Lacunar infarcts contribute to nearly 20% (12% to

Transcription

1 25 Longterm Prognosis of Symptomatic Lacunar Infarcts A HospitalBased Study Isabelle Clavier, MD; Marc Hommel, MD; Gerard Besson, MD; Benedicte elle, MD; J.E. Ferjus Perret, MD Background and Purpose This study concerns the longterm prognosis of lacunar infarcts. Methods We report the analysis of our hospitalbased series of 178 patients consecutively admitted for a lacunar syndrome due to a lacunar infarct diagnosed with computed tomography and magnetic resonance imaging. Demographic data, medical history, vascular risk factors, and imaging data were recorded for each patient. The followup was 35 ±22 months. Results The lacunar syndrome was pure motor hemiparesis in 69 patients (39%), ataxic hemiparesis in 45 patients (25.4%), pure sensory stroke in 15 patients (8.5%), sensorimotor stroke in 14 patients (7.9%), and miscellaneous syndrome in 34 patients (19.2%). The 4year survival rate was 8±4% and the 4year survival rate without recurrent stroke was 85 ±3.5%. Using Cox proportionalhazards analysis, the predictors of death were age (P<.2), diabetes mellitus (P<.5), and cigarette smoking (P<.5). We did not find any predictors of recurrence. After 1 year, 74% of the patients had mild or no disability. Using logistic regression analysis, the predictive factors of disability were age more than 7 years (P<.1), diabetes (P<.1), history of stroke or transient ischemic attack (P<.5), and type of lacunar syndrome (P<.1). Imaging data, number of lacunes, and presence of leukoaraiosis were not predictors of outcome. Conclusions Our study suggests that with a high survival rate, a low recurrence rate, and a relatively good functional recovery, lacunar infarcts have a relatively favorable prognosis. (Stroke. 1994;25:2529.) Key Words lacunar infarction prognosis survival stroke Lacunar infarcts contribute to nearly 2% (12% to 24%) of all strokes. 17 Previous studies have described the prognosis of lacunar infarcts 1 " 6 ' 8 and have shown low early case fatality and recurrent stroke rates compared with other subgroups of strokes. However, there are only a few series reporting the longterm prognosis of lacunar infarcts. 3 6 Using multivariate analysis, we studied the prognostic factors in a series of 178 patients with symptomatic lacunar infarcts admitted consecutively to the stroke unit and followed up over an average of 35 ±22 months. Subjects and Methods Patients Between June 1985 and December 1991, all 178 patients admitted consecutively to the stroke unit for a symptomatic lacunar infarct diagnosed by a neurologist with clinical examination, computed tomography (CT), and magnetic resonance imaging (MRI) were included in the study. The method for collecting patients and analysis of the images was reported previously. 9 We have distinguished the classic lacunar syndromes with either proportional or partial pure motor stroke, pure sensory stroke, ataxic hemiparesis including dysarthria, clumsy hand syndrome, and sensorimotor stroke, 1 ' 1 11 and the miscellaneous syndromes described by Fisher Demo Received March 14, 1994; final revision received July 7, 1994; accepted July 9, From the Stroke Unit, Clinique Neurologique, Department of Clinical and Biological Neurosciences, Grenoble, France. Correspondence to Dr Marc Hommel, Stroke Unit, Clinique Neurologique, Department of Clinical and Biological Neurosciences, BP , Grenoble Cedex, France American Heart Association, Inc. graphic data, medical history, and vascular risk factors were collected. Definitions of risk factors are reported in the "Appendix." MR images were analyzed independently by two observers unaware of patient status. Only concordance between the two observers was taken into account. We collected the number of lacunes defined with MRI as hyperintense, signal present in the first and second echo images, with sharp margins <2.5 cm in diameter, and located in the deeper parts of the brain supplied by the perforating arteries. To test the role of