high Lp(a) increases risk of recurrent ASCVD

 A recent study found that in patients with atherosclerotic cardiovascular disease (ASCVD), there was a progressive increase in recurrent cardiac events as lipoprotein(a) levels increased: see Lp(a) recurrent ASCVD EuropHeartJ 2025 in dropbox. or doi.org/10.1093/eurheartj/ehaf297)

 

Details:

-- they accessed the Family Heart Database of US medical claims from 340 million adults between 2012-2022, representing at least 1/2 of the national census population, with lipid data available for about 1/3 of them

-- these medical claims included medication use, diagnoses, procedures, and surgeries

    -- 273,770 people had diagnosed ASCVD as well as lipoprotein(a) levels, also referred to as Lp(a), measured

        -- ASCVD was defined as: evidence of an MI, other acute coronary syndromes, percutaneous coronary interventions (PCI), CABG, stable angina, ischemic stroke, other cerebral vascular disease, TIA, peripheral vascular disease, and "general ASCVD"

            -- 21% of these patients were hospitalized with an acute cardiovascular event, and 32% had inpatient reports of other qualifying ASCVD events; 56% had outpatient reports of ASCVD

 

-- cholesterol-lowering med classification:

    -- high impact meds: either PCSK9 monotherapy or in combination with high intensity statin monotherapy, or statin plus ezetimibe or bempedoic acid (that being a new non-statin for LDL lowering: https://www.ncbi.nlm.nih.gov/books/NBK594232/)

    -- low/moderate impact meds: low/moderate intensity statin monotherapy or bempedoic acid and/or ezetimibe

-- median LP(a) levels, per the defined Lp(a) buckets used:

    -- <15 nmol/L: 10 nmol/L

    -- 15-79 nmol/L: 34 nmol/L

    -- 80-179 nmol/L: 128 nmol/L

    -- 180-299 nmol/L: 217 nmol/L

    -- >299 nmol/L: 365 nmol/L

-- median LDL levels, by Lp(a) levels:

    -- Lp(a) <15 nmol/L: 78 mg/dL

    -- Lp(a) 15-79 nmol/L: 81 mg/dL

    -- Lp(a) 80-179 nmol/L: 83 mg/dL

    -- Lp(a) 180-299 nmol/L: 83 mg/dL

    -- Lp(a) >299 nmol/L:  88 mg/dL

        -- ie, the LDL did increase somewhat along with the Lp(a) levels

-- median triglyceride levels: 110 mg/dL for all of the Lp(a) levels

  

-- demographics in the study, as stratified by lipoprotein(a) levels, had some internal variability: age was consistently 64, 38 to 54% were women (higher percent women in those with higher Lp(a) levels), race/ethnicity, 5 to 15% were Black individuals (higher percentage Black individuals in those with higher Lp(a) levels)/Hispanic 8%/white 60%

-- baseline ASCVD event: coronary 70%/cerebrovascular 26%/peripheral vascular 20%

-- Charlson comorbidity index: zero in 45%, 1-2 in 34%, at least 3 in 21% (the higher, the more comorbidities)

-- risk factors: hypertension 75% (53% treated with meds), diabetes 33% (no comment on A1c levels), family history 1%

-- baseline LDL meds:

    -- high impact: 29-44% (higher percent in those with higher Lp(a) levels), 2% on PCSK9s

    -- low/moderate impact: 28-22% (lower percent in those with higher Lp(a) levels)

    -- none in 43-34% (lower percent in those with higher Lp(a) levels)

 

-- Principal outcomes of the study: do people who have had a prior ASCVD clinical event (who are already at higher absolute risk of another one) have recurrent ASCVD events stratified by their baseline Lp(a) level?

-- other outcomes:

    -- does lowering the LDL level prevent further ASCVD even in those with high Lp(a) levels?

