Association of Parental History of Stroke With Clinical Parameters in Patients With Ischemic Stroke or Transient Ischemic Attack
Background and Purpose— Data regarding hereditary influences on stroke remain controversial. We investigated possible associations of a family history of stroke with clinical parameters in a large cohort of well-documented patients with ischemic cerebrovascular events, with special reference to sex-specific differences.
Methods— We analyzed the association between a maternal and/or paternal history of stroke and clinical parameters in 1564 patients with known maternal and paternal history of stroke and suffering from ischemic stroke or transient ischemic attack within the Vienna Stroke Registry.
Results— Patients with a maternal history of stroke were significantly more often female (54%) than those without (44%; P=0.003). Hypertension was more prevalent in female patients with than in those without a maternal history of stroke (87% versus 74%; P=0.001). These associations remained significant after multivariate adjustment (adjusted odds ratio, 1.9; 95% CI, 1.1 to 3.5; P=0.024). Of those female patients with an echocardiogram (n=225), those with a maternal history of stroke more often had left ventricular hypertrophy (48%) than those without (20%) (adjusted odds ratio, 3.6; 95% CI, 1.5 to 8.2; P=0.003). In contrast, hypertension was equally prevalent in male patients with or without a maternal history of stroke (75% versus 74%; P=0.754). We found no significant associations of clinical parameters with a paternal history of stroke.
Conclusions— Our results indicate a sex-specific relationship between a maternal history of stroke and the prevalence of hypertension and left ventricular hypertrophy in female patients with ischemic cerebrovascular events.
It is widely acknowledged that, besides acquired risk factors, stroke has an important hereditary component. Although several studies indicated a strong association between family history of stroke and stroke risk,1–7 other studies failed to do so.8–12 These discrepancies are probably due to different and partly inconsistent study design. Stroke is a heterogeneous disease, and most studies did not account for the various stroke subgroups, assuming that the same hereditary factors would influence different types of stroke equally. Only a few studies have attempted a detailed clinical evaluation of probands. Additionally, sex-specific differences with respect to paternal or maternal history of stroke and stroke risk in male or female offspring were rarely considered.
The present study therefore sought to investigate possible associations of a family history of stroke and clinical parameters in a large cohort of well-documented patients with ischemic stroke or transient ischemic attack (TIA). We were particularly interested in possible sex-specific differences between patients with or without a maternal or paternal history of stroke.
Subjects and Methods
This study was nested in a prospective stroke registry of patients admitted to 8 neurological departments in Vienna, Austria (Vienna Stroke Registry), serving a community of 1.9 million people.13 All patients with TIA/stroke who were admitted to 1 of the participating centers within 72 hours of symptom onset between October 1998 and December 2001 were prospectively documented, on the basis of informed consent, with respect to clinical and neurological parameters (National Institutes of Health Stroke Scale, Scandinavian Stroke Scale, modified Rankin Scale, Barthel Index), results of technical investigations, presumed etiology, and follow-up investigations at 3, 12, and 24 months. Medical history with special reference to vascular diseases and family history of stroke, myocardial infarction, peripheral artery disease, diabetes mellitus, and hypertension was assessed by a structured personal interview. Data on parental history did not include the stroke type. However, it can be assumed that most of the affected parents suffered from ischemic stroke.14,15 As part of laboratory parameters, serum levels of triglycerides, total cholesterol, LDL cholesterol (LDL-C), and HDL cholesterol (HDL-C) were investigated in the majority of the patients. The study was approved by the local ethics committees.
For the present cohort study, we investigated data of all patients with acute ischemic stroke or TIA whose available clinical data had been entered into the database at the time of analysis. At that time, 3615 patients had been admitted because of suspected ischemic or hemorrhagic stroke, and data of 2378 patients were entered into the database. Patients for whom detailed clinical information was still missing (n=1237) were excluded from the analysis. These excluded patients were comparable in terms of age (67.4 [SD 14.6] versus 66.6 [SD 18.5] years; P=0.5) and sex (47% female versus 45% female; P=0.4). It was estimated that approximately 15% of these patients would not fulfill our inclusion criteria because of a nonischemic cerebrovascular event (intracerebral hemorrhage, subarachnoid hemorrhage, sinus vein thrombosis) or post hoc verification of a noncerebrovascular diagnosis (eg, epileptic seizure, hypertensive crisis). Entering of data into the database was performed in batches, and there was no selection mechanism other than order of admission. From the patients with available clinical information (n=2378), those with hemorrhagic stroke (n=203), those for whom no information about paternal and/or maternal history of stroke could be obtained (n=597), and those with a rare cause of stroke (n=14) were also excluded. Thus, 1564 patients were available for the present analysis.
