Article Text
Abstract
Background Information on outcome of patients with occlusion of the internal carotid artery (ICA) is limited by the short duration of follow-up and lack of haemodynamic studies on the brain.
Methods The authors prospectively investigated 117 consecutive patients with transient or moderately disabling cerebral or retinal ischaemia associated with ICA occlusion between September 1995 and July 1998, and followed them until June 2008. The authors determined the risk of recurrent ischaemic stroke and other vascular events and prognostic factors, including collateral pathways and transcranial Doppler CO2 reactivity.
Results Patients (mean age 61±9 years; 80% male) were followed for a median time of 10.2 years; 22 patients underwent endarterectomy for contralateral ICA stenosis and 16 extracranial/intracranial bypass surgery. Recurrent ischaemic stroke occurred in 23 patients, resulting in an annual rate of 2.4% (95% CI 1.5 to 3.6). Risk factors for recurrent ischaemic stroke were age (HR 1.07, 1.02 to 1.13), cerebral rather than retinal symptoms (HR 8.0, 1.1 to 60), recurrent symptoms after documented occlusion (HR 4.4, 1.6 to 12), limb-shaking transient ischaemic attacks at presentation (HR 7.5, 2.6 to 22), history of stroke (HR 2.8, 1.2 to 6.7) and leptomeningeal collaterals (HR 5.2, 1.5 to 17) but not CO2 reactivity (HR 1.01, 0.99 to 1.02). The composite event of any vascular event occurred in 57 patients, resulting in an annual rate of 6.4% (95% CI 4.9 to 8.2).
Conclusion The prognosis of patients with transient ischaemic attack or minor stroke and ICA occlusion depends on age, several clinical factors and the presence of leptomeningeal collaterals. The long-term risk of recurrent ischaemic stroke is much lower than that of other vascular events.
- Carotid artery diseases
- stroke
- follow-up studies
- cerebral blood flow
- cerebrovascular disease
Statistics from Altmetric.com
Introduction
Patients with transient ischaemic attack (TIA) or ischaemic stroke associated with an occlusion of the internal carotid artery (ICA) have a risk of recurrent stroke of approximately 5–6% per year.1 In the subgroup of patients with symptomatic ICA occlusion in whom a compromised haemodynamic state of the brain has been demonstrated, this risk is around 12% per year.2 3 The information on long-term outcome in patients with symptomatic ICA occlusion is limited by the short duration of follow-up2–4 or by a lack of information on the flow state of the brain.5–10 Since the large extracranial/intracranial (EC/IC) bypass trial did not show any benefit of bypass surgery in preventing recurrent stroke in patients with symptomatic ICA occlusion in general,10 the current standard treatment involves antithrombotic medication and management of vascular risk factors. Whether an EC/IC bypass operation or carotid endarterectomy (CEA) of a contralateral ICA stenosis can prevent stroke in a subgroup of patients with symptomatic ICA occlusion is uncertain.3 11–14 We previously reported the short-term outcome of a prospectively collected cohort of patients with symptomatic ICA occlusion.15 The aim of the current follow-up study was to investigate the long-term outcome of this cohort and to study risk factors including haemodynamic characteristics for recurrent ischaemic stroke.
