Abstract
Background
Cervical artery dissection (CeAD) is a common cause of stroke in young adults but is frequently misdiagnosed due to its nonspecific symptoms. This study examines factors associated with possible diagnostic error in CeAD and evaluates its impact on patient outcomes.
Methods
We conducted a secondary analysis of the STOP‐CAD (Antithrombotic Therapy for Stroke Prevention in Cervical Artery Dissection) study, which is a multicenter international study of adult patients admitted with CeAD. Possible diagnostic error was defined as the presence of CeAD symptoms within 30 days before the index CeAD diagnosis. The comparison group included patients diagnosed on their first medical encounter. Multivariable regression was used to identify factors associated with possible diagnostic error. Primary and secondary outcomes included ischemic stroke, death, and modified Rankin Scale score <2.
Results
Of 4012 patients (mean age 47.5 years, 44.6% female), 663 (16.5%) experienced possible diagnostic error. Among these, 224 (33.8%) reported to have ischemic stroke before CeAD diagnosis. Patients with possible diagnostic error were younger (odds ratio [OR], 0.89, P<0.001), more likely to have a history of migraines (OR, 1.35, P=0.007), and more likely to present with headaches (OR, 1.43, P<0.001), but less likely to show focal neurologic signs (OR, 0.68, P<0.001). There was no significant difference between groups in ischemic stroke after diagnosis (adjusted OR, 0.96, P=0.86), 90‐day modified Rankin Scale score <2 (adjusted OR, 1.04, P=0.80), or death (adjusted OR, 0.58, P=0.34).
Conclusions
One in 6 patients with CeAD experienced a possible diagnostic error, particularly those who were younger, had migraines, or presented with headaches and nonfocal symptoms.
Keywords: dissection, misdiagnosis, stroke
Subject Categories: Cerebrovascular Disease/Stroke
Nonstandard Abbreviations and Acronyms
- STOP‐CAD
Antithrombotic Therapy for Stroke Prevention in Cervical Artery Dissection
Clinical Perspective.
What Is New?
In a large international STOP‐CAD (Antithrombotic Therapy for Stroke Prevention in Cervical Artery Dissection) cohort, ~1 in 6 patients with cervical artery dissection had a prior health care encounter within 30 days before diagnosis for symptoms attributable to dissection.
Despite diagnostic delay, possible diagnostic error was not associated with a higher risk of ischemic stroke after diagnosis, worse functional outcomes or increased mortality.
What Are the Clinical Implications?
Cervical artery dissection should remain a key diagnostic consideration in younger patients presenting with new or atypical headache or neck pain, even in the absence of focal neurologic deficits.
Cervical artery dissection (CeAD) accounts for nearly one quarter of ischemic strokes in young adults under the age of 50 years. 1 Around one third of patients with CeAD develop symptoms of cerebral ischemia, the majority being ischemic stroke. 2 However, many patients develop local signs and symptoms such as headache or neck pain without signs of cerebral ischemia. These can be mistaken for a primary headache disorder such as migraine or other head and neck pathologies, leading to a misdiagnosis of CeAD. Indeed, data from a large administrative claims study indicated that ~1 in 30 dissection patients were probably misdiagnosed in the emergency department (ED) in the 2 weeks before their index diagnosis. 3 However, administrative claims data has limitations with inaccurate coding, lack of detailed data on correlates of misdiagnosis, and misdiagnosis was studied only in ED settings. To our knowledge, no prior studies have evaluated possible diagnostic error in CeAD in both ED and outpatient settings. It is also unclear whether certain patient characteristics are associated with an increased risk of possible misdiagnosis of CeAD and whether a delay in CeAD diagnosis leads to adverse outcomes.
We sought to evaluate the incidence of possible diagnostic error among patients with CeAD in a large multicenter international cohort, to identify factors associated with possible diagnostic error, and to determine whether it affected outcome in patients with CeAD. According to the National Academies of Science, Engineering, and Medicine, diagnostic error is defined as the failure to (1) establish an accurate and timely explanation of the patient's health problem(s) or (2) communicate that explanation to the patient; it encompasses terms like missed, delayed, and misdiagnosis. 4
METHODS
Aggregate data are available from the corresponding author upon reasonable request.
