Article Text

Original research
Carotid web: an under-recognized and misdiagnosed ischemic stroke etiology
  1. Ehizele M Osehobo1,2,
  2. Raul G Nogueira1,2,
  3. Sitara Koneru1,2,
  4. Alhamza R Al-Bayati1,2,
  5. Catarina Perry de Camara1,
  6. Fadi Nahab1,
  7. Bernardo Liberato1,2,
  8. Michael R Frankel1,2,
  9. Jason W Allen1,3,
  10. Charlie Chulhyun Park3,
  11. Diogo C Haussen1,2
  1. 1 Department of Neurology, Emory University School of Medicine, Atlanta, Georgia, USA
  2. 2 Marcus Stroke and Neuroscience Center, Grady Health System, Atlanta, GA, USA
  3. 3 Department of Radiology, Emory University School of Medicine, Atlanta, GA, USA
  1. Correspondence to Dr Diogo C Haussen, Neurology, Neurosurgery and Radiology, Emory University School of Medicine / Marcus Stroke & Neuroscience Center - Grady Memorial Hospital, Atlanta, GA 30303-3073, USA; diogo.haussen{at}


Background Carotid web (CaW) constitutes a possible cause of ischemic stroke, particularly large vessel occlusion syndromes. We aim to evaluate misdiagnosis rates and diagnosis trends for CaW.

Methods Based on CT angiography (CTA), we prospectively identified a cohort of patients with symptomatic CaW treated at two comprehensive stroke centers (CSC) from 2014 to 2020 to assess misdiagnosis. Official CTA reports from the CSCs and referring hospitals were then reviewed for mention of CaW. For diagnosis trends, we retrospectively analyzed a CSC electronic medical record, identifying patients with CaW mentioned in an official CTA report from 2011 to 2020.

Results For misdiagnosis, 56 patients with symptomatic CaW were identified in the CSCs; 16 (28%) had bilateral CaW, totaling 72 CaWs. Only one CaW (5.5%) was reported at referring facilities, from 14 patients/18 CaWs imaged with CTA. Conversely, 43 (69%) CaWs were reported from 49 patients/62 CaWs at the CSC (p<0.01). For diagnosis trends, from 2011 to 2020, 242 patients at a CSC accounted for 266 CTA reports mentioning CaW. The majority of these reports (n=206, 77%) were associated with stroke/transient ischemic attack (TIA) ICD-9/ICD-10 codes. The rate of CaW diagnosis adjusted per 1000 patients with stroke/TIA increased over time, 2015 being the most significant point of change ('joinpoint'; p=0.01). The analysis of CaW mentions normalized per 1000 CTA reports also showed increasing rates of diagnosis over time (joinpoint:2014; p<0.02).

Conclusion CaW was predominantly identified in patients with strokes/TIAs rather than asymptomatic patients. CaW was commonly overlooked in facilities with lower levels of cerebrovascular certification. Recognition of CaW at a CSC has significantly increased over time, independent of overall imaging and stroke patient volume.

  • stroke
  • vessel wall
  • stent
  • CT angiography
  • artery

Data availability statement

The data that supports the findings of this study are available from the corresponding author upon reasonable request.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Carotid web (CaW) is a variant of fibromuscular dysplasia in which a non-atheromatous, non-inflammatory distinct shelf-like focal area of fibrous dysplasia occurs at the posterior aspect of the carotid bulb.1 Its thromboembolic potential has been reported in younger stroke populations with otherwise unremarkable vascular risk factors.2 3 CaW may be considered as a possible etiology for stroke patients with ipsilateral stroke who fit embolic stroke of undetermined source criteria.2

Although CaW has often been considered to be overlooked or misdiagnosed during stroke workup, data is lacking to corroborate this suggestion. Misdiagnosis could be a plausible explanation for CaW to be perceived as a rare lesion, since studies indicate that it may explain 0.5% of all ischemic strokes and 2.5% of patients treated with large vessel occlusion, and it is found in 9–39% of young patients with cryptogenic stroke.2–5

We aim to (1) evaluate the rate of CaW recognition in referring hospitals compared with comprehensive stroke centers (CSCs) and (2) evaluate the temporal trends in the diagnosis of CaW at a local CSC.


