Milhaud D, de Freitas GR, van Melle G, Bogousslavsky J. Occlusion Due to Carotid Artery DissectionA More Severe Disease Than Previously Suggested. Arch Neurol. 2002;59(4):557-561. doi:10.1001/archneur.59.4.557
Stroke due to internal carotid artery dissection is considered to have a good prognosis.
To determine whether the prognosis of internal carotid artery dissection is worse than classically reported by comparing the characteristics of patients who had an acute ischemic stroke admitted to a population-based primary care center with internal carotid artery occlusion due to either dissection (DO) or atherothrombosis (AO).
Patients and Methods
Among 3502 patients admitted to our population-based primary care center, DO (n = 73) was diagnosed by angiography or magnetic resonance imaging, while AO (n = 81) was diagnosed by angiography. The characteristics of patients with DO or AO were compared using univariate and multivariate analysis.
Patients with DO were younger (mean [SD] age, 44.6  vs 60.1  years, P<.001), had fewer vascular risk factors, and presented more frequently with global middle cerebral artery territory involvement (42% vs 17%, P<.05) and less frequently with watershed infarcts (3% vs 19%, P<.05) than patients with AO. Unexpectedly, patients with DO were noted to have more severe clinical impairment, with an increased frequency of decreased consciousness, and a poorer outcome at 1 month. Multivariate analysis showed that the independent factors associated with DO were age younger than 55 years, nonsmoker, no history of hypertension, headache at presentation, and global aphasia.
Patients with DO are younger and are initially seen with fewer risk factors than patients with AO, but their clinical features and prognosis are worse. Large infarcts involving the whole middle cerebral artery territory that may be due to the lack of collateral circulation are responsible for the bad prognosis of patients with DO.
INTERNAL CAROTID artery dissection (ICAD) is a clearly identified cause of stroke and accounts for about 20% of strokes in younger patients.1 The clinical picture of ICAD is variable, with either pure local symptoms, such as Horner syndrome associated with headache or cervical pain, or cerebral ischemia in up to 80% of patients with transient ischemic attack or brain infarction.1- 3 Internal carotid artery dissection can be diagnosed using conventional angiography, which shows irregular, and often tapered, stenosis with the characteristic "string sign," "flame-shaped" occlusion, or aneurysmal dilatation.4- 8 More recently, ICAD has been diagnosed using magnetic resonance imaging, which detects a crescent-shaped hyperintensity surrounding the narrowed lumen (T1- and T2-weighted images of the neck), and usingmagnetic resonance angiography.8
Although certain authors consider ICAD to be a serious disease with a potentially grave outcome,9 several studies have shown excellent or good recovery in 70% to 90 % of these patients without significant stroke sequelae.3,4,10,11 The rate of early death in ICAD is directly due to brain infarct and is estimated at about 2% to 5%.3,5,10 However, in a previous study focusing on ICAD with occlusion, we reported mortality as high as 23%,12 and Pozzati et al13 reported that 37% of patients with occlusion and ICAD have a bad outcome. To determine whether the prognosis of ICAD is worse than classically reported, we compared the characteristics of patients with acute ischemic stroke admitted to our population-based primary care center with ICA occlusion due to either dissection (DO) or atherothrombosis (AO).
From the Lausanne Stroke Registry, a prospective primary care center–based registry,14 we selected all consecutive patients initially seen with angiography-proven internal carotid artery occlusion related to DO or AO and with first-ever ischemic stroke. Internal carotid artery dissection with occlusion was diagnosed either using conventional angiography (64 cases), which showed a tapered artery occlusion with wall artery irregularities, double lumen, or outpouching, or, when available, using magnetic resonance angiography and axial magnetic imaging (9 cases), showing wall artery hematoma or intra-arterial thrombus. Traumatic and spontaneous dissections were grouped together because of the difficulty in determining whether a dissection was spontaneous or secondary to a minor trauma and because their presentation did not differ significantly.15 In all 81 cases, the internal carotid artery occlusion type was diagnosed as AO using only conventional angiography. To further clarify the mechanism of brain infarct and the outcome in patients with ICAD, we also studied the infarct territory and outcome in patients with ICAD without occlusion diagnosed using angiography or magnetic resonance imaging and compared the results with those for patients with DO.
The patients were studied using a standard protocol including at least one computed tomographic scan within 7 days of the stroke. We also assessed the risks factors of hypertension (blood pressure >160/90 mm Hg at least twice before the stroke), known diabetes mellitus (2 or more fasting blood glucose levels >117 mg/dL [>6.5 mmol/L] before the stroke), elevated venous hematocrit, regular cigarette smoking, oral contraceptive use, hypercholesterolemia (fasting cholesterol levels >232 mg/dL [>6 mmol/L]), and a history of ischemic heart disease, atrial fibrillation, vascular claudication, or migraine. A history of previous transient ischemic attack in the same territory as the infarct was also recorded.
