| Summary: | 碩士 === 國立陽明大學 === 公共衛生研究所 === 107 === Background
Cardiovascular disease (CVD) has been ranked as No. 2 of the top ten leading causes of death in Taiwan for many years. It is an important public health issue in the modern society. Dyslipidemia was found to be one of the risk factors for cardiovascular disease. The LPL gene and APOE gene were found to be involved in lipid metabolism. If the protein had functional deficiencies, it may cause atherosclerosis and hypercholesterolemia. Though researchers have found the association between dyslipidemia and cardiovascular disease, conventional observational studies may suffer from un-measured confounders and from reverse causality, thus may limit the ability to identify a causal relationship between them. We aimed to explore the causal relationship between dyslipidemia and cardiovascular disease by using conventional observational design, as well as by using Mendelian randomization, which used rs328 of LPL gene and rs157580 of APOE gene as the instrumental variables to establish a causal relationship between dyslipidemia and cardiovascular disease.
Methods
A prospective design was used for the study. We recruited participants who joined the Matsu Community-Based Integrated Screening (MA-CIS) in both year 2015 and year 2017. We included 1951 Chinese participants with DNA sample available. The rs328 of LPL gene and rs157580 of APOE gene were selected as the instrumental variable, and were genotyped by using the TaqMan® SNP Genotyping Assay. Cardiovascular disease was defined according to self-reported disease history or medication use. The relationship between blood lipids and cardiovascular disease were estimated by logistic regression and instrumental variable analysis. The instrumental variable analysis applies two stage least square method. The results between logistic regression and instrumental variable analysis were compared by using Hausman Specification test. In addition, the F-statistics was used to evaluate the strength of the instrumental variables.
Results
In the conventional observational study, we found a significantly protective effect of the increasing LDL-C levels on cardiovascular disease (OR=0.987, 95% CI=0.980-0.994), but we did not find a significant association between triglycerides and HDL-C with cardiovascular disease (TG: OR=1.002, 95% CI=1.000-1.004; HDL-C: OR=0.984, 95% CI=0.969-1.000). In the instrumental variable analysis, we found an increasing risk of LDL-C levels on cardiovascular disease by using rs157580 and allele score as the instrumental variables (rs157580: OR=1.028, 95% CI=1.018-1.038; allele score: OR=1.028, 95% CI=1.019-1.038). Triglyceride and HDL-C were also found to be associated with cardiovascular disease risk. Though we observed rs328 was significantly associated with decreased triglyceride levels (β=-9.105，p=0.012) and increased HDL-C levels (β=1.317，p=0.037), the association between instrumental variable and lipid parameters are not strong, indicating rs328, rs157580 and allele score may not be a suitable instrument variable in this study. The result from instrumental variable analysis couldn’t reflect an actual relationship between dyslipidemia and cardiovascular disease.
Conclusions
We found a protective effect of the increasing LDL-C and HDL-C levels on cardiovascular disease, and the raising triglyceride levels will increase the risk of cardiovascular disease through conventional observational study. Although we found a causal relationship between dyslipidemia and cardiovascular disease by using the rs157580 and allele score as the instrumental variable, there is no strong association between instrument variable and lipid parameters. Thus, we are unable to establish an actual relationship between dyslipidemia and cardiovascular disease by using Mendelian randomization. We can only establish the relationship between dyslipidemia and cardiovascular disease by conventional observational design in this study.
|
|---|
