Determination of Clinically Acceptable Analytical Variation of Cardiac Troponin at Decision Thresholds.

BACKGROUND: Clinical decision-making for risk stratification for possible myocardial infarction (MI) uses high-sensitivity cardiac troponin (hs-cTn) thresholds that range from the limit of detection to several-fold higher than the upper reference limit (URL). To establish a minimum analytical variation standard, we can quantify the effect of variation on the population clinical measures of safety (sensitivity) and effectiveness [proportion below threshold, or positive predictive value (PPV)]. METHODS: From large datasets of patients investigated for possible MI with the Abbott hs-cTnI and Roche hs-cTnT assays, we synthesized datasets of 1 000 000 simulated patients. Troponin concentrations were randomly varied several times based on absolute deviations of 0.5 to 3 ng/L and relative changes of 2% to 20% around the low-risk threshold (5 ng/L) and URLs, respectively. RESULTS: For both assays at the low-risk thresholds, there were negligible differences in sensitivity (<0.3%) with increasing analytical variation. The proportion of patients characterized as low risk reduced by 30% to 29% (Roche) and 53% to 44% (Abbott). At the URL, increasing analytical variation also did not change sensitivity; the PPV fell by less than 3%. For risk stratification, increased delta thresholds (change between serial troponin concentrations) increased sensitivity at the cost of a decreased percentage of patients below the delta threshold, with the largest changes at the greatest analytical variation. CONCLUSIONS: At the low-risk threshold, analytical variation up to 3 ng/L minimally impacted the safety metric (sensitivity) but marginally reduced effectiveness. Similarly, at the URL even relative variation up to 25% minimally impacted safety metrics and effectiveness. Analytical variation for delta thresholds did not negatively impact sensitivity but decreased effectiveness.

Performance of the European Society of Cardiology 0/1-Hour, 0/2-Hour, and 0/3-Hour Algorithms for Rapid Triage of Acute Myocardial Infarction : An International Collaborative Meta-analysis.

BACKGROUND: The 2020 European Society of Cardiology (ESC) guidelines recommend using the 0/1-hour and 0/2-hour algorithms over the 0/3-hour algorithm as the first and second choices of high-sensitivity cardiac troponin (hs-cTn)-based strategies for triage of patients with suspected acute myocardial infarction (AMI). PURPOSE: To evaluate the diagnostic accuracies of the ESC 0/1-hour, 0/2-hour, and 0/3-hour algorithms. DATA SOURCES: PubMed, Embase, Cochrane Central Register of Controlled Trials, Web of Science, and Scopus from 1 January 2011 to 31 December 2020. (PROSPERO: CRD42020216479). STUDY SELECTION: Prospective studies that evaluated the ESC 0/1-hour, 0/2-hour, or 0/3-hour algorithms in adult patients presenting with suspected AMI. DATA EXTRACTION: The primary outcome was index AMI. Twenty unique cohorts were identified. Primary data were obtained from investigators of 16 cohorts and aggregate data were extracted from 4 cohorts. Two independent authors assessed each study for methodological quality. DATA SYNTHESIS: A total of 32 studies (20 cohorts) with 30 066 patients were analyzed. The 0/1-hour algorithm had a pooled sensitivity of 99.1% (95% CI, 98.5% to 99.5%) and negative predictive value (NPV) of 99.8% (CI, 99.6% to 99.9%) for ruling out AMI. The 0/2-hour algorithm had a pooled sensitivity of 98.6% (CI, 97.2% to 99.3%) and NPV of 99.6% (CI, 99.4% to 99.8%). The 0/3-hour algorithm had a pooled sensitivity of 93.7% (CI, 87.4% to 97.0%) and NPV of 98.7% (CI, 97.7% to 99.3%). Sensitivity of the 0/3-hour algorithm was attenuated in studies that did not use clinical criteria (GRACE score <140 and pain-free) compared with studies that used clinical criteria (90.2% [CI, 82.9 to 94.6] vs. 98.4% [CI, 88.6 to 99.8]). All 3 algorithms had similar specificities and positive predictive values for ruling in AMI, but heterogeneity across studies was substantial. Diagnostic performance was similar across the hs-cTnT (Elecsys; Roche), hs-cTnI (Architect; Abbott), and hs-cTnI (Centaur/Atellica; Siemens) assays. LIMITATION: Diagnostic accuracy, inclusion and exclusion criteria, and cardiac troponin sampling time varied among studies. CONCLUSION: The ESC 0/1-hour and 0/2-hour algorithms have higher sensitivities and NPVs than the 0/3-hour algorithm for index AMI. PRIMARY FUNDING SOURCE: National Taiwan University Hospital.

