Emergency Department Electrocardiogram Research Articles

Assessment of current pediatric electrocardiogram (ECG) interpretation practices

BackgroundDespite improvements in digital electrocardiograms (ECGs), current standard of care requires physician confirmation. Mismatched expectations between ordering providers and ECG readers, often pediatric cardiologists and electrophysiologists (EPs), are common, especially since there are no standardized practices for pediatric ECG reading. ObjectivesThe aim of this study was to understand current practices in pediatric ECG reading. MethodsAn electronic survey was sent to members of the Pediatric and Congenital Electrophysiology Society (PACES). Participation was optional; results were recorded from 12/19/22–1/9/23. Only complete and non-duplicate entries were included. ResultsA total of 127 responses were received, 93 were analyzed. Most responses were from centers in North America (n = 65, 70 %), including the US (n = 58, 62 %), Canada (n = 6, 6 %), and Mexico (n = 1, 1 %). The remaining were from Europe (n = 18, 19 %), Asia (n = 7, 8 %), Australia (n = 2, 2 %), and South America (n = 1, 1 %). Most (n = 46, 49 %) were from small centers (0–25 ECGs read per day), 27 respondents (29 %) were from medium centers (26–50 ECGs read per day), 20 respondents (22 %) were from large centers (>50 ECGs read per day). The majority (n = 65, 70 %) reported >3 readers/day for inpatient and emergency department ECGs. 49 % (n = 46) of centers read ECGs >2 times/day on weekdays with more variable practice on weekends. For critical/time sensitive findings, most centers (n = 90, 97 %) used verbal communication +/− the EMR. There was consensus (≥50 % agreement) that the following findings are critical/time sensitive: QTc >500 ms, T-wave alternans, narrow complex tachycardia, wide complex tachycardia, pre-excited atrial fibrillation, focal ischemic changes, second degree heart block type II, complete heart block, and pacemaker malfunction. ConclusionReading practices are variable. Critical/time sensitive findings are most often communicated verbally, however, there is no agreed upon standard. There was consensus in critical/time-sensitive findings. Improved understanding of common practices and resource allocation may lead to increased consistency in pediatric ECG reading.

Abstract 10966: Autoimmune Limbic Encephalitis Mimicking Inferior Myocardial Infarction on ECG

Case Presentation: A 64 year-old man was brought to the emergency department (ED) after being found down, awake, but unresponsive by family. In the ED, he denied chest pain and shortness of breath. His past medical history included coronary artery disease status-post coronary artery bypass grafting, heart failure with reduced ejection fraction, and paroxysmal atrial fibrillation. On physical exam, he was hemodynamically stable, oriented to self and place with no other significant neurological findings. The field and ED electrocardiograms (ECGs) (Figure 1) were consistent with an acute inferior ST elevation myocardial infarction (STEMI). The patient underwent emergent coronary angiography via left heart catheterization which showed no acute lesions requiring intervention. Initial pertinent lab work included a peak troponin-I of 0.095 ng/mL and serum sodium of 126 mmol/L. Collateral information revealed that the patient has progressive memory loss and intermittent episodes of hand tremors with non-purposeful eye movements. He also had a witnessed faciobrachial dystonic seizure, which is pathognomonic for Leucine-rich, Glioma Inactivated-1 (LGI-1) autoimmune encephalitis. Serum and CSF analyses were positive for LGI-1 IgG antibodies, confirming the diagnosis. The patient was treated with high-dose steroids, IVIG, and valproic acid with improvement in seizure activity and cognitive status, and discharged on prednisone and levetiracetam. Discussion: Per review of the current literature, central nervous system pathology may induce sympathetic overactivity causing transient myocardial injury, referred to as neurogenic stunned myocardium. This can manifest as left ventricular dysfunction, myocardial injury, and/or ECG changes consistent with myocardial infarction. This case presentation highlights an example of a seizure leading to neurogenic stunned myocardium with an ECG showing an acute inferior STEMI.

Change in out-of-hospital 12-lead ECG diagnostic classification following resuscitation from cardiac arrest

