Biomarkers · Clinical Literacy

    Your Lp(a) Is High. Here's What Still Works.

    You Can't Change the Multiplier. You Can Change What It Multiplies.

    Reviewed by Dr. Sanjeev GoelNovember 13, 202612 min read
    A row of mechanical slider controls on a dark surface, rim-lit in sage green, with one slider locked in place and the others adjustable

    A fixed, inherited number sits on top of your cardiovascular risk and multiplies it. You cannot change that number yet. What you can change is everything it multiplies — and the evidence says that is where most of your risk still lives.

    The 60-Second Answer

    Lipoprotein(a) is genetically set, stable through life, and — as of November 2026 — has no approved treatment. In Lp(a)HORIZON, lowering it by 78% in 8,350 patients changed nothing: 12.3% of events on drug, 12.1% on placebo, hazard ratio 1.02.[1] That is not a reason to ignore a high Lp(a). It is a reason to treat everything around it harder. Total cardiovascular risk is roughly a product of several terms, and the others — apoB-containing cholesterol, blood pressure, smoking, glucose, fitness — are all modifiable, with large randomised trials sizing exactly how much each is worth. In the FOURIER trial, 40 patients with high Lp(a) had to be treated with a PCSK9 inhibitor for three years to prevent one cardiovascular event, against 105 patients with lower Lp(a).[3] None of this lowers your Lp(a). All of it lowers your risk.

    Why a fixed number still leaves most of your risk modifiable

    Lipoprotein(a) is an LDL-like particle with an extra protein, apolipoprotein(a), wrapped around it. How much of it you make is set almost entirely by the LPA gene. It is stable from childhood, barely moves with diet or exercise, and about one in five Canadians carries a level high enough to matter.[9] The 2021 Canadian Cardiovascular Society guidelines recommend measuring it once in a lifetime as part of initial lipid screening — a strong recommendation on high-quality evidence — precisely because it does not change, so there is nothing to track.[8]

    For twenty years the reasonable inference was: if a high Lp(a) causes disease, lowering it should prevent disease. On 7 November 2026 that inference was tested properly and failed. Pelacarsen, an antisense drug, cut Lp(a) from a median of 236 nmol/L to 64.4 nmol/L — a 78% reduction — in 8,350 people with established cardiovascular disease, over a median 4.7 years. A primary event occurred in 511 patients (12.3%) on the drug and 503 (12.1%) on placebo: hazard ratio 1.02, 95% CI 0.90 to 1.15, P = 0.76. In the sickest subgroup, those starting above 90 mg/dL, it was 12.3% in both arms.[1] Every re-cut of the data — on-treatment, uninterrupted treatment, total recurrent events, people who achieved the lowest Lp(a) — gave the same answer.[2] We covered those results in detail here, and the pre-read here.

    So the honest position is this. Your Lp(a) remains a real, independently measured risk marker — the Canadian framework published in January 2026 is unambiguous that it is a genetically determined, independent risk factor for coronary disease, peripheral arterial disease, stroke and calcific aortic stenosis.[9] What has not been demonstrated is that moving the number moves the outcome. Those are different claims, and a null trial of one does not retire the other.

    Which leaves a practical question that the trial does not answer and this article does: if the multiplier is fixed, what is left?

    The clearest number we have: who benefits most from lowering apoB

    The most actionable result on this topic is nearly seven years old and comes from a trial that was not about Lp(a) at all.

    FOURIER randomised 27,564 patients with established atherosclerotic disease to evolocumab, a PCSK9 inhibitor, or placebo. Lp(a) was measured in 25,096 of them; the median was 37 nmol/L. Two things came out. First, in the placebo arm, patients in the highest Lp(a) quartile had a higher risk of coronary death, myocardial infarction or urgent revascularisation than those in the lowest — adjusted hazard ratio 1.22, 95% CI 1.01 to 1.48 — independently of their LDL cholesterol. Second, and more useful: evolocumab cut that endpoint by 23% (HR 0.77, 0.67–0.88) in patients whose baseline Lp(a) was above the median, and by 7% (HR 0.93, 0.80–1.08) in those at or below it.[3]

    Because the high-Lp(a) group started at higher risk, the absolute arithmetic is stark. Over three years the absolute risk reduction was 2.49% versus 0.95%, and the number needed to treat was 40 versus 105.

