The Course of Patients With Lacunar Infarcts and a Parent Arterial Lesion: Similarities to Large Artery vs Small Artery Disease

Background  The significance of occlusive lesions of the parent artery in patients with lacunar syndrome (LS) and small deep infarcts (SDIs) on diffusion-weighted imaging remains unclear.

Objective  To compare the recurrence of stroke in patients with LS and SDIs between those with vs without a parent arterial lesion.

Design  Analysis of data from a prospective acute stroke registry.

Setting  University hospital.

Patients  Using clinical syndrome, diffusion-weighted imaging, and vascular studies, we divided 173 patients into 3 groups: (1) parent arterial disease occluding deep perforators (PAD), LS with SDIs, and a parent arterial lesion (n = 32); (2) small artery disease (SAD) (n = 70); and (3) large artery disease (LAD) (n = 71).

Main Outcome Measures  Recurrent strokes and the prognosis were registered for 1 year, and the outcome of the PAD group was compared with that of the SAD and LAD groups.

Results  During follow-up, there were 9 deaths (6 vascular) and 18 recurrent strokes. The recurrence rate in the PAD group (16%) was significantly higher than that in the SAD group (1%) (P = .01) but similar to that in the LAD group (17%) (P = .87). The presence of the parent arterial lesion was the only independent predictor of stroke recurrence in patients with LS and SDIs (odds ratio, 13.8; 95% confidence interval, 1.5-123.9; P = .02).

Conclusions  Although LS on examination, SDIs on diffusion-weighted imaging, and a stable hospital course suggest lacunar stroke of benign course, our results indicate that the PAD group represents an intracranial type of LAD.

Lacunar strokes, that is, lacunar syndrome (LS) with symptomatic small deep infarcts (SDIs) on neuroimaging, are associated with low mortality rates and low stroke recurrence rates in hospital- and community-based studies.^1-7 Moreover, intracranial stenosis has been reported to be a negligible cause of SDI.⁸ For these reasons, physicians treating patients with LS and SDIs are apt to be negligent in investigating the underlying cause, such as parent arterial disease (PAD). In fact, in the TOAST (Trial of ORG 10172 in Acute Stroke Treatment) criteria, a widely used classification scheme for ischemic stroke subtypes, patients with traditional LS and ipsilateral intracranial stenosis are classified as having "2 or more causes identified" and are lumped together with patients with an "incomplete or negative evaluation" as having so-called ischemic stroke of unknown cause.⁹

However, middle cerebral artery (MCA) stenosis is recognized increasingly in patients with SDIs.^10,11 Previously, Madden et al¹² reported that approximately one third of patients who were initially diagnosed as having small artery disease (SAD), based on clinical features and baseline computed tomographic findings, had abnormalities on diagnostic study and suggested that a careful evaluation of the etiology is justified in all patients with stroke, regardless of the presumed subtype. In a previous study by our group,¹³ we recently found that atherosclerotic MCA lesions were a common cause of striatocapsular SDIs in Asian patients, and it was reported^14,15 that approximately half of the patients with MCA stenosis also had SDIs. Owing to the high risk of stroke if occlusive lesions are present,¹⁰ the prognosis or recurrence rate in patients with LS and SDIs might depend on the existence of a parent arterial lesion. However, no long-term follow-up data on patients with occlusive lesions of the parent artery, such as the MCA or the basilar artery (BA), are available.

We hypothesized that the course of patients with lacunar stroke (with LS and SDIs on neuroimaging) and an occlusive lesion of the parent artery is similar to that of large artery disease (LAD) rather than that of SAD. Therefore, we compared the long-term prognosis of patients with LS and SDIs in terms of the stroke recurrence rate, type of recurrent stroke, and prognosis and its relation to possible clinical prognostic factors and the existence of a relevant large arterial lesion. Furthermore, we compared these patients with those with non-LS.

