Differences Between Ischemic Stroke Subtypes in Vascular Outcomes Support a Distinct Lacunar Ischemic Stroke Arteriopathy
A Prospective, Hospital-Based Study
Background and Purpose— Whether and how the arterial pathology underlying lacunar ischemic stroke differs from the atherothrombotic processes causing most other ischemic strokes is still debated. Different risks of recurrent stroke and MI after lacunar versus nonlacunar ischemic stroke may support a distinct lacunar arteriopathy.
Methods— We prospectively followed a hospital-based cohort of 809 first-ever ischemic stroke patients for 1 to 4 years. We compared risks of death, recurrent stroke, and MI in patients with lacunar versus nonlacunar stroke, and performed an updated meta-analysis of recurrent stroke subtype patterns.
Results— During 1725 person-years of follow-up, 109 patients had a recurrent stroke and 31 had MI. All patients at baseline, and 93% with recurrent stroke, had brain imaging and more than half with recurrent stroke had diffusion-weighted MRI. Overall, there was no difference in recurrence risk after lacunar vs nonlacunar stroke, although there was a trend toward a lower recurrence risk in the early weeks after lacunar stroke. Lacunar recurrence was more likely after lacunar than nonlacunar stroke (OR, 6.5; 95% CI, 2.4–17.5; updated meta-analysis OR, 6.8; 95% CI, 4.2–11.2). MI risk was nonsignificantly lower after lacunar than nonlacunar stroke (rate ratio, 0.5; 95% CI, 0.2–1.1; rate ratio after excluding patients with previous ischemic heart disease: 0.3; 95% CI, 0.1–0.9).
Conclusions— Our finding of a trend toward a lower MI risk after lacunar vs nonlacunar stroke and confirmation of both a lower early recurrence risk after lacunar stroke and a tendency of recurrent stroke subtypes to “breed true” support the notion of a distinct nonatherothrombotic lacunar arteriopathy.
Approximately one-quarter of ischemic strokes are lacunar, presumed to result from occlusion, or perhaps leakiness,1 of one of the perforating arteries supplying the deep, subcortical areas of the brain. The underlying arterial pathology is poorly understood, but in most cases it is thought to be an intrinsic small vessel disease distinct from the atherothromboembolic processes causing most other ischemic strokes.1,2 In support of this is evidence that patients with lacunar ischemic stroke have a lower frequency than those with nonlacunar ischemic stroke of potential sources of thromboemboli (carotid stenosis and sources of cardiac emboli) and of previous ischemic heart disease (IHD), ie, coronary atherothrombosis.3,4
If lacunar stroke is mainly caused by a distinct nonatherothrombotic arteriopathy, then we would also expect a lower early recurrent stroke rate compared with nonlacunar ischemic stroke (attributable to a lower frequency of active sources of thromboemboli); a tendency for recurrent stroke subtypes to “breed true” (ie, a further lacunar ischemic stroke would be more likely after a lacunar than a nonlacunar ischemic stroke); and a lower risk of MI among patients with lacunar vs nonlacunar ischemic stroke. We tested these hypotheses in a large, prospective, hospital-based cohort of well-characterized ischemic stroke patients.
Although previous studies have suggested a lower early recurrence rate among patients with lacunar ischemic stroke,5 and a tendency for recurrent stroke subtypes to breed true, their reliability was limited by small numbers of events, variable and sometimes biased definitions of recurrent stroke, and low rates of brain imaging among recurrent strokes, with no reported use of diffusion-weighted MRI,6 which is particularly helpful in differentiating new from old lesions, especially if small or in patients who present late.7 Several cohort studies have assessed MI risk after ischemic stroke,8 but only 1 reported this for lacunar and nonlacunar ischemic stroke patients separately, recording only 6 MI.9 We aimed to overcome these limitations.