microangiopathy in outcome, we separated our population into patients with 1 lacune and patients with >2 lacunes. Presence of leukoaraiosis was assessed with the scale of Zimmerman et al 14 and rated present if a score >3 was obtained. Followup Since December 1991, all 178 patients have been called back to the outpatient clinic to be examined clinically. If a patient did not attend followup, data were collected by mail or by telephone from his general practitioner. When no response was obtained, a postal inquiry was made to the town hall of birth to collect information concerning occurrence and date of death. All followup was performed by a neurologist. When a patient had died, date and, if possible, cause of the death were recorded. When a recurrence had taken place, the date and type of the recurrent stroke were noted. Functional ability was assessed with the Barthel Index. Statistical Analysis Discrete data were compared using the x 1 test ar d continuous data were compared using the Student's / test. Kaplan Meier estimates were used to determine total survival and eventfree survival (defined as absence of recurrent stroke). The univariate association between patient characteristics and survival was evaluated with the logrank test. Potentially

2 26 Stroke Vol 25, 1 October 1994 significant independent predictors of death or stroke recurrence were evaluated with Cox proportionalhazards regression. We performed a backward stepwise selection of variables. The assumption of proportional hazards for the final Cox model was verified graphically, and interactions were tested. In subgroups for which there were <15% (n=27), patients were not considered in the univariate and multivariate analyses for binary variables. We evaluated the effect of baseline characteristics on longterm functional ability with stepwise multivariate logistic regression analysis. For this study, we took into account only the patients followed for at least 1 year, for whom functional capacity could be considered stable. Thus, this multivariate analysis was performed in the 14 patients whose lacunar infarct occurred between 1985 and The population was separated into two groups, patients having a normal Barthel Index (Barthel=1; group 1) and patients having a residual disability (Barthel <1; group 2). We considered values of P<.5 in bidirectional testing as indicating statistical significance. For the survival analysis, univariate comparisons, and logistic regression, we used SPSS/ PC +14. The receiver operating characteristics (ROC) curve was drawn for the model obtained with multivariate logistic regression analysis. Results General Data Between June 1985 and December 1991,178 patients participated in the study. We obtained the followup data for a period averaging 35 ±22 months (ranging from 23 days date of an early death up to 81 months) in 172 of the 178 patients (96.6%). One hundred twelve patients (63%) were examined in hospital outpatient clinics. For 33 patients (18.5%), data were obtained from the general practitioners. Thirty patients had died. Six patients (3%) dropped out of followup, and we obtained vital information by administrative inquiry for all except one patient, who was a foreigner and for whom we have no data. Therefore, we could determine a vital status for 177 (99.5%) patients. The characteristics of the population are reported in Table 1. Mortality Thirty patients died: 1 from cardiovascular diseases, 9 from cancer, 3 from various other causes, and 8 from undetermined causes. The mean age of these 3 patients was 73 ±12 years. After the qualifying lacunar infarct the survival rate decreased steadily to reach 9±2.5% at 2 years and 8±4% at 4 years (Fig 1). Univariate predictors of death were diabetes mellitus (P=.48), cardiomegaly (P=.46), and age over 7 years at the onset of the stroke (P=.25). The other baseline data, sex, type of lacunar syndromes, number of lacunes, or presence of leukoaraiosis, were not significant univariate predictors of death (Table 2). In the