Cholesterol, Lipids & ApoB

Think of the bloodstream as a delivery system. Cholesterol itself (a waxy molecule the liver manufactures and the body genuinely needs for cell membranes, hormones, and bile) cannot dissolve in blood. It needs a carrier. Low-density lipoprotein (LDL) is the carrier that deposits cholesterol in tissues, including arterial walls. High-density lipoprotein (HDL) is the carrier that runs the return route, picking up cholesterol from peripheral tissues and taking it back to the liver for processing. Triglycerides are not cholesterol at all; they are fat molecules stored in adipose tissue and released into the bloodstream after meals or during metabolic stress.

The "bad and good cholesterol" shorthand that has dominated patient education for three decades is useful but limited. LDL is not bad because it is malevolent; it is dangerous when it is elevated and when it accumulates in the arterial intima, triggering the inflammatory cascade that builds plaque. HDL is not simply good; at extremely high levels (above roughly 90 mg/dL in men), HDL loses its protective correlation and may reflect a dysfunctional HDL particle rather than a healthy one (Annualized risk data, AHA 2021, DOI: 10.1161/CIR.0000000000000950). Triglycerides above 150 mg/dL begin to predict residual cardiovascular risk beyond LDL, particularly when accompanied by low HDL and elevated fasting glucose, the constellation we call metabolic syndrome.

What most patients do not know is that the standard lipid panel reports concentrations, not particle counts. This distinction matters enormously, and the following questions address it directly.

A 62-year-old professor came into my clinic last year with an LDL of 108 and a note from his previous cardiologist that said "cholesterol well-controlled." His ApoB was 162. Those two numbers do not contradict each other; they are measuring different things. LDL-cholesterol measures the mass of cholesterol carried inside LDL particles. ApoB counts the particles themselves. You can have a low amount of cholesterol per particle and still have a very large number of particles. More particles mean more chances for one of them to lodge in an arterial wall and begin the inflammatory process that becomes plaque.

The analogy I find most useful: imagine traffic risk on a highway. LDL-cholesterol tells you how many cars are on the road by weight. ApoB tells you how many cars are on the road by number. A highway carrying 10,000 electric Mini Coopers has lower cargo weight than a highway carrying 1,000 semi-trucks, but it has far more collision potential. ApoB is the collision counter.

The data supporting ApoB as a superior predictor of cardiovascular events over LDL-C is now substantial. Sniderman and colleagues reviewed the evidence comprehensively and found that ApoB consistently outperforms LDL-C when the two are discordant, which occurs in roughly 20-30% of patients (Sniderman AD et al, JACC 2019, DOI: 10.1016/j.jacc.2019.03.529). The 2019 European Society of Cardiology guidelines made ApoB a recommended secondary target for the first time; the 2022 Canadian Cardiovascular Society guidelines went further and stated ApoB should be used as the primary treatment target. The 2018 ACC/AHA guidelines in the United States acknowledge ApoB but still position LDL-C as the primary metric, which is a methodological conservatism I find increasingly hard to defend clinically.

I have had patients with LDL of 95 and ApoB of 140. I have had patients with LDL of 155 and ApoB of 101. Standard treatment algorithms would treat the second patient and not the first. From a particle standpoint, both warrant a serious conversation about risk.

The discordance between LDL-cholesterol and ApoB is most pronounced in three populations: people with high triglycerides (who tend to have smaller, more numerous LDL particles that carry less cholesterol per particle), people with type 2 diabetes or insulin resistance, and people who are lean but metabolically unfavorable (the pattern sometimes called metabolic obesity with normal weight). If you belong to any of these groups and your physician has never mentioned ApoB, the question worth asking specifically is: "What is my ApoB, and is it concordant with my LDL?"

The test itself is inexpensive (typically $15-40 when ordered as a standalone) and available through standard laboratory panels. It is not exotic or experimental. It is a protein measurement by immunoassay that has been available for decades. The reason it is not routinely ordered is inertia, not evidence.

A reasonable threshold for clinical concern: ApoB above 100 mg/dL in a primary prevention patient, and above 80 mg/dL in a patient with established cardiovascular disease or diabetes. These are the targets embedded in the 2022 Canadian guidelines (Anderson TJ et al, Can J Cardiol 2023, DOI: 10.1016/j.cjca.2022.10.010).

Lp(a) is essentially an LDL particle with an extra protein called apolipoprotein(a) attached, making it both atherogenic and prothrombotic: it promotes both plaque buildup and clot formation simultaneously. Tsimikas described the dual mechanism elegantly in the most-cited Lp(a) review in recent years: it drives foam cell formation in plaque and inhibits plasminogen activation, impairing the body's ability to dissolve clots (Tsimikas S, NEJM 2017, DOI: 10.1056/NEJMra1605384).

Your doctor has likely never tested for it because, until the arrival of RNA-based therapies in late-stage trials, nothing short of apheresis (a plasmapheresis-type procedure available at about 30 centers in the US) could lower it meaningfully. Statins do not lower Lp(a). PCSK9 inhibitors lower it roughly 20-30%. Niacin lowers it but at a cost (see Q20). This therapeutic gap created a logical but unfortunate clinical habit: do not test for what you cannot treat. The habit persists even though knowing you have elevated Lp(a) changes your entire risk conversation, your statin target, and your CAC score interpretation.

Elevated Lp(a) is defined differently depending on the assay used. The threshold most cardiologists use is above 50 mg/dL or above 125 nmol/L (these are not directly interchangeable; nmol/L is preferred because it is particle-count-based). Roughly one in five people of any ancestry carries an elevated Lp(a), with higher prevalence in people of African and South Asian descent.

