In this MR study, we demonstrated that a genetically predicted rise in apoB levels increased the risk of most cardiovascular diseases (CVD) in European populations. Our findings were based on publicly available summary statistics-level data, and we ensured the results' consistency through rigorous variant selection and various sensitivity analyses. Previous studies have primarily investigated the link between serum lipid levels and typical CVD cases. However, to our knowledge, our study stands out as one of the few to present evidence of an association between apoB and CVD risk in European populations using the MR framework at the genetic level[22].
Reducing the number of apoB particles in the arterial lumen is crucial in preventing the entry of these particles into the arterial wall and being trapped under the arterial wall endothelium. Furthermore, larger amounts of cholesterol trapped in apoB particles within the arterial wall leads to more cholesterol being released at that site, potentially leading to more damage to the arterial wall. Thus, diminishing the number of apoB particles through treatment can be beneficial as it reduces the damage caused to the arterial wall, increasing the likelihood of healing[23].
Cholesterol is the primary component of all apoB particles, which are commonly associated with LDL-C and non-HDL-C. Changes in lipid levels, therefore, alter the cardiovascular risk associated with apoB. However, it is important to note that the quality of cholesterol in each apoB particle varies, and LDL-C and non-HDL-C are not always equivalent to apoB. Clinical laboratories can more accurately measure apoB compared to LDL-C or non-HDL-C. In 2019, the European Society of Cardiology/European Atherosclerosis Society emphasized that apoB is a more accurate marker of cardiovascular risk than LDL-C and non-high-density lipoprotein cholesterol[24]. Therefore, guidelines propose using apoB as a secondary treatment target to reduce the residual cardiovascular risk of LDL-targeted therapies[25].
ApoB simplifies messages from regular lipid markers of atherogenic risk. However, in a study of UK Biobank participants with baseline cardiovascular disease and not taking statins, measuring total and HDL cholesterol in the non-fasting state was sufficient to capture lipid-related risk in the prediction of cardiovascular disease, while increasing lipoprotein, direct or calculated LDL cholesterol unlikely to be beneficial for predictive purposes[26][27]. A previous prospective cohort study showed that 165,544 individuals without known CVD were recorded and it was determined that the addition of lipoprotein B testing to total cholesterol and high-density lipoprotein cholesterol (HDL-C) slightly improved CVD risk prediction[28]. However, multivariate Mendelian randomization studies have revealed that the association of other lipid components with CHD risk significantly attenuated to zero after considering the impact of lipoprotein B[14]. MR studies have shown that the clinical benefits of reducing triglyceride and LDL-C levels for CHD may be directly proportional to the absolute changes of apoB[29]. Similarly, the variable selection method is applied to CVD risk prediction. Using two complementary data-driven variable selection methods (LASSO stability selection, parametric, and survival random forests, non-parametric), it is found that apoB has a high degree of choice for CVD prediction[30]. Moreover, genetic differences in apoB suggest that the Mendelian randomization of apoB in association with CVD is consistent with clinical investigations. These findings are significant for preventing CVD.
For the relevant hypotheses, after adjustment for confounders such as age and sex, we included 183 SNPs to investigate associations with CVDs. The exclusion restriction hypothesis suggests that horizontal pleiotropy may occur because multiple apoB-associated SNPs can affect other traits. In order to simulate the apoB-lowering effect of PCSK9 inhibitors, we used SNPs within PCSK9. However, despite these efforts, MR-Egger and IVW consistently showed correlations between the studied variables. Also, MR-Egger, which can detect overall pleiotropy, did not detect any significant pleiotropy in our case. Furthermore, the balanced pleiotropy was supported by the symmetric distribution in the funnel plot.
Our discovery adds to the current body of evidence suggesting that apoB plays a crucial role in the development of many CVDs. And genetically proxied inhibition of PCSK9 associated with apoB was associated with reduced risk of CHD and HF. Numerous observational studies have already established a link between increased apoB levels and the onset of CVD. Unfortunately, conducting randomized controlled studies is challenging due to the extended period it takes for CVD to emerge, and many influencing factors cannot be removed during clinical investigations. As a result, we conducted an MR study utilizing apoB predicted by genetic variation, which analyzed the situation in terms of genetic variation, making it more likely to avoid the influence of external factors, improving the quality of evidence. However, the biological function of many of the SNPs used as IVs remains unknown. Therefore, the need for additional experimental studies to elucidate the underlying mechanisms remains crucial to strengthen the credibility of the evidence.
The current study represents pioneering research as it is the first to utilize a two-sample MR approach to determine the genetic correlation between apoB levels and diverse types of CVD. To ensure precise and unbiased results, we utilized the largest and up-to-date dataset available from the UKBB and limited our analysis to individuals of European descent, thereby mitigating potential population stratification. Furthermore, we carefully selected SNPs for analysis utilizing a stringent criterion (r2 < 0.001), and optimized F value calculations to maximize statistical power. By leveraging the random assignment of genotypes, MR minimizes the influence of confounding factors and reverse causality. To strengthen our results, we conducted multiple sensitivity analyses utilizing MR-Egger and weighted median techniques, thereby minimizing the probability of bias. While epidemiological studies have previously implicated apoB concentrations in coronary heart disease risk, this research expands the scope of the investigation to encompass various CVD types[10, 23].
According to this research, genetically predicted apoB may be a causal contributor to a variety of CVDs but is unlikely to be a causal risk factor for VTE. A previous population-based cohort study also showed no association between lipoproteins and VTE risk[31]. A nonlinear effect cannot be ruled out, so additional MR studies are required even though our research found no relationship between apoB and the risk of VTE. However, similar to previous MR studies, we were failed to rule out the chance that SNPs as IVs violated the hypothesis. The analysis of apoB and results showed significant heterogeneity, which may have an impact on the stability of the results.