Does Observation for ACS Make Sense? Part 2: Identifying ACS

ACS is an acute thrombosis in a coronary vessel causing varying degrees of occlusion (14). Classically, ACS was divided into ST-segment elevation myocardial infarction (STEMI), non-ST-segment elevation myocardial infarction (NSTEMI), and unstable angina (UA). STEMI is thought to result from sudden complete occlusion of a coronary vessel with subsequent transmural infarction, whereas NSTEMI and UA involve thrombosis that either does not entirely occlude the vessel or occurs in a vessel with good collateral circulation. NSTEMI and UA are distinguished by the presence or absence of infarction, i.e. elevated cardiac biomarker level.

Several studies since the late 1980s have shown that EDs are actually quite good at identifying MI (15, 16, 17). Each study found a 2% rate of missed MI in the ED. However, the conclusion from all three of these papers was that ED providers are not good at identifying MI. To support this argument the authors ignored data supporting good outcomes and embellished on findings suggesting bad outcomes for patients. Who wants to publish an article suggesting everything is just fine?

As an example, the NEJM study from 2000 (16) found that of the 19 patients with missed MI, 2 of them (11%) had ECGs that were later interpreted as ischemic by a blinded cardiologist. In their conclusion, they state that there was an “11% incidence of failure by the emergency department clinician to detect ST-segment elevations… in the electrocardiograms of patients with myocardial infarction”. That is a gross misrepresentation of the data. That sentence sounds like emergency physicians miss 11% of STEMIs. The reality is that 11% of missed MIs showed some ECG changes suggestive of ischemia. In actuality, they should have concluded: “Out of 10,000 study patients, emergency physicians only missed 2 MIs (0.02%) whose ECG suggested ischemia”. 

Another example is that in all three studies, the mortality rate between missed MI and those admitted with identified MI were the same – about 10%. The authors attribute this to insufficient data. They say the outcomes must be worse for patients with missed MI, and they just didn’t have large enough numbers to show it.

These studies led to a good amount of fear among emergency physicians and hospitals, as those scary, albeit misleading, conclusions rightly should. However, I’d argue it’s just another example of  publication bias giving us bad information.

This is what observation units were originally designed for in the 1980s, but are these still necessary? As mentioned in the prior post, troponins are more sensitive than CK-MB and become positive within 3-4 hours, with 6 hours being the maximum amount of time it would take for us to see a positive test result (18, 19). Given that we already did (hopefully) serial ECGs and “0 and 3-hour troponin” in the ED, is further testing going to add anything? If we are really worried, can we not repeat the ECG and troponin in 6 hours in the ED? This isn’t a prohibitively long length of stay. “Observation unit crowding” and handoffs to other providers in the observation unit who may not know the patient are a common source of medical errors and patient harm (20), so this must factor into our thinking of the risks and benefits of observation for the patient.

Let’s think about it another way. We often use the HEART pathway to discharge chest pain patients after 3 hours, if they are deemed low-risk. What we are essentially saying is we have ruled out ACS. Why wouldn’t that timeline also apply to “high-risk” patients? We can either rule out ACS in 3 hours or we cannot – the level of risk should not change this fact.

Image courtesy of echobasics.de

Most observation units these days will include a resting echocardiogram as part of their observation pathway for patients with chest pain. This may help further risk stratify a patient for ischemic heart disease and congestive heart failure, but does it help with our fundamental question of identifying current ischemia or infarction? Looking for regional wall motion abnormalities would be the primary means for this. Studies have shown that rest echo to identify regional wall motion abnormality is on average only about 87% sensitive for ACS or MI, and not at all specific (21, 22, 23, 24). In the study that showed the highest sensitivity of echo (93%) for identifying ischemia, 2.3% of the patients who did not have regional wall motion abnormality were actually having MI (21).

This conclusion intuitively makes sense. Echo can be limited by patient’s body habitus, the ability of the cardiologist to interpret the images, variants in anatomy, etc. Also, if you see a wall motion abnormality, how do we know it is linked to what is happening now? Unless we have an echo from one week ago, identifying an abnormality as one data-point is unlikely to be helpful. Just as an ECG without comparison to a prior has limited utility.

Let’s do this thought experiment. You work in a perfect ED where the providers accurately interpret every ECG and all causes of non-ACS chest pain have been ruled out. In this ED, STEMIs will be identified by ECG and NSTEMIs will be identified by a positive troponin. STEMI and NSTEMI are not candidates for most observations units. Therefore, the only patients going to the observation unit are those with unstable angina and who do not have ACS.

So let’s take a deep dive into the murky world of unstable angina.

Probably not. Look at it this way: the person who created the first guidelines for the diagnosis and treatment of unstable angina (UA) no longer believes it to be a real entity (25). The definition was created at a time when we were still using CK-MB as evidence of myocardial necrosis. As mentioned earlier, troponin is far more sensitive and specific than CK-MB. By 2000, AHA guidelines estimated that more than one-third of patients previously thought to have UA would be reclassified as NSTEMI, given they would have had a positive troponin (26).

Since then, troponins have become even more sensitive with a lower cut point for an abnormal result, further classifying more patients as NSTEMI rather than UA. And now, we have “high-sensitivity” troponins. Almost all patients with chest pain due to ischemia will have detectable high sensitivity troponin (27), and even slight elevations in high-sensitivity troponins have been associated with adverse outcomes (28). It is likely that the upper limit of normal will become progressively lower with time. While this will result in lower specificity for diagnosing ACS, it will lead to higher sensitivity of identifying NSTEMI, and possibly result in the end of UA as a diagnosis. Patients will either be classified as having stable angina of varying severity or MI.

