American Journal of Emergency Medicine (2012) 30, 1622-1626

Brief Report

Acute aortic dissection in the ED: risk factors and predictors for missed diagnosis^?

Muiteng Chua MD ^a, Irwani Ibrahim MD ^a,?, Xinyi Neo MD ^a, Vitaly Sorokin MD ^b,

Liang Shen PhD ^c, Shirley B.S. Ooi MD ^a

^aEmergency Medicine Department, National University Health System (NUHS), Singapore

^bDepartment of Cardiac, Thoracic and Vascular Surgery, National University Health System, Singapore

^cBiostatistic Unit, National University Health System, Singapore

Received 2 April 2011; revised 16 November 2011; accepted 23 November 2011

Abstract

Objective: This study aims to explore the risk factors and predictors involved in the missed diagnosis of Acute aortic dissection among patients in the emergency medicine department (EMD).

Methods: This is a single-center retrospective chart review conducted over a 10-year period (January 1998 to December 2008). Records with a diagnosis of “dissection of aorta” (International Classification of Diseases, Ninth Revision code 441.0) from the hospital discharge database and hospital death register were selected. Acute aortic dissection was defined as missed if diagnostic imaging to diagnose AAD or cardiothoracic surgeon consult was not elicited while in the EMD. We compared the history, clinical findings, and investigations between patients who had the diagnosis of AAD missed in the EMD and those who did not.

Results: A total of 68 patients were included in the analysis during the study period, of which 38.2% had a missed diagnosis. There was 63.2% of type A AAD by Stanford classification. Neither age, sex, nor a history of hypertension were significant risk factors for missed diagnosis of AAD. The likelihood of missed diagnosis was significantly higher in the absence of a pulse deficit (odds ratio, 35.76; 95% confidence interval, 3.70-345.34) and absence of widened mediastinum on chest radiography (odds ratio, 33.16; 95% confidence interval, 5.74-191.49).

Conclusion: Well-known risk factors for AAD such as age, male sex, and hypertension were not risk factors for missed diagnosis for AAD presenting in the EMD. The absence of pulse deficit or widened mediastinum does not exclude the diagnosis of AAD.

(C) 2012

Introduction

Acute aortic dissection is a catastrophic cardio- vascular disease, which can be associated with high mortality

^? Conflict of interest statement: The authors declare no conflict of interest.

* Corresponding author.

E-mail address: [email protected] (I. Ibrahim).

if undiagnosed (and untreated); the mortality is 1% to 2% per hour after symptom onset [1]. Timely diagnosis is particu- larly challenging because the presenting clinical features are diverse and may mimic many other diseases [1-4]. Mortality is higher among cases of Stanford type A AAD [1,4,5]. Any Delay in diagnosis or missed diagnosis results in a lack of prompt surgical intervention to prevent lethal complications such as aortic rupture, cardiac tamponade, and visceral ischemia. It has been reported that the diagnosis of AAD was

0735-6757/$ - see front matter (C) 2012 doi:10.1016/j.ajem.2011.11.017

missed in up to 38% of patients on initial evaluation and was only first established at the postmortem examination in up to 28% of patients [2,6,7].

The clinical profiles and outcomes of AAD have previously been well discussed in multicenter studies [8,9]; however, these studies did not focus on patients presenting to the emergency medicine department (EMD). The incidence of AAD has been reported to be 0.3% in patients who attended the EMD for chest or back pain [9]. To our knowledge, there is no study available that explored the risk factors and predictors of missed diagnosis of AAD in the EMD. The few available studies of AAD involving EMD presentations have only explored Clinical predictors that diagnose AAD [10,11]. In these studies, several risk factors such as hypertension, Marfan syndrome, male sex, and advanced age have been reported to be associated with AAD. The identification of factors and predictors of missed diagnosis may assist in decreasing the rate of missed diagnosis of AAD in the EMD and, hence, lead to better outcomes. The aim of this study is to determine the risk factors and predictors involved in the missed diagnosis of AAD in the EMD.