Institutional review board approval was obtained from each participating center to perform the study. Patient consent was waived due to the retrospective design of the study.
This article was written in adherence to the Strengthening the Reporting of Observational Studies in Epidemiology checklist for observational studies.
Study Design
This was a planned secondary analysis of the multicenter, cross‐sectional, international, retrospective STOP‐CAD (Antithrombotic Therapy for Stroke Prevention in Cervical Artery Dissection) study, which included patients presenting to an acute care hospital with a diagnosis of CeAD without major trauma. 5 Adult patients aged 18 and older diagnosed with CeAD were identified using International Classification of Diseases, Ninth Revision (ICD‐9; 443.21, 443.24) and Tenth Revision (ICD‐10) codes (I77.71, I77.74 and I77.75). Details of the study including inclusion and exclusion criteria were previously described. 5
Possible Diagnostic Error
For the purposes of this analysis, possible diagnostic error of CeAD was defined as presence of possible CeAD symptoms (such as headache, cervical or neck pain, Horner's syndrome, focal neurologic deficit, transient neurologic symptoms, peripheral cranial nerve palsy, dysphagia, tinnitus, retinal infarct, amaurosis) in a medical encounter within 30 days before the index CeAD diagnosis. The comparison group without possible diagnostic error consisted of patients with CeAD diagnosis made on first medical encounter. We chose the 30‐day window based on prior literature, as this period is reported to be high risk for recurrent stroke after cerebrovascular events and large studies investigating possible diagnostic error have used a similar time definition. 6 , 7 , 8
The STOP‐CAD data abstraction forms included questions to ascertain whether a patient may have presented with CeAD symptoms to a medical setting before their index CeAD diagnosis (possible diagnostic error). These questions were presented in the following sequential order in the data abstraction form using branching logic, whereby subsequent questions would appear only if the answer to Question 1 was “yes”:
Question 1: Was the patient seen by a health care professional within the past 30 days for similar symptoms?
Question 2: In which setting was the [CeAD] diagnosis not made [possible diagnostic error]?
Question 3: Did the patient have an ischemic stroke after the initial encounter, but before [CeAD] diagnosis? 4
Question 4: In what setting was the diagnosis made?
Possible diagnostic error was defined as a “yes” answer to Question 1. For Question 2, the misdiagnosis setting was divided into ED, outpatient neurology clinic, outpatient primary or urgent care, or other. We defined the possible diagnostic error of CeAD as having any medical encounter in the ED or office settings within 30 days before CeAD hospitalization for any neurological symptom reported above attributable to CeAD.
Covariates
The following covariates were included:
Demographic variables: age, biological sex, Race (White, Black, Asian, American Indian or Alaska, or Unknown), Hispanic ethnicity.
Comorbidities/medications: hypertension, diabetes, hyperlipidemia, migraine, active smoking, recent COVID infection, recent upper respiratory infection, known connective tissue disease, cervical manipulation, pregnancy, postpartum, use of selective serotonin reuptake inhibitor.
Clinical presentation: ischemic versus nonischemic symptoms (headache, neck pain, tinnitus, Horner's syndrome) on presentation during both possible diagnostic error encounter and during encounter for index hospitalization, National Institutes of Health Stroke Scale score during encounter for index hospitalization.
Imaging findings: normal baseline computed tomography scan, dissection location (single vertebral dissection, single carotid dissection, single intracranial dissection, multivessel dissection), acute infarct on imaging, occlusive dissection, presence of partially occlusive thrombus, intracranial extension, or pseudoaneurysm. All imaging variables were evaluated during the index diagnosis encounter, as no imaging data are available for encounter with possible diagnostic error.
Study Outcomes
Our primary outcome was ischemic stroke after index CeAD diagnosis within 180 days. Strokes occurring between the initial possible diagnostic error encounter and the index diagnosis date were reported by sites but not included as outcome as strokes occurring in this time interval were not confirmed. Secondary outcomes included death and excellent functional outcome defined as 90‐day modified Rankin Scale score <2 from the time of CeAD diagnosis.