multivariate stepwise Cox regression analysis, the baseline variables predicting death were diabetes, age over 7 years, and cigarette smoking, with an interaction between age and cigarette smoking. It is likely that cardiomegaly was not taken into account in the multivariate analysis because there was comorbidity with diabetes mellitus (cardiomegaly was significantly more frequent in patients with diabetes than without: P=.O186). On the other hand, the survival curve of cigarettesmoking patients did not differ significantly from that of nonsmokers (P=.17). However, the age of the smokers (63±11 years) was lower than that of nonsmokers (68±13 years; P=.O112). It is likely, there TABLE 1. Sex Male Female Risk factors Hypertension Dyslipidemia Baseline Characteristics of the 177 Patients Cigarette smoking. of Patients Frequency % (95% Cl) 62.7 (55.57.) 37.3 (3.44.5) 75.7 ( ) 64.4 ( ) 45.8( ) Cardiomegaly Arrhythmia Lower limb claudication Left valvular diseases History of ischemic heart disease History of stroke or TIA Lacunar syndromes Pure motor hemiparesis Ataxic hemiparesis Pure sensory stroke Sensorimotor stroke Miscellaneous syndromes Imaging data Number of lacunes >2 Leukoaraiosis ( ) 36.9 ( ) 6.2 (2.79.7) 14.8 (9.52.1) 13.(8.18.) 15.8( ) 26.6(2.33.1) 39. ( ) 25.4( ) 8.5 ( ) 7.9 ( ) 19.2 ( ) 36 (2943.) 24(183) Cl indicates confidence interval; TIA, transient ischemic attack. fore, that cigarette smoking was a predictor of death when adjusted for age (Table 3). Imaging variables were not taken into account in the multivariate model. Recurrent Stroke During followup, 16 of the 172 patients had at least one recurrent stroke (3 had two recurrent strokes); 13 patients had a CT scan or an MRI after this recurrence. There were four lacunar infarcts, five cortical infarcts, two transient ischemic attacks, and two intracerebral so Survival (% ± s.d.) I 1 t i Years after the lacunar infarct FIG 1. KaplanMeier survival curve after a lacunar infarct. I

3 Clavier et al Prognosis of Symptomatic Lacunar Infarcts 27 TABLE 2. Markers of Death or Recurrent Stroke: Bivariate Comparisons (LogRank Test) Age >7 y Cardiomegaly Sex Cigarette smoking Ischemic heart disease Hypertension Oyslipidemia History of stroke Type of lacunar syndrome Number of lacunes >2 Leukoaraiosis Deaths (P) (n=3) Recurrences (P) (n=16) hemorrhages. In 2 patients, CT scans showed no new abnormal image; however, the clinical features suggested largeartery infarction. The second recurrent stroke was a cortical infarct in 1 patient, an intracerebral hemorrhage in 1 patient, and unknown cause in 1 patient. The survival rate without recurrence decreased steadily over the 6 years after the lacunar infarct. It was 94±2% at 2 years and 85±3.5% at 4 years (Fig 2). Bivariate and multivariate analyses of the clinical and imaging baseline data did not show any predictor of recurrent stroke (Table 2). Functional Recovery The proportion of patients with a Barthel Index equal to 1 at 1 year or more was 74%. In their daily lives, 8% of the patients needed minor help (8 < Barthel Index <95) and 18% needed considerable help (Barthel Index <8). Multivariate logistic regression analysis showed that the main independent predictors of disability were age (P=.OO86), diabetes (P=.O1O6), history of stroke or transient ischemic attack (P=.O4O3), and type of lacunar syndrome (pure motor hemiparesis, P=.OO13; sensorimotor stroke, P=.O15O) (Table 4). Imaging data had no influence in bivariate and multivariate analyses. The ROC curve is shown in Fig 3. Discussion In our study, the survival rate decreased progressively over the 6 years after lacunar infarct by 5% per year. TABLE 3. Estimates From Cox Model Predicting Death Coefficient (±SD) Age <7 years 2.11 ±.78 Cigarette smoking 2.1 ± ±.39 Hazard Ratio (95% Cl) 8.28 (1.838) (1.634) (.612).5 Age/cigarette smoking* 1.97±.94.14(.2.88).5 Cl indicates confidence interval. Each binary variable was coded if absent, 1 if present. *Age/cigarette smoking indicates interaction between age and cigarette smoking. 