Methods

The setting of our study was in the EMD of National University Hospital, a tertiary hospital in Singapore that has approximately 100 000 EMD attendances per year. The period of the study was from January 1998 to December 2008. The study was approved by the local institutional ethics review board.

We selected records with a diagnosis of “dissection of aorta” (International Classification of Diseases, Ninth Revision, code 441.0) from the hospital discharge database, and hospital death register were collected. Emergency medicine department attendance (index admission) that resulted in admission with the discharge diagnosis of AAD or EMD attendance within 14 days of death because of AAD was included. The 14-day period from symptom onset was designated as the acute phase of aortic dissection [1].

Patients who were transferred from other hospitals or emergency departments were excluded from the study because the authors were unable to obtain EMD records, and patients with known aortic dissection, elective admis- sions for aortic dissection surgery, and aortic dissection that occurred as a complication of Coronary artery bypass graft surgery were also excluded.

Data collected from the EMD records included patient demographics, medical history, clinical presentation, clinical examination, electrocardiogram (ECG) and imaging find- ings, and EMD diagnosis. For clinical examination, pulse deficit was defined as the absence or decrease in peripheral pulses as noted by clinicians. All chest radiographs were reviewed by EMD physicians, and on-duty radiologists were consulted when necessary. The mediastinum was classified

as widened if its width at its point of maximum convexity was greater than 30% of the width of the chest. For ECG findings, Ischemic changes were defined as new or old ST elevations or depressions, T wave inversions, and Q waves. Left ventricular hypertrophy was considered when the amplitudes of R waves in lead V₅ or V₆ plus S waves in lead V₁ or V₂ were greater than 35 mm [12].

Data were extracted by 2 independent observers. In cases where there was ambiguity on whether the diagnosis of AAD was missed, a discussion was held among the investigators to review the charts and investigation results before a final decision was made.

A missed diagnosis of AAD was defined by the following criteria: AAD was not a differential diagnosis, diagnostic imaging to diagnose AAD was not performed, or expert (cardiologist or cardiothoracic surgeon) consult was not obtained while the patients were in the EMD. The characteristics, clinical findings, and imaging findings between both groups of patients (“AAD diagnosed” and “AAD missed”) were subsequently compared to determine the risk factors and predictors for missed diagnosis of AAD in the EMD.

Statistical methods

Data were analyzed using SPSS version 15.0 (SPSS Inc, Chicago, IL). All the P values calculated were 2 sided, and the statistical significance was set at P b .05. Age was presented as mean with its standard deviation and compared using the independent Student t test. Categorical data were presented as proportions and compared using ?2 test or Fisher exact test. Stepwise logistic regression was performed to determine factors that predict missed diagnosis [13]. Clinically important variables (age, sex) and variables that showed the association, defined as P b .1, (hypertension, pulse deficit, and widened mediastinum on chest radiograph) were entered into the model. The adjusted odds ratio (OR) is expressed as OR with its corresponding 95% confidence interval (95% CI).

Results

A total of 133 records with a diagnosis of “dissection of aorta” (International Classification of Diseases, Ninth Revision, code 441.0) from the hospital discharge database and hospital death register were found. Twenty-two percent of records could not be retrieved because of migration process of older records for digitalization. A total of 68 patients were included in the analysis during the 10-year study period, of which 38.2% (26/68 patients) had a missed diagnosis. The demographics and past medical profiles of our study population are shown in Tables 1 and 2. The mean ages of patients in the “AAD missed” and “AAD diagnosed” groups were 54.4 and 55.2 years, respectively. Our study