    -- does more aggressive lowering of LDL level seem to help lower the increased ASCVD risk in those with high Lp(a) levels

-- subgroup analyses were done for sex, race/ethnicity, baseline ASCVD risk, baseline diabetes, LDL therapy used

 

-- median followup: 5.4 years (30 days to 10.6 yrs)

 

Results:

-- Distribution of Lp(a) values: note the large skew to lower Lp(a) levels in the population, as well as the finding that women and Black individuals had a somewhat higher prevalence of higher LP(a) levels; note that having high Lp(a) levels is pretty uncommon overall, as opposed to the quite common high LDL levels:

 

-- Compared to individuals with lipoprotein(a) <15 nmol/L, the adjusted hazard ratios for recurrent ASCVD, as adjusted for sex, race/ethnicity, age, presence of diabetes, and impact of LDL-lowering therapy events (all were statistically significant):

    -- Lp(a) 15–79 nmol/L: 4% more, aHR 1.04 (1.01–1.07)

    -- Lp(a) 80–179 nmol/L: 15% more, aHR 1.15 (1.12–1.19)

    -- Lp(a) 180–299 nmol/L; 29% more, aHR 1.29 (1.25–1.33)

    -- Lp(a) ≥300 nmol/L: 45% more, aHR 1.45 (1.39–1.51)

-- full adjustment of all covariates did not change the ASCVD outcomes significantly over the unadjusted comparison, suggesting that these various potential confounders added little to the overall relationship between Lp(a) and ASCVD events  

 

 

--comparing outcomes by the patient’s LDL therapy:

    -- note that there was a significant increase in recurrent ASCVD events as Lp(a) levels increased

    -- also, the increased ASCVD events were much less prominent in those on more aggressive LDL therapy (especially if Lp(a) level was at least 180), and the use of the PCSK9 meds was associated with the same low frequency of ASCVD results independent of the Lp(a) level

 

 

Commentary:

-- studies over the past 4-5 decades have consistently shown that higher lipoprotein(a), ie Lp(a), levels are associated in a continuous way with increased atherosclerotic cardiovascular disease events (ASCVD) in otherwise healthy individuals, regardless of sex or race/ethnicity

    -- Lp(a) is a structural combination of an LDL-like particle which has apolipoprotein B (as does LDL) along with pretzel-like "krinkles", which has sheaths of plasminogen-like components, and functionally acts as a bad combination of being atherogenic from the LDL-like component and pro-thrombotic by the krinkle component (the number of krinkles is genetically-determined, and they act as plasminogen antagonists: https://www.ahajournals.org/doi/10.1161/01.atv.10.2.240 )

    -- apolipoproteins are on the surface of lipid moieties and provide structural support as well as binding  to receptors (ie apoB binds to LDL receptors)

-- Lp(a) levels are >90% genetically determined and are largely stable throughout life; they therefore are unlikely to have lots of variability by lifestyle or socioeconomic conditions

-- a recent study using Mendelian randomization found that Lp(a) was more atherogenic than LDL: https://gmodestmedblogs.blogspot.com/2024/03/lipoprotein-a-bad-actor.html

-- the concern is that there are no approved medications to treat high Lp(a) levels

    -- however, it is known that PCSK9 inhibitors (found to lower Lp(a) levels by 25-30%), niacin, and estrogen therapy are all associated with lower Lp(a) levels, though no clear evidence that these decrease clinical ASCVD events

        -- of these 3 meds that lower Lp(a) levels, PCSK9s are the best choice, given the potential adverse effects of niacin and estrogens, and PCSK9s have much more profound ability to lower ASCVD risk vs niacin and estrogens

    --  there are a few drugs in development to directly address high Lp(a) levels, assessing their clinical benefits from reducing Lp(a) levels

    -- as noted in Lp(a) AHA statement AtheroThrombVasc Biology2022 in dropbox, or https://www.ahajournals.org/doi/epub/10.1161/ATV.0000000000000147, an American Heart Association Scientific Statement on Lp(a) recommended to treat high Lp(a) levels by being more aggressive in treating high LDL levels. several European guidelines suggest the same [the argument being that we do know that lowering LDL levels decreases ASCVD events and that the lower, the better; so more aggressive LDL lowering should at least partially compensate for at least some of the adverse effect of high Lp(a) levels]

