JUCM Article Review: The HEAR Score

JUCM Article Review: The HEAR Score

Josh Russell,MD, and Michael Weinstock, MD

Based on the JUCM editorial: 

No Troponin, No Problem: Reimagining Chest Pain Assessment in Urgent Care
Joshua Russell, MD, MSc, FCUCM, FACEP; Mike Weinstock, MD

General:

• Many UC providers loathe chest pain patients because we are worried we don’t have the tools to exclude myocardial infarction (MI).
• Chest pain is a common complaint in urgent care, but MI is rare. A study by Hess et al JAMA showed the patients with nSTEMI only had a mortality for sudden cardiac death of 0.79% at 6 months Hess PL, Wojdyla DM, Al-Khatib SM, Lokhnygina Y, Wallentin L, Armstrong PW, Roe MT, Ohman EM, Harrington RA, Alexander JH, White HD, Van de Werf F, Piccini JP, Held C, Aylward PE, Moliterno DJ, Mahaffey KW, Tricoci P. Sudden Cardiac Death After Non-ST-Segment Elevation Acute Coronary Syndrome. JAMA Cardiol. 2016 Apr 1;1(1):73-9. doi: 10.1001/jamacardio.2015.0359. PMID: 27437658.
• Chest pain is concerning to patients predominantly due to the possibility of MI, which represents between 1% and 3% of ambulatory visits for acute complaints (Kleton et al.). 
• Only about 10% to 12% of patients presenting to an ED with concerns for acute coronary syndrome (ACS) will go on to have a major adverse cardiac event (MACE) within the subsequent 30 days (Smith, Weinstock). 
• Rates of immediate ACS in ED populations are even lower (5%-10%; Farnoff et al.).
• The best estimate from the recent literature that can be extrapolated to the UC comes from a European study of acute primary care visits, where the investigators found the 6-week risk of MACE to be <5% (Kleton et al.).

Most Studies of Chest Pain Measure the Wrong Outcomes

• Most studies reporting outcomes of patients with acute chest pain are not directly relevant to the UC clinician, not only because they are ED based, but also because they report MACE over the subsequent weeks as their primary endpoint. 
• MACE was developed in the late 1990s by cardiologists as a composite endpoint for measuring outcomes after coronary interventions (percutaneous coronary intervention [PCI]; Serruys et al.). Patients are classified as having a MACE if they die, have an MI, or have a repeat PCI during some specified period, usually 4 to 6 weeks. While convenient for statistical analysis, these composite endpoints are difficult to interpret because death and “needing to have a procedure” are far from equivalent outcomes. However, research using MACE counts these events equally.
• A second problem is that the timeline for cardiac events measured in many studies is not relevant to our clinical predicament (Weinstock, 2019). We seek to know the near-term safety of the patient.  If we knew they could make it to outpatient follow-up, we’d feel much more comfortable foregoing an immediate ED referral. 
• Many studies examining the various ACS clinical decision rules (CDR) look at MACE over a longer period (usually around 1 month) than is relevant to UC providers (Weinstock, 2019). This is problematic because immediate risk for UC patients of sudden death or serious MI has not been specifically studied. It’s certainly lower, but how much lower we can’t really say without UC-specific data.

Clinicians Do Not Tolerate Uncertainty when Considering ACS

• Fear of litigation is a valid concern. “Failure to diagnose MI” remains a leading cause of US malpractice claims. (Brown et al.).
• A recent study by Samuels et al. (Weinstock, Journal of Urgent Care Medicine; ahead of print) found that among 126 emergency providers in varying roles, half were uncomfortable with missing an acute MI even 0.1% of the time. Even though the American College of Emergency Physicians (ACEP) has stated that an ACS miss rate of 1% to 2% is acceptable (and perhaps even unavoidable), acute care providers continue to approach patients with chest pain with an overabundance of caution. 
• If we could say with confidence that there is a <2% chance of MI, we’d be well protected by the ACEP policy, and the current stream of excessive ED referrals, testing, and admissions could be significantly mitigated. 
• Over the past decade, several CDRs have been developed to address this very conundrum, the HEART score being the most prominent and well validated. But there’s a catch.

