lipoprotein (a): a bad actor

 A recent Mendelian randomization trial found a profound atherogenic effect for lipoprotein(a), significantly more so than LDL (see lipids Lp(a) more atherogen than LDL JACC2024 in dropbox or doi.org/10.1016/j.jacc.2023.10.039)

 

Details:

-- these researchers used 2 genomic databases to assess the role of lipoprotein(a), referred to as Lp(a), versus LDL in terms of atherogenicity,  on a per-particle basis. They were able to assess the individual particles in both of these lipid moieties by measuring the apolipoproteinB (apoB) concentrations (each one of these lipid moieties has a single apoB on its membrane)

    -- they accessed the UK Biobank population of 502,413 UK residents of mainly European ancestry and 54% female

    -- as a replication cohort, they accessed the CARDIoGRAMplusC4D data sets (Coronary ARtery DIsease Genome-wide Replication and Meta-analysis, plus The Coronary Artery Disease Genetics), with 98 of the Lp(a) cluster SNPs (single nucleotide polymorphisms) and 130 of the LDL cluster SNPs

-- both of these genome-wide association studies (GWAS) were adjusted for age, sex, and several genomic principal components of SNPs

 

Results:

-- the GWAS analyses found these genetic variations in apoB:

    -- for the Lp(a) cluster: 107 SNPs were found, all located in a region around the LPA gene

        -- variants of this LPA gene have been associated with risk of myocardial infarction, stroke, and aortic valve stenosis

    -- for the LDL cluster: 143 SNPs were found, with more widely dispersed loci

        -- there was essentially no overlap between the SNPs associated with each of these clusters (i.e. the polymorphisms associated with both of these lipid moieties were basically independent of each other)

 

-- Findings:

   -- one mole of Lp(a)-apoB was indeed equivalent to one mole of Lp(a) mass concentration, supporting the validity of Lp(a)-apoB as a genetic instrument to assess the association with CHD (coronary heart disease) risk

    -- comparing the per-particle atherogenicity of Lp(a) versus LDL, per 50 nmol/L increase in apoB levels, for the UK Biobank:

        -- Lp(a): 1.28 (1.24-1.33),p=4.06x10^-47

        -- LDL: 1.04 (1.03-1.05), p=3.23x10^-17

    -- comparing the per-particle atherogenicity of Lp(a) versus LDL, per 50 nmol/L increase in apoB levels for the replication analysis (CARDIoGRAMplusC4D):

        -- Lp(a): 1.17 (1.13-1.20), p=4.53x10^-21

        -- LDL: 1.04 (1.03-1.05), p=3.72x10^-13

            -- i.e.: this replication data set basically confirms the results from the UK BioBank

 

-- relative atherogenicity of CHD risk, comparing Lp(a)-apoB versus LDL-apoB:

    -- UK Biobank: 6.6-fold (5.1- 8.8)

    -- replication analysis: 3.8-fold (2.7-5.4)

this graph highlights the increases in CHD incidence rate by the 2 clusters according to their polygenic scores (PGS)

 

and here is a graphic representation of the related CHD risk for both moieties

 

-- in this analysis they did try to identify potential SNP pleiotropic effects (these are the effects that are above and beyond the direct lipid effects, such as inflammation or endothelial dysfunction or oxidative stress; they found that controlling for potential pleiotropic effects did not seem to affect their results much

 

Commentary:

--Lp(a) is effectively a combination an LDL-like particle with a single apoB on its cellular membrane and is covalently linked with apo(a); apo(a) is structurally similar to plasminogen and thereby might inhibit plasminogen binding and lead to impaired fibrinolysis. So structurally Lp(a) is both pro-atherogenic and potentially pro-thrombotic. There are an inconsistent number of repeated apo(a) units on the Lp(a), and these units are referred to as krinkles (for recent paper on many of these Lp(a) findings, see https://academic.oup.com/eurheartj/article/43/39/3925/6670882 )

    --Lp(a) can be oxidized, as with LDL, and these oxidized particles attach to macrophages most easily and thereby induces atherosclerosis, as well as triggering inflammation

    -- Lp(a) is largely genetically determined and varies widely by ethnic groups, and 30-70% of the variation in the number of krinkles seems to be related to ethnicity