Machine Learning to Predict the Likelihood of Acute Myocardial Infarction.

BACKGROUND: Variations in cardiac troponin concentrations by age, sex, and time between samples in patients with suspected myocardial infarction are not currently accounted for in diagnostic approaches. We aimed to combine these variables through machine learning to improve the assessment of risk for individual patients. METHODS: A machine learning algorithm (myocardial-ischemic-injury-index [MI3]) incorporating age, sex, and paired high-sensitivity cardiac troponin I concentrations, was trained on 3013 patients and tested on 7998 patients with suspected myocardial infarction. MI3 uses gradient boosting to compute a value (0-100) reflecting an individual's likelihood of a diagnosis of type 1 myocardial infarction and estimates the sensitivity, negative predictive value, specificity and positive predictive value for that individual. Assessment was by calibration and area under the receiver operating characteristic curve. Secondary analysis evaluated example MI3 thresholds from the training set that identified patients as low risk (99% sensitivity) and high risk (75% positive predictive value), and performance at these thresholds was compared in the test set to the 99th percentile and European Society of Cardiology rule-out pathways. RESULTS: Myocardial infarction occurred in 404 (13.4%) patients in the training set and 849 (10.6%) patients in the test set. MI3 was well calibrated with a very high area under the receiver operating characteristic curve of 0.963 [0.956-0.971] in the test set and similar performance in early and late presenters. Example MI3 thresholds identifying low- and high-risk patients in the training set were 1.6 and 49.7, respectively. In the test set, MI3 values were <1.6 in 69.5% with a negative predictive value of 99.7% (99.5-99.8%) and sensitivity of 97.8% (96.7-98.7%), and were ≥49.7 in 10.6% with a positive predictive value of 71.8% (68.9-75.0%) and specificity of 96.7% (96.3-97.1%). Using these thresholds, MI3 performed better than the European Society of Cardiology 0/3-hour pathway (sensitivity, 82.5% [74.5-88.8%]; specificity, 92.2% [90.7-93.5%]) and the 99th percentile at any time point (sensitivity, 89.6% [87.4-91.6%]); specificity, 89.3% [88.6-90.0%]). CONCLUSIONS: Using machine learning, MI3 provides an individualized and objective assessment of the likelihood of myocardial infarction, which can be used to identify low- and high-risk patients who may benefit from earlier clinical decisions. CLINICAL TRIAL REGISTRATION: URL: https://www.anzctr.org.au. Unique identifier: ACTRN12616001441404.

ICare-ACS (Improving Care Processes for Patients With Suspected Acute Coronary Syndrome): A Study of Cross-System Implementation of a National Clinical Pathway.

BACKGROUND: Efforts to safely reduce length of stay for emergency department patients with symptoms suggestive of acute coronary syndrome (ACS) have had mixed success. Few system-wide efforts affecting multiple hospital emergency departments have ever been evaluated. We evaluated the effectiveness of a nationwide implementation of clinical pathways for potential ACS in disparate hospitals. METHODS: This was a multicenter pragmatic stepped-wedge before-and-after trial in 7 New Zealand acute care hospitals with 31 332 patients investigated for suspected ACS with serial troponin measurements. The implementation was a clinical pathway for the assessment of patients with suspected ACS that included a clinical pathway document in paper or electronic format, structured risk stratification, specified time points for electrocardiographic and serial troponin testing within 3 hours of arrival, and directions for combining risk stratification and electrocardiographic and troponin testing in an accelerated diagnostic protocol. Implementation was monitored for >4 months and compared with usual care over the preceding 6 months. The main outcome measure was the odds of discharge within 6 hours of presentation RESULTS: There were 11 529 participants in the preimplementation phase (range, 284-3465) and 19 803 in the postimplementation phase (range, 395-5039). Overall, the mean 6-hour discharge rate increased from 8.3% (range, 2.7%-37.7%) to 18.4% (6.8%-43.8%). The odds of being discharged within 6 hours increased after clinical pathway implementation. The odds ratio was 2.4 (95% confidence interval, 2.3-2.6). In patients without ACS, the median length of hospital stays decreased by 2.9 hours (95% confidence interval, 2.4-3.4). For patients discharged within 6 hours, there was no change in 30-day major adverse cardiac event rates (0.52% versus 0.44%; P=0.96). In these patients, no adverse event occurred when clinical pathways were correctly followed. CONCLUSIONS: Implementation of clinical pathways for suspected ACS reduced the length of stay and increased the proportions of patients safely discharged within 6 hours. CLINICAL TRIAL REGISTRATION: URL: https://www.anzctr.org.au/ (Australian and New Zealand Clinical Trials Registry). Unique identifier: ACTRN12617000381381.

