Predictive Values of Lacunar Transient Ischemic Attacks
Background and Purpose— We postulated that a lacunar syndrome occurring with transient ischemic attacks (TIAs) or progressive nonsudden onset predicts a brain infarction (BI), presumably caused by a small artery disease (ie, lacunar BI) better than a lacunar syndrome with sudden onset.
Methods— We included 510 patients with BI. BI was classified into etiologic groups including lacunar BI group. We identified the patients with lacunar or nonlacunar syndrome, and those with TIAs preceding the BI or with symptoms of nonsudden onset.
Results— Nonlacunar syndrome had a negative predictive value for a lacunar BI of 95%. A lacunar syndrome had a positive predictive value (PPV) of 57% for lacunar infarction (n=109), and the PPV increased to 79% in the case of recent TIAs preceding the lacunar syndrome. Hypertension was present in 95% of cases with lacunar TIAs (odds ratio: 10.69; 95% confidence interval: 1.34 to 84.82; P=0.02).
Conclusions— Lacunar TIAs are almost always associated with history of arterial hypertension and have a high PPV for lacunar BI. This subgroup of patients may reflect different underlying mechanisms than the group of patient with lacunar syndrome of sudden onset.
Lipohyalinosis causes ischemic lacunar infarction by involving small arteries <300 μm.1 Other disorders, such as atherosclerotic disease of the parental artery or very small emboli of various origins, may also occlude a single perforator.2,3 This may lead to infarction of the same size (<15 mm) and therefore the same clinical presentation as lipohyalinosis. Hence, the challenge of clinicians is to presume the underlying disorder on the basis of other clinical–radiological characteristics. They classically include traditional risk factors such as hypertension and diabetes, the presence of multiple lacunes on magnetic resonance imaging (MRI) (probably representing such a widespread disease as lipohyalinosis), the absence of severe stenosis of the homolateral carotid artery, or cardiac and aortic sources of embolism. These underlying disorders causing lacunar stroke have been widely debated and even negated over the past 2 decades.4–6
It has already been shown that lacunar infarction may be preceded by stereotypically repeated clinical events of brief duration, also called lacunar transient ischemic attacks (TIAs), and by symptoms of stepwise, progressive, or abrupt onset.1,7–10 These heterogeneous presentations may account for specific underlying stroke mechanisms of small vessel occlusion such as thrombosis formed locally or vasospasm of the small artery, with transient or permanent distal hemodynamic failure.9,11 MRI may also show hyperintensities of white matter (ie, leukoaraiosis on computed tomography [CT] scan) and état criblé in association with multiple lacunes.12,13 Our hypotheses were that these clinical and radiological characteristics could improve identification of small artery disease (ie, the phenotype of lipohyalinosis) in cases of lacunar syndrome. Our aim was to determine if lacunar syndromes occurring with preceding TIAs or in a progressive manner would predict a brain infarction most likely caused by small artery disease better than lacunar syndromes of sudden onset without TIA. We then looked for particular clinical and radiological characteristics of lacunar infarcts with abrupt onset compared with those preceded by transient, stepwise, or progressive symptoms.
Materials and Methods
Patients in this analysis had cases of brain infarction (BI) and participated in the GENIC (Génétique de l’Infarctus Cérébral) study, a case–control study that examined genetic susceptibility for BI and involved 12 French neurology departments. Details of the protocol have been reported elsewhere.14 The patients with consecutive cases (n=510) were aged from 18 to 85 years and were recruited in the week after the onset of clinical symptoms if they fulfilled the following criteria: (1) clinical symptoms indicative of stroke; (2) no brain hemorrhage on CT scan; (3) infarct proven by MRI; and (4) both parents of white origin. Patients reporting a previous cardiovascular or cerebrovascular history were eligible.
Data Collection, Risk Factor, and TIA
Information on demographic characteristics and risk factors was collected using a structured questionnaire. Hypertension was defined as a history of treated hypertension. Smoking history was coded as never, previous, and current smoker. The subjects were classified as diabetic subjects when treated for insulin-dependent or noninsulin-dependent diabetes. Use of lipid-lowering drugs was assessed. History of myocardial infarction, angioplasty, coronary artery bypass surgery, or lower-limb arterial disease was recorded. History of stroke or TIAs was obtained.