Patients and methods
Patients
We prospectively included 117 consecutive patients with symptoms of transient or at most moderately disabling (modified Rankin scale, mRS ≤3)16 cerebral or retinal ischaemia associated with an ipsilateral ICA occlusion, who were referred to the Department of Neurology at the University Medical Center Utrecht, The Netherlands between September 1995 and July 1998. All patients had had symptoms within 6 months prior to inclusion in the study. The presence of the ICA occlusion was confirmed by digital subtraction angiography showing absence of filling of the extracranial ICA or common carotid artery (CCA). Patients were excluded if the ICA occlusion was caused by dissection or radiation therapy. All patients were interviewed about clinical characteristics and risk factors as listed in table 1, with special attention for clinical characteristics suggesting a haemodynamic cause of symptoms such as limb-shaking,17 precipitation of symptoms by rising from a sitting or lying position, exercise, transfer from a cold to a warm environment, a decrease in blood pressure or retinal claudication.18 Furthermore, we documented whether patients had had any ongoing symptoms after occlusion of the carotid artery had been demonstrated (but before inclusion in the study). All patients underwent MRI or CT scan of the brain. Cerebral infarcts were considered symptomatic if their location corresponded with the patients' symptoms and were classified as territorial, watershed, lacunar (diameter ≤15 mm) or large subcortical.19 The degree of stenosis of the contralateral ICA was measured on the angiograms according to the NASCET criteria.20 The presence of collateral pathways was assessed by the combined information from the angiogram and transcranial Doppler (TCD). Collateral pathways via either the anterior communicating artery or the posterior communicating artery were considered present if either of these showed at least filling of the middle cerebral artery (MCA) branches on the angiogram or if TCD showed reverse flow in the first part of the anterior cerebral artery ipsilateral to the symptomatic ICA occlusion. Reversal of flow in the ipsilateral ophthalmic artery on TCD was considered a sign of collateral circulation via the external carotid artery. Leptomeningeal collaterals were considered present if pial branches from the posterior cerebral artery extending as far as the vascular territory of the MCA or anterior cerebral artery (beyond the usual posterior cerebral artery territory) were visualised on the angiogram after selective catheterisation of one of the vertebral arteries beyond normal variability.21
Patients underwent TCD with measurement of the CO2 reactivity to investigate cerebrovascular reserve capacity. The CO2 reactivity after carbogene inhalation was measured as the relative change in blood-flow velocity in the MCA and expressed as a percentage as described previously.15
Treatment consisted of antithrombotic medication and management of vascular risk factors. Patients with a 70–99% stenosis of the contralateral ICA were offered CEA.
Patients with symptoms of cerebral ischaemia that continued after documentation of the ICA occlusion and evidence of presumably haemodynamic origin (ie, the presence of specific symptoms associated with a haemodynamic cause, a watershed infarct or low CO2 reactivity) were offered treatment with the high-flow EC/IC bypass.22 The study was approved by the institutional review board of the University Medical Center Utrecht, and written informed consent was obtained from all patients.
Follow-up and outcome
Patients were followed at regular intervals in the outpatient clinic until November 1999. Final follow-up information was obtained by structured telephone interviews of the patients in June 2008. If the patient had died, we obtained information of their relatives or their general practitioner. In case of a possible outcome event, we determined the mRS by a standardised questionnaire23 and retrieved medical records and CT or MRI scans of the brain in case of recurrent stroke. The primary outcome was any first recurrent ischaemic stroke defined as the acute onset of new focal neurological deficit of cerebral origin persisting for more than 24 h without haemorrhage on CT or MRI scan of the brain. The secondary outcome event was defined as the composite event of a non-fatal ischaemic or haemorrhagic stroke, non-fatal myocardial infarction (MI) or death due to vascular causes, whichever happened first.15 Vascular death was defined as death from fatal stroke, fatal MI, sudden death, terminal heart failure, systemic bleeding, pulmonary embolism or complications after vascular surgery. Fatal stroke was defined as death within 30 days after stroke or death after major stroke (mRS≥4), in the absence of other clear causes.24 Any death that was not clearly non-vascular, or if no information was available about the cause of death, was classified as ‘other vascular.’25
Data analysis
We determined annual rates of recurrent ischaemic stroke and of the combined outcome of vascular events with the Kaplan–Meier method. The survival analysis started from the time of inclusion in the study. We used a Cox proportional hazards model for univariable analysis of predefined risk factors for the primary and secondary outcome events, resulting in HRs with 95% CIs. The primary outcome recurrent ischaemic stroke was subdivided into any recurrent ischaemic stroke and only ipsilateral ischaemic stroke. Variables with a p value of <0.15 in the univariable analysis were included in a multivariable model. To exclude the influence of surgical treatment on the effect of predefined risk factors on the occurrence of the primary outcome event, we also performed the Cox proportional hazards analysis with censoring of patients at the time of CEA of a contralateral ICA stenosis or at the time of an EC/IC bypass.