Detectable High-Sensitivity Cardiac Troponin within the Population Reference Interval Conveys High 5-Year Cardiovascular Risk: An Observational Study.

BACKGROUND: Increased cardiac troponin I or T detected by high-sensitivity assays (hs-cTnI or hs-cTnT) confers an increased risk of adverse prognosis. We determined whether patients presenting with putatively normal, detectable cTn concentrations [> limit of detection and < upper reference limit (URL)] have increased risk of major adverse cardiovascular events (MACE) or all-cause mortality. METHODS: A prospective 5-year follow-up of patients recruited in the emergency department with possible acute coronary syndrome (ACS) and cTn concentrations measured with hs-cTnI (Abbott) and hs-cTnT (Roche) assays. Cox regression models were generated with adjustment for covariates in those without MACE on presentation. Hazard ratios (HRs) for hs-cTn were calculated relative to the HRs at the median concentration. RESULTS: Of 1113 patients, 836 were without presentation MACE. Of these, 138 incurred a MACE and 169 died during a median 5.8-year follow-up. HRs for MACE at the URLs were 2.3 (95% CI, 1.7-3.2) for hs-cTnI and 1.8 (95% CI, 1.3-2.4) for hs-cTnT. Corresponding HRs for mortality were 1.7 (95% CI, 1.2-2.2) for hs-cTnI and 2.3 (95 % CI, 1.7-3.1) for hs-cTnT. The HR for MACE increased with increasing hs-cTn concentration similarly for both assays, but the HR for mortality increased at approximately twice the rate for hs-cTnT than hs-cTnI. Patients with hs-cTnI ≥10 ng/L or hs-cTnT ≥16 ng/L had the same percentage of MACE at 5-year follow-up (33%) as patients with presentation MACE. CONCLUSIONS: Many patients with ACS ruled out and putatively normal but detectable hs-cTnI concentrations are at similar long-term risk as those with MACE. hs-cTnT concentrations are more strongly associated with 5-year mortality than hs-cTnI.

Acute Kidney Injury and mortality prognosis in Acute Coronary Syndrome patients: A meta-analysis.

AIM: The aim of this study is to provide a robust estimate of mortality risk in acute coronary syndrome (ACS)-associated acute kidney injury (AKI) to inform clinical practice and policy. METHODS: A meta-analysis of cohort studies evaluating outcomes of ACS and which reported AKI and AKI-associated mortality. Studies were excluded if they incorporated patients not admitted through the emergency department (i.e. for elective procedures), were limited to cardiogenic shock or cardiac arrest, or relied on registry data for outcomes without further adjudication. The predictor was ACS-associated AKI and outcomes early (30 day or in-hospital) mortality and late-mortality (post-hospital discharge). RESULTS: Thirty-six studies with 37 unique cohorts comprising 100 476 patients were included. The pooled rate of ACS-associated AKI was 15.8%. In 32 cohorts reporting early mortality, the crude early mortality rate was 15.0% amongst those with AKI compared with 2.0% amongst those without AKI. The pooled estimate of the relative risk of AKI-associated early mortality was 4.1 (95% confidence interval: 3.3 to 5.0) with high heterogeneity between studies (I 2  = 84% (61% to 88%)). When heterogeneity was accounted for mathematically using credibility ceilings, the risk of mortality was lower, but still clinically significant (3.1 (2.6 to 3.6)). In 19 cohorts reporting late mortality (1 to 10 years), the relative risk of AKI-associated mortality was 2.6 (2.0 to 3.3) with moderate heterogeneity (I 2  = 65 % [35% to 88%]). Following application of credibility ceiling relative risk estimate dropped to 2.2 (1.9 to 2.6). CONCLUSIONS: Acute coronary syndrome-associated AKI is associated with more than a three-fold increase in early mortality and more than two-fold in long-term mortality.

Rapid Rule-out of Acute Myocardial Infarction With a Single High-Sensitivity Cardiac Troponin T Measurement Below the Limit of Detection: A Collaborative Meta-analysis.

