The Journal of Emergency Medicine, Vol. -, No. -, pp. 1–7, 2015 Copyright Ó 2015 Elsevier Inc. Printed in the USA. All rights reserved 0736-4679/$ - see front matter

http://dx.doi.org/10.1016/j.jemermed.2014.12.068

Original Contributions

SYNCOPE IN THE PEDIATRIC EMERGENCY DEPARTMENT – CAN WE PREDICT CARDIAC DISEASE BASED ON HISTORY ALONE? David Hurst, MD,* Daniel A. Hirsh, MD,† Matthew E. Oster, MD, MPH,*‡ Alexandra Ehrlich, MPH,* Robert Campbell, MD,* William T. Mahle, MD,* Michael Mallory, MD,† and Heather Phelps, DO* *Sibley Heart Center at Children’s Healthcare of Atlanta, Emory University School of Medicine, Atlanta, Georgia, †Pediatric Emergency Medicine Associates, LLC, Atlanta, Georgia, and ‡Emory University Rollins School of Public Health, Atlanta, Georgia Reprint Address: David Hurst, MD, Department of Pediatric Cardiology, Children’s Healthcare of Atlanta at Egleston, 1405 Clifton Road NE, Atlanta, GA, 30322

, Abstract—Background: The American Heart Association recommends a ‘‘meticulous history’’ when evaluating patients with an initial episode of syncope. However, little is known about which historical features are most helpful in identifying children with undiagnosed cardiac syncope. Objectives: Our objectives were 1) to describe the cardiac disease burden in Emergency Department (ED) syncope presentations, and 2) to identify which historical features are associated with a cardiac diagnosis. Methods: Using syncope presentations in our ED between May 1, 2009 and February 28, 2013, we 1) performed a cross-sectional study describing the burden of cardiac syncope, and 2) determined the sensitivity and specificity of four historical features identifying cardiac syncope. Results: Of 3445 patients, 44.5% were male presenting at 11.5 ± 4.5 years of age. Of patients with a cardiac diagnosis (68, 2%), only 3 (0.09%) were noted to have a previously undiagnosed cardiac cause of syncope: 2 with supraventricular tachycardia and 1 with myocarditis. Among the three cases and 100 randomly selected controls, the respective sensitivity and specificity of the historical features were 67% and 100% for syncope with exercise, 100% and 98% for syncope preceded by palpitations, and 67% and 70% for syncope without prodrome. The presence of at least two features yielded a sensitivity of

100% and specificity of 100%. Conclusions: Our study, which represents the largest published series of pediatric syncope presenting to the ED, confirms that newly diagnosed cardiac causes of syncope are rare. Using a few specific historical features on initial interview can help guide further work-up more precisely. Ó 2015 Elsevier Inc. , Keywords—cardiac syncope; electrocardiogram; pediatric emergency department; screening; supraventricular tachycardia; long QT syndrome

INTRODUCTION Syncope is a common complaint prompting evaluation in pediatric emergency departments (EDs). The authors of a recent review describing over 70 million pediatric ED encounters reported that syncope is the chief complaint of just less than one of every 100 patients presenting to the ED (1). In this group of patients, cardiac causes, both previously recognized and newly diagnosed, have been estimated to provide the etiology for syncope in 1–5% of patients (2–4). Although cardiac causes of syncope are rare, they have a high recurrence rate and may be associated with significant morbidity. The extent of the diagnostic evaluation needed for children with syncope in the ED has been a focus of discussion and research in the past, with the preponderance

The data collection on human patients under the auspices of Children’s Healthcare of Atlanta has been performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments.

RECEIVED: 25 July 2014; FINAL SUBMISSION RECEIVED: 10 December 2014; ACCEPTED: 22 December 2014 1

