Cerebral Small Vessel Disease and Risk of Death, Ischemic Stroke, and Cardiac Complications in Patients With Atherosclerotic Disease
The Second Manifestations of ARTerial disease-Magnetic Resonance (SMART-MR) Study
Background and Purpose—Cerebral small vessel disease may be related to vascular and nonvascular pathology. We assessed whether lacunar infarcts and white matter lesions on MRI increased the risk of vascular and nonvascular death and future vascular events in patients with atherosclerotic disease.
Methods—Brain MRI was performed in 1309 patients with atherosclerotic disease from the Second Manifestations of ARTerial disease-Magnetic Resonance (SMART-MR) study. Infarcts were scored visually and volumetric assessment of white matter lesion was performed. Patients were followed for a median of 4.5 years (range, 0.2 to 7.1 years) for death, ischemic stroke, and ischemic cardiac complications.
Results—Cox regression models showed that presence of lacunar infarcts (n=229) increased the risk of vascular (hazard ratio, 2.6; 95% CI, 1.4 to 4.9) and nonvascular death (hazard ratio, 2.7; 95% CI, 1.3 to 5.3), adjusted for age, sex, vascular risk factors, nonlacunar infarcts, and white matter lesion. These risks were similar for patients with silent lacunar infarcts. White matter lesion volume (relative to total intracranial volume) increased the risk of vascular death (hazard ratio per milliliter increase, 1.03; 95% CI, 1.01 to 1.05) and white matter lesions in the upper quintile compared with lower quintiles increased risk of ischemic stroke (hazard ratio, 2.6; 95% CI, 1.3 to 4.9).
Conclusions—Cerebral small vessel disease, with or without a history of cerebrovascular disease, is associated with increased risk of death and ischemic stroke in patients with atherosclerotic disease.
In cerebral small vessel disease (SVD), ischemic lesions are located in the supplying areas of the small perforating arteries in the basal ganglia or in the deep white matter of the brain. Both macrovascular disease such as atherosclerosis and hypertension and microvascular disease such as endothelial dysfunction and leakage of the blood–brain barrier have been associated with SVD.1–4 On MRI, markers of cerebral SVD are visible as white matter lesions (WML) and lacunar infarcts (LI). They increase the risk of stroke, cognitive decline, dementia, and death, both in the general population and in patients with stroke.5–8 Patients with atherosclerosis have a high risk of vascular events. It is not known whether the presence of LI or WML imposes an additional risk of death or vascular events on top of pre-existent vascular disease. The only study that investigated WML in patients with established atherosclerotic disease showed that WML increased the risk of ischemic stroke and myocardial infarction.9 There are no studies concerning adverse outcomes in patients with atherosclerotic disease in relation to lacunar infarcts.
In population-based studies, cerebral SVD has also been associated with retinopathy and nephropathy.10,11 It is assumed that underlying generalized SVD is responsible for this association, which is supported by the observation that similar pathological changes (eg, hyaline arteriolosclerosis) are found in kidneys of patients with hypertensive nephropathy and in brains of patients with cerebral SVD.12,13 Nephropathy itself is associated with increased all-cause mortality,14 and because WMLs are associated with nonvascular pathology such as pneumonia,15 it can be hypothesized that patients with LI or WML are at risk for both vascular and nonvascular events.
We investigated whether WML and LI increased the risk of vascular and nonvascular death and the risk of future vascular events in a cohort of patients with atherosclerotic disease.
Second Manifestations of ARTerial Disease-Magnetic Resonance Study
Data were used from the Second Manifestations of ARTerial disease-Magnetic Resonance (SMART-MR) study, a prospective cohort study on brain changes on MRI in patients with symptomatic atherosclerotic disease.16 Between May 2001 and December 2005, all patients newly referred to the University Medical Center Utrecht with manifest coronary artery disease, cerebrovascular disease, peripheral arterial disease, or an abdominal aortic aneurysm and without MR contraindications were invited to participate. Coronary artery disease was defined as myocardial infarction, coronary artery bypass graft surgery, or percutaneous transluminal coronary angioplasty in the past or at inclusion. Cerebrovascular disease was defined as a transient ischemic attack or stroke at inclusion diagnosed by a neurologist or self-reported stroke in the past. Peripheral arterial disease was defined as surgery or angioplasty of the arteries supplying the lower extremities in the history or intermittent claudication or rest pain at inclusion. Aneurysm of the abdominal aorta (AAA) was defined as present AAA (distal aortic diameter ≥3 cm) or previous AAA surgery. All patients underwent MRI of the brain, a physical examination, ultrasonography of the carotid arteries, and blood and urine sampling. Risk factors, medical history, and functioning were assessed with questionnaires. The SMART-MR study was approved by the ethics committee of our institution and written informed consent was obtained from all patients.