IntroductionWe evaluated the incidence of change in serial 12-lead electrocardiogram (ECG) diagnostic classifications in patients resuscitated from out-of-hospital (OH) cardiac arrest (OHCA) comparing OH to emergency department (ED) ECGs. MethodsThis retrospective case series included: 1) adults (≥ 18 years old), 2) resuscitated from OHCA, 3) ≥ 1 OH and 1 ED ECG/patient, and 4) emergency medical services (EMS) transport to the study hospital. OH and ED ECGs were classified as: 1) STEMI (ST-segment Elevation Myocardial Infarction), 2) Ischemic, and 3) Non-ischemic. Two ED physicians and one cardiologist independently classified all ECGs, then generated a consensus opinion classification for each ECG based on American Heart Association’s 2018 Expert Consensus criteria. The most ischemic OH ECG classification was compared with the last ED ECG classification. ResultsFrom 7/27/12 to 7/18/19, 176 patients were entered with a mean age of 61.2 ± 16.6 years; 102/176 (58%) were male. Overall, 504 OH and ED 12-lead ECGs were acquired (2.9 ECGs/patient). ECG classification inter-rater reliability kappa score was 0.63 ± 0.02 (substantial agreement). Overall, 86/176 (49%) changed ECG classification from the OH to ED setting; 69/86 (80%) of these ECGs changed from more to less ischemic classifications. Of 49 OH STEMI ECG classifications, 33/49 (67%) changed to a less ischemic (non-STEMI) ED ECG classification. ConclusionsChange in 12-lead ECG classification from OH to ED setting in patients resuscitated from OHCA was common (49%). The OH STEMI classification changed to a less ischemic (non-STEMI) ED classification in 67% of cases.

Using ECG-To-Activation Time to Assess Emergency Physicians' Diagnostic Time for Acute Coronary Occlusion.

There is no quality metric for emergency physicians' diagnostic time for acute coronary occlusion. We sought to quantify diagnostic time associated with automated interpretation, classic ST-elevation myocardial infarction (STEMI) criteria, STEMI-equivalents, and subtle occlusions, using electrocardiogram (ECG)-to-activation of catheterization laboratory time. This multicenter retrospective study reviewed all code STEMI patients from the emergency department (ED) with confirmed culprit lesions from January 2016 to December 2018. We measured door-to-ECG (DTE) time and ECG-to-activation (ETA) time. We examined the first ED ECGs to determine whether automated interpretation labeled "STEMI," and they met classic STEMI criteria, STEMI-equivalents, or rules for subtle occlusion. ECG analysis was performed by two emergency physicians blinded to clinical scenario, automated interpretation, and angiographic outcome. There were 177 code STEMIs with culprit lesions, with a median DTE time of 9.0min and a median ETA time of 16.0min. Automated interpretation labeled 55.4% of first ECGs "STEMI" (ETA 6.5min) and 44.6% not "STEMI" (ETA 66min, p<0.0001). Of first ECGs, 63.8% met classic STEMI criteria (ETA 8.0min), 8.5% had STEMI-equivalents (ETA 32.0min, p=0.0026), 16.4% had subtle occlusions (ETA 89.0min, p=0.045), and 11.3% had no diagnostic sign of occlusion (ETA 68.0min, p=0.20). STEMI criteria missed more than one-third of occlusions on first ECG, but most had STEMI-equivalents or rules for subtle occlusion. ETA time can serve as a quality metric for emergency physicians to promote new ECG insights and assess quality improvement initiatives.

Getting to the heart of the issue: senior emergency resident electrocardiogram interpretation and its impact on quality assurance events.

ObjectiveElectrocardiogram (ECG) interpretation skills are of critical importance for diagnostic accuracy and patient safety. In our emergency department (ED), senior third-year emergency medicine residents (EM3s) are the initial interpreters of all ED ECGs. While this is an integral part of emergency medicine education, the accuracy of ECG interpretation is unknown. We aimed to review the adverse quality assurance (QA) events associated with ECG interpretation by EM3s.MethodsWe conducted a retrospective study of all ED ECGs performed between October 2015 and October 2018, which were read primarily by EM3s, at an urban tertiary care medical center treating 56,000 patients per year. All cases referred to the ED QA committee during this time were reviewed. Cases involving a perceived error were referred to a 20-member committee of ED leadership staff, attendings, residents, and nurses for further consensus review. Ninety-five percent confidence intervals (CIs) were calculated.ResultsEM3s read 92,928 ECGs during the study period. Of the 3,983 total ED QA cases reviewed, errors were identified in 268 (6.7%; 95% CI, 6.0%–7.6%). Four of the 268 errors involved ECG misinterpretation or failure to act on an ECG abnormality by a resident (1.5%; 95% CI, 0.0%–2.9%).ConclusionA small percentage of the cases referred to the QA committee were a result of EM3 misinterpretation of ECGs. The majority of emergency medicine residencies do not include the senior resident as a primary interpreter of ECGs. These findings support the use of EM3s as initial ED ECG interpreters to increase their clinical exposure.

Poor Agreement but Good Predictive Value Between Automated and Manual QTc Intervals in Pediatric Emergency Department Electrocardiograms.