Misdiagnosis at CSCs and referring hospitals

The study period spanned 2014–2020 and included symptomatic CaW, defined as a shelf-like linear filling defect in the posterior aspect of the carotid bulb on CT angiography (CTA) in patients with an ischemic event (transient ischemic attack (TIA) or acute ischemic stroke (AIS)) involving the vascular territory of the CaW-affected internal carotid artery. Etiology of the ischemic event was undetermined, as defined by Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria.6 Patients were prospectively identified at two CSCs during the initial evaluation by the stroke and/or interventional teams. Minimum stroke workup included non-contrast brain CT, MRI brain, vessel imaging with head and neck CTA, coagulation tests, inpatient telemetry (for at least 20 hours), ECG, and transthoracic and/or transesophageal echocardiography in accordance with AHA Stroke Council and Stroke Association guidelines.7 All CTA vascular images were evaluated independently by a fellowship-trained neurointerventionist and a neuroradiologist blinded to the laterality and details of the clinical event. Thin-cut axial images and maximum intensity projections (MIPs) in sagittal and coronal planes were systematically reviewed. Discrepancies were resolved with a consensus read. Patients with a superimposed thrombus associated with CaW were only included if they had a repeat CTA demonstrating an underlying CaW after resolution of the clot. Patients with two or more possible stroke etiologies (ie, cardioembolism with confirmed atrial fibrillation and an ipsilateral CaW) were excluded, so all patients in the database had AIS or TIA presumed to be caused by CaW.

Official CTA reports from the CSCs and from referring hospitals (not CSC certified by Joint Commission standards) were retrospectively reviewed for any mention of CaW, including the search terms “carotid web”, “carotid shelf”, “web”, “shelf”, “web-like”, and the plural forms of these terms. The rate of accurate CaW diagnosis via CTA was compared between the CSCs together and the referring hospitals in aggregate. The reports of digital subtraction angiography (DSA), magnetic resonance angiography with contrast (MRA), and carotid duplex ultrasound performed at the CSCs were also evaluated.

Trends in CaW diagnosis at CSC

Patients with CaW as part of their differential imaging diagnosis were identified by searching a CSC electronic medical record database (EPIC Systems Corporation, Madison, Wisconsin, USA) for head and neck CTA reports from 2011 to 2020. The aforementioned search terms for description of CaW were used to create a dataset, which was manually reviewed to ensure that the terms were appropriately linked to lesions involving the carotid vasculature. The number of CaW mentions was subsequently annualized; results reported from the year 2020 were extrapolated from the data available in January to March of that year.

The number of CaW mentions with an associated stroke or TIA diagnosis code per ICD-9/ICD-10 was standardized based on the annual number of strokes and TIAs diagnosed at the CSC. The number of CaW mentions was normalized based on the annual number of head and neck CTAs performed at the CSC. These normalizations were incorporated to limit the influence of an annual increase in patient volume on the final analyses.

Statistical analysis

Categorical variables were reported as proportions and compared by χ2 or Fisher exact test as appropriate using IBM SPSS Statistics Version 26 (IBM Corp, Armonk, New York, USA). Statistical analysis of trend change was performed with Joinpoint Trend Analysis Software, Version, April 2020 (Statistical Research and Applications Branch, National Institutes of Health, Bethesda, Maryland, USA). Joinpoint regression considers multiple permutations in a given timeline in order to model one or more joinpoints, each of which demarcates two adjacent time segments with a statistically significant change in trend.8 The analysis tests a null hypothesis in which the entire timeline can be modeled by one simple linear or logarithmic regression model against an alternative hypothesis in which the presence of one or more joinpoints suggests a piecewise or segmented regression model for best fit of the data. At each joinpoint, the change in trend was described as 'accelerated' if the earlier time segment had a slope coefficient (β) that was either significantly negative or insignificant, and if the later time segment had a statistically significant positive β. With only 10 years available for analysis, the maximum number of joinpoints was set at 1, in accordance with best practice recommendations by the National Institutes of Health.9 Years with zero (0) CaW mentions (2012–2014) were adjusted to have one mention to allow for ease of statistical computation. Significance was set at p<0.05.


