Stuck in Limbo: The Intermediate HEART Score

Author: Jamie Pospishil, MD

Editor: Philippe Ayres, MD; Esteban Davila, MD

Case:

A 68-year-old male with a past medical history of hypertension, insulin-dependent diabetes, and dyslipidemia presents to the ED with chest pain that is intermittent, sharp, “stabbing”, lasting up to one minute at a time, and occurring several times over the past three days. The pain is non-exertional, non-radiating, and the patient is able to walk several blocks without symptoms. He denies associated shortness of breath, abdominal pain, nausea, vomiting, fever, or cough. 

He is afebrile with BP 165/87 mm Hg. On exam, he is in no acute distress with clear lungs, regular heart sounds, and no chest or abdominal tenderness. His ECG demonstrates normal sinus rhythm at 73/min with left ventricular hypertrophy/strain pattern. This ECG is unchanged from one 8 months ago.

Figure 1: ECG demonstrating left ventricular hypertrophy with strain pattern [1]

The patient's initial troponin is < 0.010 ng/mL. You decide to use the HEART Score to determine an appropriate disposition for this patient:

Figure 2: Sample HEART Score

The patient's HEART Score is 5 which is associated with a 12% to 16% risk of Major Adverse Cardiac Event (MACE) within 6 weeks. You decide to place the patient on observation for “intermediate-risk chest pain”. The patient will undergo serial troponin and ECG testing and possibly be evaluated by cardiology for provocative testing.

You used appropriate testing and a decision-making tool, but your clinical gestalt is that this patient is not having acute ischemia. You can’t help but wonder, “Was observation the most appropriate disposition for this patient?”

The Problem

Chest pain is the second most common complaint in the ED with 7,194,000 visits per year in 2020.[2] With etiologies as benign as musculoskeletal pain or gastroesophageal reflux and as dangerous as acute myocardial infarction (AMI) or aortic dissection, it proves to be a difficult complaint to evaluate. ED visits for chest pain incur a substantial health care cost burden, especially when patients are admitted for cardiac workups but ultimately diagnosed with non-cardiac etiologies.[3,4] For patients in whom AMI is being considered, multiple different prognostic and risk-stratifying tools have been developed. A common risk stratification scoring system used in the ED is the HEART Score.[5] This tool uses the following components to stratify patients into risk categories: history of symptoms, ECG evaluation, patient age, risk factors (hypertension, dyslipidemia, diabetes, obesity, smoking, atherosclerotic disease – i.e. prior AMI, percutaneous coronary intervention (PCI), coronary artery bypass graft (CABG) – cerebral vascular accident, transient ischemic attack, or peripheral artery disease) and initial troponin level. Risk categories correspond to the following percent chance of MACE within 6 weeks are:

1 - Low: 0 to 3 (1% to 2%)

2 - Intermediate: 4 to 6 (12% to 16%)

3 - High: 7 to 10 (50% to 65%)

The American Heart Association and The American College of Cardiology has created the following disposition recommendations in relation to HEART Score (and other scoring systems)[6]:

1 - Low-risk patients without other clinical suspicion for active ischemia can safely be discharged

2 - Intermediate-risk patients should be observed or admitted for further evaluation and treatment

3 - High-risk patients should be admitted for further evaluation and treatment

This begs the question: can it truly be that simple?

Let’s Review the Data

There is overall good consensus on the disposition of low- and high-risk HEART Score patients. Derivation and validation studies have shown a MACE risk of 1.0% to 2.5% in patients with a HEART Score of 0 to 3.[5,7-9] The GRACE review for recurrent, low-risk chest pain looked at over 300 papers with outcome recommendations and suggests that it is safe to discharge the low-risk HEART score patient.[9]

Review of the data from high-risk patients suggests that hospitalization may be indicated. However, in this group, the data is more convoluted; MACE risk ranges from 4.2% to 72.7%.[5,7,8] Despite the wide range in the data, the patients with HEART Score 7 to 10 may often be admitted based on presumed, unacceptable risk (even at 4.2%) or their clinical picture. For example, for a patient to reach a HEART Score of 7, they must have at least one of the following: highly suspicious history, significant ST-deviation on ECG, or troponin elevated one to three times the normal limit. Regardless of stratification tool, these findings would prompt many physicians to recommend hospitalization.

Although the data is equally inconsistent for intermediate-risk patients as it is for those who are high-risk, the clinical picture in these cases becomes more complex, presenting providers with a uniquely difficult disposition decision. The initial validation studies for the HEART Score showed MACE rate of 11.6% to 20.3% for intermediate-risk patients which drove initial recommendations to admit these patients.[5,7] Other studies since have shown rates ranging more widely from 1% to 28%.[8-12] The question stands: how can we best assess these intermediate-risk patients to determine who needs hospitalization for more testing and who can be spared? To better answer this question, we can start with investigating the sources of these inconsistencies.

