Skip main navigation

Antithrombotic Treatment for Stroke Prevention in Cervical Artery Dissection: The STOP-CAD Study

Originally publishedStroke. 2024;55:908–918

Abstract

BACKGROUND:

Small, randomized trials of patients with cervical artery dissection showed conflicting results regarding optimal stroke prevention strategies. We aimed to compare outcomes in patients with cervical artery dissection treated with antiplatelets versus anticoagulation.

METHODS:

This is a multicenter observational retrospective international study (16 countries, 63 sites) that included patients with cervical artery dissection without major trauma. The exposure was antithrombotic treatment type (anticoagulation versus antiplatelets), and outcomes were subsequent ischemic stroke and major hemorrhage (intracranial or extracranial hemorrhage). We used adjusted Cox regression with inverse probability of treatment weighting to determine associations between anticoagulation and study outcomes within 30 and 180 days. The main analysis used an as-treated crossover approach and only included outcomes occurring with the above treatments.

RESULTS:

The study included 3636 patients (402 [11.1%] received exclusively anticoagulation and 2453 [67.5%] received exclusively antiplatelets). By day 180, there were 162 new ischemic strokes (4.4%) and 28 major hemorrhages (0.8%); 87.0% of ischemic strokes occurred by day 30. In adjusted Cox regression with inverse probability of treatment weighting, compared with antiplatelet therapy, anticoagulation was associated with a nonsignificantly lower risk of subsequent ischemic stroke by day 30 (adjusted hazard ratio [HR], 0.71 [95% CI, 0.45–1.12]; P=0.145) and by day 180 (adjusted HR, 0.80 [95% CI, 0.28–2.24]; P=0.670). Anticoagulation therapy was not associated with a higher risk of major hemorrhage by day 30 (adjusted HR, 1.39 [95% CI, 0.35–5.45]; P=0.637) but was by day 180 (adjusted HR, 5.56 [95% CI, 1.53–20.13]; P=0.009). In interaction analyses, patients with occlusive dissection had significantly lower ischemic stroke risk with anticoagulation (adjusted HR, 0.40 [95% CI, 0.18–0.88]; Pinteraction=0.009).

CONCLUSIONS:

Our study does not rule out the benefit of anticoagulation in reducing ischemic stroke risk, particularly in patients with occlusive dissection. If anticoagulation is chosen, it seems reasonable to switch to antiplatelet therapy before 180 days to lower the risk of major bleeding. Large prospective studies are needed to validate our findings.

See related article, p 919

Cervical artery dissection (CAD) accounts for ≈2% of ischemic strokes overall1 but up to 25% in patients <50 years of age.2,3 CAD results from vessel wall injury and can lead to an intraluminal thrombus or steno-occlusion, increasing ischemic stroke risk. Thus, antithrombotic treatment is an important aspect of stroke prevention in patients with CAD.

Results from small, randomized trials4,5 comparing anticoagulation with antiplatelet therapy showed mixed results. In the context of studies that were underpowered to demonstrate the efficacy of one treatment strategy over another, clinical equipoise persists.6

Our aim was to compare the risks of ischemic stroke and major hemorrhage in individuals with CAD treated with antiplatelet versus anticoagulation in a large multicenter international retrospective study.

METHODS

Ethical Considerations, Data Sharing, and Reporting Guideline

Approval of the institutional review board at Lifespan was obtained (1894800-5) to perform the study. All study sites obtained ethical approval per local regulations. Aggregate data can be shared upon reasonable request to the corresponding author. The study follows the Strengthening the Reporting of Observational Studies in Epidemiology reporting guideline.7

Study Population and Design

This was a multicenter international retrospective observational study (63 sites from 16 countries; Figure S1) that included patients presenting to an acute care hospital and diagnosed with CAD without concomitant major trauma. We identified adult patients aged ≥18 years with CAD based on International Classification of Diseases, Ninth Revision codes (443.21 and 443.24),8,9International Classification of Diseases, Tenth Revision codes (I77.71, I77.74, and I77.75),10 or from institutional registries. These codes have been used or validated in prior studies.8–10

The patients’ vascular neuroimaging studies were reviewed by site principal investigators, and only those with clinical suspicion for CAD and imaging confirmation were included. Imaging confirmation required the presence of at least one of the following imaging features: crescent-shaped hyperintensity in the vessel wall indicating an intramural hematoma; a double lumen sign; the presence of a dissecting pseudoaneurysm, intimal flap, or vessel irregularity; or flame-shaped or tapering stenosis or occlusion of the artery at a typical dissection site and without evidence of atherosclerotic changes. Imaging reports, when available, were reviewed by neurologists at the lead site to confirm a dissection diagnosis.

We excluded patients with incidental chronic dissection, those with major head or neck trauma within the previous 4 weeks (eg, causing skull or cervical fractures or hemorrhage), those with a dissecting aneurysm causing primary subarachnoid hemorrhage, and those with iatrogenic dissection.

Primary Exposure

Exposures were defined based on treatment onset within 180 days. The primary exposure was the antithrombotic regimen used, divided into anticoagulation (parenteral anticoagulation with heparin or low-molecular-weight heparin or oral anticoagulation with either vitamin K antagonist [VKA] or direct oral anticoagulant [DOAC]) and antiplatelet therapy (dual or single antiplatelet therapy). Dual antiplatelet therapy was defined as a combination of aspirin ≥81 mg, clopidogrel with a loading dose (300 or 600 mg), ticagrelor, cilostazol, cangrelor, prasugrel, dipiridamole, or triflusal started within 2 days from dissection diagnosis and continued for at least 21 days, potentially followed by a single antiplatelet agent. Patients who stopped dual antiplatelet therapy before 21 days due to an event on dual antiplatelet therapy and those treated with dual antiplatelet therapy but died or were lost to follow-up before 21 days were considered to be treated with dual antiplatelet therapy. Outcomes occurring after the initial course of dual antiplatelet therapy were counted against dual antiplatelet therapy. The antiplatelet-only group included patients who were treated exclusively with antiplatelets until censoring. The anticoagulation-only group included patients who were treated exclusively with anticoagulation until censoring.