Background: High-sensitivity assays for cardiac troponin T (hs-cTnT) are sometimes used to rapidly rule out acute myocardial infarction (AMI). Purpose: To estimate the ability of a single hs-cTnT concentration below the limit of detection (<0.005 µg/L) and a nonischemic electrocardiogram (ECG) to rule out AMI in adults presenting to the emergency department (ED) with chest pain. Data Sources: EMBASE and MEDLINE without language restrictions (1 January 2008 to 14 December 2016). Study Selection: Cohort studies involving adults presenting to the ED with possible acute coronary syndrome in whom an ECG and hs-cTnT measurements were obtained and AMI outcomes adjudicated during initial hospitalization. Data Extraction: Investigators of studies provided data on the number of low-risk patients (no new ischemia on ECG and hs-cTnT measurements <0.005 µg/L) and the number who had AMI during hospitalization (primary outcome) or a major adverse cardiac event (MACE) or death within 30 days (secondary outcomes), by risk classification (low or not low risk). Two independent epidemiologists rated risk of bias of studies. Data Synthesis: Of 9241 patients in 11 cohort studies, 2825 (30.6%) were classified as low risk. Fourteen (0.5%) low-risk patients had AMI. Sensitivity of the risk classification for AMI ranged from 87.5% to 100% in individual studies. Pooled estimated sensitivity was 98.7% (95% CI, 96.6% to 99.5%). Sensitivity for 30-day MACEs ranged from 87.9% to 100%; pooled sensitivity was 98.0% (CI, 94.7% to 99.3%). No low-risk patients died. Limitation: Few studies, variation in timing and methods of reference standard troponin tests, and heterogeneity of risk and prevalence of AMI across studies. Conclusion: A single hs-cTnT concentration below the limit of detection in combination with a nonischemic ECG may successfully rule out AMI in patients presenting to EDs with possible emergency acute coronary syndrome. Primary Funding Source: Emergency Care Foundation.

Assessment of the European Society of Cardiology 0-Hour/1-Hour Algorithm to Rule-Out and Rule-In Acute Myocardial Infarction.

BACKGROUND: The new European Society of Cardiology guidelines to rule-in and rule-out acute myocardial infarction (AMI) in the emergency department include a rapid assessment algorithm based on high-sensitivity cardiac troponin and sampling at 0 and 1 hour. Emergency department physicians require high sensitivity to confidently rule-out AMI, whereas cardiologists aim to minimize false-positive results. METHODS: High-sensitivity troponin I and T assays were used to measure troponin concentrations in patients presenting with chest-pain symptoms and being investigated for possible acute coronary syndrome at hospitals in New Zealand, Australia, and Canada. AMI outcomes were independently adjudicated by at least 2 physicians. The European Society of Cardiology algorithm performance with each assay was assessed by the sensitivity and proportion with AMI ruled out and the positive predictive value and proportion ruled-in. RESULTS: There were 2222 patients with serial high-sensitivity troponin T and high-sensitivity troponin I measurements. The high-sensitivity troponin T algorithm ruled out 1425 (64.1%) with a sensitivity of 97.1% (95% confidence interval [CI], 94.0%-98.8%) and ruled-in 292 (13.1%) with a positive predictive value of 63.4% (95% CI, 57.5%-68.9%).The high-sensitivity troponin I algorithm ruled out 1205 (54.2%) with a sensitivity of 98.8% (95% CI, 96.4%-99.7%)) and ruled-in 310 (14.0%) with a positive predictive value of 68.1% (95% CI, 62.6%-73.2%). CONCLUSIONS: The sensitivity of the European Society of Cardiology rapid assessment 0-/1-hour algorithm to rule-out AMI with high-sensitivity troponin may be insufficient for some emergency department physicians to confidently send patients home. These algorithms may prove useful to identify patients requiring expedited management. However, the positive predictive value was modest for both algorithms.

Renal Function and Scaled Troponin in Patients Presenting to the Emergency Department with Symptoms of Myocardial Infarction.

BACKGROUND: Cardiac troponins are often found to be elevated in patients with renal dysfunction, even in the absence of acute myocardial injury. The objective of this report was to characterize the scaled troponin values and proportion of adjudicated acute myocardial infarction (AMI) among patients with and without renal dysfunction. METHODS: The data was from a multicenter prospective study including patients presenting to the emergency department with symptoms of AMI. Troponin measurements were standardized across various assays by calculating the observed results as multiples of the assay-specific 99th percentile upper limit of normal. Patients with an estimated glomerular filtration rate (eGFR; calculated by the Chronic Kidney Disease Epidemiology Collaboration formula) <60 mL/min/1.73 m2 were considered to have renal dysfunction. RESULTS: Of 430 included patients, 249 (58%) were male and 181 (42%) were female, with a mean age of 55.9 ± 12.3 and 57.3 ± 12.8 years, respectively. Eighty-seven (20.2%) had renal dysfunction. The proportions of patients with at least one scaled troponin value above the 99th percentile cut-off point among patients with and without renal dysfunction were 40 (45.9%) and 81 (23.6%) respectively (p < 0.001). The proportions of patients with an adjudicated diagnosis of AMI among those with and without renal dysfunction were 20.7 and 18.7%, respectively (p = 0.67). Using scaled troponins, by the second test there was >5X and by the third test >15X separation in the excursion of troponin among those with AMI compared to those without. CONCLUSIONS: One or more elevated troponin values are common in those with renal dysfunction. Scaled troponins for eGFR groups were similar, indicating that the use of this interpretative technique is applicable in discerning AMI for those with and without renal dysfunction.

