| Objective: Cardiovascular disease(CVD)is the leading cause of death in China and the world.In China,40% of deaths are attributed to cardiovascular diseases,which bear a heavy burden.Hypertension has always been recognized as an independent risk factor for cardiovascular events.More than half of cardiovascular diseases in China are related to hypertension.The latest national data showed that the population with hypertension in China had exceeded 300 million,and the prevalence of hypertension in rural areas of Northeast China was as high as 50%.Therefore,the prevention and control of cardiovascular disease for Chinese rural hypertensive population is imminent,and the overall cardiovascular risk assessment and risk stratification are important strategies for the prevention and control of cardiovascular disease.At present,many risk-assessment models have been used to predict cardiovascular risk in the general population,but their applicability in hypertensive population needs to be evaluated.Moreover,the risk-assessment model has disadvantages,because most obvious cardiovascular diseases occur in very large populations with low to medium risk.Recent studies have shown that in individuals at moderate risk of cardiovascular disease,metabolomic characteristics increase the incremental identification ability beyond standard clinical variables.In addition,the identification of metabolomic risk spectrum can improve the risk stratification of cardiovascular disease.Therefore,based on the prospective cohort study of a rural hypertensive population in Northeast China,this study aims to find out the prevalence of cardiovascular disease in hypertensive population in rural areas of Northeast China;The cardiovascular disease risk assessment model of hypertensive population is constructed based on Lasso-Cox analysis;The application of non-targeted metabolomics aims to explore new risk factors of cardiovascular disease in hypertensive population,and further improve the risk stratification and treatment decision-making of cardiovascular disease in hypertensive patients in China.Methods:1.Based on the natural population cohort established by the Northeast China Rural cardiovascular and cerebrovascular health study(NCRCHS)conducted in rural areas of Northeast China from 2012 to 2013,this study screened the population with hypertension according to the definition of the 2017 American College of Cardiology /American Heart Association guidelines for hypertension.The prevalence of cardiovascular disease in hypertensive population at baseline was analyzed and the risk factors of cardiovascular disease were analyzed by Logistic regression.2.The NCRCHS study was followed up from 2015 to 2018 to collect new incident cardiovascular events.The hypertensive population without cardiovascular disease at baseline were included and randomly divided into development cohort and validation cohort according to 2:1.The candidate variables were screened by lasso regression to realize the dimensionality reduction and optimization of the model and prevent over fitting.All selected variables were included to establish Cox regression model and nomogram model of cardiovascular disease in hypertensive population.The validation of nomogram model was evaluated by consistency index(C-index),calibration degree and area under receiver operating curve(AUC).The blood pressure classification was used to update the model.The original model and the updated model were compared by calculating the net classification improvement index(NRI)and the comprehensive discrimination improvement index(IDI).3.The hypertensive population without cardiovascular disease at baseline were included based on the NCRCHS study,which was a prospective cohort study.And 30 cases of new cardiovascular disease were randomly selected during a median follow-up of 4.66 years.Hypertensive participants without cardiovascular disease were selected as the control group according to gender,age,blood pressure grade and region,which were 1:1 matching to the case group.The difference of metabolomic characteristics between the two groups of baseline plasma samples was analyzed by ultra-high performance liquid chromatography-mass spectrometry.After obtaining the sorted data,we carry out a series of multivariate pattern recognition analysis: principal component analysis(PCA)and orthogonal partial least squares discriminant analysis(OPLS-DA).The criteria for screening differential metabolisms was which p-value of student’s t-test was less than 0.05 and the variable projection importance(VIP)of the first principal component of OPLSDA model was greater than 1.The metabolic pathways of different metabolites were screened by enrichment analysis and topological analysis.A group of metabolites with the strongest association of cardiovascular disease in hypertensive population were screened according to the criteria that the differential metabolites with area under ROC curve(AUC)> 0.8.After adjusting the traditional risk factors by logistic regression model,the relationship between these differential metabolites and hypertensive cardiovascular disease was further analyzed.Pearson method was used to analyze the correlation between these metabolites and clinical indexes(blood pressure,blood lipid,etc.).Results: 1.At baseline,8723 hypertensive patients were included in the study.And the crude prevalence rate of cardiovascular disease in hypertensive population was 13.7%(95%CI 13.0%-14.5%).With the increase of blood pressure and age,the prevalence rate of cardiovascular disease in hypertensive population increased gradually.The prevalence of cardiovascular disease in hypertensive participants with blood pressure between 130-139/ 80-89 mm Hg was 9.0%(95% CI 8.0%-10.2%).The prevalence rate of coronary heart disease in female participants with hypertension was 7.6%,which was significantly higher than that in male(4.9%)(P < 0.001).Multivariate logistic regression analysis showed that the risk of cardiovascular disease increased with each increase of age,BMI,mean SBP and mean DBP(age: OR 1.052,95% CI 1.053-1.065;BMI: OR 1.026,95% CI 1.007-1.045;mean SBP: OR 1.005,95% CI 1.001-1.009;mean DBP: OR 1.010,95% CI 1.002-1.017). Diabetes,Low HDL-C,High TG and High UA increased the risk of