Electrocardiogram (ECG), extracranial duplex, and transcranial Doppler were performed on all cases. The presence of plaques, arterial stenosis, and occlusions was assessed. Transthoracic echocardiography (TTE) results were available for 464 patients (91%) and transesophageal echocardiography (TEE) was performed on 358 (77%). X-ray cerebral angiogram or magnetic resonance cerebral angiogram was performed on 208 patients (41%). Blood was drawn from fasting subjects for DNA extraction and lipid profile determination in 1 centralized laboratory. Low-density lipoprotein (LDL) cholesterol level was calculated using the Friedewald formula.
Conducting the Study
The ethics committee of Hôpital Cochin approved the research protocol. All subjects signed the informed consent form.
Inclusions were reported to the GENIC study coordination center. After inclusion, the study case report form (CRF) had to be completed by the local investigator, including known risk factors, associated diseases, or comorbid conditions; familial history, timing of the qualifying event (including date and hour of symptom onset, arrival at the hospital, first clinical examination and inclusion), symptoms and signs and their course, and antecedent TIAs. We also recorded important variables such as systolic blood pressure, diastolic blood pressure, and antihypertensive and antithrombotic treatments before and after inclusion on day 10 (or at discharge if the patient was discharged before day 10), at 6 months, and at a follow-up home visit by research nurses after 24 months. Other important biological parameters such as fasting glycemia, hematocrit, and platelet count were available.
Study CRFs, obligatory appendices to CRFs (ECG, carotid B-mode ultrasound with its videotape, transthoracic echocardiography, transesophageal echocardiography, frozen blood samples at −80°C, and source files) were all monitored on-site by the clinical research assistant after the inclusion of each patient. They were then reviewed together with the local investigator to answer queries.
An independent external audit has been performed by the research department of Assistance Publique-Hôpitaux de Paris, which had legal responsibility for the study. The result was that the collection of the study data, in accordance with the research protocol, was “extremely sound.”
Clinical Syndrome Classification
After review by one of us (D.H.) of the case report forms, discharge summaries, and follow-up visit reports, the patients were classified into clinical syndromes according to prespecified criteria. Lacunar syndromes were divided into pure motor hemiparesis, pure sensory syndrome, ataxic hemiparesis, dysarthria clumsy hand syndrome, and sensorimotor syndrome. Other clinical syndromes were classified as nonlacunar syndrome (which included atypical lacunar syndromes). Stroke onset was classified as sudden onset and nonsudden onset. Nonsudden onset corresponded to symptoms that increased gradually, neurological symptoms that fluctuated, progressive, stepwise, and stuttering worsening, or symptoms that extended from 1 limb to the other or to the face. The 1-minute cutoff was arbitrary chosen to distinguish symptoms of sudden onset to progressive symptoms.
We differentiated TIAs according to the territory involved and their time course. They were classified as recent when they preceded the brain infarct by ≤1 month, and we noted whether there had been >2 or <2 recurrences before the BI. We distinguished between those attributable to the same territory as the subsequent BI and those located in a different territory. To be included in the analysis, they had to be recent and in the same territory as the subsequent lacunar syndrome. TIAs referring to cortical involvement (eg, aphasia) led to classify the case as nonlacunar BI.
MRI scans were reviewed by 2 neurologists (P.A. and F. Chedru) to determine the anatomical location and extent of the BI, its size in millimeters (the largest diameter was chosen), the territory of the artery or arteries involved, and the presence of old infarcts, with special reference to single and multiple lacunes, “presence of leukoaraiosis,” or “presence of état criblé.” Another review was performed to specifically grade état criblé (P.A. and F.Chedru). État criblé was noted as present or absent, based on T2-weighted images and, if available, fluid-attenuated inversion recovery (FLAIR) sequence, when Wirshof Robin spaces appeared to be dilated with multiple, round, small hypersignals <1 mm on axial sections in the lenticulostriate area and with multiple, long, thin (<1 mm) hypersignals being present in the cortical–subcortical area with a comb-like aspect. When FLAIR was available, these hypersignals had to disappear. Finally, after our prespecified protocol, 2 investigators from the GENIC study (D. Leys and P. Scheltens) read the MRI scans, independently, to grade white matter hyperintensities in the accordance with the Leys and Scheltens scale.15 Scores, depending on side and number of the lesions in the frontal, temporal, parietal, and occipital white matter, were summed in a total score.