Subjects and Methods
Between 2002 and 2005 we prospectively recruited consecutive, consenting patients with stroke (defined according to WHO criteria10), admitted to or seen as outpatients at our hospital. In this study, we included patients with a first-ever-in-a-lifetime clinically evident stroke, demonstrated by brain imaging (CT or MRI) to be ischemic (ie, primary hemorrhage excluded). We assigned ischemic stroke subtypes according to the presumed site and size of the causative infarct (anterior circulation lacunar or cortical [including striatocapsular] infarction, or posterior circulation infarction) using the clinical features of the stroke,11 modified if necessary by the findings on brain imaging if an infarct considered relevant to the presenting stroke was present. We excluded cases of posterior circulation infarction, which include lacunar and nonlacunar events that are often difficult to distinguish, patients with an uncertain subtype, and patients with an unusual cause of stroke (eg, arterial dissection).
We defined recurrent stroke as for index stroke, additionally requiring a period of neurological stability of ≥24 hours between index and recurrent stroke, and exclusion of other potential causes of neurological deterioration. We defined definite MI as either autopsy evidence or at least 2 of the following; symptoms of myocardial ischemia (eg, chest pain); enzyme changes indicative of MI (generally raised troponin); and ECG changes suggesting new ischemia (new ST-T wave changes, Q waves, or left bundle branch block). We defined probable MI as sudden death without evidence of an alternative cause.
We followed-up patients for between 1 and 4 years using multiple overlapping methods, including regular patient questionnaires, contact with patients’ general practitioners, and linkage to the national death register. We reviewed all relevant records to confirm the cause of death for more than two-thirds of deaths, including all those for which part I or II of the death certificate mentioned either stroke or MI.
Whenever possible, we arranged specialist review for patients with a suspected recurrent stroke, and CT or MR brain imaging with diffusion-weighted imaging, T2, FLAIR, and gradient echo, aiming particularly to perform diffusion-weighted MRI if CT revealed no visible relevant lesion. For patients with a suspected MI, or those unable to attend a clinical assessment for suspected recurrent stroke, we sought confirmation of the event by reviewing all relevant medical records.
We used STATA version 8. We compared baseline characteristics (lacunar vs nonlacunar) using the χ2 test for dichotomous variables, Student t test for normally distributed continuous variables, the Mann-Whitney U test for non-normally distributed continuous variables, and the χ2 test for trend for ordered categorical variables. We calculated median length of follow-up as median observation time (from study entry to date of death or date censored). We compared cumulative probability plots (lacunar vs nonlacunar) with the log rank test for each of death, recurrent stroke, and MI (definite or probable), censoring patients at time of the event of interest, death, or end of follow-up. We used Cox regression to obtain unadjusted and age- and sex-adjusted hazard ratios for death and for recurrent stroke for the entire follow-up period and for prespecified periods: 0 to 1 and 1 to 4 years; 0 to 1 month and 1 month to 4 years; and (for recurrent stroke only) 0 to 1 week and 1 week to 4 years. We could not adjust hazard ratios for recurrent stroke at 1 month or 1 week because of low numbers of early recurrences. Multivariate survival analysis techniques for MI were precluded by the relatively small number of events and the complexities of analyzing survival curves that cross multiple times. Therefore, we calculated an unadjusted rate ratio, repeating this analysis after excluding patients with previous IHD, because a higher rate of MI at follow-up could be attributable to established IHD.
In prespecified sensitivity analyses we restricted the nonlacunar comparison group to patients with mild cortical ischemic stroke only (associated with a partial, as opposed to total, anterior circulation stroke syndrome11), and compared small vs large vessel disease ischemic stroke using a modified TOAST classification (Supplemental Figure I, available online at http://stroke.ahajournals.org).12
We analyzed recurrent stroke subtype patterns by calculating the odds of a lacunar recurrence after lacunar vs nonlacunar index stroke, using logistic regression to adjust for potential confounding by age, sex, and antithrombotic therapy at onset of recurrence. We updated to the end of 2007 our previous meta-analysis of published studies of recurrent stroke subtype patterns,6 including unadjusted data from the current study.