Study Outcomes

The primary outcome was a subsequent ischemic stroke during follow-up. Subsequent ischemic stroke was defined as new or worsening neurological symptoms lasting for at least 24 hours or <24 hours but with imaging evidence of new or enlarging acute infarction and referable to the territory of the affected artery.

The safety outcome was major hemorrhage during follow-up. Major hemorrhages included both major extracranial hemorrhages and symptomatic intracranial hemorrhages. Major extracranial hemorrhage was defined as a noncerebral hemorrhage requiring blood transfusion or leading to a drop in hemoglobin level of ≥2 g/dL. Symptomatic intracranial hemorrhage was defined as a new or worsening intracranial hemorrhage, causing new or worsening neurological symptoms.

Outcomes were abstracted from all available medical records and reviewed by site principal investigators. All outcome definitions were shared with sites and included in the data entry case report forms. Sites were instructed to review all pertinent medical records including available outside hospital records for outcome abstraction. Furthermore, sites were asked to confirm whether ischemic stroke outcomes were confirmed by brain imaging, and follow-up imaging reports were included when available and reviewed by neurologists from the lead site to help confirm the occurrence of study outcomes. In addition, before data set finalization, site principal investigators were asked to re-adjudicate study outcomes and confirm that events were in keeping with study-specific definitions.

Study Variables

We collected the following covariates:

Baseline demographics: Age, sex, race (White, Black, Asian, and other), and ethnicity (Hispanic versus non-Hispanic).

Comorbidities: History of hypertension, history of diabetes, history of hyperlipidemia, history of smoking, recent upper respiratory infection including coronavirus disease, and history of migraine.

Clinical variables: Clinical presentation as ischemic stroke, stroke severity (National Institutes of Health Stroke Scale score) on admission, and days from dissection diagnosis to initiation of antithrombotic treatment.

Baseline imaging variables: Location of dissected artery (carotid, vertebral, or both), site of dissection (extracranial only, intracranial only, and extracranial extending intracranially), intraluminal thrombus at or distal to the dissection site, presence of an acute infarct, presence of hemorrhagic transformation on baseline brain imaging, and presence versus absence of occlusion at the dissection site. For the purpose of this study, intracranial extension was defined as intradural extension (V4 segment of the vertebral artery and supraclinoid internal carotid artery), given that this could pose a theoretical risk of intracranial hemorrhage. When available, complete imaging reports were reviewed by the lead site to verify the imaging variables.

Treatments: Endovascular treatment (mechanical thrombectomy, stenting of dissection, or dissected vessel sacrifice) and intravenous thrombolysis.

Site Training and Data Management

All sites underwent a site initiation call to review the inclusion and exclusion criteria, study variables, and study-specific definitions. Two data entry tip sheets were shared with sites to aid with data collection. The study team provided ongoing support to sites via email or phone throughout the data collection process to ensure that all questions and concerns were addressed.

Data were entered into RedCap (Nashville, TN). After data entry was completed, individual records entered by each site, including imaging reports, were reviewed by a team of neurologists (S. Yaghi, E.G., N.K., F. Khan, C.S., and N.M.). Queries were sent as needed to confirm data accuracy and validity. After the initial round of queries were addressed, we then sent algorithm-generated queries to confirm important variables. Follow-up phone calls or meetings with sites were conducted as needed.

Analytical Plan

Main Analysis

We divided patients into 2 groups based on whether they received (1) anticoagulation (therapeutic heparin or low-molecular-weight heparin, DOAC, or VKAs) or (2) antiplatelet therapy (single or dual antiplatelet therapy).

To understand baseline differences between the 2 groups, we compared the study covariates between patients who received exclusively antiplatelets and those who received exclusively anticoagulation using the χ2 test (or Fisher exact test) for categorical variables and the Wilcoxon rank-sum test for continuous variables.

For each treatment, the follow-up period started on the day of treatment initiation. We then utilized Cox regression models with clustered standard error to account for intracluster correlation and heterogeneity when comparing study outcomes across groups. Patients were censored at the time of an outcome of interest occurring at any point after dissection diagnosis, death, loss to follow-up, or stent placement at the dissection site. Patients cross over when switching treatment groups. We used inverse probability of treatment weighting (IPTW), which is an application of propensity scores that calculates the probability of being exposed to one treatment versus the other and creates a pseudo-population based on the above probabilities so that potential confounders are equally distributed across the 2 treatment groups.11 Models were weighted for prespecified variables that have been shown or are thought to potentially alter the stroke risk and possibly affect treatment choice8,12–15 or increase major bleeding risk.16–19 Variables used for weighting and adjustment are shown in Table S1. We also performed propensity-matched analyses with replacement (a caliper of 0.05) for the ischemic stroke outcome. The propensity score was calculated using the prespecified variables weighted in model 2 (Table S1).

In all primary analyses, we used an as-treated approach; patients who switched antithrombotic groups were considered as crossovers moving from one group to the other at the time of the switch. Outcomes were only considered if they occurred while on either antiplatelet or anticoagulation. Separate, Cox models were generated for the time points 30, 90, and 180 days, respectively.

We performed sensitivity analyses, weighting variables that are significantly different between the 2 groups (antiplatelet only versus anticoagulation only) and adjusting for variables in model 3.

We also performed univariate Cox regression and generated Kaplan-Meier survival curves comparing study outcomes based on the initial antithrombotic treatment regimen patients received with censoring patients at 180 days or last follow-up (if before), death, study outcome occurring after dissection diagnosis, stent placement at dissection site, switching to a different treatment group, or stopping treatment.

Proportionality was tested using Schoenfeld residual, and accelerated failure-time parametric models were used if a model did not meet assumptions of proportionality. STATA/SE version 18.0 (StataCorp, College Station, TX) and R version 4.3.0 (R Core Team, 2023) were utilized to create visualizations of site maps in the Supplemental Material. Missing data were not imputed.

Preplanned Exploratory Subgroup Analyses

We performed the following subgroup analyses:

  1. Comparing the study outcomes between anticoagulation and dual antiplatelet therapy. This analysis was performed using an intention-to-treat approach (grouping patients based on the first treatment they received). In sensitivity analyses, we used the above approach but censored patients at the time of switching or stopping the antithrombotic agent.

