Early neurologic deterioration with symptomatic isolated internal carotid artery occlusion: a cohort study, systematic review, and meta-analysis

Abstract

Background

Acute endovascular revascularization for isolated internal carotid occlusion without tandem intracranial occlusion has been proposed to prevent early neurologic deterioration (END) and improve outcome, but has not been shown to be more effective than medical therapy. We aimed to evaluate prognosis with initial medical therapy alone, and also performed a systematic review to put these results in a broader context.

Methods

We performed a retrospective cohort study of patients admitted over a 2-year period with acute stroke/TIA due to isolated internal carotid artery occlusion. Subjects with tandem intracranial occlusion or ASPECTS≤5 were excluded. The primary outcome was END within 48 hours (NIHSS increase ≥4 persisting for ≥24 hours). Secondary outcomes included discharge NIHSS and disposition. We also performed a systematic review and meta-analysis of published studies along with the data from our cohort.

Results

Twenty-three patients met our inclusion criteria. Median age was 69 years, initial ASPECTS 10, and NIHSS score 3. END attributed to recurrent ischemia occurred in 5/23 patients (22%, 95%CI: 7-44%). At discharge, 78% had a favorable outcome with a median NIHSS of 2 (IQR 1-3). END appeared more frequent in those with higher baseline NIHSS. In our systematic review, 7 prior studies met our inclusion criteria. END occurred in 17% (95%CI:12-23%) of patients, 18% with medical therapy versus 13% with endovascular therapy, with substantial heterogeneity among studies.

Conclusions

In patients with acute stroke or TIA due to isolated internal carotid occlusion, END is relatively common (occurring in about 1 out of 6 patients). Further research is needed to evaluate the roles of maximal medical management or acute endovascular thrombectomy in these patients.

Graphical Abstract

Introduction

Symptomatic internal carotid artery occlusion (ICAO) can present with symptoms ranging from transient neurologic deficits to devastating hemispheric stroke. Further, these patients are at high risk for recurrent stroke.¹ When they present acutely with disabling symptoms, concomitant intracranial occlusion, and favorable brain imaging, current standard treatment is endovascular thrombectomy (EVT). However, the role of EVT is less clear for those with isolated ICAO without tandem intracranial artery occlusion, particularly if the initial symptoms are relatively mild. In these cases, collateral supply may be sufficient to avoid major ischemia, and there is considerable uncertainty as to whether the potential advantages of intervention, such as avoidance of recurrent or progressive ischemia, outweigh the procedural risks of distal embolization and/or vessel injury.

Currently, prospective studies in this area are lacking. However, retrospective studies suggest that EVT in selected patients with isolated ICAO may be safe and associated with low risk of recurrent ischemia.^2-4 In one case series of 107 patients with isolated ICAO treated with EVT at a single center, the rate of early neurologic deterioration (END) was 13%, 83% of these patients had improved or stable National Institute of Health Stroke Scale (NIHSS) scores at discharge, and 65% achieved independence at 3 months.⁵

At our center patients with ICAO and without acutely disabling deficits or concomitant acute intracranial occlusion are initially managed medically. We aimed to assess the clinical course and outcomes of consecutive patients with symptomatic isolated ICAO following this conservative strategy. We also performed a systematic review of the literature to put these results in a broader context.

Methods

Cohort study

This was a single center retrospective cohort study. After institutional review board approval of the study and waiving of informed consent, we searched our radiology database and Get with the Guidelines (GWTG)-Stroke database to identify all patients admitted between 7/1/2017-7/1/2019 who satisfied our inclusion criteria: age >18 years, admission diagnosis of acute stroke or TIA, vessel imaging confirming 100% occlusion of common or internal carotid artery (by CT, MR or digital subtraction angiography), absence of tandem intracranial large vessel occlusion, and Alberta Stroke Program Early CT Score (ASPECTS) >5.

We collected baseline information including demographic characteristics, vascular risk factors, stroke clinical and imaging features, and initial treatment. Medical therapy was at the discretion of the clinical team. The primary outcome was onset of END within 48 hours of admission, defined as NIHSS score increase ≥4 points persisting for ≥24 hours. This was further categorized as due to new/extended infarction, brain edema, hemorrhage, or other. Secondary outcomes included discharge NIHSS and discharge disposition. The data that supports the findings of this study are available upon reasonable request from the corresponding author.

We planned a priori for continuous and categorical group comparisons using non parametric tests (such as Kruskal Wallis and Wilcoxon rank sum) and Pearson’s χ2 (or Fisher’s test), respectively. Continuous data are presented as medians (IQR) and categorical data as percentages. Analysis was performed using Stata 15.0.