The LPA gene on chromosome 6 governs your Lp(a) level, and it governs it with unusual dominance. Unlike LDL, which responds meaningfully to diet change, statin therapy, exercise, and weight loss, Lp(a) does not budge with lifestyle change. Losing thirty pounds will not move it. A Mediterranean diet will not move it. Running six days a week will not move it. This is both discouraging and, in a strange way, clarifying: if your Lp(a) is elevated, the treatment question is not "what did I do wrong?" but "how aggressively do I need to manage everything else?"

The practical implication: elevated Lp(a) should be treated as a risk multiplier. It does not act alone. It amplifies the damage done by elevated LDL. A patient with Lp(a) of 180 nmol/L and LDL of 60 is meaningfully different from a patient with Lp(a) of 180 nmol/L and LDL of 140. The combination is the clinical emergency. This is why knowing your Lp(a) changes your LDL treatment target: most lipidologists would push the LDL below 70 mg/dL in a patient with elevated Lp(a), and some would push to below 55.

The gene-silencing therapies in development represent the first real opportunity to lower Lp(a) substantially in a pharmacologic way. Pelacarsen, an antisense oligonucleotide, lowered Lp(a) by roughly 80% in phase 2 trials (Tsimikas S et al, NEJM 2020, DOI: 10.1056/NEJMoa1905239). Phase 3 cardiovascular outcomes data are expected in 2025-2026. Olpasiran, a small interfering RNA approach, showed similar results. If outcomes data confirm benefit, Lp(a) testing will shift from niche to standard of care within a decade.

Numbers first. The population distribution of Lp(a) is highly skewed: most people cluster between 5 and 30 mg/dL, but roughly 20% carry levels above 50 mg/dL and about 5% carry levels above 150 mg/dL. The risk relationship is approximately log-linear at lower levels and steepens sharply above 100 mg/dL. A Mendelian randomization analysis by Clarke et al found that each 3.5-fold increase in Lp(a) was associated with a 1.5-fold increase in coronary artery disease risk, and this relationship was independent of LDL (Clarke R et al, NEJM 2009, DOI: 10.1056/NEJMoa0902604).

What does this mean practically? If your Lp(a) is 60 mg/dL, you should know your ApoB, get a CAC score if you have not had one, and discuss whether your LDL target should be set lower than it would be for a patient without elevated Lp(a). If your Lp(a) is 200 nmol/L, the conversation is more urgent: your cardiologist should be treating your LDL aggressively, considering PCSK9 inhibition, and screening first-degree relatives.

One practical note on units: many US labs report Lp(a) in mg/dL using mass-based assays, while most research and guidelines use nmol/L. The two do not convert cleanly because the Lp(a) particle varies in size among individuals. When in doubt, ask your laboratory which assay was used and whether the result is in mass or particle-count units. The nmol/L measurement is more reproducible across labs.

The test goes by several names depending on the ordering platform: NMR LipoProfile (LabCorp, branded as LipoScience), Cardio IQ Advanced Lipid Panel (Quest), or simply "LDL particle number." All are versions of nuclear magnetic resonance spectroscopy applied to a blood sample, which measures the size and concentration of lipoprotein particles in different density ranges.

The clinical value is analogous to ApoB: both measure particle number rather than cholesterol mass, and both identify the patient whose LDL-C is normal but whose particle burden is high. The MESA (Multi-Ethnic Study of Atherosclerosis) investigators found that LDL-P predicted coronary artery disease events significantly better than LDL-C, particularly in patients with metabolic syndrome (Cromwell WC et al, J Clin Lipidol 2007, DOI: 10.1016/j.jacl.2007.02.003). The discordance pattern is the same as with ApoB: high LDL-C with low LDL-P is less dangerous than high LDL-P with low LDL-C.

Why do I usually recommend ApoB over LDL-P for most patients? Cost and availability. ApoB is a direct protein immunoassay that nearly every laboratory runs for $15-40. NMR-based LDL-P requires specialized platforms and costs significantly more. The clinical information is largely redundant. If your insurer covers NMR lipid testing without cost, there is modest additional information in the full panel (particularly small LDL-P), but for the core question of "how many atherogenic particles am I carrying," ApoB answers it more affordably.

The distinction emerged from Krauss et al in the 1980s and 1990s, who used gradient gel electrophoresis to separate LDL particles by size. Pattern B (small dense LDL predominance) was associated with roughly a threefold increase in heart attack risk compared to pattern A in the Physicians' Health Study cohort (Stampfer MJ et al, JAMA 2000, DOI: 10.1001/jama.283.17.2237). The mechanism is straightforward: small dense LDL particles slip more easily through the arterial endothelium, stay in the subintimal space longer because they bind less readily to LDL receptors, and oxidize more readily because their antioxidant content (vitamin E, carotenoids) is lower than large buoyant LDL.

What drives pattern B? The strongest determinants are elevated triglycerides, insulin resistance, and low HDL. Again, the metabolic syndrome cluster. A man with triglycerides of 230, HDL of 38, and LDL of 115 is very likely running pattern B, even though his LDL looks acceptable by standard thresholds. This is why metabolic syndrome and insulin resistance reframe the entire lipid conversation: the LDL number is the same, but the particle character is fundamentally more dangerous.

Practically, pattern B predicts risk best when LDL-C and ApoB are discordant. If your ApoB is proportionate to your LDL-C, the pattern question adds less information. If your ApoB is elevated relative to your LDL-C, the underlying mechanism is almost certainly pattern B with elevated particle number.

A 54-year-old architect in my practice told me he had stopped taking rosuvastatin because he had read that statins cause liver damage. He had been off his statin for eight months. His LDL had climbed from 68 to 142. I spent fifteen minutes with him going through the data.