The MR investigations were performed on an 1.5-Tesla whole-body system (Gyroscan ACS-NT; Philips Medical Systems, Best, the Netherlands). The protocol consisted of transversal T1-weighted (TR/TE=235/2 ms), T2-weighted (TR/TE=2200/11 ms and 2200/100 ms), fluid-attenuating inverse recovery (TR/TE/inversion time=6000/100/2000 ms) and inversion recovery (TR/TE/inversion time=2900/22/410 ms) sequences. Field of view was 230×230 mm, matrix size 180×256, slice thickness 4.0 mm, no gap, and 38 slices.
We used the T1-weighted, inversion recovery, and fluid-attenuating inverse recovery sequence for the probabilistic segmentation technique.17,18 It distinguishes cortical gray matter, white matter, cerebrospinal fluid, and lesions. The results of the segmentation analysis were visually checked for the presence of infarcts and adapted if necessary to make a distinction between WML and infarct volumes. Total brain volume was calculated by summing the volumes of gray and white matter and, if present, the volumes of WML and infarcts. All volumes cranial to the foramen magnum were included. Total intracranial volume was calculated by summing total brain volume and cerebrospinal fluid volume. WML volumes were divided by intracranial volume and multiplied by the study population mean intracranial volume.
Infarcts were visually rated by an investigator and neuroradiologist blinded to clinical characteristics and re-evaluated in a consensus meeting. Infarcts were defined as focal hyperintensities on T2-weighted images of ≥3 mm in diameter. Hyperintensities in the white matter had to be hypointense on T1-weighted and fluid-attenuating inverse recovery images to distinguish them from WML. Dilated perivascular spaces were distinguished from infarcts on the basis of their location, form, and absence of gliosis. Infarcts were categorized as lacunar (sized 3 to 15 mm in diameter in plane and located in the subcortical white matter, thalamus, or basal ganglia) and nonlacunar (cortical infarcts, large subcortical infarcts, infratentorial infarcts).
Glucose and lipid levels were determined in an overnight fasting venous blood sample. Height and weight were measured to calculate body mass index. Blood pressure was measured twice with a sphygmomanometer and the average was calculated. Hypertension was defined as mean systolic blood pressure ≥160 mm Hg or mean diastolic blood pressure ≥95 mm Hg or antihypertensive drug use. Diabetes mellitus was defined as a history of diabetes mellitus, glucose ≥7.0 mmol/L, or oral antidiabetic drugs or insulin use. Hyperlipidemia was defined as total cholesterol >5.0 mmol/L, low-density lipoprotein cholesterol >3.2 mmol/L, or lipid-lowering drug use. Smoking (pack-years) and alcohol intake (never, former, current) were assessed with questionnaires.
Of the 1309 patients participating in the SMART-MR study, MR data were irretrievable for 19 patients; 14 had no fluid-attenuating inverse recovery sequence; and 44 patients had no brain volume data due to motion or artifacts. Four patients were lost to follow-up. Consequently, the analyses were performed in 1228 patients.
Patients received a questionnaire every 6 months to provide information on hospitalization and outpatient clinic visits. If a cardiovascular event was reported, original source documents were retrieved and reviewed to determine the occurrence of cardiovascular disease. All possible events were audited independently by 3 physicians of the End Point Committee. Patients were followed until death or refusal of further participation. The outcomes used in this study were death, ischemic stroke, and ischemic cardiac complications (see Supplemental Table I for definitions; http://stroke.ahajournals.org).
Patients were followed from date of MRI scan until death, loss to follow-up, or end of follow-up (March 2009), whichever came first. Cox regression analysis was used to estimate associations of the presence of LI with death, ischemic stroke, or ischemic cardiac complications adjusted for age and sex (Model 1); hypertension, diabetes mellitus, body mass index, smoking, alcohol consumption, and hyperlipidemia (Model 2); and presence of non-LI on MRI or a history of clinically evident cerebrovascular disease (Model 3). History of clinically evident cerebrovascular disease was defined as self-reported stroke or carotid operation in the clinical history or inclusion in the study with stroke or transient ischemic attack as a diagnosis. Analyses were repeated with silent LI, defined as LI on MRI but no history of clinically evident cerebrovascular disease.
The same models were used to analyze associations of WML with death, ischemic stroke, and cardiac complications. WML was analyzed as a continuous variable per milliliter relative to intracranial volume and as a dichotomized variable (upper quintile [>4.2 mL] versus the 4 lower quintiles). The same adjustments were made as in the models for LI, except for Model 3 in which we adjusted for all infarcts or history of clinically evident cerebrovascular disease.
Finally, all analyses were additionally adjusted for diagnosis of atherosclerotic disease at inclusion. Furthermore, interaction terms were tested between LI and diagnosis at inclusion and between WML and diagnosis at inclusion for all outcomes. SPSS 15.0 (Chicago, IL) was used to analyze the data.
