Prognosis of Untreated Strokes Due to

Anterior Circulation Proximal Intracranial

Arterial Occlusions Detected by Use of
Computed Tomography Angiography FREE
ABSTRACT
Importance Limited data exist regarding the natural history of proximal intracranial arterial
occlusions.
Objective To investigate the outcomes of patients who had an acute ischemic stroke attributed
to an anterior circulation proximal intracranial arterial occlusion.
Design, Setting, and Participants A prospective cohort study at 2 university-based hospitals
from 2003 to 2005 in which nonenhanced computed tomography scans and computed
tomography angiograms were obtained at admission of all adult patients suspected of having an
ischemic stroke in the first 24 hours of symptom onset.
Exposure Anterior circulation proximal intracranial arterial occlusion.
Main Outcomes and Measures Frequency of good outcome (defined as a modified Rankin
Scale score of 2) and mortality at 6 months.
Results A total of 126 patients with a unilateral complete occlusion of the intracranial internal
carotid artery (ICA; 26 patients: median National Institutes of Health Stroke Scale [NIHSS]
score, 11 [interquartile range, 5-17]), of the M1 segment of the middle cerebral artery (MCA; 52
patients: median NIHSS score, 13 [interquartile range, 6-16]), or of the M2 segment of the MCA
(48 patients: median NIHSS score, 7 [interquartile range, 4-15]) were included. Of these 3
groups of patients, 10 (38.5%), 20 (38.5%), and 26 (54.2%) with ICA, MCA-M1, and MCA-M2
occlusions, respectively, achieved a modified Rankin Scale score of 2 or less, and 6 (23.1%), 12
(23.1%), and 10 (20.8%) were dead at 6 months. Worse outcomes were seen in patients with a
baseline NIHSS score of 10 or higher, with a modified Rankin Scale score of 2 or less achieved
in only 7.1% (1 of 14), 23.5% (8 of 34), and 22.7% (5 of 22) of patients and mortality rates of
35.7% (5 of 14), 32.4% (11 of 34), and 40.9% (9 of 22) among patients with ICA, MCA-M1, and
MCA-M2 occlusions, respectively. Age (odds ratio, 0.94 [95% CI, 0.91-0.98]), NIHSS score
(odds ratio, 0.73 [95% CI, 0.64-0.83]), and strength of leptomeningeal collaterals (odds ratio,
2.37 [95% CI, 1.08-5.20]) were independently associated with outcome, whereas the level of
proximal intracranial arterial occlusion (ICA vs MCA-M1 vs MCA-M2) was not.
Conclusions and Relevance The natural history of proximal intracranial arterial occlusion is
variable, with poor outcomes overall. Stroke severity and collateral flow appear to be more

important than the level of proximal intracranial arterial occlusion in determining outcomes. Our
results provide useful data for proper patient selection and sample size calculations in the design
of new clinical trials aimed at recanalization therapies.
A proximal intracranial arterial occlusion is an independent factor associated with poor
functional outcomes and high mortality rates in patients with acute ischemic stroke.1- 3 Yet limited
data exist about the natural history of proximal intracranial arterial occlusions. Most of the
available information about the course of this disease comes from large intervention trials that
might have limitations (such as limited generalizability) owing to their intrinsic design.1,2
Currently, the only approved pharmacological therapy for the treatment of acute ischemic stroke
is intravenous (IV) tissue plasminogen activator (tPA) administered within 4.5 hours of symptom
onset. Intra-arterial techniques, including mechanical thrombectomy, are rapidly evolving and
may represent an option for those patients who have contraindications to IV tPA or for those
patients for whom IV tPA is not effective. However, its efficacy remains to be proved in
randomized trials. Further information on the natural history of proximal intracranial arterial
occlusions is essential for the proper design of clinical trials to test the efficacy of endovascular
approaches. In the present study, we sought to establish the rates and predictors of long-term
outcomes of patients who had an acute ischemic stroke attributed to an anterior circulation
proximal intracranial arterial occlusion and who did not undergo any reperfusion therapy.