Statistical analyses were performed with the use of SPSS version 10.0. Univariate comparisons of continuous variables were performed with the unpaired t test or with the Mann-Whitney U test, as appropriate. Binary and categorical data were analyzed with χ2 statistics. Probability values of <0.05 were considered significant.
To assess the relation between family history and >1 clinical variable simultaneously, multivariate logistic regression was applied. All variables that were at least weakly associated with a maternal or paternal history of stroke were included (P<0.2 in univariate analyses). The Cox and Snell R was used to assess the variability explained by each model. The Hosmer-Lemeshow test was used to assess the model fit.
Interaction was assessed with the use of multiplicative interaction terms and log likelihood ratio χ2 tests.
Statistical analyses comprised the following steps. First, we compared patients in whom information about a history of maternal and paternal stroke could be obtained (“family history available”) and those patients for whom this information could not be obtained for 1 or both parents (“family history not available”) with particular reference to sex-specific recall bias. For all further analyses, only patients with an available maternal and paternal history of stroke were included. Clinical parameters were then compared in (1) patients whose father and/or mother had suffered a stroke (“positive family history”) versus patients in whom neither father nor mother had suffered a stroke (“negative family history of stroke”); (2) patients whose mother had suffered a stroke (“positive maternal history of stroke”) versus patients whose mother had not suffered a stroke (“negative maternal history of stroke”); and (3) patients whose father had suffered a stroke (“positive paternal history of stroke”) versus patients whose father had not suffered a stroke (“negative paternal history of stroke”).
The following clinical parameters were analyzed: age (quintiles); sex; stroke severity: (1) TIA (duration of symptoms <24 hours); (2) minor stroke (duration of symptoms >24 hours and Rankin Scale score at 1 week ≤2); and (3) major stroke (Rankin Scale score at 1 week >2, including patients with fatal stroke); presumed etiology classified according to prespecified criteria into the following groups: (1) large-vessel disease (ipsilateral carotid stenosis ≥70%, presumable local thrombosis of a large intracranial vessel, arterio-arterial embolism from aortic plaques/thrombi); (2) small-vessel disease (clinical lacunar syndrome and no lesion or subcortical lesion <1.5 cm on CT or MRI); (3) cardioembolic (high-risk source of cardiac embolism); and (4) no determined etiology; hypertension (defined as a history of hypertension reported by patient or relative or documented in previous medical records or antihypertensive treatment at discharge or blood pressure on admission >180 mm Hg systolic or >110 mm Hg diastolic; yes versus no); diabetes (defined as a history of diabetes as reported by patient or relative or a fasting blood glucose level >8.325 mmol/L [>150 mg/dL]; yes versus no); current cigarette smoking (yes versus no); previous stroke (yes versus no); clinically manifest coronary artery disease (no/angina pectoris only/myocardial infarction); chronic or paroxysmal atrial fibrillation (yes versus no); and total cholesterol, LDL-C, HDL-C, and triglyceride levels in quintiles. One or more items of the medical history could not be obtained in 3% to 5% of the patients. In this case the respective item was classified as “no.” For the subgroup of patients in whom echocardiography was available (n=555), the presence of left ventricular hypertrophy (LVH) was also investigated.
In 597 of 2161 patients, no information about maternal and/or paternal history of stroke could be obtained. The reasons for lack of availability of family history were (1) that the patient did not know whether his/her father/mother had suffered a stroke or (2) that the patient could not be interviewed. The latter was mainly due to severe medical condition of the patient or disturbances of speech or language.
Multivariate logistic regression revealed that higher age (odds ratio [OR] for the fifth versus the fourth quintile for lack of availability of family history: 2.2; 95% CI, 1.6 to 3.2; P<0.001) and stroke severity (major stroke versus minor stroke: OR=1.4; 95% CI, 1.1 to 1.8; P=0.010) were the most important factors for lack of availability of a family history of stroke.
Because the information about family history was obtained by a personal structured interview, differences between male and female patients regarding their knowledge of family history were examined to anticipate confounding by sex-specific family history recall bias. After adjustment for stroke severity as determined by the modified Rankin Scale on days 5 to 7, no significant difference between male and female patients emerged. For all further analyses, patients with an unknown family history of stroke were excluded.