Intermediate-Risk Patients: Why So Inconsistent?

When reviewing the studies that reported starkly different MACE rates in intermediate-risk patients, a few discrepancies stand out. First, several papers differ on their definition of “MACE”. Papers reporting higher percentages of MACE typically included AMI, left heart catheterization, PCI, CABG, and death as endpoints. [5,7-10,12]  In contrast, papers reporting lower percentages of MACE tended to focus on the incidence of ST-elevated myocardial infarction (STEMI), cardiac arrest, life-threatening arrhythmias, and death as endpoints.[11,13]

Studies that report higher MACE rates are reporting outcomes that may be clinically inconsequential, such as non-STEMI or left heart catheterization that is normal; in these cases, patients may have undergone non-beneficial and potentially harmful, invasive testing. In contrast, studies reporting lower MACE rates may be deemphasizing the clinical benefit of hospitalization by excluding patients who, despite not having AMI, had diagnostic, positive left heart catheterization resulting in the initiation of optimal medical management. Note that there is controversy about the true benefit of PCI/CABG for stable CAD.[14-17]

There also appears to be inter-provider variability in ratings of patients in the intermediate HEART Score category, perhaps further contributing to inconsistencies.[18] While the overall inter-provider variability was small, the greatest degree of variability was the scoring of the patient’s history and ECG interpretation among the intermediate-risk patients. There is also evidence that many subjective components of a history that we traditionally deem more or less ‘concerning’ for AMI do not consistently correlate with confirmed AMI.[19] Even within a relatively objective component of “risk factors”, there can be variability. The cardiovascular risk profile of a patient with well-controlled hypertension, minimal hypercholesterolemia on a statin, and well-controlled diabetes on metformin is very different from a patient who is non-adherent with medications with systolic blood pressure of 180 and a hemoglobin A1c of 13%. A spectrum of risk within one factor cannot be reflected in the risk stratification score.

Lastly, the components that comprise a HEART Score 4 to 6 can have varying clinical importance. The following patients both have a HEART Score of 5, but the clinical pictures painted by their scores tell two different stories:

Figures 3 and 4: Sample HEART Scores

One patient qualifies due to age, significant ECG findings, a few chronic risk factors, and a positive troponin while the other qualifies only due to age, chronic risk factors, and a moderately suspicious story. When considering the inter-provider variability on judgment of history and ECG, it becomes clear why it is much more difficult to definitively risk stratify these “intermediate risk” patients based on a number scoring system.

Let’s look into this a little more.

Score Components Matter

The initial validation study done by Backus et al. showed evidence that troponin and ECG changes were the most critical components of the HEART Score, with 14 of the 16 reported AMI outcomes having an initial troponin elevation.[7] Some studies have shown that the only factors that are independent predictors of MACE were history, ECG, and troponin level.[7,8]    

Weinstock et al. evaluated 11,230 ED patients presenting with chest pain and 2 negative troponins and found a 0.18% incidence of MACE (defined as life-threatening arrhythmia, STEMI, cardiac or respiratory arrest, or death during hospitalization).[13] To take it a step further, they then narrowed the sample size to patients without concerning factors by excluding those with abnormal vital signs, ischemic ECG patterns, new LBBB, or paced rhythm. These results showed a MACE in only four (0.05%) of 7,266 patients. Weinstock et al. included a secondary endpoint of “possible or definite MI” which occurred in 28 (0.39%, 95% CI 0.26% to 0.56%) of these 7,266 patients. 

In light of Weinstock et al., which provided data on patients with intermediate heart scores solely from age, history, and risk factors, this now begs the question: how important are risk factors? Several studies have found little correlation between the presence of cardiovascular risk factors and incidence of AMI; rather, these factors have higher correlation with CAD, a chronic condition that may be diagnosed in an outpatient setting.[20,21] Interestingly, Han et al. found that among the >65 age group, the number of risk factors had no correlation with diagnosis of AMI.[20] 

Another consideration for the patients identified in Weinstock et al. is that they may not have warranted assessment with the HEART Score in the first place. Validation and review studies included all patients presenting to the ED with “chest pain” or, in some cases, “chest pain equivalents”.[7-14,18,19] However, when a patient presentation is clearly inconsistent with AMI (low MI prevalence and low pre-test probability), the positive predictive value of a highly sensitive/less specific decision tool will be lower. When the HEART Score is applied to a patient older than 65 with multiple risk factors presenting with clear, non-cardiac chest pain, they will immediately fall into the intermediate-risk category which may influence a physician’s disposition decision. This may have occurred to the patient in the case presentation. 

