Authors: Steve Goodacre, Valerie Lechene, Graham Cooper, Sarah Wilson, Jim Zhong
Categories: Practice, 1565
Source: The BMJ
A healthy woman in her mid 50s experiences sudden, tearing pain, like a lightning bolt from her neck to her chest, radiating to her back, coming in waves, with severity fluctuating over subsequent hours. At times she is able to talk and even walk, but she feels that her consciousness level is mostly reduced, and she has difficulty breathing. She feels dizzy and nauseous. Her mother survived a type A aortic dissection, three years previously, at the age of 77.
An ambulance is called and arrives 90 minutes later. A paramedic makes a tentative diagnosis of aortic dissection based on the presenting features. The woman is given oral morphine and transported to hospital, arriving 45 minutes later. In the emergency department she receives an initial diagnosis of panic attack and is managed conservatively, until reassessment some hours later triggers computed tomography angiography, which shows an aortic dissection.
Acute aortic syndrome (AAS) is a life threatening condition where a tear in the thoracic aorta can lead to rupture of the aorta and death. It encompasses three acute aortic dissection; intra-mural haematoma; and penetrating ulcer,^1^ and is commonly classified into Stanford type A (involving the ascending aorta) and type B (sparing the ascending aorta) or DeBakey classification, with type 1 involving ascending and descending aorta and type 2 involving ascending aorta alone (fig 1). Without treatment, AAS can progress to aortic rupture, with rapid deterioration and death.
Fig 1 Classification of aortic dissection
AAS is uncommon. Meta-analysis of population based studies from North America, Europe, Asia, and Australasia estimated a pooled incidence of 4.8 per 100 000 individuals per year, with 3.0 per 100 000/year type A and 1.6 per 100 000/year type B aortic dissections.^2^ Mean patient age in the studies varied from 58.9 to 77.3 years and the proportion of men varied from 50% to 84%. Hospital episode statistics for England in 2022-23 reported 1542 admissions with dissection of the aorta out of 6 million emergency admissions.^3^ Aortic dissection accounts for around three quarters of AAS.^4^
AAS is easily missed because similar symptoms are reported by patients with other much more common diagnoses, such as acute coronary syndrome, gastro-oesophageal reflux, and panic attacks. Chest pain is the most common presenting symptom of AAS,^5^ but was also the chief presenting reason for 6% of emergency department attendances in England in 2022-23.^6^ A US retrospective cohort study of 33 emergency departments estimated that one aortic dissection was diagnosed in every 980 attendances with atraumatic chest pain.^7^ Low rates of exposure to a diagnosis of AAS may mean that clinicians fail to consider it as a possible diagnosis alongside other more common causes of chest pain. Our case presentation illustrates the diagnosis of AAS initially being overlooked in the emergency department in favour of a more common diagnosis (panic attack). Clinicians who assess acute chest pain need to be aware of AAS and how it is investigated, to avoid misdiagnosis.