    Figure 1 · The same drug, two very different returns

    Number needed to treat over 3 years

    Absolute risk reduction over 3 years (%)

    Read down each panel, not across. Evolocumab versus placebo in 25,096 FOURIER patients with established atherosclerotic disease, split at the median baseline Lp(a). A lower number needed to treat means more benefit. The p-value for the interaction between baseline Lp(a) and treatment effect was 0.07 — not conventionally significant — so this is a tendency the trial observed, not a proven difference in drug efficacy.

    Source: O'Donoghue et al., Circulation 2019;139(12):1483–1492.

    One caveat has to travel with that figure, and it is the authors' own: the p-value for the interaction was 0.07, which is not conventionally significant. The claim the data support is the one O'Donoghue and colleagues made — that patients with higher baseline Lp(a) "tended to derive greater coronary benefit" — not that PCSK9 inhibitors are more effective if your Lp(a) is high. But a near-threefold difference in the number needed to treat is the strongest practical signal available on this question, and it points one way.

    The honest tension: two trials that disagree

    Here the evidence genuinely conflicts, and the conflict is worth printing rather than smoothing over.

    ODYSSEY OUTCOMES tested alirocumab, another PCSK9 inhibitor, in 18,924 patients already on high-intensity statins after an acute coronary syndrome, with a median follow-up of 2.8 years. A pre-specified analysis asked whether the drug's effect on Lp(a) and on LDL cholesterol each independently predicted events. Alirocumab lowered Lp(a) by a median of only 5.0 mg/dL and corrected LDL-C by 51.1 mg/dL, and reduced major adverse cardiovascular events (HR 0.85, 0.78–0.93). After adjustment for baseline levels of both lipoproteins and for demographic and clinical characteristics, both reductions independently predicted lower risk: each 1 mg/dL fall in Lp(a) carried a hazard ratio of 0.994 (0.990–0.999, P = 0.0081). The authors concluded that Lp(a) "should be an independent treatment target after ACS."[4]

    In 2020 that was a reasonable conclusion. In 2026 a randomised trial designed to test exactly that proposition — lower Lp(a), see whether events fall — reduced it by 78% and found nothing.[1]

    Both results are real. They differ in design, and the difference is the whole lesson. ODYSSEY's Lp(a) finding is an observational analysis conducted inside a randomised trial: patients were not randomised to a particular degree of Lp(a) lowering, so whatever else predicts who responds well to alirocumab is baked into the association. Lp(a)HORIZON randomised the thing itself. When an observational analysis and a randomised comparison of the same question disagree, the randomised comparison is the stronger design. That does not make the ODYSSEY analysis worthless — it generated a hypothesis precise enough to be tested, which is what good observational work does — but it does not survive the test.

    How much is lowering LDL cholesterol actually worth?

    This is the part that is not in doubt, and it is where the numbers are largest.

    The Cholesterol Treatment Trialists' Collaboration pooled individual participant data from 26 randomised trials: 129,526 people in 21 statin-versus-control trials and 39,612 in five trials of more-versus-less-intensive treatment. Per 1.0 mmol/L reduction in LDL cholesterol, major vascular events fell by 22% (rate ratio 0.78, 95% CI 0.76–0.80) and all-cause mortality by 10% (RR 0.90, 0.87–0.93). More-intensive regimens produced a further 15% reduction (11–18) for a mean further LDL-C fall of 0.51 mmol/L. Crucially, no threshold was found anywhere in the range studied, including in people whose LDL-C was already below 2 mmol/L — and the authors estimated that a 2–3 mmol/L reduction would cut risk by about 40–50%.[6]

    Note the framing. It is not "statins cut risk by 22%." It is 22% per mmol/L, proportionally, with no floor yet found. That proportionality is what makes it a lever rather than a yes/no decision.

    A note on which number to track. The 2021 CCS guidelines are explicit that randomised trials have not identified a clear LDL-C, non-HDL-C or apoB target, and set thresholds for intensifying therapy instead: in secondary prevention, on a maximally tolerated statin, LDL-C ≥ 1.8 mmol/L, non-HDL-C ≥ 2.4 mmol/L, or apoB ≥ 0.7 g/L. For anyone with triglycerides above 1.5 mmol/L, the guideline recommends non-HDL-C or apoB instead of LDL-C as the preferred screening parameter — a strong recommendation on high-quality evidence. ApoB is an insured laboratory test in every province except Ontario.[8] You can read more about the individual markers in our cardiovascular biomarker database.

    None of this lowers Lp(a). All of it lowers the number of atherogenic particles that the Lp(a) multiplier is acting on.

    The counterintuitive part: statins raise Lp(a)

    This is the fact most likely to be misread, so read the whole paragraph.

