Donate Help Contact The AHA Sign In Home
American Heart Association
Stroke
Search: search_blue_button Advanced Search
« Previous Article | Table of Contents | Next Article »
Stroke. 2004;35:212-227
Published online before print December 18, 2003, doi: 10.1161/01.STR.0000107187.84390.AA
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
35/1/212    most recent
01.STR.0000107187.84390.AAv1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Floßmann, E.
Right arrow Articles by Rothwell, P. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Floßmann, E.
Right arrow Articles by Rothwell, P. M.

(Stroke. 2004;35:212.)
© 2004 American Heart Association, Inc.

Comments, Opinions, and Reviews

Systematic Review of Methods and Results of Studies of the Genetic Epidemiology of Ischemic Stroke

Enrico Floßmann, MRCP; Ursula G.R. Schulz, MD, MRCP Peter M. Rothwell, MD, PhD, FRCP

From the Stroke Prevention Research Unit, Department of Clinical Neurology, Radcliffe Infirmary, Oxford, UK.

Correspondence to Dr P.M. Rothwell, Stroke Prevention Research Unit, Department of Clinical Neurology, Radcliffe Infirmary, Woodstock Rd, Oxford 0X2 6HE, UK. E-mail peter.rothwell{at}clneuro.ox.ac.uk


*    Abstract
up arrowTop
 *Abstract
down arrowIntroduction
 down arrowMaterials and Methods
 down arrowResults
 down arrowDiscussion
 down arrowReferences
 
Background and Purpose— To design appropriate molecular genetic studies, we first need to understand the genetic epidemiology of stroke. We therefore performed a systematic review of the literature to assess the heritability of stroke according to methodological quality of studies and to determine any heterogeneity in findings between studies and possible publication bias.

Methods— We searched for twin studies and studies of family history of stroke using bibliographic databases and by hand-searching reference lists and journals. Odds ratios (ORs) for family history as a risk factor for stroke were calculated within studies and combined by meta-analysis. Heterogeneity between studies, methodological quality of studies, and the influence of the age at which stroke occurred in both probands and relatives were assessed.

Results— We identified 53 independent studies (3 twin, 33 case-control, and 17 cohort). Monozygotic twins were more likely to be concordant than dizygotic twins (OR, 1.65; 95% CI, 1.2 to 2.3; P=0.003). A positive family history was a risk factor for stroke in both case-control (OR, 1.76; 95% CI, 1.7 to 1.9; P<0.00001) and cohort (OR, 1.30; 95% CI, 1.2 to 1.5; P<0.00001) studies. However, there was major heterogeneity between studies (cohort P=0.0001; case-control P<0.00001), with much stronger associations in small studies and methodologically less rigorous studies. Moreover, studies that reported insufficient data to allow meta-analysis tended to have found weaker associations. Family history of stroke was more frequent in studies that were confined to probands or relatives aged <70 years. However, few studies considered the number of affected and unaffected relatives, only 2 studies considered stroke phenotypes in detail, and only 19 studies (38%) adjusted associations for intermediate phenotypes. No twin study, only 5 cohort studies (26%), and 20 case-control studies (61%) differentiated between ischemic and hemorrhagic stroke in the proband. Family history of stroke was more frequent in large- and small-vessel stroke than in cardioembolic stroke. There were very few data on the influence of family history on stroke severity and no data on stroke recovery.

Conclusions— Twin studies suggest a small genetic contribution to stroke, but reliable interpretation of published family history studies is undermined by major heterogeneity, insufficient detail, and potential publication and reporting bias. More detailed large-scale genetic epidemiology is required.


Key Words: family • meta-analysis • publication bias • review literature • stroke, ischemic • twin studies


*    Introduction
up arrowTop
 up arrowAbstract
 *Introduction
down arrowMaterials and Methods
 down arrowResults
 down arrowDiscussion
 down arrowReferences
 
Afamily history of stroke is regarded as an important risk factor for the development of cerebrovascular disease, and strokes cluster in families.1 Animal models suggest that susceptibility to ischemic stroke is influenced by genetic factors,2 and there are a number of rare mendelian stroke syndromes in humans.3,4 However, several dozen candidate gene studies have produced no consistent results, and data on the genetic epidemiology of stroke are conflicting.4–6 One reason for this is that although ischemic stroke is a highly complex trait, few studies have assessed stroke subtypes, and no studies have been properly powered to do so. Moreover, many studies have combined ischemic and hemorrhagic strokes. It is unlikely that these very different pathological conditions are under the same genetic influences. Of equal importance, little account has been taken of the role of intermediate phenotypes (eg, hypertension, hyperlipidemia, diabetes, carotid stenosis), many of which have a substantial genetic component themselves. Furthermore, a positive family history could be the result of shared genes, shared environment, or both.

Several genome screens and other detailed molecular genetic studies of ischemic stroke are planned. However, if these are to be properly targeted, it is essential that we first understand the basic genetic epidemiology. Recent reviews of the genetics of ischemic stroke have not sought to be systematic and have cited only 11 to 16 family history studies and 1 twin study.4–6 We performed a systematic review of all published twin studies and studies of family history as a risk factor for stroke. We sought to estimate effect sizes, to determine the extent of heterogeneity in the strength of associations between studies, and to explain any differences by assessing the quality of studies, searching for possible publication bias, and exploring whether effects varied between stroke subtypes, age of stroke onset in probands and relatives, and the presence of intermediate phenotypes.


*    Materials and Methods
up arrowTop
 up arrowAbstract
 up arrowIntroduction
 *Materials and Methods
down arrowResults
 down arrowDiscussion
 down arrowReferences
 
Search Strategy
We sought to identify articles that reported on family history as a potential risk factor for stroke. Studies were identified by 2 independent observers from MEDLINE+ and EMBASE (Silverplatter Winspirs 4.0 online and Entrez PubMed NIH 06/08/2001 for 1966 to May 2003) with the following search terms: family history AND (stroke OR CVA OR TIA OR cerebrovascular) and twin AND (stroke OR CVA OR TIA OR cerebrovascular). No restriction was made on the language of publication. Journals that yielded >10% of all studies identified electronically were systematically hand-searched for further relevant studies published after 1980. The reference lists of all articles that met the inclusion criteria were searched. We contacted authors personally if their publications were unavailable in the United Kingdom.

