Providing High Value Care to Patients with Acute Chest PainUniversity of Texas Southwestern Medical CenterInternal Medicine Grand RoundsDr. Rebecca VigenJune 14, 2019This is to acknowledge that Rebecca Vigen has disclosed that she does not have any financialinterests or other relationships with commercial concerns related directly or indirectly to theprogram. Dr. Vigen will not be discussing off-label uses in her presentation.

Rebecca Vigen, MD, MSCSAssistant Professor of Internal MedicineDivision of CardiologyRebecca Vigen, MD, MSCS is an Assistant Professor of Internal Medicine in the Division ofCardiology. She is a general cardiologist and health services researcher focused on high valuecare for patients with cardiovascular disease. Dr. Vigen received her MD from Louisiana StateUniversity and completed her internal medicine training at UT Southwestern. She thencompleted a research fellowship in cardiovascular outcomes research at the University ofColorado where she earned her MSCS. She then completed her clinical cardiology fellowship atUTSW in 2015 and since then has continued on faculty with a practice in general cardiology,echocardiography, as well as remaining active in health services research and qualityimprovement.Purpose and Overview:The purpose of this presentation is to educate our department about the use of protocols forsuspected acute coronary syndrome among patients presenting with chest pain. We will reviewthe potential that new protocols have in improving the efficiency of care delivery.Educational Objectives:1. Understand the importance of differentiating cardiac from non-cardiac chest pain amongpatients presenting to the emergency department.2. To differentiate the high-sensitivity troponin from the conventional 4th generationtroponin.3. To describe different rule out myocardial infarction protocols that incorporate the use ofhigh-sensitivity troponin.4. To understand the rationale for the creation of the PHHS/UTSW hs-cTnT Protocol.5. Gain an awareness of how this protocol has affected resource utilization in the health caresystem.

Chest pain in the emergency department and ED Overcrowding:Chest pain is the second most common complaint among patients presenting to theemergency department (ED) in the United States and in 2013 there were over 6 million ED visitsfor chest pain.[1] The large majority of patients who present with chest pain do not havemyocardial infarction, with a prevalence of MI rate that varies based on the population studied.The high volume of chest pain is a contributor to emergency department overcrowding. There isa need to increase the efficiency in the evaluation of patients presenting with chest pain.Emergency department overcrowding is a major public health problem in the US and hasbeen associated with poor outcomes, increased resource utilization, and restricted access to care.In a study of two large Canadian hospitals of patients presenting with chest pain or shortness ofbreath, investigators found that during periods of ED crowding, the mean time to physicianassessment was 107.3 minutes as compared to 76 minutes during periods in which the ED wasnot crowded.[2] In a study of 187 hospitals in California in 2007, ED crowding as defined bytop quartile of diversion hours per facility was associated with higher odds of inpatient death,longer length of stay, and increased costs per admission.[3] Finally, overcrowding has beenassociated with delays in care among patients with STEMI.[4]Despite the low prevalence of ACS among patients presenting with chest pain and newdiscoveries in the diagnosis and management of acute chest pain, missing myocardial infarctionsstill occur and are a frequent cause of litigation.[5] In a prospective trial of 10,689 patientspresenting to 10 US hospitals in 1993, 17% of patients had either AMI or unstable angina. Ofthe patients with AMI, 2.1% were discharged from the ED and of those with unstable angina,2.3% were discharged from the ED. Failure to hospitalize these patients was associated withhigher risk-adjusted mortality.[6] In a more contemporary study of ED patients in North Texasbetween 2009 and 2015, the rates of missed ACS (patient diagnosed with ACS 7 days after anindex ER visit) was 3.2%. Additionally, the rates of missed ACS did not change throughout thestudy period.[7]The annual ED volume at PHHS is high. For the fiscal year of 2017 – 2018, there were244,209 arrivals to the PHHS ED. The monthly proportion of patients who undergo cardiacevaluation is also high and from 2017 to September 2019, there were on average 2,604 monthlyencounters in which patients underwent ECG and troponin testing. Given the high volume ofpatients with chest pain evaluated in the ED and the overcrowding that can ensue without rapidtriage of patients, it is important to establish efficient and effective ED protocols for the saferule-out of low risk chest pain. The ideal protocol for chest pain rule out would be fast, highlysensitive and have the ability to identify patients at very low risk for 30 day adverse cardiacevents such that these patients can be discharged home instead of being admitted for observationor additional inpatient testing.