Association of High-Sensitivity Cardiac Troponin I Concentration With Cardiac Outcomes in Patients With Suspected Acute Coronary Syndrome.

Importance: High-sensitivity cardiac troponin I testing is widely used to evaluate patients with suspected acute coronary syndrome. A cardiac troponin concentration of less than 5 ng/L identifies patients at presentation as low risk, but the optimal threshold is uncertain. Objective: To evaluate the performance of a cardiac troponin I threshold of 5 ng/L at presentation as a risk stratification tool in patients with suspected acute coronary syndrome. Data Sources: Systematic search of MEDLINE, EMBASE, Cochrane, and Web of Science databases from January 1, 2006, to March 18, 2017. Study Selection: Prospective studies measuring high-sensitivity cardiac troponin I concentrations in patients with suspected acute coronary syndrome in which the diagnosis was adjudicated according to the universal definition of myocardial infarction. Data Extraction and Synthesis: The systematic review identified 19 cohorts. Individual patient-level data were obtained from the corresponding authors of 17 cohorts, with aggregate data from 2 cohorts. Meta-estimates for primary and secondary outcomes were derived using a binomial-normal random-effects model. Main Outcomes and Measures: The primary outcome was myocardial infarction or cardiac death at 30 days. Performance was evaluated in subgroups and across a range of troponin concentrations (2-16 ng/L) using individual patient data. Results: Of 11 845 articles identified, 104 underwent full-text review, and 19 cohorts from 9 countries were included. Among 22 457 patients included in the meta-analysis (mean age, 62 [SD, 15.5] years; n = 9329 women [41.5%]), the primary outcome occurred in 2786 (12.4%). Cardiac troponin I concentrations were less than 5 ng/L at presentation in 11 012 patients (49%), in whom there were 60 missed index or 30-day events (59 index myocardial infarctions, 1 myocardial infarction at 30 days, and no cardiac deaths at 30 days). This resulted in a negative predictive value of 99.5% (95% CI, 99.3%-99.6%) for the primary outcome. There were no cardiac deaths at 30 days and 7 (0.1%) at 1 year, with a negative predictive value of 99.9% (95% CI, 99.7%-99.9%) for cardiac death. Conclusions and Relevance: Among patients with suspected acute coronary syndrome, a high-sensitivity cardiac troponin I concentration of less than 5 ng/L identified those at low risk of myocardial infarction or cardiac death within 30 days. Further research is needed to understand the clinical utility and cost-effectiveness of this approach to risk stratification.

Effectiveness of EDACS Versus ADAPT Accelerated Diagnostic Pathways for Chest Pain: A Pragmatic Randomized Controlled Trial Embedded Within Practice.

STUDY OBJECTIVE: A 2-hour accelerated diagnostic pathway based on the Thrombolysis in Myocardial Infarction score, ECG, and troponin measures (ADAPT-ADP) increased early discharge of patients with suspected acute myocardial infarction presenting to the emergency department compared with standard care (from 11% to 19.3%). Observational studies suggest that an accelerated diagnostic pathway using the Emergency Department Assessment of Chest Pain Score (EDACS-ADP) may further increase this proportion. This trial tests for the existence and size of any beneficial effect of using the EDACS-ADP in routine clinical care. METHODS: This was a pragmatic randomized controlled trial of adults with suspected acute myocardial infarction, comparing the ADAPT-ADP and the EDACS-ADP. The primary outcome was the proportion of patients discharged to outpatient care within 6 hours of attendance, without subsequent major adverse cardiac event within 30 days. RESULTS: Five hundred fifty-eight patients were recruited, 279 in each arm. Sixty-six patients (11.8%) had a major adverse cardiac event within 30 days (ADAPT-ADP 29; EDACS-ADP 37); 11.1% more patients (95% confidence interval 2.8% to 19.4%) were identified as low risk in EDACS-ADP (41.6%) than in ADAPT-ADP (30.5%). No low-risk patients had a major adverse cardiac event within 30 days (0.0% [0.0% to 1.9%]). There was no difference in the primary outcome of proportion discharged within 6 hours (EDACS-ADP 32.3%; ADAPT-ADP 34.4%; difference -2.1% [-10.3% to 6.0%], P=.65). CONCLUSION: There was no difference in the proportion of patients discharged early despite more patients being classified as low risk by the EDACS-ADP than the ADAPT-ADP. Both accelerated diagnostic pathways are effective strategies for chest pain assessment and resulted in an increased rate of early discharges compared with previously reported rates.