METHODS
We analyzed data from 741 consecutive patients enrolled in a prospective cohort study at 2
university-based hospitals from 2003 to 2005, the Screening Technology and Outcomes
Project in Stroke (STOPStroke), in which nonenhanced computed tomography (CT) scans
and CT angiograms (CTAs) were obtained at admission of all patients suspected of having
an ischemic stroke in the first 24 hours of symptom onset. Patients were excluded if the
administration of an iodinated contrast agent was contraindicated (ie, history of allergies to
contrast agents, pregnancy, congestive heart failure, or renal insufficiency) or if there was
evidence of intracranial hemorrhage on nonenhanced CT scans. The STOPStroke study
received institutional review board approval at both participating institutions and was compliant
with the Health Insurance Portability and Accountability Act. All participants or their proxies
provided informed consent at enrollment.
Data on clinical history, laboratory results, demographics, and stroke risk factors of all
patients were collected at baseline by direct interview or by review of the medical record by
trained staff. National Institutes of Health Stroke Scale (NIHSS) scores were obtained at
baseline as part of the patient admission workup. Time to hospital arrival was calculated as the
amount of time elapsed between the onset of symptoms (last time seen normal for nonwitnessed
symptom-onset patients) and the time of arrival to the emergency department. Time to CTA was
calculated in a similar fashion. The STOPStroke study was designed to evaluate long-term
outcome of ischemic stroke. Follow-up modified Rankin Scale (mRS) scores were obtained
at 6 months in order to maximize stroke recovery while minimizing loss to follow-up and
nonstroke-related events. For the present study, patients with an acute complete occlusion
of the intracranial internal carotid artery (ICA) and/or of the M1 and/or M2 segments of

the middle cerebral artery (MCA) were selected. Patients with bilateral and/or posterior
circulation strokes, as well as those treated with IV tPA and/or endovascular therapy, were
excluded from the analysis.

Image Protocol and Review

The detailed STOPStroke nonenhanced CT and CTA protocol is described elsewhere. 4,5 All
patients underwent nonenhanced CT, CTA, and postcontrast CT imaging. The CTA images were
reconstructed as thick maximum intensity projection reconstructions. Image review was
independently performed by a board-certified neuroradiologist and a clinical neurologist
experienced in stroke imaging interpretation. Disagreements in readings were resolved by
consensus. Reviewers were blinded to follow-up clinical and imaging findings but had
information with regard to the patients ages, sex, and presenting clinical symptoms. Neither of
the reviewers had participated in the selection of the patients. Variable window width and center
level settings were used for optimal ischemic hypoattenuation detection with nonenhanced CT
and CTA images. In all cases, the nonenhanced CT images obtained for acute stroke were
reviewed first, followed by the CTA images. Reviewers rated the ischemic lesion on the
nonenhanced CT scans according to the Alberta Stroke Program Early Computed Tomography
Score (ASPECTS).6 Readers assessed the presence of complete occlusion of the intracranial ICA
and MCA by thrombus after reviewing both the CTA source images and the thick maximum
intensity projection reconstructions. The CTA source images were used to determine the presence
of collateral vessels in the region of the leptomeningeal convexity. Contrast in leptomeningeal
vessels distal to the occlusion was scored as 1 (absent), 2 (less than the contralateral unaffected
side), 3 (equal to the contralateral unaffected side), 4 (more than the contralateral unaffected
side), and 5 (exuberant). Because of the very small number of patients with extreme scores, the
scale was collapsed into 3 ordinal groups: less than contralateral unaffected side (scores 1-2),
equal to contralateral unaffected side (score 3), and greater than contralateral unaffected side
(scores 4-5).
Statistical Analysis

Continuous variables were reported as mean (standard deviation) or median (interquartile range
[IQR]) values. Categorical variables were reported as proportions. A good clinical outcome was
defined as having an mRS score of 2 or lower at 6 months of follow-up. Clinical and
neuroimaging characteristics were compared according to the site of anterior circulation
proximal intracranial arterial occlusion. Differences in continuous variables were assessed by 1way analysis of variance and by use of the Kruskal-Wallis test in the case of nonnormally
distributed data. Differences between proportions were assessed by use of the Fisher exact test or
the 2 test when appropriate.
We described rates of good clinical outcome and mortality according to the most proximal site of
anterior circulation proximal intracranial arterial occlusion. The same analysis was performed for
the selection of patients with baseline NIHSS scores of 10 or higher. Univariable analysis was
used to test the association between different variables and the follow-up mRS scores.
Differences in continuous variables were assessed by use of the independent samples t test or the
Mann-Whitney U test in the case of nonnormally distributed data. Differences between