cardiovascular disease in hypertensive participants(diabetes: OR 1.320,95%CI 1.108-1.573;Low HDL-C: OR1.245,95%CI 1.039-1.491;High TG: OR 1.222,95%CI 1.040-1.437;High UA: OR 1.352,95%CI 1.081-1.690).The risk of cardiovascular disease in patients with hypertension who drink alcohol,eat more fish and lean meat,and engage in moderate and severe labor intensity was reduced.Hypertensive participants with family history of hypertension,family history of coronary heart disease and family history of stroke had an increased risk of cardiovascular disease(family history of hypertension: OR 1.683,95% CI 1.449-1.954; family history of stroke: OR 1.489,95% CI 1.272-1.742;family history of coronary heart disease: OR 1.524,95% CI 1.286-1.807).2.The baseline characteristics of the developing cohort(n = 4055)were similar to those of the validation cohort(n =2027).Of the 6082 hypertensive participants included in the study,331(5.4%)had cardiovascular events during a median follow-up of 4.66 years.Nine independent variables were screened by the Lasso method(when lambda was by a standard error lambda.1se): gender,age,transient ischemic attack(TIA),family history of diabetes,family history of hypertension,family history of stroke,glomerular filtration rate(e GFR),high low density lipoprotein cholesterol(LDL-C)and low serum magnesium.A nomogram model was established.The nomogram model was updated by blood pressure classification.The c-index of the primary model and the update model were 0.736(95% CI 0.706 – 0.766)and 0.766(95% CI 0.737 – 0.795)in the developing cohort,respectively.And 0.711(95%CI 0.665 – 0.757)and 0.751(95% CI 0.707 – 0.795)in the validation cohort,respectively.The AUCs of the prediction model in the developing cohort and the validation cohort were as follows: the AUC of the 2-year cumulative CVD risk of the primary model was 0.744 of the developing cohort and 0.719 of the validation cohort,respectively;the AUC of 4-year cumulative CVD risk of the primary model was 0.746 of the developing cohort and0.742 of the validation cohort,respectively;the AUC of 2-year cumulative CVD risk of the updated model was 0.779 of the developing cohort and 0.782 of the validation cohort;the AUC of 4-year cumulative CVD risk of the updated model was 0.778 of the developing cohort and 0.780 of the validation cohort,respectively.The calibration curves of the primary model and the updated model both indicated a good calibration degree.Net reclassification index(NRI)and comprehensive discriminant improvement index(IDI)suggested that the reclassification of the updated model was more accurate and the performance of the model was better.3.There was no significant difference in baseline characteristics between case group(n =30)and control group(n = 30).PCA analysis and OPLS-DA analysis of plasma metabolites in positive and negative ion combination mode showed that there were differences between the two groups.The VIP value > 1 of the OPLS-DA model and the P < 0.05 of student’s ttest were used as the criteria for screening differential metabolites.A total of 57 differential metabolites were screened.36 metabolic pathways were found to be involved in the pathogenesis of cardiovascular disease in patients with hypertension,of which 7 were the main metabolic pathways.Seven main metabolic pathways including D-arginine and Dornithine metabolic pathways(Raw P = 0.077,Impact 0.500);Cyano-amino acid metabolic pathways(Raw P = 0.150,Impact 0.333);Alanine,aspartate and glutamate metabolic pathways(Raw P = 0.023,Impact 0.207);Pyruvate metabolic pathway(Raw P = 0.276,Impact 0.183);Arginine and proline metabolic pathways(Raw P < 0.001,Impact 0.123);Glycerol phospholipid metabolic pathway(raw P = 0.006,Impact 0.065);Taurine and low taurine metabolic pathways(Raw P = 0.016,Impact 0.022).Fifty-seven differential metabolites were analyzed by ROC curve.Finally,nine significant differential metabolites with AUC > 0.8 under the ROC curve were selected,and the traditional cardiovascular risk factors were adjusted for logistic regression analysis.We found that seven differential metabolites were related to the increased risk of cardiovascular disease: 4-Hydroxybenzaldehyde,N-cyclopropyl-trans-2-cis-6-nonanediamine,Octadecyl-amine,Phosphatidylcholine(16:0 / 16:0),Sphingomyelin(d16:1 / 24:1(15z)),Sphingomyelin(d18:0 / 18:1(9z))and Sphingomyelin(d18:1 / 20:0);two metabolites were associated with a reduced risk of cardiovascular disease:(3xi,6xi)-Cyclo-(alanyl valyl)and 5-Aminoimidazole nucleotides.The AUC of the prediction model constructed by 9significantly different metabolites was 0.969,95% CI(0.926-1.000).Among them,4-Hydroxybenzaldehyde was significantly correlated with SBP(r = 0.265,P < 0.05),5-Aminoimidazole nucleotide was significantly correlated with DBP(r = 0.335,P < 0.05),Phosphatidylcholine(16:0 / 16:0)was significantly correlated with DBP(r =-0.296,P <0.05),Sphingomyelin(d16:1 / 24:1(15z))was significantly correlated with LDL-C(r =0.402,P < 0.05).Conclusions: 1.This study reported for the first time that the prevalence of cardiovascular disease in hypertensive population based on the new diagnostic criteria of hypertension which was 130 / 80 mm Hg in rural areas of Northeast China was high,and the prevalence of coronary heart disease in female hypertensive population was higher than that in male.2.Poor blood pressure control,diabetes,dyslipidemia,hyperuricemia and elevated BMI all increased the risk of cardiovascular disease in hypertensive patients.3.Based on the combination of traditional cardiovascular risk factors and clinical biochemical indexes by applying Lasso-Cox regression model,a new prediction model and nomogram were established for the first time to predict the short-term risk of cardiovascular disease in hypertensive population(2 and 4 years).4.At the same time,the study of plasma non-targeted metabolomics found that there were significant differences in plasma metabolic spectrum between hypertensive patients with cardiovascular disease and hypertensive patients without cardiovascular disease.5.Nine different metabolites and seven main metabolic pathways were found.The combination of the nine metabolites can well predict the occurrence of cardiovascular disease in hypertensive population. |
PDF Full Text Request