2

D. Hurst et al.

of the literature asserting that the history and physical examination are of greatest value when attempting to discern between cardiac and noncardiac causes of syncope. The American Heart Association published a statement on the evaluation of syncope in 2006, which stated that a ‘‘meticulous history’’ is of paramount importance when evaluating patients with an initial syncopal episode (5). However, no studies to date have clarified the precise historical features that have been associated with cardiovascular causes in pediatric patients presenting to the ED with syncope. The electrocardiogram (ECG) has been an important adjunct test in studies of syncopal children; its utility has been described previously in various cohorts, although not specifically in new ED cardiac diagnoses (1,3,5–7). Further, evaluations beyond a detailed history, physical examination, and ECG have been shown to be high cost and low yield in the evaluation of pediatric syncope. Echocardiography, as well as Holter and event monitoring, have low utility in the syncope population, with only 0.6% of echocardiograms demonstrating a potentially causative abnormality, and no abnormal monitors yielding a cardiac diagnosis in this population (8,9). Given that there is no standard work-up for the evaluation of syncope in the ED, our goals included providing data to inform such a standardized evaluation and clarifying more precisely which historical features are associated with cardiovascular causes to predict appropriate testing. Our objectives were 1) to describe in a pediatric ED the burden of syncope due to a new cardiac diagnosis, and 2) to identify the historical features that are associated with an underlying cardiac diagnosis. We hypothesized that patients with cardiac syncope would present with at least two of the following four historical features: syncope with exercise, syncope preceded by palpitations, syncope without prodrome, or syncope with exercise preceded by chest pain. MATERIALS AND METHODS Data Source With approval of the Institutional Review Board of Children’s Healthcare of Atlanta (GA), we reviewed the electronic medical record of the two primary EDs (Children’s Healthcare of Atlanta at Egleston and Scottish Rite) of our large pediatric tertiary care health care system from May 1, 2009 through February 28, 2013. Inclusion criteria were age eighteen years or younger with a chief complaint of syncope or near syncope (ICD-9 codes 780.2 and 780.4). Data were collected on each patient by an experienced and trained abstractor (D.A.H.) using a standard software application, including demographics, the date of service, length of time spent in the ED, the time until the next ED

visit, the chief complaint, a history of present illness, the ED diagnosis(es), the disposition from the ED, and the final hospital diagnosis(es) if admitted. ED diagnoses included ECG interpretations by the attending emergency physicians with review by a pediatric cardiologist within 24 hours. Final diagnoses were extracted from the attending of record documentation and listed in accordance with International Classification of Diseases, Ninth edition, by medical coders. Patients with a known history of structural, acquired, or electrical heart disease were not considered to represent new cases of cardiac syncope. For those with a potentially new case of cardiac syncope, outpatient charts were also reviewed where follow-up was performed. Finally, the ED and inpatient (where applicable) records of 100 controls from those diagnosed with noncardiac syncope were reviewed to evaluate the sensitivity and specificity of the historical features. The initial chart review was performed by one reviewer (D.H.), with an additional reviewer involved in the categorization of each patient with a cardiac diagnosis with no discrepancies. Outcomes and Variables of Interest Our primary outcome of interest was an ED presentation of syncope in the patient with a new cardiac diagnosis. Our primary exposure variables were the historical features of presentation including exercise-related syncope with or without chest pain, absence of a prodrome, and palpitations associated with syncope. Because the utility of ECGs in the evaluation of syncope has been reported in the past, we sought to describe previously unreported historical features that might help predict a cardiac cause of syncope in ED presentations (3,6,7). Statistical Analysis Data were summarized using mean (SD) and median (range) for continuous variables, and n (%) for categorical variables. After the descriptive analysis was completed, sensitivity and specificity, as well as positive and negative predictive values for historical features of syncope, were calculated using ED cases of syncope with a newly diagnosed cardiac etiology and 100 randomly selected patients from the same ED population with no cardiac etiology of their syncope. For all variables, we have reported 95% confidence intervals. We performed data analysis using Statistical Analysis Software (version 9.3; SAS Institute Inc., Cary, NC). RESULTS During this defined study period, 3445 patients who met inclusion criteria presented identifying a chief complaint

The Importance of History: Syncope in the Pediatric ED Table 1. Demographic Features of ED Presentations with Syncope Characteristic Gender: n (%) Male Female Race/ethnicity: n (%) White Hispanic African American Asian Unknown/declined Other* Age in years: (Mean 6 SD) Admitted: n (%)†

n (%) or Mean 6 SD 1533 (44.5) 1912 (55.5) 1714 (49.8) 419 (12.2) 1286 (37.3) 55 (1.6) 74 (2.1) 72 (2.1) 11.5 6 4.5 306 (8.9)

* American Indian = 4; Hawaiian/Pacific Islander = 3; Multi-racial = 65. † Admitted to any hospital bed from the emergency department (ED).