For now though, many of us do not have high-sensitivity troponin assays available, so we may still see some people with “unstable angina”.

Unstable angina is typically defined as angina that is either worsening, occurs at rest or is new, coupled with either transient ST elevation (>1 mm and <20 minutes), ST depression (>0.5 mm) or TWI (>3 mm) in at least 2 contiguous leads OR in a patient with known CAD (as diagnosed by angiography or prior MI) (29, 30).

As you can see from this definition, a patient with chest pain, known CAD, and no ECG changes could technically have UA. Add a negative troponin and that seems like the patients who we would typically see in an observation unit.

By definition, patients with unstable angina are patients with ACS, and they should presumably be treated aggressively for ACS. The following is a brief, up to date review of the literature on treating unstable angina. Of note, unstable angina is often grouped with NSTEMI and called NSTE-ACS.

Nitroglycerin

A 2009 Cochrane Review concluded that nitrates reduce short-term mortality in patients with ACS (31). The studies in this review used IV nitrates, primarily IV nitroglycerin.

2014 AHA guidelines recommend sublingual nitroglycerin to be administered to all patients with chest pain concerning for ACS, followed by IV nitroglycerin if the pain is refractory to sublingual medication (11).

Aspirin

In 2002, the Antithrombotic Trialists Group performed a meta-analysis of aspirin use in various cardiovascular disorders including acute stroke and acute MI (32). They found that aspirin at initial presentation, and then low dose daily, reduced serious vascular events in every subset of patients. One of these subsets was patients with unstable angina.

2014 AHA Guidelines recommend giving 162 to 325 mg non-enteric coated aspirin, chewed, to all patients with suspected ACS (11).

Dual Antiplatelet Therapy

The CURE trial established the benefit of clopidogrel, with aspirin, given immediately on patient presentation with ACS (33). Patients presenting within 24 hours of suspected NSTE-ACS – defined by ECG changes or elevated cardiac biomarkers – were randomized to receive aspirin plus a loading dose of 300 mg clopidogrel, and then 75 mg daily thereafter, compared to control receiving just aspirin. They found reduced recurrent ischemia, MI, heart failure, and death in the group receiving aspirin + clopidogrel. These benefits were seen early (in the first 30 days) and long-term (over the next 9 months). 

2014 AHA guidelines recommend dual antiplatelet therapy for treatment of NSTE-ACS (11).

Anticoagulation

A 2008 Cochrane Review concluded that heparins do not improve mortality, but do reduce the risk of MI in the first week – from 5% to 2% – in patients presenting with ACS (34). The studies included both low molecular-weight heparin (LMWH) and unfractionated heparin (UFH). It should be noted that this risk reduction in MI disappeared at 30 days and at 3-6 month follow up.

There is considerable controversy as to whether heparin benefits patients with NSTE-ACS, given its known, significant bleeding risk. For more reading see this great post by the website NNT: NNT – Heparin for ACS.

Regardless of the data, 2014 AHA guidelines recommend the use of heparin in all NSTE-ACS (11).

Fibrinolytic Therapy

The TIMI IIIB trial found no benefit and increased risk of severe hemorrhage for fibrinolytic therapy in patients with NSTE-ACS, (35).

2014 AHA Guidelines do not recommend fibrinolytics as a treatment for NSTE-ACS (11).

Invasive strategy

A 2016 Cochrane review concluded that an early invasive strategy (within 2 hours) in patients with unstable angina or NSTEMI does not improve mortality when compared to a selective invasive strategy (36). A selective invasive strategy involves performing catheterization only on patients with ongoing ischemia – recurrent infarction, angina at rest, dynamic ST changes on ECG, and definitive inducible ischemia on provocative testing. They found that patients who underwent early invasive therapy had a 2% decreased risk of MI at one year, but a 3% increased risk of MI during the actual procedure. It should also be noted that most patients in this meta-analysis had positive biomarkers, i.e. had NSTEMI. When the authors analyzed only those patients with negative biomarkers, i.e. unstable angina, they found increased mortality with an early invasive strategy.

2014 AHA guidelines recommend risk stratifying patients to see who would benefit from an early invasive versus delayed invasive strategy (11). This may change as the Cochrane Review came out after these recommendations.

After evaluating the evidence and recommendations for treatment of unstable angina, it seems clear that patients whom we suspect have unstable angina, i.e. ACS, may not be suitable for the observation unit. That may include the older patient with a concerning history and known CAD, but a non-specific ECG and negative initial troponin (Editor’s note: A patient with “clinically important” unstable angina or ACS – one in whom there may be a significant risk of short-term complications and aggressive treatment may benefit – will likely have ongoing/recurrent symptoms and demonstrate dynamic ECG changes and/or detectable troponin on serial testing). Regardless of current evidence to the contrary, AHA guidelines still recommend aggressive medical therapy, and possibly even an early invasive strategy for patients with unstable angina.

In conclusion, observation and the interventions and tests that come with it are unlikely to help us identify ACS any better than our initial ED evaluation. Troponin will almost always become positive within 6 hours in patients with NSTEMI, so this can reasonably be identified within an ED visit. For those patients who are high-risk – concerning stories or ECGs or ongoing chest pain – consider the possibility of unstable angina and remember that the treatment for this should be timely and relatively aggressive. I would argue that placing a concerning patient on observation is not appropriate. Cardiology should be consulted early in such patients. For patients with non-concerning stories and ECGs, ACS can be ruled out in an ED visit, regardless of the patient’s age and risk factors.