Heart Fatty Acid Binding Protein and cardiac troponin: development of an optimal rule-out strategy for acute myocardial infarction.

BACKGROUND: Improved ability to rapidly rule-out Acute Myocardial Infarction (AMI) in patients presenting with chest pain will promote decongestion of the Emergency Department (ED) and reduce unnecessary hospital admissions. We assessed a new commercial Heart Fatty Acid Binding Protein (H-FABP) assay for additional diagnostic value when combined with cardiac troponin (using a high sensitivity assay). METHODS: H-FABP and high-sensitivity troponins I (hs-cTnI) and T (hs-cTnT) were measured in samples taken on-presentation from patients, attending the ED, with symptoms triggering investigation for possible acute coronary syndrome. The optimal combination of H-FABP with each hs-cTn was defined as that which maximized the proportion of patients with a negative test (low-risk) whilst maintaining at least 99 % sensitivity for AMI. A negative test comprised both H-FABP and hs-cTn below the chosen threshold in the absence of ischemic changes on the ECG. RESULTS: One thousand seventy-nine patients were recruited including 248 with AMI. H-FABP < 4.3 ng/mL plus hs-cTnI < 10.0 ng/L together with a negative ECG maintained >99 % sensitivity for AMI whilst classifying 40.9 % of patients as low-risk. The combination of H-FABP < 3.9 ng/mL and hs-cTnT < 7.6 ng/L with a negative ECG maintained the same sensitivity whilst classifying 32.1 % of patients as low risk. CONCLUSIONS: In patients requiring rule-out of AMI, the addition of H-FABP to hs-cTn at presentation (in the absence of new ischaemic ECG findings) may accelerate clinical diagnostic decision making by identifying up to 40 % of such patients as low-risk for AMI on the basis of blood tests performed on presentation. If implemented this has the potential to significantly accelerate triaging of patients for early discharge from the ED.

Cost and outcomes of assessing patients with chest pain in an Australian emergency department.

OBJECTIVES: We sought to characterise the demographics, length of admission, final diagnoses, long-term outcome and costs associated with the population who presented to an Australian emergency department (ED) with symptoms of possible acute coronary syndrome (ACS). DESIGN, SETTING AND PARTICIPANTS: Prospectively collected data on ED patients presenting with suspected ACS between November 2008 and February 2011 was used, including data on presentation and at 30 days after presentation. Information on patient disposition, length of stay and costs incurred was extracted from hospital administration records. MAIN OUTCOME MEASURES: Primary outcomes were mean and median cost and length of hospital stay. Secondary outcomes were diagnosis of ACS, other cardiovascular conditions or non-cardiovascular conditions within 30 days of presentation. RESULTS: An ACS was diagnosed in 103 (11.1%) of the 926 patients recruited. 193 patients (20.8%) were diagnosed with other cardiovascular-related conditions and 622 patients (67.2%) had non-cardiac-related chest pain. ACS events occurred in 0 and 11 (1.9%) of the low-risk and intermediate-risk groups, respectively. Ninety-two (28.0%) of the 329 high-risk patients had an ACS event. Patients with a proven ACS, high-grade atrioventricular block, pulmonary embolism and other respiratory conditions had the longest length of stay. The mean cost was highest in the ACS group ($13 509; 95% CI, $11 794-$15 223) followed by other cardiovascular conditions ($7283; 95% CI, $6152-$8415) and non-cardiovascular conditions ($3331; 95% CI, $2976-$3685). CONCLUSIONS: Most ED patients with symptoms of possible ACS do not have a cardiac cause for their presentation. The current guideline-based process of assessment is lengthy, costly and consumes significant resources. Investigation of strategies to shorten this process or reduce the need for objective cardiac testing in patients at intermediate risk according to the National Heart Foundation and Cardiac Society of Australia and New Zealand guideline is required.

The clinical utility window for acute kidney injury biomarkers in the critically ill.

