Clinical benefit of lowering TG and LDL-c proportional to absolute change in ApoB
Association of Triglyceride-Lowering LPL Variants and LDL-C–Lowering LDLR Variants With Risk of Coronary Heart Disease
Introduction and methods
The CV risk-lowering effects of lowering LDL-c levels by reducing LDL particles through upregulation of the LDL receptor (LDLR) is well established. Indeed, all major clinical guidelines recommend treatment to lower plasma LDL-c levels. Guidelines do not, however, recommend treatment to lower plasma triglyceride (TG) levels, because the randomized trial evidence on its effect is less consistent [1,2].
Rare loss-of-function (LOF) mutations in the lipoprotein lipase (LPL) gene have been found to be associated with higher plasma TG levels and a higher CV risk. LOF mutation in the genes encoding for natural inhibitors of LPL such as APOC3, ANGPTL3 and ANGPTL4, on the other hand, are associated with lower TG levels and a lower CV risk [3-6]. It is, however, unknown whether lowering TG levels via targeting the LPL pathway will result in lower CV risk.
Apolipoprotein B (ApoB)-containing lipoprotein particles carry both TG and cholesterol in plasma. ApoB-containing lipoproteins such as TG-rich lipoprotein particles and LDL particles all have a single ApoB molecule. That allows comparison of the clinical benefit of lowering TG levels to the benefit of lowering LDL-c per unit change in ApoB.
This study used mendelian randomization to compare the associations of TG-lowering LPL variants and LDL-c-lowering LDLR variants with the risk of CVD per unit difference in ApoB, to make inferences about the potential clinical benefit of lowering plasma TG levels as compared with lowering LDL-c levels. Various data sources were used, totaling individual participant data from 470.478 participants (30.328 cases of CHD) and summary-level data of 184.305 participants (60.801 CHD cases). The LPL genetic score was based on 5 independently inherited variants that were associated with lower plasma TG levels. Similarly, the LDLR score was based on 3 independently inherited variants.
- For each 10-mg/dL (to convert to mmol/L, multiply by 0.0113) lower level of ApoB-containing lipoproteins, the LPL score was associated with 69.9 mg/dL (95%CI: 68.1-71.6, P=7.1x10^-1363) lower plasma TG level and 0.7 mg/dL (95%CI: 0.03-1.4, P=0.04) higher plasma LDL-c level.
- Similarly, for each 10-mg/dL lower level of ApoB-containing lipoproteins, the LDLR score was associated with 14.2 mg/dL (95%CI: 13.6-14.8, P=1.4x10^-465) lower plasma LDL-c and 1.9 mg/dL (95%CI: 0.1-3.9, P=0.4) lower TG level.
- The LPL and LDLR scores were associated with similarly lowered risk of CHD per 10 mg/dL lower level of ApoB-containing lipoproteins (LPL-score: OR: 0.771, 95%CI: 0.741-0.802; LDLR-score: OR: 0.773, 95%CI: 0.747-0.801).
- Stratified analyses suggested that the associations of the LPL and LDLR scores with plasma lipids, lipoproteins and the risk of CHD were independent of each other.
- 2x2 Factorial mendelian analyses revealed that combined exposure to both the LPL and LDLR genetic scores was associated with linearly additive lower levels of TG, LDL-c and ApoB, and a log-linearly additive decrease in the risk of CHD.
- Sensitivity analyses of variants in genes for therapeutic targets that lower TG via the LPL pathway, and variants in targets of LDL-c lowering therapies through the LDLR pathway, were associated with similarly lowered risk of CHD per unit difference in ApoB as compared with the LPL and LDLR scores, and an APOB genetic score.
This mendelian randomization analysis showed that TG-lowering LPL variants and LDL-c-lowering LDLR variants were associated with similar reductions of CHD risk per unit lower level of ApoB-containing lipoproteins. Their effect on CHD risk was similar, despite the observation that the LPL and LDLR gene scores were associated with different associated lipid levels. The analyses suggested that the associations between lower TG level and lower LDL-c level with risk of CHD are independent, additive and proportional to the absolute change in ApoB.
Thus, these findings suggest that the clinical benefit of lowering TG levels is similar to that of lowering LDL-c levels, per unit change in ApoB-containing lipoproteins. Consequently, all APoB-containing lipoprotein particles, including TG-rich VLDL particles and their metabolic remnants and LDL particles, appear to have about the same effect on the risk of CVD per particle. It may be anticipated that the clinical benefit of lowering TG or LDL-c levels, or both, is proportional to the absolute change in ApoB-containing lipoproteins, irrespective of the observed changes in plasma TG or LDL-c levels.
Navar  notes that inconsistent findings have been published on the relation between TG levels and CHD risk. That is not unexpected given the complexities of TG metabolism. Serum TG measurements assess the total mass of TG rather than the type of particles carrying the TG, such as chylomicrons, or VLDL particles (of which either the number can be elevated, or the TG content therein). Hypertriglyceridemia should therefore not be considered a single disease, but rather a heterogeneous set of disorders, possibly with different CV risk profiles.
About the study by Ference et al., she states that ‘Given the different ApoB-containing particles affected by LDLR and LPL, this finding provides some of the first evidence that VLDL particles are as
atherogenic as LDL particles.’ (…)
‘The use of ApoB as a lens through which to examine the benefit of genetic variants associated with lower LDL-C or TG levels is a biologically informed approach that helps simplify the complexity of lipid biology. This is particularly relevant for understanding TGs.’ The data suggest that the magnitude of TG lowering needed to reduce a given amount of ApoB is about five time that of the extent of LDL-c lowering for the same amount of ApoB lowering. This may explain why results have been mixed for trials involving TG-lowering therapies.
Navar also notes some limitations of the approach of Ference et al.. While mendelian randomization assesses the potential effect of genetically altered exposure to lipids from birth, participants in clinical trials are exposed to therapy for a much shorter period and often only once atherosclerotic disease has developed. This makes extrapolation from mendelian randomization studies to treatment effects problematic. Moreover, mendelian randomization does not account for pleiotropic, non-lipid related effects. This point she illustrates by describing the results of the REDUCE-IT trial, in which high-dose ω-3 oil eicosapentaenoic acid (EPA) lowered the relative risk of CV events. EPA did lower TG levels, but had a negligible effect on ApoB, suggesting that EPA yielded benefits through non-TG, non-ApoB-related, off-target effects.
Navar concludes that no direct causal link between TGs and CHD risk is proven by this study. ‘ When an ApoB-containing particle is trapped in the arterial wall, all of the contents of that particle, including TGs, ApoB, phospholipids, and cholesterol esters, are present. Whether the TGs alone cause atherosclerosis, or are inert bystanders in that process, still remains to be determined.’
Nevertheless, important messages arise from this work by Ference et al.. Treatments that lower LDL-c or TG levels will lead to CHD risk reduction proportional to the ApoB reduction. Giving the accumulating evidence on the importance of ApoB, guidelines should consider giving routine ApoB measurement more attention.