We estimated extent of misclassification of ischemic stroke subtypes by calculating the proportion of patients with a visible relevant infarct on their brain scan whose final classification placed them in a different comparison group from their clinical syndrome classification. We then applied this proportion to the patients with no visible relevant infarct on brain imaging to estimate the possible extent of residual misclassification among both index and recurrent ischemic stroke subtypes.
We recruited and followed-up 809 patients with a brain imaging-confirmed first-ever-in-a-lifetime anterior circulation ischemic stroke (282 lacunar, 527 nonlacunar) for a median of 2.4 (IQR, 1.7–3.2) years, giving 1725 person-years follow-up time, with all patients followed-up by at least 1 of our overlapping methods.
Lacunar patients were slightly younger than nonlacunar patients (mean, 69 vs 73 years), more often male (60% vs 47%), had lower frequencies of IHD, atrial fibrillation, and ipsilateral carotid stenosis, and were more often smokers (Table 1).
Thirty-eight (13%) lacunar patients and 153 (29%) nonlacunar patients died (adjusted hazard ratio, lacunar vs nonlacunar: 0.46; 95% CI, 0.33–0.66; Figure 1A). Analyses by time period showed that the difference was most prominent early after stroke (0–1 month deaths, lacunar vs nonlacunar 1/282 vs 34/527; hazard ratio, 0.07; 95% CI, 0.01–0.51).
One hundred nine patients had at least 1 recurrent stroke; 36 (13%) were in the lacunar and 73 (14%) were in the nonlacunar group. We found no statistically significant difference in recurrence risk between lacunar and nonlacunar patients overall (Figure 1B; age- and sex-adjusted hazard ratio, 0.88; 95% CI, 0.59–1.31) but found a trend toward a lower early risk of recurrence among lacunar patients, especially in the first week (1/282 lacunar vs 7/527 nonlacunar; hazard ratio, 0.31; 95% CI, 0.04–2.53). Eight (3%) patients with lacunar vs 25 (5%) with nonlacunar ischemic stroke had MI, giving a nonsignificantly lower rate of MI in the lacunar group (Figure 1C; rate ratio, 0.51; 95% CI, 0.23–1.14; P=0.09), which was significantly lower when we excluded patients with previous IHD (rate ratio, 0.28; 95% CI, 0.08–0.95; P=0.03). Sensitivity analyses for all outcomes gave similar results (data not shown).
93% of patients with recurrent stroke had brain imaging. We were able to assign a clinical syndrome and imaging-based subtype in 97 of 98 cases of recurrent ischemic stroke, more than three-quarters of whom had a visible relevant infarct on brain imaging (Table 2). The distribution of recurrent subtypes in the lacunar vs nonlacunar groups was significantly different (Supplemental Figure II, available online at http://stroke.ahajournals.org), and the odds of an ischemic lacunar recurrence were ≈6-times greater after lacunar vs nonlacunar stroke at baseline (OR, 6.50; 95% CI, 2.43–17.52; OR adjusted for age, sex, and antithrombotic treatment, 5.39; 95% CI, 1.79–16.2).
Estimate of Ischemic Stroke Subtype Misclassification
Among 509 patients with a visible relevant infarct on their scan, we corrected the clinical classification in 64 patients. Applying this proportion to the remaining 300 patients, we estimated that, overall, at least 42 of the total 809 patients (5%) may have been residually misclassified at baseline. Slightly fewer (4/97 [4%]) recurrent strokes were potentially misclassified. Among baseline and recurrent strokes, the estimated proportion misclassified was similar in the lacunar and nonlacunar comparison groups.
