American Heart Association

epidemiology and genetics

Risk Factors Associated with Presence of Unruptured Intracranial Aneurysms in a Large, Cross-sectional Imaging Study

The detection of unruptured intracranial aneurysms (UIAs) has increased as the use of noninvasive cerebral angiography has becomes more widespread. It has been posited that UIAs in different locations may have different pathophysiological mechanisms and consequently, different risk factor profiles. Kang and Kim et al. conducted a cross-sectional study of 18,954 consecutively enrolled subjects who had magnetic resonance angiography (MRA) of the cerebral vessels as part of a routine health examination over an approximately ten year period.

MRA was performed with 1.5-T and 3-T scanners. Subjects with aneurysms less than 3 mm or having fusiform, mycotic, traumatic, or treated aneurysms were excluded. The location of aneurysms was classified as distal internal carotid artery (dICA), middle cerebral artery (MCA), MCA bifurcation, anterior cerebral artery (ACA), anterior communicating artery (Acom), posterior communicating artery (Pcom), or posterior circulation artery (including vertebral, basilar, posterior cerebral, and anterior and posterior inferior cerebellar arteries). Overall age, sex, and vascular risk factors were compared for those with and without an UIA and separately for each aneurysm location subgroup. UIA distribution in the study population was also compared with a group of 12,702 subjects who had MRA as part of an outpatient evaluation for headache at the same institution.

Compared to patients having an MRA for headache, those in the health screening group tended to be younger (57.2 ± 8.5 vs 63.1 ± 12.5) and less frequently women (39% vs 64.4%). The prevalence of UIA was higher in the headache evaluation group as compared to the health screening group (2.38% vs 1.94% p=0.007). The prevalence of UIA was similar in men and women in the two groups and similarly increased in both groups in women over age 50 years.

Patients in the health screening group with an UIA tended to be older (57.2 ± 8.5 vs 55.8 ±9.0 p=0.005) and more frequently women (50.7% vs 38.8% p<0.001) than those without aneurysms. Interestingly, coronary artery disease was less prevalent in the aneurysm group (10.1% vs 24.2% p<0.001). In multivariable analysis, advanced age, female sex, hypertension, smoking and CAD each had independent associations with UIAs.

In order of descending frequency, UIAs were most commonly found in the distal ICA (45.2%), followed by the MCA bifurcation (13.4%) and the Acom (13.4%). Pcom aneurysms were observed in 10.4% of patients, whereas MCA and ACA aneurysms were found in 6.5% and 5.7% of subjects, respectively. Female sex, hypertension, and CAD were independently associated (CAD was negatively associated) with dICA, MCA bifurcation and Pcom aneurysms, whereas MCA aneurysms were associated with advanced age, smoking and negatively associated with CAD. Posterior circulation aneurysms were only associated with hypertension.

The findings of this study are consistent with previous work finding associations between vascular risk factors and the presence of UIAs, particularly with respect to age, female sex and hypertension. The analysis of the relationship between risk factors and UIAs by location supports the possibility of varying pathophysiologic mechanisms for aneurysm formation depending on site. The negative association with CAD is particularly intriguing; it is somewhat counter intuitive to assume that one form of vascular degeneration could be protective against another. Additional research aimed at understanding the mechanical and molecular basis of IA formation would be helpful to further elucidate the issue.

The study does have limitations. The possibility for selection bias should be acknowledged. It is also possible that the presumed race-ethnic homogeneity of the subjects (recruited from sites in Korea) limits the generalizability. Additional studies in other sites using a similar protocol will be needed to validate these findings.

Alcohol Intake and the Risk of Stroke

Alexander E. Merkler, MD 
Jones SB, Loehr L, Avery CL, Gottesman RF, Wruck L, Shahar E, and Rosamond WD. Midlife Alcohol Consumption and the Risk of Stroke in the Atherosclerosis Risk in Communities Study. Stroke. 2015

The association between alcohol use and stroke has been well studied. Based on prior reports and meta-analyses, light alcohol consumption is thought to be protective against ischemic stroke whereas heavy alcohol consumption may increase the risk of ischemic stroke. On the other hand, alcohol consumption only seems to increase the risk of intracerebral hemorrhage (ICH). As the authors point out, previous studies had multiple limitations including lack of information regarding previous alcohol use, socioeconomic data, or were retrospective analyses. 