Here is what that data says. Statins were tested in some of the largest and most rigorous cardiovascular trials ever conducted. The West of Scotland Coronary Prevention Study, JUPITER, CARDS, TNT, PROVE-IT, and the Heart Protection Study all enrolled over 150,000 patients and followed them for up to seven years. The consistent finding: statins reduce major cardiovascular events by 25-35% per mmol/L reduction in LDL-C, with no increase in cancer, no meaningful increase in liver failure, and a small but real increase in new-onset type 2 diabetes in patients who were already predisposed (Collins R et al, Lancet 2016, DOI: 10.1016/S0140-6736(16)31357-5).

The liver concern is based on transaminase elevations seen in early studies at high doses. Clinically significant liver damage from statins is extraordinarily rare. The FDA removed the routine liver monitoring requirement in 2012 precisely because the evidence did not support it. The diabetes signal is real: rosuvastatin in JUPITER increased diabetes rates by roughly 25% relative risk, specifically in patients who already had elevated fasting glucose, (Ridker PM et al, NEJM 2012, DOI: 10.1056/NEJMoa1118567). If you do not have prediabetes, this concern is not relevant to you.

The nocebo effect (the opposite of placebo, where a patient experiences side effects because they expect to) is a real physiological phenomenon, not a dismissal of patient experience. The SAMSON trial (Statistical Approach to Mechanism of Statin Side Effects, Wood E et al, NEJM Evidence 2020, DOI: 10.1056/EVIDoa2000003) gave patients statin pills, identical-looking placebo pills, and empty bottles, then asked them to rate symptoms daily using a smartphone app. When patients took statins versus placebo, symptom scores were elevated by only 8%. When patients took statins versus nothing, symptom scores were elevated by a larger margin, driven almost entirely by the knowledge they were taking a statin rather than any pharmacologic effect.

This is a clinically significant finding, not a rhetorical one. It means that the majority of patients who stop statins because of "intolerable" muscle pain may be experiencing something real but not pharmacologically caused. The expectation of pain, primed by alarming online articles, by friends who said they had bad experiences, by a general cultural suspicion of pharmaceuticals, generates the pain.

I say this not to dismiss any individual patient's experience. Muscle pain is muscle pain, and if it happens to you it is real to you. What I say to patients is that the origin of the pain changes the treatment approach. True statin myopathy (rare, with CK elevation) requires stopping the drug. Nocebo-related symptom pain can often be addressed by switching formulation, switching statin, adjusting dose timing, or trying a rechallenge with a different agent. An n-of-1 approach (blind rechallenge) is now recommended in ACC/AHA guidance for exactly this reason.

The spectrum of statin muscle effects runs from mild, diffuse myalgia (the most common) through myositis (inflammation with enzyme elevation) to rhabdomyolysis (muscle breakdown with CK above 10x normal, myoglobinuria, and potential renal failure). True rhabdomyolysis is rare: in the large statin trials, the rate was roughly 1 in 10,000 patient-years (Armitage J, Lancet 2007, DOI: 10.1016/S0140-6736(07)60526-9). It is more common with high-dose statins and with drug interactions, particularly drugs that inhibit CYP3A4 enzymes: certain antibiotics, antifungals, and some HIV medications can raise simvastatin or lovastatin levels to dangerous concentrations.

If you develop significant muscle symptoms on a statin, the appropriate sequence is: stop the statin, measure a serum CK level within a few days, and document whether symptoms resolve over two to four weeks. If they do resolve completely, the evidence supports a rechallenge with a different statin at a lower dose or with a different dosing frequency (see Q13). Rosuvastatin and pravastatin have the lowest rates of myopathy in head-to-head comparison studies; fluvastatin and pitavastatin are also considered lower-risk. Simvastatin at doses above 40 mg carries the highest myopathy risk.

Risk factors that increase the likelihood of true statin myopathy include: hypothyroidism, kidney disease, age over 70, female sex, lower body mass, and concurrent use of certain interacting medications. If any of these apply, start at the lowest available statin dose and monitor.

The mechanistic argument for CoQ10 is coherent: statins block HMG-CoA reductase, which inhibits not only cholesterol synthesis but also ubiquinone (CoQ10) synthesis, since both share the mevalonate pathway. Skeletal muscle CoQ10 concentrations are measurably lower in statin-treated patients. The hypothesis that this reduction causes mitochondrial dysfunction and therefore muscle pain has intuitive appeal.

The problem is that intuitive mechanism has not translated into clinical benefit in controlled trials. The largest meta-analysis examining CoQ10 for statin myopathy (Qu H et al, J Am Heart Assoc 2018, DOI: 10.1161/JAHA.118.009835) pooled twelve randomized controlled trials and found no statistically significant reduction in muscle pain scores, muscle weakness, or statin discontinuation rates compared to placebo. Individual trials have been underpowered and heterogeneous, but the pattern of results does not support routine use.

My clinical practice: I do not recommend CoQ10 for statin myopathy as first-line management. If a patient is already taking it and reports it helps them, I do not argue. The supplement is safe and the nocebo effect works in both directions. But if a patient asks me whether they should add CoQ10 to tolerate their statin, my honest answer is that the evidence says it probably will not make a pharmacologic difference, and the better approach is to try a different statin or a different dosing strategy.

Rosuvastatin has a plasma half-life of approximately 19 hours, and its pharmacologic effect on hepatic HMG-CoA reductase persists beyond that because statins work inside liver cells rather than depending on continuous plasma concentration. Atorvastatin has a half-life of 14 hours with active metabolites extending efficacy further. Both are reasonable candidates for alternate-day dosing in patients who experience consistent myalgia with daily use.