Inclusion Criteria
Articles were included in the review if they fulfilled the following criteria: (1) study was a prospective cohort, case-control, or twin study; and (2) study reported on the frequency of a positive family history of stroke for patients with stroke and a control group for case-control studies, or the risk of stroke (either undifferentiated or ischemic) in patients with and without a positive family history of stroke for prospective cohort studies, or the concordance for stroke in monozygotic versus dizygotic twins.

Data Extraction
The following data were extracted from eligible reports by 2 independent observers with a structured questionnaire: (1) type of study (case-control, cohort, or twin); (2) setting (hospital or population based); (3) length of follow-up (for cohort studies); (4) details of patient selection (Were patients recruited consecutively or by type of stroke? Were patients with transient ischemic attack [TIA] or subarachnoid hemorrhage included?); (5) details of collected family history (Did studies also collect information on family history of ischemic heart disease, hypertension, peripheral vascular disease, TIA, or diabetes mellitus? Were only fatal events included? Which relatives were included in a positive family history? Were the age and number of affected relatives and the size of the family taken into account? What method was used to collect a family history?); (6) details of stroke classification in patients (ischemic versus hemorrhagic; subtyping of ischemic strokes according to the TOAST7 [or similar] criteria); (7) consideration of whether other potential risk factors (age at event, personal history of hypertension, diabetes mellitus, ischemic heart disease, peripheral vascular disease, carotid disease, cardioembolism, smoking, cholesterol, alcohol consumption, body mass index, use of hormone replacement therapy and/or oral contraceptive pill, social class/education, employment status, marital status, amount of exercise) collected and related to the presence of family history of stroke; (8) results and conclusion of each study (Was a positive family history found to be a risk factor for stroke? If so, was there a difference for the various stroke subtypes? Was a positive family history associated with any confounders or established risk factors or environmental factors? Did the study correct for these potential confounders? Did the study consider genetic versus environmental factors in a positive family history?); and (9) consideration of whether studies collected information on the influence of family history of stroke on stroke severity and recovery.

Studies that gave absolute numbers for patients with family history of stroke were identified. Only patients with ischemic stroke were counted if the subtypes were differentiated. A parental family history of stroke was used for the meta-analysis in which family history was reported separately for whether strokes had occurred in parents or siblings and in which it was not possible to derive an overall estimate for family history in first-degree relatives (FDR).

Statistical Analysis
Odds for a positive family history as a risk factor for stroke were calculated within individual studies. Where appropriate, odds ratios (ORs) from separate studies were combined by fixed-effects meta-analysis according to the Mantel-Haenszel method. Heterogeneity between studies was calculated with the {chi}2 method. The possibility of publication bias was assessed by checking funnel plots of the inverse standard error (precision) of studies versus the OR for asymmetry and by simple linear regression analysis of the standard normal deviate (lnOR/SE) versus precision (1/SE). The predicted regression line assuming no bias was compared with the actual simple regression fit (P<0.10 was regarded as significant).8 Both cohort and case-control studies measure the odds of stroke that are conferred by a positive family history of stroke. We therefore tested for potential publication bias of both study types individually and combined.

We predefined a simple score to assess the methodological quality of studies, giving 1 point each if a study (1) recruited patients consecutively; (2) defined the relatives who contributed to the family history; (3) separated family histories of stroke, hypertension, or myocardial infarction (MI); (4) took the age of onset of stroke in relatives into account; (5) took the number of affected relatives into account; (6) distinguished between ischemic and hemorrhagic strokes; and (7) subtyped ischemic strokes by etiology. Studies were then stratified into 3 strata if they met <3, 3 to 5, and >=6 quality criteria. Heterogeneity of OR estimates between the resulting strata was calculated. To detect heterogeneity in relation to the power of studies, we stratified them into tertiles according to the inverse of their standard errors. A possible correlation between study quality and precision was also tested for by the rank correlation test of Spearman.

To identify other sources of heterogeneity, year of publication (divided into studies published before and after 1990), type of study (prospective cohort or case-control), median age of subjects, maximum age of subjects, maximum age of relatives, whether hemorrhagic strokes were included or not, and family history of fatal stroke only were tested versus the natural logarithm of the OR weighted by the inverse variance.


*    Results
up arrowTop
 up arrowAbstract
 up arrowIntroduction
 up arrowMaterials and Methods
 *Results
down arrowDiscussion
 down arrowReferences
 
The electronic literature search yielded 889 publications with the search terms family history AND (stroke OR CVA OR TIA OR cerebrovascular) and 105 publications for the search terms twin AND (stroke OR CVA OR TIA OR cerebrovascular). Duplicate records were removed, and abstracts were reviewed. Ninety-seven publications were judged potentially relevant by 1 or both observers. Review of reference lists of these articles and a hand-search of the journal Stroke for the years 1980–2003, which was the only journal that met the criteria for hand-searching, yielded an additional 127 publications. Thus, a total of 224 articles (9 in languages other than English) were considered in detail. Two of the articles were published in Spanish, 2 in Italian, 2 in German, and 1 each in Russian, Chinese, and Hungarian.

Sixty publications satisfied our inclusion criteria. They consisted of 3 twin studies that investigated the concordance of stroke in monozygotic versus dizygotic twins (1 group published a follow-up report on their original article as an abstract),9–12 19 articles that published data on 17 independent cohort studies,13–28,65–67 and 37 articles that published data on 33 independent case-control studies.29–64,68 Of these, 9 cohort studies15,18,20,22–25,28,66 and 6 case-control studies33,34,36,40,43,54 only reported relative risks or ORs and gave no absolute numbers and therefore could not be included in any meta-analysis. However, details of these studies are given in Table 1Down. Thus, 3 twin studies,9–12 9 cohort studies,* and 27 case-control studies{dagger} reported sufficient data to allow inclusion in the meta-analysis. One additional cohort study was not included because outcome strokes in probands were defined only as MRI evidence of infarction without data on clinical stroke outcomes.69


View this table:
[in this window]
[in a new window]
  		TABLE 1. Characteristics and Results of Cohort and Case-Control Studies Not Included in the Formal Meta-Analysis


View this table:
[in this window]
[in a new window]
  		TABLE 1. Continued

Twin Studies
The characteristics and main results of the studies are summarized in Table 2. The 2 Scandinavian twin studies9,12 used national death or hospital discharge registries to identify twins with stroke. The original report of the American study employed a mailed questionnaire10; the follow-up report additionally used a telephone health screen and a mortality review.11 None of the twin studies distinguished between stroke subtypes or assessed confounding by other risk factors. In the meta-analysis, monozygotic twins were more likely to be concordant for stroke than dizygotic twins (OR, 1.65; 95% CI, 1.2 to 2.3; P=0.003; heterogeneity P=0.3; Figure 1).