Statistical Analysis
We compared the aforementioned covariates among patients with and without possible diagnostic error using rank sum test for continuous variables and χ2 test for categorical variables. We used multivariable regression to identify factors associated with possible diagnostic error. The following covariates included in multivariable regression models were selected a priori based on clinical relevance and prior literature on CeAD and diagnostic delay: age, female sex, race, hyperlipidemia, migraine, postpartum state, selective serotonin reuptake inhibitor use, presenting with headache, presenting with neck pain, presenting with focal deficit, dissection vessel location, acute infarct on imaging, occlusive dissection, and pseudoaneurysm. Univariable analyses were performed for descriptive purposes only and were not used to determine covariate inclusion. Multivariable logistic regression analysis was used to assess the association between possible diagnostic error and the primary and secondary outcomes of interest. Specifically, models evaluating ischemic stroke and mortality were adjusted for demographic factors and imaging markers of dissection severity (age, female sex, occlusive dissection, partially occlusive thrombus, multivessel dissection, intracranial extension; mortality also adjusted for National Institutes of Health Stroke Scale score). Models evaluating functional outcome were adjusted for age, female sex, and baseline stroke severity (National Institutes of Health Stroke Scale score). These covariates were selected a priori based on clinical relevance and factors known to influence stroke outcomes or dissection severity. Analyses were performed using STATA SE 18.0 (StataCorp, College Station, TX). Analyses were conducted using a 2‐sided significance level of 0.05. Missing data were not imputed.
RESULTS
Baseline Characteristics
Of 4023 STOP‐CAD study patients, 4012 were included in this substudy after excluding 11 patients with unavailable variables about possible diagnostic error. The mean±SD age of included subjects was 47.5 (13.3) years and 1788 (44.6%) were women. A total of 663 (16.5%) had a medical encounter for symptoms attributable to CeAD within 30 days preceding the index CeAD diagnosis (the possible diagnostic error group). Of these, 314 were seen in the ED, 272 in primary care/urgent care clinics, 65 in the inpatient setting, 33 in outpatient neurology clinics, and 32 in other settings. Fifty patients had >1 encounter with possible diagnostic error. Out of 663 patients in the possible diagnostic error group, 584 had an encounter 0 to 15 days before the date of index diagnosis and 78 had an encounter 16 to 30 days before index diagnosis (encounter date missing on 1 patient). Among patients with a preceding encounter with possible diagnostic error, 224 (33.8%) had an ischemic stroke reported between the initial encounter with possible diagnostic error and the CeAD index diagnosis encounter and 11 (1.7%) had an additional ischemic stroke within 180 days after index diagnosis. The median time from any symptom onset to diagnosis was 1 day for the entire cohort (interquartile range, 0–6) and 75% of patients were diagnosed within 6 days from symptom onset (Figure S1). The median time from symptom onset to diagnosis was significantly higher in patients with possible diagnostic error (7 days [interquartile range, 3–14]) compared with patients without possible diagnostic error (1 day [interquartile range, 0–4], P<0.001).
Factors Associated With Possible Diagnostic Error
Table 1 depicts the univariate comparisons of patients with versus without possible diagnostic error. After adjustment, younger age (adjusted odds ratio [aOR], 0.89 with a 10‐year increase [95% CI, 0.83–0.95], P<0.001), history of migraine headache (aOR, 1.35 [95% CI, 1.09–1.67], P=0.007), presentation with headache (aOR, 1.43 [95% CI, 1.19–1.71], P<0.001), and absence of focal neurologic signs or symptoms (absence of focal deficit, Horner's or cranial nerve palsy) (aOR, 0.68 [95% CI, 0.56–0.83], P < 0.001) were independently associated with possible diagnostic error (Table 2).
Table 1.