QTc interval is significant because prolongation may lead to ventricular dysrhythmia. Computerized electrocardiogram machines typically measure QT interval length and, using an algorithm assessment of multiple leads, calculate a QTc value. Manual measurement of the QT interval used to calculate the QTc value is more time-consuming but potentially more accurate. In this study, we compare the automated QTc calculation with the QTc value calculated using manual QT measurements. We prospectively obtained 350 resting 12-lead electrocardiograms (ECGs) in children aged 2 to 14 years in an academic pediatric emergency department. Manual measurement of the QT interval was performed and the QTc was calculated using the 2 most commonly used correction methods, Bazzet and Fridericia formulas. The paired values were used to perform a Bland-Altman analysis and create a receiver operating characteristic curve. Bland-Altman analysis determined that QT-automated and QTc-Bazett had an average difference of 3.8 milliseconds, with a standard deviation of 86 milliseconds (95% confidence interval = -161 to 176). An automated QTc value of 455 milliseconds was sensitive to detect manual QTc values of greater than 480 milliseconds. In children with resting ECGs, there is a poor agreement between the automated QTc produced by a computerized electrocardiogram and the QTc value obtained using manual QT measurement. Statistically and clinically relevant discrepancy between the automated QTc and QTc values calculated after manual QT measurement was present. Automated QTc values may be used as a screening tool to detect prolonged QTc, but for accurate determination of QTc, manual measurement is necessary.

MP06: Using electrocardiogram-to-activation time to assess emergency physicians’ diagnostic delay of acute coronary occlusion

Introduction: Electrocardiogram (ECG) diagnosis of acute coronary occlusion has been broadening in recent years, from classic ST-Elevation Myocardial Infarction (STEMI) criteria to STEMI-equivalents and rules for subtle occlusions. However, there is no quality metric focused on emergency physicians’ decision-making. We hypothesized that the time from initial emergency department (ED) ECG to activation of Code STEMI could quantify diagnostic delay associated with automated interpretation, classic STEMI criteria, and other signs of occlusion. Methods: This multi-centre retrospective study reviewed all ED Code STEMI patients with confirmed culprit lesions from two urban academic EDs over a three-year period (Jan 2016 to Dec 2018). We reviewed charts to calculate ECG-to-Activation (ETA) time, measured from the time stamp on the initial ED ECG to the time a Code STEMI was activated (based on the hospital call centre log). We examined ECGs to determine: 1) if automated computer interpretation labelled “STEMI” or not; and 2) whether they met classic STEMI criteria, STEMI-equivalent patterns, or rules for subtle occlusion, based on a priori criteria from published guidelines or studies. All ECGs were reviewed by the lead author (JTTM) and those not obviously meeting classic STEMI criteria were independently reviewed by the other author. Results: There were 180 Code STEMIs from the ED with culprit lesions, including 177 with complete information. Average ETA time was 46.5 minutes (95% Confidence Interval [CI] 36.3-56.7min). Automated interpretation labelled 55.4% of initial ECGs as “STEMI” (ETA 13.9 min, 95%CI 9.8-18.0min), and 44.6% not as “STEMI” (ETA 86.9min, 95%CI 67.9-105.9min). Initial ECGs included 62.1% with classic STEMI criteria (ETA 17.3min, 95%CI 12.8-21.8min), 11.3% with STEMI-equivalents (ETA 49.5min, 95%CI 29.5-69.5min), 18.1% with subtle occlusions (ETA 118.3min, 95%CI 81.5-155.1min) and 8.5% with no initial sign of occlusion (ETA 102.9min, 95%CI 53.9-151.9min). Inter-rater reliability was very good (Cohen's kappa 0.84). Conclusion: Over 90% of Code STEMI patients with culprit lesions had initial ECGs diagnostic of acute coronary occlusion, but automated interpretation and classic STEMI criteria only identified 55.4% and 62.1%, respectively. STEMI-equivalents and subtle occlusions were associated with significant diagnostic delays. ETA time can serve as a quality metric for emergency physicians and may help guide ED quality improvement initiatives.

A deep neural network learning algorithm outperforms a conventional algorithm for emergency department electrocardiogram interpretation