View this table:
[in this window]
[in a new window]
  		TABLE 2. Characteristics of Twin Studies



View larger version (9K):
[in this window]
[in a new window]
  		Figure 1. Odds for concordance for stroke in monozygotic (MZ) vs dizygotic (DZ) twins. Left, Meta-analysis with latest data for all published twin studies, including follow-up report for the Veterans Affairs cohort (Brass et al,11 1996). Right, Meta-analysis with original report from the Veterans Affairs twin cohort (Brass et al,10 1992). Sig indicates significance; het, heterogeneity.

Case-Control Studies
Characteristics of the case-control studies included in the meta-analysis are summarized in Table 3Down. Of these, 15 independent studies differentiated between ischemic and hemorrhagic strokes,{ddagger} and 4 studies subtyped ischemic strokes further into cardioembolic, large-artery disease, small-artery disease, and undetermined,49,59 into cortical artery occlusion and perforating artery occlusion,60 or included only patients with large-vessel stroke.46 Most studies defined which relatives were considered in a family history (mostly FDR: parents and/or siblings), but 3 studies did not.29,55,58 Two studies used a pooled family history of stroke or hypertension or MI.48,53


View this table:
[in this window]
[in a new window]
  		TABLE 3. Characteristics of Case-Control Studies Included in the Meta-Analysis


View this table:
[in this window]
[in a new window]
  		TABLE 3. Continued

The combined OR for a positive family history of stroke as a risk factor for stroke was 1.76 (95% CI, 1.7 to 1.9; P<0.00001; 5991 patients). However, there was significant heterogeneity between studies (P<0.00001; Figure 2). The funnel plot of the OR versus the variance was asymmetrical (Figure 3), and there was evidence of possible publication bias on simple linear regression of the standard normal deviate against precision (P=0.003). However, studies cited in recent reviews4–6 had an overall similar association of family history of stroke with stroke risk (OR, 1.56; 95% CI, 1.4 to 1.8; P<0.00001; heterogeneity P=0.05; 1665 patients). The OR did not change significantly when only the 15 studies exclusively investigating ischemic stroke were combined (OR, 1.75; 95% CI, 1.6 to 2.0; P<0.00001; heterogeneity P=0.005; 3800 patients).



View larger version (39K):
[in this window]
[in a new window]
  		Figure 2. Odds for stroke patients vs controls to have a positive family history (FHx) of stroke (ordered by variance).



View larger version (10K):
[in this window]
[in a new window]
  		Figure 3. Funnel plot of OR of family history of stroke as a risk factor for stroke vs precision (ie, inverse of the standard error of the OR) in case-control (full circles) and cohort studies (empty circles). Note the asymmetry of the plot due to a lack of low precision estimates with odds <1 (ie, small negative studies).

Studies that fulfilled at least 6 methodological quality criteria found a weaker association with family history of stroke (OR, 1.28; 95% CI, 1.1 to 1.5; P=0.01; heterogeneity P=0.27; 1221 patients) than studies that fulfilled more than half but <6 (OR, 1.88; 95% CI, 1.7 to 2.0; P<0.00001; heterogeneity P=0.0001; 4280 patients) and studies that fulfilled less than half the methodological quality criteria (OR, 2.40; 95% CI, 1.8 to 3.2; P<0.00001; heterogeneity P=0.0001; 490 patients). The difference between the 3 strata was significant (P=0.00006; Figure 4). There was no correlation (r=0.085, P=0.67) between the inverse standard error (precision) of a study and the quality score.



View larger version (20K):
[in this window]
[in a new window]
  		Figure 4. Odds for stroke patients to have a positive family history (FHx) of stroke. Studies are stratified by quality criteria (A) and inverse of their standard error (B). Sig indicates significance; het, heterogeneity.

Ten studies adjusted the ORs for a family history of stroke at least partially for the presence of other established risk factors of stroke in patients,§ and 1 study62 adjusted the ORs for the presence of risk factors for stroke in relatives. They found with 3 exceptions35,49,58 that adjustment either attenuated the association with family history of stroke or that family history was no longer a significant risk factor after adjustment (Table 4). Two studies that investigated the influence of environmental factors other than smoking on family history of stroke as a risk factor for stroke found no significant association.35,64 Ten|| of 12 studies56,59 investigating a possible interaction between family history of stroke and hypertension in patients found a significant association; 3 studies found an association with smoking,35,44,52 3 found an association with diabetes mellitus,37,44,52 4 found an association with the presence of ischemic heart disease or vascular disease,31,37,46,52 and 1 found an association with the presence of hyperlipidemia.42 Two studies found that family history of stroke was significantly related to a cluster of vascular risk factors.42,37


View this table:
[in this window]
[in a new window]
  		TABLE 4. Comparison of Crude, Age- and Sex-Adjusted, and Multivariate Adjusted (for Stroke Risk Factors) OR in Cohort and Case-Control Studies Included in Meta-Analysis

Only 2 recent studies49,59 compared the association of a positive family history of stroke between ischemic stroke subtypes. Their findings and the findings of 1 additional study75 that did not include a control group and could therefore not be included in the meta-analysis are summarized in Table 5. No case-control study investigated the influence of family history of stroke on stroke severity or recovery.


View this table:
[in this window]
[in a new window]
  		TABLE 5. Two Case-Control Studies and 1 Observational Study That Compared Frequency of Family History of Stroke by Subtypes of Ischemic Stroke in the Proband According to the TOAST Criteria7

Prospective Cohorts
Characteristics of the cohort studies included in our meta-analysis are summarized in Table 6. One study21 only reported ischemic stroke outcomes, and 1 study examined a cohort of TIA patients.14 None of the remainder distinguished between hemorrhagic and ischemic strokes. All but 3 studies13,14,67 defined which relatives were considered in the family history (FDR, mainly parents).