Systematic review

We searched the PubMed database for all relevant studies from inception until March 2021 using the following terms: ("Stroke"[MeSH Terms] OR "Stroke"[All Fields] or "TIA"[MeSH terms] or "TIA"[all fields]) AND ("Carotid"[All Fields] OR "Carotid"[MeSH Terms]) AND "Occlusion"[All Fields] AND "Isolated"[All Fields]). Then we reviewed the references of all eligible articles to find other potential studies. Studies were selected if they included patients with isolated symptomatic carotid occlusion. They were also required to have reported either END, recurrent ischemic events, or any clinical worsening attributed to the stroke within 10 days. Since not all studies used the same definition for neurologic decline, we pooled these events as evidence of END. Either medical management, EVT, or both was permitted.

1Two reviewers (OK and NN) independently conducted the same search strategy to identify eligible articles and extracted the outcome data of interest. Disagreements were resolved by discussion. If agreement could not be reached, a third senior reviewer (SK) was involved for a final decision. Data collected included: Author name, year of publication, type of study, number of patients, time from symptom onset to presentation, type of treatment, admission diagnosis, NIHSS cutoff, primary outcome of neurologic decline, and timeframe of decline. No formal assessment for risk of bias was used since the available literature was almost entirely comprised small cohorts with a high risk of bias.

We performed a meta-analysis of these studies along with the data from our present cohort. Quantitative data synthesis and forest plot construction were performed using the meta R package in R version 4.0.2 (R Foundation for Statistical Computing).⁶ Meta-analysis was conducted using a random intercept logistic regression model, which is a recommended approach for meta-analysis of single proportions.⁷ Confidence intervals for individual studies were derived based on the Clopper-Pearson method. Statistical heterogeneity between studies was expressed in terms of the Cochran Q and I² statistics. An Egger test was performed to assess for publication bias.

Results

Cohort study

We identified 253 potential subjects; after detailed review of medical records and imaging results, 24 patients met our inclusion criteria. Of these, 1 patient with a major hemispheric syndrome received EVT on presentation, while the other 23 were treated with initial medical therapy. These 23 patients were included in the final analysis of outcome with medical therapy (Figure 1). Baseline characteristics are shown in table 1. Of the 23 patients, 7 also had other competing mechanisms besides large artery atherosclerosis (e.g. high risk source of cardioembolism or hypercoagulable state).

Figure 1:

Flow diagram of search terms and criteria used to query GWTG and hospital radiology database in order to identify eligible patients.

GWTG: Get with the guidelines; ASPECTS: Alberta stroke program early CT score; TIA: Transient ischemic attack

Table 1:

Baseline medical and clinical characteristics of patients with isolated carotid occlusion.

	Total	END
	N=23	N=5
Baseline demographics N (%)
Age, median (IQR) (years)	69 (63-77)	75 (67-84)
Female sex	11 (48)	2 (40)
Right side symptomatic	10 (43)	3 (60)
Presenting as TIA	5 (22)	0
Hypertension	20 (87)	5 (100)
Hyperlipidemia	16 (70)	3 (60)
Coronary artery disease	7 (30)	2 (40)
Diabetes mellitus	7 (30)	1 (20)
Active tobacco use	8 (35)	0
Atrial fibrillation	8 (35)	1 (20)
Congestive heart failure	5 (22)	1 (20)
Chronic kidney disease	8 (35)	2 (40)
Received tPA	2 (9)	2 (40)
Medication use at presentation N (%)
Statins	12 (52)	3 (60)
Antiplatelets	9 (39)	2 (40)
Anticoagulation	4 (17)	0
Clinical/Imaging admission data, median (IQR)
Admission ASPECTS (range, 0-10)	10 (9-10)	9 (8-9)
Admission NIHSS Score (range, 0-32)	3 (2-7)	8 (7-13)
Time from last known normal to presentation (hours)	7 (2.5-24)*	4 (1.5-12)
CTP core volume (mL)	0†	0 (0-13)
CTP mismatch volume (mL)	69 (56-145) †	61 (56.5-154.5)

^*1 value missing.

^†10 values missing. END: Early neurologic decline; tPA: Tissue plasminogen activator; ASPECTS: Alberta stroke program early CT score; NIHSS: National institute of health stroke scale; CTP: CT perfusion, cerebral blood flow <30% and 6 second cut offs were used for core volume and mismatch volume calculations, respectively.