proportions were assessed by use of the 2 test. A logistic regression model was used to identify
independent predictors of outcome. Given their known strong association with outcome, the
variables of age, baseline NIHSS score, sex, ASPECTS, and site of intracranial occlusion were
selected a priori and forced into the final model. Other variables were selected based on their
association with the outcome in the univariable analysis. Those with P .10 in the univariate
analysis were included in the multivariable model and were selected using a backward
elimination process (P .10 for elimination). After the final model was obtained, the pattern of
leptomeningeal collaterals was included to verify its effect on the models fit. The site of
intracranial occlusion and the pattern of leptomeningeal collaterals were tested as ordinal
variables in the regression analysis. The Hosmer-Lemeshow test was used to assess the goodness
of fit of the models. To test whether the inclusion of the pattern of leptomeningeal collaterals
improved the model, we used the log-likelihood ratio test. To test the possible effect of those
patients lost to follow-up on the results, we performed a sensitivity analysis. The last known
mRS score was carried forward and used as the final outcome score. The same analysis for the
multivariable logistic regression modeling was performed. A 2-sided P<.05 was considered to
be statistically significant. All statistical analyses were performed using SPSS software version
17.0 (SPSS Inc).

RESULTS
Unilateral anterior circulation proximal intracranial arterial occlusion was identified in 215
patients. Fifty-one patients (23.7%) were treated with IV tPA, 10 patients (4.6%) were treated
with endovascular therapy, and 16 patients (7.4%) received both IV tPA and endovascular
treatment. Treated patients and untreated patients were similar with regard to age (mean [SD]
age, 68 [16] years for treated patients vs 71 [17] years for untreated patients; P=.19), sex (40%
of treated male patients vs 47% of untreated male patients; P=.37), and ASPECTS (median
[IQR] score, 8 [7-9] for treated patients vs 8 [5-10] for untreated patients; P=.93), but treated
patients had higher NIHSS scores at baseline (median [IQR] score, 16 [12-19] vs 11 [5-16]; P
<.001) and shorter times from stroke onset to hospital arrival (median [IQR] time, 1 hour [1-3
hours] vs 4 hour [2-12 hours]; P <.001) than did untreated patients. Of the 138 untreated
patients, 12 (8.7%) were excluded owing to a lack of outcome data at 6 months. The remaining
126 patients met the inclusion criteria for the study.
Among these 126 patients, the mean (SD) age was 67.9 (16.5) years, the median (IQR)
admission NIHSS score was 11 (5-16), 31 (24.6%) had minor strokes (NIHSS score of <5), 100
(79.4%) were white, and 51 (40.5%) were male patients. Twenty-six (20.6%), 52 (41.3%), and
48 (38.1%) patients had occlusion of the intracranial ICA, MCA-M1, and MCA-M2,
respectively. Additional baseline clinical and imaging characteristics are shown in Table 1.
Table 1. Baseline Characteristics of Patients With Anterior Circulation Proximal Intracranial
Arterial Occlusion

View Large | Save Table | Download Slide (.ppt)

The 12 patients who were lost to follow-up had a higher mean (SD) age, 74 (13.1) years, a lower
median admission NIHSS score, 7 (IQR, 1-12), and a higher median ASPECTS, 9.5 (IQR, 9-10).
Comparable distributions of the sites of intracranial occlusions were observed with 3 (25.0%), 5
(41.7%), and 4 (33.3%) of the 12 patients lost to follow-up with occlusions of the intracranial
ICA, MCA-M1, and MCA-M2, respectively.
As expected, the patients with a more proximal occlusion had lower ASPECTSs on nonenhanced
CT scans at admission (with median scores of 7, 7, and 8, respectively; P=.01) despite similar
presentation times. Fourteen of 26 patients with intracranial ICA (53.8%), 34 of 52 patients with
MCA-M1 (65.4%), and 22 of 48 patients with MCA-M2 (45.8%) occlusions presented with
baseline NIHSS scores of 10 or higher. Additional clinical and imaging characteristics according
to the most proximal site of occlusion are shown in Table 2.
Table 2. Clinical and Imaging Characteristics According to the Most Proximal Site of Occlusion

View Large | Save Table | Download Slide (.ppt)

The median 6-month mRS score of the study cohort was 3 (IQR, 1-5). The rate of good outcome
(6-month mRS score, 0-2) was 44.4% (56 of 126 patients) in the overall group and 38.5% (10 of
26), 38.5% (20 of 52), and 54.2% (26 of 48) in the group of patients with ICA, MCA-M1, and
MCA-M2 occlusions, respectively. The good outcome rates were notably lower for patients with
NIHSS scores of 10 or higher at admission (7.1% [1 of 14 patients], 23.5% [8 of 34 patients],
and 22.7% [5 of 22 patients], respectively) than for patients with NIHSS scores of less than 10
(75.0% [9 of 12 patients], 66.7% [12 of 18 patients], and 80.8% [21 of 26 patients], respectively)
(P<.01, determined by use of the 2 test). Rates of good outcome according to the site of
intracranial occlusion and other factors associated with outcome are shown in Table 3.
Table 3. Univariable Analysis of Predictors of Good Functional Outcome at 6 Monthsa