Misdiagnosis at CSCs and referring hospitals

Of the 56 patients who were confirmed to have symptomatic CaW, 16 patients had bilateral webs, accounting for a total of 72 webs imaged with CTA. Fourteen of these patients were initially scanned with CTA at a referring hospital prior to transfer to the CSC. Four had bilateral webs, leading to an aggregate of 18 webs imaged at the outside facility. Review of the referring hospital CTA official reports for mentions of these 18 webs found only one (5.5%) diagnosis of CaW.

Conversely, 49 patients were either only scanned at the CSC or re-scanned at the CSC following transfer from a referring hospital. Thirteen of these patients had bilateral webs, leading to a total of 62 webs imaged at the CSCs. Of these webs, 43/62 (69.3%) were accurately diagnosed as CaW in the official CTA read, a significantly higher rate compared with the referring hospital official reads (p<0.0001, table 1). In the CSCs and referring hospitals, misdiagnosed CaWs were either read as normal or mischaracterized as another entity, such as atherosclerosis or dissection. Diagnostic performance of other imaging modalities, including DSA, MRA with contrast, and ultrasound are shown in table 1.

Table 1

Diagnostic performance of various imaging modalities for the diagnosis of carotid web

Trends in CaW recognition at CSC

In the CSC electronic medical records, 242 patients met the inclusion criteria for analysis, accounting for 266 CTA reports with mention of CaW. Of the 242 included patients, 183 (75.6%) had an ICD-9 or ICD-10 diagnosis code associated with stroke or TIA, accounting for 206 CTA reports with mention of CaW. In this cohort, the number of CTA reports with mention of CaW was first annualized and then normalized per 1000 patients diagnosed with stroke per year at the CSC (figure 1A). The annual rate of CaW diagnosis per 1000 stroke/TIA patients increased over time, with the year 2015 being the most significant point of change (joinpoint). The 2015–2020 period was found to have a β of 8.453, a significant change from the β of 0.508 in the prior period of 2011–2015 (p=0.01, figure 1B). From 2015 to 2020, the annual rate of patients with CaW mentions per 1000 diagnosed strokes/TIAs at the CSC ranged from 3 to 43; this number excludes mentions from repeat or follow-up scans.

Figure 1

(A) Annual volume of patients with carotid web mentions and stroke/TIA patients diagnosed. (B) Annual rate of carotid web diagnosis adjusted per stroke/TIA volume. Data from January 1, 2020 to March 3, 2020 were extrapolated to model the entire year. *Statistically significant change in slope (joinpoint) identified at 2015, at alpha level <0.05. CaW, carotid web; TIA, transient ischemic attack.

The annual number of CTAs with CaW reported was subsequently normalized per 1000 head and neck CTAs performed per year (figure 2A). The analysis of CaW mentions normalized per 1000 CTAs also demonstrated increasing rates of diagnosis over time. The year 2014 was determined to be the significant point of change in trend (joinpoint) from the observed time period. Regression modeling estimated a significant slope coefficient (β) of 5.071 from 2014 to 2020, a notable increase from the 2011–2014 β of −0.135 (p<0.02, figure 2B). From 2014 to 2020, the annual rate varied from 1 to 29 patients with CaW mentions per 1000 CTAs performed at the CSC, excluding repeat or follow-up scans.

Figure 2

(A) Annual volume of patients with carotid web mentions and head and neck CTAs performed. (B) Annual rate of carotid web diagnosis adjusted per CTA volume. Data from January 1, 2020 to March 3, 2020 were extrapolated to model the entire year. *Statistically significant change in slope (joinpoint) identified at 2014, at alpha level <0.05. CaW, carotid web; CTA, CT angiography.

Sensitivity analysis was performed excluding all repeat CTAs. The raw data for both stroke-associated and diagnosis-independent CaW mentions were reviewed and any repeat scans of patients were removed. Joinpoint regression on the reviewed datasets reproduced the same joinpoints as the full datasets: 2015 with p=0.037 for CaW adjusted for stroke/TIA volume, and 2014 with p=0.033 for CaW adjusted for CTA volume (online supplemental material, figures I and II, available online).