Between December 1, 2000, and May 23, 2002, we prospectively studied consecutive patients with ischemic stroke, who had been examined within a week of symptom onset, and those who had a relevant lesion on diffusion-weighted imaging (DWI).

Patients were evaluated according to a protocol that included demographic data, medical history, vascular risk factors, and the National Institutes of Health Stroke Scale at days 0 to 7.¹⁶ We used the same criteria for vascular risk factors as in our previous study¹⁷: hypertension, diabetes mellitus, hypercholesterolemia, and smoking. On day 7 after hospital admission, all the patients were evaluated according to the National Institutes of Health Stroke Scale, the modified Rankin Scale,¹⁸ and the Barthel Index.¹⁹ Laboratory tests included brain magnetic resonance imaging (1.5 T), vascular study, echocardiography, electrocardiography, and routine blood testing. Hemostatic markers of prothrombic tendency, including protein C, protein S, antithrombin III, and antiphospholipid antibodies (lupus anticoagulant and anticardiolipin IgG and IgM) were checked in all patients younger than 50 years.

During the study, 396 consecutive patients with symptomatic ischemic stroke were admitted to the Department of Neurology at Ajou University Hospital (Figure 1). The group included 136 patients with acute symptomatic lacunar infarcts. Patients with acute symptomatic lacunar infarcts were defined as those who were examined within a week of symptom onset who showed focal neurologic deficits lasting longer than 24 hours that were consistent with LS and those who had relevant small deep lesions on DWI. All patients provided informed consent to participate in this study.

Digital subtraction or magnetic resonance angiography, electrocardiography, and transthoracic echocardiography were performed in all patients, and transesophageal echocardiography and Holter monitoring were performed in selected patients, especially when the possibility of a cardioembolic source was suggested or when no other cause of stroke was found. Patients with potential sources of cardioembolism (atrial fibrillation, mitral stenosis, prosthetic valve, myocardial infarction within 6 weeks, intracardiac clot, ventricular aneurysm, and bacterial endocarditis)²⁰ were excluded from the study (n = 4) because a stenotic lesion of an intracranial artery on digital subtraction or magnetic resonance angiography might represent a partially recanalized clot rather than an atherosclerotic lesion. We also excluded (1) patients with other determined causes, such as carotid dissection, hypercoagulable state, vasculitis, and complicated migraine (n = 4); (2) those who did not undergo a complete workup (n = 8); and (3) those who were not followed up for 1 year after the index stroke (n = 18).

Based on their clinical syndrome, infarct size on DWI, and the results of vascular studies, we divided the patients with acute symptomatic lacunar infarcts into 2 groups: (1) patients with LS and SDIs on DWI and no occlusive lesion of the relevant artery (the SAD group) and (2) patients with LS and SDIs on DWI and an occlusive lesion of the parent artery occluding the deep perforators (the PAD group). For comparison, patients with non-LS, a larger lesion on DWI, and an occlusive lesion of the relevant artery (the LAD group)⁹ were included; these lesions were intracranial (MCA, BA, and the intracranial portion of the internal carotid artery or vertebral artery) or extracranial (carotid sinus and extracranial portion of the vertebral artery). Lacunar syndrome was defined as 1 of the 5 classic syndromes: pure motor hemiplegia, pure sensory stroke, sensory-motor stroke, ataxic-hemiparesis, and dysarthria–clumsy hand syndrome^21,22; SDIs as subcortical or pontine hyperdense circular or oval lesions with a diameter of 15 mm or less on DWI²³; and occlusive lesions as stenosis exceeding 50% or occlusion of the large intracranial vessels and internal carotid artery according to the method of the Warfarin-Aspirin Symptomatic Intracranial Disease Study Group²⁴ and the North American Symptomatic Carotid Endarterectomy Trial,²⁵ respectively. Two of us (P.H.L. and U.S.J.), masked to the patient's clinical presentation, interpreted the DWI and angiographic features. Interobserver agreement was 91% for the presence of SDIs or larger infarcts and 85% for the presence of an occlusive lesion on angiography.