The current study by Jones et al, evaluates the relationship between mid-life alcohol use and stroke using the Atherosclerosis Risk in Communities Study (ARIC). More than 15,000 patients were enrolled in ARIC, a prospective study of middle aged adults, performed at four socioeconomically diverse locations in the United States. Patients were followed for up to 25 years. In this study, patients included had no history of prior stroke and either were current alcohol consumers or lifetime abstainers. Thus, patients who were former alcohol consumers were excluded from the study. Alcohol use was assessed using questionnaires at the initial enrollment visit. Alcohol use was divided into 4 groups: never drinkers, light drinkers (less than 3 drinks / week), moderate drinkers (4-17 drinks/week) and heavy drinkers (>18 drinks/week).

Of the 12,000 ARIC subjects included in the analysis, more than one third were light drinkers, one-third reported lifetime abstinence from alcohol, and just 5% were heavy alcohol consumers. After adjustment for confounders, any amount of alcohol use was neither harmful nor protective against the development of ischemic stroke, which does not support the classic “J-shaped” relationship of alcohol use and risk of ischemic stroke. There was a trend that light alcohol consumption may be protective against ischemic stroke in white women and harmful in black men, but the confidence intervals were wide and crossed 1. On the other hand, moderate-to-heavy alcohol use was significantly associated with ICH (HR 1.99, 95% CI

The main limitation is that alcohol consumption was only evaluated at one time during the study – the initial visit. Thus, it is very possible that each patient’s alcohol consumption varied as the study progressed. Other limitations include the lack of knowledge regarding type of alcohol consumed, the reliance of questionnaires, and the fact that the study power was too low to detect significant differences in certain study subgroups such as heavy alcohol use.

Overall, the results of this large, prospectively gathered, heterogeneous population show neither a protective nor harmful effect of alcohol consumption on ischemic stroke risk. On the other hand, the results are consistent with prior reports showing a significant association between moderate-heavy alcohol consumption and development of ICH. Future research will be necessary to tease out the relationship between alcohol use and ischemic stroke.

Left atrial cardiomyopathy may be a cause of thromboembolism in the absence of atrial fibrillation

Up to one-third of ischemic strokes are of unknown cause. Although implantable loop recorders are helpful at detecting atrial fibrillation, many cardioembolic appearing strokes remain without an etiology. For example, in a recent study, despite three years of heart monitoring, 70% of patients with cryptogenic stroke had no evidence of atrial fibrillation.[1] Heretofore we have believed that the relationship between atrial fibrillation and stroke is straightforward, but Dr. Kamel et al challenge this traditional model by postulating that perhaps atrial fibrillation is merely a marker or consequence of atrial disease and that it is actually atrial disease that is a risk factor for stroke.

In the current study, Dr. Kamel et al evaluate the relationship between left atrial disease, as measured by P-wave terminal force in lead V1, and the risk of stroke while adjusting for atrial fibrillation and other stroke risk factors. P-wave terminal force is an established electrocardiographic marker of left atrial cardiomyopathy without necessarily having atrial fibrillation and can be easily calculated on a standard EKG. In order to test the hypothesis that atrial disease may lead to thromboemoblism without atrial fibrillation, atrial disease should be specifically associated with cardioembolic or cryptogenic stroke subtypes as opposed to non-cardioembolic stroke subtypes.

The authors measured the P-wave terminal force in 241 patients with ischemic stroke and a randomly selected cohort of 798 patients without ischemic stroke. P-wave terminal force in lead V1was indeed associated with the composite of cryptogenic and cardioemoblic stroke, but not non-cardioembolic stroke subtypes. In addition, in a sensitivity analysis that excluded any patient with atrial fibrillation, the results were unchanged or stronger. However, adjustment for left atrial size on echocardiography attenuated the results.

Limitations include the small sample size, the fact that two people manually calculated the P-wave terminal force and the inter-rater reliability was only moderately reliable, and the lack of more definitive methods to rule out atrial fibrillation.

Overall, the manuscript questions the long believed understanding that atrial fibrillation causes stroke. Instead, atrial fibrillation may merely represent part of the spectrum of atrial cardiomyopathy, which in itself may lead to thromboembolism. Furthermore, if proven true, the ease and cost-effectiveness of simply checking an EKG may prove to be an efficient way to assess for thromboembolic risk as compared to years of heart-rhythm monitoring.