In a large retrospective study, Spiegel et al. focused specifically on chest pain patients with HEART Scores 4 to 6, a negative initial troponin, and a “non-ischemic” ECG. Among 1,118 patients, six (0.5%, 95% CI 0.2% to 1.2%) had MACE.[11] In this study, endpoints were defined as life-threatening arrhythmia, STEMI, cardiac arrest, or death. A secondary outcome of non-STEMI was reported in six patients, six of whom had troponin elevation within 6 hours of presentation. Importantly, this study also looked at the interventions that were performed in these admitted patients. 216 (19.3%) of 1,118 patients had provocative testing (mostly nuclear stress tests). Of these, 36 had positive stress tests findings leading to left heart catheterization. Five left heart catheterizations were entirely clear, and only five patients had subsequent PCI or CABG. This sheds light on the over-testing and over-catheterization, exposing patients to an unnecessary procedure that is not without risk.[22-24]

Studies support the idea that not all patients with a HEART Score of 4 to 6 carry the same MACE risk.[11,13] The AHA guidelines accept a 1% to 2% MACE rate in the decision to discharge patients, in line with ACEP’s acceptance of a missed diagnosis rate for 30-day MACE in NSTEMI.[6,25] These acceptable miss rates are somewhat higher than surveyed emergency physician acceptable miss rates, which range from 0.01% to 1%.[26,27] This data shows that when taking the score components and clinical picture into consideration, intermediate-risk patients with nonischemic ECGs and negative initial troponins may fall within this widely accepted risk rate, even when including non-clinically relevant outcomes. Despite this, current AHA protocols suggest observation or admission for all of these patients.[6]

Back to the Case 

Let us now consider: How does this data apply to our patient with an unchanged, nonischemic ECG and a negative troponin?

We owe it to our patient to think beyond a number scoring system. Instead of letting the score  dictate our disposition decision, we must think critically about the patient in front of us within the realms of both beneficence and nonmaleficence. How concerning is their story? Does this patient have confirmed CAD? What specific testing do I think this patient needs urgently? Are there any need for urgent testing? Beyond the immediate clinical questions, there are several other factors that may go into making the disposition decision for our patient. To name a few:

1 - Does the patient have reliable primary care follow-up to assess resolution or evolution of their symptoms and provide appropriate treatment and referral for testing if indicated?

2 - When will the patient realistically be able to receive an outpatient cardiology evaluation for provocative testing, echocardiography, or medication management as needed?

3 - Is the patient at high-risk for contracting other illnesses or having iatrogenic adverse outcomes while admitted to the hospital (eg. immunocompromised)?

4 - How distressed is the patient regarding their symptoms?

5 - How engaged is the patient in shared decision-making?

The evidence to guide further management of this patient with intermediate-risk with further cardiac testing such as coronary CT angiography (CCTA) remains controversial. CCTA may offer a path forward with this patient, and notably, CCTA is the only recommended test supported by a high level of evidence for moderate risk patients without known CAD and an initial negative workup.[6] If the CCTA returns negative for CAD, then you can safely discharge this patient. Previous meta-analyses have shown pooled sensitivity of 98% to 99% for obstructive CAD, and prior randomized controlled trials have demonstrated higher rates of discharge from the ED, shorter lengths of stay, and higher CAD detection rates without increased risk of MACE when compared to “traditional care”.[28-30] 

While it is true that a hospitalization may expedite access to confirmatory testing for CAD (eg. CCTA, invasive coronary angiography), it will also subject the patient to the risk of iatrogenic events and potentially unnecessary and harmful invasive procedures, add cost to the patient and hospital system, and consume hospital resources.[6,31,32] 

Ultimately, disposition decisions for intermediate-risk patients with chest pain will vary, but ought to be based on shared decision-making with the patient that considers the risks and benefits of hospitalization and additional testing.

Summary:

 - Though validated in all patients with chest pain, the HEART Score’s positive predictive value will be much lower for the patient whose history of present illness suggests a non-cardiac etiology

 - For those patients that fall into the intermediate category (HEART Score 4 to 6) with a nonischemic ECG and 1 or 2 negative troponins (at time 0 and 3 hr), the clinically relevant cardiac event rates are both < 1%, respectively [11,13]

- If one is concerned about potential undiagnosed underlying CAD, a CCTA is an effective, time-saving test to potentially exclude CAD

- Disposition decisions for intermediate-risk patients will ultimately vary and should consider shared decision-making between the patient and individual provider regarding the risks and benefits of additional testing