Same-territory chronic infarcts are common in intracranial atherosclerosis and are associated with future recurrence

Abstract

Background

Intracranial atherosclerotic disease (ICAD) has a high risk of symptomatic recurrence. We investigated the prevalence of same-territory chronic infarcts among symptomatic ICAD patients, and their ability to discriminate between absence and presence of future same-territory ischaemic events.

Methods

In a retrospective analysis of prospectively collected data from two Australian stroke centres, patients were followed for recurrent ischaemia over 12 months. Ability to discriminate between absence and presence of recurrent stroke/TIA (Area-under-curve (AUC) in receiver-operating-characteristic analysis) and degree of model fit (Bayesian information criterion (BIC)) were compared for prognostic characteristics of chronic same-territory stroke, >70% stenosis and CT-perfusion hypoperfusion.

Results

From 2019 to 2020, 135 patients had symptomatic ICAD, mean age 73.8 (SD±13.2) and 57/135 (42%) female, of whom 29/135 (21%) had chronic same-territory infarcts. Over a 12-month follow-up period, recurrent cerebrovascular events occurred in 9/29 (31%) patients with, and in 11/106 (10%) without, chronic same-territory stroke (OR 3.9 95% CI 1.4 to 10.6, p=0.008). In discrimination and model fit analysis, BIC indicated strong evidence for improved model fit, for same-territory chronic infarction (AUC 0.64) vs >70% stenosis (AUC 0.58, p=0.59; BIC difference 4.7). In n=99 with CT-perfusion, chronic infarct showed better fit than CT-perfusion hypoperfusion (0.64 vs 0.61, p=0.79; BIC difference 2.8). Chronic infarct alone also outperformed all combinations of chronic infarct, hypoperfusion and >70% stenosis in model fit.

Conclusions

Chronic same-territory infarcts are common among patients presenting with symptomatic ICAD and can better discriminate future same-territory ischaemic cerebrovascular events over a 12-month period than stenosis>70% and hypoperfusion on CTP.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Intracranial atherosclerosis (ICAD) is likely the most common cause of stroke with the highest recurrence rate. Consequently, identifying patients at the greatest risk of secondary events remains a critical challenge in stroke medicine. Previous research into risk stratification has focused on biomarkers that require specialised imaging.

WHAT THIS STUDY ADDS

• This study found that among patients with ICAD, there is an association between established same-territory stroke on routine index imaging and recurrent stroke.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These findings suggest that standard-of-care scans can be used to identify high-risk ICAD patients who may benefit from additional medical or interventional therapies.

Introduction

Recurrent same-territory ischaemia is common in intracranial atherosclerotic disease (ICAD).¹ Beyond intensive medical therapy, there is uncertainty about how to best treat ICAD, given negative randomised trials of intracranial stenting.²

Given the high risk of recurrence and uncertainty regarding treatment, a focus of ICAD research has been on how to best identify patients at particularly high risk of recurrent stroke. ICAD lesions causing >70% stenosis³ and hypoperfusion on CT-perfusion⁴ have been shown to be associated with elevated future risk of ischaemia. Research into computational flow dynamics, transcranial Doppler and phase contrast MRI has demonstrated that these may be useful in risk stratification.⁵ Many of these modalities, however, require additional imaging and expertise that are not available in the routine work-up of stroke patients.

In this study, we aimed to investigate the prevalence of chronic same-territory stroke in patients with stroke secondary to ICAD. We also aimed to investigate the discriminative ability of the presence of chronic same-territory infarction for future cerebrovascular events over a 12-month follow-up period. We hypothesised that the presence of same-territory chronic infarct may improve discriminative ability for future stroke and TIA over a 12-month period compared with other known prognostic characteristics.

Methods

This was a multicentre retrospective analysis of prospectively collected data. Consecutive patients admitted to the Royal Melbourne Hospital and Austin Hospital Stroke Units with ischaemic stroke over a 2-year period (2019–2020) were reviewed. Patients in whom no CT-angiogram was performed were excluded.

Intracranial stenosis was assessed on CT-angiogram. Stenosis was defined as vessel narrowing >30%, as per the Comparison of Warfarin and Aspirin for Symptomatic Intracranial Arterial Stenosis (WASID) trial.⁶ Patients with >70% stenosis were identified using the same criteria. Patients who had a CT perfusion scan were assessed for same-territory hypoperfusion defined as Tmax >6 s as per Yaghi et al⁴ Chronic same-territory infarcts were identified on CT and MRI scans obtained at the time of acute stroke admission. Covert chronic infarct was defined as having chronic infarct on initial imaging but no documentation of previous stroke attributable to that arterial territory. Images were independently reviewed by two blinded, vascular neurologists. Cases of disagreement were adjudicated by a third vascular neurologist.