View this table:
[in this window]
[in a new window]
  		TABLE 6. Characteristics of Prospective Cohort Studies Included in the Meta-Analysis

Individuals with a positive family history of stroke had a slightly higher risk of subsequent stroke than those without a family history (OR, 1.30; 95% CI, 1.2 to 1.5; P<0.00001; 1906 stroke outcomes). However, there was major heterogeneity between studies (P=0.0001; Figure 5). The funnel plot of the OR versus the variance was asymmetrical (Figure 3), and there was borderline significant evidence of potential publication bias (P=0.10). However, those studies cited in the recent reviews4–6 reported significantly (P=0.00007) higher odds (OR, 1.76; 95% CI, 1.5 to 2.1; P<0.00001; heterogeneity P=0.18) than studies that were not cited (OR, 1.12; 95% CI, 1.0 to 1.3; P=0.16; heterogeneity P=0.05). There was also possible reporting bias in that the majority of cohort studies that we could not include in the formal meta-analysis because they gave no detailed numbers failed to find significant associations of family history of stroke with stroke risk.15,20,22–24,28,66



View larger version (19K):
[in this window]
[in a new window]
  		Figure 5. Odds for subsequent stroke in subjects with a positive family history (FHx) of stroke in cohort studies (ordered by variance).

Two studies examined the influence of family history of stroke on stroke severity and found a greater influence on less severe strokes,21,26 but no study investigated an association between family history of stroke and stroke recovery.

Five studies adjusted at least partially for potential confounding by established stroke risk factors in the subjects. One study did not report the effect of adjustment.17 The association was attenuated after adjustment in 2 studies21,26 but did not show any major change in the other 2 studies16,27 (Table 4). Two studies also adjusted for potential environmental confounding other than smoking. One of these found that the risk conferred by a positive family history of stroke was independent of the socioeconomic status,16 and the other26 found that patients with a family history of stroke were more likely to be manual workers but found no interaction between family history of stroke and heavy alcohol drinking or physical activity.

Three studies found a significant interaction between family history of stroke and presence of hypertension in the studied probands.16,17,26 One study also found a positive association between family history of stroke and increased body mass index and cholesterol levels but failed to find an association with smoking or diabetes mellitus.26

Influence of Age
Four case-control and 3 prospective cohort studies reported a family history of stroke in relatives at young age. Jerrard-Dunne et al49 and Fonte et al39 provided details about stroke onset in FDR before the age of 65 years, Margaglione et al53 reported about stroke or MI in male FDR aged <55 years and female FDR aged <60 years, and Diaz et al37 reported family history of stroke onset in siblings aged <70 years. The combined OR for these 4 studies was nonsignificantly (P=0.35) higher (OR, 1.82; 95% CI, 1.4 to 2.3; P<0.00001; heterogeneity P=0.78; 1311 patients; Figure 6) than the remainder (OR, 1.61; 95% CI, 1.5 to 1.7; P<0.00001; heterogeneity P<0.00001). Both cohort studies26,27 that stratified their analyses by whether parental stroke occurred before or after age 70 years found a larger effect on stroke risk for parental stroke before 70 years. It was not possible, however, to perform a meta-analysis on these studies because they did not report absolute numbers for each stratum separately. A third study that only considered family history of stroke in parents aged <60 years16 found an overall OR for stroke of 1.89 in men and 1.80 in women.



View larger version (29K):
[in this window]
[in a new window]
  		Figure 6. Odds for stroke patients vs controls to have a positive family history (FHx) in FDR aged <65 years (top) and odds for stroke patients aged <70 years to have a positive family history of stroke vs controls (bottom).

Five case-control studies41,48,50,63,64 recruited only patients aged <70 years, 1 study recruited patients with stroke onset at <65 years,60 1 study recruited patients with stroke onset at <50 years,55 and 1 study reported family histories separately for patients aged <54 and >=54 years.68 The combined OR for these younger patient groups was nonsignificantly (P=0.10) higher (OR, 1.93; 95% CI, 1.7 to 2.2; P<0.00001; heterogeneity P=0.0001; 1240 patients; Figure 6) than the remainder (OR, 1.69; 95% CI, 1.6 to 1.8; P<0.00001; heterogeneity P=0.00001).

Two prospective cohort studies14,17 reported the influence of family history of stroke on the subgroup of individuals who had their strokes when aged <65 or 70 years, respectively, during follow-up. They found a reduced effect in younger probands (OR, 0.41; 95% CI, 0.2 to 0.8; P=0.29; heterogeneity P=0.85; 38 stroke outcomes); however, the number of strokes was very small. In addition, Wannamethee et al26 reported that the increased risk of stroke with parental death from stroke was apparent in older men (aged 50 to 59 years at baseline) but not in younger men (aged 40 to 49 years at baseline). In contrast, Jousilahti et al16 found a higher relative risk of family history of stroke for both men and women aged <50 years compared with older individuals.

Influence of Number of Affected Relatives
Two case-control studies46,64 and 1 prospective cohort study26 reported the effect if both parents had been affected by stroke26,64 or vascular disease.46 The combined OR for a family history of stroke in the case-control studies increased significantly (P=0.018) when they were analyzed separately for a family history in only 1 parent (OR, 1.08; 95% CI, 0.7 to 1.6; P=0.77; heterogeneity P=0.35) versus a family history of both parents (OR, 2.45; 95% CI, 1.4 to 4.2; P=0.002; heterogeneity P=0.18). The number of subjects with 2 affected parents in the cohort study was too small to allow meaningful analysis.

Test for Heterogeneity
Cohort studies (P=0.069), higher maximum age of patients (P=0.071), and higher maximum age of relatives (P=0.059) were associated with a lower OR on univariate analysis of potential sources of heterogeneity. These parameters still had a significant inverse association if tested together in a multivariate model (type of study, P=0.032; maximum age of relatives, P=0.021; maximum age of patients, P=0.018) and accounted for approximately 34% of the total heterogeneity between studies (adjusted R2=0.344).