Results
In total, we included 117 patients (mean age 61 years, range 35–79, 80% male). Ninety-three (80%) patients had presented with cerebral ischaemic symptoms, whereas 24 (20%) patients had had only retinal ischaemic symptoms (retinal infarction in three, transient monocular blindness in 19 and chronic ocular ischaemic syndrome in five patients; three patients had two of these symptoms). Symptoms with haemodynamic characteristics were present in 16 patients; limb-shaking in nine, retinal claudication in three, symptoms after rising in five, after exercise in three and after low blood pressure in two patients. Seventy-eight (66%) patients had a symptomatic infarct on their CT or MRI, which we classified as territorial in 31 patients, watershed in 32, lacunar in 11 and large subcortical in four patients. Of 117 patients, 40 (34%) patients underwent a surgical revascularisation procedure after a median of 46 days after inclusion (range 3–623 days). Twenty-two patients underwent CEA of a 70–99% contralateral ICA stenosis. One of them also had angioplasty for severe stenosis of the proximal part of the CCA on the side of the occlusion 14 days after the CEA because of recurrent TIAs. One patient underwent stenting of the contralateral ICA 20 months after inclusion in the study because of recurrent TIAs due to a contralateral ICA stenosis that had progressed from moderate to severe (>70%). Another patient underwent endarterectomy of a 90% stenosis of the ECA. Sixteen patients underwent EC/IC bypass surgery. Six patients with a 70–99% stenosis of the contralateral ICA chose not to be operated on, and in another patient CEA was not possible because the stenosis extended to the carotid siphon.
Risk of recurrent ischaemic stroke
The median follow-up time until recurrent ischaemic stroke or death was 10.2 years (range 7 days to 12.8 years). None of the patients was lost to follow-up. Of the 117 patients, 52 (44%) died during follow-up. Any recurrent ischaemic stroke occurred in 23 (20%) patients, resulting in an annual rate of 2.4% (95% CI 1.5 to 3.6). The risk of any recurrent ischaemic stroke was highest in the first 1.5 years after presentation (figure 1) (annual rate 8.0%, 95% CI 4.4 to 13.0). As shown in table 2, 15 of the 23 patients had an ipsilateral ischaemic stroke, whereas eight ischaemic strokes occurred in another vascular territory. Nine of the 23 recurrent ischaemic strokes were major, resulting in an mRS≥4 (of which six were fatal), and 13 were at most moderately disabling strokes (mRS≤3). In one patient, we could not reliably determine the mRS because of comorbidities. In three patients, the (minor) ischaemic stroke occurred during the time they were waiting for the planned operation. Recurrent ischaemic stroke after the intervention occurred in six patients after an EC/IC bypass operation (four within 30 days, one after 8 months and one almost 12 years after the operation), in two patients after CEA of the contralateral ICA (3 and 4 years after surgery) and in one patient immediately after angioplasty of the CCA. Eleven of the 77 (14%) patients who did not undergo a revascularisation procedure had a recurrent ischaemic stroke. Two of them were known to have a contralateral ICA stenosis. The annual rate of recurrent ischaemic stroke in patients who did not undergo a revascularisation operation was 2.1% (95% CI 1.2% to 3.6%).
Risk factors for recurrent ischaemic stroke
Table 1 shows the relationship between baseline characteristics and the recurrence of ischaemic stroke. Patients with cerebral symptoms had an eightfold (age- and sex-adjusted HR 8.0, 95% CI 1.1 to 60) higher risk of recurrent ischaemic stroke than patients with retinal symptoms only. Continuation of symptoms after documented occlusion was associated with a fourfold (HR 4.5, 95% CI 1.6 to 12) higher risk of recurrent ischaemic stroke. Other risk factors were older age (HR 1.07, 95% CI 1.02 to 1.13), the presence of limb-shaking (HR 7.7, 95% CI 2.6 to 22), history of stroke (HR 2.9, 95% CI 1.2 to 6.7) and the presence of leptomeningeal collaterals on angiography (HR 5.1, 95% CI 1.5 to 17). TCD CO2 reactivity was not predictive for the risk of recurrent ischaemic stroke (HR 1.01, 95% CI 0.99 to 1.02). Bilateral ICA occlusion tended to be associated with a lower risk of recurrent ischaemic stroke (HR 0.2, 95% CI 0.02 to 1.3). Because of the relatively small number of recurrent ischaemic strokes, we refrained from further multivariable analysis. When we restricted the analysis to ipsilateral recurrent ischaemic stroke, age- and sex-adjusted risk factors were continuation of symptoms after documented carotid occlusion (HR 7.3, 95% CI 1.7 to 33), limb-shaking (HR 7.8, 95% CI 2.4 to 26) and leptomeningeal collaterals (HR 5.3, 95% CI 1.2 to 23). The analysis with censoring at the time of the first revascularisation procedure showed that the age- and sex-adjusted factors continuation of symptoms after documented carotid occlusion (HR 3.7, 95% CI 1.2 to 11) and older age (HR 1.07, 95% CI 1.00 to 1.13) remained significantly associated with recurrent ischaemic stroke.