Chest pain work-up and role of hs-cTnT in the diagnosis of AMI:The fourth universal definition ofmyocardial infarction emphasizes thedifference between myocardial injury andinfarction. Injury is defined by an elevation introponin with at least one level above the 99thpercentile upper reference limit (URL). Thiscan be either acute, if there is a rise/fall introponin, or chronic, if the elevation does notchange significantly over serial measurement.Myocardial infarction is diagnosed when thereis acute myocardial injury with evidence ofischemia.[8] There is a broad differentialdiagnosis for myocardial injury. Acutemyocardial injury can occur in the setting ofacute heart failure or myocarditis. ChronicFigure 1: Interpreting Myocardial Injury, Fourth Universal Definitionmyocardial injury can occur in the setting ofof Myocardial Infarction. Thygesen, K., et al., Circulation, 2018.structural heart disease or chronic kidneydisease. Finally, there are several differenttypes of acute myocardial infarction. Type I occurs secondary to plaque rupture or plaqueerosion. Type 2 can occur secondary to oxygen supply mismatch states such as hypertension orarrhythmias. Type 3 is defined by patients who suffer cardiac death before biomarker elevationoccurs or when MI is detected on autopsy. Type 4 and 5 are related to PCI and CABG.Potential algorithms for chest pain rule out:After the history, physical, and ECG, cardiac biomarkers are needed to evaluate formyocardial injury. The ACC/AHA NSTEMI guidelines recommend that they be obtained atbaseline and 3-6 hours after symptom onset to evaluate for rise/falling pattern.[9] Thedisadvantage of this protocol is that patients typically have to wait in the ER for prolongedmonitoring to rule out MI.Newer, high-sensitivity assays have been developed and tested in many different rule outalgorithms outside the U.S., and the U.S. FDA first approved the use of the hs-cTnT assay in2017. This assay is both highly sensitive and precise. Because of the sensitivity, small increasescan be detected and in some patients, a single blood draw can exclude ACS. Because of theprecision of the assay, small changes over time can be monitored such that MI can be ruled outby sequential blood draws in which there are no significant dynamic changes. By definition, ahigh-sensitivity assay is one in which the coefficient of variance (CV) is 10% at the 99thpercentile upper reference limit in the population. Additionally, concentrations below the 99thpercentile should be above the assay’s minimum limit of detection for 50% of healthyindividuals in the population.[10, 11]

Despite the widespread use of the hs-cTnT in Europe, there is no standardized protocolthat has been endorsed by the ESC for the evaluation of chest pain. The following are anoverview of several different options for implementation.I. Replace the 4th generation cTnT with the hs-cTnT assay and use 99th percentile ULN:One strategy is to simply replace the current troponin test with the 99th percentile ULN ofthe hs-cTnT. The diagnostic accuracy of several hs-cTnT assays and the 4th generation assaywere compared in a multicenter observational study. The APACE (The Advantageous Predictorsof Acute Coronary Syndrome Evaluation) study is a prospective, international multicenter studythat evaluated 786 patients who presented to EDswith symptoms concerning for AMI. In thestudy, investigators measured different troponinassays and adjudicated the final diagnosis in allpatients. The sensitivity of the standard 4thgeneration cTnT assay was 83%, the negativepredictive value was 97%, and the positivepredictive value was 72%. In comparison, theRoche hs-cTnT had a sensitivity of 95%, negativepredictive value of 99%, and positive predictivevalue of 50%. Additionally, this figuredemonstrates the higher accuracy of the highsensitivity assays in the diagnosis of MI which ismore pronounced earlier in the presentation ofFigure 2: Diagnostic Accuracy of Cardiac Troponin Assays atchest pain.[12]Presentation According to Time since Onset of Chest Pain, Reichlin, T.,et al, N Engl J Med, 2009.As demonstrated by this study by Reichlin, the higher sensitivity and negative predictivevalue of the hs-cTnT assays also comes with the expense of a lower positive predictive value.This problem is likely to be magnified in populations in which the prevalence of myocardialinfarction is lower. This tabledemonstrates this problem. Asthe MI probability goes down, theadditional positive test resultswith the hs-cTnT assay vs. thestandard assay increases (falsepositives). Therefore, theproportion of positive tests withthe high-sensitivity assay who donot have myocardial infarctionTable 1: Estimated Proportion of False Positive MI Diagnoses with hs-cTnT assay, De Lemos, JA, Clin Chemwill increase.[13]2011.