8 4 Survival without recurrent stroke (% ± s.d.) Years after the lacunar infarct FIG 2. KaplanMeier survival curve without recurrent stroke after a lacunar infarct. This survival rate is very close to those observed in other series: 9% in Oxford, 1 97% in Rochester, 4 and 87% in Dijon. 5 6 However, concerning the longterm prognosis of lacunar infarct, the data are scant because there are only a few studies, thus making comparisons more difficult. At 5 years the survival rate was 76±4.5% in our study, 75 ±11% in Rochester, 4 and about 65% in the hospitalbased study carried out by Gandolfo et al. 3 This relatively favorable shortterm prognosis, when compared with the prognosis of other subgroups of strokes, may be explained by a combination of the low fatality rate directly due to lacunar infarct itself and the low occurrence of intercurrent diseases due to better preservation of functional abilities. The main explanation for the differences observed in the longterm prognosis between the series is probably less likely to lie in the mean age of the patients, which was very close throughout the series, than in the types of lacunar infarcts included in the studies. The lacunar subtypes were very similar in our series and in Rochester 4 ; TABLE 4. Estimates From Logistic Regression Model Predicting Functional Handicap Age <7y >7y History of stroke or TIA Pure motor hemiparesis Sensorimotor stroke Coefficient (±SD) 1.17± ± ± ± ±.82 TIA indicates transient ischemic attack. Relative Risk P <.1 <.1 <.O5 <.1 <.2

4 28 Stroke Vol 25, 1 October " 7 I «& so 2 S 4 & « sensltlvlty (%) FIG 3. Receiver operating characteristics curve (logistic regression model for analysis of residual disability). however, Gandolfo et al 3 reported a high proportion of pseudobulbar palsies. In our study, cardiovascular diseases explained one third of the deaths occurring during the followup period. Among these, only one death was directly due to a recurrent stroke and another to a complication after a recurrent stroke. Gandolfo et al 3 reported a higher proportion of deaths (73.4%) due to cardiovascular diseases during the 7year followup period. Among these, seven were due to fatal strokes. Among 26 recorded deaths in patients with lacunar infarction, Sacco et al 4 reported 9 deaths (35%) due to heart disease and 3 deaths (12%) due to recurrent stroke. In our study, cancer, which was the second cause of death (3%), appeared to be higher than in other studies (13% for Gandolfo et al 3 ). Nevertheless, the high proportion of deaths of undetermined cause may be at the origin of some discrepancies between the series. Concerning prediction of death, the differences between the methods used in the studies may be at the origin of differences in the results. In particular, some studies were population based and others were hospital based. The clinical inclusion and exclusion criteria were different (inclusion after the firstever strokes or after recurrence, inclusion of miscellaneous syndromes, inclusion of partial syndromes, inclusion of lacunar transient ischemic attack). Moreover, the inclusion of patients without CT scan or MRI in some series may have caused diagnostic uncertainty. Predictors of death have rarely been studied 316 in lacunar infarcts. Moreover, the low number of these events necessitates great care when commenting on the results. Gandolfo et al 3 reported an excess of mortality in patients with pseudobulbar palsy, poor functional recovery, badly controlled hypertension, and age over 65 years. We only studied characteristics present at the onset of the qualifying lacunar infarct. We found a significant difference in survival in cigarette smokers, in patients with diabetes mellitus, and in patients over 7 years of age. Our results are close to those of Brainin et al' 6 (risk ratio for death, 2.43 for age and 2.27 for diabetes). The other characteristics, history of stroke, type of lacunar syndrome, sex, and arterial risk factors, were not predictors of death in