Introduction: Chest pain is a common reason or presentation to the Emergency Department (ED). Emergency physicians are often challenged to identify patients who are at high risk for an acute coronary syndrome (ACS), from those with low risk, who might benefit from early discharge without extensive diagnostic workup. The introduction of high sensitivity cardiac troponins have improved early rule out strategies.
Previous studies on high sensitivity troponins (hs-cTnT) have yielded two risk stratification strategies for rule in / rule out of myocardial infarction (MI); using an initial hs-cTnT below the limit of quantification (LOQ) alone and using an algorithm with initial and 1-hour hs-cTNT (0/1-h algorithm), as currently recommended by the European Society of Cardiology.
These strategies are often combined with a risk score, of which in the United States (US) the HEART score is the most commonly used. However, due to the relatively late introduction of hs-cTnT in the US, multicenter US data for these risk stratification strategies using hs-cTnT are lacking.
The goal of the current study was to prospectively evaluate the diagnostic performance of a hs-cTnT assay (Roche, Basel Switzerland) for the detection of 30-day MACE using two different strategies: an initial hs-cTnT <LOQ alone and a 0/1-h algorithm, both with and without clinical variables (ECG interpretation and HEART score).
Paper: Allen et al. Diagnostic Performance of High Sensitivity Cardiac Troponin T Strategies and Clinical Variables in a Multisite United States Cohort. Circulation 2021. PMID 33474976
Clinical Questions:
- What is the diagnostic performance (safety and efficacy) of the Roche hs-cTnT assay for the detection of 30-day MACE (plus the composite of cardiac death or MI) at 30 days using an initial hs-cTnT below the limit of quantification (LOQ) and a 0/1-h algorithm?
- Same question as above both with and without the inclusion of ECG interpretation and HEART score?
What They Did:
- Prospective observational cohort study of ED patients with acute chest pain or other symptoms suggestive of ACS
- Eight US EDs between January 25, 2017 to September 6, 2018
Inclusion:
- Age ≥ 21
- Presenting with chest discomfort or other symptoms consistent with possible ACS (defined by the ED provider ordering serial troponins)
- Enrollment < 1 hour of site’s first clinical blood draw
- Have a 2nd standard of care cTnT measurement within the next 2-12 hours
Exclusion:
- ST-segment elevation MI at presentation
- SBP < 90 mmHg
- Life expectancy < 90 days
- Non-cardiac illness requiring admission
- Lack of capacity to provide consent
- Inability to be contacted for follow up
- Non-English speaking
- Pregnancy
- Prior enrollment in this study
Outcomes:
- Primary:
- 30-day major adverse cardiac events (cardiac death, MI, and coronary revascularization)
- Secondary:
- Index and 30-day composite of cardiac death or myocardial infarction
- Individual components of the MACE outcomes
Results:
- 1,462 patients were eligible for analysis
- 46.4% women, 53.6% men
- 37.1% African American
- Average age of 57.6 years (SD 12.9 years)
- Hs-cTnT at t = 0 in 1460 patients (99.9%), at t = 1 in 1432 patients (97.9%)
- Median time from ED arrival to baseline study blood drawn was 82 minutes (63 – 106 minutes)
- 30-day follow up completed in 1406 patients (96.2%), loss to follow up 3.8%
- MACE at 30-day occurred in 210 patients (14.4%)
- 9 patients (0.6%) with cardiac death
- 185 patients (12.7%) with MI
- 24 patients (1.6%) with coronary revascularization without MI
- MACE at index visit occurred in 181 patients (12.4%)
- 2 patients (0.1%) with cardiac death
- 169 patients (11.6%) with MI
- 11 patients (0.8%) with coronary revascularization without MI
- MACE after index visit (during 30-day follow up) occurred in 46 patients (3.1%)
- 7 patients (0.5%) with cardiac death
- 26 patients (1.8%) with MI
- 14 patients (1.0%) with coronary revascularization without MI
Critical Results:
Diagnostic Accuracy of hs-cTnT below LOQ (< 6 ng/L) Alone:
- Initial hs-cTnT <LOQ occurred in 32.8% (95% CI: 30.4-35.3%) of cohort
- NPV 98.3% for 30-day MACE (95% CI: 96.7-99.3%), 99.0% for 30-day Cardiac death and MI (95% CI: 97.6-99.7%), 99.2% for MI at presentation (95% CI: 97.9-99.8%)