[1] Sanna T, Diener HC, Passman RS, Di Lazzaro V, Bernstein RA, Morillo CA, et al. Cryptogenic stroke and underlying atrial fibrillation. N Engl J Med. 2014;370:2478-2486. Tweet: Atrial cardiomyopathy may be an independent risk factor for stroke

Interleukin 16 may lead to carotid plaque stabilization and reduce risk of stroke

Alexander E. Merkler, MD

Grönberg C, Bengtsson E, Fredrikson GN, Nitulescu M, Asciutto G, Persson A, et al. Human Carotid Plaques With High Levels of Interleukin-16 Are Associated With Reduced Risk for Cardiovascular Events. Stroke. 2015

Over the years, we have learned that carotid artery disease is more complex than degree of stenosis. Two patients with identical severities of carotid artery stenosis may have entirely different phenotypes: one patient may develop recurrent strokes or TIAs and the other may remain asymptomatic. Ongoing research evaluating “vulnerable plaque” characteristics such as intraplaque hemorrhage, necrosis, or lipid core volume will hopefully guide physicians to better predict which patients are at risk for stroke and may benefit from a carotid artery intervention.

Interleukin 16 (IL-16) is a cytokine with both pro and anti-inflammatory properties. In the current study, Grönberg et al evaluate whether high levels of IL-16 found in human carotid artery plaques may confer protection against stroke. The authors evaluated 206 asymptomatic and symptomatic carotid artery plaques from human subjects who received carotid endarterectomies. Symptomatic patients had >70% stenosis and asymptomatic patients had >80% stenosis. 

The authors found that IL-16 mRNA expression was significantly higher in plaques from patients with asymptomatic carotid disease as compared to plaques from symptomatic carotid artery disease. In addition, high levels of IL-16 protein in carotid plaques were associated with a decreased incidence of future cerebrovascular events. Furthermore, high levels of IL-16 were associated with markers of plaque stability including elastin and collagen.

Some limitations to consider include 1) the lack of MRI data on radiographical plaque morphology –ie: amount of hemorrhage or calcification and 2) lack of causality – instead of high levels of IL-16 being protective against stroke, perhaps, in patients with symptomatic carotid disease, stroke may lead to decreased levels of IL-16 through post-stroke immunodepression and subsequent plaque destabilization.

Overall, IL-16 seems to be associated with carotid plaque stabilization and a reduced risk of stroke. If these findings are proven true, IL-16 could be a useful target to help reduce complications from large artery atherosclerosis.

Stroke and uncontrolled risk factors in Sri Lanka, a sample of the urban developing world

Neal S. Parikh, MD

Chang T, Gajasinghe S, and Arambepola C. Prevalence of Stroke and Its Risk Factors in Urban Sri Lanka: Population-Based Study. Stroke. 2015

In this issue of Stroke, researchers seek to establish the prevalence of stroke in Colombo, Sri Lanka. Though the aim may appear narrow at first glance, the results are striking and deserve the attention of those interested in global health and health disparities.

The researchers conducted a cross-sectional, survey-based study of the adult inhabitants of densely-populated Colombo, Sri Lanka. The survey they used is a validated tool and depends on patient report of clinical symptoms consistent with stroke. Their methodology was designed to comprehensively sample the diversity of Colombo. A remarkable 96.9% response rate yielded 2,313 people in 782 households. 

The prevalence was 10.4 strokes per 1000 adults. Stroke occurred in a 2:1 ratio favoring men and at a young mean age of 58.2 years. Two-thirds of stroke survivors experienced changes or cessation of employment. It is noteworthy that 62.5% of these patients had hypertension, 33.3% had diabetes and 45.8% were smokers.

If we are to extrapolate from this data and other data from India and China cited by the authors, young men and women in the urbanizing developing world are subject to high rates of stroke at young age and often face unemployment subsequently. In parts of the world with limited acute stroke treatment due to resource and infrastructure limitations, the importance of addressing readily modifiable risk factors such as hypertension and smoking cannot be overstated.

Serum leptin is not associated with stroke risk.

Dr. Saber and colleagues examine the longitudinal association between serum leptin levels and incident stroke in older individuals in the Framingham Heart Study.

The authors cite evidence for leptin as having a role in multiple metabolic processes, inflammation and platelet aggregation. It also seems that leptin has been variably associated with MI risk but also protection from coronary disease. Their purpose was to add to the small number of existing, limited studies exploring the relationship between leptin and stroke.

757 patients (mean age 79, 470 women) recruited from 1992 to 1994 were included and followed for incident stroke for up to 10 years. There were 119 strokes, of which 99 were ischemic. Atrial fibrillation was not a well-established stroke risk factor at that time, so baseline vascular risk factor covariates were limited to age, sex, systolic blood pressure, anti-hypertensive therapy, smoking status, diabetes and cardiovascular disease. Dyslipidemia was not included, either.

Multivariate Cox regression tested the relationship between sex-standardized log-leptin levels (to account for right-shifted distribution) and incident stroke. There was no association between baseline leptin and incident stroke or ischemic stroke. In patients with high waist-to-hip ratios, there was an inverse relationship between baseline leptin and stroke risk.

Their findings are consistent with two recent prospective studies. As they state, the observed inverse association between leptin and stroke risk in the highest waist-to-hip ratio subjects is intriguing. However, it is challenging to draw any mechanistic or clinical conclusions from this data except that serum leptin does not appear to be associated with stroke risk. Overall, leptin does not seem to be a reliable marker of cardiovascular or cerebrovascular risk.

Cerebral venous thrombosis may be associated with anemia.

Russell Mitesh Cerejo, MD

Coutinho JM, Zuurbier SM, Gaartman AE, Dikstaal AA, Stam J, Middeldorp S, and  Cannegieter SC. Association Between Anemia and Cerebral Venous Thrombosis: Case–Control Study. Stroke. 2015
In the paper titled “Association between anemia and cerebral venous thrombosis (CVT): a case-control study”, Jonathan Coutinho and colleagues set out to evaluate the association of CVT and anemia. They evaluated 156 cases at their institution between 2006 and 2004 and compared their characteristics to 2916 controls, which were taken from the Dutch MEGA study (Multiple Environmental and Genetic Assessment of risk factors for venous thrombosis) from 1999 to 2004.

Mean hemoglobin concentration was lower in cases than in controls (8.06 vs. 8.68 mmol/l, p<0.001). Anemia was present in 27.0 % (41/152) of the cases, significantly more frequently than in controls (6.5%, 189/2916; absolute difference 20.5%, 95% CI 14 – 28%). Patients with CVT more often had been diagnosed with cancer (9.2 vs. 3.7%) compared to controls (table 1). Oral contraceptive use (69.7% vs. 21.1%) and pregnancy/puerperium (5.3 vs. 1.4%) were more frequent in female cases. After adjustment for potential confounders, anemia was significantly associated with CVT (adjusted OR 4.4, 95% CI 2.8-6.9). Exclusion of patients with a recent infection, neurosurgical intervention, or inflammatory bowel disease, did not change the results (adjusted OR 3.9, 95% CI 2.4-6.4). When they excluded patients with a history of malignancy, the association between anemia and CVT remained significant (adjusted OR 3.6, 95% CI 2.2- 5.9). They also found that anemia is a stronger risk factor for CVT in men than in women.

This study sheds light on the possible association between CVT and anemia, which have been reported in prior case reports and small case series. However any causal correlations are difficult to extrapolate from this data. None the less this important association may be useful in monitoring patients at higher risk for CVT.

Genetics of Lacunar Stroke

Russell Mitesh Cerejo, MD

Traylor M, Bevan S, Baron JC, Hassan A, Lewis CM, and Markus HS. Genetic Architecture of Lacunar Stroke. Stroke. 2015 
From Dechambre to CM Fisher, we have come along way in understanding the pathogenesis of lacunar strokes. While vascular risk factors play a major role, little is known about its genetics. Dr. Traylor and colleagues in their paper “The genetic architecture of lacunar stroke” set out to investigate the genetics and heritability of lacunar strokes and its subtypes.