Table 1 shows the baseline characteristics of the study sample. Two hundred twenty-nine patients had one or more LI. Of these, 127 had a history of clinically evident cerebrovascular disease and 102 had no such history. Among the 999 without LI, 155 had a history of clinically evident cerebrovascular disease and 844 had no such history. Patients in the upper quintile of WML had a median WML volume of 7.7 mL (10th to 90th percentile=4.6 to 22.5 mL). In total, 106 patients died during a median follow-up of 5.3 years (range, 0.2 to 8.1 years).
One or more LIs increased the risk of all-cause death in model 1 (hazard ratio [HR], 3.0; 95% CI, 2.0 to 4.4), which remained significant in Models 2 and 3 (Table 2). In patients without a history of clinically evident cerebrovascular disease (n=946), LI also increased the risk of all-cause death after adjusting for age and sex (HR, 3.6; 95% CI, 2.2 to 6.0, P<0.001) and also after additional adjustment for vascular risk factors, non-LI, and WML (HR, 3.2; 95% CI, 1.8 to 5.5; P<0.001). WML significantly increased the risk of all-cause death in Models 1, 2, and 3 (per milliliter increase as well as upper quintile of WML; Table 2).
Fifty-seven patients (53.8%) had a vascular cause of death (sudden death n=18, stroke n=10, congestive heart failure n=7, myocardial infarction n=4, AAA rupture n=4, other vascular causes n=14). Presence of LI significantly increased risk of vascular death in Models 1, 2, and 3 (Table 2). Silent LI also significantly increased risk of vascular death (HR, 4.1; 95% CI, 1.9 to 8.7; P<0.001) adjusted for all covariates. WML significantly increased the risk of vascular death in Models 1, 2, and 3 (per milliliter increase as well as upper quintile of WML; Table 2).
LI significantly increased risk of nonvascular death after adjusting for all covariates (Table 2) as did silent LI after adjusting for all covariates (HR, 2.6; 95% CI, 1.1 to 6.0; P=0.028). WMLs were not significantly associated with nonvascular death (Table 2).
Symptomatic as well as silent LI significantly increased risk of ischemic stroke; however, in Model 3, this risk decreased and was no longer statistically significant. Similarly, WML per milliliter was no longer significant after adjustment for infarcts on MRI or history of stroke, although severe WML remained significant in Model 3 (Table 3).
Ischemic Cardiac Complications
Symptomatic as well as silent LI and WML were not significantly associated with risk of ischemic cardiac complications (Table 3).
In all analyses, additional adjustment for atherosclerotic disease at inclusion resulted in similar effect estimates and did not change the significance levels (data not shown). All interactions tested were nonsignificant, except for the interaction between LI with AAA for the outcome “ischemic cardiac complications” (HR, 4.90; 95% CI, 1.20 to 20.00). However, the number of patients with AAA was very small and this could also be a chance finding.
Twenty-two patients were included with symptomatic LI without clinically apparent concomitant coronary artery disease, peripheral artery disease, AAA, or carotid artery stenosis of >50%. Exclusion of these patients did not change the results (data not shown). Likewise, exclusion of 37 patients with non-LI without other clinically apparent manifestations of atherosclerotic disease did not change the results (data not shown).
We found that presence of LI on MRI, whether symptomatic or asymptomatic, increased the risk of vascular and nonvascular death in patients with symptomatic atherosclerotic disease. LI also increased the risk of future ischemic stroke, but this was explained by concomitant infarcts on MRI and a history of clinically evident cerebrovascular disease. Severe WML load increased the risk of vascular death and ischemic stroke irrespective of concomitant cerebrovascular disease.
Our finding that WML and LI led to an increased risk of all-cause death is consistent with studies in different study populations.5,8,15 More interesting was that LIs were not only associated with vascular death, but also with nonvascular death. In our cohort, the majority of nonvascular deaths were the result of fatal malignancy (72.5%). As previously shown for nephropathy,14 these results show that the presence of LI is associated with increased morbidity and mortality not exclusively resulting from vascular events. In this respect, LI could be a marker of overall increased vulnerability to adverse outcomes, indicating the clinical relevance of LI in patients with manifestations of atherosclerotic disease outside the brain. Furthermore, our results indicate that silent LIs, which are generally not treated, may actually be important for the patients' prognosis. Previous studies showed that treatment with statins can improve vascular function in patients with symptomatic LI and WML.19 However, further research is needed to investigate whether this treatment also improves clinical outcomes of patients with apparently asymptomatic LI and WML.