- The addition of a low HEART score (0 – 3) to a hs-TnT <LOQ had an efficacy of 20.1% (95% CI: 18.1-22.3%)
- NPV at 30-day
- 99.0% for MACE (95% CI: 97.0-99.8%)
- 99.7% for MI or cardiac death (95%CI: 98.1-100.0%)
- NPV at index-visit
- 99.7% for MI (95% CI: 98.1-100.0%), sensitivity was 99.4% (95% CI: 96.7-100.0%)
- The addition of a non-ischemic ECG has no additional efficacy over the HEART score on a hs-cTnT below LOQ
- NPV at 30-day
Diagnostic Accuracy of the 0/1-h Algorithm:
- The 01/-h algorithm was completed in 1430 patients
- 826 patients (57.8%) were ruled out by the algorithm (hs-cTnT: 0 hr <12 ng/L and Δ 0-1hr <3 ng/L)
- 23 patients (1.6%) had MACE at 30-day (1 cardiac death, 13 MI, 9 index/30-day revascularization), all had non-ischemic ECG and 7 had a low-risk HEART score, 9 were early presenters
- NPV at 30-day
- 97.2% for MACE (95% CI: 95.9-98.2%), sensitivity 88.7% (95% CI: 83.5-92.7%)
- 98.4% for MI or cardiac death (95% CI: 97.3-99.2%), sensitivity was 92.9% (95%CI: 88.2-96.2%)
- Index visit MI occurred in 8 patients (1.0%)
- NPV at index-visit
- 99.0% for MI (95%CI: 98.1-99.6%), sensitivity was 95.1% (95%CI: 90.6-97.9%)
- 604 patients (42.2%) were unable to be ruled out by the algorithm
- PPV at 30-day
- 29.8% for MACE (95%CI: 26.2-33.6%), specificity was 65.4% (95%CI: 62.7-68.1%)
- 28.2% for MI or cardiac death (95%CI: 24.6-31.9%), specificity was 65.2% (95%CI: 62.5-67.8%)
- The addition of a low-risk HEART score to the rule out range of the 0/1h algorithm was effective in in 441 patients (30.8%)
- NPV at 30-day
- 98.4% for MACE (95%CI: 96.8-99.4%), sensitivity was 96.6% (95%CI: 93.0-98.6%)
- 99.3% for MI or cardiac death (95%CI: 98.0-99.9%), sensitivity was 98.4% (95%CI: 95.3-99.7%)
- NPV at index-visit
- 99.5% for MI (95%CI: 98.4-99.9%), sensitivity was 98.8% (95%CI: 95.7-99.9%)
- The addition of a non-ischemic ECG has no additional efficacy over the 0/1h algorithm in the rule out range
- NPV at 30-day
- PPV at 30-day
- 826 patients (57.8%) were ruled out by the algorithm (hs-cTnT: 0 hr <12 ng/L and Δ 0-1hr <3 ng/L)
The 0/1-h high-sensitivity troponin algorithm for 30 Day MACE, cardiac death or MI, and index-visit MI
ACS = acute coronary syndrome, CV = cardiovascular, ED = emergency department, MACE= major adverse cardiac event, MI= myocardial infarction, NPV = negative predictive value, PPV = positive predictive value
The 0/1-h high-sensitivity troponin algorithm combined with the HEART score for 30-day MACE, cardiac death or MI, and index-visit MI
ACS = acute coronary syndrome, CV = cardiovascular, ED = emergency department, MACE= major adverse cardiac event, MI= myocardial infarction, NPV = negative predictive value, PPV = positive predictive value
Early Presenters:
- 516 patients (35%) presented early (<3h) after onset of chest pain
- 79 (15.3%) patients had MACE
- Early presenters with initial hs-TnT measure <LOQ alone did only receive high NPV of 99.3% for MI at index-visits (95%CI: 96.3-100.0%)
- Early presenters with initial hs-TnT measure <LOQ and a HEART score of 0-3:
- NPV at 30-day
- 100.0% for MACE (95%CI: 97.2-100.0%)
- 100.0% for MI or cardiac death (95%CI: 97.2-100.0%)
- NPV at index-visit
- 100.0% for MI (95%CI: 97.2-100.0%), sensitivity was 100.0% (95%CI: 95.3-100.0%)
- Early presenters with hs-cTnT 0/1h algorithm in rule-out range alone did only receive high NPV of 99.4% for MI at index-visits (95%CI: 96.7-100.0%)
- Early presenters with hs-cTnT 0/1h algorithm in rule-out range and a HEART score of 0-3:
- NPV at 30-day
- 98.2% for MACE (95%CI: 94.9-99.6%), sensitivity was 95.9% (95%CI: 88.6-99.2%)
- 98.8% for MI or cardiac death (95%CI: 95.8-99.9%), sensitivity was 97.1% (95%CI: 89.9-99.6%)
- NPV at index-visit
- 99.4% for MI (95%CI: 96.7-100.0%), sensitivity was 98.3% (95%CI: 90.9-100.0%)
- NPV at 30-day
- NPV at 30-day
Strengths:
- This is the first and largest prospective multisite US cohort study to evaluate hs-cTnT strategies to date
- In this cohort the differences in time of chest pain onset were small and nonsignificant to the diagnostic performance of hs-cTn cut points
- The treating providers were blinded to the hs-cTnT and the HEART scores.