Clinical information collected from the acute admission included demographic factors; vascular risk factors including hypertension, diabetes, smoking history, hyperlipidaemia; comorbidities including atrial fibrillation, ischaemic heart disease, peripheral vascular disease; and medication use including lipid lowering agents, anticoagulants and antiplatelets. Recurrent cerebrovascular event was defined as stroke or transient ischaemic attack that occurred >24 hours after initial presenting event. Patients were monitored for recurrent events during acute admission and followed up for 12 months by the stroke units, as per routine clinical practice.

Statistical analysis

The analyses were performed for all patients (n=135) and repeated for the subset of patients who had a CT-perfusion scan (n=99). To investigate a potential for selection bias, the baseline and imaging features of patients with CT-perfusion were compared with those without.

Univariate analysis was performed to compare ICAD patients with same-territory chronic stroke and patients without same-territory chronic stroke. Adjustment for confounders was not performed as this was not a causal study. Fisher’s exact test was used to compare categorical variables and Mann-Whitney U test for continuous variables.

Ability to discriminate between absence and presence of recurrent ischaemia within 12 months was investigated using area-under-curve (AUC) in receiver-operating-characteristic (ROC) analysis. Degree of model fit was evaluated using Bayesian information criterion (BIC), where lower values indicate improved model fit. By convention, differences in BIC of 2–6 are regarded as positive, 6–10 as strong and >10 as very strong. AUCs and BICs were compared for prognostic characteristics of >70% stenosis, chronic same-territory stroke and CT-perfusion hypoperfusion, alone and in combination. Association between chronic same-territory stroke and presence of recurrent ischaemia within 12 months was investigated using univariate logistic regression with the magnitudes reported as OR with respective 95% CI.

Statistical analysis was performed using SPSS 29.0.0.0 (IBM Statistics) and Stata 18BE (StataCorp, College Station, Texas, USA).

Results

Of 1535 patients screened, 381 were removed as there was no CT angiogram and 1019 were removed as they did not have intracranial atherosclerosis. 135 patients were included in the final analysis. The most frequent site of disease was the middle cerebral artery, observed in 61 patients (45%). This was followed by the basilar artery in 27 patients (20%), the vertebral artery in 19 (14%), the internal carotid artery in 17 (13%) and the anterior cerebral artery in 11 patients (8%).

Mean (SD) age was 73.8 (SD±13.2) and 57/135 (42%) were female. No significant differences were observed in baseline characteristics between patients with and without prior chronic infarction. Baseline characteristics are summarised in table 1 and figure 1.

Table 1. Baseline characteristics of patients.

	Intracranial atherosclerosis with same-territory chronic infarction	Intracranial atherosclerosis with no same-territory chronic infarction	OR	P value
Age (mean)	73 (SD±12.6)	74 (SD±11.2)	NA	0.98
Sex (female)	11/29 (38%)	46/106 (43%)	1.24 (0.45 to 3.41)	0.827
Hypercholesterolaemia	19/29 (66%)	64/106 (60%)	0.37 (0.13 to 1.08)	0.672
Diabetes	11/29 (38%)	41/106 (39%)	0.42 (0.12 to 1.45)	0.942
Smoking	8/29 (28%)	30/106 (28%)	0.95 (0.32 to 2.94)	0.939
Hypertension	17/29 (59%)	87/106 (82%)	0.56 (0.17 to 1.8)	0.130
Atrial fibrillation	2/29 (7%)	13/106 (12%)	0.53 (0.11 to 2.49)	0.524
Comorbid ischaemic heart disease	5/29 (17%)	23/106 (22%)	1.63 (0.39 to 6.87)	0.797
Comorbid peripheral vascular disease	0/29 (0%)	6/106 (6%)	Not estimable	0.340
Antiplatelet use prior to admission	14/29 (48%)	35/106 (33%)	1.06 (0.37 to 3.09)	0.190
Lipid lowering agent use prior to admission	13/29 (45%)	44/106 (42%)	1.65 (0.57 to 4.72)	0.833
Anticoagulant use prior to admission	0/29 (0%)	8/106 (8%)	Not estimable	0.201
Intracranial atherosclerotic disease lesion >70% stenosis	19/29 (66%)	61/106 (58%)	0.4 (0.13 to 1.12)	0.525

Figure 1. Baseline characteristics of patients with and without same-territory chronic infarction. ICAD, intracranial atherosclerotic disease.

Death within 12 months of presenting with stroke secondary to intracranial atherosclerosis occurred in 27/135 (20%) of patients, with death due to stroke occurring in 16/135 (12%) of patients. 17/135 (13%) patients were lost to follow-up. The remaining 91/135 (67%) were followed up in stroke outpatient clinic for 12 months.

The majority (111/135, 82%) of patients with intracranial atherosclerosis had MRI scans in addition to CT scans, but 79/111 (71%) of MRI scans only included a diffusion-weighted sequence. The majority (99/135, 73%) had CT-perfusion.