    A subject-level meta-analysis of 5,256 patients from six randomised trials — 1,371 on placebo, 3,885 on statin, all measured with the same validated assay — found that statins raise Lp(a). The ratio of geometric means for statin versus placebo was 1.11 (95% CI 1.07–1.14, P < 0.0001). Mean percent change from baseline ran from +8.5% to +19.6% on statin against −0.4% to −2.3% on placebo. Atorvastatin raised it more than pravastatin (ratio 1.09, 1.05–1.14), and HepG2 hepatocytes incubated with atorvastatin showed increased LPA mRNA and apolipoprotein(a) protein.[5]

    What this is not is a reason to stop or avoid a statin. Every outcome trial that established statin benefit — including all 26 in the CTT analysis — was run in populations whose Lp(a) was doing this the entire time, and the benefit was measured anyway. The drug is lowering the thing that carries most of the risk while nudging up a number that, on the best current evidence, does not respond to being lowered. A rise in Lp(a) after starting a statin is expected, is not treatment failure, and is not a signal to change therapy. Nobody should start, stop or alter any lipid-lowering medication on the basis of this article. That is a conversation with your own physician, who can see the rest of your chart.

    Blood pressure: the second large, measurable lever

    The Blood Pressure Lowering Treatment Trialists' Collaboration pooled individual participant data from 51 randomised trials and 358,707 participants, excluding anyone with prior heart failure. Per 5 mm Hg reduction in systolic blood pressure, the hazard ratio for major cardiovascular events was 0.82 (0.76–0.88) in people under 55, 0.91 (0.88–0.95) at 55–64, 0.91 (0.88–0.95) at 65–74, 0.91 (0.87–0.96) at 75–84, and 0.99 (0.87–1.12) at 85 and older (adjusted P = 0.050).[7]

    Figure 2 · What 5 mm Hg is worth, by age

    Hazard ratios for major cardiovascular events per 5 mm Hg reduction in systolic blood pressure, with 95% confidence intervals, from individual participant data on 358,707 people in 51 randomised trials. Bars left of the dashed line at 1.0 mean fewer events. The oldest band is greyed because only 4,788 participants (1.3% of the total) were 85 or over and its interval crosses 1.0 — no benefit should be claimed there.

    Source: Blood Pressure Lowering Treatment Trialists' Collaboration, Lancet 2021;398(10305):1053–1064.

    Two things make this unusually useful. First, the collaboration found no clinically meaningful heterogeneity by baseline blood pressure in any age group — the proportional benefit held down below 120/70 mm Hg. Second, absolute risk reductions were larger in the older groups, because they started at higher risk. The oldest band is the honest exception: only 4,788 participants (1.3% of the total) were 85 or over, and its confidence interval comfortably includes 1.0, so no benefit should be claimed there.

    If you want context on how much a single blood pressure reading can actually tell you, that is a separate and surprisingly large problem, which we covered in this piece.

    The levers with no trial but a large effect

    Three more things belong on the list, and honesty requires describing them without inventing effect sizes.

    Smoking is the obvious one. No one will ever randomise people to smoke, so there is no trial of the kind quoted above; the evidence is observational and overwhelming, and stopping is almost certainly the single largest modifiable change available to anyone who smokes.

    Glycaemic control and cardiometabolic health matter for the same structural reason — they feed into the same product of risk terms — but the trial evidence for how much a given change buys is more mixed and more condition-specific than the lipid and blood-pressure numbers above, so no figure is offered here.

    Cardiorespiratory fitness is strongly and consistently associated with lower mortality in large cohorts, but it has not been randomised to a hard-outcome endpoint in the way LDL cholesterol and blood pressure have. It belongs on the list; it does not belong in the figure. You can find our broader evidence coverage of these in the longevity section.

    Figure 3 · Three levers, to scale

    A scale illustration, not a meta-analysis. These three figures come from different trials, with different primary endpoints, in different populations, and are not directly comparable — the point is the order of magnitude, not a head-to-head. The LDL-C figure is a rate ratio of 0.78 per 1.0 mmol/L; the blood-pressure figure is a hazard ratio of 0.91 per 5 mm Hg in the 55–84 age bands; the Lp(a) figure is Lp(a)HORIZON's hazard ratio of 1.02, which is a 2% increase and is plotted below zero for that reason.

    Sources: CTT Collaboration, Lancet 2010; BPLTTC, Lancet 2021; Cho et al., N Engl J Med 2026.