Hospital course was determined after 1 week and was defined as follows: improved, when the National Institutes of Health Stroke Scale score decreased 2 points or more; stable, when the score decreased less than 2 points; worsening, when the score increased after admission; and fluctuating, when the score increased and then decreased episodically or vice versa. Patients were evaluated by one of us (O.Y.B.) at the outpatient clinic at 3-month intervals for up to 12 months after stroke onset. During 1-year follow-up, information on the attainment of end points was obtained through direct clinical examination and through information given by family members in cases of death (n = 4). Recurrent stroke was defined as a focal neurologic deficit occurring suddenly in a vascular territory, lasting more than 24 hours, and occurring at any time after the acute phase of the index stroke.^7,26 The type of recurrent stroke was based on clinical and magnetic resonance imaging findings. To define the outcome of patients with stroke, Barthel Index scores during follow-up after the index stroke were divided into 2 classes at a cutoff value of 60, and the modified Rankin Scale scores were divided at a cutoff value of 3.²⁷

Between-group differences were examined using the χ² test, Fisher exact test, unpaired t test, or 1-way analysis of variance with post hoc analysis. Kaplan-Meier curves were used to estimate survival free of recurrent stroke or death. Potentially significant predictors of poor prognosis and stroke-free recurrence after acute symptomatic lacunar stroke (SAD or PAD groups) were evaluated using multiple logistic regression analysis. Initially, the 15 independent variables regarding the clinical and radiologic features were tested for their individual associations with the outcome of interest. Those that were significant at the 0.2 level in the univariate model were entered into the initial multivariate model. When the most parsimonious model was obtained by backward stepwise elimination of the nonsignificant factors, each excluded variable was again entered into the model separately to test its contribution to the final model. Results are given as odds ratios as estimates of relative risk with a 95% confidence interval. Statistical significance was established at P <.05.

The study population consisted of 173 patients. There were 107 men (62%) and 66 women (38%) (mean ± SD age, 62 ± 10 years; age range, 37-85 years). According to the clinical, DWI, and angiographic findings, 70 patients were classified in the SAD group, 32 in the PAD group, and 71 in the LAD group. The patients' clinical characteristics are given in Table 1.

The age, sex ratio, and territory of infarcts did not differ among the groups. The National Institutes of Health Stroke Scale score at hospital admission was significantly higher and the Barthel Index score 7 days after admission was significantly lower in the LAD group than in the other groups (P<.001 for both). However, these indices did not differ between the SAD and PAD groups. The risk factors for stroke and the types of LS did not differ between the SAD and PAD groups. Most patients were discharged from the hospital taking antiplatelet agents, and they showed good compliance (98%).

Angiographic findings of the relevant and nonrelevant vessels are given in Table 2. The angiographic characteristics of the patients in the PAD group included the presence of occlusive lesions of either the MCA or the BA, which were the parent arteries of the small perforators supplying the territory of the DWI lesion. Angiographic stenosis of the relevant intracranial large artery (MCA or BA) was more prevalent in the PAD group (29 patients, 91%) than in the LAD group (36 patients, 51%), whereas a stenotic lesion of the ipsilateral internal carotid artery (either extracranial or intracranial) or vertebral artery was more frequent in the LAD group (35 patients, 49%) than in the PAD group (4 patients, 13%). In the PAD group, the prevalence of stenotic lesions of the large intracranial artery (MCA or BA) was significantly higher at relevant sites than at nonrelevant sites (P <.001).

Unstable clinical courses (fluctuating or worsening) during the first 7 days of admission were significantly more common in the LAD group (22 patients, 31%) than in the other groups (P = .001). However, the frequency of such an unstable temporal profile was similar in the PAD (4 patients, 13%) and SAD (5 patients, 7%) groups (Table 3).