--this current study found that Black individuals tend to have higher Lp(a) levels. This has been found in other studies as well. I am not sure how to explain this, given that there are lots of genetic mixtures in people of all "races", and prior studies have found that much of the differences by race have been explained by social conditions;  we have evolved into an understanding that race is fundamentally a social construct. Lp(a) levels are determined by genetics >90% of the time. Of course, that does leave 10% related to other issues and perhaps this small variability in Lp(a) levels is explained by the differences found by social conditions

    -- that being said, potential confounders such as diabetes and hypertension were not found to contribute the recurrent ASCVD events in this study, as would be expected since Lp(a) is largely genetically determined, though there were no granular data in the study about the specifics of the diabetes (meds, A1c levels, diet and exercise) or of the hypertension (level of BP, treatments)

-- women have also been found to have higher Lp(a) levels. this is likely because of their decreased estrogen with menopause, and estrogens do lower Lp(a) levels

-- this study found a few things:

    -- in the large number of people accessed in this database, there was an increasing risk of recurrent ASCVD events as the Lp(a) level increased

    -- this increase was independent of sex, race/ethnicity, or baseline diabetes

    -- high impact LDL-lowering meds seemed to significantly decrease the high ASCVD rate associated with the high Lp(a) levels

    -- adding PCSK9 inhibitors seemed to nullify this effect (we know that PCSK9s decrease Lp(a) levels; this study found that they were associated with decreased clinical events)

        -- but, this is a bit hard to interpret since part of this PCSK9 benefit may from also lowering the LDL levels more and not only attributable to lowering Lp(a) levels.

-- and, an added issue to the complexity of the above is the role of apolipoprotein B levels:

    -- several studies have found that apo B levels correlate with cardiovascular outcomes more than closely than LDL levels

    -- one clear reason is that there is one apo B on each LDL particles, so for a given LDL level, there are more apo Bs if the LDLs are the small/dense ones that are 3-fold as atherogenic as the big/fluffy ones

    -- but, there is an added issue of Lp(a)s, since they have apo B as well: the total of apo B levels is a combo of those on LDL and Lp(a): https://www.amjmed.com/article/S0002-9343(24)00127-X/fulltext

    -- several countries (not including the US) promote using apo B measurements over LDL ones: this led to the European Society of Cardiology in 2019 to recommend testing for apo B levels as opposed to LDLs: https://europepmc.org/article/med/33870931 and https://pubmed.ncbi.nlm.nih.gov/33870931/; canadian guidelines:  https://www.cmaj.ca/content/195/33/E11 , and a more recent review: https://academic.oup.com/eurheartj/article/45/27/2410/7663778 and doi: 10.1097/MOL.0000000000000754

-- so, the issue of increased Lp(a) is really enmeshed in the issue of apo B

    -- and, to me, it does make sense to measure both: the Lp(a) could probably be measured once since it iseems to be stable over one's lifetime, and apoB levels are much better than LDL levels in predicting atherosclerotic risk......

    -- and it seems that using meds to lower Lp(a) and apo B (eg, higher doses of statins, etc) does seem to help

Limitations:

-- this was a retrospective data-mining study based on claims data, which does not include all of the clinical information that a planned study would have

    -- for example, the categorization of “diabetes” in this study is basically a yes-no binary formulation. The yes-no part is problematic because an A1c of 6.4% is not really significantly different from 6.5%, though patients would be put into counterposing buckets for analysis; and diabetes with an A1c of 6.5% vs 13.5% might well have a different outcome since ASCVD risk does increase with higher A1c levels. so, if the patients with diabetes were largely well-controlled, generalization of the results from this study to those less well controlled may be problematic (ie, we cannot definitively conclude that diabetes did not affect outcomes, as found in the above study). Also, we only know the baseline prevalence of diabetes and no followup assessments

