Clinical characteristics and outcome in expansive compared with steno-occlusive mural hematoma in spontaneous cervical artery dissection

Abstract

Background:

Spontaneous cervical artery dissection (sCeAD) is one of the prime causes of ischemic stroke in young adults. Based on vessel wall imaging, steno-occlusive or expansive wall hematomas can be distinguished. It is unclear whether these two distinct morphological phenotypes reflect different pathophysiological processes.

Aim:

We aim to evaluate differences in clinical characteristics and long-term recurrence between patients with expansive and steno-occlusive mural wall hematoma in the acute phase.

Methods:

Participants of the ReSect-study, one of the largest single-center cohort studies with long-term follow-up of sCeAD patients, with sufficient magnetic resonance imaging (MRI) were included. All available MRI scans were retrospectively evaluated for patients dichotomized to two groups: (1) mural hematoma causing steno-occlusive pathologies without expansion of total vessel diameter (steno-occlusive hematoma), and (2) mural hematoma causing vessel diameter expansion without lumen stenosis (expansive hematoma). Patients with mixed steno-occlusive and expansive vessel pathologies were excluded from the analysis.

Results:

In total, 221 individuals were available for analysis. The pathognomonic vessel wall hematoma was steno-occlusive in 187 (84.6%) and expansive in 34 (15.4%). No difference was seen in patient demographics, clinical status at admission, laboratory parameters, family history, or the frequency of clinical stigmata for connective tissue disorders. Both patients with expansive and steno-occlusive mural hematoma had a high likelihood of suffering cerebral ischemia (64.7 vs 79.7). Still, time from symptom onset to diagnosis was significantly longer in those with expansive dissection (17.8 vs 7.8 days, p = 0.02). Those with expansive dissections were more likely to have upper respiratory infection within 4 weeks prior to dissection (26.5% vs 12.3%, p = 0.03). Upon follow-up, functional outcome was identical and groups did not differ in rate of sCeAD recurrence, but those with expansive mural hematoma at baseline more frequently had residual aneurysmal formation (41.2% vs 11.5%, p < 0.01).

Conclusions:

As cerebral ischemia was frequent in both, our clinical results do not advise for differential treatment or follow-up based on the acute morphological phenotype. There was no clear evidence of a different aetiopathogenesis between patients with steno-occlusive or expansive mural hematoma in the acute phase. More mechanistic approaches are needed to elucidate potential differences in pathomechanism between both entities.

Data access:

Anonymized data not published within this article will be made available by request from any qualified investigator.

Introduction

The pathogenesis for one of the prime reasons for ischemic stroke in young adults, spontaneous cervical artery dissection (sCeAD), remains unclear to date.^1,2 It is well known that the pathognomonic vessel wall hematoma can be located either subintimally or subadventitially. Subintimal mural hematomas are believed to be associated with intimal tear and anterograde blood flow from the vessel to the false lumen (inside-out theory) resulting in vessel stenosis or occlusion, while subadventitial sCeAD may arise from rupture of the vasa vasorum (outside-in theory) and rather lead to expansive vessel pathologies.^3–8 It is tempting to speculate that these two distinct phenotypes, which can be distinguished in magnetic resonance imaging (MRI), have non-identical pathomechanisms that might be reflected by differences in patient characteristics as well as the long-term outcome. To date, clinical features and outcomes have only been assessed in patients with either steno-occlusive or aneurysmal post-sCeAD pathologies.^7–9

Aims and/or hypothesis

It was our goal to investigate whether individuals with either steno-occlusive or expansive sCeAD-related mural hematoma in the acute phase differ in clinical characteristics, functional or imaging outcome, or show clinical signs of connective tissue disorder. Furthermore, we investigated whether either one of the types of vessel wall hematomas is associated with sCeAD recurrence in the long run.