Figure 2 presents the Kaplan-Meier estimates of recurrent stroke or death (Figure 2A) and recurrent stroke (Figure 2B) for the different groups. During follow-up, 18 patients (10%) had at least 1 recurrent stroke (Table 3); recurrent strokes were hemorrhagic in 1 patient and ischemic in 17. The recurrence rate differed among the subtypes (P = .01). The recurrence rate in the PAD group (16%) was higher than that in the SAD group (1%; P = .01) and similar to that in the LAD group (17%; P = .87). The type and site of recurrent stroke were correlated with those of the index stroke. Five of the 32 patients in the PAD group had recurrent stroke: 3 of these 5 patients had an LS as their index stroke, and the recurrent stroke was in the same territory as the index stroke in all these patients. These findings were similar to those for the LAD group: 12 recurrent strokes occurred in 71 patients in the LAD group, most (10 patients, 83%) were non-LS index strokes, and new lesions were located in the same vascular distribution as the index stroke in all but 1 patient. In the SAD group, 1 patient had recurrence of pure motor hemiparesis in the same territory.

For patients with lacunar stroke (the SAD and PAD groups), a multivariate logistic regression analysis of factors obtained on hospital admission was performed, and the results showed that the presence of stenosis of the relevant artery was the only independent predictor of stroke recurrence in patients with LS and SDIs (odds ratio, 13.8; 95% confidence interval, 1.5-123.9; P = .02). Other clinical and radiologic results, including epidemiologic data and risk factors, the severity of neurologic deficits, the presence of stenosis of nonrelevant vessels, and the vascular territory involved, were not associated with an increased risk of stroke recurrence (P >.05).

During the study, there were 9 deaths (5%): 3 were due to different medical complications, such as upper gastrointestinal tract bleeding or infection (2 in the LAD group and 1 in the PAD group); 1 patient each in the LAD and PAD groups had a myocardial infarction; and 4 patients in the LAD group had massive brain swelling.

At the end of 1-year follow-up, the PAD group had more disable states than the SAD group (P = .002). None of the patients in the SAD group had a poor prognosis, that is, Barthel Index scores less than 60 or modified Rankin Scale scores greater than 3, compared with the LAD (20 patients, 28%) and PAD (5 patients, 16%) groups.

This study compared the recurrence and prognosis of stroke patients with classic LS and the presence of SDIs and an occlusive lesion of the relevant intracranial artery with those of SAD and LAD. To our knowledge, this is the first prospective study of a series of patients with classic LS and the presence of an occlusive lesion of the relevant large intracranial artery, evaluated according to a protocol that included DWI, echocardiography, and vascular studies, in which information on stroke recurrence and the prognosis at 1-year follow-up were available.

The role of potential cardioembolic sources, carotid artery disease, and intracranial atherosclerotic disease in the pathogenesis of lacunar stroke is still controversial.^5,11,28-33 Therefore, the presence of a cardiac or carotid embolic source and intracranial stenosis of the relevant artery in patients with lacunar stroke was thought to represent opportunistic disease or to be a marker of diffuse atherosclerosis.²⁸ Recently, no difference was found in the prevalence of ipsilateral and contralateral MCA disease in patients with LS or SDIs.²⁸ However, our data showed that a stenotic lesion of a large intracranial artery was more prevalent on the ipsilateral side than on the contralateral side in the PAD group, indicating a role of these stenotic lesions of large intracranial vessels in the development of LS and SDIs on DWI. Such a difference might arise from the fact that intracranial atherosclerosis is more common in Asians than in Westerners, and it plays an important role in developing LS in Asians. We found a high prevalence of intracranial stenosis of the relevant artery in patients with lacunar stroke: 29 (28%) of 102 patients had stenosis of the relevant intracranial artery, which is higher than in Westerners²⁸ and similar to the rate in Chinese patients.^11,34 Moreover, our findings, in agreement with published data,³⁴ suggest that parent arterial stenosis located adjacent to the orifice of perforators produces a small, deep, lacunalike infarct rather than a nonlacunar-type infarct; angiographic stenosis of the relevant intracranial large artery was more prevalent in the PAD group, and extracranial large artery stenosis was more prevalent in the LAD group.