Univariate Analysis of Characteristics Between Patients With and Without Possible Diagnostic Error
| With possible diagnostic error | Without possible diagnostic error | P value | |
|---|---|---|---|
| Mean age, y (IQR) | 44 (36–52) | 47 (38–56) | <0.001* |
| Female sex, n (%) | 338/663 (51.0%) | 1450/3349 (43.3%) | <0.001* |
| Race† | 0.03* | ||
| White, n (%) | 468/663 (70.6%) | 2477/3349 (74.0%) | |
| Black, n (%) | 42/663 (6.3%) | 200/3349 (6.0%) | |
| Asian, n (%) | 18/663 (2.7%) | 130/3349 (3.9%) | |
| Other/unknown, n (%) | 135/663 (20.4%) | 542/3349 (16.2%) | |
| Hispanic ethnicity, n (%) | 35/349 (10.0%) | 138/1758 (7.8%) | 0.18 |
| Medical history | |||
| Hypertension, n (%) | 221/663 (33.3%) | 1199/3349 (35.8%) | 0.23 |
| Diabetes, n (%) | 54/663 (8.1%) | 286/3349 (8.5%) | 0.74 |
| Hyperlipidemia, n (%) | 132/663 (19.9%) | 787/3349 (23.5%) | 0.05 |
| Migraine, n (%) | 151/663 (22.8%) | 515/3349 (15.4%) | <0.001* |
| Active smoking, n (%) | 115/663 (17.3%) | 667/3349 (19.9%) | 0.13 |
| Recent COVID infection, n (%) | 5/663 (0.8%) | 44/3349 (1.3%) | 0.33 |
| Recent upper respiratory infection, n (%) | 50/663 (7.5%) | 201/3349 (6.0%) | 0.14 |
| Connective tissue disease, n (%) | 15/663 (2.3%) | 67/3349 (2.0%) | 0.66 |
| Neck trauma, n (%) | 154/663 (23.2%) | 734/3348 (21.9%) | 0.46 |
| Pregnancy, n (%) | 1/663 (0.2%) | 8/3349 (0.2%) | 1 |
| Postpartum, n (%) | 17/663 (2.6%) | 36/3349 (1.1%) | <0.01* |
| Selective serotonin reuptake inhibitor SSRI use, n (%) | 75/663 (11.3%) | 289/3349 (8.6%) | 0.03* |
| Presenting symptoms/signs | |||
| National Institutes of Health Stroke Scale score on admission, median, IQR | 0 (0–3) | 1 (0–7) | <0.001 |
| Headache, n (%)‡ | 390/663 (58.8%) | 1528/3349 (45.6%) | <0.001* |
| Neck pain, n (%)‡ | 242/663 (36.5%) | 976/3349 (29.1%) | <0.001* |
| Tinnitus, n (%)‡ | 20/663 (3.0%) | 78/3349 (2.3%) | 0.30 |
| Presenting with focal deficit, n (%) | 458/663 (69.1%) | 2625/3347 (78.4%) | <0.001* |
| Imaging findings | |||
| Vessel location | 0.05 | ||
| Single vertebral artery dissection, n (%) | 313/663 (47.2%) | 1694/3349 (50.6%) | |
| Single carotid artery dissection, n (%) | 255/663 (38.5%) | 1294/3349 (38.6%) | |
| Single intracranial dissection, n (%) | 2/663 (0.3%) | 12/3349 (0.4%) | |
| Multivessel artery dissection, n (%) | 93/663 (14.0%) | 348/3349 (10.4%) | |
| Acute infarct on imaging, n (%) | 357/663 (53.8%) | 1972/3349 (58.9%) | 0.02* |
| Occlusive dissection, n (%) | 206/656 (31.4%) | 1241/3315 (37.4%) | <0.01* |
| Partially occlusive thrombus, n (%) | 56/663 (8.4%) | 353/3349 (10.5%) | 0.10 |
| Intracranial extension, n (%) | 64/626 (10.2%) | 275/3190 (8.6%) | 0.20 |
| Pseudoaneurysm, n (%) | 113/663 (17.0%) | 433/3349 (12.9%) | 0.01* |
| Acute treatment | |||
| Intravenous thrombolysis, n (%) | 35/663 (5.3%) | 601/3347 (18.0%) | <0.001* |
| Endovascular treatment, n (%) | 86/663 (13.0%) | 773/3349 (23.1%) | <0.001* |
P values <0.05 were considered statistically significant.
Only US sites.
Variables in possible diagnostic error group are from the same encounter and in group without possible diagnostic error—from index hospitalization encounter for cervical artery dissection.
Table 2.