Medical management involved antithrombotic therapy, tissue plasminogen activator (tPA, 2 patients), and blood pressure augmentation (3 patients). Antithrombotic use involved dual antiplatelets therapy (DAPT) in 11 patients, single antiplatelet therapy in 6 patients, and anticoagulation in 6 patients. Indications for anticoagulation included atrial fibrillation, mechanical heart valve, concomitant malignancy, and in 2 cases, treating physician preference.

The primary outcome of END occurred in 5 patients (22%, 95% CI: 7–44%), with median time from admission to END of 17.2 hours. None of the patients treated with blood pressure augmentation had END. END occurred in both patients who were treated with tPA, 2/6 treated with single antiplatelet therapy, and 1/11 treated with DAPT. Three patients had confirmed extension of infarct on repeat imaging and the other two had a sustained worsening without evidence of new or extended infarct (only one eventually returned to baseline). Only one of the five patients experienced major worsening from mild weakness to a full hemispheric syndrome requiring a rescue EVT (which was successful and resulted in minimal deficits). Of the other 4 patients who had END, none had repeat vessel imaging done.

END was detected in 4/10 patients (40%) with admission NIHSS>4 compared to 1/13 patients (8%) with NIHSS≤4 (p=0.13). Similarly, the median baseline NIHSS was higher in patients with END compared to those without (8 vs 3). Moreover, the median last known well to presentation was numerically lower in those with END that those without (4 hrs vs 8.9 hrs, p = 0.31). None of the 5 patients who presented initially as a TIA had END. The stroke mechanism was reported as large artery atherothromboembolism in 4/5 patients and atrial fibrillation in 1/5 with END.

The median discharge NIHSS of the total sample was 2 (IQR 1-3), and 18/23 patients (78%) had stable or improved NIHSS on discharge. Favorable discharge disposition was observed in 87% of patients, including 12 patients (52%) discharged home and 8 (35%) to acute/inpatient rehab facility.

Systematic review and meta-analysis

We identified 7 articles that met our inclusion criteria (Figure 2); data extracted from these articles are summarized in Table 2. Five prior studies plus our cohort reported outcomes in 415 patients treated with medical management, and two studies reported outcomes in 116 patients who underwent EVT. END occurred in 17% (95%CI: 12-23%) of patients, 18% (95% CI: 12-26%) with medical management and 13% (95% CI: 8-20%) with EVT (Figure 3). There was substantial heterogeneity among studies and overlap between these two treatment strategies. Risk of bias among studies was high given inherent design methods. There was no evidence of major publication bias (p=0.60).

Figure 2:

Flow diagram outlining literature review search results

Table 2:

Summary of the available literature by study. Most studies included patients who were medically managed only. Two studies only included patients that were initially managed with endovascular therapy

Author	Type of study	Number of patients	Time from symptom onset	Treatment details	Patients included	Requirement for enrollment	Outcomes reported	Worsening within (days)
Damania, 20158	Retrospective cohort	33	<24 hrs	Medical management	IS/TIA	Mild symptoms not eligible for tPA/EVT	24.3% had recurrent stroke	7
Jadhav, 20185	Retrospective cohort	107	37 patients <24 hrs; 70 patients >24 hrs	EVT	IS/recurrent TIA/Progressive symptoms	No NIHSS cutoff	14% with END and 28% with new or enlarged infarct	2
Mazya, 20199	Subgroup analysis of prospective cohort (SITS-MOST)	96	<4.5 hrs	Medical management (tPA only)	IS	NIHSS≤5	16.7% nhEND	1
Hause, 20191	Retrospective cohort	133	Median 0.6 days	Medical management	IS/TIA	No NIHSS cutoff	7.5% had recurrent ischemic event	9
Castro-Afonso, 201910	Retrospective cohort	9	≤31 hrs	EVT (5 successful)	IS	No NIHSS cutoff	None had worsening NIHSS at discharge	NA
Boulenoir, 202011	Retrospective cohort	74	≤4.5 hrs	Medical management +/− EVT	IS	NIHSS≤5	30% had END	1
Schiphorst, 202012	Retrospective cohort	56	≤24 hrs	Medical management	IS	No NIHSS cutoff	20% had END	7
Present study	Retrospective cohort	23	Mean 7 hrs	Medical management	IS/TIA	No NIHSS cutoff	22% had END	2

EVT – Endovascular therapy; END – Early neurologic decline (NIHSS increase ≥4); nhEND – Non-hemorrhage related END; IS: Ischemic stroke; TIA – Transient ischemic attack; NIHSS – National institute of health stroke scale; IV tPA – Intravenous tissue plasminogen activator

Figure 3:

Systematic review and meta-analysis of the 7 identified studies in addition to the current study. We compared proportion of END in patients who were initially managed medically to those who were managed initially managed with endovascular therapy

Discussion

In our single center cohort study, END was seen in 22% of patients with isolated ICAO who were managed medically. Patients with END were older, had lower ASPECTS, and higher NIHSS scores on admission. The rate of END seen in our cohort is consistent with the reported range (8-30%) in other studies examining medically managed patients. This broad range of frequency of END likely reflects variability in study inclusion criteria and design. Our cohort also included patients with low NIHSS scores (median 3), which even with an intracranial large vessel occlusion may have not been eligible for mechanical thrombectomy.