Clinical variables associated with paternal or maternal history (referred to as family history) of stroke are shown in Table 1. LDL-C (which was available for only 703 patients, however) was significantly higher in patients with a positive paternal or maternal history of stroke than in subjects with negative family history of stroke. None of the other variables differed significantly between the 2 groups.
Stratification according to sex of the affected parent revealed differences of the association between clinical parameters and a maternal or paternal history of stroke, respectively.
Patients with a positive maternal history of stroke were significantly more often female than those with negative maternal history (Table 2). Moreover, patients with a positive maternal history of stroke had a higher prevalence of hypertension and LVH, were less often cigarette smokers, and had significantly higher total cholesterol levels and LDL-C levels than patients with a negative maternal history of stroke. Stroke severity and etiology did not differ significantly according to maternal history of stroke.
In contrast, a paternal history of stroke was not associated with any of the investigated parameters except for a trend of a higher prevalence of atrial fibrillation in patients with a negative paternal history of stroke (19% versus 14%; P=0.055).
The interaction between sex, hypertension, and maternal history of stroke was highly significant (χ2=10.1; df=1; P<0.01). Further analyses of the relation between maternal history of stroke and clinical parameters were therefore performed for men and women separately. We found no significant interaction between age, hypertension, and a maternal history of stroke.
Women with a positive maternal history of stroke showed a higher prevalence of hypertension (87% versus 74%; P=0.001) and higher total cholesterol levels (median, 234 mg/dL; interquartile range, 203 to 264 mg/dL versus median, 221 mg/dL; interquartile range, 186 to 248 mg/dL; P=0.011) than women with a negative maternal history of stroke. Of those women with an echocardiogram (n=225), those with a positive maternal history of stroke had a markedly higher prevalence of LVH (48%) than those with a negative maternal history of stroke (20%; P<0.001).
In male patients there were no significant differences between patients with negative or positive maternal history of stroke except for a trend of a lower prevalence of diabetes among those men with a positive maternal history of stroke (28%) compared with those with a negative maternal history of stroke (20%; P=0.069).
To investigate the relation between maternal history of stroke and several clinical variables simultaneously, we applied multivariate logistic regression. Since data for LDL-C and LVH were available for only 45% and 35% of the population, respectively, these parameters were not included in the multivariate model. However, total cholesterol, which was available for the majority (87%) of the patients, was included in the analysis.
Multivariate analysis in female patients revealed hypertension as the only clinical variable significantly associated with a positive maternal history of stroke (Table 3). The OR for hypertension in women with a positive maternal history of stroke was 1.9 (95% CI, 1.1 to 3.5; P=0.024) compared with those with a negative maternal history of stroke. Tests for interaction between hypertension and age and between hypertension and cholesterol were not significant. However, there was a highly significant interaction between age and cholesterol in that younger women with a positive maternal history of stroke had a higher risk of cholesterol levels in the highest quintile than those with a negative maternal history of stroke or older women with a positive maternal history of stroke (test for interaction: P<0.001).
When the analysis was restricted to those patients with a documented echocardiography, the presence of LVH was even more strongly associated with a positive maternal history of stroke than clinically known hypertension (Table 4). The OR for having LVH was 3.6 (95% CI, 1.5 to 8.2; P=0.003) for those female patients with a positive maternal history of stroke compared with those with a negative maternal history of stroke. Since women with an echocardiogram were significantly younger than those without, the relation between positive maternal history of stroke and cholesterol levels was also stronger in this subgroup.
In contrast to female patients, no significant associations between maternal history of stroke and clinical parameters were present in male patients, and consequently no multivariate analyses were performed.
Univariate analyses had revealed no significant associations between paternal history of stroke and clinical parameters. Therefore, we did not perform multivariate analysis with respect to paternal history of stroke.
The association of a family history of stroke with stroke risk has been investigated in several studies but with diverging results.1–12 In contrast, the relation between family history of stroke and clinical parameters in stroke patients and particularly sex-specific aspects of this relation have not been addressed until now. The results of our study demonstrate marked sex-specific differences of the relation between a maternal or paternal history of stroke and clinical parameters in male and female stroke patients.
First, patients with a positive maternal history of stroke were more often women than those with a negative maternal history of stroke, whereas a paternal history of stroke was not associated with specific sex distribution in the patients. It seems, therefore, that a maternal history of stroke affects male and female offspring differently.