BackgroundCardiologs® has developed the first electrocardiogram (ECG) algorithm that uses a deep neural network (DNN) for full 12‑lead ECG analysis, including rhythm, QRS and ST-T-U waves. We compared the accuracy of the first version of Cardiologs® DNN algorithm to the Mortara/Veritas® conventional algorithm in emergency department (ED) ECGs. MethodsIndividual ECG diagnoses were prospectively mapped to one of 16 pre-specified groups of ECG diagnoses, which were further classified as “major” ECG abnormality or not. Automated interpretations were compared to blinded experts'. The primary outcome was the performance of the algorithms in finding at least one “major” abnormality. The secondary outcome was the proportion of all ECGs for which all groups were identified, with no false negative or false positive groups (“accurate ECG interpretation”). Additionally, we measured sensitivity and positive predictive value (PPV) for any abnormal group. ResultsCardiologs® vs. Veritas® accuracy for finding a major abnormality was 92.2% vs. 87.2% (p < 0.0001), with comparable sensitivity (88.7% vs. 92.0%, p = 0.086), improved specificity (94.0% vs. 84.7%, p < 0.0001) and improved positive predictive value (PPV 88.2% vs. 75.4%, p < 0.0001). Cardiologs® had accurate ECG interpretation for 72.0% (95% CI: 69.6–74.2) of ECGs vs. 59.8% (57.3–62.3) for Veritas® (P < 0.0001). Sensitivity for any abnormal group for Cardiologs® and Veritas®, respectively, was 69.6% (95CI 66.7–72.3) vs. 68.3% (95CI 65.3–71.1) (NS). Positive Predictive Value was 74.0% (71.1–76.7) for Cardiologs® vs. 56.5% (53.7–59.3) for Veritas® (P < 0.0001). ConclusionCardiologs' DNN was more accurate and specific in identifying ECGs with at least one major abnormal group. It had a significantly higher rate of accurate ECG interpretation, with similar sensitivity and higher PPV.

Abstract 244: ECG Recording Errors Prevalence in a Hospital Setting

Background: Electrocardiographic recording can have many potential errors due to erroneous recording which may impact their proper interpretation and patient care. OBJECTIVE: To assess the prevalence of reported ECG errors in a university hospital database with a mix of inpatient and outpatient electrocardiograms (ECG). Methods: The 1212 ECGs performed during one month were downloaded and individually analyzed for reported errors and artifacts. There were 309 (25.5%) outpatient ECGs and 903 (74.5%) inpatient or emergency department ECGs. The types of errors are reported as absolute numbers and as percentages of the total errors. Results: Among 1212 ECGs analyzed, 71 ECGs (5.9%) contained error or artifact statements. Among those, 12 ECGs (16.9%) were precordial lead misplacements. There were 2 ECGs (2.8%) with limb lead reversals. A total of 57 ECGs (80.3%) contained baseline artifacts; 12 ECGs (16.9%) had AC noise artifacts, and 45 ECGs (63.3%) had baseline wander artifacts. Out of the total ECGs with reported errors, 60 ECGs (84.5%) were performed in the inpatient or emergency department setting, while 11 ECGs (15.5%) were performed in the outpatient setting. Discussion: Electrocardiography remains the cornerstone of early diagnosis and treatment of cardiac disease in the inpatient and outpatient setting. ECG recording artifacts have been reported with variable frequency, with inpatient recordings likely suffering greater errors, depending on the acuity of the situation. We observed that in the outpatient setting, there was 3.6% (11/309) prevalence of recording error, while in the inpatient and emergency department setting there was 6.6% (60/903) prevalence of ECG recording errors. Despite the overall small percentages, this is an alarming problem, especially with regards to inpatient ECGs when the results often prompt critical treatment decisions. The reason for the discrepancy between inpatient and outpatient ECG errors is not clear. Potential causes may include the intensity of the situation with compromise of proper lead placement, such as use of Mason-Likar configuration or placing leads above the left breast in female patients or unavailability of proper recording surfaces as in trauma or burn patients, but may also be caused by the lack of availability of adequately trained staff especially in the after-hour situations. Adequate training of staff to perform proper ECGs, and creation of robust electrocardiographic interpretation algorithms to recognize more recording errors on-screen before printing is called for.

Terminal QRS distortion is present in anterior myocardial infarction but absent in early repolarization

BackgroundEarly repolarization (ER) and acute left anterior descending artery occlusion (LADO) may be difficult to distinguish. Terminal QRS distortion (TQRSD), defined by the absence of both an S wave and J wave in either of leads V2 or V3, is often present in anterior ST-segment elevation myocardial infarction. We hypothesized that this finding would always be absent in ER. MethodsThis was a retrospective analysis of electrocardiograms (ECGs) of consecutive patients who presented to the emergency department with ischemic symptoms and had a cardiologist interpretation of “benign ER” on the initial emergency department ECG. All ECGs were scrutinized for the presence of an S wave and a J wave in leads V2 and V3. Differences in S-wave amplitudes between complexes with and without J waves were analyzed using nonparametric Mann-Whitney testing and confidence intervals around a proportion. ResultsOne hundred seventy-one patients were identified with benign ER. Zero of 171 had TQRSD (specificity for LADO, 100%; 95% confidence interval, 97.8-100). In lead V2, S waves were absent in only 1 of 171 ECGs; however, in that ECG, a J wave measuring 0.5 mm was present. In lead V3, S waves were absent in 16 ECGs, but all of these ECGs had J waves. When J waves were absent in leads V2 or V3, the corresponding S waves were deeper than S waves in QRS complexes with J waves. ConclusionTerminal QRS distortion was never observed in benign ER. Based on previous studies indicating the presence of TQRSD in LADO, it was, thus, 100% specific to LADO when the differential diagnosis was acute myocardial infarction vs ER.