of syncope or near syncope, representing 0.7% of 526,715 total ED visits. Of the 3445 total patients, 44.5% were male presenting at a mean age of 11.5 6 4.5 years (Table 1). After analyzing the data set as described above, 68 patients were discovered with a documented cardiac diagnosis as part of the chief complaint, history of present illness, or final diagnosis. After a detailed chart review of each of these patients, 53 were found to have a prior diagnosis of cardiac disease (structural, electrical, or acquired) as previously described, and thus were excluded from final analysis related to our variables of interest. After exclusion of those with a prior cardiac diagnosis, there remained 15 patients (0.4%) for analysis. A more detailed chart review of these patients, including their inpatient record and outpatient charts, was performed. Of these 15 patients, there were 11 with what we have termed ‘‘incidental’’ cardiac diagnoses, that is, diagnoses that were not thought to be causative of the syncopal event at either cardiology follow-up (8/11, 73%) or, in the case of those lost to cardiology follow-up (3/11, 27%), based on documentation by the emergency physician. An a priori list of incidental diagnoses included sinus bradycardia and tachycardia, as well as first-degree atrioventricular block. In all of the latter three cases, the final diagnosis was neurocardiogenic syncope in the ED. The 11 patients with incidental diagnoses had firstdegree atrioventricular block (5/11), sinus bradycardia (5/11), and incidentally diagnosed atrial fibrillation (1/ 11) in a patient with a clinical picture strongly suggestive of vasovagal syncope, who was subsequently admitted and cardioverted into sinus rhythm after intracardiac thrombus was excluded by transesophageal echocardiogram. One final patient with a condition predisposing to vasomotor instability and a history of vasovagal syncope presented with a typical vasovagal event. He was noted to

3

have a prolonged QT interval and was subsequently lost to follow-up. This patient is noted in the legend for Figure 1. After exclusion of prior cardiac disease and incidental cardiac diagnoses, there remained three patients (0.09%) with a cardiac diagnosis as the proximate cause of their syncopal event. Of these patients, 33% were male presenting at a mean age of 14.9 6 1.6 years. For those with a cardiac cause of their syncope, 2 patients presented with supraventricular tachycardia and one with myocarditis and ventricular tachycardia. Utilizing the three cases of syncope with a cardiac etiology and 100 randomly selected controls from the same ED cohort with documented noncardiac syncope, we reviewed each of the four historical features chosen in terms of their sensitivity and specificity (with 95% confidence intervals) for a cardiac cause of syncope (Table 2). Reported syncope during exercise was 67% sensitive (95% CI 13–98) as this feature was endorsed by 2/3 of patients with cardiac syncope. This feature was 100% specific (95% CI 91–100) as none of the controls reported syncope during exercise. The negative predictive value (NPV) was 98% (95% CI 88–99) and the positive predictive value (PPV) was 100% (95% CI 20–100). Chest pain preceding exercise-related syncope also demonstrated the same sensitivity and specificity, as it was reported by 67% of patients and was absent in all of the controls. No prodrome or warning symptoms were noted by 67% (95% CI 13–98) of patients with cardiac syncope, and the specificity of this symptom was 70% (95% CI 55–82). The NPV of an absent prodrome was 97% (95% CI 84–99) and the PPV was 12% (95% CI 2–38). The historical feature with the best sensitivity and specificity was syncope preceded by palpitations, which was 100% sensitive (95% CI 30–100) and 98% specific (95% CI 88–99). The NPV for the presence of palpitations was 100% (95% CI 91–100) and the PPV was 75% (95% CI 22–99). Finally, using at least two of these features yielded a 100% sensitivity and specificity for predicting cardiac disease. DISCUSSION In this study, the largest single-center cohort of pediatric patients presenting with syncope to the ED, we found that new cardiac etiologies represented 0.4% of all presentations of syncope, with the cardiac diagnosis as the proximate cause being approximately 0.1% (1,3,4,10,11). Additionally, we found historical features to have excellent specificity, namely, syncope with exercise, chest pain preceding syncope during exercise, the absence of a prodrome, and palpitations preceding syncope. Taking any two of these features together resulted in a 100% sensitive screen for cardiac disease in our syncope population.

4

D. Hurst et al.

Figure 1. Flow diagram demonstrating identification of cardiac diagnoses. SVT = supraventricular tachycardia.