View Large | Save Table | Download Slide (.ppt)

The 6-month mortality rate was 22.2% (28 of 126 patients) in the overall group and 23.1% (6 of
26 patients), 23.1% (12 of 52 patients), and 20.8% (10 of 48 patients) in the group of patients
with ICA, MCA-M1, and MCA-M2 occlusions, respectively. Most of fatalities (19 of 28
[67.9%]) occurred during the hospitalization. The mortality rates were considerably higher for
patients with NIHSS scores of 10 or higher at admission (35.7% [5 of 14 patients], 32.4% [11 of
34 patients], and 40.9% [9 of 22 patients], respectively) than for patients with NIHSS scores of
less than 10 at admission (8.3% [1 of 12 patients], 5.6% [1 of 18 patients], and 3.8% [1 of 26
patients]) (P=.17, .04, and .003, respectively, determined by use of the Fisher exact test).
In the logistic regression model, only younger age (odds ratio, 0.94 [95% CI, 0.91-0.98]) and
lower baseline NIHSS score (odds ratio, 0.73 [95% CI, 0.64-0.83]) were independently
associated with a good outcome (Table 4). There was a trend for lower glucose levels at
admission to be associated with better outcomes (P=.08). The model provided an adequate fit
for the data (P =.84, determined by use of the Hosmer-Lemeshow test). In the second model,
besides age and baseline NIHSS, a favorable pattern of leptomeningeal collaterals was also
independently associated with good outcome (odds ratio, 2.37 [95% CI, 1.08-5.20]) but did not
change the effect of the other variables. The second model also provided an adequate fit for the
data (P=.87, determined by use of the Hosmer-Lemeshow test). Notably, the level of proximal
intracranial arterial occlusion (ICA vs MCA-M1 vs MCA-M2) was not independently associated
with outcomes. Including the pattern of leptomeningeal collaterals improved the model fit (P
=.01, determined by use of the log-likelihood ratio test) ( Table 4). Carrying forward the last
known mRS score for those patients who were lost to follow-up at 6 months, as a sensitivity
analysis, essentially did not change the results from the previous analysis.
Table 4. Multivariable Analysis of Predictors of Good Outcome (Logistic Regression Models)

View Large | Save Table | Download Slide (.ppt)

DISCUSSION
In the present study, using a large population of patients with anterior circulation proximal
intracranial arterial occlusion determined by CTA, we found that more than a half of the patients
(56%) did not achieve functional independence and that almost one-fourth (22%) of those