Category	AAD missed (n = 26)	(%)	AAD diagnosed (n = 42)		(%)	P
Demographics
Mean age (y)	54.4 (+-12.0)	55.2 (+-15.5)			.80
Sex
Male	18 (69.2%)	36		(85.7%)	.10
Ethnic group
Chinese	16 (61.5%)	31		(73.8%)
Malay	8 (30.8%)	9		(21.4%)
Others	2 (7.7%)	2		(4.8%)	.56
Hypertension	12 (46.2%)	29		(69.0%)	.06
Marfan	0 (0.0%)	2		(4.8%)	.27
syndrome
Previous aortic	0 (0.0%)	4		(9.5%)	.11
surgery

population comprised 79.4% males, and 60.3% had a history of hypertension. Age, sex, and hypertension were not significant risk factors for missed diagnosis in EMD (P =

Table 1 Demographics of study participants

.80, .10, and .06, respectively). Alternative diagnoses for missed AAD included non-ST-elevation myocardial infarc- tion, angina, transient cerebral ischemia, stroke, central cord syndrome, unspecified abdominal pain, unspecified chest pain, syncope, and sepsis.

At presentation, 60.3% (41/68 patients) of the study population presented with chest pain, of whom only 2.9% (2/68 patients) had the typical history of tearing, ripping chest pain of acute onset and 8.8% (6/68 patients) had severe, worst-ever pain. Interestingly, 2 other patients had tearing, ripping abdominal pain but did not experience any chest pain.

Table 2 Clinical features of study participants

Table 3 Investigations/imaging of study participants

The other patients reported varying characteristics of chest pain, and 42.6% (29/68 patients) were unable to characterize the nature of their pain or had nonspecific pain characteristics (Table 2). There was no statistical difference in the type of chest pain between the “AAD missed” and the “AAD diagnosed” groups (P = .68). There were 63 patients who had their pulses examined. Among these, most (68.3%) did not display a pulse deficit on clinical examination (Table 2).

As for investigations, 2 patients in our study did not have any chest radiograph done. Of the 66 chest radiographs performed, 54.5% did not show widening of the mediasti- num (Table 3). Chest radiographs were not done in those 2 patients because of the following reasons: 1 patient had presented with syncope and cardiopulmonary collapse, and initial efforts were focused on resuscitation, whereas the other patient had typical features of acute ST-elevation myocardial infarction and was referred for emergency coronary angiography.

In multivariate logistic regression analysis, EMD physicians are likely to miss the diagnosis of AAD when pulse deficit was not detected on examination (OR, 35.76; 95% CI, 3.70-345.34) or when chest radiograph did not show widened mediastinum (OR, 33.16; 95% CI, 5.74-

191.49). Overall, 63.2% (43 patients) had Stanford type A AAD. An interesting thing to note is that there was no significant difference (P = .58) between the proportions of types of AAD in “AAD missed” and “AAD diagnosed” groups (proportions of type A AAD were 61.5% and 64.3%, respectively).

Category	AAD missed (n = 26)	(%)	AAD diagnosed (n = 42)		(%)	P
Clinical features Chest pain type (n = 41) Sharp Tearing Severe, worst ever Others/unable to characterize Back pain Abdominal pain Syncope Paraplegia Pulse deficit (n = 63 patients) AR murmur Congestive cardiac failure	15 (57.7%)	26		(61.9%)
	1 (6.7%)	3		(11.5%)
	0 (0.0%)	2		(7.7%)
	3 (20.0%)	3		(11.5%)
	11 (42.3%)	18		(42.9%)	.68
	6 (23.1%)	17		(40.5%)	.14
	6 (23.1%)	17		(40.5%)	.12
	3 (11.5%)	5		(11.9%)	.99
	3 (11.5%)	4		(9.5%)	.99
	1 (3.8%)	19		(45.2%)	b.001
	3 (11.5%)	4		(9.5%)	.71
	5 (19.2%)	3		(7.1%)	.24
AR indicates aortic regurgitation.