The literature on pediatric syncope has described the incidence, causes (including cardiac diagnoses that may be implicated), and the utility of diagnostic testing in both patients admitted to the hospital and those in the specialists’ office (6,8,9). There have been no studies to date that report the incidence of newly diagnosed cardiovascular causes of syncope in pediatric patients

acutely presenting to the ED for evaluation; currently available data in the ED setting include patients with prior cardiac diagnoses (10–12). Patients with underlying heart disease, when presenting with a syncopal episode, require a thoughtful and thorough evaluation. What is unclear is how extensive an evaluation is necessary for patients presenting with syncope who do

Table 2. Demographics and Presence of Historical Features in Patients with a New Cardiac Diagnosis Demographics

Historical Features

Diagnosis Age Weight Gender Race Past med history Syncope during exertion Exertion + chest pain* Palpitations† Without prodrome

Patient 1

Patient 2

Patient 3

AVNRT 13 1/12 46.6 kg Female Asian None

Myocarditis/VT 15 7/12 59.2 kg Male Black Asthma

+ + + +

+ +

AVRT 16 2/12 77.7 kg Female White Asthma UTIs + + +

AVNRT = atrioventricular nodal re-entrant tachycardia; VT = ventricular tachycardia; AVRT = atrioventricular re-entrant tachycardia; UTI = urinary tract infection. * Chest pain preceding exertional syncope. † Onset of palpitations immediately prior to syncope.

The Importance of History: Syncope in the Pediatric ED

not carry a prior cardiac diagnosis. We sought to determine what factors may help decide who is at higher risk for a cardiac cause and who ultimately may need further cardiological assessment. Although there are data describing the frequency of cardiac consultation for patients presenting with syncope, there has not been any statistical evaluation of what factors may predict the need for further testing or consultation (2). We also sought to clarify which cardiac diagnoses were ultimately thought to be causative as opposed to incidental. Descriptive studies have, in the past, grouped patients with potentially incidental findings including sinus bradycardia, low-grade atrioventricular block, or a transiently prolonged QT interval without ventricular dysrhythmia in the same category as patients with a more definitively causative cardiac etiology (12). Other studies have included neurocardiogenic syncope as cardiac in origin as opposed to describing it in a separate category indicating a transient alteration in underlying physiology (4). As nonneurocardiogenic, cardiac etiologies have a more ominous outcome, we aimed to determine the incidence of these potentially serious or life-threatening causes. The specific question regarding the utility of the ECG was not specifically addressed in our study. It has been posited that an ECG should be a part of any pediatric syncope evaluation (1). Several groups have reported the diagnostic yield of the ECG in determining a cardiac cause for syncope with a diagnostic rate of 0.004% to 1% (6,7). In another large cohort, Massin and others found 17 ECGs in 25 patients with cardiac syncope had pathologic findings, but when one excludes those with previous diagnoses, the ratio decreases to 2/25, and both cases would have been identified as potentially cardiac in origin using the screening proposed in this study (3). Certainly, further studies examining the utility of ECGs in combination with utilizing the precise historical features noted in our study could provide further insight into evaluating this patient population. At a minimum, careful attention to family medical history and risk factors for cardiomyopathies and inherited channelopathies are essential in the evaluation of pediatric patients with syncope in any setting, whether or not the patient has a normal ECG. Given that the resting ECG is an integral part of evaluating for these potentially lethal disorders, its continued use is recommended. The association of syncope related to exercise as a marker for cardiac cause has been described in the literature (2,3,13). The lack of a prodrome prior to a syncopal event and the presence of preceding palpitations have also been reported as concerning historical features (3,5,12). Whereas one study documented the presence or absence of all three of these features in each of their syncopal patients, no statistical analysis of their individual predictive qualities was undertaken (3).

5

Limitations There are several limitations to our study. We utilized coded diagnoses to develop our list of potential cardiac causes as opposed to a detailed chart review of all patients. We minimized this limitation in two ways. One was by a thorough review of the chief complaint and history of present illness of all cases, which should serve to improve identification of any additional potential cardiac concerns. The second was by detailed review of 100 randomly selected controls to determine if there were potential cardiac etiologies that were not identified. Similarly, there may have been some patients with a cardiac diagnosis that was not made by the ED or inpatient attending, as there was no review of follow-up in patients with a noncardiac diagnosis. Although it is possible that a cardiac cause was missed, given the broad time range of the study, we would expect patients with recurrent concerning episodes to return for repeat evaluation; particularly because this center is the primary provider of pediatric emergency and tertiary care in the region. Also, in review of those with incidental cardiac diagnoses, there was strong correlation between the emergency physician and the cardiologist regarding the lack of causality. Of those charts reviewed, no patient with an ED diagnosis of neurocardiogenic syncope who was seen by a cardiologist subsequently was determined to have a cardiac etiology for their event. A final limitation would be that our sensitivity estimates are limited by our small number of new cardiac diagnoses. Additional studies may be needed to verify the validity of these historical features. The strengths of this study include the size of the cohort, as well as the ability to capture cardiac diagnoses at a facility that is not only the primary provider of pediatric care in the area, but also the region’s tertiary referral center. CONCLUSION The vast majority of patients presenting with syncope in an acute care setting have a noncardiac etiology for their event. Our study clarifies the prior estimates of cardiac disease in syncopal patients and also establishes the incidence of new cardiac diagnoses in children presenting with syncope in the ED, which is quite low. Certainly, providers should continue to utilize a detailed patient medical history, a multigenerational family history, a comprehensive physical examination, and a thoughtful review of the ECG in their syncope evaluations. These data, however, provide an additional screening tool that ED providers can use to help safely and efficiently rule out potentially ominous cardiovascular causes while, at the same time, determining which patients need further cardiovascular evaluation. In this way, we hope our data