Most ACS Prediction Tools Don’t Work in Urgent Care 

• Most prediction tools require serum troponin testing, which is available in <10% of UC centers in the United States. 
• For ED patients, on the other hand, the recent development of the HEART and EDACS (Emergency Department Assessment of Chest Pain Score) scores has proven to be highly useful in identifying a large proportion of patients presenting with concerns for ACS who can safely be discharged without further immediate workup. 
• These tools, especially the HEART score, have been widely adopted by emergency clinicians, who now can discharge many more patients with chest pain and still sleep well at night (Mahler et al.).

A HEAR(-T) Score for the Rest of Us

• The HEART score, first developed in 2008 (Six et al.), is a clever acronym that combines (1) History, (2) ECG findings, (3) Age, (4) coronary artery disease (CAD) Risk factors, and (5) Troponin values to categorize patients as low, moderate, or high risk for ACS. Since being introduced, the HEART score has been validated by multiple investigators and found to be a reliable means of risk-stratifying patients with chest pain for risk of MACE over the subsequent weeks, with a sensitivity >98% for low HEART score patients (Backus et al.). 
• The need for troponin testing in the calculation of a HEART score has left UC providers feeling (somewhat appropriately) resigned to continue the status quo practice of near-automatic ED referrals for all but the lowest risk patients (ie, anxious adolescents). This has resulted in an abundance of low-risk ED referrals with an accompanying line in the chart: “Cannot r/o ACS without troponin.” But do we actually need a troponin to exclude ACS in low-risk patients with chest pain? 
• While the HEART score may be the best-known clinical decision tool for chest pain presentations, its lesser known cousin, the “HEAR” or “HEAR(-T)” score, has recently been externally validated with promising results. It seems the dogma of mandatory troponin testing when considering ACS may not be as ironclad as we’ve thought in recent decades – especially for the very low-risk patients. 
• In 2020, Smith et al. first described the use of a HEART score without troponin testing (ie, HEAR score) applied retrospectively to over 4,000 ED patients from the original HEART score study population. They found that a HEAR score of 0 or 1 occurred in 9% of patients and was 97.8% sensitive for ruling out 30-day MACE in this population. As ACEP has codified a 2% acceptable miss rate for ACS, this sensitivity almost exactly meets the minimum necessary for an acceptable “test” to be clinically useful in this situation. (Interestingly, the addition of a single troponin test in this study did not improve the sensitivity of the rule either.)
• More recently, O’Reilly and colleagues published the results of external validation of the HEAR score. They performed a secondary analysis of data collected in a prospective cohort study of 820 patients presenting in an urban Canadian ED with symptoms concerning for ACS. Improving on the clinical utility of the original HEAR study, O’Reilly et al. included patients with known CAD (who were excluded from the initial study) and used both 30-day MACE and immediate risk of MI diagnosed within 24 hours of ED presentation as co-primary endpoints. Importantly, patients with ischemic changes or new arrhythmia on ECG, advanced renal failure, MI within the prior month, and those <25 years of age were excluded. 
• They found that nearly 25% of patients had a HEAR score of 0 or 1. Confirming that low-risk patients are indeed low risk for bad near-term outcomes, only 1 patient in the low-risk group (score of 0 or 1) had an MI or 30-day MACE event. This yielded a sensitivity of HEAR <2 for 30-day risk of MACE or immediate MI of 98.3% to 99.2%. Better yet, for patients with a HEAR score of 0, sensitivity was 100%.
• Given that UC centers tend to see younger, healthier, lower-acuity patients with chest pain compared to the ED population, it’s likely that an even greater proportion of UC patients will actually fall into this low risk (ie, score 0 or 1) group. This means that by applying the HEAR rule, there is now an evidence base for discharging low-risk patients directly from UC. Coupled with the support of ACEP’s clinical policy on acceptable ACS miss rates, UC providers should feel confident that this is a reasonable practice. 