  2. Comparing the study outcomes between anticoagulation with parenteral anticoagulation or DOAC versus any antiplatelet therapy.

  3. Comparing the study outcomes between anticoagulation with parenteral anticoagulation or VKA versus any antiplatelet therapy.

Interaction Analyses

We performed preplanned adjusted interaction analyses for model 2 variables, investigating the association between treatment type and ischemic stroke risk in the following subgroups: occlusive versus nonocclusive dissection; partially occlusive intraluminal thrombus versus no partially occlusive intraluminal thrombus; and presentation with versus without ischemic stroke.

RESULTS

Of the 4023 patients included in STOP-CAD (Antithrombotic Treatment for Stroke Prevention in Cervical Artery Dissection; Table S2), 3636 patients met our inclusion criteria (2453 received antiplatelets, 402 received anticoagulation, and 781 received both treatments). Reasons for exclusion are shown in Figure 1. The mean (±SD) age was 47.0±13.3 years, and 53.8% were men. There were 162 subsequent ischemic strokes (4.4%) and 28 major hemorrhages (0.8%; 19 major extracranial hemorrhages and 9 symptomatic intracranial hemorrhages) occurring up to 180 days of follow-up; 87% (95% CI, 80.9%–91.8%) of ischemic strokes occurred in the first 30 days and 98.1% were confirmed by imaging.

Figure 1.

Figure 1. Study flow chart.

Most (90.1%) of the 3636 patients had at least 30 days of follow-up data or had died by day 30. A comparison of the baseline characteristics of patients with versus without 30-day follow-up is shown in Table S3. The median (interquartile range) follow-up duration was not significantly different between patients treated exclusively with antiplatelets when compared with those treated exclusively with anticoagulation (298 [99–786] versus 220 [98–934]; P=0.999), but slightly more patients in the anticoagulation-only group had ≥30-day follow-up (92.3% [371/402] versus 88.7% [2177/2453]; P=0.034).

The baseline characteristics between patients treated only with antiplatelets and those treated only with anticoagulation are shown in Table 1. Patients treated with anticoagulation were more likely to be female (51.0% versus 45.3%; P=0.033), White (80.8% versus 73.2%; P=0.001) have a partially occlusive intraluminal thrombus (16.4% versus 6.5%; P<0.001). Patients treated with antiplatelet therapy were more likely to be of Asian race (4.0% versus 1.7%; P=0.022), have intracranial dissection (6.2% versus 2.7%; P=0.006), have received endovascular treatment (15.5% versus 10.0%; P=0.003), and have a higher initial median (interquartile range) National Institutes of Health Stroke Scale score (1 [0–5] versus 0 [0–3]; P=0.001); although both groups were mildly affected overall at presentation.

Table 1. A Comparison of Baseline Characteristics Between Patients Treated With Antiplatelet Agents Only and Those Treated With Anticoagulation Only

Antiplatelet (n=2453)Anticoagulation (n=402)P value
Age, y; median (IQR)n=285546 (37–56)45 (37–54)0.072
Female sexn=28551111/2453 (45.3%)205/402 (51.0%)0.033
White racen=28551795/2453 (73.2%)325/402 (80.8%)0.001
Black racen=2855139/2453 (5.7%)21/402 (5.2%)0.721
Asian racen=285599/2453 (4.0%)7/402 (1.7%)0.022
Non-Hispanic ethnicityn=25861993/2197 (90.7%)345/389 (88.7%)0.211
Migrainen=2855408/2453 (16.6%)80/402 (19.9%)0.107
Hypertensionn=2855865/2453 (35.3%)137/402 (34.1%)0.645
Diabetesn=2855207/2453 (8.4%)26/402 (6.5%)0.181
Hyperlipidemian=2855544/2453 (22.2%)89/402 (22.1%)0.987
Active smokingn=2855486/2453 (19.8%)80/402 (19.9%)0.967
Recent COVID-19 infectionn=285530/2453 (1.2%)4/402 (1.0%)1.000
Recent upper respiratory infectionn=2855157/2453 (6.4%)21/402 (5.2%)0.366
Ischemic stroke presentationn=28531492/2452 (60.8%)239/401 (59.6%)0.635
NIHSS score on admission, median (IQR)n=26591 (0–5)0 (0–3)0.001
Acute infarct on imagingn=28551349/2453 (55.0%)205/402 (51.0%)0.136
Baseline hemorrhagic conversionn=285594/2453 (3.8%)12/402 (3.0%)0.405
Occlusive dissectionn=2824816/2425 (33.6%)116/399 (29.1%)0.072
Partially occlusive thrombusn=2855159/2453 (6.5%)66/402 (16.4%)<0.001
Intracranial extensionn=2855219/2453 (8.9%)38/402 (9.5%)0.733
Intracranial dissectionn=2855152/2453 (6.2%)11/402 (2.7%)0.006
Vertebral dissectionn=28551220/2453 (49.7%)198/402 (49.3%)0.858
Intravenous thrombolysisn=2855321/2453 (13.1%)48/402 (11.9%)0.526
Endovascular treatmentn=2855381/2453 (15.5%)40/402 (10.0%)0.003
Days from diagnosis to treatment, median (IQR)n=28550 (0–1)0 (0–1)0.564
Aspirin treatmentn=28552416/2453 (98.5%)NA
Clopidogrel treatmentn=2855731/2453 (29.8%)NA
Ticagrelor treatmentn=285531/2453 (1.3%)NA
Cilostazol treatmentn=28551/2453 (0.0%)NA
Vitamin K antagonist treatmentn=2855231/402 (57.5%)NA
Apixaban treatmentn=2855103/402 (25.6%)NA
Rivaroxaban treatmentn=285521/402 (5.2%)NA
Dabigatran treatmentn=285511/402 (2.7%)NA
Edoxaban treatmentn=28551/402 (0.2%)NA
Low-molecular-weight heparin usen=2855208/402 (51.7%)NA
Intravenous heparin usen=2855168/402 (41.8%)NA

IQR indicates interquartile range; NA, not available; and NIHSS, National Institutes of Health Stroke Scale.