Multivariable Logistic Regression Evaluating the Relationship Between Patient Characteristics and Possible Diagnostic Error of CeAD. Variables included age, female sex, race, hyperlipidemia, migraine, postpartum state, selective serotonin reuptake inhibitor use, presentation with headache, presentation with neck pain, presentation with focal deficits, dissection vessel location, acute infarct on imaging, occlusive dissection, pseudoaneurysm
| Misdiagnosis | No.=663 |
|---|---|
| Age | OR, 0.89 [95% CI, 0.83–0.95], P<0.001 |
| History of migraines | OR, 1.35 [95% CI, 1.09–1.67], P=0.01 |
| Headache on presentation | OR, 1.43 [95% CI, 1.19–1.71], P<0.001 |
| Focal deficit on presentation | OR, 0.68 [95% CI, 0.56–0.83], P<0.001 |
Association Between Possible Diagnostic Error and Outcomes
In adjusted analyses, there was no significant difference in the 180‐day ischemic stroke incidence (aOR, 0.96 [95% CI, 0.63–1.48], P=0.86), excellent 90‐day functional outcome (aOR, 1.04 [95% CI, 0.76–1.44], P=0.80), and death at 90 days (aOR, 0.58 [95% CI, 0.19–1.8], P=0.34) among patients with CeAD with and without possible diagnostic error (Table 3).
Table 3.
Unadjusted and Adjusted Logistic Regression to Assess Association of Possible Diagnostic Error With Ischemic Stroke at 180 Days, Excellent mRS Score at 90 Days, and Death
| Outcome | Sample size (N) | Unadjusted OR (95% CI) | P value | Adjusted OR (95% CI) | P value |
|---|---|---|---|---|---|
| Ischemic stroke at 180 d | 4012/3792 | 0.86 (0.56–1.33) | 0.49 | 0.96 (0.63–1.48) * | 0.86 |
| Excellent mRS at 90 d | 3221 | 1.51 (1.13–2.01) | 0.005 | 1.04 (0.76–1.44)† | 0.797 |
| Death | 4012/3792 | 0.28 (0.09–0.83) | 0.021 | 0.58 (0.19–1.80)‡ | 0.34 |
mRS indicates modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; and OR, odds ratio.
Adjusted for age, female sex, partially occlusive thrombus, occlusive dissection, multivessel dissection, and intracranial extension of dissection.
Adjusted for age, female sex, and NIHSS score.
Adjusted for age, female sex, partially occlusive thrombus, occlusive dissection, multivessel dissection, intracranial extension of dissection, and NIHSS score.
DISCUSSION
In this large, multicenter, international STOP‐CAD substudy, 1 in 6 patients with CeAD experienced possible diagnostic error, defined as a prior health care encounter attributable to CeAD within 30 days before their index diagnosis. Younger age, history of migraine headaches, presence of headache, and clinical presentations without focal neurologic deficits were associated with a higher likelihood of possible diagnostic error. Importantly, we did not find a significant difference in 180‐day ischemic stroke risk, excellent functional outcome, or death among patients with CeAD with and without preceding possible diagnostic error.
Our rate of possible diagnostic error of CeAD was greater than in a prior study using administrative claims data from a large heterogeneous cohort of patients from 2 states, which reported a CeAD misdiagnosis (type of possible diagnostic error) in 1 in 30 patients in ED settings. 3 Our higher rate of possible diagnostic error may be due to a longer time interval (30 days in our study versus 14 days in the prior study) and the inclusion of encounters in a variety of non‐ED outpatient settings. Additionally, our use of detailed clinical data through chart reviews may have allowed for the inclusion of more comprehensive neurological symptoms, including nonfocal presentations that are known to be diagnostically challenging, potentially leading to the identification of more encounters with possible diagnostic error.