Risk of composite outcome of any vascular event
Vascular outcome events occurred in 57 (49%) patients. This included 28 vascular deaths (including six fatal ischaemic strokes, one fatal intracerebral haemorrhage, one fatal MI, six terminal heart failures, 11 sudden deaths, three other vascular causes), 17 non-fatal ischaemic strokes, one non-fatal intracerebral haemorrhage in the basal ganglia and 11 non-fatal MIs. Overall, of the 52 patients who died during follow-up the causes of death were vascular in 40 patients (in 28 vascular death as first event, in 12 vascular death after non-fatal stroke or MI as a first event) and non-vascular in 12 patients (malignant disease in seven, infectious disease in three, renal disease in one and an undiagnosed gradually progressive disease in one patient). The annual rate of the composite vascular endpoint was 6.4% (95% CI 4.9 to 8.2; figure 1). Multivariable analysis showed that age (HR 1.06, 95% CI 1.02 to 1.09; table 3), a history of stroke (HR 1.9, 95% CI 1.1 to 3.5) and a history of ischaemic heart disease (HR 1.8, 95% CI 1.1 to 3.2) were independent risk factors for the occurrence of any vascular event.
Discussion
This 10-year follow-up study shows that patients with TIA or minor ischaemic stroke and ICA occlusion, treated with EC/IC bypass in case of recurrent symptoms of presumably haemodynamic origin or treated with CEA in case of contralateral ICA stenosis, have an annual risk of recurrent ischaemic stroke of 2.4%, with the highest risk in the first 1.5 years after presentation. Elderly patients, those who present with cerebral instead of retinal ischaemic symptoms, patients with limb-shaking, recurrent symptoms after documented occlusion, a history of stroke, and those who have leptomeningeal collaterals on their angiogram have an increased risk of recurrent ischaemic stroke. CO2 reactivity as measured with TCD did not predict recurrence of ischaemic stroke.
We were able to obtain a complete and very long duration of follow-up of a large cohort of well-documented prospectively studied patients. The largest study on outcome in patients with symptomatic ICA occlusion to date is the EC/IC bypass trial,10 in which 423 patients who received best medical treatment were followed for 4.5 years. This study reported an annual rate of ischaemic stroke of 6.3%, but measurements of the haemodynamic state of the brain were not performed. Other prospective follow-up series of patients with symptomatic ICA occlusion had a mean duration of follow-up of at most 4 years and found annual rates of ischaemic stroke between 5 and 10%,3 4 6 26 27 except for one study that found an annual rate of 2%.9 Some of these studies also included patients who presented with major stroke.4 6 The comparatively low recurrent stroke rate that we found can be explained in three ways. First, we showed that patients who present with symptomatic ICA occlusion have the highest risk of recurrent ischaemic stroke in the first 1.5 years, and the risk of stroke is relatively low thereafter. Second, a considerable proportion of the patients in our study had presented with retinal ischaemic symptoms only. Patients with retinal symptoms only have a lower risk of ischaemic stroke than patients who present with cerebral ischaemic symptoms.3 28 Third, in previous decades, medical secondary prevention has improved with the introduction of statins,29 dipyridamole in combination with aspirin30 and more rigorous control of blood pressure.31
In contrast to other studies,26 32 TCD CO2 reactivity in this study was not predictive of recurrent ischaemic stroke. In previous studies, the results may have been confounded by including patients without symptoms as well as patients with symptoms. Asymptomatic patients have a relatively low stroke risk33 and a relatively high cerebrovascular reactivity.26 Similar to our findings after 2 years' follow-up,15 we found that the presence of leptomeningeal collaterals on the angiogram was predictive of recurrent ischaemic stroke. Although the role of these collaterals in the flow state of the brain needs further investigation,34 35 it is suggested that the finding of leptomeningeal collaterals is indicative of haemodynamic compromise.36–38 The risk of the composite endpoint of any vascular event of 6.4% per year is relatively high when compared with the risk of 4% per year in patients with TIA or minor stroke not selected because of carotid disease,30 probably indicating that patients with a symptomatic ICA occlusion have relatively severe generalised vascular disease.