    Because Lp(a) is inherited, it is a family finding

    A high Lp(a) is not only information about you. It is genetically determined, so a first-degree relative has roughly a one-in-two chance of carrying the same elevation, and most of them will never have been tested — the January 2026 Canadian framework describes Lp(a) as under-recognised and undermeasured despite affecting an estimated 20% of Canadians, and gives specific attention to the role of cascade screening in families.[9]

    In primary prevention, the CCS guidelines recommend earlier and more intensive health-behaviour-modification counselling for anyone with an Lp(a) at or above 50 mg/dL (or 100 nmol/L) — though note this is explicitly a strong recommendation based on expert consensus, not on trial evidence.[8] Telling your siblings, parents and adult children to have the test once is, on current evidence, probably the highest-yield thing a high result enables.

    What is not worth doing — and what is coming

    Niacin lowers Lp(a) by roughly 20–30% and has been tested against hard outcomes twice, in trials of more than 25,000 and 3,400 people, and failed both times while producing a long list of adverse effects. It is not a treatment for Lp(a).

    Lipoprotein apheresis is the one intervention that substantially removes Lp(a) from the circulation today. It is not a casual option: it is a repeated, usually weekly, extracorporeal procedure. In a multicentre observational study of 113 patients across eight German apheresis centres, cardiovascular events per patient fell from 1.22 ± 1.16 in the two years before treatment to 0.33 ± 0.75 in the two years after, with the largest reductions in those with isolated Lp(a) elevation (−77%, P < 0.001).[10] That is a striking number, and it has to carry a large caveat: this is an uncontrolled before-and-after design in patients selected for having just had a run of events, which is precisely the situation in which regression to the mean produces impressive-looking improvements. The authors say so themselves, calling for prospective trials. In Canada access is a further problem — implementation barriers have been described in the Canadian literature for more than a decade.[11]

    What is actually coming is more trials. Pelacarsen was the first of several Lp(a)-lowering drugs; olpasiran, lepodisiran and muvalaplin all have large phase 3 outcome trials running, reading out between 2028 and 2031. Lp(a)HORIZON tells us that one drug, at one dose, in one population of people with established disease, did not help. It does not tell us that a larger absolute reduction, started earlier, in people without established disease, would not.

    The practical bottom line

    • A high Lp(a) is fixed, inherited and currently untreatable. Measure it once; there is nothing to track.
    • The first drug to lower it substantially — by 78%, in 8,350 people — did not reduce events.
    • That shifts the job from treating the number to treating everything it multiplies.
    • Lowering LDL cholesterol has the largest, best-quantified effect: 22% fewer major vascular events per 1.0 mmol/L, with no threshold found.
    • In FOURIER, the high-Lp(a) group needed 40 patients treated for three years to prevent one event, against 105 — though the interaction p-value was 0.07.
    • Lowering systolic blood pressure by 5 mm Hg cuts major cardiovascular events by about 9% in most adult age groups, regardless of starting level.
    • Expect your Lp(a) to rise slightly on a statin. That is pharmacology, not failure, and not a reason to stop.
    • Tell your first-degree relatives to get tested once.
    • None of this is medical advice, and nothing here is a reason to start, stop or change a medication on your own.

    Frequently asked questions

    Can anything actually lower my Lp(a)?

    PCSK9 inhibitors lower it by roughly a quarter — in FOURIER, evolocumab reduced Lp(a) by a median of 26.9% at 48 weeks. Lipoprotein apheresis lowers it a great deal more, but it is a repeated extracorporeal procedure with observational evidence only and limited availability in Canada. Niacin lowers it and failed two large outcome trials. The purpose-built drugs that lower it by 70–80% are still in trials, and the first one to report did not reduce events. Diet and exercise essentially do not move it.

    If the drug failed, does my Lp(a) result still matter?

    Yes, but as information rather than as a target. A high Lp(a) still identifies you as being at higher cardiovascular risk independently of your LDL cholesterol, and it is still a reason to treat the modifiable risks more aggressively and earlier, and to have your family tested. What has not been shown is that lowering the number itself helps.

    Should I ask for a PCSK9 inhibitor because my Lp(a) is high?

    That is a decision for you and your physician, and it depends on your overall risk, what you are already taking, and funding and eligibility criteria that vary by province and insurer. What the evidence shows is that in a high-risk secondary-prevention population, the patients with higher Lp(a) tended to get more absolute benefit from a PCSK9 inhibitor. It does not define an Lp(a) threshold above which anyone "should" be on one.

    My Lp(a) went up after I started a statin. Is the statin making things worse?