Electrocardiographic indicators of acute coronary syndrome are more common in patients with ambulance transport compared to those who self-transport to the emergency department journal of electrocardiology

IntroductionThe American Heart Association recommends individuals with symptoms suggestive of acute coronary syndrome (ACS) activate the Emergency Medical Services' (EMS) 911 system for ambulance transport to the emergency department (ED), which enables treatment to begin prior to hospital arrival. Despite this recommendation, the majority of patients with symptoms suspicious of ACS continue to self-transport to the ED. The IMMEDIATE AIM study was a prospective study that enrolled individuals who presented to the ED with ischemic symptoms. ObjectivesThe purpose of this secondary analysis was to determine differences in patients presenting the ED for possible ACS who arrive by ambulance versus self-transport on: 1) time-to-initial hospital electrocardiogram (ECG), 2) presence of ischemic ECG changes, and 3) patient characteristics. MethodsInitial 12-lead ECGs acquired upon patient arrival to the ED were evaluated for ST-elevation, ST-depression, and T-wave inversion.ECG signs of ischemia were analyzed both individually and collapsed into an independent dichotomous variable (ED ECG ischemia yes/no) for statistical analysis. Patient characteristics tested included: gender, age, race, ethnicity, English speaking, living alone, mode of transport, and presenting symptoms (chest pain, jaw pain, shortness of breath, nausea/vomiting, syncope, and clinical history). ResultsIn 1299 patients (mean age 63.9, 46.7% male), 384 (29.6%) patients arrived by ambulance to the ED. The mean time-to-initial ECG was 47minutes for ambulance patients versus 53minutes for self-transport patients (p<0.001). Mode of transport was found to be an independent predictor for time-to-initial ECG controlling for age, gender, and race (p=0.004). There were significantly higher rates of ECG changes of ischemia for patients who arrived by ambulance versus self-transport (p=0.02), and patient characteristics differed by mode of transport to the ED. DiscussionOur findings indicate that less than 30% of individuals with symptoms of ACS activate the EMS ‘911’ system for ambulance transport to the ED. Individuals more likely to activate 911 have timelier ECG but higher rates of ischemic changes, specifically ST-depression and T-wave inversion. Individuals least likely to activate 911 are women, younger individuals, Latino ethnicity, live with a significant other, and those experiencing chest or jaw pain.

Abstract 236: Prevalence of ECG Recording Errors in a University Hospital ECG Database

Background: Electrocardiographic recording can suffer from several potential errors which may impact their proper interpretation and patient care. Objective: To assess the prevalence of reported ECG errors in a university hospital database with a mix of inpatient and outpatient electrocardiograms (ECG). Methods: The first 1000 ECGs performed during one month were downloaded and individually analyzed for reported errors and artifacts. There were 237 (23.7%) outpatient ECGs and 763 (76.3%) inpatient or emergency department ECGs. The types of errors are reported as absolute numbers and as percentages of the total errors. Results: Among 1000 ECGs analyzed, 65 ECGs (6.5%) contained error or artifact statements. Among those, 15 ECGs (23.1%) were perceived precordial lead misplacements simulating anterior myocardial infarctions. There were 5 ECGs (7.7%) with lead reversals; 3 of which were precordial lead reversals (4.6%) and 2 were limb lead reversals (3.1%). A total of 45 ECGs (69.2%) contained baseline artifacts; 27 ECGs (41.5%) had AC noise artifacts, and 18 ECGs (27.7%) had baseline wander artifacts. Out of the total ECGs with reported errors, 60 ECGs (92.3%) were performed in the inpatient or emergency department setting, while 5 ECGs (7.7%) were performed in the outpatient setting. Discussion: Electrocardiography remains the cornerstone of early diagnosis and treatment of cardiac disease in the inpatient and outpatient setting. ECG recording artifacts have been reported with variable frequency, with inpatient recordings likely suffering greater errors, depending on the acuity of the situation. We observed that in the outpatient setting, there was 2.1% (5/237) prevalence of recording error, while in the inpatient and emergency department setting there was 7.9% (60/763) prevalence of ECG recording errors. Despite the overall small percentages, this is an alarming problem, especially with regards to inpatient ECGs when the results often prompt critical treatment decisions. The reason for the discrepancy between inpatient and outpatient ECG errors is not clear. Potential causes may include the intensity of the situation with compromise of proper lead placement, such as use of Mason-Likar configuration or placing leads above the left breast in female patients or unavailability of proper recording surfaces as in trauma or burn patients, but may also be caused by the lack of availability of adequately trained staff especially in the after-hour situations. Adequate training of staff to perform proper ECGs, and creation of robust electrocardiographic interpretation algorithms to recognize more recording errors on-screen before printing is called for.