There was highly significant (P<0.0001; Figure 3) evidence of possible publication bias if both case-control and cohort studies were tested together. When stratified into tertiles according to the inverse of the standard error (precision) of their individual estimates, the most power-ful (largest) studies found significantly (P<0.00001) smaller ORs (OR, 1.48; 95% CI, 1.4 to 1.6; P<0.00001; heterogeneity P<0.00001) than studies of intermediate power (OR, 1.82; 95% CI, 1.6 to 2.0; P<0.00001; heterogeneity P=0.005) and low power (OR, 2.30; 95% CI, 1.9 to 2.8; P<0.00001; heterogeneity P=0.24; Figure 4).


*    Discussion
up arrowTop
 up arrowAbstract
 up arrowIntroduction
 up arrowMaterials and Methods
 up arrowResults
 *Discussion
down arrowReferences
 
A number of epidemiological studies of twins, affected sibling pairs, and family history data have suggested that there are significant genetic influences in stroke. Previous nonsystematic reviews of genetic epidemiology of stroke had identified 11 to 16 family history studies and 1 twin study. Ours is the first systematic review, and we identified 50 independent family history studies and 3 independent twin studies.

Twin studies provide the most reliable evidence of genetic influence in complex diseases, and they are best suited to disentangle genetic influences from influences of a shared environment. However, there are difficulties in conducting twin studies in stroke patients. Stroke most commonly affects old people, which makes it challenging to recruit enough twin pairs and increases the chance of twins dying of other unrelated diseases. Thus far, only 3 twin studies have been reported, and they reached different conclusions. Our meta-analysis showed a small genetic influence on the risk of stroke, with monozygotic twins being 1.6 times more likely to be concordant for stroke than dizygotic twins. This should be compared with a 3- to 5-fold better correlation of blood pressure among monozygotic twins versus dizygotic twins70,71 and a 10-fold higher concordance of monozygotic versus dizygotic twins in studies of multiple sclerosis.72 The most frequently quoted twin study of stroke, which used a mailed questionnaire for case ascertainment, found a 17.7% concordance in monozygotic twins versus 3.6% in dizygotic twins.10 This higher concordance might partially be explained by the younger age of the twins in this study compared with the other twin studies, which mainly explored concordance of fatal stroke, suggesting an increase in the relative influence of hereditary factors on stroke risk in younger patients. However, this initial study was based on only 8 stroke-concordant twin pairs, and on 10-year follow-up this difference had diminished to 12.8% and 8.0%, respectively, which is in good agreement with the other 2 studies. Interestingly, in a study on heritability of coronary heart disease among Swedish twins, Marenberg et al73 found a very strong association of age and genetic liability with a strong genetic component among young twins and a weak component among old twins. None of the twin studies of stroke differentiated between stroke subtypes or collected information on other risk factors to assess potential confounders.

We also found that a family history of stroke was only a moderate risk factor for stroke in case-control and cohort studies. Case-control studies found a higher risk compared with prospective cohort studies, but the results of the highest quality case-control studies were very similar to the findings of the prospective cohort studies (OR, 1.28; 95% CI, 1.1 to 1.5 versus OR, 1.30; 95% CI, 1.2 to 1.5). Prospective cohort studies eliminate recall bias, which is one of the most important biases in family history studies, and they are therefore more likely to determine the true underlying effect. Nevertheless, there was major heterogeneity between studies within both the case-control and the prospective cohort groups.

There was evidence of possible publication bias in both case-control and prospective cohort studies, with larger studies reporting more conservative estimates than small studies. There was also evidence of reporting bias. However, it is likely that there was true heterogeneity between studies. First, there is some evidence that genetic factors are less important in strokes that occur later in life. Studies that only considered a family history of stroke to be positive if the affected relatives were younger than a certain age16,39,53 found a stronger association than the pooled estimate. Moreover, studies that stratified their analysis by the age at which strokes occurred in relatives found a stronger effect of family history of stroke in the group of relatives affected at a younger age.16,25–27,37,49 There is more uncertainty in studies investigating young stroke patients. Studies investigating only young and middle-aged patients with stroke41,48,50,55,60,63,64 found a greater influence of family history of stroke compared with studies including older patients, but there was disagreement between the studies that stratified their patients by age and reported on the effect of family history of stroke in different age strata.

Second, not all stroke types are likely to be equally heritable, and case mix is therefore likely to influence the results. For instance, subarachnoid hemorrhage, which was included in some of the studies, might have a different genetic component compared with other strokes.74,50 Many studies did not differentiate between ischemic and hemorrhagic stroke, and only 2 studies49,59 subtyped ischemic stroke sufficiently to compare a family history between subtypes. Large- and small-vessel disease was more strongly associated with a positive family history of stroke than the remainder of ischemic strokes. One additional study75 gave details of a family history of stroke in carefully subtyped stroke patients but did not include a control group. The exact stroke phenotype among affected family members is very difficult to ascertain, and this could mask stronger associations of less frequent stroke subtypes.

Third, few studies investigated the potential influence of the number of affected relatives or controlled for family size. The impact of a family history of stroke increased with the number of affected FDR in the studies25,46,64 that reported on this. Few studies corrected their findings for this effect.

Fourth, the influence of family history on the risk of stroke might vary in different populations and ethnic groups and may also vary over time as environmental factors, dietary habits, and levels of deprivation change. A considerable degree of the heritability of stroke appears to be conferred by the heritability of risk factors and intermediate phenotypes (see below). The introduction of preventive treatments such as antihypertensive drugs, statins, and carotid endarterectomy could have attenuated the penetrance of stroke. Interestingly, the strongest associations were found in studies from the 1960s,41,46 mainland China,51,58 and Russia.48

Fifth, genetic factors could have an influence on stroke severity. Only 2 studies21,26 investigated the influence of family history of stroke on stroke severity. Both found an increased influence on less severe strokes. However, there was no consistency in the results between studies that only considered a family history of fatal stroke. Four studies26,27,41,46 found a positive association with family history of fatal stroke, whereas 4 others18,23,54,62 failed to find any association. Reed et al69 found a positive correlation between a family history of stroke and MRI stroke volume on purely imaging-defined strokes in probands.