INTRODUCTION: Acute Kidney Injury (AKI) biomarker utility depends on sample timing after the onset of renal injury. We compared biomarker performance on arrival in the emergency department (ED) with subsequent performance in the intensive care unit (ICU). METHODS: Urinary and plasma Neutrophil Gelatinase-Associated Lipocalin (NGAL), and urinary Cystatin C (CysC), alkaline phosphatase, γ-Glutamyl Transpeptidase (GGT), α- and π-Glutathione S-Transferase (GST), and albumin were measured on ED presentation, and at 0, 4, 8, and 16 hours, and days 2, 4 and 7 in the ICU in patients after cardiac arrest, sustained or profound hypotension or ruptured abdominal aortic aneurysm. AKI was defined as plasma creatinine increase ≥ 26.5 µmol/l within 48 hours or ≥ 50% within 7 days. RESULTS: In total, 45 of 77 patients developed AKI. Most AKI patients had elevated urinary NGAL, and plasma NGAL and CysC in the period 6 to 24 hours post presentation. Biomarker performance in the ICU was similar or better than when measured earlier in the ED. Plasma NGAL diagnosed AKI at all sampling times, urinary NGAL, plasma and urinary CysC up to 48 hours, GGT 4 to 12 hours, and π-GST 8 to 12 hours post insult. Thirty-one patients died or required dialysis. Peak 24-hour urinary NGAL and albumin independently predicted 30-day mortality and dialysis; odds ratios 2.87 (1.32 to 6.26), and 2.72 (1.14 to 6.48), respectively. Urinary NGAL improved risk prediction by 11% (IDI event of 0.06 (0.002 to 0.19) and IDI non-event of 0.04 (0.002 to 0.12)). CONCLUSION: Early measurement in the ED has utility, but not better AKI diagnostic performance than later ICU measurement. Plasma NGAL diagnosed AKI at all time points. Urinary NGAL best predicted mortality or dialysis compared to other biomarkers. TRIAL REGISTRATION: Australian and New Zealand Clinical Trials Registry ACTRN12610001012066. Registered 12 February 2010.

Derivation and Validation of Thresholds Using Synthetic Data Methods for Single-Test Screening of Emergency Department Patients with Possible Acute Myocardial Infarction Using a Point-of-Care Troponin Assay.

BACKGROUND: Single-sample (screening) rule-out of acute myocardial infarction (AMI) with troponin requires derivation of a single-test screening threshold. In data sets with small event numbers, the lowest one or two concentrations of myocardial infarction (MI) patients dictate the threshold. This is not optimal. We aimed to demonstrate a process incorporating both real and synthetic data for deriving such thresholds using a novel pre-production high-precision point-of-care assay. METHODS: cTnI concentrations were measured from thawed plasma using the Troponin I Next (TnI-Nx) assay (i-STAT; Abbott) in adults on arrival to the emergency department with symptoms suggestive of AMI. The primary outcome was an AMI or cardiac death within 30 days. We used internal-external validation with synthetic data production based on clinical and demographic data, plus the measured TnI-Nx concentration, to derive and validate decision thresholds for TnI-Nx. The target low-risk threshold was a sensitivity of 99% and a high-risk threshold specificity of >95%. RESULTS: In total, 1356 patients were included, of whom 191 (14.1%) had the primary outcome. A total of 500 synthetic data sets were constructed. The mean low-risk threshold was determined to be 5 ng/L. This categorized 38% (95% CI, 6%-68%) to low-risk with a sensitivity of 99.0% (95% CI, 98.6%-99.5%) and a negative predictive value of 99.4% (95% CI, 97.6%-99.8%). A similarly derived high-risk threshold of 25 ng/L had a specificity of 95.0% (95% CI, 94.8%-95.1%) and a positive predictive value of 74.8% (95% CI, 71.5%-78.0%). CONCLUSIONS: With the TnI-Nx assay, we successfully demonstrated an approach using synthetic data generation to derive low-risk thresholds for safe and effective screening.

Machine learning for diagnosis of myocardial infarction using cardiac troponin concentrations.

Although guidelines recommend fixed cardiac troponin thresholds for the diagnosis of myocardial infarction, troponin concentrations are influenced by age, sex, comorbidities and time from symptom onset. To improve diagnosis, we developed machine learning models that integrate cardiac troponin concentrations at presentation or on serial testing with clinical features and compute the Collaboration for the Diagnosis and Evaluation of Acute Coronary Syndrome (CoDE-ACS) score (0-100) that corresponds to an individual's probability of myocardial infarction. The models were trained on data from 10,038 patients (48% women), and their performance was externally validated using data from 10,286 patients (35% women) from seven cohorts. CoDE-ACS had excellent discrimination for myocardial infarction (area under curve, 0.953; 95% confidence interval, 0.947-0.958), performed well across subgroups and identified more patients at presentation as low probability of having myocardial infarction than fixed cardiac troponin thresholds (61 versus 27%) with a similar negative predictive value and fewer as high probability of having myocardial infarction (10 versus 16%) with a greater positive predictive value. Patients identified as having a low probability of myocardial infarction had a lower rate of cardiac death than those with intermediate or high probability 30 days (0.1 versus 0.5 and 1.8%) and 1 year (0.3 versus 2.8 and 4.2%; P < 0.001 for both) from patient presentation. CoDE-ACS used as a clinical decision support system has the potential to reduce hospital admissions and have major benefits for patients and health care providers.