The data supporting QOD dosing is primarily observational and from small trials rather than large outcomes studies, which is an important limitation. Matalka et al found that rosuvastatin 5 mg QOD reduced LDL by approximately 30% in patients who had discontinued daily statin therapy for intolerance (Matalka MS et al, Pharmacotherapy 2010, DOI: 10.1592/phco.30.6.605). Multiple subsequent studies confirmed that QOD dosing achieves 60-80% of the LDL reduction of daily dosing at much lower rates of muscle symptoms. The muscle symptoms appear to correlate with peak plasma concentration; alternate-day dosing lowers peak levels while maintaining meaningful hepatic efficacy.

The practical protocol I use: for a patient who cannot tolerate rosuvastatin 5 mg daily, I will start rosuvastatin 5 mg every other day, check a lipid panel in six to eight weeks, and add ezetimibe 10 mg daily (which does not cause myopathy) to reach the LDL target. This combination often achieves a 40-50% LDL reduction with minimal symptom burden.

Bempedoic acid works differently from statins in one clinically important way: it is a prodrug that is converted to its active form by the enzyme ACLY only in liver tissue. Skeletal muscle lacks this enzyme, which means the drug cannot cause the mitochondrial dysfunction in muscle that underlies true statin myopathy. In the CLEAR Harmony trial, muscle-related adverse events were not increased compared to placebo (Laufs U et al, N Engl J Med 2019, DOI: 10.1056/NEJMoa1816038).

The CLEAR Outcomes trial, published in 2023, provided the cardiovascular outcomes data the field had been waiting for. In approximately 14,000 patients who were statin-intolerant, bempedoic acid 180 mg reduced major adverse cardiovascular events by 13% compared to placebo (Nissen SE et al, N Engl J Med 2023, DOI: 10.1056/NEJMoa2215539). This is a smaller effect than high-intensity statins but is clinically real and confirmed in a well-powered trial. The result positioned bempedoic acid as a genuine primary option for statin-intolerant patients, not merely a weak add-on.

Practical limitations: gout attacks are more frequent on bempedoic acid (incidence roughly doubled in trials), and the drug may transiently raise uric acid. Patients with a history of gout should use it with caution. The drug also increases the risk of tendon rupture modestly, though this was a rare event in trials.

PCSK9 (proprotein convertase subtilisin/kexin type 9) is a protein the liver secretes that binds to and destroys LDL receptors on liver cell surfaces. Fewer receptors mean less LDL clearance from the bloodstream. Blocking PCSK9 preserves receptors and dramatically accelerates LDL removal. The biology was identified through studies of people with naturally occurring loss-of-function PCSK9 mutations, who have lifelong LDL levels of 40-60 mg/dL and virtually no coronary artery disease.

The FOURIER trial (evolocumab) and ODYSSEY Outcomes trial (alirocumab) both demonstrated that adding a PCSK9 inhibitor to statin therapy in very high-risk patients (those with recent acute coronary syndrome or established atherosclerosis) reduced major cardiovascular events by 15-20% beyond what statins achieved alone (Sabatine MS et al, NEJM 2017, DOI: 10.1056/NEJMoa1615664; Schwartz GG et al, NEJM 2018, DOI: 10.1056/NEJMoa1801174). This is consistent with the "lower is better" hypothesis for LDL: there does not appear to be a floor below which LDL reduction stops providing benefit.

When is it actually needed? I use PCSK9 inhibitors in four situations: patients with established ASCVD (prior MI, stent, or stroke) who remain above LDL 70 despite maximal tolerated statin plus ezetimibe; patients with familial hypercholesterolemia who cannot reach LDL target; patients who are truly statin-intolerant with very high baseline LDL; and patients with elevated Lp(a) and established disease where the risk-burden justifies maximal LDL reduction.

The distinction between inclisiran and evolocumab/alirocumab is worth understanding because patients often conflate them. Evolocumab and alirocumab are proteins (antibodies) that circulate in the blood and bind to the PCSK9 protein before it can damage LDL receptors. They require subcutaneous injection every 2-4 weeks. Inclisiran is a different technology: it is a synthetic strand of RNA that is delivered to liver cells, where it binds to the messenger RNA encoding PCSK9 and triggers its degradation. The liver then cannot produce PCSK9 for roughly six months. One injection, six months of LDL reduction.

The ORION-10 and ORION-11 trials demonstrated that inclisiran 300 mg, given twice a year after an initial loading dose, reduced LDL by approximately 50% sustained over two years, with a side-effect profile not different from placebo except for mild injection-site reactions (Ray KK et al, NEJM 2020, DOI: 10.1056/NEJMoa1912387). The ORION-4 outcomes trial, expected to report around 2025-2026, will provide definitive cardiovascular event reduction data. The mechanism and LDL magnitude are consistent with the antibody PCSK9 inhibitors, giving confidence in outcome benefit.

Clinically, inclisiran is most useful for patients who struggled with the injection frequency of monoclonal antibody PCSK9 inhibitors. Adherence to twice-a-year injections administered in a clinic setting is demonstrably better than self-injection every two weeks.

The dismissal of ezetimibe dates from its early years, when critics pointed out that it did not appear to reduce carotid intima-media thickness in the ENHANCE trial. That intermediate endpoint trial generated significant controversy, and many cardiologists became skeptical. The IMPROVE-IT trial, published in 2015, settled the question. In 18,144 patients after acute coronary syndrome, simvastatin plus ezetimibe versus simvastatin alone reduced the composite cardiovascular endpoint from 34.7% to 32.7% over a median of six years, a modest but statistically significant and clinically real benefit, (Cannon CP et al, NEJM 2015, DOI: 10.1056/NEJMoa1410489). The trial also confirmed the "lower is better" hypothesis: LDL was 53 mg/dL in the combination arm versus 69 mg/dL in the statin-only arm, and the difference in events tracked the difference in LDL.

