Identifying Very-Low-Risk Patients for Future Myocardial Infarction or Death

Abstract

High-sensitivity cardiac troponin (hs-cTn) testing has enabled shorter time intervals between serial measurements when assessing patients with possible acute coronary syndrome (ACS) with the use of a single sample strategy proposed for early risk stratification. Specifically, a rapid rule-out for myocardial infarction (MI) is suitable for a pathway if the sensitivity for 30-day cardiac events is ≥ 99%. At the population level (n = 131,095 emergency department [ED] patients), low hs-cTn results alone yielded sensitivities < 99%. Test performance was improved with additional laboratory tests, with the combination of glucose, estimated glomerular filtration rate (eGFR), and hs-cTnT or hs-cTnI yielding sensitivities > 99.0% in patients ≥ 65 years of age. A recent clinical practice guideline from the American College of Cardiology/American Heart Association for the evaluation and diagnosis of acute chest pain lists different clinical decision pathways (CDPs) that may be used to further classify patients with suspected ACS into different risk categories.1Gulati M. Levy P.D. Mukherjee D. et al.2021 AHA/ACC/ASE/CHEST/SAEM/SCCT/SCMR guideline for the evaluation and diagnosis of chest pain: executive summary: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines.Circulation. 2021; 144: e368-e454Google Scholar These guidelines place emphasis on time of symptom onset for CDPs that use the limit of detection or low cutoff for hs-cTn assays for rapid rule-out.1Gulati M. Levy P.D. Mukherjee D. et al.2021 AHA/ACC/ASE/CHEST/SAEM/SCCT/SCMR guideline for the evaluation and diagnosis of chest pain: executive summary: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines.Circulation. 2021; 144: e368-e454Google Scholar However, from a clinical perspective, time of symptom onset may be misinterpreted and electrocardiographic (ECG) changes can be nonspecific in many ED patients. From the laboratory perspective, gaps affecting hs-cTn interpretation on what constitutes an undetectable or low hs-cTn concentration exist.2Neumann J.T. Twerenbold R. Ojeda F. et al.Application of high-sensitivity troponin in suspected myocardial infarction.N Engl J Med. 2019; 380: 2529-2540Crossref PubMed Scopus (194) Google Scholar For example, laboratories may select different lower limits for reporting, either using the instructions for use/package insert from the respective test manufacturer or deriving their own lower limit. Other times, a regulatory body (eg, the Food and Drug Administration in the United States) mandates that the lower limit of reporting be set to a specific metric for hs-cTn assays (ie, the limit of quantification). Collectively, there is no universal laboratory standardisation regarding what lower limit should be used, and as such is a source of variation and misclassification. Our objective was to assess laboratory testing alone in a large ED population with the use of low hs-cTn cutoffs in isolation and in combination with other common laboratory tests for future risk stratification. Patients who presented to the ED with acute chest pain and evaluated for ACS from April 2013 to December 2020 were identified at hospitals in the province of Ontario via databases (ie, the discharge abstract database for hospitalisations, the National Ambulatory Care Reporting System for ED visits, the Registered Persons Database for demographics and death, the Ontario Health Insurance Plan, and the Ontario Laboratory Information System for laboratory data). During this time period, the same hs-cTnT assay generation was used (Roche Diagnostics) with other hs-cTnI assays gradually being approved by Health Canada (Abbott Diagnostics, Beckman Coulter, and Siemens Healthcare Diagnostics) and being used by hospital laboratories within Ontario. These data sets were linked by means of unique encoded identifiers and analysed at ICES, formerly known as the Institute for Clinical Evaluative Sciences.3Lau G. Koh M. Kavsak P.A. et al.Clinical outcomes for chest pain patients discharged home from emergency departments using high-sensitivity vs conventional cardiac troponin assays.Am Heart J. 2020; 221: 84-94Crossref PubMed Scopus (8) Google Scholar Patients were included if they presented to the ED with chest pain during the study timeline (International Classification of Disease 10th Revision: R071-R074, I200, I201, I208, and I209) (n = 1,708,034). Patients were excluded if they were admitted to hospital or presented to a paediatric or mental health