Methods

Patient recruitment, selection, and variable definition

During the ReSect study, each sCeAD patient admitted to the Department of Neurology of the Medical University of Innsbruck from January 1990 until December of 2016 was identified and invited to a standardized outpatient follow-up visit, which had to have been at least 1 year after sCeAD event. sCeAD diagnosis had to be secured through mural hematoma visualization in T1-fat saturated MRI images at baseline. Furthermore, known monogenetic tissue disorders and a timely association to major head/neck trauma or trauma causing additional signs of separate external or internal injury were exclusion criteria. For the current analysis, all available MRI images of ReSect patients were retrospectively evaluated. To enable group allocation, the raw magnetic resonance angiography (MRA) images and T1 fat-saturated MRI sequence at dissection-diagnosis (baseline) with multiplanar-reconstructions were used. Individuals were dichotomized as follows: (1) steno-occlusive hematoma—mural hematoma causing either lumen stenosis or occlusion without expansion of total vessel diameter at the dissection site, and (2) expansive hematoma—mural hematoma causing vessel diameter expansion without lumen stenosis or occlusion at dissection site (Figure 1). Vessel and lumen diameter of the dissection site were compared to proximal, non-dissected sections of the same artery. A graduation of vessel stenosis was not performed. The group allocation was performed by one rater (LM-S). To account for intra-rater variability, a random sample of 30 ReSect participants was re-rated 4 months after the initial rating. Patients with single-vessel sCeAD causing simultaneous stenosis and diameter expansion as well as those with multiple-vessel sCeAD with both traits being evident were excluded. Variables used for detailing patient and sCeAD characteristics were defined as follows: (a) ischemia was a composite variable encompassing transient ischemic attack and ischemic stroke; (b) vascular risk factors were assessed in accordance with current guidelines;^10,11 (c) comorbidities were identified through patient history taking and retrospective chart review; (d) family history was reserved for first-degree relatives; (e) patients with multiple single-time point sCeAD had more than one cervical artery dissected at dissection diagnosis (baseline); and (f) symptom to diagnosis was defined as the duration between perceived symptom begin as stated by each patient to MRI-based visualization of the mural hematoma. The ReSect-study-specific follow-up included a detailed patient history taking, 3T-whole-body contrast-enhanced MRI, and a structured clinical examination. The examination included the National Institutes of Health Stroke Scale (NIHSS), Barthel Index, modified Rankin Scale (mRS), and an assessment of potential clinical stigmata for connective tissue disorders, which were done by an experienced dermatologist and mirrored the investigation proposed by Dittrich et al. ¹² Missing values were completed through retrospective chart review of all available electronic medical files. At follow-up, the raw MRA images and T1 fat-saturated sequence of the study-specific 3T-MRI were used to assess residual vessel pathologies. Further details on the ReSect study have already been published.^13–17

Figure 1.

Examples for (a) steno-occlusive and (b) expansive mural hematoma due to sCeAD of the right internal carotid artery.

Statistical methodology

IBM SPSS Statistics 27 was used for the statistical analysis. Differences in categorical or continuous variables were calculated using Pearson chi-square or Mann–Whitney U test, respectively.

Standard protocol approvals, registration, and patient consents

All analyses were approved by the local ethics committee at the Medical University Innsbruck (EK#UN5072,325/4.1) and appropriate informed consent of patients who took part in the ReSect study was obtained.

Data availability

Study data that support the findings of this study are available from the corresponding author upon reasonable request after ethics approval and receipt of a signed material transfer agreement.

Results

In total, 279 individuals were included in the ReSect study. After excluding five patients with insufficient baseline imaging and 53 individuals with simultaneous expansive and steno-occlusive sCeAD-related mural hematoma or multiple dissections, 221 individuals were available for analysis. sCeAD was steno-occlusive in 187 (84.6%) and expansive in 34 (15.4%). Concerning intra-rater variability, group allocation in a re-rating was identical. Patient-related and dissection-related baseline as well as follow-up information are given within Tables 1 and 2. In short, clinical presentation was similar in steno-occlusive and expansive sCeAD, which included likelihood of suffering cerebral ischemia. Local symptoms differed concerning frequency of peripheral cranial nerve palsies between the two entities. Furthermore, after excluding those with central Horner’s syndrome due to brainstem infarction, those with expansive sCeAD were more likely to have peripheral Horner’s syndrome compared to those with steno-occlusive sCeAD (14.7% vs 48%, p = 0.02). Still, time from symptom onset to sCeAD diagnosis was significantly longer in those with expansive sCeAD. Concerning baseline characteristics, those with expansive sCeAD were more likely to have upper respiratory infection within 4 weeks prior to sCeAD diagnosis. No difference was seen in patient demographics, clinical status at admission, laboratory parameters, vessel segment involvement of dissected artery (i.e. C1a–C3 in internal carotid and V0–V4 in vertebral artery dissection, data not shown), family history, or the frequency of clinical stigmata for connective tissue disorders. Upon follow-up, functional outcome was identical and groups did not differ in rate of sCeAD recurrence. Only residual aneurysmal formation was significantly more frequent in patients with expansive mural hematoma at baseline. Irrespective of group allocation, all 11 patients suffering stroke during follow-up were treated with antiplatelets at the time of onset. In addition, we evaluated the antithrombotic treatment in all patients. Of those with residual aneurysm formation, 13 of 18 (72.2%) were on antiplatelets and 1 of 18 (5.5%) took vitamin K antagonists. In patients with residual stenosis or occlusion, 25 of 32 (78.1%) took antiplatelets and 1 of 32 (3.1%) received direct oral anticoagulant (DOAC) treatment at the time of follow-up. Patients being without antithrombotic treatment were predominantly due to medication noncompliance, as our local standard operating procedures (SOPs) state that a residual vessel pathology warrants continuous treatment. We have added this information to the results section.