Univariable Analyses Comparing the Anticoagulation-Only Versus Antiplatelet-Only Groups

Of the 162 subsequent strokes, 130 occurred on antiplatelet therapy and 32 occurred on anticoagulation. Of the 28 major hemorrhages, 16 occurred on antiplatelet therapy and 12 occurred on anticoagulation.

In univariate analyses, when compared with patients treated exclusively with antiplatelets, those treated exclusively with anticoagulation had nonsignificantly lower ischemic stroke (1.5% versus 3.3%; P=0.059) but nonsignificantly higher major hemorrhagic events (0.7% versus 0.5%; P=0.484) by day 180. Kaplan-Meir survival analyses and univariate Cox regression analyses assigning patients to the initial treatment they received and censoring at the time of loss to follow-up or 180 days, stopping the treatment, stent at the dissection site, outcome of interest occurring after dissection diagnosis, or death are shown in Figure 2. These show a nonsignificantly lower risk of ischemic stroke by day 180 with anticoagulation (hazard ratio [HR], 0.73 [95% CI, 0.35–1.55]; P=0.411) but a higher major hemorrhage risk with anticoagulation by day 180 (HR, 8.23 [95% CI, 1.52–44.69]; P=0.015).

Figure 2.

Figure 2. Kaplan-Meier Survival curves for ischemic stroke and major bleeding outcomes. Kaplan-Meier survival curve for ischemic stroke (A) and major hemorrhage (B) comparing antiplatelets vs anticoagulation and assigning patients to the initial treatment strategy they received and censoring at the time of loss to follow-up or 180 d, stopping the treatment, stent at the dissection site, outcome of interest occurring after dissection diagnosis, or death. Unadjusted hazard ratios (HRs) were calculated with survival regression with antiplatelet as the reference group. Note that accelerated failure-time parametric models were used due to proportionality not being met for the major hemorrhage comparison. Patients with an outcome after dissection diagnosis but before starting treatment were dropped at day 0.

Association Between Anticoagulation and Subsequent Ischemic Stroke Risk in All Patients

Differences in ischemic stroke risk were not significant between the 2 groups. In adjusted Cox regression analyses with IPTW (using the crossover approach), anticoagulation treatment was associated with a nonsignificantly lower risk of subsequent ischemic stroke by day 30 (adjusted HR, 0.71 [95% CI, 0.45–1.12]; P=0.145), day 90 (adjusted HR, 0.59 [95% CI, 0.22–1.58]; P=0.293), and day 180 (adjusted HR, 0.80 [95% CI, 0.28–2.24]; P=0.670; Table 2). In propensity-matched analyses with replacement, anticoagulation treatment was associated with a nonsignificantly lower ischemic stroke risk by day 30 (HR, 0.62 [95% CI, 0.33–1.15]; P=0.129), day 90 (HR, 0.55 [95% CI, 0.30–1.02]; P=0.056), and day 180 (HR, 0.34 [95% CI, 0.07–1.55]; P=0.163).

Table 2. Unadjusted and Adjusted Cox Regression Analyses With IPTW Comparing Ischemic Stroke and Major Hemorrhage Between Anticoagulation to Antiplatelet Therapy

Unadjusted
HR (95% CI); P value
IPTW model 1
HR (95% CI); P value
IPTW model 2
HR (95% CI); P value
IPTW model 3
HR (95% CI); P value
Ischemic stroke
 30 dHR, 0.68 (95% CI, 0.44–1.05); P=0.084HR, 0.74 (95% CI, 0.47–1.15); P=0.178HR, 0.70 (95% CI, 0.45–1.11); P=0.131HR, 0.71 (95% CI, 0.45–1.12); P=0.145
 90 dHR, 0.75 (95% CI, 0.50–1.12); P=0.155HR, 0.82 (95% CI, 0.54–1.24); P=0.340HR, 0.78 (95% CI, 0.51–1.18); P=0.240HR, 0.59 (95% CI, 0.22–1.58); P=0.293*
 180 dHR, 0.83 (95% CI, 0.56–1.21); P=0.329HR, 0.90 (95% CI, 0.61–1.33); P=0.589HR, 0.77 (95% CI, 0.27–2.18); P=0.626*HR, 0.80 (95% CI, 0.28–2.24); P=0.670*
Major hemorrhage
 30 dHR, 1.11 (95% CI, 0.40–3.08); P=0.841HR, 1.12 (95% CI, 0.40–3.16); P=0.825HR, 1.36 (95% CI, 0.35–5.36); P=0.659*HR, 1.39 (95% CI, 0.35–5.45); P=0.637*
 90 dHR, 1.56 (95% CI, 0.66–3.69); P=0.310HR, 1.53 (95% CI, 0.64–3.65); P=0.334HR, 1.65 (95% CI, 0.69–3.95); P=0.264HR, 1.64 (95% CI, 0.68–3.92); P=0.269
 180 dHR, 7.46 (95% CI, 1.83–30.33); P=0.005*HR, 5.47 (95% CI, 1.60–18.72); P=0.007*HR, 5.51 (95% CI, 1.55–19.51); P=0.008*HR, 5.56 (95% CI, 1.53–20.13); P=0.009*

IPTW indicates inverse probability weighting; and HR, hazard ratio.

* Accelerated failure-time parametric models used due to proportionality not met. Variables used for weighting and adjustment in all models are listed in Table S1.

Association Between Anticoagulation and Major Hemorrhage Risk in All Patients

In Cox regression analyses with IPTW (using the crossover approach), anticoagulation treatment was associated with an increased risk of major hemorrhage by day 180 (adjusted HR, 5.56 [95% CI, 1.53–20.13]; P=0.009) but not day 90 (adjusted HR, 1.64 [95% CI, 0.68–3.92]; P=0.269) or day 30 (adjusted HR, 1.39 [95% CI, 0.35–5.45]; P=0.637; Table 2).

