The Nested Case-control Study Of Trimethylamine N-oxide And Cardiovascular Disease Risk

Objective: Recent studies have shown that trimethylamine N-oxide(TMAO)was a risk factor for cardio-cerebral vascular diseases(CVD)in different clinical settings,but few studies confirmed the association in community-based general population,especially in Chinese population.We therefore tested the association between TMAO and risk of future CVD in a nested case-control study from a prospective cohort study within a large cohort of community-based general population who were followed up a near 5-year period.Method: This study was a nested case-control study from a prospective cohort design.From June 2012 to September 2017,a total of 2,454 patients without baseline CVD completed a median follow-up of 4.83 years of follow-up.During the follow-up period,a total of 86 incident cases of newly diagnosed CVD(48 CHD and 38 stroke cases)were identified.Using risk-set sampling,we then randomly selected controls who remained free from CVD and matched them to the CVD patients in a 1:1 ratio,according to age(within 3 year),sex(man/woman),duration of the follow-up period(within 1 month),and hypertension(yes/no).Stable isotope dilution liquid chromatography-tandem mass spectrometry(LC-MS/MS)was used for quantification of TMAO according to previously described method.We constructed 2 models: in model 1,we adjusted for age and ethnicity;and in model 2,we further adjusted for age,ethnicity,and a propensity score.The propensity score was calculated with a linear regression model entering TMAO as the dependent variable and the independent variables included SBP,BMI,antihypertensive medications,current smoking,current drinking,diabetes mellitus,TC,TG,HDL-C,and e GFR.In these multivariable logistic models,TMAO was categorized as quartiles(Q4,≥1.89 μmol/L;Q3,1.05-1.89 μmol/L;Q2,0.43-1.05 μmol/L;Q1,<0.43 μmol/L)and the lowest quartiles chosen as the reference group.Meanwhile,TMAO were separately entered on a continuous scale(1 μmol/L increment)in these models.To assess the impact on the predictive value of the model of adding in risk factors,the change in-2 log likelihood was calculated for each consecutive model and compared with a χ2distribution,with the degrees of freedom reflecting the parameters in each model.In addition,receiver operating characteristic(ROC)curves were constructed,and the areas under the curves(AUC)were calculated to assess the additional discriminant power of each consecutive logistic model.Result: There was no significant difference for baseline SBP,BMI,fasting serum glucose,TC,TG,LDL-C,HDL-C,and proportions of ethnicity,diabetes mellitus,antihypertensive medications,current smoking,and current drinking between CVD cases and controls except for e GFR.After adjusted for age and ethnicity,participants with TMAO of ≥1.89 μmol/L(Q4)and 1.05 to 1.89 μmol/L(Q3)versus <0.43 μmol/L(Q1)had ORs for CVD of 2.898(95% CI: 1.421-5.909)and 2.537(95% CI:1.248-5.518),respectively.After further adjusted for other potential confounders,TMAO of ≥1.89μmol/L(Q4)and 1.05 to 1.89 μmol/L(Q3)versus <0.43 μmol/L(Q1)had ORs for CVD of 2.735(95% CI: 1.328-5.630)and 2.544(95% CI: 1.251-5.172),respectively.With CVD as the end points,the first model incorporating SBP,BMI,antihypertensive medications,current smoking,current drinking,diabetes mellitus,TC,TG,HDL-C,and e GFR had an AUC of 0.664(95% CI: 0.583-0.746),a change in-2 log likelihood of 224.408.The second model was supplemented by the TMAO and led to a significant increase(P < 0.001)in the fit of the model [change in-2 log likelihood of 13.167 with 1 df,AUC was 0.732(95% CI: 0.658-0.806)].Conclusion: In the community-based general population,there was a positive association between TMAO and future risk of CVD.Addition of TMAO improved the prediction of CVD beyond traditional risk factors.We recommend that TMAO be considered as a potential novel preventive target in the management of low-risk CVD adults.