The HEART Pathway Randomised Trial. Identifying Emergency Department Patients With Acute Chest Pain for Early Discharge

Mahler et al. Circulation Cardiovascular Outcomes 2015; 8:00-00. doi:10.1161/CIRCOUTCOMES.114.001384

Clinical Question

• In patients presenting to the Emergency Department with chest pain, does the HEART Pathway compared with usual care, reduce objective cardiac testing?

Design

• Randomised control trial
• Block randomisation stratified by the presence of known coronary artery disease
• Intention to treat analysis
• Patients randomised to the HEART pathway received a second HEART score assessment by a study investigator blinded to the initial physician score
• Consensus of 2 reviewers blinded to treatment arm adjudicated subjective outcomes with disagreement settled by consensus or involvement of 3rd reviewer
• Follow up at 30 days with telephone interview using a validated questionnaire. Patients with incomplete follow up considered to be free of 30 day major adverse cardiac outcomes (MACE)
• Convenience sample (data collected on 6 days per week, 80 hours).
• Power calculation: Powered at 90% to detect a 15% reduction in objective cardiac testing within 20 days, with expected loss to follow up rate of 10%, and a 2 sided alpha of 5%

Setting

• Single centre, tertiary academic Emergency Department, USA.
• Data collected: 9/2012 to 2/2014.

Population

• Inclusion criteria: >21 years old presenting with symptoms suggestive of acute coronary syndrome (ACS)
  • Both an ECG and troponin were ordered for the evaluation of ACS
• Exclusion criteria: new ST elevation >1mm, hypotension, life expectancy <1 year, requiring admission for non-cardiac cause, previous enrolment, non English speaking
• 282 patients randomized
• Comparing baseline characteristics of the HEART group vs. control group:
  • Female: 57% vs. 57%
  • Race:
    • White: 64% vs. 66%
    • Black: 34% vs. 33%
  • Risk factors:
    • Smoking: 30% vs. 24%
    • Hypertension 53% vs. 58%
    • Hyperlipidemia: 43% vs. 43%
    • Diabetes mellitus: 22% vs. 19%
  • TIMI risk score >1: 43% vs. 45%
  • Previous coronary disease: 20% vs. 21%

Intervention

• The HEART pathway (n=141)
  • HEART score determined with serial Troponin measurements at 0 and 3 hours
  • The HEART score is based on 5 components: history (clinical suspicion of cardiac sounding history), ECG, age, risk factors and troponin. Each component is assessed on a scale of 0-2 and then added to produce a final score
  •  47% (n=66) of patients had a HEART score <=3 and negative Troponins at 0 and 3 hours
    • Classified as low risk and suitable for early discharge without further cardiac testing
    • 32% (n=21) of low risk patients went on to have objective cardiac testing within 30 days of whom 2 patients had a positive exercise stress echocardiogram. One of these presumed to be false positive result by cardiologist, the second had a negative coronary angiogram
  • 53% (n=75) of patients had a HEART score >3 +/or positive Troponin
    • Classified as high risk and admitted/referred to cardiology
    • 79% went on to have objective cardiac testing within 30 days
  • 4 patients lost to 30 day follow up

Control

• Usual care (n=141)
  • Clinicians encouraged to follow American College of Cardiology guidelines recommending serial cardiac biomarkers and objective cardiac testing before discharge from hospital
  • 6 patients lost to 30 day follow up

For both intervention and control groups:

• Serum Troponins were performed using the ADVIA Centaur platform TnI-Ultra high sensitivity troponin assay.
• Cardiac testing modalities routinely available included exercise stress echo, dobutamine stress echo, coronary CT angiogram, stress nuclear imaging, stress cardiac MRI or invasive coronary angiogram
• Care delivered/investigations performed at physicians discretion

Outcomes

• Primary outcome: rate of objective cardiac testing (stress test, CT angiography or cardiac angiography) within 30 days
  • Significantly reduced in the HEART pathway compared with usual care
    • 56.7% vs. 68.8%, absolute risk reduction 12.1%, p=0.048
    • NNT = 9
    • Fragility index = 1
• Secondary outcomes: comparing HEART pathway with standard care
  • Early discharge rate (discharge from ED without objective cardiac testing) – significantly increase in HEART pathway
    • 39.7% vs. 18.4%, p<0.001
  • Length of stay – significantly reduced in HEART pathway
    • median 9.9 hours vs. 21.9 hours, p=0.013
  • Cardiac related recurrent ED visits – no significant difference
    • 2.8% vs. 4.3%, p=0.75
  • Cardiac related non-index hospitilisation – no significant difference
    • 3.6% vs. 2.8%, p>0.999
  • MACE (defined by composite of all-cause mortality, myocardial infarction or coronary revascularisation within 30 days) – no significant difference
    • 5.7% vs. 6.4%, p=1
  • Test characteristics of HEART pathway for determining MACE
    • Sensitivity: 100% (95% C.I. 63.1-100)
    • Specificity: 49.6% (95% C.I. 40.8-58.4)
    • +ve Likelihood Ratio 1.99
    • -ve Likelihood Ratio 0 (No patients indexed as low risk by the HEART pathway had a MACE)
• Interobserver agreement of HEART score was acceptable, k=0.63

Authors’ Conclusions

• Use of the HEART pathway significantly decreased objective cardiac testing, resulted in an early discharge rate of nearly 40% and cut median length of stay by 12 hours. No patients identified for early discharge had MACE at 30 days and the HEART pathway was not associated with increased cardiac-related return ED visits or non-index hospitalizations.

Strengths

• Randomised control trial
• Clinically relevant question and endpoints investigated
• Pragmatic study that allowed provider discretion in adherence to the HEART pathway yet still analysed with an intention to treat method
• Registered with clinicaltrials.gov

Weaknesses

• Single centre US study with small sample size
• Convenience sample rather than consecutive patients with no description of what times patients were eligible for enrolment.
• Study not powered to detect differences in MACE between groups
• Level of agreement for assessment of HEART score was only 0.63
• Loss to follow up of 4% of patients
• Fragility index of 1 indicates that although statistically significant the difference between groups in relation to the primary endpoint was very small.
• A large proportion of patients across the trial received objective cardiac testing (68% in the usual care group) and this may not be reflective of UK practice in which low risk patients are probably less likely to receive objective cardiac testing.

The Bottom Line

• The results of this study show that the HEART pathway may provide a useful risk stratification tool for assessing patients with potential acute coronary syndromes in the emergency department. There was a reduction in objective cardiac testing and length of stay while maintaining a very low risk of missing significant cardiac events. However before I start adopting the pathway in my practice I would like to see further larger studies looking at its use in a UK population.

External Links

• [article] The HEART Pathway Randomized Trial Identifying Emergency Department Patients With Acute Chest Pain for Early Discharge
• [further reading] The original HEART score article with a description of how points are scored
• [further reading] MD Calc: HEART Score Calculator
• [further reading] R.E.B.E.L.EM: Is it Time to Start Using the HEART Pathway in the Emergency Department?
• [further reading] The SGEM: Groove is in the HEART Pathway

Metadata

Summary author: Ben Short
Summary date: 25th November 2016
Peer-review editor: @davidslessor