Investigations	AAD missed (n = 26)		(%) AAD diagnosed (n = 42)		(%)	P
Chest radiography	24	(92.3%)	42	(100%)
findings (n = 66)
Abnormal aortic	4	(16.7%)	8	(19.0%)	.99
configuration
Widened	3	(12.5%)	27	(64.3%)	b.001
mediastinum
Pleural effusion	3	(12.5%)	5	(11.9%)	.99
Aortic	0	(0.0%)	2	(4.8%)	.53
calcification
Cardiomegaly	7	(29.2%)	4	(9.5%)	.08
ECG findings	21	(80.8%)	39	(92.9%)
(n = 60)
No abnormalities	7	(26.9%)	12	(28.6%)	.46
LVH	2	(2.2%)	6	(14.3%)	.63
Ischemic changes	7	(26.9%)	17	(40.5%)	.41
Type of AAD a
Type A	16	(61.5%)	27	(64.3%)	.58
Type B	8	(30.8%)	14	(33.3%)
Indeterminate	2	(7.7%)	1	(2.4%)
LVH indicates left ventricular hypertrophy. a By Stanford classification, from imaging studies, that is, magnetic resonance imaging and CT scans.

Discussion

Characteristic physical and radiologic findings are not always present in AAD. In our study, there is a significantly higher likelihood of missed diagnosis when physical examination did not detect pulse deficit and when the chest radiograph did not show widened mediastinum. These 2 factors are classically associated with AAD and predict a missed diagnosis of AAD in their absence. From our review of previous studies, only 15.1% to 27.0% of patients with AAD have a pulse deficit detected clinically, whereas 60.0% to 61.6% have a widened mediastinum on chest radiographs [1,14]. The presence of pulse deficit may be more commonly found in type A AAD [15]. However, its absence does not completely rule out the diagnosis of AAD. Furthermore, identifying pulse deficit is not an objective test. It is prone to observer bias and human error. Interpretation of widened mediastinum on chest radiograph is also observer dependent. Its sensitivity and specificity vary among different observers; it ranges from 77% to 97% and 62% to 89%, respectively, among different radiologists [16]. The agreement between observers reading chest radiographs is only moderate (? = 0.49) [16]. The absence of widened mediastinum also does not rule out AAD because dissection may have occurred in the distal aorta.

Apart from the absence of classic features, missed diagnoses can also be attributed to the varied presentation of AAD that can mimic other disease entities such as acute coronary syndrome or Cerebrovascular events. For instance, a positive troponin test and ischemic changes on ECG may prompt the clinician to make a diagnosis of acute coronary syndrome. These atypical clinical features are possible cofounders that may lead to a delayed diagnosis [4]. On the other hand, the frequency of characteristic findings is also low [1]. The typical clinical features of sudden onset of tearing, ripping chest pain, and others [2] may not be present [1]. Some patients present with complications of AAD such as stroke or syncope from cerebral hypoperfu- sion and may not have any preceding chest pain [1,3]. From our study population, only 2.9% of patients presented with the typical features of central tearing chest pain of acute onset, whereas 8.8% had their worst-ever pain. Data from other studies report that 71% to 85% of the study population with AAD had presented with chest pain, with 39.0% to 50.6% having the typical tearing or ripping pain [1,4,14,17] and 90% to 90.6% having severe worst-ever chest pain [1,14]. A large proportion of our patients were unable to characterize their chest pain, and most of them (66.7%) were in their seventh decade of life and above. This could explain the deficiencies in History taking because older patients may face difficulty trying to express the type of chest pain they are experiencing because of language barriers or cognitive impairment [18].

Patients often present with a combination of the characteristics [2,4] rather than a single feature. Hence, it is important to evaluate and decide on further imaging modality

based on a combination of characteristic findings. Further investigations such as computed tomography (CT) or echocardiography may be required even in the absence of classic clinical features. This is especially so when patients have comorbidities and risk factors associated with AAD. The main risk factors include a history of hypertension, Marfan syndrome, male sex, and advanced age [1,4,14,17], with the single most important factor being hypertension. The incidence of hypertension ranges from 54% to 78% among patients with AAD in a few centers [1,2,4,5,9,19,20]. This incidence was also demonstrated in our study (60.3%). Although a history of hypertension was not statistically significant (P = .06), its high incidence and a P value close to

.05 may indicate its clinical relevance.