Ezetimibe's profile makes it particularly useful in combination strategies. It does not cause muscle pain. It does not interact with CYP3A4 metabolism. It can be added to any statin at any dose. It is now generic and costs $4-10 per month. For patients who cannot tolerate high-intensity statin doses, simvastatin or rosuvastatin at moderate doses plus ezetimibe often reaches LDL targets while maintaining tolerability.

One underappreciated feature: ezetimibe works through a different mechanism than statins (blocking intestinal NPC1L1 receptor versus blocking hepatic synthesis), so the combination is genuinely additive rather than redundant.

Plant sterols and stanols are structural analogs of cholesterol found in vegetable oils, nuts, and seeds, and added to functional foods such as certain margarines, yogurts, and orange juices. They compete with dietary and biliary cholesterol for absorption in the small intestine via the NPC1L1 receptor, the same receptor blocked by ezetimibe. Regular consumption of 2 grams per day reduces LDL by approximately 8-14% in a dose-response relationship, a finding confirmed across multiple meta-analyses (AbuMweis SS et al, J Am Diet Assoc 2008, DOI: 10.1016/j.jada.2007.10.015).

The practical question is whether this 8-14% matters clinically. For a low-to-intermediate risk patient whose LDL is mildly elevated and who is not yet at the threshold for statin therapy, dietary sterols and stanols represent a legitimate first-line strategy. For a high-risk patient already on a statin, the additional 8-14% is real but modest, useful as part of a broader dietary approach but insufficient to reach aggressive LDL targets alone.

Sources matter. The most consistent clinical data come from fortified foods rather than supplements. The fortified margarines and yogurts in European markets have the longest track record. Plant sterol supplements in capsule form are available but less consistently studied. There is no evidence of harm at the recommended doses, and they do not interact with statins.

The distinction between soluble and insoluble fiber is not just biochemical pedantry. Insoluble fiber (mostly cellulose and lignin in vegetable skins, wheat bran) adds bulk and supports bowel regularity. It has minimal effect on LDL. Soluble fiber, found in oats (beta-glucan), psyllium, legumes, barley, apples, and citrus, forms a viscous gel in the intestinal lumen that traps bile acids and prevents their reabsorption. This is the mechanism with the LDL effect.

The numbers: a meta-analysis of 67 controlled trials found that 2-10 grams of soluble fiber per day reduced LDL by 1.6-4.2 mg/dL on average (roughly 2-4% per gram of soluble fiber consumed) (Brown L et al, Am J Clin Nutr 1999, DOI: 10.1093/ajcn/69.1.30). Daily consumption of 70-80 grams of dry oats (about 3 servings) provides roughly 3 grams of beta-glucan soluble fiber and reduces LDL by 5-8% in most studies. Psyllium husks at 10-15 grams per day reduce LDL by 6-10%.

In absolute terms, this is smaller than a statin's effect. But it is additive to statins, it has zero pharmacological side effects, and it compounds with other dietary interventions. A patient who eats 3 servings of oats, 2 grams of plant sterols, and follows a Mediterranean diet pattern may reduce LDL by 20-30% through diet alone, which is meaningful in a primary prevention context.

Niacin's pharmacologic profile is genuinely impressive in isolation: at doses of 1-3 grams per day, it reduces LDL by 15-25%, raises HDL by 15-35%, lowers triglycerides by 20-50%, and reduces Lp(a) by 20-30%. For two decades after its introduction in the 1960s, it was the most broadly active lipid drug available. The Coronary Drug Project, published in 1975, showed that niacin reduced nonfatal MI and five-year mortality in men with prior MI (Canner PL et al, JACC 1986, DOI: 10.1016/S0735-1097(86)80293-5).

The problem is what happened when niacin was tested against the backdrop of modern statin therapy. AIM-HIGH (niacin added to simvastatin in patients already at LDL goal) was stopped early because niacin added no cardiovascular benefit and showed a trend toward increased stroke. HPS2-THRIVE (extended-release niacin plus laropiprant added to statin therapy in 25,000 patients) showed no reduction in major vascular events and a significant increase in myopathy, new-onset diabetes, gastrointestinal complications, and infection (HPS2-THRIVE Collaborative Group, N Engl J Med 2014, DOI: 10.1056/NEJMoa1300955).

The likely explanation: the incremental HDL and LDL benefit from niacin, when added to a statin already at LDL goal, provides no additional atherosclerotic regression. The drug's benefit in the pre-statin era was real but cannot be replicated in an era of statin-treated background risk.

Over-the-counter fish oil supplements at typical doses (1-2 grams per day) produce modest triglyceride lowering, roughly 5-10%. At the prescription dose of 4 grams per day, both EPA and DHA-containing formulations lower triglycerides by 20-50%, with the magnitude depending on baseline triglyceride level: the higher the starting triglycerides, the larger the absolute reduction. This is one of the most reliable dose-response relationships in lipid pharmacology.

What the form question comes down to is cardiovascular outcomes rather than triglyceride lowering. See Q23 for the detailed icosapent ethyl (Vascepa) story. The clinical bottom line: if your goal is triglyceride reduction alone (triglycerides above 500 mg/dL with concern for pancreatitis), any prescription omega-3 formulation will achieve it. If your goal is cardiovascular event reduction in a patient with elevated triglycerides despite statin therapy, the evidence supports icosapent ethyl specifically.