II. 0/3 hour Rule Out Strategy:The European Society ofCardiology 2011 guidelines recommend a0/3 hour algorithm using the 99thpercentile ULN at 0 and 3 hours as seen inthis figure. The disadvantages of thisprotocol are that it requires many patientsto remain in the ED for the 3 hour blooddraw, relies on the 99th percentile ULNwhich will lead to higher rates of falsepositives in some populations, and doesnot specify the change values that are to beconsidered abnormal.[14]Figure 3: 0/3 Rule Out Strategy, Roffi M., et al,. Eur Heart J, 2016.III. 0/1 hour rule out strategy:Investigators developed a 0/1 hour ruleout strategy from a prospective cohort study ofpatients presenting with chest pain to an ED inBasel, Switzerland. In this study, a 0/1 rule outalgorithm was derived with optimal thresholdsfor rule-in and rule-out of myocardial infarctionin a derivation subset of the study and thenvalidated in the other half of the cohort. Theinvestigators found that the a rule-out in whichthe initial troponin was 12 ng/L and the deltaFigure 4: 0/1 hour rule out strategy, Reichlin, T., et al., Arch Intern Med,between the 0 and 1 hour values was 3, had a2012sensitivity and negative predictive value of100%. A rule-in was considered when the 0 hour value was 52 ng/L and the delta between the0 and 1 hour value was 52 ng/L.[15] The advantages of using a 0/1 hour algorithm are that ittakes advantage of the assay precision, is rapid, and well validated. Disadvantages include thecomplex timing of the blood draws and that the algorithm cannot be applied to early presenters.Finally, it leaves a proportion of patients in an indeterminate zone in which further evaluation isrequired.IV. 0 Hour rule out strategy:Several studies have evaluated a 0 hour rule out strategy in which MI is ruled out on asingle blood draw. First, in a prospective cohort study of 703 patients from the UK with chest

pain who were evaluated with both the standard and hs-cTnT, investigators adjudicated finaldiagnosis of myocardial infarction and calculated the diagnostic performance of both tests. Theinvestigators found that the sensitivity of a hs-cTnT 3ng/l which was the limit of detection ofthe assay was 100% (95% CI 97.2 – 100.0) and negative predictive value was 100% (95% CI98.1 – 100.0).[16] In a second retrospective study in which 7, 130 patients underwent testingwith the hs-cTnT, a initial hs-cTnT 6ng/L had a sensitivity of 99.8% for AMI (95% CI 98.7 –100) and a negative predictive value of 99.9 (95% CI 99.8 – 100).[17] Therefore, the 0 hour ruleout strategy is safe, but it only applies to a small proportion of patients presenting with chest pain(in the study by McRae, et al, only 42% had a initial value of 6 ng/L). Therefore many patientswill require subsequent testing. The European society of cardiology NSTEMI guidelinesendorse a 0 hour rule out strategy that can be applied to patients who have had chest pain for 3hours prior to presentation.[18]V. Use of Risk Scores/Accelerated Diagnostic Pathways:Numerous risk stratification protocols have been developed to identify patients at lowrisk for adverse outcomes who may be safely discharged from the ED after presenting with chestpain. The HEART score is an algorithm that was previously used in the PHHS ED for patientspresenting with chest pain, but not ACS andis predictive of MACE at 6 weeks (ACS,PCI/CABG, death). This score wasoriginally developed in a cohort of 122 chestpain patients and categorizes individualsinto low, intermediate, and high risk. Thescore takes into account history, ECG, age,risk factors, and troponin levels.[19] Ascore of 0-3 confers a risk of 2.5%, score of4-6 points, 20.3%, and 7 points, risk of72.7%.Figure 5: HEART Score Components, Six, AJ, et al., Neth HeartJ, 2008.Further studies have evaluated the modified HEART score which combines the history,ECG, age, risk factors elements of the traditional heart score with the hs-cTnT early rule outalgorithms. The TRA