A systematic review of 12 studies (1663 patients) estimated that one in three patients with an eventual diagnosis of aortic dissection was initially misdiagnosed.^8^ The most common misdiagnoses were acute coronary syndrome, stroke, and pulmonary embolism. A more recent estimate from a population based retrospective cohort study of 1299 patients diagnosed with AAS in Ontario, Canada between 2003 and 2018, identified that 13% had attended an emergency department in the previous 14 days with symptoms suggesting AAS.^9^
Missed diagnosis can lead to delayed surgery for type A aortic dissection and missed opportunities for medical management (blood pressure control) or emergency intervention for type B aortic dissection. Missed diagnosis of type A dissection is associated with an approximate doubling of mortality (hazard ratio 2.14, 95% confidence interval 0.89 to 5.13)^10^ and delayed surgery is associated with increased mortality (67% at 8-12 hours versus 20% at 0-4 hours after diagnosis).^11^ Blood pressure control using beta blockers is associated with an approximate halving of mortality in type B dissection.^12^
NHS Resolution, an organisation of the UK Department of Health and Social Care that provides expertise on resolving healthcare based concerns and disputes, identified aortic disease, including dissection, as a common cause of fatality related negligence claims.^13^ A study of 135 medical practice litigations across the US involving aortic dissection cited failure to diagnose as the reason for litigation in 64%.^14^ A review by the Healthcare Safety Investigation Branch found that half of patients with acute aortic dissection die before reaching any specialist centre in the UK^15^ and a systematic review of 14 studies of out-of-hospital cardiac arrest identified that the 7% caused by aortic dissection had 100% mortality.^16^
AAS is definitively diagnosed by computed tomographic angiography scanning of the aorta (fig 2), or other imaging techniques, such as electrocardiography gated computed tomographic angiography or magnetic resonance angiography. Computed tomographic angiography incurs costs and incurs small risks of radiation induced malignancy and reaction to contrast media. Clinicians therefore use clinical assessment and biomarkers (if appropriate) to assess AAS risk and select patients for imaging. If the patient is unwell with typical features of AAS and AAS is strongly suspected, then arrange a computed tomographic angiography without delay.
Fig 2 Computed tomographic angiography showing aortic dissection with true lumen (TL) and false lumen (FL)
The diagnostic challenge of AAS is well recognised^17^ but recent research has clarified the role of clinical assessment and biomarkers.^18^ ^19^ ^20^
Consider risk factors, symptoms, and signs to estimate the probability of AAS. Assessment may be structured, using a clinical score or algorithm, or unstructured, using clinical judgment. Several scores or algorithms have been developed for AAS but only the aortic dissection detection risk score (ADD-RS) has been widely studied.^18^ ADD-RS gives a score between zero (low risk) and three (high risk) by allocating one point each if the patient has a risk factor for AAS, a symptom suggesting AAS, or a sign of AAS (table 1).
A meta-analysis of 11 cohort studies of ADD-RS^18^ reported that a score greater than zero had 94.6% (95% credible interval 90% to 97.5%) sensitivity and 34.7% (95% credible interval 20.7% to 51.2%) specificity for AAS, while a score greater than one had 43.4% (95% credible interval 31.2% to 57.1%) sensitivity and 89.3% (95% credible interval 80.4% to 94.8%) specificity. The included studies were heterogeneous and had variable methodological quality. The low prevalence of AAS in the clinically relevant population means that sensitivity of 95% could be sufficient to rule out AAS, while specificity of 90% is required to avoid over-investigation. These findings could be interpreted as suggesting that patients with a score of two or three should be selected for imaging while those with a score of zero would not benefit from further testing. How patients with a score of one should be managed is uncertain. The risk of AAS associated with each ADD-RS score depends upon the prevalence of AAS in the population being tested. If ADD-RS is used in a very low prevalence population, such as all patients with chest pain, then even scores of two or three will be associated with a low risk of AAS.
The patient in our case presentation had a high risk condition (family history) and high risk pain features, giving a score of two, and suggesting computed tomographic angiography would be of benefit.
Electrocardiography can diagnose acute coronary syndrome and other causes of acute chest pain but does not assist with diagnosis of AAS.
Blood tests (biomarkers) can be used to select patients with suspected AAS for imaging. D-dimer is the only biomarker that has been extensively studied for diagnosing AAS. Many other biomarkers have had limited evaluation, but none are ready for clinical use.^19^ A meta-analysis of 18 cohort studies of D-dimer using a threshold of 500 ng/mL reported 96.5% (95% credible interval 94.8% to 98%) sensitivity and 56.2% (95% credible interval 48.3% to 63.9%) specificity for AAS, but noted uncertainty around estimates due to risk of bias and heterogeneity.^20^ This is similar to the sensitivity and specificity of D-dimer for diagnosing venous thromboembolism^21^ and suggests that D-dimer could rule out AAS in patients with a low or intermediate clinical risk (as determined by the ADD-RS or unstructured assessment), but indiscriminate use in patients with a very low clinical risk of AAS could lead to over-use of computed tomographic angiography. D-dimer sensitivity does not appear to be time dependent. A cohort study of 273 patients diagnosed with AAS estimated that D-dimer sensitivity was 97% within one hour of symptom onset and did not vary with time from symptom onset.^22^ Age adjusted D-dimer may offer improved specificity compared with a fixed threshold but requires further evaluation.