the study of Brainin et al, 16 nor were they in our series. In our study, the shortterm recurrent rate was similar to those reported in other studies,' 7 17 whereas the midterm recurrent rate appeared to be lower (6.5% at 2 years in our study, 12% in the study of Gandolfo et al, 3 and 11% in the study of Giroud et al 5 6 ). The longterm recurrent rate was 15% at 4 years in our study, 26% at 5 years in the study of Sacco et al, 4 and 34% at 7 years in the study of Gandolfo et al. 3 The recurrent type of stroke was not reported in most of the studies, especially in populationbased studies. Recurrent lacunar infarcts accounted for about one quarter of the recurrences (4 of 21 for Sacco, 4 7 of 24 for Gandolfo, 3 4 of 16 in our series). Moreover, the rate of ischemia versus hemorrhage was very close to the usual 8% rate reported in stroke.' 7 However, in one series, the recurrences were mainly lacunar infarcts (6 of 7), 8 suggesting recurrences of the identical pathophysiology of the qualifying stroke. Therefore, the discrepancy between the results of these studies is not a major contribution in the controversy concerning the hypothesis of a single mechanism being at the origin of lacunar infarcts that could be indirectly approached by recurrent stroke subtype analysis. In our series, no baseline variable predicted recurrences. Treatment of the risk factors, regular use of antiaggregants, and the small number of events may contribute to this negative result. Favorable functional prognosis has been reported after a lacunar infarct ' 9 However, this finding has not been confirmed in other studies. In the OCSP, 1 the proportion of independent patients was 66.3% after 1 year and 64% in Dijon after 2 years. 56 In our study, it was 74% after 1 year. The functional prognosis depends on the type of lacunar syndrome, age, and history of stroke or transient ischemic attack. In the study of Gandolfo et al, 3 as in ours, patients with motor deficit (sensorimotor strokes, pure motor hemiparesis) had the worst functional prognosis, whereas patients with ataxic hemiparesis and other lacunar syndromes recovered almost completely. Diabetic patients needed more help in their daily lives. Boiten et al 2 recently reported that patients with one or more asymptomatic lacunes had a higher frequency of hypertension and leukoaraiosis than patients with one symptomatic lacune. They suggested that the vasculopathies may be different, the former group having leukoaraiosis. In our results, we found an association betvcen hypertension and leukoaraiosis (^ 2 =11.6; P=.1) and between multiple lacunes and leukoaraiosis (* 2 =4.45; / J =.O34). However, the association between hypertension and the number of lacunes s2 did not reach statistical significance (A 2 =2.7; P=.1). They did not report the effects of their hypothesis on prognosis. Miyao et al 2 ' reported that leukoaraiosis was a predictor of higher mortality, stroke recurrence, and disability. Our results did not support their findings. Conclusions The longterm prognosis of lacunar infarcts is favorable. They have a higher survival rate, a lower recurrence rate, and a relatively good functional prognosis compared with the other subgroups of strokes. Some predictive factors could be isolated (cigarette smoking, age, and diabetes for death; type of lacunar syndrome, age, history of stroke or transient ischemic attack, and diabetes for disability). It appears that a hospitalbased study, ensuring high diagnostic precision and facilitating