They investigated 1029 subjects with MRI confirmed lacunar strokes and 964 controls. The dataset was genotyped on the Illumina HumanExomeCore array, which contains both exome content (~250,000 SNPs) and common tag SNPs (~250,000 SNPs) found on conventional GWAS arrays, and imputed to 1000 Genomes phase 1. The lacunar strokes were sub-typed as follows: isolated lacunar infarcts (ILI) and multiple infarcts and leukoaraiosis (MLI/LA). They used genetic restricted maximum likelihood (GREML) methods to estimate the proportion of phenotypic variance on the liability scale explained by the genetic relationships between individuals based on common SNPs. They also investigated whether the heritability of lacunar stroke was enriched for regulatory sites in the autosome.

They found that lacunar stroke and its subtypes were significantly heritable with estimates being higher for the MLI/LA subtype and slightly lower for the ILI subtype. This heritable component was significantly enriched for sites affecting expression of genes. The two subtypes of lacunar stroke in isolation, but not in combination, was associated with rare variation in the genome, suggesting that they may have distinct genetic susceptibility factors. Such studies are the stepping-stone to understanding the genetic composition of complex diseases and pave way for future studies.

Using CHADS2 and CHA2DS2-VASc Scores to Predict Atrial Fibrillation after Stroke

Neal S. Parikh, MD

Fauchier L, Clementy N, Pelade C, Collignon C, Nicolle E, and LipGYH. Patients With Ischemic Stroke and Incident Atrial Fibrillation: A Nationwide Cohort Study. Stroke. 2015
In this issue of Stroke, Fauchier and colleagues seek to identify predictors of incident atrial fibrillation (AF) after ischemic stroke. The prediction and therefore timely diagnosis of atrial fibrillation is of significant public health importance given that, as the authors point out, one in five of stroke can be attributed to AF.

The authors utilize a robust dataset that includes all hospitalizations in France. They identified 48,992 patients without known AF discharged after ischemic stroke during 2009 and calculated their CHADS2 and CHA2DS2-VASc scores. Patients with primary cardiac conditions served as a control, though only for a part of the analysis.

Over 15±5 months of follow-up, 4,828 patients were diagnosed with AF during a hospitalization. 7.88 patients per 100-person years developed incident AF after stroke, whereas the rate was 5.91% after a cardiac diagnosis. As a continuous variable, the CHADS2 and CHA2DS2-VASc scores were associated with incident AF with HR 1.70 (95% CI, 1.66-1.75) and 1.45 (1.42-1.48), respectively. The c statistic was a moderate 0.7. Unsurprisingly, of the CHADS2 and CHA2DS2-VASc variables, only age, hypertension, heart failure and vascular disease independently predicted AF.

The major limitations of this study are as follows: 1. They did not determine if CHADS2 and CHA2DS2-VASc scores predict AF in their control group; 2. To call AF diagnosed after a stroke “incident” is misleading – AF may have been present prior to the stroke as well; 3. AF is often asymptomatic – many patients with “incident” AF may never be hospitalized with AF; 4. Cryptogenic stroke patients in 2009 did not routinely undergo prolonged cardiac rhythm monitoring, the results of which would have been an ideal outcome measure.

The authors hope that their findings will help identify patients after a stroke who may benefit from prolonged monitoring for AF. The CHADS2 and CHA2DS2-VASc scores are inadequate for this purpose. If the goal is identification of patients for prolonged cardiac monitoring upon discharge after stroke, the use of CHADS2 and CHA2DS2-VASc scores presents an unnecessary handicap. Stroke patients routinely have additional data including echocardiogram (left atrial dilatation), MRI (e.g. embolic pattern of strokes), telemetry (frequent premature atrial complexes), labs (B-type natriuretic peptide) that may be helpful.

However, when it comes to stroke, secondary prevention is too late; primary prevention is the Holy Grail. It is for this purpose – for general population AF screening purposes – that the CHADS2 and CHA2DS2-VASc scores may be helpful. Unfortunately, the authors did not assess the test characteristics of these scores in their control group.