6

D. Hurst et al.

will help providers find the high-risk patients, decrease variability in evaluations, and eliminate unnecessary testing. This study provides the background data to justify a prospective study utilizing this screening algorithm to further elucidate the findings.

6.

REFERENCES

8.

1. Anderson JB, Czosek RJ, Cnota J, Meganathan K, Knilans TK, Heaton PC. Pediatric syncope: National Hospital Ambulatory Medical Care Survey results. J Emerg Med 2012;43:1–9. 2. Geggel RL. Conditions leading to pediatric cardiology consultation in a tertiary academic hospital. Pediatrics 2004;114:e409–17. 3. Massin MM, Malekzadeh-Milani S, Benatar A. Cardiac syncope in pediatric patients. Clin Cardiol 2007;85:81–5. 4. Bo I, Carano N, Agnetti A, et al. Syncope in children and adolescents: a two-year experience at the Department of Paediatrics in Parma. Acta Biomed 2009;80:36–41. 5. Strickberger SA, Benson DW, Biaggioni I, et al. AHA/ACCF scientific statement on the evaluation of syncope: from the American Heart Association Councils on Clinical Cardiology, Cardiovascular

9.

7.

10. 11. 12. 13.

Nursing, Cardiovascular Disease in the Young, and Stroke, and the Quality of Care and Outcomes Research Interdi. J Am Coll Cardiol 2006;47:473–84. Steinberg LA, Knilans TK. Syncope in children: diagnostic tests have a high cost and low yield. J Pediatr 2005;146:355–8. Zhang Q, Du J, Wang C, Du Z, Wang L, Tang C. The diagnostic protocol in children and adolescents with syncope: a multi-centre prospective study. Acta Paediatr 2009;98:879–84. Ritter S, Tani LY, Etheridge SP, Williams RV, Craig JE, Minich LL. What is the yield of screening echocardiography in pediatric syncope? Pediatrics 2000;105:e58. Saarel EV, Stefanelli CB, Fischbach PS, Serwer GA, Rosenthal A, Dick M. Transtelephonic electrocardiographic monitors for evaluation of children and adolescents with suspected arrhythmias. Pediatrics 2004;113:248–51. Driscoll DJ, Jacobsen SJ, Porter CJ, Wollan PC. Syncope in children and adolescents. J Am Coll Cardiol 1997;29:1039–45. Tretter JT, Kavey R-EW. Distinguishing cardiac syncope from vasovagal syncope in a referral population. J Pediatr 2013;163: 1618–16231. Zhang Q, Zhu L, Wang C, et al. Value of history taking in children and adolescents with cardiac syncope. Cardiol Young 2013;23:54–60. Johnsrude CL. Current approach to pediatric syncope. Pediatr Cardiol 2000;21:522–31.

The Importance of History: Syncope in the Pediatric ED

ARTICLE SUMMARY 1. Why is this topic important? Cardiac syncope, although uncommon in the pediatric emergency department (ED), is associated with significant morbidity when unidentified. 2. What does this study attempt to show? Our review attempts 1) to describe in a pediatric ED the burden of syncope due to a new cardiac diagnosis, and 2) to identify the historical features associated with an underlying cardiac diagnosis. 3. What are the key findings? Our study, which represents the largest published series of pediatric syncope presenting to the ED, confirms that newly diagnosed cardiac causes of syncope are rare. Using a few specific historical features on initial interview can help guide further work-up more precisely. 4. How is patient care impacted? This screening tool provides a mechanism by which ED providers can safely and efficiently rule out potentially ominous cardiovascular causes while, at the same time, determining which patients need further cardiovascular evaluation.

7

Syncope in the Pediatric Emergency Department - Can We Predict Cardiac Disease Based on History Alone?

The American Heart Association recommends a "meticulous history" when evaluating patients with an initial episode of syncope. However, little is known...
315KB Sizes 0 Downloads 8 Views