For LI, the increased risk of ischemic stroke was explained by other infarcts on MRI and a history of clinically evident cerebrovascular disease. Patients with severe WML load did have an increased risk of ischemic stroke after adjusting for other infarcts and a history of stroke. This is consistent with 2 other studies in older populations that found a relation between WML and risk of ischemic stroke, although those were not adjusted for other infarcts.9,20
Although it is thought that LI and WML are both caused by small vessel changes in the brain, the differential associations with nonvascular death suggest that the prognosis may be different and we tentatively hypothesize that LI and WML may actually be 2 separate forms of SVD. Further studies within 1 study population and in population-based studies are needed to investigate this hypothesis.
Strengths of our study are the large number of patients included, the virtually complete follow-up, the rigorous assessment of clinical outcomes, the analyses within patients without a history of clinically evident cerebrovascular disease, the automated brain segmentation, and the adjustment of confounders. Also, the large sample size enabled us to also analyze silent LI. Previously published studies investigated silent infarcts in general and did not look specifically at LI.7 Furthermore, our study is the first to specifically investigate nonvascular death.
A limitation is that it is difficult to verify if silent infarcts are really clinically asymptomatic. We tried to minimize misclassification by excluding patients who had a transient ischemic attack or stroke at inclusion or reported a transient ischemic attack or stroke in the past. Furthermore, infratentorial infarcts were scored as nonlacunar, although 14 patients had infratentorial infarcts <15 mm without having any other lacunar infarcts supratentorial. This could have led to some underestimation of the results for LI. Furthermore, not all LIs show cavitation, and noncavitating LIs may resemble WML.21 This could also be a source of underestimation of LI. Finally, we investigated multiple outcomes and adjusted for many potential confounders with a relatively small number of events and the results should be interpreted with caution.
In conclusion, LIs and WMLs on MRI are risk factors for adverse outcomes in patients with atherosclerotic diseases. Further research is needed to investigate whether the presence of LIs or WMLs have added value in models to predict prognosis in patients with atherosclerotic disease.
Sources of Funding
The research is supported by the Dutch Ministry of Welfare and Health and the Netherlands Organization for Scientific Research (project no. 917-66-311). They had no involvement in study design; collection, analysis, or interpretation of data; writing of the report; or in the decision to submit paper for publication.
Y.v.d.G. received support from the Dutch Ministry of Welfare and Health. M.I.G. received support from the Netherlands Organization for Scientific Research.
Members of the SMART Study Group of UMC Utrecht: A. Algra, MD, PhD, Julius Center and Rudolf Magnus Institute for Neurosciences, Department of Neurology; P.A. Doevendans, MD, PhD, Department of Cardiology; Y. van der Graaf, MD, PhD, D.E. Grobbee, MD, PhD, G.E.H.M. Rutten, MD, PhD, Julius Center for Health Sciences and Primary Care; L.J. Kappelle, MD, PhD, Department of Neurology; W.P.Th.M. Mali, MD, PhD, Department of Radiology; F.L. Moll, MD, PhD, Department of Vascular Surgery; and F.L.J. Visseren, MD, PhD, Department of Vascular Medicine.
Joanna M. Wardlaw, MD, was the Guest Editor for this paper.
The online-only Data Supplement is available at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.111.594853/-/DC1.
- Received June 30, 2010.
- Revision received May 18, 2011.
- Accepted June 2, 2011.
- © 2011 American Heart Association, Inc.
- Breteler MM,
- van Swieten JC,
- Bots ML,
- Grobbee DE,
- Claus JJ,
- van den Hout JH,
- et al
- Wardlaw JM,
- Sandercock PA,
- Dennis MS,
- Starr J
- de Jong G,
- Kessels F,
- Lodder J
- Debette S,
- Beiser A,
- DeCarli C,
- Au R,
- Himali JJ,
- Kelly-Hayes M,
- et al
- Ikram MA,
- Vernooij MW,
- Vrooman HA,
- Hofman A,
- Breteler MM
- Ikram MA,
- Vernooij MW,
- Hofman A,
- Niessen WJ,
- van der Lugt A,
- Breteler MM
- Kwa VI,
- van der Sande JJ,
- Stam J,
- Tijmes N,
- Vrooland JL
- van Swieten JC,
- van den Hout JH,
- van Ketel BA,
- Hijdra A,
- Wokke JH,
- van Gijn J
- Rifkin DE,
- Katz R,
- Chonchol M,
- Fried LF,
- Cao J,
- de Boer I,
- et al
- Briley DP,
- Haroon S,
- Sergent SM,
- Thomas S
- Sterzer P,
- Meintzschel F,
- Rosler A,
- Lanfermann H,
- Steinmetz H,
- Sitzer M
- Vermeer SE,
- Hollander M,
- Van Dijk EJ,
- Hofman A,
- Koudstaal PJ,
- Breteler MM
- Potter GM,
- Doubal FN,
- Jackson CA,
- Chappell FM,
- Sudlow CL,
- Dennis MS,
- et al