Brain Infarction Subtypes Classification
The patients were classified into etiologic subtypes by 2 neurologists (P.A. and F. Chedru) according to prespecified criteria after a review of the clinical files, discharge summaries, follow-up visit reports, and results of investigations. These subtypes were lacunar stroke, nonlacunar stroke, cardioembolic stroke, arterial dissection, other cause, undetermined cause, and unknown cause.
Lacunar stroke (presumably small artery disease) is defined as a small deep infarct measuring <15 mm (MRI) in the appropriate territory in a patient presenting a clinical syndrome compatible with the diagnosis of lacune (either classical lacunar syndrome or atypical lacunar syndrome such as hemichorea, hemiballism, isolated dysarthria, etc) without any finding in favor of an atherothrombotic or cardioembolic stroke (see later).
Nonlacunar strokes were divided into atherothrombotic stroke, defined as ipsilateral internal carotid stenosis ≥30%, or ipsilateral stenosis ≥50% of another intracranial/extracranial artery, or plaques >4 mm in the aortic arch with a mobile component.
Cardioembolic stroke was indicated when a cardiac source was recognized, such as myocardial infarction within the previous 3 weeks, atrial fibrillation, mitral stenosis, cardiac thrombus, valvular vegetations, atrial myxoma, prosthetic mitral or aortic valve, left ventricular aneurysm, or dilated cardiomyopathy.
Arterial dissection was diagnosed in patients with typical clinical–angiographic patterns of carotid or vertebral artery dissection.
Other causes were rare causes such as polycythemia vera, cerebral arteritis, or thrombocythemia.
Undetermined cause was indicated when ≥2 causes as defined previously coexisted in the same individual.
Unknown cause was indicated when no cause was identified. Patients with an isolated elevation of antiphospholipid antibodies, patent foramen ovale, atrial septal aneurysm, valvular strands, mitral valve prolapse, or mitral annulus calcifications belonged to this group.
Comparisons of the clinical and radiological characteristics of patients with and without lacunar infarction were performed by using ANCOVA (continuous variables) and logistic regression (qualitative variables) after adjustment for age and gender. We also examined if these characteristics were significantly different between lacunar infarction (reference group) and other stroke subtype.
Further analyses were performed among the lacunar infarction. In this group, we used Student t tests and χ2 tests (for small numbers, the Fisher exact test was used) to compare clinical and radiological characteristics between patients with progressive onset or those preceded by TIA and patients with sudden onset and no TIA. Odds ratios (ORs) were estimated by using logistic regression.
Concerning the positive predictive value (PPV) for lacunar infarction, 95% confidence intervals (CIs) were estimated by using the exact binomial procedure with continuity corrections.
All statistical testing was performed at the 2-tailed α level of 0.05. The data were analyzed with the SAS package.16
Among the 109 cases (21.4%) classified as lacunar infarction, 52 were identified as pure motor hemiparesis (47.7%), 12 as pure sensory syndrome (11%), 10 as ataxic hemiparesis (9.2%), 2 as dysarthria clumsy hand syndrome (1.8%), 15 as sensorimotor syndrome (13.8%), and 18 as nonlacunar syndrome (16.5%). The baseline characteristics in lacunar and nonlacunar infarction group are shown in Table 1. We found no significant differences between the lacunar and nonlacunar group for sex, body mass index, diabetes, or smoking. Hypertension was as frequent in the lacunar infarction as in the nonlacunar infarction group when considering treated hypertension by history or after the stroke (until 2.7 years on average). LDL cholesterol ≥130 mg/dL and LDL ≥160 mg/dL or lipid-lowering treatment were more frequent in the lacunar group (Table 1). Recent TIA (19.3% versus 8.7%, P=0.002) and nonsudden onset (25.7% versus 15%, P=0.008) were more frequent in the lacunar than in the nonlacunar group, except for atherothrombotic group (Table 2). Patients with lacunar BI had more frequent white matter hyperintensities (approximately corresponding to leukoaraiosis on CT), état criblé, and multiple lacunes (Table 2). Hypertension was not significantly different in patients with single or multiple lacunes on MRI scan (53% versus 68%, P=0.07). Among the 510 cases, 161 patients with lacunar syndrome were identified; 54 (33.5%) of them had TIA or nonsudden onset. PPV of lacunar syndrome according to the presence or absence of clinical characteristics and radiological findings are shown in Table 3. Lacunar syndrome had a PPV of 57% for lacunar infarction. The PPV increased to 79% in the case of recent TIAs preceding the lacunar syndrome and to 69% in the case of nonsudden onset of the lacunar syndrome. A history of hypertension, diabetes, or smoking did not modify the PPV of lacunar syndrome. The PPV at various stages of the clinical diagnosis of a lacunar BI are shown in Figure 1. Nonlacunar syndrome had a negative predictive value of 95%. The characteristics of lacunar infarction with TIA or nonsudden onset (ie, progressive or stepwise onset) and sudden onset lacunar infarction, not preceded by TIA, are shown in Table 4. Current smokers were less frequent in the TIA and nonsudden onset groups. Lacunar infarcts preceded by TIA were more frequently associated with treated hypertension (95% of cases) than those not preceded by TIA, with an OR of 10.69 (95% CI: 1.34 to 84.8; P=0.02) (Table 4).