We previously identified 3 studies reporting on patterns of recurrent ischemic stroke subtypes.6,13–15 We identified 2 subsequent studies,16,17 one of which overlapped with and superseded a previously included study.16 Pooling data from these with data from our own study gave an increased odds of a lacunar recurrence after lacunar vs nonlacunar ischemic stroke at baseline (OR, 6.83; 95% CI, 4.18–11.18; Figure 2). There was substantial heterogeneity between studies (I2=77%), largely explained by 1 study with a much higher proportion of lacunar recurrences after nonlacunar index stroke (23%) than in the other studies (3% to 11%).13 This could be attributable to the definition of recurrent stroke used in this study; recurrences within 21 days and in the same part of the brain as the index event were excluded, leading to underestimation of the extent to which recurrences breed true.
In our large, prospective stroke cohort study, we found no difference in long-term risk of recurrent stroke between lacunar and nonlacunar patients. However, the very early risk of recurrence was almost certainly lower in lacunar patients, confirming previous results,5 and probably reflecting a lower prevalence of an active proximal embolic source. We also confirmed that recurrent stroke subtypes tend to breed true, even after adjusting for potential confounding effects of age, sex, and antithrombotic treatment. The similarity between adjusted and unadjusted results provides methodological justification for our meta-analysis of unadjusted data from published studies, the results of which confirmed and strengthened the findings from this study. Finally, we report for the first time to our knowledge a potentially lower MI risk after lacunar vs nonlacunar ischemic stroke, in keeping with our previous work suggesting that previous IHD is less common among lacunar patients.4
Our study benefits from a number of methodological strengths. First, the large study population generated a relatively large number of outcome events, >5-times as many MI events as in the only previous published study of MI risk among both lacunar and nonlacunar ischemic stroke subtypes.9 Second, a high proportion of patients had brain imaging after index and recurrent strokes (100% and 93%, respectively, with more than half of recurrences undergoing diffusion-weighted MRI). This allowed more accurate classification of recurrent stroke subtypes than in previous similar studies, in which only 40% to 65% of patients with recurrent stroke had brain imaging, none with diffusion-weighted MRI.6 The extent of misclassification in our study was low and similar in both comparison groups, so that we are unlikely to have overestimated, but may have underestimated, any true differences in outcome. Finally, our unbiased definition of recurrent stroke did not exclude patients with an early recurrence in the same territory of the brain as the index event, as many previous studies have. This limited underestimation of the early recurrence risk19 and potential bias in determining recurrent stroke subtypes.
Our study does have some limitations. First, our early stroke recurrence rates were lower than those of previous similar studies.5 This may partly reflect our recruitment of patients several days (median, 9) after stroke onset, excluding some patients who had already had a very early recurrence and so underestimating the very high recurrence rate in the first few days after nonlacunar (mainly large artery) ischemic stroke and diminishing the apparent difference in early recurrence risk between nonlacunar and lacunar stroke. In addition, we may have been more likely to miss nonfatal outcome events occurring among patients with a severe index stroke, because ascertainment and reporting of further events can be reduced in such patients. Reassuringly, however, our results were unchanged in sensitivity analyses comparing patients with lacunar vs mild cortical strokes, and the incidence of MI among ischemic stroke patients in our study was comparable to that in previous published studies.8
Second, although this is the largest study of MI risk among different ischemic stroke subtypes to date, the relatively low incidence of MI limited the reliability and precision of comparisons between lacunar and nonlacunar patients, and prevented full adjustment for potential confounders. Third, in our primary analyses, we will have included a small number of patients in the lacunar group whose stroke may have been caused by embolism from proximal sources rather than intrinsic intracranial small vessel disease. However, the results of our primary analyses remained unchanged when we compared stroke subtypes according to their presumed etiologic cause. Fourth, there was substantial heterogeneity between studies in our meta-analysis of recurrent stroke subtypes, attributable to results of 1 study,13 which may have underestimated the extent to which recurrent stroke subtypes breed true. Finally, the requirement for consent meant that not all eligible patients were included. However, as described previously, we recruited 88% of all eligible patients and found no difference in age, sex, or stroke subtype distribution between participants and nonparticipants. We did find that participants were more likely to be admitted to a stroke unit and were more affluent.20
In conclusion, our findings considerably strengthen existing evidence for a lower risk of stroke recurrence early after lacunar ischemic stroke, and for a tendency of recurrent ischemic stroke subtypes to breed true. Our additional finding of a lower risk of MI after lacunar vs nonlacunar ischemic stroke provides further epidemiological evidence to suggest that many lacunar ischemic strokes are caused by a distinct, nonatherothrombotic, small vessel arteriopathy. Because this finding was based on small numbers of MI events (albeit a much larger number than previously published), it would be helpful to confirm it with a pooled analysis of data from similar stroke cohorts, with well-characterized baseline ischemic stroke subtypes and follow-up for MI.