Sensitivity Analyses

Weighting Based on Differences in Baseline Characteristics

We performed sensitivity analyses, limiting the weighting to variables significantly different between the 2 groups in Table 1. In the fully adjusted model, anticoagulation was associated with a nonsignificantly lower risk of ischemic stroke by day 30 (adjusted HR, 0.79 [95% CI, 0.46–1.36]; P=0.397), day 90 (adjusted HR, 0.68 [95% CI, 0.22–2.11]; P=0.509), and day 180 (adjusted HR, 0.87 [95% CI, 0.28–2.71]; P=0.810). Anticoagulation was associated with an increased risk of major hemorrhage by day 180 (adjusted HR, 6.77 [95% CI, 1.43–32.01]; P=0.016) but not day 30 (adjusted HR, 1.89 [95% CI, 0.41–8.66]; P=0.415) or day 90 (adjusted HR, 3.27 [95% CI, 0.57–18.80]; P=0.185).

Exploratory Analyses

Dual Antiplatelet Therapy Versus Anticoagulation

There were 535 patients treated with dual antiplatelet therapy. The median (interquartile range) duration of dual antiplatelets was 77 days (22–121 days). In adjusted analyses, outcomes did not significantly differ between the 2 groups. After IPTW (model 3), the risk of subsequent ischemic stroke was nonsignificantly higher in patients treated with dual antiplatelet therapy when compared with anticoagulation therapy by day 30 (adjusted HR, 1.87 [95% CI, 0.77–4.54]; P=0.164), day 90 (adjusted HR, 1.97 [95% CI, 0.80–4.82]; P=0.138), and day 180 (adjusted HR, 2.03 [95% CI, 0.90–4.59]; P=0.09). Overall effect sizes were consistent with the main analyses and persisted in propensity-matched analyses (Table S4).

The risk of major bleeding was nonsignificantly lower with dual antiplatelet therapy than anticoagulation by day 30 (adjusted HR, 0.89 [95% CI, 0.18–4.36]; P=0.886), day 90 (adjusted HR, 0.60 [95% CI, 0.15–2.44]; P=0.472), and day 180 (adjusted HR, 0.59 [95% CI, 0.18–1.95]; P=0.385; Table S5).

Overall findings were unchanged when censoring patients at the time of switching treatment groups.

Anticoagulation With a VKA Oral Agent or Parenteral Anticoagulation Compared With Antiplatelet Therapy

There were 674 patients treated with a VKA or parenteral anticoagulation. When compared with antiplatelet therapy, after IPTW (model 3), anticoagulation with a VKA as the oral agent or parenteral anticoagulation was not associated with a lower risk of ischemic stroke by day 30 (adjusted HR, 0.95 [95% CI, 0.32–2.85]; P=0.931), day 90 (adjusted HR, 0.98 [95% CI, 0.22–4.32]; P=0.975), or day 180 (adjusted HR, 1.84 [95% CI, 0.46–7.28]; P=0.386). In propensity-matched analyses, however, there was a nonsignificantly lower ischemic stroke risk with VKA use by day 30 (HR, 0.31 [95% CI, 0.08–1.18]; P=0.086), day 90 (HR, 0.20 [95% CI, 0.03–1.24]; P=0.083), and day 180 (HR, 0.18 [95% CI, 0.03–1.34]; P=0.095; Table S4).

In adjusted Cox regression analyses, the risk of major hemorrhage was nonsignificantly higher with VKAs or parenteral anticoagulation by day 180 (adjusted HR, 8.74 [95% CI, 0.67–114.79]; P=0.099; Table S5).

Anticoagulation With a DOAC or Parenteral Anticoagulation Compared With Antiplatelet Therapy

There were 565 patients treated with a DOAC or parenteral anticoagulation. When compared with antiplatelet therapy, after IPTW (model 3), anticoagulation with a DOAC or parenteral anticoagulation was associated with a nonsignificantly lower risk of ischemic stroke by day 30 (adjusted HR, 0.66 [95% CI, 0.37–1.18]; P=0.162), day 90 (adjusted HR, 0.52 [95% CI, 0.13–2.03]; P=0.348), and day 180 (adjusted HR, 0.59 [95% CI, 0.14–2.48]; P=0.468). Overall findings were similar with propensity score matching (Table S4).

In adjusted Cox regression analyses, the risk of major hemorrhage was nonsignificantly higher with DOAC or parenteral anticoagulation by day 180 (adjusted HR, 4.08 [95% CI, 0.99–16.71]; P=0.051; Table S5).

Interaction Analyses

The findings on ischemic stroke outcome in our study did not vary in the subgroups of partially occlusive thrombus and presentation with ischemic stroke (Table S6). However, the effect of anticoagulation on the risk of subsequent ischemic stroke by day 180 depended on the presence of occlusion (Pinteraction=0.009), indicating an association with reduced ischemic stroke risk in patients with occlusive dissection (HR, 0.40 [95% CI, 0.18–0.88]) but not with nonocclusive dissection (HR, 1.34 [95% CI, 0.83–2.14]). No interactions were found for the major hemorrhage outcome (Table S6).

DISCUSSION

The STOP-CAD study includes a large population of patients with CAD and provides real-world contemporary data on the association of an antithrombotic regimen with the risk of subsequent ischemic stroke and major hemorrhage. Consistent with prior observations, overall event rates were low; the risk of subsequent ischemic stroke up to 180 days was 4.4% and that of major hemorrhage was 0.8%. We found a nonsignificantly lower ischemic stroke risk with anticoagulation treatment but a higher bleeding risk when used for up to 180 days. We found a potential benefit for reducing ischemic stroke risk with anticoagulation in patients with occlusive dissection.

Our findings align with the TREAT-CAD study (Aspirin Versus Anticoagulation in Cervical Artery Dissection), which showed numerically fewer ischemic events with VKA anticoagulation compared with aspirin.5 Furthermore, a meta-analysis of the TREAT-CAD study and CADISS (Cervical Artery Dissection in Stroke Study) trial showed fewer ischemic strokes but more frequent major bleeding events with anticoagulation.20 The STOP-CAD study provides evidence of nonsignificantly lower subsequent ischemic strokes but higher major hemorrhage risk in patients treated with anticoagulation in a real-world setting. The majority of subsequent ischemic strokes in our study and other studies occurred in the first 4 weeks after dissection diagnosis.20 Therefore, limiting the more aggressive antithrombotic strategy to the first 30 days could possibly result in a lower risk of ischemic stroke without significantly increasing the major bleeding risk.

