Chest pain in the ED: HEART score, the HEART pathway and who can go home

Calculator for this article: our HEART score calculator.

Chest pain is a daily problem in emergency medicine, and the work is identifying, safely and quickly, the patients who can go home. The HEART score was built for that, and the HEART pathway turned it into a protocol. They are related, but they are not the same thing.

The HEART score

HEART scores five elements from 0 to 2 points each: History, ECG, Age, Risk factors and Troponin, for a total of 0 to 10.^[1]

The original 2008 paper prints both "45-65" and "≥65" for the age bands, so age 65 is ambiguous in the source; our calculator scores 65 as 2 points and says so.

In the 2008 derivation cohort of 120 patients, the endpoint occurred in 2.5% with a score of 0 to 3, 20.3% with 4 to 6, and 72.7% with 7 to 10.^[1] The prospective validation in 2,440 unselected patients across ten hospitals found 6-week major adverse cardiac events (MACE) in 1.7%, 16.6% and 50.1% of the same three groups, with 36.4% of patients in the low group. Its C-statistic (0.83) was higher than TIMI (0.75) and GRACE (0.70) in that cohort.^[2]

Two things to notice. The score uses a single troponin at presentation. And 1.7% is low, but it is not zero.

A worked example

A 58-year-old with a moderately suspicious history, nonspecific repolarization changes, hypertension and hyperlipidemia, and a normal first troponin scores 1 + 1 + 1 + 1 + 0 = 4. That is intermediate risk, with a 6-week MACE rate of 16.6% in the validation cohort,^[2] even though nothing about the presentation feels alarming. The same patient at 44 would score 3. Age alone moves patients across the line, which is one reason the score should be calculated rather than estimated. Record the components, not just the total, so the next clinician can see where each point came from, especially the history element, which rests on judgment.

The HEART pathway

The HEART pathway adds what the score does not have: serial troponins. In the randomized trial that defined it, patients with a low-risk HEART score (0 to 3) and negative troponins at 0 and 3 hours were recommended for discharge without further testing.^[3]

Results in 282 patients, pathway versus usual care:^[3]

• Objective cardiac testing at 30 days: 56.7% vs 68.8%, a 12.1 percentage-point reduction.
• Length of stay: 9.9 vs 21.9 hours.
• Early discharge: 39.7% vs 18.4%.
• No patient identified for early discharge had a major adverse cardiac event within 30 days.

The paper's text defines low risk as a score of 0 to 3; one of its tables prints "less than 3". If you are writing a local protocol from the primary source, use the text definition and say so.

The pathway was then implemented across a health system in 8,474 adult emergency patients. It classified 30.7% as low risk, 0.4% of whom had death or MI within 30 days, and hospitalization fell from 61.6% to 55.6% (adjusted odds ratio 0.79).^[4]

Does using the score change outcomes?

A stepped-wedge cluster randomized trial in 3,648 patients compared HEART-based care with usual care. Six-week MACE during HEART care was 1.3% lower than during usual care, within the noninferiority margin, but there were no statistically significant differences in early discharge or readmissions. In low-risk patients the MACE incidence was 2.0%.^[5]

That result is worth sitting with. Introducing a score did not, by itself, send more people home. The pathway trials, which tied the score to serial troponins and an explicit discharge recommendation, did. A score changes practice when it is wired to a decision.

Who can go home

Reading the evidence together:

• A low HEART score alone is not a discharge rule. It was validated with a single presentation troponin and a 6-week MACE rate of 1.7% in the low group.^[2]
• Low HEART score plus negative serial troponins is the pathway that was tested for early discharge, with 0.4% death or MI at 30 days in the implementation cohort.^[4]
• The rule assumes the question is ACS. A HEART score does nothing for aortic dissection, pulmonary embolism or other dangerous causes of chest pain. Those need their own consideration and their own tools.
• Local troponin assays and protocols vary. The pathway trials used 0- and 3-hour troponins with the assays of their time. Your institution's high-sensitivity protocol may use different timings; follow it.

Where decision support helps

The failure mode in chest pain is rarely the arithmetic. It is a score calculated before the second troponin, a dissection not considered because the HEART score was low, or a discharge without the follow-up that the pathway assumed. Pogosh reads the note, offers the HEART score as a calculator to run (never automatically), and returns the next steps as structured items: the troponins, the imaging that fits the differential, the escalation criteria, and a post-ED disposition with follow-up instead of "ED" as the answer. A separate, strict emergency check looks for time-critical presentations in the same note. The clinician decides who goes home.