- Combination of a low-risk HEART score and an initial hs-cTnT measure < LOQ yielded high sensitivity and NPV for MACE even among early presenters
- Sensitivity analyses testing a variety of assumptions yielded similar diagnostic performance of the LOQ and the 0/1-h algorithm when used with or without clinical variables
Limitations:
- Although the study was conducted across eight US EDs, the sites were mostly urban academic medical centers and therefore results may not be generalizable to all US ED settings
- The cohort had a higher MACE rate than has been reported in many prior US studies
- The low PPV with the addition of the HEART score could lead to over-triage as well as over-testing
- Lost to follow-up rate was small, but the investigators were unable to contact <4% (56/1442) of the cohort, which may have caused misclassification and underestimation of MACE
- This study utilized only the Roche hs-cTnT assay, and results cannot be extrapolated to other hs-cTn assays
- Classification of patients as early or late presenters was susceptible to patient recall bias
- Subgroup analysis of early presenters was limited to 516 patients (35% of the total population), meaning the confidence intervals for this group are somewhat broader than the overall population
- No patient was managed according to the use of the hs-cTnT assay or the strategies evaluated, which means efficacy and safety of hs-cTnT strategies still need to be prospectively evaluated in the setting of implementation into clinical practice in a US population.
Discussion:
- This multisite, prospective study of hs-cTnT early rule out strategies in US ED patients suggests that adding a HEART score to these strategies improves safety
- When used alone, a single hs-cTnT measure <LOQ had insufficient NPV and sensitivity to rule-out 30-day MACE
- Interestingly, the addition of an ECG alone did not improve the diagnostic performance of either of the hs-cTnT strategies
- The HEART score, which includes ECG findings in its scoring along with other historical features, was able to improve the diagnostic performance of both hs-cTnT strategies in this analysis.
- The addition of a low-risk HEART score to an initial hs-cTnT value <LOQ, increases the NPV for 30-day cardiovascular death and MI to 99.7% and to 100% in early presenters, but with fewer patients to be ruled out
- The addition of a low-risk HEART score to a hs-cTnT 0/1-h algorithm in the rule-out range, increases the NPV for 30-day cardiovascular death and MI to 99.3%
- Among late presenters (those with chest pain onset > 3 hours), use of the LOQ at arrival for the cutoff used (<6 ng/l) yielded a 98.2% NPV for 30-day MACE
- This is the first large multicentre US study, which demonstrates that the hs-cTnT 0/1-h algorithm had a NPV of 99.0% for index visit MI, similar to prior (European) studies
- None of the patients with 30-day MACE who had initial hs-cTnT measures < LOQ or were in the rule-out range of the 0/1-h algorithm had an ischemic ECG as determined by their treating provider
- It’s worth mentioning that many patients had (near) normal levels of creatinine, raising the question what these results would look like in patients with decreased renal function
Author Conclusion: “In a prospective multisite US cohort, an initial hs-cTnT <LOQ combined with a low-risk HEART score was 99% NPV for 30-day MACE. The 0/1-h hs-cTnT algorithm did not achieve a NPV >99% for 30-day MACE when used alone or with a HEART score.”
Clinical Take Home Point: Addition of the HEART score to an initial hs-TnT below the limit of quantification improves sensitivity and NPV for cardiac events increasing patient safety, but this strategy does rule-out fewer patients.
Guest Post By:
Benjamin M. Gerretsen, MD
Emergency Medicine
Erasmus University Medical Center
Leiden, Netherlands
Twitter: @bmgerretsen
Barbra Backus, MD
Emergency Physician
Erasmus University Medical Center
Netherlands
Twitter: @barbrabackus
References:
- Allen et al. Diagnostic Performance of High Sensitivity Cardiac Troponin T Strategies and Clinical Variables in a Multisite United States Cohort. Circulation 2021 (ahead of print). PMID 33474976
Post Peer Reviewed By: Salim R. Rezaie, MD (Twitter: @srrezaie)