Unlocking the Genetics of Lacunar Stroke through Genome-wide Association Studies

Danny R. Rose, Jr., MD

Traylor M, Bevan S, Baron JC, Hassan A, Lewis CM, and Markus HS. The Genetic Architecture of Lacunar Stroke. Stroke. 2015
The identification of genetic risk factors has opened new avenues for understanding disease pathogenesis. With the advent of genome-wide association studies (GWAS) over the last decade, insights have been gained into the contribution of single nucleotide polymorphisms (SNPs) to a variety of disease states. The potential for these types of studies to identify therapeutic targets and implicate particular biological pathways has made them particularly enticing to researchers, but previous GWAS for lacunar stroke have been inconclusive. It has been suggested through pathologic and radiographic analyses that lacunar stroke is not a pathologically homogeneous disease, and this phenotypic heterogeneity could obscure potential genetic association in GWAS analyses. In particular, it has been suggested that there may be two distinct lacunar subtypes: isolated lacunar infarcts (ILI) associated with microatheroma and multiple lacunar infarcts and leukoaraiosis (MLA/LA) associated with diffuse small vessel arteriopathy. Traylor et al. investigated the potential heritability of lacunar stroke and its subtypes by conducting a GWAS using a highly phenotyped cohort of MRI confirmed lacunar strokes.

This study included 1,029 Caucasian subjects less than 70 years of age with a clinical lacunar syndrome related to an anatomically compatible lesion on MRI.
Subjects with other potential causes of stroke were excluded, which included those with >50% stenosis of intracranial or extracranial vessels, previous carotid endarterectomy, moderate to high probability of cardioembolic stroke by TOAST criteria, cortical infarcts and subcortical infarcts >15mm in diameter, an identified etiology of their stroke (dissection, vasculitis, hypercoagulable state) and with NOTCH3 CADASIL or Fabry mutations.Those with lacunar disease were compared to age and sex stratified controls free of clinical cerebrovascular disease. Lacunar subjects were subtyped into two groups based on the presence and severity of leukoaraiosis using the semiquantitative Fazekas scale. Both subjects with lacunar disease and controls were assessed for known risk factors associated with lacunar stroke including hypertension, diabetes, hyperlipidemia, and smoking status.

Heritability was determined using genetic restricted maximum likelihood (GREML) analyses to estimate the proportion of phenotypic variance explained by the genetic relationships among subjects with common SNPs. This was performed versus controls for all MRI defined lacunar strokes as well as the two subtypes described previously. The authors also analyzed identified SNPs using information from a database of genotype-tissue expression studies that seeks to identify “functional” sites in the genome that influence expression, as well as sites where regulatory factors are thought to bind to the genome as evidenced by matched transcription factor binding sites and DNAse footprints. Lastly, the authors tested for the potential polygenic contribution from rare variation by calculating the ratio of risk to protective variants for all lacunar stroke cases and both subtypes versus controls.

It was found that lacunar stroke verified by MRI in the selected subjects was highly heritable for the population in aggregate (h2=0.2 0-0.25, p=0.00054) as well as the ILI (h2=0.15-0.18, p=0.029) and MLI/LA subtypes (h2=0.23-0.28, p=0.0035). The point estimate suggested a higher degree of heritability for the MLI/LA subtype, suggesting an increased contribution of genetic factors as compared to the ILI subtype, although this was viewed as speculative and needs to be confirmed. The heritability of lacunar stroke from this study is comparable to other neurologic diseases with strong genetic influences such as Alzheimer’s Disease and multiple sclerosis and should prompt additional research into using GWAS to gain more insights into the pathophysiology and genetic risk factors for this disease. This is further strengthened by the finding that a significant portion of the SNPs identified were found to have roles in either gene expression or regulation. Their analysis regarding polygenic contribution of rare variation showed an increase in the ratio of risk variants to protective variants at low allele frequencies in both subtypes of lacunar stroke, but not in both subgroups combined. This suggests a distinct role for the variants implicated for each of the pathologic subtypes.

This study represents a significant step forward in our understanding of the genetic basis of lacunar stroke and leukoaraiosis. The authors’ choice of including only MRI confirmed disease as well as the decision to divide lacunar stroke into distinct pathologic subtypes for analysis appears to be a large contributing factor to the success of this study compared to previous GWAS evaluating the genetic component of lacunar stroke. Additionally, their findings showing significant contribution of rare variants to both subtypes individually, but not as a whole, provides more evidence that these do represent distinct subtypes with potentially different pathologic mechanisms. Their successful precedent for separating lacunar strokes into these groups will likely influence future study design when evaluating the lacunar stroke population as a whole to determine novel associations and findings that may be specific to one group. Although this study was limited by a somewhat small sample size and the inherently incomplete nature of functional genetic databases, it represents a promising glimpse into the genetic basis of small vessel disease.