Characteristics and prediction model of high-risk population of stroke in 40~75 years old residents in Tianjin

Objective To understand the detection status of stroke high-risk groups in Tianjin, analyze the general demographic distribution characteristics and the prevalence status of major related risk factors, and build a prediction model, so as to provide strategic basis for the primary prevention of stroke high-risk groups in Tianjin.Methods From 2019 to 2023, a cluster sampling method was used to select the monitoring sites in 16 project areas of Tianjin, and questionnaires, physical examination and biochemical examination were conducted among the permanent residents who lived in the monitoring sites for more than half a year. χ2 test, multiple logistic regression and ROC curve were used for statistical analysis.Results A total of 181888 subjects were included in this study, with an average age of (56.36±8.64) years, and a total of 53286 high-risk patients with stroke were detected, with a detection rate of 29.30%. The average age of high-risk groups was (57.90±8.14) years, the average BMI was (26.93±3.69) kg/m2, the average SBP level was (132.49±13.00) mmHg, the average DBP was (81.60±8.12) mmHg, and the average FBG level was (6.52±2.06) mmol/L. TC level was (5.31±1.21) mmol/L, TG level was (2.20±1.63) mmol/L, HDL-C level was (1.31±0.44) mmol/L, LDL-C level was (3.02±1.02) mmol/L, all of which were significantly higher than those in non-high-risk groups (P<0.001). The proportion of male high-risk group (37.13%) was significantly higher than that of female high-risk group (23.34%). Male, old age (≥50 years old), drinking (including abstaining from alcohol), smoking (including quitting smoking), family history (including hypertension, diabetes, hyperlipidemia, stroke, coronary heart disease) and retirement were the risk factors for stroke high-risk detection. High education level (high school/secondary school and above) and regular exercise were protective factors for stroke risk detection. The AUC for BMI alone was 0.725 (95%CI: 0.722-0.728, P<0.001), and the AUC for systolic blood pressure was 0.686(95%CI 0.684~0.689, P<0.001). The AUC for predicting stroke risk was 0.692(95%CI:0.689~0.6954, P<0.001). Gender, old age, low educational level, retirement status, smoking status, amateur sports status, family history, BMI, SBP, DBP, FBG, TC, TG and LDL-C were included as independent variables to construct a stroke risk prediction model. The model fit was as follows: Y=-0.123× sex +0.493× old age +0.109× low education +0.127× retired +2.487× smoking status -2.215× amateur sports status +2.425× family history +0.283×BMI +0.027×SBP +0.016×DBP +0.513×FBG +0.162×TC +0.453×TG +0.139×LDLC. The prediction model was used to predict the risk of stroke. The area under ROC curve was 0.931 (95%CI: 0.930 ~ 0.932, P<0.001), indicating that the model had high prediction accuracy.Conclusion The detection rate of stroke risk group in residents aged 40 to 75 years old in Tianjin is relatively high, and the risk prediction model established by using self-judged and self-measured physical indicators has a very high prediction accuracy for residents in this area.