Finally, several other methodological issues could potentially affect the strength of association between family history of stroke and risk of stroke. The choice of control groups (eg, whether they included patients with other vascular diseases) is likely to affect the magnitude of association. In addition, the ascertainment of family history is difficult, and different methodologies to collect family history are prone in varying degrees to bias and inaccuracies.76,77

Most established risk factors and intermediate phenotypes for stroke, such as hypertension, diabetes mellitus, ischemic heart disease, hypercholesterolemia, smoking, and carotid stenosis, are likely to run in families.78–85 A family history of stroke was associated with a higher proportion of hypertension# and other conventional risk factors in patients, including smoking, and most studies found that adjustment for vascular risk factors diminished the associations between family history and stroke. In addition, Nicolaou et al86 found a higher occurrence of hypertension and stroke in parents of hypertensive probands compared with controls; Williams et al1 found that positive family histories of ischemic heart disease, stroke, hypertension, and diabetes mellitus were significantly associated with each other; and finally, Lestro-Henriques et al87 found that a family history of cardiac or cerebrovascular disease was more common in hypertensive stroke patients than in normotensive stroke patients. Furthermore, in 2 other studies, a family history of stroke was significantly related to a cluster of vascular risk factors.37,42 We were unable, however, to precisely estimate the contribution of the heritability of risk factors to the heritability of stroke. Most studies reported that heritability of stroke was independent of hypertension (or at least independent of a history of hypertension or a single measurement of blood pressure).16,17,26,27,49,52,58,59 The collection of family histories of risk factors poses further methodological difficulties since hypertension, type II diabetes mellitus, and hypercholesterolemia were significantly underdiagnosed in the past, and diagnostic criteria have changed. The contribution of these risk factors to the heritability of stroke may therefore be greater than previously thought.

We were not able to assess the role of environmental factors, eg, smoking, eating, and drinking habits or social status in a positive family history of stroke and to what degree they modulate genetic influences.

Conclusions
We identified 53 independent studies of the genetic epidemiology of ischemic stroke. Taken together, these studies suggested a small genetic contribution to stroke occurrence, but there was major heterogeneity between studies, with much stronger associations in small studies and methodologically less rigorous studies. Moreover, studies that reported insufficient data to allow meta-analysis tended to have found weaker associations. It is possible that the genetic contribution is larger for certain subtypes of stroke. However, only 2 studies considered stroke phenotype in detail, and many studies did not differentiate between ischemic and hemorrhagic stroke in the proband. There was some evidence that the genetic contribution is greater for stroke occurring at a younger age, with stronger associations in analyses confined to probands or relatives aged <70 years. However, few studies considered the number of affected and unaffected relatives, and only a minority of studies adjusted associations for intermediate phenotypes. There were few data on the influence of family history on stroke severity and no data on stroke recovery. More detailed large-scale genetic epidemiological studies are required.


*    Acknowledgments
 
This study was supported by the United Kingdom Medical Research Council (Drs Floßmann and Rothwell) and the Wellcome Trust (Dr Schulz). We would like to thank Robert Cuffe for advice on statistical analysis. We are very grateful to Professor Hugh S. Markus for providing us with the raw numbers from his study of family history in subtypes of ischemic stroke.49


*    Footnotes
 
*References 13,14,16,17,19,21,26,27,67. Back

{dagger}References29–31,35,37–39,41,42,46–53,55,56,58–64,68. Back

{ddagger}References 29,37–39,42,44,46,49–51,53,55,57–60,64. Back

§References 31,35,38,39,49,50,52,58,59,64. Back

||References 29,31,35,37,41,42,49,52,64,68. Back

¶References 9,10,14,16–18,26,27,35,44,47,51,52,59,64,68. Back

#References 16,17,26,31,35,37,41,42,47,49,51,52,55,59,64. Back

Received June 28, 2003; revision received September 10, 2003; accepted September 26, 2003.


*    References
up arrowTop
 up arrowAbstract
 up arrowIntroduction
 up arrowMaterials and Methods
 up arrowResults
 up arrowDiscussion
 *References
 