Cautions in Applying the HEAR Score

• It is important to remember the limitations of CDRs in clinical practice. First, CDRs (including the HEAR score) are developed to exclude conditions, not to make diagnoses (Adams et al.). Patients with HEAR scores of 0 or 1 can be safely presumed to be low enough risk for discharge from UC without immediate ED referral, but patients with scores >1 do not necessarily warrant immediate 911 activation. 
• It is not appropriate to use the HEAR score to justify your disposition decision in such patients. In other words, a “negative” HEAR score is meaningful but a “positive” result is not. In fact, the specificity of a score >1 for one of the adverse cardiac outcomes was an unimpressive 19% to 26% in the O’Reilly validation study. 
• Second, a CDR can only be applied validly to the same type of patients as those included in the studies from which it was derived. For example, patients <25 years and with end-stage renal disease were excluded in the HEAR validation study. Therefore, the rule can’t be relied upon in these patients unless a subsequent study that does not exclude these patients produces similar results. 

A New Approach When Considering ACS in UC

• Although most patients with chest pain who present to UC are exceptionally low risk for ACS (and even more so for sudden cardiac death), providers are extremely intolerant of missing an MI. A recent ACEP policy statement, however, provides top cover for an approach to evaluation for ACS that results in a miss rate <2%. 
• While the original HEART score is inaccessible to most UC clinicians due to lack of troponin testing, the ability to obtain an ECG is nearly universal.
• When patients present to a UC center with chest pain or symptoms that are concerning for ACS, they can be approached initially in the standard fashion: rapid rooming, vitals, and ECG. If the patient has a ST-elevation MI (STEMI) or other clear signs of ischemia, 911 activation is appropriate.
• For the vast majority of patients, the ECG will be reassuring and you’ll be able to take some time to look up and apply the HEAR score. With a reassuring history and ECG, a large proportion of patients can safely be ruled out for immediate and 30-day MACE (provided the HEAR is score <2). 
• For the rest of the patients, we can continue to use our clinical gestalt, appreciating its shortcomings, as well as shared decision-making regarding the necessity of immediate versus as-needed ED referral and 911 activation. 
• Applying this strategy in chest pain management, rather than quickly dismissing patients due to lack of troponin testing, will be appreciated by your patients, who certainly want to avoid the ED if possible. Most importantly, it will achieve this in an evidence-based fashion, avoiding bad outcomes not only for our patients but for ourselves as well. 

References:

Clinical prediction rules
Adams ST, Leveson SH. BMJ 2012;344:d8312. doi: 10.1136/bmj.d8312.

A prospective validation of the HEART score for chest pain patients at the emergency department
Backus BE, Six AJ, Kelder JC, et al. Int J Cardiol. 2013;168(3):2153-2158.

An epidemiologic study of closed emergency department malpractice claims in a national database of physician malpractice insurers
Brown TW, McCarthy ML, Kelen GD, et al. Acad Emerg Med. 2010;17(5):553-560. doi: 10.1111/j.1553-2712.2010.00729.x.

Young patients with chest pain: 1-year outcomes
Collin MJ, Weisenthal B, Walsh KM, et al. Am J Emerg Med. 2011;29(3):265-270. doi: 10.1016/j.ajem.2009.09.031.

Does this patient with chest pain have acute coronary syndrome? The rational clinical examination systematic review
Fanaroff AC, Rymer JA, Goldstein SA et al. JAMA. 2015;314(18):1955-1965.

Development of a clinical prediction rule for 30-day cardiac events in emergency department patients with chest pain and possible acute coronary syndrome
Hess EP, Brison RJ, Perry JJ, et al. Ann Emerg Med. 2012;59:115-125.e111.