-- several studies over the past 40 to 50 years have identified Lp(a) as a potentially strong atherosclerotic risk factor. Prior studies have indeed found that Lp(a) is both pro-inflammatory and pro-atherosclerotic, and that it is a risk factor for cardiovascular disease even in those with low LDL levels. The observational clinical studies during much of the last several decades, however,  have had somewhat inconsistent results, some finding no additional atherogenicity for Lp(a) over LDL, others finding significantly higher atherosclerotic risk that was independent of the LDL levels

-- as a disclaimer, I am no expert in Mendelian randomization (MR), and I did not go into all of the extensive details in this study, but in brief:

    -- MR is basically a technique that relies on specific genetic variations called single nucleotide polymorphisms (SNPs) that occur naturally over time and are found to be strongly associated with a specific outcome (in this case: apoB), with the assumption that these SNPs/genetic variations are randomly represented in the population. The goal is to see if there is a causal relationship with an observational clinical outcome (in this case: cardiovascular diseases) with these pieces of genetic code. 

        -- so, this technique represents a relatively easy-to-do analysis, sort of like a "randomized controlled trial" (since these SNPs are assumed to be randomly represented in the population), and pretty much eliminates reverse causation (this is really an issue with a behavior, eg the possibility that cardiovascular disease leads people to drink instead of vice versa)

-- the conclusions of this study is that Lp(a) is a highly atherogenic lipid moiety, and this is important for several reasons:

    -- we focus on LDL as the marker of atherogenesis as well as our lipid marker for decreasing cardiovascular risk through pharmacologic and nonpharmacologic means

    -- we do know that LDL is an electrophoretic classification, that there are several different LDL particles within the range identified as LDL, and that small dense LDL particles are roughly three-fold more atherogenic than the larger fluffier LDLs

        -- these small dense LDL particles are associated with diabetes and metabolic syndrome, for example. Which does raise the important issue of us clinicians being more aggressive in lipid management in patients with these particular syndromes. For more information on this see lipid dense LDL more atherogenic MBCcardiovasc2023 in dropbox or doi.org/10.1186/s12872-023-03578-0; as well as https://gmodestmedblogs.blogspot.com/2023/10/update-ascvd-risk-factor-critique.html

    -- there are other lipid moieties also that may affect atherogenesis other than LDL, including remnant cholesterol:

        -- another recent Mendelian randomization study found that remnant cholesterol (the cholesterol carried in chylomicrons, chylomicron remnants, VLDL, and IDL lipoproteins) conferred increased risk of coronary artery disease, myocardial infarction, and stroke independent of the effects of LDL cholesterol levels: see lipid remnant cholesterol causal not LDL ArtRhrombVascBio2023 in dropbox, or DOI: 10.1161/ATVBAHA.123.319297)

-- a few comments about apoB:

    -- focusing on apoB is a great way to look at the specific number of particles in the blood, since there is one apoB on each of the LDL and Lp(a) particles, allowing for a per-particle comparison

    -- apoB does have a well-standardized laboratory assays, which decrease the variability from laboratory to laboratory 

    – the apoB itself maintains a pretty stable and uniform protein structure between individuals

 

-- A few comments about Lp(a):

-- Lp(a), as opposed to apoB, has much more significant variations by assays. Part of the issue is that the size of the Lp(a) can vary a lot from person to person, depending on the number of krinkles attached 

-- the only drugs that we have now that really decrease the Lp(a) levels are PCSK9 inhibitors, nicotinic acid (niacin) in the 2 to 4 g/d range, and estrogens (niacin and estrogens are pretty much not being used for cardioprotection anymore)

    -- PCSK9 inhibitors (FOURIER trial with evolocumab; ODYSSEY trial with alirocumab) lower the Lp(a) levels roughly 23%-29%, in addition to lowering the LDL by 50%-60%. These drugs lead to especially large reductions in the relative risk of cardiovascular disease in those with high baseline Lp(a) levels (these patients experienced the greatest decrease in these cardiovascular events, independent of LDL changes)

    -- there are two drugs in development that specifically target Lp(a): pelacarsen and olpasiran

--  however, on a population basis, high LDL levels are much more commonly encountered in than high Lp(a) levels.

Limitations:

-- this was a huge study but with a relatively uniform population (white, European). do these results apply to other areas of the world?