5 Clavier et al Prognosis of Symptomatic Lacunar Infarcts 29 the obtention of data, is helpful in the study of the prognosis of one stroke subgroup. It highlights the possibility that both populationbased and hospitalbased studies have complementary uses. Appendix Hypertension was defined as a previous history of known hypertension with antihypertensive therapy or three or more blood pressure recordings >16 mm Hg for systolic and/or >95 mm Hg for diastolic pressure (WHO criteria) before or after the first week after the stroke. mellitus was defined as a known and treated diabetes, a fasting serum glucose concentration >1.4 g/l, or a serum glucose concentration >2 g/l 2 hours after the ingestion of 75 g of glucose. Dyslipidemia was defined as a known and treated dyslipidemia or a total fasting cholesterol level 22.4 g/l in men (>2.2 g/l in women) and/or a triglyceride level >1.4 g/l in men (>1.2 g/l in women). Cigarette smoking was defined as a total consumption of a 2 cigarette packs per year. History of ischemic heart disease included previous myocardial infarction and/or angina pectoris. Arrhythmia included chronic or paroxysmal atrial fibrillation or flutter. Cardiomegaty is defined as a cardiothoracic index >.5 on chest radiography. Cardiac valve diseases included left valve diseases, prosthetic aortic or mitral valve, and aortic and/or mitral valve stenosis. Lower limb claudication is defined as either lower limb claudication or as being due to atheromatous lower limb arterial disease. References 1. Bamford J, Sandercock P, Jones L, Warlow C. The natural history of lacunar infarction: the Oxfordshire Community Stroke Project. Stroke. 1987;18: Arboix A, MartiVilalta JL, Garcia JH. Clinical study of 227 patients with lacunar infarcts. Stroke. 199;21: Gandolfo C, Moretti C, Dall Agata D, Primavera A, Brusa G, Loeb C. Longterm prognosis of patients with lacunar syndromes. Ada Neurol Scand. 1986;74: Sacco SE, Whisnant JP, Broderick JP, Phillips SJ, O'Fallon WM. Epidemiological characteristics of lacunar infarcts in a population. Stroke. 1991;22: Giroud M, Gras P, Milan C, Arveux P, Beuriat P, Vion PH, Dumas R. Histoire naturelle des syndromes lacunaires: apport du registre dijonnais des accidents vasculaires cerebraux. Rev Neurol. 1991; 147: Giroud M, Milan C, Beuriat P, Gras P, Essayagh E, Arveux P, Dumas R. Incidence and survival rates during a twoyear period of intracerebral and subarachnoid haemorrhages, cortical infarcts, lacunes and transient ischaemic attacks: the stroke registry of Dijon, IntJ Epidemiol 1991:2: Ricci S, Celani MG, Guercini G, Rucireta P, Vitali R, La Rosa F, Duca E, Ferraguzzi R, Paolotti M, Seppoloni D, Caputo N, Chiurulla C, Scaroni R, Signorini E. Firstyear results of a communitybased study of stroke incidence in Umbria, Italy. Stroke. 1989;2: Boiten J, Lodder J. Prognosis for survival, handicap and recurrent stroke in lacunar patients and superficial infarction. Cerebrovasc Dis. 1993;3: Hommel M, Besson G, Le Bas JF, Gaio JM, Pollak P, Borgel F, Perret J. Prospective study of lacunar infarction using magnetic resonance imaging. Stroke. 199;21: Donan GA, Tress BM, Bladin PF. A prospective study of lacunar infarction using computerized tomography. Neurology. 1982;32: Mohr JP. Lacunes. Stroke. 1982;13: Fisher CM. Lacunar strokes and infarcts: a review. Neurology. 1982;32: Fisher CM. Lacunar infarcts: a review. Cerebrovasc Dis. 1991;1: Zimmerman RD, Fleming CA, Lee BCR, SaintLouis LA, Deck MDF. Periventricular hypertension as seen by magnetic resonance: prevalence and significance. Am J Neuroradiol. 1986;7: rusis MJ. SPSS/PC+ 4. Manual. Chicago, 111: SPSS Inc; Brainin M, Foulkes MAA, Pauly E, Seiser A, Dastmaltschi J, Steinert M. The role of hypertension and diabetes for the survival of lacunar stroke patients. Cerebrovasc Dis. 1992;2:22. Abstract. 17. Hier DB, Foulkes MA, Swiontoniowski M, Sacco RL, Gorelik PB, Mohr JP, Wolf PA. Stroke recurrence within 2 years after ischemic infarction. Stroke. 1991;22: rrving B, Staaf G. Pure motor stroke from presumed lacunar infarct: incidence, risk factors and initial course. Cerebrovasc Dis. 1991;l: Libman RB, Sacco RL, Shi T, Tatemichi TK, Mohr JP. Neurologic improvement in pure motor hemiparesis: implication for clinical trials. Neurology. 1992;42: Boiten J, Lodder J, Kessels F. Two clinically distinct lacunar infarct entities? A hypothesis. Stroke. 1993;24: Miyao S, Takano A, Teramoto J, Takahashi A. Leukoaraiosis in relation to prognosis for patients with lacunar infarction. Stroke. 1992;23: