| Objective: To investigate the association between plasma homocysteine(Hcy)and risk of hypertension.Methods: Individuals aged ≥15 free of hypertension in Emin Xinjiang were investigated with a stratified multi-stage random sampling method between January and December 2014,who were divided into three groups based on tertiles of baseline Hcy(T1/T2/T3)for following up 5 years,with incidence of hypertension compared.Kaplan Meier curve was plotted to compare the risk of hypertension among three groups and multivariate Cox proportional hazards model was used to analyze the association between Hcy and hypertension.Adjusted confounders were selected based on univariate Cox regression analysis and directed acyclic graphs(DAGs).Time-dependent receiver operating characteristic(ROC)curve was applied to determine the best cut-off of Hcy for predicting the 5-year risk of hypertension.Subgroup analysis was further performed among confounders screened by DAGs,with Cox regression model analyzing the interaction between these factors and Hcy.Results:(1)A total of 1926 subjects aged ≥15 were investigated at baseline in 2014,of whom 1463 individuals free of hypertension were included in the cohort.After a median follow-up of 4.83 years(interquartile range: 4.82-4.85),1352 subjects accomplished this work(missing rate: 7.59%)and finally included in analysis,with an average age of 48.67 years old and 58.58% of women.Individuals were divided into three groups(458 in T1 group,444 in T2 group and 450 in T3 group)based on the baseline Hcy tertiles.Compared with T1 and T2,individuals in T3 group were older,having higher systolic blood pressure and more likely to be men and drinkers,while the folic acid level was lower(P for all value<0.017).At end of the follow-up,517 patients developed into new-onset hypertension,and the risk of hypertension in T3 was higher than T2 and T1 groups(log rank test: T3 vs T2,P<0.001;T3 vs T1,P<0.001;T2 vs T1,P=0.001).Model 1 was adjusted with confounders screened by univariate Cox regression and model 2 was constructed by DAGs.In model 1,compared with T1,the risk of hypertension was increased by 24.8%(95%CI: 0.956-1.631)in T2 and 72.9%(95%CI: 1.323-2.259)in T3,respectively(P for trend<0.001),and the risk of hypertension was increased with 11.0%(95%CI: 1.044-1.180)for every5umol/L increase of Hcy.In model 2,compared with T1,the risk of hypertension was increased by 18.6%(95%CI: 0.895-1.570)in T2 and 61.9%(95%CI: 1.215-2.157)in T3,respectively(P for trend=0.001)and the risk of hypertension was increased with 8.5%(95%CI: 1.014-1.160)for every 5umol/L increase of Hcy.(2)Time-dependent ROC demonstrated the best cut-off value of Hcy for predicting 5-year risk of hypertension was 12.01umol/L(sensitivity and specificity were 58.42% and 65.02%,respectively).Subjects with Hcy>12.01umol/L had a higher risk of hypertension of 64.0%(95%CI: 1.315-2.046)than those with Hcy≤12.01umol/L,after adjusted for confounders.(3)Interaction analysis showed that gender,age,smoking,drinking,overweight/obesity,abdominal obesity and folic acid status had no interaction with Hcy.Conclusion:(1)Elevated plasma Hcy was an independent risk factor of hypertension,suggesting that reducing Hcy may be one of the effective measures to prevent and control hypertension.(2)Individuals with Hcy >12.01umol/L at baseline were at high risk of hypertension after 5 years,suggesting that it may be a potential intervention threshold of Northern Xinjiang population.(3)Traditional risk factors such as age,gender and lifestyles did not modify association of Hcy and hypertension. |
PDF Full Text Request