Table 1.

Differences in patient characteristics between those with steno-occlusion or diameter expansion due to sCeAD-related mural hematoma at baseline.

Baseline	Steno-occlusive sCeAD	Expansive sCeAD	p
	n = 187	n = 34
Female a	72 (38.5)	9 (26.5)	0.18
Age (years) b	44.6 (15.3)	45.8 (15.6)	0.96
Weight (kg) b	78.2 (22.5)	75.5 (22.1)	0.42
Height (cm) b	172.1 (14.0)	171.1 (12.0)	0.49
Local signs and symptomsa
Head/neck pain	142 (75.9)	26 (76.5)	0.97
Horner’s syndrome	25 (13.4)	5 (14.7)	0.85
Cranial nerve palsy	0 (0.0)	5 (14.7)	<0.01
Pulsatile tinnitus	11 (5.9)	4 (11.8)	0.22
Ischemia a	149 (79.7)	22 (64.7)	0.06
NIHSS admission b	2.8 (3.0)	2.4 (2.3)	0.18
mRS admission b	1.9 (2.0)	1.6 (2.0)	0.12
Vascular risk factors and comorbidities a
Type 2 diabetes	7 (3.7)	0 (0.0)	0.25
Hypertension	65 (34.8)	12 (35.3)	0.95
Dyslipidemia	84 (44.9)	12 (35.3)	0.30
Active smoking	46 (24.6)	11 (32.4)	0.28
Peripheral arterial disease	1 (0.5)	0 (0.0)	0.67
COPD	3 (1.6)	0 (0.0)	0.46
Laboratory markers b
HbA1c (%)	5.4 (0.5)	5.3 (0.2)	0.96
Cholesterol (mg/dL)	189.6 (57.0)	182.4 (48.5)	0.71
LDL (mg/dL)	118.0 (52.0)	111.0 (45.0)	0.58
Family history a
Stroke	(n = 104)32 (30.8)	(n = 20)7 (35.0)	0.71
Dissection	(n = 97)1 (1.0)	(n = 20)1 (5.0)	0.20
Migraine	(n = 97)15 (15.5)	(n = 20)3 (15.8)	1.00
Follow-up
Age at follow-up (years) b	52.0 (14.5)	53.1 (17.4)	0.94
Follow-up duration (years) b	7.9 (7.3)	7.8 (5.4)	0.74
Weight (kg) b	77.6 (18.8)	81.5 (20.3)	0.96
Waist circumference (cm) b	92.0 (19.5)	93.0 (15.0)	0.90
Hip circumference (cm) b	97.7 (13.0)	96.6 (18.0)	0.40
NIHSS b	0.7 (1.0)	0.2 (0.0)	0.07
mRS b	0.5 (1.0)	0.3 (0.5)	0.33
Barthel index b	99.5 (0.0)	98.5 (0.0)	0.06
Novel diagnoses since sCeAD a
Ischemic stroke	10 (5.3)	1 (2.9)	0.31
Migraine	28 (15.0)	3 (8.8)	0.34
Hypertension	(n = 166)71 (42.8)	18 (52.9)	0.08
Depression	24 (12.8)	5 (14.7)	0.52
Laboratory markers b
HbA1c (%)	5.7 (0.6)	5.4 (0.5)	0.42
Cholesterol (mg/dL)	184.4 (62.5)	189.2 (62.0)	0.64
LDL (mg/dL)	114.8 (51.0)	121.2 (64.0)	0.48
Alpha1Antitrypsin (mg/dL)	132.7 (24.3)	132.5 (41.5)	0.77

sCeAD: spontaneous cervical artery dissection; NIHSS: National Institutes of Health Stroke Scale; COPD: chronic obstructive pulmonary disease; mRS: modified Rankin Scale; LDL: low-density lipoproteins; IQR: interquartile range.

Bold indicates: level of significance <0.05.

^aN (%).

^bMedian (IQR).

Table 2.

Differences in sCeAD characteristics between those with sCeAD-related vessel steno-occlusion or diameter expansion at baseline.