  1. Williams RR, Hunt SC, Heiss G, Province MA, Bensen JT, Higgins M, Chamberlain RM, Ware J, Hopkins PN. Usefulness of cardiovascular family history data for population-based preventive medicine and medical research (the Health Family Tree Study and the NHLBI Family Heart Study). Am J Cardiol. 2001; 87: 129–135.[CrossRef][Medline] [Order article via Infotrieve]
  2. Huang Z, Huang PL, Panahian N, Dalkara T, Fishman MC, Moskowitz MA. Effects of cerebral ischemia in mice deficient in neuronal nitric oxide synthase. Science. 1994; 265: 1883–1885.[Abstract/Free Full Text]
  3. Tournier-Lasserve E, Iba-Zizen MT, Romero N, Bousser MG. Autosomal dominant syndrome with strokelike episodes and leukoencephalopathy. Stroke. 1991; 22: 1297–1302.[Abstract/Free Full Text]
  4. Hassan A, Markus HS. Genetics and ischaemic stroke. Brain. 2000; 123: 1784–1812.[Abstract/Free Full Text]
  5. Hademenos GJ, Alberts MJ, Awad I, Mayberg M, Shepard T, Jagoda A, Latchaw RE, Todd HW, Viste K, Starke R, et al. Advances in the genetics of cerebrovascular disease and stroke. Neurology. 2001; 56: 997–1008.[Abstract/Free Full Text]
  6. Brass LM, Alberts MJ. The genetics of cerebrovascular disease. Baillieres Clin Neurol. 1995; 4: 221–245.[Medline] [Order article via Infotrieve]
  7. 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: TOAST: Trial of Org 10172 in Acute Stroke Treatment. Stroke. 1993; 24: 35–41.[Abstract/Free Full Text]
  8. Egger M, Smith GD, Schneider M, Minder C. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997; 315: 629–634.[Abstract/Free Full Text]
  9. Bak S, Gaist D, Sindrup SH, Skytthe A, Christensen K. Genetic liability in stroke: a long-term follow-up study of Danish twins. Stroke. 2002; 33: 769–774.[Abstract/Free Full Text]
  10. Brass LM, Isaacsohn JL, Merikangas KR, Robinette CD. A study of twins and stroke. Stroke. 1992; 23: 221–223.[Abstract/Free Full Text]
  11. Brass LM, Carrano D, Hartigan PM, Concato J, Page WF. A follow-up study of the National Academy of Science/Veterans Administration twin registry. Neurology. 1996; 46: A212. Abstract.
  12. de Faire U, Friberg L, Lundman T. Concordance for mortality with special reference to ischaemic heart disease and cerebrovascular disease: a study on the Swedish Twin Registry. Prev Med. 1975; 4: 509–517.[CrossRef][Medline] [Order article via Infotrieve]
  13. Berger K, Schulte H, Stögbauer F, Assmann G. Incidence and risk factors for stroke in an occupational cohort: the PROCAM study. Stroke. 1998; 29: 1562–1566.[Abstract/Free Full Text]
  14. Brass LM, Shaker LA. Family history in patients with transient ischemic attacks. Stroke. 1991; 22: 837–841.[Abstract/Free Full Text]
  15. Harmsen P, Rosengren A, Tsipogianni A, Wilhelmsen L. Risk factors for stroke in middle-aged men in Göteborg, Sweden. Stroke. 1990; 21: 223–229.[Abstract/Free Full Text]
  16. Jousilahti P, Rastenyte D, Tuomilehto J, Sarti C, Vartiainen E. Parental history of cardiovascular disease and risk of stroke: a prospective follow-up of 14 371 middle-aged men and women in Finland. Stroke. 1997; 28: 1361–1366.[Abstract/Free Full Text]
  17. Khaw KT, Barrett-Connor E. Family history of stroke as an independent predictor of ischemic heart disease in men and stroke in women. Am J Epidemiol. 1986; 123: 59–66.[Abstract/Free Full Text]
  18. Kiely DK, Wolf PA, Cupples LA, Beiser AS, Myers RH. Familial aggregation of stroke: the Framingham Study. Stroke. 1993; 24: 1366–1371.[Abstract/Free Full Text]
  19. Lindenstrøm E, Boysen G, Nyboe J. Risk factors for stroke in Copenhagen, Denmark, I: basic demographic and social factors. Neuroepidemiology. 1993; 12: 37–42.[Medline] [Order article via Infotrieve]
  20. Menotti A, Giampaoli S. A single risk factor measurement predicts 35-year mortality from cardiovascular disease. G Ital Cardiol. 1998; 28: 1354–1362.[Medline] [Order article via Infotrieve]
  21. Morrison AC, Fornage M, Liao D, Boerwinkle E. Parental history of stroke predicts subclinical but not clinical stroke: the Atherosclerosis Risk in Communities Study. Stroke. 2000; 31: 2098–2102.[Abstract/Free Full Text]
  22. Sesso HD, Lee I-M, Gaziano JM, Rexrode KM, Glynn RJ, Buring JE. Maternal and paternal history of myocardial infarction and risk of cardiovascular disease in men and women. Circulation. 2001; 104: 393–398.[Abstract/Free Full Text]
  23. Shaper AG, Philips AN, Pocock SJ, Walker M, Macfarlane PW. Risk factors for stroke in middle-aged British men. BMJ. 1991; 302: 1111–1115.[Medline] [Order article via Infotrieve]
  24. Simons LA, McCallum J, Friedlander Y, Simons J. Risk factors for ischemic stroke: Dubbo Study of the elderly. Stroke. 1998; 29: 1341–1346.[Abstract/Free Full Text]
  25. Voko Z. Genetic research at the ASA International Stroke Conference. Stroke Newsletter. 2000;summer 2000: 10. Abstract.
  26. Wannamethee SG, Shaper AG, Ebrahim S. History of parental death from stroke or heart trouble and the risk of stroke in middle-aged men. Stroke. 1996; 27: 1492–1498.[Abstract/Free Full Text]
  27. Welin L, Svärdsudd K, Wilhelmsen L, Larsson B, Tibblin G. Analysis of risk factors for stroke in a cohort of men born in 1913. N Engl J Med. 1987; 317: 521–526.[Abstract]
  28. Wilhelmsen L. Synergistic effects of risk factors. Clin Exp Hypertens. 1990; A12: 845–863.
  29. Abu-Zeid HAH, Choi NW, Maini KK, Hsu PH, Nelson NA. Relative role of factors associated with cerebral infarction and cerebral hemorrhage: a matched pair case-control study. Stroke. 1977; 8: 106–112.[Abstract/Free Full Text]
  30. Alter M. Genetic factors in cerebrovascular accidents. Trans Am Neurol Assoc. 1967; 92: 205–208.[Medline] [Order article via Infotrieve]
  31. Alter M, Kluznik J. Genetics of cerebrovascular accidents. Stroke. 1972; 3: 41–48.[Abstract/Free Full Text]
  32. Alvarez J, Matias-Guiu J, Sumalla J, Molins M, Insa R, Moltó JM, Martin R, Codina A, Martinez-Vazquez JM. Ischemic stroke in young adults, I: analysis of the etiological subgroups. Acta Neurol Scand. 1989; 80: 28–34.[Medline] [Order article via Infotrieve]