Personalized diagnosis in suspected myocardial infarction.

BACKGROUND: In suspected myocardial infarction (MI), guidelines recommend using high-sensitivity cardiac troponin (hs-cTn)-based approaches. These require fixed assay-specific thresholds and timepoints, without directly integrating clinical information. Using machine-learning techniques including hs-cTn and clinical routine variables, we aimed to build a digital tool to directly estimate the individual probability of MI, allowing for numerous hs-cTn assays. METHODS: In 2,575 patients presenting to the emergency department with suspected MI, two ensembles of machine-learning models using single or serial concentrations of six different hs-cTn assays were derived to estimate the individual MI probability (ARTEMIS model). Discriminative performance of the models was assessed using area under the receiver operating characteristic curve (AUC) and logLoss. Model performance was validated in an external cohort with 1688 patients and tested for global generalizability in 13 international cohorts with 23,411 patients. RESULTS: Eleven routinely available variables including age, sex, cardiovascular risk factors, electrocardiography, and hs-cTn were included in the ARTEMIS models. In the validation and generalization cohorts, excellent discriminative performance was confirmed, superior to hs-cTn only. For the serial hs-cTn measurement model, AUC ranged from 0.92 to 0.98. Good calibration was observed. Using a single hs-cTn measurement, the ARTEMIS model allowed direct rule-out of MI with very high and similar safety but up to tripled efficiency compared to the guideline-recommended strategy. CONCLUSION: We developed and validated diagnostic models to accurately estimate the individual probability of MI, which allow for variable hs-cTn use and flexible timing of resampling. Their digital application may provide rapid, safe and efficient personalized patient care. TRIAL REGISTRATION NUMBERS: Data of following cohorts were used for this project: BACC ( www. CLINICALTRIALS: gov ; NCT02355457), stenoCardia ( www. CLINICALTRIALS: gov ; NCT03227159), ADAPT-BSN ( www.australianclinicaltrials.gov.au ; ACTRN12611001069943), IMPACT ( www.australianclinicaltrials.gov.au , ACTRN12611000206921), ADAPT-RCT ( www.anzctr.org.au ; ANZCTR12610000766011), EDACS-RCT ( www.anzctr.org.au ; ANZCTR12613000745741); DROP-ACS ( https://www.umin.ac.jp , UMIN000030668); High-STEACS ( www. CLINICALTRIALS: gov ; NCT01852123), LUND ( www. CLINICALTRIALS: gov ; NCT05484544), RAPID-CPU ( www. CLINICALTRIALS: gov ; NCT03111862), ROMI ( www. CLINICALTRIALS: gov ; NCT01994577), SAMIE ( https://anzctr.org.au ; ACTRN12621000053820), SEIGE and SAFETY ( www. CLINICALTRIALS: gov ; NCT04772157), STOP-CP ( www. CLINICALTRIALS: gov ; NCT02984436), UTROPIA ( www. CLINICALTRIALS: gov ; NCT02060760).

Machine learning with D-dimer in the risk stratification for pulmonary embolism: a derivation and internal validation study.

AIM: To develop a machine learning model to predict the diagnosis of pulmonary embolism (PE). METHODS AND RESULTS: We undertook a derivation and internal validation study to develop a risk prediction model for use in patients being investigated for possible PE. The machine learning technique, generalized logistic regression using elastic net, was chosen following an assessment of seven machine learning techniques and on the basis that it optimized the area under the receiver operator characteristic curve (AUC) and Brier score. Models were developed both with and without the addition of D-dimer. A total of 3347 patients were included in the study of whom, 219 (6.5%) had PE. Four clinical variables (O2 saturation, previous deep venous thrombosis or PE, immobilization or surgery, and alternative diagnosis equal or more likely than PE) plus D-dimer contributed to the machine learning models. The addition of D-dimer improved the AUC by 0.16 (95% confidence interval 0.13-0.19), from 0.73 to 0.89 (0.87-0.91) and decreased the Brier score by 14% (10-18%). More could be ruled out with a higher positive likelihood ratio than by the Wells score combined with D-dimer, revised Geneva score combined with D-dimer, or the Pulmonary Embolism Rule-out Criteria score. Machine learning with D-dimer maintained a low-false-negative rate at a true-negative rate of nearly 53%, which was better performance than any of the other alternatives. CONCLUSION: A machine learning model outperformed traditional risk scores for the risk stratification of PE in the emergency department. However, external validation is needed.