hospital (n = 82,883), were aged < 40 or > 105 years (n = 480,674), were not enrolled in the Ontario Health Insurance Plan (n = 2,340), had no ECG performed (n = 130,524), and did not have hs-cTn reported and other laboratory tests (eg, creatinine, glucose) (n = 880,518) that are useful for risk stratification. The final population included 131,095 patients (Supplemental Fig. S1).3Lau G. Koh M. Kavsak P.A. et al.Clinical outcomes for chest pain patients discharged home from emergency departments using high-sensitivity vs conventional cardiac troponin assays.Am Heart J. 2020; 221: 84-94Crossref PubMed Scopus (8) Google Scholar The first ED visit with the necessary laboratory results was used as the index to assess the primary outcome, which was defined as all-cause mortality or MI after ED discharge up to 365 days. Two single blood draw approaches were evaluated for classifying low-risk patients. First, for a low hs-cTn cutoff alone, low hs-cTnT was defined as < 6 ng/L (lowest reportable concentration in the United States) and low hs-cTnI defined as < 5 ng/L, a concentration that has been used for identifying low-risk patients with several different hs-cTnI assays (approach 1).4Kavsak P.A. Cerasuolo J.O. Ko D.T. et al.Using the clinical chemistry score in the emergency department to detect adverse cardiac events: a diagnostic accuracy study.CMAJ Open. 2020; 8: E676-E684Crossref PubMed Scopus (14) Google Scholar,5Westwood M.E. Armstrong N. Worthy G. et al.Optimising the use of high-sensitivity troponin assays for the early rule-out of myocardial infarction in patients presenting with chest pain: a systematic review.Clin Chem. 2021; 67: 237-244Crossref PubMed Scopus (14) Google Scholar A second method was used in which patients with normal kidney function (eGFR ≥ 90 mL/min/1.73 m2 with eGFR calculated via the Chronic Kidney Disease–Epidemiology Collaboration equation), normal glycemia (glucose < 5.6 mmol/L), and normal hs-cTn (hs-cTnT < 8 ng/L or hs-cTnI < 4 ng/L) were classified as low risk (ie, low risk defined by Clinical Chemistry Score < 1) (approach 2).4Kavsak P.A. Cerasuolo J.O. Ko D.T. et al.Using the clinical chemistry score in the emergency department to detect adverse cardiac events: a diagnostic accuracy study.CMAJ Open. 2020; 8: E676-E684Crossref PubMed Scopus (14) Google Scholar Descriptive (nonparametric) and diagnostic parameters (ie, sensitivity and negative predictive value [NPV] with 95% confidence intervals [CIs]) were calculated with the Clinical Chemistry Score and the hs-cTn respective cutoffs in the overall population. Forest plots on important subgroups (ie, sex, age) and the most common conditions/procedures (ie, history of hypertension, diabetes, and previous stress test) for the sensitivity of the primary outcome at 30 days with both approaches were also performed. Analyses were performed using SAS version 9.1.3 software (SAS Institute, Cary, NC) and MedCalc Statistical Software version 20.109 (MedCalc Software, Ostend, Belgium) for the primary outcome (death/MI). Ethics approval was granted by the Hamilton Integrated Research Ethics Board (HiREB: 4717-D). Of the 131,095 patients discharged home from the ED, 52.7% were evaluated with hs-cTnT and 47.3% with hs-cTnI. The 30-day incidence of death/MI was < 0.50% with NPV ≥ 99.5% for hs-cTn alone (approach 1) and for the Clinical Chemistry Score (approach 2) (Table 1). However, the sensitivities were higher for approach 2 (range 97.3%-98.4%) than for approach 1 (range 53.9%-92.7%) over the first year (Supplemental Table S1). Only approach 2 yielded an incidence rate < 1.00% at 180 days and at 365 days. Subgroup analyses at 30 days yielded sensitivity point estimates > 99% only for approach 2 (Fig. 1). Here, using approach 2 with hs-cTnT, the sensitivity for hs-cTnT was > 99% for women, those ≥ 65 years old, and patients with a history of hypertension or diabetes.Table 1Specific characteristics and 30-day death/MI for patients discharged home and evaluated with hs-cTnT (n = 69,133) and hs-cTnI (n = 61,962) in the province of Ontario from 2013 to 2020Approach 1 with hs-cTnT n = 30,365 (44%)Approach 2 with hs-cTnT n = 9107 (13%)Approach 1 with hs-cTnI n = 37,182 (60%)Approach 2 with hs-cTnI n = 6005 (10%)Age, y52 (46-60)49 (44-56)57 (49-67)49 (44-56)Female20,169 (66.4)5833 (64.0)19,923 (53.6)3306 (55.1)History Hypertension10,325 (34.0)2377 (26.1)17,825 (47.9)1663 (27.7) Stress test10,277 (33.8)2717 (29.8)14,648 (39.4)1808 (30.1) Diabetes4173 (13.7)619 (6.8)8093 (21.8)401 (6.7)30-day death/MI37 (0.12)8 (0.09)156 (0.42)9 (0.15)Sensitivity, %92.7 (90.1-94.8)98.4 (96.9-99.3)57.8 (52.6-62.9)97.6 (95.4-98.9)Negative predictive value, %99.9 (99.8-99.9)99.9 (99.8-99.9)99.6 (99.5-99.6)99.9 (99.7-99.9)Values are presented as median (interquartile range), n (%), or mean (95% confidence interval).hs-cTn, high-sensitivity cardiac troponin; MI, myocardial infarction. Open table in a new tab Values are presented as median (interquartile range), n (%), or mean (95% confidence interval). hs-cTn, high-sensitivity cardiac troponin; MI, myocardial infarction. Over 8 years at the population level in Ontario, patients discharged home from the ED had a 30-day death/MI incidence of < 0.50%, using either a low hs-cTn cutoff alone or with inclusion of normal kidney function and normal glycemia. The overall sensitivities with either approach for death/MI was < 99%, lower than the benchmark used in some guidelines.1Gulati M. Levy P.D. Mukherjee D. et al.2021 AHA/ACC/ASE/CHEST/SAEM/SCCT/SCMR guideline for the evaluation and diagnosis of chest pain: executive summary: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines.Circulation. 2021; 144: e368-e454Google Scholar However, in patients ≥ 65 years old, the sensitivity with approach 2 (the Clinical Chemistry Score) was > 99% with the use of either hs-cTnT or hs-cTnI. Also, approach 2 yielded higher and similar diagnostic estimates with the use of all hs-cTn assays (ie, sensitivity > 97%) over 365 days and exceeded another proposed sensitivity benchmark for rule-out of MI in chest pain patients from a large systematic review undertaken for the National Institute for Health and Care Excellence guidance in the United Kingdom.5Westwood M.E. Armstrong N. Worthy G. et al.Optimising the use of high-sensitivity troponin assays for the early rule-out of myocardial infarction in patients presenting with chest pain: a systematic review.Clin Chem. 2021; 67: 237-244Crossref PubMed Scopus (14) Google Scholar There are limitations to this large population-based study. First, laboratory data obtained from the Ontario Laboratory Information System does not include the actual manufacturer name of the test, only that the test is a high-sensitivity assay (authors PK and DK oversaw this process). Accordingly, no subanalyses were performed using specific assay cutoffs. However, some guidelines and key opinion leaders have advocated for a “low” troponin cutoff. For example, the 5 ng/L cutoff for the Abbott and Siemens hs-cTnI assays have been suggested as a universal cutoff for rule-out. Applying one such common low cutoff, as was done in this paper, yields sensitivity aspects that are inferior for hs-cTnI vs hs-cTnT (for which there is only 1 manufacturer). We think that our approach re-emphasises the importance of 1) using assay specific cutoffs or 2) using the Clinical Chemistry Score, which has been shown to yield higher diagnostic estimates regardless of the assay. Second, sex-specific cutoffs for hs-cTn were not used for rule-out. Sex-specific cutoffs have been advocated to detect myocardial injury (eg, the on-going CODE-MI [hs-cTn—Optimizing the Diagnosis of Acute Myocardial Infarction/Injury in Women; NCT03819894] trial in Canada) and not to rule out MI. There are minor differences in hs-cTn concentrations between the sexes within the low normal range, which are actually within the imprecision of the assay at the low end. Because laboratories do not regularly monitor the low end, and laboratory recommendations suggesting only 1 quality-control level in the normal range, analytically only 1 cutoff was used. Third, the largest group (n = 880,518) excluded from our analyses represented those who did not have all 3 laboratory tests reported. It is unclear if there are differences in risk and test performance between those with only hs-cTn tested vs those with 2 or 3 laboratory tests reported. Accordingly, the performance of approach 1 in those patients that had only hs-cTn tested is unknown, and caution is warranted in extrapolating findings from this study assessing all 3 laboratory tests to that group. Fourth, the percentage of patients deemed to be of low risk is smaller using approach 2 vs approach 1. This difference may result in more patients using approach 2 being subjected to a second blood drawing and a second hs-cTn measurement, which is useful to document acute myocardial injury. Overall, the Clinical Chemistry Score (approach 2) yields higher sensitivity and represents a standardised single blood drawing approach for classifying very-low-risk ED patients whose incidence of MI/death is < 1% over 1 year.