The largest identified cohort of medically treated patients ICAO to date, a subgroup analysis from the MINOR-STROKE study examining patients with isolated ICAO treated with tPA, found that 22/74 (30%) had END.¹¹ Supra-bulbar carotid occlusion was the only predictor of END on admission. Interestingly, in 76% of the patients who had END the attributed mechanism was an intracranial occlusion. The authors postulated that this was likely due to the underlying mechanism of carotid occlusion, being that non-chronic, non-atherosclerotic occlusions are more likely to lyse with tPA and embolize distally. In our cohort, the mechanism of decline was presumed to be due to collateral failure. Repeat vessel imaging was done in only 1 patient who did not have a stump embolism. The 2 patients who received tPA also had END. We were unable to look at predictors of END in our cohort given our small sample size and relatively low number of events. Notably, we could not adequately assess the role of time from onset to presentation, and it is possible that those who present early, including those given tPA, are more likely to have clinical instability. Nevertheless, our study adds to the available evidence of this uncommon phenotype.

Medical management for ICAO was highly variable in both our study and across studies included in the systematic review, and management during the time periods of the included studies may not reflect current medical management. Indeed, there is now strong evidence for a benefit of dual antiplatelet therapy (DAPT) in patients with minor stroke or TIA due to large artery atherosclerosis,^13-14 with a significant reduction in early stroke recurrence. Such advances in medical therapy over time complicate the use of historical comparisons to estimate outcomes with medical therapy alone and compare this to outcomes seen with EVT. On the other hand, revascularization procedures too may have improved with time, resulting in greater vessel patency and decreased complications. While no significant conclusions can be drawn from this, none of the patients who were management with blood pressure augmentation had END.

Interpretation of prior studies examining END and recurrent ischemia in acutely symptomatic patients with isolated ICAO is also challenging due to variable time windows from symptom onset, patient selection criteria, inclusion of tandem occlusions, antithrombotic treatments, and outcome measures. Our systematic review suggested a risk of END in about 17% of patients. END was numerically lower in the EVT group (13%), however, this was largely driven by 1 study, and END rates with both approaches were overlapping (i.e. no significant difference). Moreover, data regarding neurologic decline secondary to endovascular therapy are also heterogeneous. Jadhav et al reported a 22% distal embolization rate and one case of intracranial perforation.⁵

Limitations of our cohort study include the retrospective nature of the analysis, data abstraction from electronic health records, lack of long-term outcomes, lack of a direct control group, single center, selection bias, and low power. However, our study adds to the available evidence of this uncommon phenotype. We systematically identified all consecutive stroke cases with isolated ICAO and for our pre-specified target cohort almost none were offered EVT, irrespective of perceived risk of END, arguing against a major effect of selection bias on our observed outcomes. Limitations of our systematic review include the inherent limitations of small observational studies, including potential biases related to patient selection, unblinded outcome assessments, and incomplete data, along with variable definitions of END and unexplained heterogeneity among the studies. Nevertheless, the systematic review provides a summary of the currently available data.

Conclusion

In patients with ICAO and stroke or TIA, END is relatively common (occurring in about 1 out of 6 patients). Those who present with stroke rather than TIA appear to be at greater risk for recurrent ischemia. Further research is needed to select the optimal therapy for individual patients with ICAO.

Clinical Perspective:

1) What is new?

We performed a single center retrospective clinical study of 23 consecutive patients presenting with isolated ICAO and present the rate of early neurologic decline (END). We also conducted a systematic review and meta-analysis (7 studies included) outlining the rate of END in patients who were initially managed medically vs with endovascular intervention.

2) What are the clinical implications?

The results of our study provide an initial frame work for an RCT comparing both treatment strategies. END is relatively common in this patient population and interpretation of prior studies is challenging due to variable time windows from symptom onset, patient selection criteria, inclusion of tandem occlusions, antithrombotic treatments, and outcome measures.

Non-standard Abbreviations and Acronyms:

END

Early neurologic decline

NIHSS

National Institute of Health Stroke Scale

ASPECTS

Alberta Stroke Program Early CT Score