Acute aortic dissection has a high mortality rate, ranging from 15% to 27% in delayed and missed diagnosis [1,4]. Hence, prompt intervention is essential. One of the challenges in diagnosing AAD, apart from its diverse clinical presentation, is the lack of one single investigation that can be available on site, and which is noninvasive, rapid, and easily accessible yet offers high sensitivity and specificity [14,21]. radiographic imaging with CT is the preferred modality of investigation now because of its high diagnostic accuracy of 92% to 96% [22,23]. It also helps to exclude or diagnose other differentials of chest pain [24]. Echocardiography is an alternative modality. How- ever, transthoracic echocardiography offers a low sensitiv- ity of 70% and is usually only useful in diagnosing type A AAD [21,25]. On the other hand, transesophageal echocardiography offers a high sensitivity of 95% to 99% and a high specificity of 98% [14,25]; it is also rapid, taking less than 15 minutes, according to a study conducted by Erbel et al [25]. The main drawback of TEE is the lack of trained personnel.

This warrants the search for better, simpler, and more objective Diagnostic investigations. There have been studies on biochemical markers such as creatine kinase-BB isozyme and circulating smooth muscle myosin [26,27] for the detection of AAD. However, these studies were conducted on small sample sizes and are not widely used in the clinical setting yet. There is no current large clinical trial to validate their practical use. Currently available biomarkers such as D-dimer are often too sensitive and nonspecific. A search for a suitable marker may promise a solution to this long-standing problem.

Our study had its limitations. Because of the low incidence of aortic dissection, the sample size recruited over a 10-year period was still relatively small. Being a retrospective study and chart review, we were limited by the availability of the records in a few ways: (1) some records were outsourced for digitalization and could not be accessible, and (2) history and physical examination were limited by what we could interpret from the documenta- tion. However, our study does demonstrate clinically important findings that absence of typical features may not be reliable in excluding such a lethal diagnosis.

Conclusion

Despite its presence since ancient times, AAD remains a diagnostic challenge. Even with current medical advances, there is no single Diagnostic modality that can accurately diagnose AAD. Our study has shown that 2 characteristics of AAD-pulse deficit and widened mediastinum-when undetected on physical examination and chest radiograph, respectively, predict a misdiagnosis in EMD. Hence, a High clinical index of suspicion is necessary, and we suggest proceeding with further imaging with CT or TEE to exclude AAD in the presence of significant risk factors and/or combination of classic findings.

Acknowledgments

Data collection is assisted by Dr Cheryl Chiang, Dr Laurel Yip, and Miss Wong Sui Hua of EMD, National University Health System (NUHS), Singapore. The authors also thank the NUHS Medical Publications Support Unit, Singapore, for assistance in the preparation of the manuscript.

References

1. Hagan PGNC, Isselbacher EM, Bruckman D, Karavite DJ, Russman PL, et al. The International Registry of Acute Aortic Dissection (IRAD): new insights into an old disease. JAMA 2000;283(7):897-903.
2. Von Kodolitsch YSA, Nienaber CA. Clinical prediction of acute aortic dissection. Arch Intern Med 2000;160(19):2977-82.
3. C Ayrik HC, O Aslan O, Karcioglu E, Yilmaz E. Seeing the invisible: painless aortic dissection in the emergency setting. Emerg Med J 2006;23:e24.
4. Claudio Rapezzi SL, Graziosi M, et al. Risk factors for diagnostic delay in acute aortic dissection. Am J Cardiol 2008;102:1399-406.
5. Huseyin Ince CAN. Diagnosis and management of patients with aortic

dissection. Heart 2007;93:266-70.