Infarct topography was classified according to vascular territory using mapping templates developed in our center and elsewhere.16 The arterial territories considered were the anterior cerebral artery; middle cerebral artery (MCA) pial territories, including the anterior and posterior pial territories; the deep territory of the MCA, including perforating and medullary branches; and the global MCA territory involving superficial and deep territories. Multiple, superficial, and deep territories and watershed infarcts (cortical or subcortical) were also considered. Neurological status and type of stroke onset were evaluated on admission by a neurologist. Level of consciousness was evaluated on admission, decreased consciousness being defined as somnolence or coma. Functional status at 1 month was measured by establishing a modified Rankin Scale score from the prospectively collected Lausanne Stroke Registry data, with a poor outcome being considered a Rankin Scale score of 3 or more.
The 2-sided t test was used to compare means. To compare proportions, the Pearson χ2 test was used unless the number of cases was too small, in which case the Fisher exact test was used. The Bonferroni-type correction attributed to Holmes was applied. To further characterize ICAD, we performed a stepwise logistic regression (stepdown) using all clinical factors identified in the univariate analysis as being significantly different between patients with DO or AO. In all tests, P<.05 was considered statistically significant.
Between January 1, 1979, and December 31, 1998, Lausanne Stroke Registry recorded 3873 first-ever stroke patients and, in 3502 of these, the stroke was diagnosed as ischemic in origin. Of this ischemic stroke group, 88 patients had ICAD and, of these, 73 had occlusion. This last group consisted of 35 men and 38 women whose mean (SD) age was 44.6 (10) years. Two patients had bilateral ICAD and occlusion. Fibromuscular dysplasia was diagnosed in 14 patients (19%) and dissection was clearly related to a history of trauma in 7 patients (9.6%). Eighty-one patients (65 men and 16 women; mean [SD] age, 60.1  years) with ischemic stroke had AO proven by conventional angiography. In this group, the contralateral carotid artery was normal in 18 cases and showed atherosclerotic lesions with less than 70% stenosis in 42 cases or with 70% or more stenosis in 16 cases, while internal carotid artery occlusion was bilateral in 5 cases. In the acute phase, 36 (49%) of the patients with DO received anticoagulant therapy using unfractionated heparin sulfate; all other patients with DO (with contraindication for full-dose anticoagulant therapy) and patients with AO were treated with a low dose of low-molecular-weight heparin and aspirin. The DO group contained a significantly higher proportion of women and patients were younger than in the AO group (Table 1).
Risk factors in patients with DO and AO are summarized in Table 1. In patients with DO, the prevalence of hypertension, diabetes mellitus, smoking, hypercholesterolemia, and a history of vascular claudication was significantly lower than in patients with AO.
The global MCA territory was significantly more frequently involved in the DO group than in the AO group (42% vs 17%, P<.05), in which anterior pial MCA territory involvement predominated (Table 2). Patients with DO presented less frequently with watershed infarcts than patients with AO (3% vs 19%, P<.05), suggesting that DO most often caused stroke via embolism. Moreover, by studying angiograms in patients with DO, we documented embolization in the M1- or M2-segments of the MCA in 12 patients and distal embolization in 13 patients. The infarct territories involved in patients with ICAD without occlusion were similar to those in the DO group, with a low rate of watershed infarcts (7%); however, these patients were seen less frequently with global MCA infarcts (20% vs 42%, P = .10).
Horner syndrome (complete or incomplete) was present in 18 of the patients with DO (24%) and in 10 of the patients with AO (12%) (P = .12). Severe clinical impairment, including motor and sensory deficits with lateral hemianopia, was more frequent in patients with DO than in patients with AO (41% vs 14%, P<.01), as was global aphasia (32% vs 16%, P<.05) (Table 3). Motor deficit involving the face, upper limb, and lower limb was more frequent in patients with DO than in patients with AO (62% vs 36%, P<.01). Decreased consciousness was also more frequent in patients with DO than in patients with AO (42% vs 16%, P<.05). Finally, a bad outcome at 1 month was seen in 37% of the patients with DO compared with 16% of the patients with AO (P<.05), and mortality at 1 month was 14% in patients with DO compared with 1% in patients with AO (P<.003). In comparison, at 1 month, 20% of patients with ICAD without occlusion had a bad outcome and 7% died (P = .14 compared with patients with ICAD with occlusion).
For the patients with DO, logistic regression (Table 4) showed that the independent factors characterizing occlusive dissection were age younger than 55 years, nonsmoker, no history of hypertension, headache at presentation, and global aphasia.
In this study, patients with DO fell into the same age range as those reported in other series of patients with ICAD6 and were significantly younger than the patients with AO. There was no clear male-female preponderance in the patients with DO, while the patients with AO presented with male preponderance, as previously reported.17,18 Except for oral contraception, patients with DO had fewer vascular risk factors (hypertension, diabetes mellitus, cigarette smoking, and vascular claudication) than patients with AO. Finally, despite the patients with DO being younger and having fewer vascular risk factors than the patients with AO, their clinical features and outcomes were more severe.