The benefit of a short course of more intensified antithrombotic regimen was also suggested in patients with minor strokes or high-risk transient ischemic attacks, irrespective of the cause. In this patient population, the greatest benefit was achieved with 21 days of dual antiplatelet therapy, while a more prolonged duration of dual antiplatelet therapy was associated with an increased risk of major bleeding without any added benefit in ischemic stroke prevention.21 Our results for ischemic stroke reduction with anticoagulation did not achieve statistical significance, thus the potential benefit of a 30-day anticoagulation approach followed by single antiplatelet therapy needs further study.

Overall, the effect size for the ischemic stroke outcome was grossly similar when dual antiplatelet therapy was compared with anticoagulation. Thus, we cannot exclude a potential benefit of anticoagulation over dual antiplatelet therapy in lowering ischemic stroke risk, particularly in patients perceived as having a higher ischemic stroke risk. This requires further study.

In this study, it remains unclear whether the possible protective effect on ischemic stroke reduction was driven by the specific anticoagulant regimen used, as the number of patients in the different anticoagulation subgroups was too small to make any firm conclusions. Furthermore, the risk of major hemorrhage with anticoagulation seemed to be increased in the subgroups of VKAs or parenteral anticoagulation or the subgroup of DOAC or parenteral anticoagulation.

In our study, patients with occlusive dissection, a high-risk subgroup,22 had a lower ischemic stroke risk with anticoagulation versus antiplatelet therapy but patients without an occlusive dissection did not. This was not shown in the other prespecified subgroups.

Altogether, it is important to weigh the risks and benefits when considering anticoagulation. If anticoagulation is chosen, it is crucial to mitigate the bleeding risk by shortening the duration of anticoagulation treatment and controlling other risk factors for major bleeding, albeit that risk was low (≈0.8%) in our study.

Strengths and Limitations

Our study has several major limitations. First, the retrospective and observational nature of our study may affect our findings through confounding by indication. Current treatment practices likely result in patients with a perceived lower risk of ischemic stroke being treated with antiplatelet therapy and those with a lower bleeding risk being treated with anticoagulation. Thus, treatment bias could potentially bias toward the use of antiplatelets in a group likelier to be predisposed to bleeding and anticoagulation in a group likelier to be at higher ischemic stroke risk. However, we found the opposite relationship in the study. Furthermore, we used IPTW and propensity score matching to mitigate bias. Despite our efforts, residual confounding may still be present. Second, our study lacked central and blinded outcome adjudication. However, trained neurologists at the central site reviewed all study outcomes and imaging reports linked to these outcomes when available and verified them with the site principal investigators. Third, ≈10% of patients were lost to follow-up after 30 days. Fourth, most sites in our study were tertiary care centers at large academic institutions from high-income countries. Thus, our findings may not be generalizable to patients treated at community hospitals or from moderate- or low-income countries. Patients treated at large tertiary centers are likely to present with a more severe disease. Fifth, due to the low number of major hemorrhagic events, we were unable to reliably perform propensity score matching for the major hemorrhage outcome. Sixth, the dual antiplatelet therapy group included different regimens, but the overwhelming majority were aspirin and clopidogrel. Thus, the findings in this analysis may not be generalizable to nonclopidogrel-based dual antiplatelet therapy. Seventh, we used an as-treated and crossover approach. Because the risk of a new ischemic stroke is highest early after the initial event, this approach may have favored treatments started at a later time point as opposed to those started initially. We sought to mitigate this bias by adjusting for the time to initiation of each treatment, but residual confounding may still be present. Finally, this study did not analyze imaging data and did not capture reasons for crossover. Further analyses, including them, may provide clinically useful information.

Our study has several strengths. First, it is a large multinational study with an ethnically and geographically diverse patient population, capturing different practice patterns, and thus the results are more generalizable than other studies of CAD. Second, the large number of patients provided sufficient power to evaluate for differences between antiplatelet and anticoagulation therapies and to perform certain subgroup analyses. Finally, we captured many key variables, allowing us to account for and minimize potential treatment bias and to investigate the effect of treatment in various subgroups.

Conclusions

The STOP-CAD study cannot exclude a benefit of anticoagulation over antiplatelet therapy in ischemic stroke risk following CAD, particularly in patients with occlusive dissection. If anticoagulation is chosen initially, it may be reasonable to switch to antiplatelet therapy before 180 days to lower the risk of major bleeding. Because our results for ischemic stroke reduction with anticoagulation versus antiplatelet treatment did not achieve statistical significance and given the abovementioned limitations of the STOP-CAD study design, our findings require validation by large prospective studies. Our study could inform the design of prospective registries with systematic data collection, pragmatic trials with decentralized collection of data,23 or pooled cohort studies and meta-analyses to help further address this important clinical question.

