Diabetes and hypertension markedly increased the risk of ischemic stroke associated with high serum resistin concentration in a general Japanese population: the Hisayama Study

In 2002, a screening survey for the present study was performed in the town of Hisayama, a suburb of the Fukuoka metropolitan area in Japan. The age and occupational distributions and nutrient intake of the population were almost identical to those of Japan as a whole based on data from the national census and nutrition survey. A detailed description of this survey was published previously [15]. Briefly, of all residents aged 40 years or over, 3,328 underwent the examination (participation rate, 77.6%). After excluding 30 subjects who did not consent to participate in the study, 86 who had already eaten breakfast, and 11 who did not have enough stored sera with which to measure resistin concentrations, a total of 3,201 (1,382 men and 1,819 women) were enrolled in the study group and underwent a comprehensive assessment. This study was approved by the Ethics Committee of Kyushu University, and written informed consent was obtained from all participants.

Cases of CVD were defined as subjects who had histories of stroke or coronary heart disease. CVD was identified using the following criteria. The diagnosis and classification of stroke were determined on the basis of clinical information, including brain computed tomography and magnetic resonance imaging, cerebral angiography, echocardiography or carotid duplex imaging. Stroke was defined as sudden onset of nonconvulsive and focal neurological deficit persisting for ≥24 hours and was classified as either ischemic or hemorrhagic [16, 17]. Hemorrhagic stroke included brain hemorrhage and subarachnoid hemorrhage. Ischemic stroke was further divided into 4 clinical categories -- lacunar infarction, atherothrombotic infarction, cardioembolic infarction, and undetermined subtype of ischemic stroke -- based on the Classification of Cerebrovascular Disease III proposed by the National Institute of Neurological Disorders and Stroke [18], as well as on the basis of the diagnostic criteria of the Trial of Org10172 in Acute Stroke Treatment Study [19] and the Cerebral Embolism Task Force [20]. Details of the diagnostic criteria for the ischemic stroke subtypes have been described previously [21]. In brief, lacunar infarction was defined as the presence of a relevant brainstem, basal ganglia, or subcortical hemispheric lesion with a diameter of < 1.5 cm demonstrated by brain imaging, and no evidence of cerebral cortical or cerebellar impairment. Atherothrombotic infarction was diagnosed when the subjects had significant stenosis (> 50%) or occlusion of a major cerebral artery with an infarct ≥1.5 cm in a brain imaging study. The diagnosis of cardioembolic infarction was made on the basis of primary and secondary clinical features suggestive of cardioembolic infarction as reported by the Cerebral Embolism Task Force [20]. The undetermined subtype of ischemic stroke included strokes that could not be classified because of insufficient clinical or morphologic information. We considered morphologic findings to be significant and used clinical features as reference information. Cases of cerebrovascular diseases that could be attributed to a distinct pathology, such as collagen disease, hematological disorder, trauma, chronic subdural hematoma, or moyamoya disease, were excluded from the evaluation.

The diagnostic criteria for coronary heart disease included acute myocardial infarction, silent myocardial infarction or coronary artery disease followed by coronary artery bypass surgery or angioplasty [16, 17]. Acute myocardial infarction was diagnosed when a subject met at least two of the following criteria: (1) typical symptoms, including prolonged severe anterior chest pain; (2) abnormal elevation in cardiac enzymes, i.e., greater than a two-fold increase in the upper limit of the normal range; (3) evolving diagnostic electrocardiogram (ECG) changes; and, (4) morphological changes, including local asynergy of cardiac wall motion on echocardiography or persistent perfusion defects observed using cardiac scintigraphy. Silent myocardial infarction was defined as the above-mentioned morphological changes without any history of clinical symptoms or abnormalities in cardiac enzymes. Among the participants, 175 subjects had a history of CVD as follows: 79 had ischemic stroke; 41 had hemorrhagic stroke; 42 had coronary heart disease; 11 had both ischemic stroke and coronary heart disease; and, 2 had hemorrhagic stroke and coronary heart disease. In the analysis stratified by type of CVD, coexisting CVDs were each stratified into their respective types. Furthermore, on the basis of the above criteria, 90 ischemic stroke cases were divided into 51 cases of lacunar infarction, 29 of atherothrombotic infarction, 7 of cardioembolic infarction, and 3 of undetermined infarction; 43 hemorrhagic stroke cases were divided into 31 of brain hemorrhage and 12 of subarachnoid hemorrhage.

At the screening examination, blood samples were obtained between 8:00 and 10:30 AM after at least a 12-hour overnight fast. An aliquot of serum from each subject was stored at -80°C. Serum resistin was measured using a human ELISA kit (R&D Systems, Inc., Minneapolis, MN) following the manufacturer's protocol. According to data supplied by the manufacturer, the limit of detection was 0.16 ng/ml, and the intra-assay coefficient of variations (CVs) were < 5% for low concentrations, and < 4% for high concentrations. The inter-assay CV was < 9% for both low and high concentrations. Recovery was greater than 90%. We confirmed that linearity was maintained at concentrations less than 0.16 ng/ml, and both intra- and inter-assay CVs were comparable to those specified by the manufacturer (2.6-10.5%). The antibodies used in the enzyme-linked immunosorbent assay (ELISA) do not cross-react with either mouse resistin or other human cytokines. The results obtained using the ELISA manufactured by R&D were highly correlated with those obtained previously using a kit supplied by LINCO (Linco Research, Inc., St. Charles, MO) (r = 0.987, y = 1.040x+0.469, y, this kit, x, LINCO's kit) [22‐24].

