The study subjects comprised 3,187 Japanese individuals (894 subjects with ischemic stroke and 2,293 controls) who either visited the outpatient clinics or were admitted to the participating hospitals (Gifu Prefectural Tajimi Hospital, Tajimi; Gifu Prefectural General Medical Center, Gifu; Inabe General Hospital, Inabe; Japanese Red Cross Nagoya First Hospital, Nagoya; Hirosaki University Hospital and Hirosaki Stroke Center, Hirosaki, Japan) between 2002 and 2012. The diagnosis of ischemic stroke was based on the occurrence of a new and abrupt focal neurological deficit, with neurological symptoms and signs persisting for >24 h; it was confirmed by positive findings in computed tomography or magnetic resonance imaging (or both) of the head. The type of stroke was determined according to the Classification of Cerebrovascular Diseases III (19). Individuals with cardiogenic embolic infarction, lacunar infarction alone, transient ischemic attack, moyamoya disease or cerebral venous sinus thrombosis were excluded from the study. Control individuals had no history of ischemic or hemorrhagic stroke, or other cerebral diseases; of coronary artery disease, aortic aneurysm, or peripheral arterial occlusive disease; or of other thrombotic, embolic or hemorrhagic disorders.

Hypertension was defined as individuals who either had systolic blood pressure ≥140 mmHg or diastolic blood pressure ≥90 mmHg, or had been administered antihypertensive medication. Dyslipidemia was defined as individuals who had either a serum concentration of triglycerides of ≥1.65 mmol/l, a serum high-density lipoprotein (HDL)-cholesterol of <1.04 mmol/l, a serum low-density lipoprotein (LDL)-cholesterol of ≥3.64 mmol/l, or had been administered antidyslipidemic medication. Diabetes mellitus was defined as individuals who either had a fasting plasma glucose level of ≥6.93 mmol/l or blood glycosylated hemoglobin (hemoglobin A1c) content of ≥6.5%, or had been administered antidiabetes medication. Chronic kidney disease (CKD) was defined as individuals who had an estimated glomerular filtration rate (eGFR) of <60 ml min⁻¹ 1.73 m⁻². The eGFR was calculated with the use of the simplified prediction equation derived from the modified version of that described in the Modification of Diet in Renal Disease study, as proposed by the Japanese Society of Nephrology: eGFR (ml/min/1.73 m²) = 194 × age (years)^−0.287 × serum creatinine (mg/dl)^−1.094 (x 0.739 if female) (20).

The study protocol complied with the Declaration of Helsinki and was approved by the Committees on the Ethics of Human Research of Mie University Graduate School of Medicine, Hirosaki University Graduate School of Medicine (Hirosaki, Japan) and the participating hospitals. Written informed consent was obtained from each participant.

The SNPs that were shown to be significantly associated with coronary artery disease or myocardial infarction were searched in Caucasian populations by the meta-analyses of GWASs (14,15). These SNPs were examined with the SNP database (dbSNP; National Center for Biotechnology Information, Bethesda, MD, USA; http://www.ncbi.nlm.nih.gov/SNP/) to find SNPs with a minor allele frequency of ≥0.015 in a Japanese population. A total of 29 SNPs (21) were selected and the possible association with ischemic stroke was examined.

Venous blood (7 ml) was collected into tubes containing 50 mmol/l ethylenediaminetetraacetic acid (disodium salt), the peripheral blood leukocytes were isolated and genomic DNA was extracted from these cells with a DNA extraction kit (Genomix; Talent, Trieste, Italy). Genotypes of SNPs were determined at G&G Science Co., Ltd. (Fukushima, Japan) by a method that combines the polymerase chain reaction and sequence-specific oligonucleotide probes with suspension array technology (Luminex, Austin, TX, USA). Primers, probes and other conditions for genotyping of SNPs examined in the study are as described previously (21). The overall call rate of 29 SNP genotyping was 99%. Detailed genotyping methodology was also as described previously (22–24).

Quantitative data were compared between two groups by the Mann-Whitney U test, given that the data were not normally distributed (P<0.01 by the Kolmogorov-Smirnov and Lilliefors test). Categorical data were compared by the χ² test. Allele frequencies were estimated by the gene counting method. Multivariable logistic regression analysis was performed with ischemic stroke as a dependent variable and independent variables including age, gender (0, female; 1, male), body mass index (BMI), smoking status (0 non-smoker; 1 current or former smoker), history of hypertension, diabetes mellitus and dyslipidemia (0 no history; 1 positive history), and each genotype; and the P-value, odds ratio and 95% confidence interval were calculated. Genotypes of each polymorphism were assessed according to dominant [0 AA; 1 AB+BB (A, major allele; B, minor allele)], recessive (0 AA+AB; 1 BB) and additive genetic models. Additive models comprised additive 1 (AB vs. AA) and 2 (BB vs. AA) models, which were analyzed simultaneously with a single statistical model. P<0.05 was considered to indicate a statistically significant difference. The statistical tests were performed with JMP version 5.1 and JMP Genomics version 6.0 software (SAS Institute Inc., Cary, NC, USA).

The characteristics of the 3,187 study subjects are shown in Table I. Age, the frequency of men, the prevalence of smoking, hypertension, dyslipidemia, diabetes mellitus and CKD, as well as systolic and diastolic blood pressure, serum concentrations of triglycerides, LDL-cholesterol and creatinine, and fasting plasma glucose level were greater, whereas serum concentrations of HDL-cholesterol and eGFR were lower in subjects with ischemic stroke compared to the controls.

The genotype distributions and allele frequencies were compared by the χ² test and revealed that 3 SNPs were significantly (P<0.05) associated with the prevalence of ischemic stroke (Table II). The genotype distributions and allele frequencies of rs2075650 of the translocase of outer mitochondrial membrane 40 homolog (TOMM40) gene and of rs273909 of the solute carrier family 22, member 4 (SLC22A4) gene were significantly associated with the prevalence of ischemic stroke. Allele frequencies of rs9319428 of the fms-related tyrosine kinase 1 gene (FLT1) were also significantly associated with ischemic stroke. These SNPs were further examined by multivariable logistic regression analysis with adjustment for covariates.

The multivariable logistic regression analysis was performed with adjustment for age, gender, BMI, smoking status and the prevalence of hypertension, diabetes mellitus and dyslipidemia and revealed that rs2075650 of TOMM40 (recessive and additive 2 models) and rs273909 of SLC22A4 (dominant and additive 1 models), but not rs9319428 of FLT1, were significantly (P<0.05) associated with ischemic stroke, with the minor G and C alleles, respectively, being protective against this condition (Table III).

Subsequently, the association of rs2075650 of TOMM40 or rs273909 of SLC22A4 to the prevalence of hypertension, dyslipidemia, diabetes mellitus or CKD was examined in all the subjects (Table IV). The rs2075650 of TOMM40 or rs273909 of SLC22A4 was significantly (P<0.05) associated with the prevalence of dyslipidemia (dominant model) or CKD (recessive model), respectively.

The associations of rs2075650 of TOMM40 or rs273909 of SLC22A4 were examined with the clinical parameters including systolic, diastolic and mean blood pressure, pulse pressure, serum concentrations of triglycerides, LDL-cholesterol, HDL-cholesterol, creatinine, fasting plasma glucose level and eGFR (Table V). To avoid the effects of medical treatment on clinical parameters, the corresponding controls were examined for hypertension, dyslipidemia, diabetes mellitus or CKD. The rs2075650 of TOMM40 was not associated with any clinical parameter, whereas rs273909 of SLC22A4 was associated with fasting plasma glucose level (recessive model).