    -- LDL levels are quite susceptible to the environment: diets with lots of saturated or trans fats, smoking, weight, physical activity, some meds (steroids, some HIV meds), CKD, even stress and depression, etc.  And these vary a lot in individuals and in different areas of the world

    -- Lp(a) is the opposite: basically genetic, though there might be some association with diet (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7400957/) and estrogens/androgens

        – which brings up another issue: genetics are clearly not determinant in LDL but are more so with Lp(a). this muddies the water a bit when a study such as this one only assesses the genetic component. the issue is undoubtedly more complicated than what was in this article

--the range of Lp(a) levels in their population were pretty much the middle levels: ie only 10% were below the lower limit of normal for the assay (<3.8 nmol/L), and 6-7% were above the upper limit (>189 nmol/L): the mean level in the study was 44.6 nmol/L and median 21.2 nmol/L. This might limit generalizability to a larger and more diverse population

--there was no information about whether the people in the study had comorbidities (including heart disease, but also the array of cardiac comorbidities), were on meds/which ones, were smokers, what their BMI was, all of which might have affected the results (especially for LDL levels)

-- as mentioned above, Lp(a) comes in many sizes, depending on the number of krinkles attached.  Do all of these Lp(a)s with different numbers of krinkles behave the same way in terms of cardiovascular risk? are those with more krinkles more pro-thrombotic and derive more benefit by decreasing Lp(a) levels? if so, what are the parameters? should we have assays assessing krinkle size (perhaps by some ratio of total Lp(a) content divided by apoB levels??)

-- as pointed out by the researchers, the Lp(a) SNPs were all on a small area of chromosome 6. Perhaps there are linkages between some of the SNPs, and that could change the calculated risk (the researchers tried to control for this)

    – but, maybe even more distant SNPs have important interactions with each other, perhaps increasing or decreasing the effectiveness of these SNPs in meaningful ways. it may be that  these interactions might be missed by assessing changes in the just the single SNPs? and perhaps the researchers are simply missing some important SNPs to test for? perhaps these are more rarely found and missed by their evaluation? there were hundreds found, though it seems that newer MR evaluations often find more SNPs than earlier studies. and, as noted, almost all of the Lp(a) levels were in the mid-range and not at the more extreme levels. perhaps a more-skewed level of Lp(a)s in a more ethnically diverse population would elicit very important unrecognized SNPs????

so, what is the utility of this study for us in primary care??

--it does tangentially promote the importance of measuring apoB levels (as is the norm in Europe, per their 2023 guidelines), and apoB levels are higher for those with the more atherogenic but small dense LDLs when compared to the total LDL concentration (since there is one apoB on each LDL, whether small or big)

--it makes it clearer that LDL is only one of many lipid markers for cardiovascular disease. for example, patients with diabetes are therefore at higher risk of having the more atherogenic small LDLs. should they be treated more aggressively to lower the LDLs further? the ADA guidelines suggest statins at age 40 in all comers, not just those with a certain LDL cutpoint; and given that metabolic syndrome is also associated with high cardiovascular risk and the more atherogenic LDLs, it seems to me that this group should be targeted as well. For a deeper analysis of the role of prediabetes (actually including A1c levels lower that 5.7%) and the associated increased cardiovascular risk, see https://gmodestmedblogs.blogspot.com/2024/03/colchicine-decreases-cardiovasc-events.html ).

--but in terms of Lp(a) and what the benefit is of decreasing its level, we clearly need actual real data of people enrolled in RCTs (which seems to be happening, though should include an array of patients for primary and secondary prevention). And, if positive, before we can even apply these results to the general population, we need a robust, accurate, reliable, uniform assay to measure Lp(a).  so,not quite ready for prime-time....

geoff

-----------------------------------

If you would like to be on the regular email list for upcoming blogs, please contact me at gmodest@bidmc.harvard.edu

  

to get access to all of the blogs:  go to http://gmodestmedblogs.blogspot.com/ to see the blogs in reverse chronological order

  -- click on 3 parallel lines top left, if you want to see blogs by category, then click on "labels" and choose a category​

  -- or you can just click on the magnifying glass on top right, then type in a name in the search box and get all the blogs with that name in them

if you would like a copy of the article, please let me know

please feel free to circulate this to others. also, if you send me their emails (gmodest@bidmc.harvard.edu), i can add them to the list