One practical note: over-the-counter fish oil products are not FDA-regulated for content, and many do not deliver the labeled dose of EPA or DHA (Kleiner AC et al, J Sci Food Agric 2015, DOI: 10.1002/jsfa.7211). Independent testing by organizations like ConsumerLab has found variability in content and oxidation levels across brands. If you are going to take fish oil for a cardiovascular purpose, the prescription formulations provide known dose and purity.

The biochemical distinction: both EPA and DHA are long-chain omega-3 polyunsaturated fatty acids. DHA is longer (22 carbons) and more abundant in brain and retinal tissue; EPA is shorter (20 carbons) and more active in inflammatory signaling pathways. At a mechanistic level, EPA competes with arachidonic acid for cyclooxygenase and lipoxygenase enzymes, reducing proinflammatory eicosanoid synthesis. DHA also has anti-inflammatory effects, but its impact on LDL particle size and LDL-C appears to differ from EPA's: DHA increases LDL-C modestly in some patients, EPA generally does not (Jacobson TA et al, J Clin Lipidol 2012, DOI: 10.1016/j.jacl.2012.04.003).

The clinical outcomes data: REDUCE-IT (icosapent ethyl, pure EPA) showed 25% relative risk reduction in MACE. STRENGTH (EPA+DHA carboxylic acid formulation) showed no significant MACE reduction. ORIGIN (low-dose EPA+DHA) showed neutral results. The pattern suggests that the cardiovascular benefit seen with EPA is specific to the EPA molecule or to high-dose pure EPA formulation, and is not a class effect of omega-3 fatty acids generally.

There is ongoing debate about whether the mineral oil placebo used in REDUCE-IT inflated the benefit by raising LDL-C in the comparator arm (the OMEMI and other studies found no such benefit in high-risk elderly patients). The debate is legitimate; the REDUCE-IT results are still the best available outcomes evidence for triglyceride-elevated patients.

The REDUCE-IT trial enrolled 8,179 patients with established cardiovascular disease or diabetes plus elevated triglycerides (150-499 mg/dL) despite statin therapy and randomized them to icosapent ethyl 4 g/day or mineral oil placebo. Over a median follow-up of 4.9 years, the primary composite endpoint (cardiovascular death, nonfatal MI, nonfatal stroke, coronary revascularization, or hospitalization for unstable angina) was reduced from 22.0% to 17.2% (absolute risk reduction 4.8%, number needed to treat 21 over five years) (Bhatt DL et al, NEJM 2019, DOI: 10.1056/NEJMoa1812792).

The controversy: the mineral oil placebo is not biologically inert and appears to have raised LDL-C and CRP in the control arm, potentially exaggerating the treatment benefit. The JELIS trial (published 2007, Japan), which used eicosapentaenoic acid added to statin therapy with a different placebo design, showed an 18% reduction in major coronary events without this concern. The JELIS data support EPA benefit but with a different magnitude and population.

My clinical position: icosapent ethyl is FDA-approved, the outcomes data are positive even accounting for placebo controversy, and for patients with established cardiovascular disease and triglycerides persistently above 150 mg/dL on statin therapy, the benefit-risk ratio supports prescribing it. Substituting fish oil from the grocery shelf is not an equivalent strategy.

Triglycerides above 500 mg/dL warrant treatment regardless of cardiovascular risk because hypertriglyceridemia at this level causes acute pancreatitis, a painful, potentially life-threatening event unrelated to cardiovascular disease. At this threshold, fibrates or prescription omega-3 are the primary pharmacologic tools, with the choice depending on whether statins are also being used (fibrate-statin combinations carry myopathy risk, particularly with gemfibrozil).

For triglycerides between 200 and 500 mg/dL in a patient without established cardiovascular disease and without diabetes, the evidence for adding prescription omega-3 is weaker. Lifestyle modification (reducing refined carbohydrates, eliminating alcohol, losing weight if overweight, increasing physical activity) can reduce triglycerides by 20-50% in this range and should be tried first. A statin, if indicated for LDL or overall risk, also reduces triglycerides by 10-30% as a secondary effect.

The triglyceride story is also fundamentally about metabolic health. Persistently elevated triglycerides above 200 mg/dL in the absence of secondary causes (hypothyroidism, renal disease, diabetes, certain medications) is a signal of insulin resistance. Treating triglycerides without addressing the underlying metabolic disorder is treating the marker, not the disease.

The timing logic is driven by three purposes: confirming treatment response, checking for new metabolic changes (new diabetes, new thyroid disease), and tracking risk evolution over time. For each purpose, the interval differs.

Treatment response: a statin lowers LDL within 4 weeks; the maximal effect is established by 6-8 weeks. A lipid panel at 6-12 weeks after starting or changing therapy gives you the information needed to adjust the dose or add a second agent. Checking too early (before 4 weeks) shows incomplete effect; checking too late (after 6 months) delays dose optimization.

Stable patients on therapy: if LDL is at target, ApoB is at target, and the patient's metabolic status has not changed, annual rechecking provides reassurance and detects the occasional patient whose response attenuates over time (rare but it happens, usually because of weight gain or new metabolic disease). Every-six-months testing for a stable patient at LDL goal adds cost and anxiety without clinical value.

Untreated patients: for someone in their forties with borderline LDL (100-129 mg/dL) who is managing through diet and exercise, a recheck every 1-2 years tracks whether the trend is stable or rising, which informs when pharmacotherapy conversation becomes necessary.

One exception: patients started on PCSK9 inhibitors often benefit from an early recheck (4-6 weeks) to confirm the drug is working, because the cost and injection burden make confirming efficacy worthwhile early.

For three decades, the standard instruction was to fast for 9-12 hours before a cholesterol test. This requirement was based on the fact that the most common LDL calculation in the United States, the Friedewald equation (LDL = Total Cholesterol - HDL - Triglycerides/5), requires an accurate fasting triglyceride value. Triglycerides rise significantly after eating, which would make the calculation inaccurate.