ARTICLE INFORMATION

Supplemental Material

Tables S1–S6

Figure S1

Nonstandard Abbreviations and Acronyms

CAD

cervical artery dissection

DOAC

direct oral anticoagulant

HR

hazard ratio

IPTW

inverse probability of treatment weighting

STOP-CAD

Antithrombotic Treatment for Stroke Prevention in Cervical Artery Dissection

VKA

vitamin K antagonist

TREAT-CAD

Aspirin Versus Anticoagulation in Cervical Artery Dissection

CADISS

Cervical Artery Dissection in Stroke Study

Disclosures Disclosures provided by Dr Nguyen in compliance with American Heart Association annual Journal Editor Disclosure Questionnaire are available at https://www.ahajournals.org/editor-coi-disclosures. Dr Arnold reports compensation from Boehringer Ingelheim, AstraZeneca, Bayer, Bristol-Myers Squibb, Covidien, Daiichi Sankyo, Novartis, Sanofi, Pfizer, Medtronic, Novo Nordisk, and Amgen for consultant services. Dr Lester reports a provisional patent for Methods and compositions for disrupting tau aggregates. Dr Touze reports compensation from Elsevier for other services and employment by Caen. J.E. Kaufman reports grants from Goldschmidt Jacobson-Stiftung. Dr Traenka reports travel support from Bayer Healthcare. Dr Aguiar de Sousa reports compensation from Daiichi Sankyo, Bayer, AstraZeneca, Johnson & Johnson, and Fundação Bial for other services; compensation from the University of British Columbia for data and safety monitoring services; compensation from Organon & Co for consultant services. Dr Rosa reports grants from Merck Sharp & Dohme Corporation. Dr Field reports compensation from HLS Therapeutics, AstraZeneca Canada, and Roche for consultant services; service as a board member for Destine Health; and compensation from the Canadian Medical Protective Association for expert witness services; and grants from Bayer. Dr Leker reports compensation from Medtronic, Ischemaview, Bayer, Abbott Diabetes Care, Biogen, Janssen Biotech, and Boehringer Ingelheim for other services. Dr Nolte reports compensation from Daiichi Sankyo Europe GmbH, Boehringer Ingelheim, Pfizer, Bristol-Myers Squibb, and Alexion Pharmaceuticals for consultant services; and compensation from AstraZeneca, Abbott Canada, Deutsches Zentrum für Neurodegenerative Erkrankungen, Novartis, Portola Pharmaceuticals, Deutsches Zentrum für Herz-Kreislaufforschung, and Novartis for other services. Dr Poppe reports grants from Foundation Brain Canada, Heart and Stroke Foundation of Canada, and Stryker; and compensation from Roche for other services. Dr Liebeskind reports compensation from Medtronic, Genentech, Cerenovus, Stryker, and Rapid Medical Ltd, for consultant services. B. Mac Grory reports grants from the National Institutes of Health; employment by Duke University Medical Center; compensation from Bayer for other services; grants from the American Heart Association, Duke Bass Connections, and the Duke Office of Physician Scientist Development. Dr Al Kasab reports compensation from Stryker for other services and employment by Medical University of South Carolina. Dr Kicielinski reports compensation from Stryker, Penumbra Inc, Medtronic, and MicroVention Inc, for other services; travel support from MicroVention Inc; and employment by Medical University of South Carolina and Elsevier. Dr de Havenon reports stock options in TitinKM and Certus; grants from the National Institutes of Health; and compensation from Novo Nordisk for consultant services. Dr Siegler reports grants from Philips and employment by the University of Chicago. Dr Willey reports compensation from Edwards Lifesciences Corporation and Abbott Fund for end point review committee services; compensation from Uptodate for other services; and compensation from the Abbott Laboratories for consultant services. Dr Sharma reports a provisional patent for a stroke etiology classifier algorithm and grants from the National Institutes of Health Clinical Center. Dr Martins reports compensation from Pfizer, Medtronic, Servier Affaires Medicales, Daiichi Sankyo, Bayer, Novo Nordisk, Novartis, Penumbra Inc, and Boehringer Ingelheim for other services. Dr Simpkins reports grants from the National Institutes of Health. Dr Stretz reports grants from Massachusetts General Hospital. Dr Furie reports compensation from Janssen Biotech for consultant services. The other authors report no conflicts

Footnotes

Presented in part at the International Stroke Conference, Phoenix, AZ, February 7–9, 2024.

The podcast and transcript are available at https://www.ahajournals.org/str/podcast.

*The full author list is available in the Article Information section.

This manuscript was sent to Marc Fisher, Senior Guest Editor, for review by expert referees, editorial decision, and final disposition.

For Sources of Funding and Disclosures, see page 917.

Supplemental Material is available at https://www.ahajournals.org/doi/suppl/10.1161/STROKEAHA.123.045731.

Correspondence to: Shadi Yaghi, MD, Department of Neurology, Brown University, 593 Eddy St, Providence, RI 02903. Email