The authors thank the patients, care givers, doctors, administrators, and programming staff who contributed to data collection; Professors Charles Warlow and Peter Sandercock for clinical expertise; Drs Andrew Farrell and Gillian Potter for neuroradiology input; Mike McDowall for programming assistance; Isabel Jennings for administrative support; and Professor Charles Warlow and Dr Rustam Al-Shahi Salman for comments on an earlier version of this manuscript.
Sources of Funding
The Edinburgh Stroke Study was funded by the Wellcome Trust (Clinician Scientist award 063668/Z/01/A to C.L.M.S.). C.A.J. received additional funding from the Binks Trust. Most of the MR scanning was performed in the SFC Brain Imaging Research Centre (www.sbircs.ed.ac.uk).
- Received May 18, 2009.
- Revision received July 1, 2009.
- Accepted August 4, 2009.
Wardlaw JM. What causes lacunar stroke? J Neurol Neurosurg Psychiatry. 2005; 76: 617–619.
Jackson C, Sudlow C. Are lacunar strokes really different? A systematic review of differences in risk factor profiles between lacunar and non-lacunar infarcts. Stroke. 2005; 36: 891–904.
Lovett KJ, Coull AJ, Rothwell PM. Early risk of recurrence by subtype of ischaemic stroke in population-based incidence studies. Neurology. 2004; 62: 569–573.
Jackson C, Sudlow C. Comparing risks of death and recurrent vascular events between lacunar and non-lacunar infarction. Brain. 2005; 128: 2507–2517.
Touze E, Varenne O, Chatellier G, Peyrard S, Rothwell PM. Mas JL. Risk of myocardial infarction and death after TIA and ischaemic stroke. A systematic review and meta-analysis. Stroke. 2005; 36: 2748–2755.
Landi G, Cella E, Boccardi E, Musicco M. Lacunar versus non-lacunar infarcts: Pathogenetic and prognostic differences. J Neurol Neurosurg Psychiatry. 1992; 55: 441–445.
Adams HP Jr, Bendixen BH, Kappelle LJ, Biller J, Love BB, Gordon DL, Marsh EE III. Classification of subtype of acute ischemic stroke: Definitions for use in a multicenter clinical trial. Stroke. 1993; 24: 35–41.
Hillen T, Coshall C, Tilling K, Rudd AG, McGovern R, Wolfe CDA. Cause of stroke recurrence is multifactorial: Patterns, risk factors, and outcomes of stroke recurrence in the South London stroke register. Stroke. 2003; 34: 1457–1463.
Hata J, Tanizaki Y, Kiyohara Y, Kato I, Kubo M, Tanaka K, et al. Ten year recurrence after first ever stroke in a Japanese community: the Hisayama study. J Neurol Neurosurg Psych. 2005; 76: 368–372.
Coull AJ, Rothwell PM. Underestimation of the early risk of recurrent stroke. Evidence of the need for a standard definition. Stroke. 2004; 35: 1925–1929.
Jackson C, Crossland L, Dennis M, Wardlaw J, Sudlow C. Assessing the impact of the requirement for explicit consent in a hospital-based stroke study. QJM. 2008; 101: 281–289.