In our study, infarct topography in patients with DO showed that almost all infarcts (97%) were territorial, involving mainly the MCA territory, whereas watershed infarcts were rare (only 2 cases). Cortical and subcortical infarcts were more likely to be of embolic origin,16,19 whereas junctional or watershed infarcts were more likely to be of hemodynamic origin.20 As previously reported for ICAD with stenosis or occlusion, in the present study, the presumed mechanism of cerebral ischemia in the patients with DO was embolic. This point was confirmed by studying the infarct territories involved in patients with ICAD without occlusion; in these patients in whom strokes only occur via embolism, roughly the same low frequency of watershed infarcts was found as in patients with DO. However, in patients with ICAD without occlusion, there was a trend toward a better prognosis than in the patients with DO; this may be due to the lower rate of global MCA infarcts in this group, which may be related to a lower frequency of intraluminal thrombus and smaller emboli.
In the study of Lucas et al,21 92% of infarcts were large cortical and subcortical infarcts that were presumed to be embolic, and, in the series of Steinke et al,22 70% of infarcts were territorial MCA infarcts or large striatocapsular infarcts with a high probability of being due to embolism. However, in 1 study23 including only 11 patients, 5 presented with 1 or more lesions in the rostral corona radiata ipsilateral to the ICAD and were interpreted as low-flow–induced ischemic brain damage. In our series of ICAD with occlusion, the presumed mechanism is the formation of a thrombus in the dissected artery favored by blood stagnation after the level of occlusion, with secondary distal embolism in the MCA territory in most cases. Moreover, 2 studies have shown that microembolic signals are detected by transcranial Doppler in patients with brain infarct due to ICAD.24,25 In our patients with AO, the prevalence of watershed infarcts was higher (16%) than in the patients with DO, the frequency being similar to that seen in another of our studies.26 However, in these patients, the predominant mechanism of ischemia was embolic. In the study of Mounier-Vehier et al,18 of the 40 patients with atherothrombotic internal carotid artery occlusion, only 5 presented with border-zone infarcts.
In the present study, large cortical and subcortical infarcts due to global MCA arterial territory involvement were significantly more frequent in patients with DO than in patients with AO, whereas there was a trend to a lower frequency of deep MCA infarcts in the DO group compared with the AO group. Previous studies have emphasized the importance of the collateral circulation in determining the size of infarcts.27- 29 The extent of brain infarction in patients with DO may be due to poor functioning and mobilization of the secondary collateral leptomeningeal pathways compared with patients with AO. Indeed, in patients with AO, parenchymal chronic hypoperfusion due to previous ipsilateral tight stenosis or contralateral stenosis greater than 50% leads to mobilization of the secondary collateral pathways, which was not the case in the patients with DO in the present study. In contrast to the primary collateral system, which responds immediately to focal circulation failure, the secondary collateral leptomeningeal pathways take longer to respond30- 32 and do not develop in patients with DO. Moreover, because dissection is a dynamic process, we cannot exclude the possibility of an underlying hemodynamic mechanism prior to thrombus formation and embolism in patients with DO, the combination of which contributes to large lesions in the subcortical area.
In the present study, because of the profile of our registry, which only includes patients with first-ever acute ischemic stroke, all patients with DO presented with completed stroke compared with the 40% to 60% of reported cases of ICAD in other series,9 emphasizing limitations in the interpretation of registry-based studies because of the use of data obtained from selected populations. However, the severity of stroke related to ICAD is clearly due to the high prevalence of infarcts involving the global MCA territory.
Although patients with DO are younger and are initially seen with fewer risk factors than patients with AO, their clinical features and prognosis are worse. Larger infarcts involving the global MCA territory, which may be due to the lack of collateral circulation, are responsible for the bad prognosis of patients with DO.
Accepted for publication December 28, 2001.
Author contributions: Study concept and design (Drs Milhaud, de Freitas, and Bogousslavsky); acquisition of data (Dr Bogousslavsky); analysis and interpretation of data (Drs Milhaud, de Freitas, van Melle, and Bogousslavsky); drafting of the manuscript (Drs Milhaud and Bogousslavsky); critical revision of the manuscript for important intellectual content (Drs van Melle and Bogousslavsky); statistical expertise (Dr van Melle); obtained funding (Dr Milhaud); administrative, technical, and material support (Drs Milhaud, de Freitas, and Bogousslavsky); study supervision (Drs Milhaud and Bogousslavsky).
Corresponding author and reprints: Julien Bogousslavsky, MD, Department of Neurology, University Hospital Lausanne, rue du Bugnon 46, Lausanne 1011, Switzerland.