Several studies^1,6,7 have conducted long-term follow-up of lacunar strokes and report mean annual stroke rates ranging from 4% to 7%. However, these studies did not use the same inclusion criteria, and stenosis of a large intracranial artery was not considered in most of the studies. In the present study, the risk of stroke recurrence after lacunar stroke was 6% (6 of 102 patients with either PAD or SAD) during the first year after the index stroke, which was similar to that reported previously. Although stroke syndrome, SDIs on DWI, and a stable hospital course suggested lacunar stroke of benign course, our data showed a high frequency of recurrent stroke in patients with stenosis of a large intracranial artery. These results, which analyzed independent predictors of stroke recurrence after an index lacunar stroke, confirm this hypothesis. Earlier studies of serial angiography showed that lesion progression occurs frequently in patients with intracranial atherosclerosis (approximately 60%), particularly in medium-sized arteries (MCA or BA),³⁵ and that the risk of stroke is high if associated intracranial stenosis is present.¹⁰ Therefore, the detection of such occlusive lesions in patients with lacunar strokes is of great importance in clinical practice. Combined with the fact that patients with LS and SDIs frequently had an associated occlusive lesion of a large intracranial artery, and these patients had a higher rate of recurrence and poorer outcome, diagnostic and therapeutic approaches different from those for SAD, which has a good prognosis, are needed.

Moreover, the high frequency of intracranial stenosis in patients with LS and SDIs on DWI may also explain the difference in the type and location of recurrent stroke observed in this study: the higher frequency of recurrent stroke with non-LS compared with others,^4,5,7 occurring more frequently in the same vascular distribution of the index stroke compared with others,^5,7 and being associated with significantly increased disability compared with others.⁷ In patients with SAD, the most important vascular pathogenic mechanism seems to be widespread intracranial SAD due to prolonged hypertension, whereas the propagation of thrombi at a preexisting intracranial stenosis may play an important role in recurrent stroke in the PAD group.

In conclusion, although LS on examination, the presence of SDIs on neuroimaging, and a stable hospital course suggest lacunar stroke of benign course, our follow-up data indicate that patients with PAD have an intracranial type of LAD. Owing to the high recurrence rate, physicians dealing with patients with LS should be alert for large intracranial stenosis, especially in populations with a higher frequency of intracranial stenosis. Further studies with more patients and longer follow-up are needed.

Corresponding author: Oh Young Bang, MD, PhD, Department of Neurology, School of Medicine, Ajou University, Woncheon-dong San 5, Paldal-ku, Suwon, Kyungki-do, 442-749, South Korea (e-mail: nmboy@unitel.co.kr).

Accepted for publication November 19, 2003.

Author contributions: Study concept and design (Drs Bang, S. Y. Joo, and Huh); acquisition of data (Drs Bang, S. Y. Joo, U. S. Joo, J. H. Lee, and I. S. Joo); analysis and interpretation of data (Drs Bang, S. Y. Joo, and P. H. Lee); drafting of the manuscript (Drs Bang, S. Y. Joo, P. H. Lee, U. S. Joo, J. H. Lee, and I. S. Joo); critical revision of the manuscript for important intellectual content (Drs Bang and Huh); statistical expertise (Drs Bang, P. H. Lee, and J. H. Lee); administrative, technical, and material support (Drs Bang, P. H. Lee, I. S. Joo, and Huh); study supervision (Drs Bang and Huh).

We thank Woon Ki Paik, MD, PhD, at Hanyang University for his advice and assistance in preparing this manuscript; Se Ho Oh, MD, Kwang Gi Heo, MD, and Jin Soo Lee, MD, for their help throughout this project; the patients and their families; and the physicians and nurses at Ajou University Hospital.