The average annual percent change (aAPC) in the number of head and neck CTAs with CaW mentions from 2014 to 2020 was 137% per year, while the aAPC of the total number of head and neck CTAs performed at the hospital was only 13.8% within the same period. Similarly, in the stroke/TIA cohort, the aAPC of CTAs with CaW mentions from 2015 to 2020 was 96.6% per year, while the aAPC of stroke and TIA volume in the CSC spanning that same time period was only 20.3% per year. For both adjusted analyses, the data indicate that the main reason for the growth in CTA diagnosis of CaW is explained by reasons other than the overall increase in stroke patient and/or vascular imaging volumes. The yearly percentage of all patients with stroke/TIA who had an associated CaW mention in the official report ranged is shown in table 2 and ranged from 0% in the first 3 years of the study period to 4.3% in 2019.

Table 2

Annualized carotid web, stroke, and TIA volume at a comprehensive stroke center


This study shows novel findings related to CaW and stroke. The data support the suggestion that CaW is an overlooked condition, especially in hospitals with lower levels of stroke certification. We also observed that CaW was reported more commonly in patients with stroke or TIA compared with asymptomatic patients, and that the rate of CaW diagnosis at the studied CSC has increased over time, independent of patient volume. We posit that an increase in education and awareness surrounding CaW led to the rise in diagnostic performance at the CSC.

Although fibromuscular dysplasia was first described over 80 years ago, there is surprisingly limited awareness about this entity within the medical community.10 11 We have shown that misdiagnosis is not an uncommon occurrence, and this may be due to the fact that the knowledge related to CaW is even more restricted.12 13 It is possible that the lack of a consensus in naming has contributed to limiting awareness in the general medical community. In 1968, Rainer et al described a CaW as “a discrete filling defect at the internal carotid origin”.14 Momose et al described it as a “web formation” in 1973,15 and Lipchik et al applied the term “diaphragm” a year later.16 In 1995, Gironell et al characterized CaW as a “pseudovalvular fold”.17 Many other terms have been used to describe what is now more commonly referred to as CaW, including “weblike septum”, “thrombotic carotid megabulb”, and “spur”.13

Limited recognition of CaW may also be attributable to the fact that it largely affects younger otherwise healthy individuals without traditional risk factors for stroke. However, it is not uncommon for patients with CaW to present with vascular comorbidities. A meta-analysis indicated that 23% of symptomatic CaW patients had hypertension, 16% had hyperlipidemia, 6% had diabetes, and 17% were smokers.1 Relating to these risk factors, differentiating atherosclerosis from CaW presents an additional diagnostic challenge. There may be radiological similarities between webs and atheromatous lesions.18 A recent report demonstrated the use of Optical Coherence Tomography to image a known CaW, and showed classic atherosclerotic changes superimposed to the expected endoluminal fibrotic protuberance.19 CaW can also be erroneously interpreted as a flap, leading to mischaracterization as a dissection.13 These diagnostic challenges help explain the lack of consensus on the treatment of symptomatic or asymptomatic CaW.20

The imaging method can also influence the rate of detection of CaW. DSA has been described to perform comparably to CTA, while ultrasonography was found to have a much more limited role in the diagnosis of CaW.21 Intravascular ultrasound may not have sufficient resolution to differentiate between CaW, thrombus, or atherosclerosis.22 Here, we corroborate the concept that DSA has a high yield and ultrasonography a low yield for diagnosis, and that contrasted MRA may have a limited role despite the small relative sample. These complexities may explain the observed finding that a higher level of care leads to higher chances of CaW diagnosis.