  33. Becher H, Grau A, Steindorf K, Buggle F, Hacke W. Previous infection and other risk factors for acute cerebrovascular ischaemia: attributable risks and the characterisation of high risk groups. J Epidemiol Biostat. 2000; 5: 277–283.[Medline] [Order article via Infotrieve]
  34. Bharucha NE, Bharucha EP, Bharucha AE, Bhise AV, Schoenberg BS. Case-control study of completed ischemic stroke in the Parsis of Bombay: a population based study. Neurology. 1988; 38: 490–492.[Abstract/Free Full Text]
  35. Caicoya M, Corrales C, Rodriguez T. Family history and stroke: a community case-control study in Asturias, Spain. J Epidemiol Biostat. 1999; 4: 313–320.[Medline] [Order article via Infotrieve]
  36. Carrieri PB, Orefice G, Maiorino A, Provitera V, Balzano G, Lucariello A. Age-related risk factors for ischemic stroke in Italian men. Neuroepidemiology. 1994; 13: 28–33.[Medline] [Order article via Infotrieve]
  37. Diaz JF, Hachinski VC, Pederson LL, Donald A. Aggregation of multiple risk factors for stroke in siblings of patients with brain infarction and transient ischemic attacks. Stroke. 1986; 17: 1239–1242.[Abstract/Free Full Text]
  38. Feigin VL, Wiebers DO, Nikitin YP, O’Fallon WM, Whisnant JP. Risk factors for ischemic stroke in a Russian community: a population-based case-control study. Stroke. 1998; 29: 34–39.[Abstract/Free Full Text]
  39. Fonte G, Bo M, Poli L, Fiandra U, Fabris F. Stroke ischemico e attacchi ischemici transitori: studio caso-controllo sui fattori di rischio in pazienti anziani ospedalizzati. Recenti Prog Med. 1993; 84: 254–262.[Medline] [Order article via Infotrieve]
  40. Gertler MM, Rusk HA, Whiter HH, Leetma HE, Ehrenkranz M. Ischemic cerebrovascular disease: the assessment of risk factors. Geriatrics. 1968; 23: 135–141.[Medline] [Order article via Infotrieve]
  41. Gifford AJ. An epidemiological study of cerebrovascular disease. Am J Public Health. 1966; 56: 452–461.[Free Full Text]
  42. Graffagnino C, Gasecki AP, Doig GS, Hachinski VC. The importance of family history in cerebrovascular disease. Stroke. 1994; 25: 1599–1604.[Abstract]
  43. Halim A, Ottman R, Hauser WA, Sacco RL. The role of hypertension and cigarette smoking in the familial aggregation of stroke in the Northern Manhattan Stroke Study. Neurology. 1998; 50: A35. Abstract
  44. Hassan A, Sham PC, Markus HS. Planning genetic studies in human stroke. Neurology. 2002; 58: 1483–1488.[Abstract/Free Full Text]
  45. Herman B, Schmitz PIM, Leyten ACM, van Luijk JH, Frenken CWGW, op de Coul AAW, Schulte BPM. Multivariate logistic analysis of risk factors for stroke in Tilburg, the Netherlands. Am J Epidemiol. 1983; 118: 514–525.[Abstract/Free Full Text]
  46. Heyden S, Heyman A, Camplong L. Mortality pattern among parents of patients with atherosclerotic cerebrovascular disease. J Chronic Dis. 1969; 22: 105–110.[CrossRef][Medline] [Order article via Infotrieve]
  47. Hu HH, Chu FL, Chiang BN, Lan CF, Sheng WY, Lo YK, Wong WJ, Luk YO. Prevalence of stroke in Taiwan. Stroke. 1989; 20: 858–863.[Abstract/Free Full Text]
  48. Issaeva I, Mikheev VF. Rol’ nasledstvennogo faktora v razvitii insul’ta [The role of hereditary factors in the development of stroke]. Zh Nevropat Psikhiat Korsakov. 1967; 67: 22–24.
  49. Jerrard-Dunne P, Cloud G, Hassan A, Markus HS. Evaluating the genetic component of ischemic stroke subtypes: a family history study. Stroke. 2003; 34: 1364–1369.[Abstract/Free Full Text]
  50. Kubota M, Yamaura A, Ono J, Itani T, Tachi N, Ueda K, Nagata I, Sugimoto S. Is family history an independent risk factor for stroke? J Neurol Neurosurg Psychiatry. 1997; 62: 66–70.[Abstract]
  51. Li S, Wang C, Fu Y, Cheng XM, Feng E, Wang W, Shu Q, Yang Q, Chen S, Su Q, et al. Risk factors for stroke in rural areas of the People’s Republic of China: results of a case-control study. Neuroepidemiology. 1990; 9: 57–67.[Medline] [Order article via Infotrieve]
  52. Liao D, Myers R, Hunt S, Shahar E, Paton C, Burke G, Province M, Heiss G. Familial history of stroke and stroke risk: the Family Heart Study. Stroke. 1997; 28: 1908–1912.[Abstract/Free Full Text]
  53. Margaglione M, di Minno G, Grandone E, Vecchione G, Celentano E, Cappucci G, Grilli M, Simone P, Panico S, Mancini M. Abnormally high circulation levels of tissue plasminogen activator and plasminogen activator inhibitor-1 in patients with a history of ischemic stroke. Arterioscler Thromb. 1994; 14: 1741–1745.[Abstract/Free Full Text]
  54. Marshall J. Familial incidence of cerebrovascular disease. J Med Gen. 1971; 8: 84–89.[Medline] [Order article via Infotrieve]
  55. Matias-Guiu J, Alvarez J, Insa R, Moltó JM, Martin R, Codina A, Martinez-Vazquez JM. Ischemic stroke in young adults, II: analysis of risk factors in the etiological subgroups. Acta Neurol Scand. 1990; 81: 314–317.[Medline] [Order article via Infotrieve]
  56. Muñiz J, Juane R, Castro-Beiras A, Fernández Fuertes I, Lorenzo Gallego S, Sánchez Herrero J. Estudio de casos y controles de factores de riesgo de accidente cerebrovascular agudo. Med Clin (Barc). 1993; 101: 401–405.[Medline] [Order article via Infotrieve]
  57. Peng D, Zhao S. Serum lipids, lipoprotein and stroke [in Chinese]. Hunan Yi Ke Da Xue Xue Bao. 1999; 24: 167–170.[Medline] [Order article via Infotrieve]
  58. Peng DQ, Zhao SP, Wang JL. Lipoprotein (a) and apolipoprotein E epsilon 4 as independent risk factors for ischemic stroke. J Cardiovasc Risk. 1999; 6: 1–6.[Medline] [Order article via Infotrieve]
  59. Polychronopoulos P, Gioldasis G, Ellul J, Metallinos IC, Lekka NP, Paschalis C, Papapetropoulos T. Family history of stroke in stroke types and subtypes. J Neurol Sci. 2002; 195: 117–122.[CrossRef][Medline] [Order article via Infotrieve]
  60. Shintani S, Kikuchi S, Hamaguchi H, Shiigai T. High serum lipoprotein(a) levels are an independent risk factor for cerebral infarction. Stroke. 1993; 24: 965–969.[Abstract/Free Full Text]
  61. Spriggs DA, French JM, Murdy JM, Bates D, James OFW. Historical risk factors for stroke: a