Left Dominant Arrhythmogenic Cardiomyopathy Causing Sustained Ventricular Tachycardia – A Case Report

A 56-year-old white man was admitted to hospital because of palpitations and dyspnoea. He had no family history of cardiomyopathy or sudden death, and the electrocardiograms (ECG) of his son and parents were normal. He had an history of palpitations, and in the emergency department ECG showed sustained ventricular tachycardia-flutter with inferior axis right bundle branch block (RBBB) in V1 and heart rate 250 beats per minute. His standard ECG showed a complete RBBB pattern with minor ST-T segment changes. 2D echo and coronary angiography were normal, while magnetic resonance imaging (MRI) revealed a reduction in left mid-lateral wall thickening with hypokinetic movement and mild reduction of ejection fraction. An Extensive sub-epicardial area of conspicuous left ventricular intra-myocardial fat and sub-epicardial left ventricular late enhancement in lateral wall, was also seen at MRI. No abnormalities of right ventricular size, function or tissue characterisation were detected. A diagnosis of atypical left dominant arrhythmogenic cardiomyopathy was made, based on clinical and images basis, as gene testing was not available.

Speed and accuracy of text-messaging emergency department electrocardiograms from a small community hospital to a provincial referral center

Background Currently, transmission of electrocardiograms (EKGs) from a small emergency department (ED) to specialists at referral hospitals can be a time-consuming and laborious process. We investigate whether text messaging by use of short message service (SMS) of EKGs from a small hospital to consultants at a large hospital is rapid and accurate. Methods This study involved a one-month prospective evaluation of consecutive EKGs recorded in a small community ED. Investigators obtained de-identified photographs of each EKG via a mobile phone camera. Each EKG picture, along with a brief patient clinical history, was sent via SMS to on-call emergency physicians located at a large referral care site. All images were evaluated solely on a mobile phone. The primary outcome was the proportion of SMS that were received within two minutes of being sent. As a secondary outcome, the intra-rater evaluation of the initial EKG and the SMS EKG image were compared on 13 standardized features. The tertiary outcome was cost of text messaging. Results A total of 298 patients (14.6%) had 409 EKGs performed and a total of 926 SMS were sent. 921 SMS (99.5%, 95% confidence interval (CI) 98.7–99.8%) arrived within two minutes with a median transmission time of nine seconds (interquartile range (IQR) 3–32 s). Between the gold standard original EKG, and the interpretation of the texted image, six out of 409 (1.5%, 95% CI 0.6–3.3%) had any differences recorded, across all 13 categories. Overall, the study cost 4.1 cents per texted image. Conclusions Systematic text messaging of ED EKGs from a small community hospital to a referral center is a rapid, accurate, portable, and inexpensive method of data transfer. This may be a safe and effective strategy to communicate vital patient information.

Derivation of a risk scale and quantification of risk factors for serious adverse events in adult emergency department syncope patients

Determining the appropriate disposition of emergency department (ED) syncope patients is challenging. Previously developed decision tools have poor diagnostic test characteristics and methodological flaws in their derivation that preclude their use. We sought to develop a scale to risk-stratify adult ED syncope patients at risk for serious adverse events (SAEs) within 30 days. We conducted a medical record review to include syncope patients age ≥ 16 years and excluded patients with ongoing altered mental status, alcohol or illicit drug use, seizure, head injury leading to loss of consciousness, or severe trauma requiring admission. We collected 105 predictor variables (demographics, event characteristics, comorbidities, medications, vital signs, clinical examination findings, emergency medical services and ED electrocardiogram/monitor characteristics, investigations, and disposition variables) and information on the occurrence of predefined SAEs. Univariate and multiple logistic regression analyses were performed. Among 505 enrolled patient visits, 49 (9.7%) suffered an SAE. Predictors of SAE and their resulting point scores were as follows: age ≥ 75 years (1), shortness of breath (2), lowest ED systolic blood pressure < 80 mm Hg (2), Ottawa Electrocardiographic Criteria present (2), and blood urea nitrogen > 15 mmol/L (3). The final score calculated by addition of the individual scores for each variable (range 0-10) was found to accurately stratify patients into low risk (score < 1, 0% SAE risk), moderate risk (score 1, 3.7% SAE risk), or high risk (score > 1, ≥ 10% SAE risk). We derived a risk scale that accurately predicts SAEs within 30 days in ED syncope patients. If validated, this will be a potentially useful clinical decision tool for emergency physicians, may allow judicious use of health care resources, and may improve patient care and safety.