Baseline—dissection-related	Steno-occlusive sCeAD	Expansive sCeAD	p
	n = 187	n = 34
Minor trauma a	71 (38.0)	8 (23.5)	0.09
Prior respiratory infection (⩽4 weeks) a	23 (12.3)	9 (26.5)	0.03
Anterior circulation sCeAD a	83 (44.4)	15 (44.1)	1.00
Multiple single time-point sCeAD a	24 (12.8)	5 (14.7)	0.78
Symptom to diagnosis (days) b	7.8 (8.0)	17.8 (21.8)	0.02
Duration hospital (days) b	10.9 (8.8)	11.5 (7.0)	0.74
Number of connective tissue stigmata b	0.8 (1.0)	0.8 (1.0)	0.68
Initial anticoagulation a	(n = 166)131 (78.9)	26 (76.5)	0.57
Follow-up
Recurrent dissection a	7 (3.7)	1 (2.9)	0.82
Follow up vessel statusa,c
Normal	(n = 96)63 (65.6)	(n = 17)8 (47.1)	0.13
Stenosis	(n = 96)12 (12.5)	(n = 17)1 (5.9)	0.23
Occlusion	(n = 96)15 (15.6)	(n = 17)4 (23.5)	0.29
Aneurysm	(n = 96)11 (11.5)	(n = 17)7 (41.2)	<0.01
Follow-up medication a
Antiplatelet	(n = 132)78 (59.1)	18 (52.9)	0.46
oAC	(n = 132)4 (3.0)	0 (0.0)	0.35

sCeAD: spontaneous cervical artery dissection; IQR: interquartile range.

^aN (%).

^bMedian (IQR); oAC: Oral anticoagulation.

^cMultiple vessel pathologies per patient possible.

Discussion

To date, studies investigating differences in patient characteristics are limited to those with residual post-sCeAD aneurysms or stenoses/occlusions after the sCeAD-related mural hematoma has already resolved.^8,9 As either steno-occlusive or expansive mural vessel hematoma may be attributed to competing pathomechanisms, it was our goal to investigate whether patient or sCeAD characteristics differed between the two.^3–8 First, it is important to note that within our cohort, expansive sCeAD was notably less frequent than steno-occlusive ones (15.4% vs 84.6%—Table 1). Concerning anthropologic factors, patient history, prior vascular risk factors, laboratory parameters and family history as well as clinical and functional follow-up, those with steno-occlusive sCeAD-related mural hematoma did not differ to those with expansive sCeAD. In terms of clinical presentation, it was interesting to find that not only an understandable difference in the frequency of peripheral cranial nerve palsy could be found but that almost two-thirds of those with solely expansive mural hematoma suffered cerebral ischemia (Table 1). This finding emphasizes that these individuals should not be considered as low-risk vascular patients and treated similarly as those with steno-occlusive pathologies. Still, clinical work-up and identification of expansive sCeAD were more challenging as it took longer to establish the diagnosis in these patients (17.8 vs 7.8 days respectively—Table 1). It is well possible that both physicians and patients do not consider classical local symptoms (e.g. progressive and continuous head/neck pain unresponsive to oral analgesics) as warning sign for urgent further evaluation.^13,16 On the contrary, non-steno-occlusive vessel pathologies can only rarely be demonstrated in extracranial vessel ultrasound, which often is the primary diagnostic modality—especially in neurological practices. The finding that post-sCeAD aneurysm formation was more common upon follow-up in patients with initial vessel diameter expansion fits the hypothesis of a different pathomechanism in these conditions. Even though patients with monogenetic connective tissue disease more frequently have expansive vessel pathologies, clinical signs of such disorders as well as rate of sCeAD-recurrence did not differ between our groups.^10,11 Still, to unequivocally invalidate this connection, more detailed analyses of possible underlying, subclinical connective tissue disorders in patients with expansive sCeAD would be necessary. ¹⁴ Furthermore, infections and inflammation have long been discussed as a potential factor in sCeAD occurrence. ¹⁷ As those with expansive mural hematoma were more likely to have recent respiratory infection prior to sCeAD occurrence (26.5% vs 12.3%—Table 2), it is tempting to speculate that this risk is limited to expansive and therefore subadventitial sCeAD. Limitations of our analysis are the potential of a recall bias as some data depend on the patient history taking at follow-up. Furthermore, as the mural hematoma is subject to change over time, it is possible that initially solely steno-occlusive sCeAD may develop vessel expansion, ensuing allocation bias. Finally, we have not performed a formal grading of vessel stenoses. In summary, when comparing expansive to steno-occlusive mural hematomas of sCeAD, there are understandable differences in clinical presentation, such as the occurrence of peripheral cranial nerve palsies being limited to expansive sCeAD, which correlates to the local mass effect of the mural hematoma. Still, as two-thirds of patients with expansive sCeAD suffered cerebral ischemia, we emphasize that these patients should not be considered as low-risk and subsequently treated differently in clinical routine. Finally, there were no clinical signs pointing toward expansive sCeAD-related mural hematomas being associated with connective tissue disorder or risk of sCeAD recurrence. Therefore, our results do not advise for differential treatment or follow-up measures. As our analysis is exploratory, confirmation and more mechanistic approaches are needed to elucidate potential differences in pathomechanism between both entities.