
The Oracle: HEART Score
The HEART (stands for History, ECG, Age, Risk Factors, Troponin) score is a prospectively validated scoring system that allows you to determine your patients risk for UA. It predicts your patient's 6-week risk of major adverse cardiac event (all-cause mortality, myocardial infarction, or need for coronary revascularization). But first a quick disclaimer:
This scoring system should only be used in patients > 21 years of age presenting with chest pain that is suggestive of ACS or in which ACS has not been ruled out as a potential diagnosis.
DO NOT use it in patients with new STEs, new ECG (or dynamic) changes, hypotension, or those that are critically ill.
Ok, back to the show. This tool helps guide your disposition. Strictly speaking, patients that are low risk by HEART score will be discharged home with an outpatient appointment for stress testing within 72 hours (although this is currently being debated, new data suggests urgent referral to pmd without giving definitive appointment for stress testing). Patients that are intermediate risk by HEART score require an observation period AND cardiology consultation prior to discharge. Patients that are high risk by HEART score basically have UA and must be treated the same way with an admission to cardiology with usual UA care.
I want to emphasize this point: please make sure you only apply the HEART score in patients with undifferentiated chest pain in which you suspect ACS. If you have done a thorough enough history, exam and have convinced yourself that you have an alternative diagnosis that is WAAAAAAY more likely than ACS then don’t use the HEART score. For example, if the patient has chest pain because he has a huge zoster rash on his chest then please do not calculate a HEART score for them. Just remember that it is difficult to predict with a single symptom or exam finding if a patient’s chest pain is due to ACS. You have to look at the whole picture and convince yourself that you have a logical and clear-cut alternative diagnosis. Don’t come to me with:
“So, the patient has exertional chest pain, that is associated with nausea, vomiting, and diaphoresis. It radiates to the right arm. But it was TOTALLY reproducible on exam. Sooooooo, I don’t think it is ACS. Send him home.”
While I will concede that it is odd that this patient, with multiple items from the high likelihood for ACS list, has reproducible chest pain. I will not agree that this could not be ACS. About 7% of chest pain that is reproducible is due to ACS. Chew on that.
You will often hear providers say any combination of the following:
- “This patient needs a rule out.”
- “I want to admit for an ACS rule out.”
- “I want to rule out ACS.”
- “We have ruled out ACS on this patient.”
- “Do we need to rule out ACS on this patient?”
- You get the idea…
I want to make sure you understand what this means colloquially. Language and words matter (ironic, I know, since my spelling and grammar are horrible, but indulge me). Just like you shouldn’t use the terms atypical chest pain and noncardiac chest pain interchangeable (THEY DO NOT MEAN THE SAME THING!), you should know what it means “to rule out ACS.”
In the ED, an ACS rule out means you have determined that your patient with undifferentiated chest pain does not have STE-ACS or NSTEMI-ACS. That is all you have ruled out. You have ruled out the possibility of an acute MI that may lead to cardiac cell death (if not taken to PCI emergently) or has already resulted in cardiac death (elevated troponin). You have not ruled out the possibility of UA (go back a few sections if you still don’t understand what I mean). Therefore, you can’t discharge your patient that looks great (continually giving you the thumbs up sign) just because you worked them up and they have a “negative” ECG and a negative troponin. You HAVE to determine their risk for UA and the only way we can do that today, in an evidence-based manner, is with the HEART score.
How Many ECGs & Troponins Do I Need?
This is another hotly debated topic. The most likely correct answer is: it depends. It depends on what troponin assay you have at your institution. It depends on how long the chest pain has been going on. Some have advocated to do 2 sets of ECGs/Trops on all comers with undifferentiated chest pain that may be due to ACS. Others are firm believers in the “one-and-done” approach. I’m not smart enough to know the correct answer, but I will tell you what I was taught and why it makes sense.
Older generation troponin assays could detect elevated troponin levels between 6-12 hours after onset of myocardial injury (i.e. onset of cardiac chest pain). This is why some old school docs may ask you to obtain troponins Q6-8 hours apart. Modern assays can detect troponin elevations as early as 3-4 hours after onset of myocardial injury (with some even earlier than that). I was taught to figure out how many ECGs/Trops were needed based on the onset of the patient’s chest pain:
- If the chest pain has been going on for LESS than 3 hours → get at least 2 sets of ECGs/Trops.
- If the chest pain has been going on for MORE than 3 hours → get one ECG/Trop and you are done.
- If you are unsure when the chest pain began → get at least 2 sets of ECGs/Trops.
That made sense to me and that is what I do. You can decide for yourself after taking a deep dive into the literature or by asking a more experienced physician.
Ugh. There Goes The ED, Giving Me Another BS Chest Pain Consult
This was my attempt to outline all of the cognitive processing that happens when a patient presents to the ER with undifferentiated chest pain. It is a very basic approach and does not take into consideration all the subtle nuances of some diseases. I am not an expert. I am not a cardiologist. The point of this is to give you a basic understanding and to have it act as a foundation to add more knowledge. This is my interpretation of how to approach undifferentiated chest pain after I spent some time picking smart people’s brains and reading the literature. You may have differing ideas on the approach and I encourage you to send us your feedback. Ultimately, you have to study this on your own time to have it make sense in your brain. Thanks for reading!
ddxof.com Algorithm for Chest Pain
Subscribe to the EM Ed by entering your email in the subscription box below. Don't rely on Facebook to get notifications for new posts. We only email when a new post is published. No spam. If you are reading this on your phone, just keep scrolling down to get to the Subscribe box.