Prehospital ECG Provides Important Information

Do you routinely track down the prehospital electrocardiogram (ECG) and compare it to tracings obtained in the emergency department (ED)? Researchers in Quebec compared prehospital and first ED ECGs for 1209 patients who ultimately received diagnoses of cardiac disorders, including ST-segment …

Information on Myocardial Ischemia and Arrhythmias Added by Prehospital Electrocardiograms

Background. The prehospital electrocardiogram (ECG) allows earlier identification of acute ST-segment elevation myocardial infarction (STEMI). Its utility for detection of other acute cardiac events, as well as for transient ST-segment abnormalities no longer present when the first hospital ECG is performed, is not well characterized. Objective. We sought to examine whether the prehospital ECG adds supplemental information to the first ECG obtained in hospital, by comparing data on possible cardiac ischemia and arrhythmias provided by the two ECGs, in ambulance patients later diagnosed as having cardiac disorders, including STEMI. Methods. Ambulance personnel acquired 12-lead ECGs for patients suspected of having an acute ischemic event, prior to transport to hospital. The first emergency department 12-lead ECG was provided by medical records at the receiving hospital, and the principal hospital diagnosis for the event was extracted from chart data. Two cardiologists, blinded to the hospital diagnosis, provided their consensus interpretation of 1,209 pairs of ECGs, noting the presence or absence of specific abnormalities on each tracing. Results. Among the 82 patients who had an eventual hospital diagnosis of STEMI, the study cardiologists identified 71 with ST-segment elevations on the ECGs they examined. The vast majority of these (97%) were observed on both ECGs, but the prehospital ECG showed ST-segment elevation for two additional patients (3%). No additional instances were seen only on the hospital ECG. Among the 116 patients with a hospital diagnosis of non–ST-segment elevation myocardial infarction (NSTEMI), the cardiologists identified 36 with ST-segment depressions: 28 (78%) of these were present on both ECGs, seven (19%) only on the prehospital ECG, and one (3%) only on the hospital ECG. Among the 567 patients with any cardiac hospital diagnosis, the cardiologists identified 87 with arrhythmias: 73 (84%) on both ECGs, 12 (14%) only on the prehospital ECG, and two (2%) only on the hospital ECG. Conclusions. Beyond identifying ST-segment elevation earlier, prehospital ECGs detect important transient abnormalities, information not otherwise available from the first emergency department ECG. These data can expedite diagnosis and clinical management decisions among patients suspected of having an acute cardiac event. The prehospital ECG should be fully integrated into emergency medicine practice.

Abstract TP203: New Onset Atrial Fibrillation Post Stroke

Background: Atrial fibrillation (AF) makes a person five times more likely to experience a stroke. According to the National Stroke Association, atrial fibrillation accounts for about 15 percent of stroke. Many of these patients do not have any previous history of AF, and are diagnosed with this condition post-stroke. We sought to characterize the frequency and features of AF in stroke patients hospitalized after acute presentation, focusing on new diagnoses. Methods: All subjects were enrolled in the Field Administration of Stroke Therapy- Magnesium (FAST-MAG) clinical trial, a phase 3 NIH-funded study of pre-hospital Magnesium Sulfate vs. placebo for patients with symptom onset &lt;2hours. General demographic information, past medical history, first electrocardiogram (ECG) in the emergency department (ED), and final diagnosis data were collected on consecutive subjects. Results: Of 1478 patients, 69 (4.7%) had no previous history of AF but were diagnosed with AF post-stroke onset. There were 274 cases with AF recorded on the ED ECG (274), only 36 (13.1%) had no documented history of AF. Patients in both groups were of similar age (age 81 vs 79 years) and had similar rates of ICH diagnosis (8.7 vs 7.1%). However, newly diagnosed patients had more severe strokes (median NIHSS 14 vs 10, p=0.008) and a lower burden of cardiac disease by history (coronary artery disease 15.9 vs 31.7%, and myocardial infarction 5.8 vs 15.9%). Female gender women (50.7 vs 48.2%) history of prior stroke (10.1 vs 8.8%), hypertension (89.9 vs 89.5%), and hyperlipidemia (56.5 vs 54.1%) were similar in both groups. Conclusion: Most new diagnoses of AF after stroke were made based on the ED ECG. Strokes in patients who would go on to have new AF diagnoses were more severe than those where AF diagnosis was already established, likely due to less likelihood of anticoagulation.