    A rise is expected. Across six randomised trials, statins raised Lp(a) by roughly 8% to 20% on average. The statin is lowering the apoB-containing particles that carry most of the risk, which is the effect measured in every outcome trial — and those trials were run in people whose Lp(a) was rising the same way. This is not a reason to stop or change treatment on your own.

    How often should I retest my Lp(a)?

    Once, in most cases. Canadian guidelines recommend measuring it once in a lifetime as part of initial lipid screening, because levels are genetically determined and stable. There is no established reason to repeat it unless a borderline result needs confirming, or a future treatment makes tracking it meaningful.

    Should my children be tested?

    It is worth discussing with their physician. Lp(a) is inherited, so first-degree relatives of someone with a high level are substantially more likely to have one themselves, and cascade screening in families is specifically highlighted in the 2026 Canadian guidance. Most people with a high Lp(a) have never been tested and do not know.

    References

    1. [1]Cho L, Nicholls SJ, Frndak S, et al. Pelacarsen and Cardiovascular Outcomes in Patients with High Lipoprotein(a). N Engl J Med. 2026 Nov 7. doi:10.1056/NEJMoa2610159
    2. [2]Nissen SE, Cho L, Frndak S, et al. Extended Analyses of Pelacarsen and Cardiovascular Outcomes in Patients with High Lipoprotein(a). NEJM Evid. 2026 Nov 7. doi:10.1056/EVIDoa2600373
    3. [3]O'Donoghue ML, Fazio S, Giugliano RP, et al. Lipoprotein(a), PCSK9 Inhibition, and Cardiovascular Risk. Circulation. 2019;139(12):1483–1492. doi:10.1161/CIRCULATIONAHA.118.037184
    4. [4]Bittner VA, Szarek M, Aylward PE, et al. Effect of Alirocumab on Lipoprotein(a) and Cardiovascular Risk After Acute Coronary Syndrome. J Am Coll Cardiol. 2020;75(2):133–144. doi:10.1016/j.jacc.2019.10.057
    5. [5]Tsimikas S, Gordts PLSM, Nora C, Yeang C, Witztum JL. Statin therapy increases lipoprotein(a) levels. Eur Heart J. 2020;41(24):2275–2284. doi:10.1093/eurheartj/ehz310
    6. [6]Cholesterol Treatment Trialists' (CTT) Collaboration. Efficacy and safety of more intensive lowering of LDL cholesterol: a meta-analysis of data from 170,000 participants in 26 randomised trials. Lancet. 2010;376(9753):1670–1681. doi:10.1016/S0140-6736(10)61350-5
    7. [7]Blood Pressure Lowering Treatment Trialists' Collaboration. Age-stratified and blood-pressure-stratified effects of blood-pressure-lowering pharmacotherapy for the prevention of cardiovascular disease and death: an individual participant-level data meta-analysis. Lancet. 2021;398(10305):1053–1064. doi:10.1016/S0140-6736(21)01921-8
    8. [8]Pearson GJ, Thanassoulis G, Anderson TJ, et al. 2021 Canadian Cardiovascular Society Guidelines for the Management of Dyslipidemia for the Prevention of Cardiovascular Disease in Adults. Can J Cardiol. 2021;37(8):1129–1150. doi:10.1016/j.cjca.2021.03.016
    9. [9]Thanassoulis G, Anand S, Arsenault BJ, et al. Clinical Care for Patients With Elevated Lipoprotein(a) Level in Canada: A Practical Framework for Screening, Risk Assessment, and Management. Can J Cardiol. 2026;42(4):867–884. doi:10.1016/j.cjca.2025.12.060
    10. [10]Schatz U, Tselmin S, Müller G, et al. Most significant reduction of cardiovascular events in patients undergoing lipoproteinapheresis due to raised Lp(a) levels — a multicenter observational study. Atheroscler Suppl. 2017;30:246–252. doi:10.1016/j.atherosclerosissup.2017.05.047
    11. [11]Filler G, Lee M, Hegele RA. Barriers to the Implementation of Lipoprotein Apheresis in Canada. Can J Cardiol. 2017;33(3):409–411. doi:10.1016/j.cjca.2017.01.008
    Disclaimer: This article is for educational purposes only and is not medical advice. It does not establish a physician–patient relationship and is not a substitute for individual assessment. Do not start, stop or change any medication — including statins, PCSK9 inhibitors or blood-pressure medication — on the basis of this article. Discuss your Lp(a) result and any treatment with your own physician.