1. Spittell PC SJ, Jr., Joyce JW, Tajik AJ, Edwards WD, Schaff HV, et al. Clinical features and differential diagnosis of aortic dissection: experience with 236 cases (1980 through 1990). 1993;68(7):642-51.
2. Bickerstaff LKPP, Hollier LH, Melton LJ, Van Peenen HJ, Cherry KJ, et al. Thoracic aortic aneurysms: a population based study. Surgery 1982;92(6):1103-8.
3. Suzuki TMR, Ince H, Nagai R, Sakomura Y, Weber F, et al. Clinical profiles and outcomes of acute Type B aortic dissection in the current

era: lessons from the International Registry of Aortic Dissection (IRAD). Circulation 2003;108(Suppl 1:II3):12-7.

1. Tsai TTTS, Nienaber CA. Acute aortic dissection: perspectives from the International Registry of Aortic Dissection (IRAD). Eur J Vasc Endovasc Surg 2009;37(2):149-59.
2. Sullivan PRWA, Leckey RD, Burke JL. Diagnosis of acute thoracic aortic dissection in the emergency department. Am J Emerg Med 2000;18(1):46-50.
3. Shirakabe AHN, Yokoyama S, Shinada T, Suzuki Y, Kobayashi N, et al. Diagnosic score to differentiate acute aortic dissection in the emergency room. Circ J 2008;72(6):986-9.
4. Hampton JR. The ECG made easy. Sixth ed. Churchill Livingstone; 2003.
5. Hosmer D, Lemeshow S. Applied logistic regression. New York: John Wiley & Sons Inc; 2000.
6. Jonathan Golledge KAE. Acute aortic dissection. Lancet 2008;372: 55-64.
7. Eduardo Bossone VR, Nienaber CA, Trimarchi S, Ballotta A, Mehta RH. Usefulness of pulse deficit to predict in-hospital complications and mortality in patients with Acute type A aortic dissection. Am J Cardiol 2002;89(7):851-5.
8. Ho RTBC, Bloch RD, Hoffer EK, Mann FA, Stern EJ, Wilson AJ. Can we rely on mediastinal widening on chest radiography to identify subjects with aortic injury? Emerg Radiol 2002;9(4):183-7.
9. Michael Klompas M. Does this patient have an acute thoracic aortic dissection? JAMA 2002;287(17):2262-71.
10. Steven A, Farmer DLR, Higginson IJ. Chest pain: communication of symptoms and history in a London emergency department. Patient Educ Couns 2006;63:138-44.
11. Ijaz A, Khan CKN. Clinical, diagnostic and management perspectives of aortic dissection. Chest 2002;122:311-24.
12. SW Kung WN, Ng MH. Aortic dissection in an accident and emergency department in Hong Kong. Hong Kong Med J 2007;13: 122-30.
13. Tom Treasure MJR. Investigation of suspected dissection of the thoracic aorta. Lancet 1991;338:490-5.
14. Luca Salvolini PR, Fiore D, Scaglione M, Piccoli GP, Giovagnoni A. acute aortic syndromes: role of multi-detector row CT. Eur J Radiol 2008;65:350-8.
15. Robert A, Novelline JTR, Rao PM, Stuk JL. Helical CT in emergency radiology. Radiology 1999;213:321-39.
16. Mariano Scaglione LS. Advances of MDCT in the evaluation of acute chest pain. Eur J Radiol 2008;65:331-2.
17. R Erbel RE, Daniel W, Roelandt J, Visser C, Rennollet H, the European cooperative study group for echocardiography. Echocardiography in diagnosis of aortic dissection. Lancet 1989;333(1989):457-60.
18. Hiroshima Katoh TS, Hiroi Y, Ohtaki E, Suzuki S, Yazaki Y, Nagai R. Diagnosis of aortic dissection by immunoassay for circulating smooth muscle myosin. Lancet 1995;345:191-2.
19. Riemenschneider KBM, Schmoike M, Kurz A, Guder WG. Biochemical diagnosis of aortic dissection by raised concentrations of creatine kinase BB-isozyme. Lancet 1997;350:784-5.