In this study, we found that the PPV of a lacunar syndrome for lacunar infarction was only 57% and increased to 79% in the case of a previous TIA. The addition of other clinical criteria such as hypertension, systolic blood pressure, and diabetes did not significantly increase the predictive value of lacunar syndrome for a lacunar infarction. The addition of MRI findings (infarction <15 mm in diameter) increased the prediction of a lacunar infarction to 80% and 89%, depending on whether previous TIAs were present, and to 93% in cases of multiple lacunes and recent TIAs. Our findings suggest that some clinical characteristics associated with lacunar syndrome help predict a lacunar infarction caused by a small artery disease (so-called lipohyalinotic phenotype) based on a triple definition including clinical, radiological, and stroke mechanism patterns. In a similar approach, Gan et al found that lacunar syndrome predicts a final diagnosis of lacunar infarction with a PPV of 65%.16 In our study, other radiological parameters strongly associated with the final diagnosis of lacunar infarction were white matter hyperintensities (leukoaraiosis), état criblé, and multiple lacunes. However, we found a somewhat surprising 95% negative predictive value of nonlacunar syndrome for lacunar brain infarction. This may be because of our definition of lacunar brain infarction together with our choice to include only the classical lacunar syndromes. Interestingly, patients with lacunar infarction had more frequent elevated LDL cholesterol than patients with nonlacunar infarction, but their blood pressure and diabetic status were not different, as has frequently been shown.4,5,16 The only subgroup that may have a more frequent hypertensive status is the group of patients with lacunar infarction preceded by TIAs, with 95% of cases being hyperintensive. Interestingly, these patients with lacunar TIAs might be at an extremity of the spectrum of lacunar small artery disease associated with high blood pressure. We speculate that this subgroup of patients with lacunar TIAs may have different vasoreactivity and may be prone to vasospasm. Alternative mechanism could be a hemodynamic compromise beyond a severely stenotic small artery caused by the lipohyalinotic processes associated with a clot, thus explaining the TIA claudication. Further studies of blood pressure monitoring during the TIAs and of cerebrovascular reactivity are required in these patients to improve our understanding of the mechanisms by which lacunar infarction is preceded by TIAs. Leukoaraiosis and état criblé were strongly associated with lacunar infarction (P=0.008 and P<0.0001, respectively), but their frequency did not differ between patients with multiple and single lacunes. This combination of radiological findings probably reflects the presence of an underlying arteriolopathy. One limitation of this study, because it was designed and performed in the 1990s, is the lack of diffusion-weighted imaging (DWI) in each case. It is likely that addition of DWI findings would have increased the PPV >80%, which we found in the present series with MRI without DWI. Future studies should include DWI. It could be argued that the high frequency of TIA preceding the lacunar syndrome in our study reflected a likely embolic mechanism underlying the lacunar small deep infarct. However, this is unlikely because our definition of lacunar small deep infarcts excluded ipsilateral internal carotid artery stenosis >29%, ipsilateral intracranial stenosis >50% (all patients underwent B-mode imaging of carotid artery as well as transcranial Doppler, and up to 40% had an angiogram), and any cardiac or aortic source of embolism (TEE was performed in 77% and TTE in 92% of patients, with a TEE performed in 68% of patients classified as having a lacunar stroke). In conclusion, lacunar syndromes with TIAs or with nonsudden onset have a high PPV for a small artery disease (lipohyalinotic phenotype); lacunar BI preceded by lacunar TIAs almost always seems to be associated with arterial hypertension. These observations may have important implications in patients with lacunar TIAs for the prevention of lacunar stroke with implementation of blood pressure-lowering.