patients were dead at 6 months, with most of the fatalities (67.9%) occurring during
hospitalization. Patients with intracranial ICA and MCA-M1 occlusions had a lower frequency of
functional independence than did patients with MCA-M2 occlusions (38% vs 54%). A
particularly unfavorable behavior was observed in patients with an NIHSS score of 10 or higher
at admission.
A previous analysis of the STOPStroke database, including both treated and untreated patients,
showed that the presence of proximal intracranial arterial occlusion was associated with poor
outcomes in acute ischemic stroke.3 Several factors might be implicated. These patients
frequently have larger ischemic lesions at presentation, and their infarcts are more likely to grow
over time compared with patients without proximal intracranial arterial occlusion. The low
frequency of good outcomes found in our study is comparable to the control arm of previous
trials. In the Middle Cerebral Artery Embolism Local Fibrinolytic Intervention Trial, 2 only 38.6%
of the patients in the control arm (with a median NIHSS score of 14) achieved good functional
outcome. In the Prolyse in Acute Cerebral Thromboembolism II trial, 1 only 25% of patients in
the control arm (with a median NIHSS score of 17) achieved good functional outcome. In the
placebo arm of the Echoplanar Imaging Thrombolytic Evaluation Trial (with a median NIHSS
score of 10),7 only 40% achieved good functional outcome (with 23% and 47% of patients with
good outcome with ICA and MCA occlusions, respectively). In the Mechanical Retrieval and
Recanalization of Stroke Clots Using Embolectomy trial, 8 only 26% and 10% of the patients in
the penumbral (with a median NIHSS score of 16) and nonpenumbral (with a median NIHSS
score of 20.5) control arms achieved good functional outcome, respectively.
Although patients with more distal occlusions presented with lower NIHSS scores, this was not
statistically significant, probably because of the great overlap of values among the 3 groups.
Also, patients with more distal occlusions had higher ASPECTSs. Interestingly, approximately
half of the patients with a proximal intracranial arterial occlusion presented with ASPECTSs
greater than 7 despite a relatively late time to CT imaging (median time, 6 hours [IQR, 3-16
hours]). Variations in the degree of collateral flow might explain, at least in part, the pleomorphic
clinical presentation of strokes due to proximal intracranial arterial occlusions and why a
significant proportion of those patients could be potential candidates for reperfusion trials.4
Previous multivariable models have consistently shown the importance of age and admission
NIHSS score as independent variables in the prognosis of ischemic stroke. 3,9 Consistent with
these observations, in our final model, age and baseline NIHSS score were the only independent
prognostic markers for patients who had an acute ischemic stroke attributable to a proximal
intracranial arterial occlusion. Other studies3,4,10 have also shown the importance of neuroimaging
markers as important prognostic factors in the acute phase of ischemic stroke. The degree of
collateralization was associated with a favorable outcome that was independent of other
important clinical and imaging variables, improving the models fit. However, including
information about collaterals in the model did not influence the effect of other variables (no
confounding effect), probably owing to the lack of association between the degree of
collateralization and the other clinical variables. It is possible that the relatively small number of
patients might have contributed to the lack of statistical significance of some imaging markers
like the ASPECTS and the site of intracranial occlusion, which were previously shown to have
prognostic value in acute ischemic stroke.

Although there is incontestable evidence about the benefit of IV tPA for acute stroke, the low
frequency of recanalization in patients who present with ICA or MCA-M1 occlusions is also
known.11 Previous studies12,13 have demonstrated a strong association between recanalization and
outcome in acute ischemic stroke. When we consider the increasing use of endovascular
therapies for acute ischemic stroke that specifically target those lesions, the appropriate
knowledge of their natural history is of paramount importance in the treatment decisions and in
the design of new clinical trials.
To date, only 1 randomized controlled trial1 showed the superiority of endovascular treatment
over medical treatment. Recently published trials on endovascular treatment failed to show the
superiority of the endovascular approach over IV tPA.14 The reasons that might explain those
failures are the selection of patients too good to be treated and the relatively low rates of
recanalization achieved with old mechanical devices and techniques compared with the new
ones. Higher rates of recanalization and improved functional outcomes with the new stent
retrievers were demonstrated in 2 recently published randomized trials. 15,16 Recanalization status
was not assessed in the present study. Also, delayed time to treatment might have also played an
important role because collateral failure tends to develop over time. Because CTA provides only
a static picture of the cerebral vasculature (without providing any information about flow), the
inclusion of perfusion imaging in the analysis might have added valuable additional information.
Our study has some limitations. Because the data were collected from 2 tertiary care centers, it
might not necessarily reflect the natural history of the greater patient population who present to
smaller centers. Also, patients who received IV tPA and endovascular treatment were excluded
from this analysis, which might bias our results toward worse outcomes. However, for the
duration of our study, endovascular stroke therapy was not commonly offered in 1 of the 2
institutions and was only typically used in the early time window (ie, the first 6-8 hours) in the
other one as reflected by the overall low number of treated patients (12% of patients with
proximal intracranial arterial occlusions). In addition, our results are comparable to the control
arm of previous trials; no significant differences in important clinical variables such as age, sex,
and ASPECTS were observed, and, on the contrary, treated patients presented with higher
NIHSS scores.
On the other hand, the prospective nature of our study contributed to increasing the quality of the
data collection while limiting the potential for misclassification. The use of CTA, which provides
a rapid assessment of the intracranial and extracranial vasculature with high accuracy and is
widely available, makes our results more robust and generalizable to smaller centers with limited
access to other methods such as magnetic resonance imaging.

CONCLUSIONS
Our results provide useful information on the natural history of anterior circulation proximal
intracranial arterial occlusions, highlighting their variable behavior (depending on the patients
initial NIHSS score and age and on the imaging features) and their overall poor prognosis. This
type of data is essential for proper selection of patients and for the calculation of sample sizes in
the design of new clinical trials focused on recanalization therapies.