Moreover, hypertension was significantly more prevalent in female but not in male patients with a positive maternal history of stroke compared with patients with a negative maternal history of stroke. Accordingly, the prevalence of LVH was markedly higher in female patients with a positive maternal history of stroke. These associations remained significant after multivariate adjustment. Interestingly, the association between a positive maternal history of stroke and LVH was even stronger than that with clinically overt hypertension in female patients. This finding indicates that a significant proportion of female patients with a positive maternal history of stroke had long-standing hemodynamic changes before they developed overt hypertension.
Conversely, paternal history showed no significant association with the prevalence of hypertension or other clinical parameters in either male or female patients.
In conclusion, the results of this study indicate that parental, particularly positive maternal, history of stroke is associated with the prevalence of hypertension, LVH, and elevated cholesterol levels in female but not in male offspring.
Several possible mechanisms may contribute to the apparent sex-specific relation between parental history of stroke and hypertension. These comprise genetic as well as environmental and psychosocial mechanisms.16–22
There are several limitations of our study. The fact that we did not analyze a control population of individuals without stroke limits the conclusion to patients who have already suffered a stroke. We also cannot infer whether a positive maternal or paternal history of stroke is associated with a different absolute risk of stroke in men and women. However, our investigation was not designed as a case-control study but sought to examine associations of a family history of stroke with clinical parameters within a stroke population. Such associations, as they were found in our population, can help to uncover specific relationships between familial burden of stroke and clinical parameters, which can then be proven in case-control studies and prospective studies. Additionally, the whole population of the stroke registry could not be analyzed because data of several patients were not available at the time of our investigation. Demographic parameters of these missing patients are similar to those included in our study, and we are not aware of a selection mechanism of data entry that might have confounded our results.
As a consequence of our results, sex-specific relationships between paternal or maternal family history and male or female offspring should be considered in future studies.
The Vienna Stroke Registry is supported by research grants of the Medizinisch-Wissenschaftlicher Fonds des Bürgermeisters der Bundeshauptstadt Wien (project Nos. 1540 and 1829), the Jubiläumsfonds der Oesterreichischen Nationalbank (project Nos. 6866 and 8281), and the Austrian Research Society (P13902-MED). The Vienna Stroke Registry is sponsored by an unrestricted educational grant of Sanofi-Synthelabo and Bristol-Myers-Squibbs. The Vienna Stroke Registry is supported by the Wiener Krankenanstaltenverbund (L. Kaspar, MD).
- Received March 14, 2003.
- Revision received May 14, 2003.
- Accepted May 19, 2003.
Marshall J. Familial incidence of cerebral hemorrhage. Stroke. 1973; 4: 38–41.
Khaw K-T, 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.
Kiely D, Wolf P, Cupples A, Beiser A, Myers R. Familial aggregation of stroke: the Framingham Study. Stroke. 1993; 24: 1366–1371.
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.
Brass L, Isaacson J, Merikangas K, Robinette D. A study of twins and stroke. Stroke. 1992; 23: 221–223.
Jousilahti P, Rastenyte D, Tuomilehto J, Sarti C, Vartiainen E. Parental history of cardiovascular disease and risk of stroke. Stroke. 1997; 28: 1361–1366.
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.
Boysen G, Nybou J, Appleyard M, Sorensen PS, Boas J, Somnier F, Jensen G, Schnohr P. Stroke incidence and risk factors for stroke in Copenhagen, Denmark. Stroke. 1988; 19: 1345–1353.
Brass L, Shaker M. Family history in a TIA cohort. Stroke. 1991; 22: 837–841.
Graffagnino C, Gasecki A, Doig G, Hachinski V. The importance of family history in cerebrovascular disease. Stroke. 1994; 25: 1599–1604.
Morrison A, Fornage M, Liao D, Boerwinkle E. Parental history of stroke predicts subclinical but not clinical stroke. Stroke. 2000; 31: 2098–2102.
Kolominsky-Rabas P, Weber M, Gefeller O, Neundoerfer B, Heuschmann PU. Epidemiology of ischemic stroke subtypes according to TOAST criteria. Stroke. 2001; 32: 2735–2740.
Koike G, Jacob HJ. Hypertension. In: Jameson JL, ed. Principles of Molecular Medicine. Totowa, NJ: Humana Press; 1998.