Two developments have changed the picture. First, many laboratories now measure LDL directly (direct LDL-C) rather than calculating it, making the triglyceride value irrelevant to LDL accuracy. Second, large epidemiological studies demonstrated that non-fasting lipid panels predict cardiovascular risk at least as well as fasting panels, because the post-meal state is where the human body actually spends most of its time (Langsted A et al, JAMA 2019, DOI: 10.1001/jamainternmed.2019.0297). The 2016 European Atherosclerosis Society/European Federation of Clinical Chemistry consensus statement explicitly endorsed non-fasting lipid testing as the default.

When fasting still matters: if your lab calculates LDL from total cholesterol and triglycerides (rather than measuring it directly), you should fast. If triglycerides are being monitored because they are elevated or you are taking omega-3 therapy, fasting provides the more clinically meaningful number. If an accurate non-HDL cholesterol is the target metric, fasting adds precision.

The Langsted and Nordestgaard studies from the Copenhagen General Population Study tracked over 90,000 individuals and showed that non-fasting LDL and non-HDL cholesterol were as strongly associated with ischemic heart disease and stroke as fasting values (Langsted A et al, JAMA 2019, DOI: 10.1001/jamainternmed.2019.0297). The biological explanation: postprandial lipemia does elevate triglyceride-rich remnant particles, but total LDL-C and direct LDL-C change relatively little after eating. The practical implication: the variance introduced by non-fasting state for LDL-C is smaller than the variance between individuals and smaller than the clinical decision threshold.

Non-HDL cholesterol, which is calculated as total cholesterol minus HDL, does not require a fasting sample because it captures all atherogenic lipoproteins including remnant particles. Non-HDL is actually argued by some to be a better primary target than LDL-C because it encompasses VLDL-C and IDL-C in addition to LDL-C, and these remnant particles are increasingly recognized as atherogenic (see Q43 on residual risk). ApoB, which counts all atherogenic particles directly, has the same advantage.

The cultural inertia around fasting is strong. Many patients have been told to fast for so long that they assume something is wrong with a non-fasting result. It is worth knowing what your lab is actually measuring: direct vs. calculated LDL, so the conversation with your physician is accurate.

This is the most clinically dangerous misconception I encounter. A man who is told his cholesterol is "fine" concludes that his heart attack risk is low. He may be right. He may also be completely wrong, depending on what was not measured.

The landmark INTERHEART study showed that elevated non-HDL cholesterol, ApoB, and the ApoB-to-ApoA1 ratio were among the strongest predictors of first myocardial infarction globally (Yusuf S et al, Lancet 2004, DOI: 10.1016/S0140-6736(04)17018-9). But the INTERHEART data also showed that hypertension, smoking, diabetes, abdominal obesity, and psychosocial stress each contributed independently. A man with normal LDL who smokes, has a waist circumference of 42 inches, has fasting glucose of 108, and has never had a CAC score has several pathways to a cardiac event that a standard lipid panel cannot detect.

The specific mechanisms: Lp(a) can drive plaque and clot formation with normal LDL (see Q4-6). Insulin resistance drives small dense LDL pattern and elevated particle number with normal LDL-C (see Q8). Inflammation (elevated hs-CRP) independently predicts events after statin therapy, as JUPITER demonstrated. Coronary artery calcium score can be positive and significant with LDL below 130; the CAC score captures decades of accumulated risk, not just the current lipid snapshot.

FH results from mutations in the LDL receptor gene (LDLR), the apolipoprotein B gene (APOB), or the PCSK9 gene, all of which impair LDL clearance from the bloodstream. Heterozygous FH (one copy of the mutation, the most common form) typically produces LDL levels of 190-400 mg/dL. Homozygous FH (two copies, extremely rare at 1 in 250,000) produces LDL above 400 mg/dL and coronary artery disease in childhood.

How would you know? The clues: LDL above 190 mg/dL without obvious secondary cause (hypothyroidism, nephrotic syndrome, obesity); first-degree relatives with heart attacks or coronary bypass before age 55 (men) or 65 (women); physical findings such as xanthomas (cholesterol deposits in Achilles tendons or hand tendons), xanthelasmas (yellow deposits around the eyelids), or corneal arcus before age 45. Genetic testing confirms the diagnosis but is not required for clinical management.

The Dutch Lipid Clinic Network score, the Simon Broome criteria, and the MEDPED criteria are the three diagnostic scoring systems used in different countries. Any of them can be applied in a clinic visit.

Why does it matter? Because FH patients have lifetime LDL exposure that starts at birth. A 50-year-old with untreated FH has had roughly 30 more years of high LDL exposure than an age-matched person who developed high LDL from diet and lifestyle at age 40. Genetic testing of first-degree relatives after a proband is identified can potentially prevent heart attacks in family members who do not yet know they carry the mutation (Khera AV et al, JACC 2016, DOI: 10.1016/j.jacc.2016.09.931).

The case for pediatric FH screening is compelling and underappreciated. Children with untreated heterozygous FH develop measurable atherosclerosis in childhood: carotid intima-media thickness is already increased by age 10 in affected children (Wiegman A et al, NEJM 2004, DOI: 10.1056/NEJMoa040805). Coronary artery events in untreated FH men occur at a median age in the 40s, meaning the arterial damage accumulates silently through the twenties and thirties.

The cascade screening approach (testing all first-degree relatives after a proband is identified) has been the most efficient strategy in countries where it has been implemented. The Netherlands and United Kingdom have national FH screening programs that have identified tens of thousands of affected individuals; both programs reduced cardiac events and improved survival in FH populations compared to opportunistic screening.