Our study identified several risk factors associated with possible diagnostic error of CeAD. We found that younger patients had a higher likelihood of having possible diagnostic error, which is consistent with prior studies examining possible diagnostic error of stroke. 3 , 6 , 9 In fact, one study showed that younger patients under the age of 44 years were 7 times more likely to have a probable stroke misdiagnosis compared with patients older than 75 years of age. 6 A recent study on misdiagnosis of CeAD using administrative claims data showed that the mean age among patients with probable misdiagnosis in the ED setting was 43.3 years compared with 53 years in patients without probable misdiagnosis. 3
Patients with a history of migraine headaches or presenting with headache were more likely to have possible diagnostic error, which is similar to what is reported in the stroke diagnostic error literature. Migraine is a known stroke chameleon, often incorrectly diagnosed in cases of missed stroke. 6 , 10 , 11 Diagnosing cerebrovascular conditions in patients who present with a history of migraines, headache, and otherwise minimal neurological deficits is challenging. 12 , 13 Headache is a common presenting symptom in patients with CeAD, it often precedes focal neurologic deficit, and at times it is the only clinical presentation along with neck pain. 14 , 15 There are no specific features that help distinguish dissection‐associated headache from other causes of headaches. Although inclusion of headache predominant presentations may overestimate true diagnostic error, excluding such presentations would risk underestimating delayed in recognition, particularly among nonischemic presentation. Thus, our findings should be interpreted as reflecting challenges in early recognition rather than inappropriate care at the initial encounter. Clinicians need to use caution when evaluating patients with headache and keep CeAD in the differential to initiate timely and appropriate workup. The goal of highlighting headache‐predominant presentations is not to suggest that all patients with headache or neck pain should undergo vascular imaging. Given the high prevalence of benign headache disorders, indiscriminate imaging could lead to substantial overtesting and false‐positive evaluations. Rather, our findings emphasize maintaining clinical awareness of CeAD in patients with new, atypical, or severe head or neck pain, particularly in younger individuals. Careful clinical judgment is required to balance timely recognition of CeAD with appropriate and judicious use of diagnostic resources.
Additionally, patients with nonischemic clinical presentation were more likely to experience possible diagnostic error, which is also consistent with prior stroke literature. Stroke protocols generally capture focal neurological symptoms, but not isolated headache or neck pain, which are often the initial symptoms in CeAD and tend to precede focal ischemic symptoms. 16 CeAD diagnosis should be considered in patients who present with new or worsening headache or neck pain. 7 Moreover, a quarter of patients with carotid dissection have an associated partial Horner syndrome, which if present with new or worsening headache could suggest carotid dissection and help with timely diagnosis. 1 , 17
In addition, we found that patients without possible diagnostic error were more likely to receive intravenous thrombolysis and endovascular therapy during index hospitalization compared with patients who experienced possible diagnostic error. Although the exact reason for this is unclear, it is possible that patients without possible diagnostic error had more typical acute stroke presentations, facilitating rapid recognition and access to acute stroke care, which contributed to be absence of possible diagnostic error.
Importantly, one third of patients with possible diagnostic error were found to have a stroke between the possible diagnostic error encounter and before the index diagnosis. These events occurred before confirmed CeAD diagnosis, were not adjudicated, and may not have been preventable. Therefore, they should be interpreted as descriptive in nature rather than as evidence of harm from diagnostic delay. Nevertheless, this finding underscores the close time proximity between symptoms onset and ischemic events and highlights the importance of timely recognition and treatment of CeAD, to reduce the risk of future ischemic stroke. 16
Lastly, we did not find any difference in outcomes among CeAD patients with and without possible diagnostic error, which is keeping in line with the previous literature. 3 We may be overestimating the number of possible diagnostic errors, as some encounters were likely not related to subsequent hospitalization for CeAD. Alternatively, we could be underestimating the rate of diagnostic errors, as some patients likely did not return for evaluation, sought care at different medical centers, or died before an accurate diagnosis could be made. It is also possible that patients who experience diagnostic errors are low‐risk individuals, whose outcomes remain unchanged despite the diagnostic error. Our findings that patients with possible diagnostic error were less likely to present with focal deficits or have acute infarcts or occlusive dissection on imaging, compared with patients without possible diagnostic error, might support this hypothesis. However, it remains unclear whether timely and accurate diagnosis of patients who initially experience a diagnostic error leads to improved outcomes or reduces harm.