A systematic review of four studies evaluating emergency physician point-of-care ultrasound for thoracic aortic dissection reported sensitivities ranging from 41% to 91% and specificities of 94% to 100% when an intimal flap was seen.^23^ A more recent prospective cohort study (n=1314) of a point-of-care ultrasound protocol combining transthoracic echocardiography with scanning of the abdominal aorta reported 93.2% sensitivity and 90.9% specificity.^24^ This suggests a possible role for point-of-care ultrasound in the emergency department diagnosis of AAS, but the role of operator experience needs to be determined. Point-of-care ultrasound is a core skill for emergency physicians, but additional training would be required for diagnosing AAS.
ADD-RS can be combined with D-dimer in various ways. A recent meta-analysis combined data from six studies of ADD-RS and D-dimer to estimate sensitivities and specificities and provide direct comparisons between alternative strategies.^18^ Table 2 outlines the sensitivities and specificities of using ADD-RS with D-dimer to select patients for imaging or using each test alone. These provide a range of trade-offs between sensitivity and specificity. International guidelines vary in their recommendations for using ADD-RS and D-dimer to select patients for computed tomographic angiography (see international guidelines box).
AAS should be considered in patients with chest, back, or abdominal pain, syncope, or symptoms related to malperfusion. However, this population has a very low prevalence of AAS.^25^ Applying diagnostic strategies for AAS to all such patients would result in low positive predictive value and very high use of computed tomographic angiography. Clinicians therefore need to apply diagnostic strategies selectively to those with a non-negligible risk of AAS, such as those with an additional feature suggesting AAS (“chest pain plus one”). A recent cohort study of 5548 patients attending the emergency department with possible symptoms of AAS found that clinicians rated the likelihood of AAS as zero in 2315/4111 (56%).^25^ Applying diagnostic strategies only to those with a non-zero likelihood of AAS could result in a more deliverable rate of computed tomographic angiography but it is currently unclear how clinicians determine a zero likelihood of AAS and whether this judgment is accurate.
In the UK, AAS is managed according to principles set out in the NHS acute aortic dissection toolkit,^26^ which NHS England produced to improve outcomes from AAS. Elsewhere, local guidelines are applicable.^5^ ^27^ ^28^ Acute management involves analgesia and reducing systolic blood pressure to 100-120 mmHg. Type A AAS is usually managed operatively in a regional aortic centre. Type B AAS is split into complicated or non-complicated by the presence of haemodynamic instability and/or malperfusion of an organ system or limb. Uncomplicated type B AAS is usually managed medically with blood pressure control. Although patients may not require transfer to a tertiary centre, they should all be discussed with a cardiothoracic or vascular specialist to agree management. Complicated type B AAS may require tertiary transfer for endovascular stent graft placement. In-hospital mortality is 22% for type A and 13% for type B aortic dissection.^29^
Research into biomarkers may produce new tests to assist with diagnosis of AAS, while further evaluation of ADD-RS, D-dimer, and point-of-care ultrasound may clarify their role. This could lead to reduced risk of misdiagnosis and reduced reliance on computed tomographic angiography to rule out AAS.
The delay in diagnosis may have been caused by initial misdiagnosis as a panic attack. The patient received appropriate treatment once the correct diagnosis was made and she recovered. However, the diagnostic delay remains a source of anxiety.
Advisers to this series are Anthony Harnden, professor of primary care, Department of Primary Care Health Sciences, University of Oxford, and Dr Kevin Barraclough, School of Social and Community Medicine, University of Bristol.