REFERENCES

  • 1. Béjot Y, Daubail B, Debette S, Durier J, Giroud M. Incidence and outcome of cerebrovascular events related to cervical artery dissection: the Dijon Stroke Registry.Int J Stroke. 2014; 9:879–882. doi: 10.1111/ijs.12154CrossrefMedlineGoogle Scholar
  • 2. Putaala J, Metso AJ, Metso TM, Konkola N, Kraemer Y, Haapaniemi E, Kaste M, Tatlisumak T. Analysis of 1008 consecutive patients aged 15 to 49 with first-ever ischemic stroke.Stroke. 2009; 40:1195–1203. doi: 10.1161/STROKEAHA.108.529883LinkGoogle Scholar
  • 3. Ekker MS, Verhoeven JI, Schellekens MMI, Boot EM, van Alebeek ME, Brouwers PJAM, Arntz RM, van Dijk GW, Gons RAR, van Uden IWM, et al. Risk factors and causes of ischemic stroke in 1322 young adults.Stroke. 2023; 54:439–447. doi: 10.1161/STROKEAHA.122.040524LinkGoogle Scholar
  • 4. Markus HS, Levi C, King A, Madigan J, Norris J; Cervical Artery Dissection in Stroke Study (CADISS) Investigators. Antiplatelet therapy vs anticoagulation therapy in cervical artery dissection: the Cervical Artery Dissection in Stroke Study (CADISS) randomized clinical trial final results.JAMA Neurol. 2019; 76:657–664. doi: 10.1001/jamaneurol.2019.0072CrossrefGoogle Scholar
  • 5. Engelter ST, Traenka C, Gensicke H, Schaedelin SA, Luft AR, Simonetti BG, Fischer U, Michel P, Sirimarco G, Kägi G, et al; TREAT-CAD Investigators. Aspirin Versus Anticoagulation in Cervical Artery Dissection (TREAT-CAD): an open-label, randomised, non-inferiority trial.Lancet Neurol. 2021; 20:341–350. doi: 10.1016/S1474-4422(21)00044-2CrossrefGoogle Scholar
  • 6. Ye SL, Wang C, Wang LL, Xu TZ, Li XQ, Tang T. Oral anticoagulant and antiplatelet therapy for cervical artery dissection: a meta-analysis of clinical trials.Clin Appl Thromb Hemost. 2021; 27:10760296211051708. doi: 10.1177/10760296211051708CrossrefGoogle Scholar
  • 7. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP; STROBE Initiative. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies.Lancet. 2007; 370:1453–1457. doi: 10.1016/S0140-6736(07)61602-XCrossrefMedlineGoogle Scholar
  • 8. Morris NA, Merkler AE, Gialdini G, Kamel H. Timing of incident stroke risk after cervical artery dissection presenting without ischemia.Stroke. 2017; 48:551–555. doi: 10.1161/STROKEAHA.116.015185LinkGoogle Scholar
  • 9. Qureshi AI, Chaudhry SA, Hassan AE, Zacharatos H, Rodriguez GJ, Suri MF, Lakshminarayan K, Ezzeddine MA. Thrombolytic treatment of patients with acute ischemic stroke related to underlying arterial dissection in the United States.Arch Neurol. 2011; 68:1536–1542. doi: 10.1001/archneurol.2011.213CrossrefMedlineGoogle Scholar
  • 10. Witsch J, Mir SA, Parikh NS, Murthy SB, Kamel H, Navi BB, Segal AZ, Fink ME, Rutrick SB, Safford MM, et al. Association between cervical artery dissection and aortic dissection.Circulation. 2021; 144:840–842. doi: 10.1161/CIRCULATIONAHA.121.055274LinkGoogle Scholar
  • 11. Chesnaye NC, Stel VS, Tripepi G, Dekker FW, Fu EL, Zoccali C, Jager KJ. An introduction to inverse probability of treatment weighting in observational research.Clin Kidney J. 2022; 15:14–20. doi: 10.1093/ckj/sfab158CrossrefGoogle Scholar
  • 12. Arnold M, Kurmann R, Galimanis A, Sarikaya H, Stapf C, Gralla J, Georgiadis D, Fischer U, Mattle HP, Bousser MG, et al. Differences in demographic characteristics and risk factors in patients with spontaneous vertebral artery dissections with and without ischemic events.Stroke. 2010; 41:802–804. doi: 10.1161/STROKEAHA.109.570655LinkGoogle Scholar
  • 13. Wu Y, Wu F, Liu Y, Fan Z, Fisher M, Li D, Xu W, Jiang T, Cheng J, Sun B, et al. High-resolution magnetic resonance imaging of cervicocranial artery dissection: imaging features associated with stroke.Stroke. 2019; 50:3101–3107. doi: 10.1161/STROKEAHA.119.026362LinkGoogle Scholar
  • 14. Lichy C, Metso A, Pezzini A, Leys D, Metso T, Lyrer P, Debette S, Thijs V, Abboud S, Kloss M, et al; Cervical Artery Dissection and Ischemic Stroke Patients-Study Group. Predictors of delayed stroke in patients with cervical artery dissection.Int J Stroke. 2015; 10:360–363. doi: 10.1111/j.1747-4949.2012.00954.xCrossrefGoogle Scholar
  • 15. von Babo M, De Marchis GM, Sarikaya H, Stapf C, Buffon F, Fischer U, Heldner MR, Gralla J, Jung S, Simonetti BG, et al. Differences and similarities between spontaneous dissections of the internal carotid artery and the vertebral artery.Stroke. 2013; 44:1537–1542. doi: 10.1161/STROKEAHA.113.001057LinkGoogle Scholar
  • 16. Paciaroni M, Agnelli G, Caso V, Tsivgoulis G, Furie KL, Tadi P, Becattini C, Falocci N, Zedde M, Abdul-Rahim AH, et al. Prediction of early recurrent thromboembolic event and major bleeding in patients with acute stroke and atrial fibrillation by a risk stratification schema: the ALESSA score study.Stroke. 2017; 48:726–732. doi: 10.1161/STROKEAHA.116.015770LinkGoogle Scholar
  • 17. Debette S, Leys D. Cervical-artery dissections: predisposing factors, diagnosis, and outcome.Lancet Neurol. 2009; 8:668–678. doi: 10.1016/S1474-4422(09)70084-5CrossrefMedlineGoogle Scholar
  • 18. Biller J, Sacco RL, Albuquerque FC, Demaerschalk BM, Fayad P, Long PH, Noorollah LD, Panagos PD, Schievink WI, Schwartz NE, et al; American Heart Association Stroke Council. Cervical arterial dissections and association with cervical manipulative therapy: a statement for healthcare professionals from the American Heart Association/American Stroke Association.Stroke. 2014; 45:3155–3174. doi: 10.1161/STR.0000000000000016LinkGoogle Scholar
  • 19. Torn M, Bollen WLEM, van der Meer FJM, van der Wall EE, Rosendaal FR. Risks of oral anticoagulant therapy with increasing age.Arch Intern Med. 2005; 165:1527–1532. doi: 10.1001/archinte.165.13.1527CrossrefMedlineGoogle Scholar
  • 20. Debette S, Mazighi M, Bijlenga P, Pezzini A, Koga M, Bersano A, Kõrv J, Haemmerli J, Canavero I, Tekiela P, et al. ESO guideline for the management of extracranial and intracranial artery dissection.Eur Stroke J. 2021; 6:XXXIX–LXXXVIII. doi: 10.1177/23969873211046475CrossrefGoogle Scholar
  • 21. Johnston SC, Elm JJ, Easton JD, Farrant M, Barsan WG, Kim AS, Lindblad AS, Palesch YY, Zurita KG, Albers GW, et al; POINT and Neurological Emergencies Treatment Trials Network Investigators. Time course for benefit and risk of clopidogrel and aspirin after acute transient ischemic attack and minor ischemic stroke.Circulation. 2019; 140:658–664. doi: 10.1161/CIRCULATIONAHA.119.040713LinkGoogle Scholar
  • 22. Traenka C, Grond-Ginsbach C, Goeggel Simonetti B, Metso TM, Debette S, Pezzini A, Kloss M, Majersik JJ, Southerland AM, Leys D, et al; CADISP-Plus Study Group. Artery occlusion independently predicts unfavorable outcome in cervical artery dissection.Neurology. 2020; 94:e170–e180. doi: 10.1212/WNL.0000000000008654CrossrefGoogle Scholar
  • 23. Yaghi S, Siegler JE, Nguyen TN. Pitfalls of randomized controlled trials in stroke: how can we do better?Stroke Vasc Interv Neurol. 2023; 3:e000807. doi: 10.1161/SVIN.123.000807LinkGoogle Scholar

eLetters(0)

eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.

Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.