Our study has several important limitations. An important limitation of this study is the inability to determine whether prior health care encounters truly represented diagnostic error, as many presentations included nonspecific symptoms, for which immediate vascular imaging may not have been indicated. Thus, our definition of possible diagnostic error may overestimate true misdiagnosis while at the same time, undercapture external encounters. Thus, the findings should be interpreted as identifying presentations associated with delayed recognition rather than definitive diagnostic error. Not all encounters within 30 days before CeAD diagnosis may be related to subsequent CeAD detection. Although some may represent instances of severely delayed or missed diagnosis of CeAD, others could have had symptoms similar but unrelated to CeAD. Differentiating recurrent migraine symptoms from musculoskeletal neck pain or symptoms that would prompt further CeAD workup is important in the allocation of imaging resources in the emergency department or outpatient settings. Also, strokes reported to have occurred between the time interval of possible diagnostic error and final diagnosis were not adjudicated or confirmed, compared with stroke outcomes that occurred after the index correct diagnosis, which were all confirmed by site investigators. These events should be considered descriptive in nature and not interpreted as evidence of harm attributable to diagnostic delay. In addition, our study is retrospective and prone to ascertainment bias. Lastly, diagnosis of CeAD was identified using ICD 10 codes and not prospectively, and detailed data with regard to a possible diagnostic error encounter such as hospital type (teaching versus nonteaching) and clinician level or experience (years in practice) were not collected. However, studying diagnostic error would not be feasible prospectively, and retrospective data sets are the only way to study root causes of diagnostic error. On the other hand, retrospective data sets may be limited by documentation, and prior medical encounters might not be systematically documented.
CONCLUSIONS
One in 6 patients eventually diagnosed with CeAD experienced a prior health care encounter for symptoms attributable to CeAD, with one ‐third having an ischemic stroke after the possible diagnostic error encounter. This was especially common in younger patients with a history of migraines and a headache or nonfocal clinical symptoms.
Sources of Funding
None.
Disclosures
None.
Supporting information
Figure S1
STROBE Checklist
Acknowledgments
Author contributions: Ekaterina Bakradze, Shadi Yaghi, and Setareh Salehi Omran contributed to conception and design of the study, acquisition and analysis of data, and drafting a significant portion of the article and tables. Liqi Shu, Thanh Lam Ngoc Tran, Ross Curtiss Crandal, Evan Lester, James Ernest Siegler, Mary K. Penckofer, Farhan Khan, Wayneho Kam, Brian MacGrory, Faddi G. Saleh Velez, Cheran Elangovan, Balaji Krishnaiah, Preet Gandhi, Reza Bavarsad Shahripour, Malik Gannam, Mirjam Rachel Heldner, Kateryna Antonenko, Aaron Rothstein, Michele Romoli, Marialuisa Zedde, Rosario Pascarella, Piers Klein, Farid Khasiyev, Ossama Khazaal, Adeel S Zubair, Marwa Elnazeir, João Pedro Marto, and James Anthony Giles EAM contributed to conception and design of the study and acquisition of data. Daniel M. Mandel, Ava Leigh Liberman, Thalia S. Field, Nils Henninger, Ronen R. Leker, Issa Metanis, Edgar Samaniego, Thanh N. Nguyen, Mohammad Almajali, Ahmad Nehme, Stefan T. Engelter, and Christoph Stretz contributed to acquisition of data and drafting a significant portion of the article and tables.
This article was sent to Fadar Oliver Otite, MD, SM, Associate Editor, for review by expert referees, editorial decision, and final disposition.
Supplemental Material is available at https://www.ahajournals.org/doi/suppl/10.1161/JAHA.125.046408
For Sources of Funding and Disclosures, see page 7.
References
- 1. Yaghi S, Engelter S, Del Brutto VJ, Field TS, Jadhav AP, Kicielinski K, Madsen TE, Mistry EA, Salehi Omran S, Pandey A, et al. Treatment and outcomes of cervical artery dissection in adults: a scientific statement from the American Heart Association. Stroke. 2024;55:e91–e106. doi: 10.1161/STR.0000000000000457 [DOI] [PubMed] [Google Scholar]