Same-territory chronic infarctions that pre-dated the acute presenting infarction occurred in 29/135 (21%) of patients, with 17/29 (59%) of these covert in nature. Recurrent ischaemic stroke occurred in 9/29 (31%, 95% CI 15% to 51%) of symptomatic ICAD patients with same-territory chronic infarction, compared with 11/106 (10%, 95% CI 5% to 18%) of symptomatic ICAD patients without same-territory chronic infarction, OR 3.9 (95%CI 1.4 to 10.6), p=0.008.

No significant differences in percent stenosis >70%, age, chronic infarct and recurrent events were found between the patients with CT perfusion and those without (see online supplemental table 1).

In ROC analysis of all patients (n=135), same-territory chronic infarction had similar AUC and there was evidence of improved model fit compared with >70% stenosis (AUC 0.64 vs 0.58, p=0.59, BIC difference 4.7). Adding chronic infarct to >70% stenosis had similar discriminative ability and improved model fit compared with>70% alone (AUC 0.71 vs 0.58, p=0.08; BIC difference 2.5) and improved discriminative ability vs chronic infarct alone (AUC 0.71 vs 0.64, p=0.04). However, chronic infarct alone was more parsimonious than the combined >70% stenosis and chronic infarct model (BIC difference 2.2). See online supplemental tables 2and 3 for more details.

In ROC analysis of patients with CTP (n=99), chronic infarct had similar AUC vs CTP hypoperfusion (AUC 0.64 vs 0.61, p=0.79) but model fit was better (BIC difference 2.8); combining CTP hypoperfusion and chronic infarct did not significantly improve the model (AUC 0.72 vs 0.64, p=0.07, BIC difference 1.82); but combining CT hypoperfusion, chronic infarct and >70% stenosis improved model discrimination (AUC 0.77 vs 0.64, p=0.04) but not fit (BIC difference 3.99). See online supplemental tables 4 and 5 for more details.

An illustrative patient who had a chronic infarct on initial imaging and then went on to have a disabling future stroke is described in figure 2.

Figure 2. A patient presented with dizziness and ataxia, and subsequently found to have diffuse posterior circulation intracranial atherosclerotic disease (A, B). MRI demonstrated a small acute right cerebellar ischaemic stroke as well as chronic infarction in the left cerebellum (C). Despite intensive medical management, the patient went on to have diffuse bilateral cerebellar infarction 2 weeks after initial presentation (D).

Discussion

We found a high prevalence of chronic infarcts (21%) among patients with acute stroke secondary to intracranial atherosclerosis, of which 59% were covert. Our findings are consistent with the myRIAD study in which 24.7% of ICAD patients had a recurrent infarct on MRI at 6–8 weeks, of which 77% were covert.⁷ These findings show that the burden of recurrent ischaemic events in ICAD is likely higher than reported by studies relying purely on clinical follow-up.⁸ They also highlight the importance of carefully screening for prior infarcts in ICAD patients, even if there is no history of previous stroke.

Screening for chronic infarcts may be important because our study found an almost four times greater odds of future same-territory stroke in patients with previous same-territory infarct. Our findings are consistent with the findings of a sub-study of a previous randomised controlled trial of intracranial stenting.⁹ We compared the presence of same-territory chronic infarct to hypoperfusion on CTP, stenosis >70% and age and found the presence of chronic infarction was more strongly associated with recurrent same-territory infarct. We also found that chronic infarction alone outperformed models that use any combination of same-territory chronic infarction, hypoperfusion on CTP, stenosis >70% and age.

Identifying chronic infarcts may be a simple and effective method to identify patients at elevated risk of future stroke. The advantage of this approach is that no additional imaging beyond standard stroke work-up is required. Additionally, other methods of evaluating risk in ICAD patients that the field is currently researching focus on a particular aspect of the pathophysiology—haemodynamics, emboli or plaque instability. However, these processes likely work in a synergistic way, and focusing on one may not sufficiently describe the underlying pathophysiology. Identifying chronic infarcts may capture the pathophysiology in a more holistic way as this likely represents the summation of multiple pathophysiological processes.

The limitations of this study stem from the fact that data was obtained from routine practice. Follow-up was missing for 13% of patients. Many patients did not receive T1 and FLAIR MRI sequences, considered to be the most sensitive techniques for identifying chronic stroke,⁸ and therefore our rates of chronic stroke and post-admission recurrent stroke were likely underestimated. Another limitation is the relatively small case numbers.

Conclusion

Chronic same-territory infarcts are common among patients presenting with acute stroke secondary to intracranial atherosclerosis. These infarcts are often covert and indicate that ICAD-related recurrent ischaemic event rates are higher than suggested by studies that rely on clinical follow-up. Chronic same-territory infarct may be important to identify as it is associated with recurrent future same-territory stroke. Future studies should investigate chronic infarcts in symptomatic ICAD using more extensive MRI sequences, including high-resolution structural MRI to identify cortical microinfarcts. The usefulness of chronic infarcts in identifying high-risk patients, and whether more intensive management benefits this patient group, should also be further investigated.

Data availability statement

Data are available upon reasonable request.