    -- as a retrospective trial, we do not know the thought process of the clinicians in terms of how they chose LDL-lowering therapy (from none and up to the high impact therapies). For example, this study found that the absolute ASCVD risk was lowest in those on no LDL therapy and highest in those on high impact therapy, which sounds like there was a large selection bias in terms of which therapy was prescribed by the clinicians (and, in this study, it was clear that the measured LDL levels increased in those with higher Lp(a) levels). and, why did some people have Lp(a) levels checked in the first place? also a likely sampling bias (bias by indication) that distorts the validity of the conclusions to the general population

    --it is rather shocking that about 40% of these individuals, all with history of ASCVD, were on no LDL-lowering meds taken. And, prescribing statins should be pretty automatic, independent of the LDL level, since these patients are predisposed to a recurrent ASCVD event, such as from having the small dense LDLs which are 3-fold as atherogenic (and, quite common in those with diabetes and metabolic syndrome) which could have been assessed better by measuring apolipoprotein B levels as well as the Lp(a) levels

-- this study was a secondary prevention study (ie people with documented ASCVD). it is quite likely that individuals at only high risk for ASCVD events (primary prevention) would also benefit from Lp(a) reduction, but this needs to be studied

-- this study and its database do not have accurate information about mortality, just the ASCVD outcomes noted above. And a more thorough evaluation would require this mortality effect in order to find a clear benefit of Lp(a) lowering

so,

--a limited study since this study was observational, but it does seem to point to some pretty obvious conclusions, adding to the overall Lp(a) literature:

    -- many studies have found a relationship between Lp(a) levels and ASCVD

        -- which is really expected, since Lp(a) is the combination of 2 bad cardiovascular particles: a pro-atherogenic LDL-like entity along with a pro-thrombotic entity (the krinkles that mimic but interfere with plasminogen). a double hit

    -- and we know that PCSK9 meds decrease Lp(a) levels

    -- and this study strongly suggested that PCSK9 inhibitors seem to nullify the adverse cardiovascular effects of even very high Lp(a) levels

-- which all leads to the question "what is a clinician to do?". Not clear from this study, but it does seem reasonable to:

    -- check Lp(a) levels in patients at high cardiovascular risk or those with already evident heart disease

        -- this approach has the benefit of potentially avoiding the high mortality and morbidity of that first ASCVD event (a 2023 study found that an acute MI, one component of ASCVD, is deadly in up to 30% of patients (more if more comorbidities) before reaching the hospital and an additional 5-12% do not survive within the first year post-MI https://www.ncbi.nlm.nih.gov/books/NBK537076/. And ASCVD survivors have a much higher absolute recurrent ASCVD risk than the ASCVD risk in people who never had a prior event

    -- consider checking Lp(a) levels routinely on all patients, given the documented deleterious effects associated with them (though zero data on this suggestion)

        -- the plus here is that Lp(a) levels do not change much over one's lifetime, so a single test may well be adequate

        -- and, if high, discuss the importance of diet/exercises in lowering the LDL level, with meds as needed

    -- consider strongly the approach of very aggressive LDL lowering in those with high Lp(a) levels, especially if above 180 nmol/L (though i would personally do so even if lower, since the long-term risk of elevated Lp(a) levels <180 would likely lead to slower but still increased atherosclerosis)

        -- this approach has been recommended by the AHA but also in European countries

    -- it certainly sounds more reasonable to measure apolipoprotein B levels than LDL levels (as done in much of the industrialized world), which provides a more accurate marker of atherosclerotic risk (and this actually includes the amount of apoB in Lp(a) particles)

    -- in the absence of formal studies on the benefit of lowering Lp(a) levels, i would personally try to get insurance company approval for prescribing PCSK9 inhibitors in people at extremely high Lp(a) levels, or those who are at high ASCVD risk but have inadequate LDL response to high impact LDL-lowering meds in combination

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

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