- 2. Griffin KJ, Harmsen WS, Mandrekar J, Brown RD Jr, Keser Z. Epidemiology of spontaneous cervical artery dissection: population‐based study. Stroke. 2024;55:670–677. doi: 10.1161/STROKEAHA.123.043647 [DOI] [PubMed] [Google Scholar]
- 3. Liberman AL, Navi BB, Esenwa CC, Zhang C, Song J, Cheng NT, Labovitz DL, Kamel H, Merkler AE. Misdiagnosis of cervicocephalic artery dissection in the emergency department. Stroke. 2020;51:1876–1878. doi: 10.1161/STROKEAHA.120.029390 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Medicine Io, National Academies of Sciences E, Medicine . Improving Diagnosis in Health Care. Washington, DC: The National Academies Press; 2015. [Google Scholar]
- 5. Yaghi S, Shu L, Mandel D, Leon Guerrero CR, Henninger N, Muppa J, Affan M, Ul Haq Lodhi O, Heldner MR, Antonenko K, et al. Antithrombotic treatment for stroke prevention in cervical artery dissection: the STOP‐CAD study. Stroke. 2024;55:908–918. doi: 10.1161/STROKEAHA.123.045731 [DOI] [PubMed] [Google Scholar]
- 6. Newman‐Toker DE, Moy E, Valente E, Coffey R, Hines AL. Missed diagnosis of stroke in the emergency department: a cross‐sectional analysis of a large population‐based sample. Diagnosis (Berl). 2014;1:155–166. doi: 10.1515/dx-2013-0038 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Dhamoon MS, Sciacca RR, Rundek T, Sacco RL, Elkind MS. Recurrent stroke and cardiac risks after first ischemic stroke: the Northern Manhattan Study. Neurology. 2006;66:641–646. doi: 10.1212/01.wnl.0000201253.93811.f6 [DOI] [PubMed] [Google Scholar]
- 8. Omran SS, Zhang C, Seitz A, Bruce SS, Liao V, Pawar A, Navi BB, Kamel H, Liberman AL. Possible misdiagnosis of pregnancy‐associated stroke in the emergency department. J Stroke Cerebrovasc Dis. 2024;34:108139. doi: 10.1016/j.jstrokecerebrovasdis.2024.108139 [DOI] [PubMed] [Google Scholar]
- 9. Richoz B, Hugli O, Dami F, Carron PN, Faouzi M, Michel P. Acute stroke chameleons in a university hospital: risk factors, circumstances, and outcomes. Neurology. 2015;85:505–511. doi: 10.1212/WNL.0000000000001830 [DOI] [PubMed] [Google Scholar]
- 10. Madsen TE, Khoury J, Cadena R, Adeoye O, Alwell KA, Moomaw CJ, McDonough E, Flaherty ML, Ferioli S, Woo D, et al. Potentially missed diagnosis of ischemic stroke in the emergency department in the Greater Cincinnati/northern Kentucky stroke study. Acad Emerg Med. 2016;23:1128–1135. doi: 10.1111/acem.13029 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Otlivanchik O, Liberman AL. Migraine as a stroke mimic and as a stroke chameleon. Curr Pain Headache Rep. 2019;23:63. doi: 10.1007/s11916-019-0801-1 [DOI] [PubMed] [Google Scholar]
- 12. Liberman AL, Lu J, Wang C, Cheng NT, Moncrieffe K, Lipton RB. Factors associated with hospitalization for ischemic stroke and TIA following an emergency department headache visit. Am J Emerg Med. 2021;46:503–507. doi: 10.1016/j.ajem.2020.10.082 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Dubosh NM, Edlow JA, Goto T, Camargo CA Jr, Hasegawa K. Missed serious neurologic conditions in emergency department patients discharged with nonspecific diagnoses of headache or back pain. Ann Emerg Med. 2019;74:549–561. doi: 10.1016/j.annemergmed.2019.01.020 [DOI] [PubMed] [Google Scholar]
- 14. Kaufmann JE, Harshfield EL, Gensicke H, Wegener S, Michel P, Kagi G, Nedeltchev K, Kellert L, Rosenbaum S, Nolte CH, et al. Antithrombotic treatment for cervical artery dissection: a systematic review and individual patient data meta‐analysis. JAMA Neurol. 2024;81:630–637. doi: 10.1001/jamaneurol.2024.1141 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Arnold M, Cumurciuc R, Stapf C, Favrole P, Berthet K, Bousser MG. Pain as the only symptom of cervical artery dissection. J Neurol Neurosurg Psychiatry. 2006;77:1021–1024. doi: 10.1136/jnnp.2006.094359 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Gottesman RF, Sharma P, Robinson KA, Arnan M, Tsui M, Ladha K, Newman‐Toker DE. Clinical characteristics of symptomatic vertebral artery dissection: a systematic review. Neurologist. 2012;18:245–254. doi: 10.1097/NRL.0b013e31826754e1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Lee VH, Brown RD Jr, Mandrekar JN, Mokri B. Incidence and outcome of cervical artery dissection: a population‐based study. Neurology. 2006;67:1809–1812. doi: 10.1212/01.wnl.0000244486.30455.71 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1
STROBE Checklist