The prevalence of renal artery fibromuscular dysplasia in the general population has been estimated at 4.4%, based on a retrospective review of healthy kidney donors.23 The prevalence of CaW in the general population is unknown, but several studies have attempted to quantify the prevalence of CaW in AIS. A recent case–control study at a CSC found that 17.6% of patients with embolic stroke of undetermined source had an ipsilateral CaW, and that the overall frequency of CaW in patients with ischemic stroke was 0.5%.2 A retrospective analysis of the 2015 MR CLEAN study (Multicenter Randomized Clinical Trial of Endovascular Treatment for Acute Ischemic Stroke in the Netherlands) reported that 2.5% of patients with AIS and intracranial large vessel occlusion had CaW on the symptomatic side.3 In the present study, 183 patients with stroke were found to have mention of CaW in their CTA report over the 10-year study period. Following the joinpoint of 2015, the annual frequency of CaW in patients with stroke/TIA at the CSC ranged from 1.3% to 4.3%, peaking in 2019. It is possible that the recognition of CaW at the CSC may have reached a plateau (figure 1A), which could indicate that the performance of detection by CTA has been optimized.

The results of this study should be considered in the context of certain limitations. The retrospective study design has inherent deficiencies. There may be CaW mentions in the CTA reports that were recognized as a direct result of suggestion by vascular neurologists or neurointerventionists during the hyperacute patient evaluation prior to the final report, which could have influenced rates of CSC CaW diagnosis. In addition, the list of search terms included the most common CaW descriptions and was not all-encompassing; this may have influenced the report yield. Differences in CTA acquisition parameters between institutions potentially may also limit the detection of CaW, with non-CSC sites possibly using larger slice thickness than the CSC site. There have been changes in these CT scanners at the studied CSC over the reported timeframe, with the installation of more advanced scanners containing a larger number of detectors, which could theoretically influence the detection rate of CaW; however, the CT slice thickness acquired on all CTA at the CSC during the study period was unchanged at 0.625 mm, mitigating against this bias. Although the local institutional protocol is to obtain a CTA head and neck at the initial evaluation, the possibility that patients with cryptogenic strokes and other stroke subtypes had a different rate of CTA imaging is a theoretical source of selection bias. Finally, joinpoint analyses have typically been used for reporting trends in clinical outcomes, as opposed to tendencies in diagnosis.24


The majority of CaWs were identified in patients with stroke/TIAs, and less commonly in asymptomatic patients. CaW remains a commonly overlooked diagnosis in facilities with lower levels of cerebrovascular certification. Recognition of CaW at a CSC has significantly increased over time, independent of overall imaging and stroke patient volume trends.

Data availability statement

The data that supports the findings of this study are available from the corresponding author upon reasonable request.

Ethics statements

Patient consent for publication

Ethics approval

Authors obtained ethical approval to conduct this study from the Emory University Institutional Review Board (approval ID IRB00091421). This is a non-interventional study with statistical analysis of radiology reports, which were obtained from the electronic medical record and without direct contribution from the studied population. As such, the data obtained for evaluation did not require individual consent per the IRB approval.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.


  • Twitter @SitaraKoneru, @diogohaussen

  • Contributors EMO, RGN, and DCH conceived and planned the study. EMO, SK, ARA-B, BL, FN, CPC, and CCP contributed to data acquisition. EMO and DCH performed statistical calculations. EMO, DCH, BL, and JWA contributed to interpretation of the results. EMO and DCH took lead in writing the paper, with all other authors providing critical feedback that helped shape the final manuscript.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests The declared interests below are associations with commercial entities that could be viewed as having an interest in the general area of the submitted manuscript. None of these entities provided direct funding or sponsorship in support of this paper. DCH: Consultant for Stryker and Vesalio, Viz-AI (stock options). RGN: Principal Investigator, Stryker Neurovascular (DAWN trial (no compensation), Trevo-2 trial), Cerenovus/Neuravi (ENDOLOW trial, no compensation); consultant to Stryker Neurovascular; steering committee member, Stryker Neurovascular (no compensation), Medtronic (SWIFT trial, SWIFT Prime trial, no compensation), Cerenovus/Neuravi (ARISE-2 trial, no compensation); angiographic core lab, Medtronic (STAR trial); executive committee member, Penumbra (no compensation); Physician advisory board, Cerenovus/Neuravi, Phenox, Anaconda, Genentech, Biogen, Prolong Pharmaceuticals, Allm Inc (no compensation), Viz-AI (stock options).

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.