Incremental Value of Computed Tomography in Triaging Emergency Department Patients With Acute Chest Pain

In North America, more than 5 million patients per yearpresent to the emergency department (ED) with acute chestpain [1]. Early triage is essential for both prognosis andtreatment; however, current management methods do noteffectively assess those patients in whom initial cardiacbiomarkers and electrocardiogram (ECG) changes areinconclusive. Although 80% of patients admitted to thehospital undergo extensive testing, which often provenoncardiac causes of chest pain, 2%-8% of dischargedpatients inadvertently have acute coronary syndrome(ACS) [2].At the same time, it is important to note that chest painmay result from a broad range of etiologies, includingcardiac, pulmonary, musculoskeletal, and psychologicalorigins. As such, a thorough history and physical examina-tion of the patient should be undertaken to narrow thedifferential diagnosis before imaging is requested. In thisway, clinicians may be able to arrive at a diagnosis withoutthe need for unnecessary imaging studies.Recent advances in multidetector computed tomography(MDCT) and the advent of the triple-rule-out (TRO) protocolprovide noninvasive visualization of coronary and/orpulmonary arteries, thoracic aorta, and the other intrathoracicstructures. By evaluating both coronary and noncoronarydisease, the TRO scan provides a cost-effective diagnosticstudy for acute chest pain [3]. The aim of this article is toprovide an overview of the imaging modalities used in theevaluation of patients presenting with acute chest pain, witha focus on the role and added value of MDCT.Initial AssessmentTraditionally, the initial assessment in the ED entailsa thorough history, physical examination, and ECG, as wellas measurement of cardiac biomarkers. However, up to 28%of patients present with atypical chest pain or delayed find-ings [4], which limit the sensitivity of traditional assess-ments. Up to 20% of patients with ACS have atypical ornonexistent chest pain, and up to 10% of patients initiallydiagnosed with an myocardial infarction are later found tohave normal or nonspecific ECG findings [4]. Moreover, themeasurement of troponin has high specificity but lowsensitivity for an ACS in the initial hours of presentation.Due to these diagnostic errors, the American College ofCardiology (ACC) and American Heart Association (AHA)have published guidelines on risk stratification for patientswith ACS. Patients in the first category (low risk) and secondcategory (high risk) pose no major difficulty to diagnosis andtreat [5]. However, patients in the third category (interme-diate risk) are more difficult to triage [5]. This group consistsof patients often in their fourth through sixth decade of life,with findings that tend to be indeterminate, nonspecific, oratypical, with few if any risk factors. Imaging has thegreatest potential for risk stratification and clinical decisionmaking in this challenging category.Standard Diagnostic TestingCurrent AHA-ACC guidelines recommend a multimodalapproach to diagnostic testing [5]. For early triage, functionaltests are of limited value because of the requirement forserial negative biomarkers, the specific expertise ofpersonnel, and the frequency of nondiagnostic tests [1]. The

The utility of the prehospital electrocardiogram

The 12-lead electrocardiogram (ECG) can capture valuable information in the prehospital setting. By the time patients are assessed by an emergency department (ED) physician, their symptoms and any ECG changes may have resolved. We sought to determine whether the prehospital electrocardiogram (pECG) could influence ED management and how often the pECG was available to and reviewed by the ED physician. A retrospective medical record review was conducted on a random sample of patients ≥ 18 years who had a prehospital 12-lead ECG and were transported to one of two tertiary care centres. Data were recorded onto a standardized data extraction tool. Three investigators independently compared the pECG to the first ECG obtained in the ED after patient arrival at the hospital. Any abnormalities not present on the ED ECG were adjudicated to ascertain whether they had the potential to change ED management. Of 115 ambulance runs selected, 47 had no pECG attached to the ambulance call record (ACR) and another 5 were excluded (one ST elevation myocardial infarction, one cardiac arrest, three ACR missing). Of the 63 pECGs reviewed, 16 (25%) showed changes not apparent on the initial ED ECG (κ = 0.83; 95% CI 0.74-0.93), of which 12 had differences that might influence ED management (κ = 0.76; 95% CI 0.72-0.82). Only one hospital record contained a copy of the pECG, despite the current protocol that paramedics print two copies of the pECG on arrival in the ED (one copy for the ACR and one to be handed to the medical personnel). None of 110 ED charts documented that the pECG was reviewed by the ED physician. The pECG has the potential to influence ED management. Improvement in paramedic and physician documentation and a formal pECG handover process appear necessary.