This study was a case-control study. Participants were recruited between May 1, 2018 and April 30, 2019 from all 8 counties of the Xiangxi Tujia and Miao Autonomous Prefecture (Jishou, Baojing, Fenghuang, Guzhang, Huayuan, Longshan, Luxi, and Yongshun). All participants were residents who had lived in Xiangxi for more than 5 years.
Case Selection
The case group included all IS patients who were admitted to the First Affiliated Hospital of Jishou University and all 8 County People’s Hospitals of the Xiangxi Tujia and Miao Autonomous Prefecture within 5 days of symptom onset and 72 h of hospital admission and with a discharge diagnosis of first-ever IS. IS was diagnosed according to the World Health Organization (WHO) clinical criteria[12]. Computed tomography(CT) or magnetic resonance imaging(MRI) of the brain was completed within 1 week of presentation. All patients also underwent a range of blood tests, chest radiography, and electrocardiography at admission. According to the Trial of Org 10172 in Acute Stroke Treatment (TOAST) criteria[13], IS was divided into 1) large-artery atherosclerosis, 2) cardioembolism, 3) small-vessel occlusion, 4) stroke of other determined etiology, and 5) stroke of undetermined etiology. Disease severity was estimated by the National Institutes of Health Stroke Scale (NIHSS) score on admission by the physician. For patients unable to communicate adequately, proxy respondents were used. We defined a valid proxy respondent as a spouse or first-degree relative who was living in the same home or was aware of the participant’s previous medical history and current treatments. The exclusion criteria included those who: (1) were unable to communicate and without a valid surrogate respondent; (2) had subdural hemorrhage, tumor, or brain abscess (i.e., non-vascular causes); (3) were under current hospitalization for acute coronary syndrome or myocardial infarction; and (4) an inability to obtain consent from the patient or surrogate.
Control Selection
Controls were community-based from all 8 counties and had no history of stroke. Each control was matched for age (± 5 years) with the case. Exclusion criteria for controls were identical to those described for cases.
Assessment Of Risk Factors
Data on occupation, education, fertility, monthly family income, smoking status, frequency of fast food consumption, frequency of hot pot consumption, moderate-intensity physical activity(MIPA), and history of HT, DM and hyperlipidemia were collected by the study physicians using structured questionnaires. Occupation was classified as either mental or manual work. Education was classified as either ༜9 years of school education or ≥ 9 years (senior high school or above). The history of fertility was classified as having given birth to ≤ 2 children or to ≥ 3 children. Smoking status was classified as never smoked and current smoking (smoking ≥ 1 cigarette per day within the year prior to the interview and included those who had quit smoking less than a year). MIPA was defined as 4 h or more per week, including brisk walking, dancing, gardening, housework and domestic chores, traditional hunting and gathering, general building tasks (e.g., roofing, thatching, painting), carrying/moving moderate loads (< 20 kg), and so on[14]. HT was defined as under treatment with antihypertensive medication, a previous HT diagnosis, or current HT according to the 2003 WHO criteria (blood pressure of 140/90 mmHg or higher). For those with a history of DM or treated diabetes preceding IS, diabetes was defined according to the 1999 WHO criteria as fasting plasma glucose (FPG) ≥ 7.0 mmol/L (126 mg/dL), a 2-h oral glucose tolerance test of ≥ 11.1 mmol/L (200 mg/dL), or glycated hemoglobin (HbA1c) ≥ 6.5%. We defined dichotomous components of hyperlipidemia as total cholesterol (TC) ≥ 6.2 mmol/L (240 mg/dL), triglyceride (TG) ≥ 2.3 mmol/L (200 mg/dL), low-density lipoprotein cholesterol (LDL-C) ≥ 4.1 mmol/L (160 mg/dL), or high-density lipoprotein cholesterol (HDL-C) < 1.0 mmol/L (40 mg/dL)[15].
Waist and hip circumferences were measured in the standing and supine positions. If patients were unable to stand because of disability, these measurements were completed only in the supine position[16]. Fasting blood samples (10 mL) were taken from cases and controls within 72 h of recruitment, separated by centrifugation, and frozen at -70 °C immediately after processing. Samples were shipped in packaging incorporating dry ice cooling agents by courier from every site to a blood storage site, where they were stored at -70 °C. FPG, HbA1c, TP, TC, TG, LDL-C, HDL-C, ApoA1, ApoB and high-sensitivity C-reactive protein (hs-CRP) concentrations were measured at Dian Diagnostics Group with a Beckman Coulter (Brea, CA) AU680 Clinical Chemistry Analyzer and Beckman Coulter reagents. Detailed cutoffs for WHR[17], HDL-C[15], ApoB/ApoA1 ratio[18] and hs-CRP[19] appear in Additional file 1: Table S1.
Statistical analysis
Statistical analyses were produced with IBM SPSS Statistics for Windows (version 23.0; Armonk, NY) and R for Windows(Version 4.0.3). All statistical tests were two-sided.
First, we performed a comparison between case and control groups using the Chi-squared tests and calculated univariate odds ratios (ORs) and 95% confidence intervals (CIs) for all dichotomized or multinomial risk factors. Second, binary multivariable logistic regression was applied, with backward stepwise (likelihood ratio) variable removal at the level of P < 0.10, and adjusted ORs and 95% confidence intervals (CIs) were calculated. The Breslow-Day test was used for the homogeneity of the association between modifiable risk factors and IS by ethnicity[20], if P < 0.05, we will detect the biological interaction between the modifiable risk factor and ethnicity. Three measures of biological interaction: RERI, the relative excess risk due to interaction; AP, the attributable proportion due to interaction; and S, the synergy[21]; index with corresponding 95% confidence intervals were calculated in R[22]. RERI and AP equal to 0 and S equal to 1 were defined as no interaction[21]. The names and values of variables in univariate and multivariate logistic regression analyses also appear in Additional file 1: Table S1.