This study was a single-centre study. Patients with a symptomatic ICA occlusion were referred from all over The Netherlands to a tertiary university hospital, and referral bias may have played a role. However, patients with frequent and ongoing symptoms are probably more likely to be referred than those with only one event and no further symptoms. Therefore, the risk of recurrent ischaemic stroke that we report would rather be over- than underestimated.
Another limitation of our study is that it does not reflect true natural history, as we advised CEA in patients with a >70% stenosis of the contralateral ICA and EC/IC bypass operation in selected patients with recurrent symptoms of cerebral ischaemia of presumed haemodynamic origin. Recently, another observational study showed that CEA of the contralateral ICA in 39 patients with symptomatic or asymptomatic ICA occlusion resulted in an improvement in vasomotor reactivity compared with 32 control patients who did not undergo CEA. However, they did not find any beneficial effect of CEA on the rate of recurrent stroke or death in the long-term.14 Although there is no evidence from randomised controlled trials that CEA of a contralateral ICA stenosis or EC/IC bypass operation can prevent stroke, we advised these operations because we estimated the risk of recurrent stroke to be high if patients were treated with the best medical treatment only. Whether this treatment strategy may have prevented ischaemic stroke in some patients remains unclear. In fact, five operated patients suffered a stroke within 30 days of the operation, and four of the 35 patients who were operated on without any complications suffered a recurrent stroke after 5 years on average. Because of the selection bias with regard to treatment, we could not statistically compare outcome in medically and surgically treated patients.
A third limitation is that the follow-up of our patients did not include serial duplex scanning or CT angiography to reconfirm the occlusion after several years. Although infrequently, it has been found that spontaneous recanalisation of an occluded ICA can occur, and a remaining high-grade stenosis would expose a patient to increased risk of an embolic event.39
Despite complete occlusion of the ICA, patients who have not suffered a major stroke at the time of occlusion have a relatively low risk of recurrent stroke, in particular once no stroke has occurred in the first 18 months. While this information is reassuring to the patients, it should be noted that their risk of other vascular events is significant, and meticulous control of vascular risk factors is of key importance. Patients who continue to have symptoms after documentation of the ICA occlusion have a relatively high risk of recurrent ischaemic stroke. Although our study was not designed to evaluate the effectiveness of several treatment strategies, we suggest that, particularly in patients with ongoing symptoms after documented occlusion and a contralateral ICA stenosis, a CEA may be considered. EC/IC bypass surgery is associated with a considerable risk of postoperative stroke, in particular in unstable patients with repeated TIAs.22 Whether or not EC/IC bypass surgery can prevent stroke in patients with demonstrated haemodynamic compromise will be answered by the ongoing Carotid Occlusion Surgery Study.40
References
Footnotes
Funding SP is supported by a grant from The Netherlands Heart Association (grant number 2003B263) and the Foundation ‘De Drie Lichten’ (grant number 41/09).CJMK was supported by a clinical fellowship from The Netherlands Organization for Health Research and Development (grant number 907-00-103).
Competing interests None.
Ethics approval Ethics approval was provided by the institutional review board of the University Medical Center Utrecht.
Provenance and peer review Not commissioned; externally peer reviewed.