Blood for the glucose assay was collected into tubes containing NaF, and plasma glucose concentrations were determined immediately after venipuncture using the glucose-oxidase method. Serum total cholesterol and high density lipoprotein (HDL) cholesterol concentrations were determined enzymatically. hs-CRP concentrations were measured using a modification of the Behring Latex-Enhanced CRP assay on a Behring Nephelometer BN-100 (Behring Diagnostics, Westwood, MA). Blood pressure was measured three times using a standard mercury sphygmomanometer in the seated position after the subject had rested for at least 5 minutes. ECG abnormalities were defined as left ventricular hypertrophy (Minnesota Code 3-1) and/or ST depression (Minnesota Code 4-1, 2, 3).

Body height and weight were measured in light clothing without shoes, and body mass index (BMI) was calculated. Each participant completed a self-administered questionnaire that assessed medical history, smoking habits, alcohol intake, and exercise. The questionnaire was checked by trained interviewers at the screening. Alcohol intake and smoking habits were classified as either presence or absence of current habitual use. Those subjects who engaged in sports or other forms of exercise ≥3 times a week during their leisure time categorized regular exercisers.

The SAS software package version 8.2 (SAS Institute, Inc., Cary, NC) was used to perform all statistical analyses. The subjects were divided into quintiles of resistin concentration as follows: < 6.8, 6.8 to 8.7, 8.8 to 11.5, 11.6 to 16.2, and ≥16.3 ng/ml. Values of possible risk factors were adjusted for age and sex using the covariance method and were compared among the quintiles using the linear regression model. Age- and sex-adjusted resistin concentrations were compared among CVD subtypes using the same methods. The frequencies of risk factors were adjusted for age and sex by the direct method using all subjects as a standard population and were tested for trends using the Cochran-Mantel-Haenszel test. Age- and sex-adjusted or multivariate-adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for CVD were calculated using logistic regression analysis. A p-value < 0.05 was considered statistically significant in all analyses.

Age- and sex-adjusted means or frequencies of potential risk factors by quintiles of serum resistin concentration are shown in Table 1. Age and hs-CRP, and the frequency of male sex increased with quintiles of resistin, while mean HDL-cholesterol and the frequencies of alcohol consumption and regular exercise were negatively associated with the quintiles of resistin. The other variables were not significantly associated with resistin quintiles.

Age- and sex-adjusted serum resistin concentrations were greater in subjects with total CVD than in those without CVD (p = 0.002) (Table 2). When CVD was divided into types, subjects with ischemic stroke had greater resistin concentrations than those without CVD (p < 0.001). Regarding subtypes of ischemic stroke, subjects with lacunar and atherothrombotic infarction had greater resistin than subjects without CVD (p = 0.02 for lacunar infarction; p < 0.001 for atherothrombotic infarction); however, no association was observed between resistin and cardioembolic infarction or undetermined subtype of ischemic stroke. Resistin concentrations in subjects with hemorrhagic stroke, including brain and subarachnoid hemorrhage, and coronary heart disease were not different from subjects without CVD.

To further evaluate the association between serum resistin concentrations and the risk of CVD, age- and sex-adjusted or multivariate-adjusted ORs were estimated by quintiles of resistin concentration (Table 3). In Model 1, the age- and sex-adjusted ORs for total CVD and ischemic stroke increased with quintile of serum resistin (p for trends, 0.02 for CVD, < 0.001 for ischemic stroke). Compared to the first quintile, the age- and sex-adjusted OR for ischemic stroke was greater in the third, fourth, and fifth quintiles (the third quintile: OR 3.54, 95%CI 1.17-10.67, p = 0.02; the fourth quintile: OR 4.48, 95%CI 1.53-13.09, p = 0.006; the fifth quintile: OR 4.70, 95%CI 1.62-13.61, p = 0.004), while there was no association between quintile of resistin and CVD. The association between quintile of resistin and ischemic stroke remained significant even after adjustment for age, sex, BMI, diabetes, total cholesterol, HDL-cholesterol, systolic blood pressure, ECG abnormalities, current drinking, current smoking, and regular exercise (Model 2). Furthermore, as shown in Model 3, after adjustment for hs-CRP and the confounding factors, this association remained essentially unchanged (Fig. 1). No significant associations were found between quintile of resistin and either hemorrhagic stroke or coronary heart disease.

To determine whether high serum resistin concentration was associated with increased risk of ischemic stroke in subjects who had either diabetes or hypertension, we examined the combined and unique effects of high resistin and the presence of either diabetes or hypertension on the risk of ischemic stroke. High resistin was defined as the third, fourth, and fifth quintile of resistin concentration. Compared to non-diabetic subjects with low resistin, the age- and sex-adjusted OR of ischemic stroke was greater in non-diabetic subjects with high resistin (OR: 2.15; 95% CI: 1.13-4.10; p = 0.02), but not in diabetic subjects with low resistin (OR: 1.27; 95% CI: 0.35-4.62; p = 0.72). The risk of ischemic stroke markedly increased in diabetic subjects with high resistin (OR: 5.88; 95% CI: 2.82-12.26; p < 0.001). Furthermore, diabetic subjects with high resistin had greater risk of ischemic stroke than diabetic subjects with low resistin (p < 0.05). Similarly, a marked increase in the age- and sex-adjusted risk of ischemic stroke was observed in hypertensive subjects with high resistin compared with normotensive subjects with low resistin (OR: 4.66; 95% CI: 1.81-11.97; p = 0.001), but not in normotensive subjects with high resistin (OR: 1.87; 95%CI: 0.69-5.10; p = 0.22), or hypertensive subjects with low resistin (OR: 1.52; 95%CI: 0.51-4.54; p = 0.45). In hypertensive subjects, high resistin was associated with risk of ischemic stroke (p < 0.05). These associations remained robust even after adjustment for the above-mentioned confounding factors (Fig. 2).