We thank Chantal Nouharet for her secretarial assistance. This study was supported by grants-in-aid from the Fondation CNP pour la Santé, Caisse Nationale d’Assurance Maladie des Travailleurs Salariés (3AM001), Institut National de la Santé et de la Recherche Médicale (INSERM), Programme Hospitalier de Recherche Clinique of the French Ministry of Health (AOA9402), and Sanofi-Winthrop Laboratories. Assistance Publique-Hôpitaux de Paris held legal responsibility for this study (P930902). Association Claude Bernard supported the Formation de Recherche en Neurologie Vasculaire at St. Antoine, Lariboisière, and Bichat hospitals. This study was supported by INSERM and Assistance Publique-Hôpitaux de Paris at the Clinical Investigation Centre of Saint-Antoine University Hospital. Work for this paper was supported by a grant from Association SOS-ATTAQUE CEREBRALE.
The names of the institutions and investigators in the GÉNIC (Étude du profil Génétique de l’Infarctus Cérébral) study are available on our web site: http://www.ccr.jussieu.fr/GENIC/Acknowledgment.html
- Received August 27, 2003.
- Revision received February 10, 2004.
- Accepted February 27, 2004.
Fisher CM. Lacunar strokes and infarcts: a review. Neurology. 1982; 32: 871–876.
Caplan LR. Intracranial branch atheromatous disease: a neglected, understudied, and underused concept. Neurology. 1989; 39: 1246–1250.
Macdonald RL, Kowalczuk A, Johns L. Emboli enter penetrating arteries of monkey brain in relation to their size. Stroke. 1995; 26: 1247–1250.
Boiten J, Lodder J. Lacunar infarcts. Pathogenesis and validity of the clinical syndromes. Stroke. 1991; 22: 1374–1378.
Lodder J, Bamford JM, Sandercock PA, Jones LN , Warlow CP. Are hypertension or cardiac embolism likely causes of lacunar infarction? Stroke. 1990; 21: 375–381.
Kappelle LJ, Koudstaal PJ, van Gijn J, Ramos LM, Keunen JE. Carotid angiography in patients with lacunar infarction. Stroke. 1988; 19: 1093–1096.
Steinke W, Ley SC. Lacunar stroke is the major cause of progressive motor deficits. Stroke. 2002; 33: 1510–1516.
Nakamura K, Saku Y, Ibayashi S, Fujishima M. Progressive motor deficits in lacunar infarction. Neurology. 1999; 52: 29–33.
Donnan GA, O’Malley HM, Quang L, Hurley S, Bladin PF. The capsular warning syndrome: pathogenesis and clinical features. Neurology. 1993; 43: 957–962.
van Swieten JC, van den Hout JH, van Ketel BA, Hijdra A, Wokke JH, van Gijn J. Periventricular lesions in the white matter on magnetic resonance imaging in the elderly. Brain. 1991; 114: 761–774.
Boiten J, Lodder J, Kessels F. Two clinically distinct lacunar infarct entities? A hypothesis. Stroke. 1993; 24: 652–656.
Elbaz A, Poirier O, Canaple S, Chedru F, Cambien F, Amarenco P. The association between the Val34Leu polymorphism in the factor XIII gene and brain infarction. Blood. 2000; 95: 586–591.
Gan R, Sacco RL, Kargman DE, Roberts JK, Boden-Albala B, Gu Q. Testing the validity of the lacunar hypothesis. Neurology. 1997; 48: 1204–1211.