Statin therapy is now approved and used in children as young as 8-10 with FH in specialized centers, with long-term safety data from the PLASC trial and the pediatric FH registry showing no developmental harm and significant reduction in carotid IMT progression. The decision to treat a child with a statin is not taken lightly, but for a child with a confirmed FH mutation and LDL above 190 mg/dL, the risk of untreated disease substantially exceeds the documented risk of childhood statin therapy.

The 2018 ACC/AHA guidelines moved away from fixed LDL thresholds as automatic treatment triggers toward a risk-based conversation. The Pooled Cohort Equations calculate 10-year atherosclerotic cardiovascular disease risk; statin therapy is clearly indicated when risk is above 7.5-10%, and a clinical discussion with "risk-enhancing factors" is recommended at lower risk levels. LDL above 190 is always an indication for statin regardless of calculated risk, bypassing the risk calculator because of the presumed genetic cause.

But the risk calculator has important limitations for younger patients. A 38-year-old with LDL of 175, no other risk factors, and no family history has a 10-year Pooled Cohort risk of perhaps 3%. The calculator says no statin. But his lifetime risk of a cardiovascular event may be very high, because LDL exposure is cumulative and he has 40 more years of it ahead. The concept of LDL-years (analogous to pack-years for smoking) captures this: the area under the LDL-time curve predicts cardiovascular risk better than any single snapshot.

For men under 40 with LDL above 160 and any risk-enhancing factor (family history, elevated ApoB, elevated Lp(a), elevated hs-CRP, or coronary calcium score above zero), I have a long conversation that includes both the 10-year calculator and the lifetime trajectory. A CAC score of zero in this group argues for watchful waiting; any calcium argues for early intervention.

This question comes up frequently because patients have read that "cholesterol is essential" and worry that lowering it too much will deprive their cells of a necessary molecule. The concern is theoretically coherent but clinically unfounded. Cholesterol synthesis occurs primarily in the liver, and the liver can meet all cellular cholesterol requirements through endogenous synthesis even when circulating LDL is extremely low. People with congenital PCSK9 loss-of-function mutations who have lifelong LDL of 15-30 mg/dL live healthy, cognitively intact lives with no increased rate of cancer, hemorrhagic stroke, or neurodevelopmental disorders (Cohen JC et al, NEJM 2006, DOI: 10.1056/NEJMoa054013).

The ODYSSEY Outcomes and FOURIER trials both had pre-specified analyses of patients who achieved very low LDL levels. In patients with LDL below 25 mg/dL or even below 15 mg/dL, event rates continued to decrease without any signal of increased non-cardiovascular harm. The cardiovascular benefit-risk calculation remained clearly favorable at these extreme levels.

Where does clinical caution apply? Hemorrhagic stroke risk is modestly increased at very low LDL in populations with high baseline hemorrhagic stroke rates (particularly East Asian populations with high intracerebral hemorrhage background rates). This signal was primarily observed in the statin era and is modest, but it informs the conversation for patients who have already had a hemorrhagic stroke.

Secondary prevention (treating patients who have already had a cardiovascular event) is where the LDL evidence is strongest and where the treatment targets are most aggressive. The pathophysiologic rationale is clear: after a plaque rupture event, the arterial lesion is unstable, vulnerable plaques remain in other locations, and the inflammatory milieu is heightened. Aggressive LDL lowering in this phase stabilizes remaining plaques, reduces neovascularization within plaques, and possibly reduces the lipid core in existing lesions.

The landmark trial history: PROVE-IT showed that atorvastatin 80 mg (achieving LDL ~62 mg/dL) was superior to pravastatin 40 mg (achieving LDL ~95 mg/dL) after ACS. TNT confirmed this in stable CAD. FOURIER showed that adding evolocumab to statin therapy, bringing LDL to ~30 mg/dL, reduced recurrent MI by 27% over 2.2 years. ODYSSEY Outcomes showed a 15% reduction in MACE with alirocumab in post-ACS patients, with greater benefit in those achieving LDL below 54 mg/dL.

Practically: after a heart attack, I start high-intensity statin therapy (atorvastatin 40-80 mg or rosuvastatin 20-40 mg) before discharge, add ezetimibe if LDL remains above 70 mg/dL at 6-week follow-up, and refer for PCSK9 inhibitor if still above 70 mg/dL on dual oral therapy. This stepwise approach is now codified in the 2022 ACC Expert Consensus Decision Pathway.

A 51-year-old endocrinologist came to clinic convinced that if he raised his HDL from 38 to 55, his cardiovascular risk would drop proportionately. He was eating a diet heavy in avocado and olive oil specifically for this purpose. His instinct is correct that those foods raise HDL; the data support it. His assumption that higher HDL concentration means lower risk is where the story gets more complicated.

The epidemiological association between high HDL and low cardiovascular risk is robust. But Mendelian randomization studies, which test whether genes that raise HDL also reduce cardiovascular events, have been consistently disappointing. Patients with genetic variants that produce chronically high HDL do not have correspondingly low cardiovascular event rates (Voight BF et al, Lancet 2012, DOI: 10.1016/S0140-6736(12)60312-2). Drug trials of HDL-raising agents (niacin, CETP inhibitors such as torcetrapib, evacetrapib, and anacetrapib) failed to show cardiovascular benefit despite raising HDL by 70-130%. This pattern suggests that HDL-C concentration is a marker of metabolic health rather than a directly modifiable therapeutic target.

Where does this leave dietary HDL-raising? The Mediterranean dietary pattern, which includes olive oil, fatty fish, nuts, and moderate wine, reduces cardiovascular events in primary prevention (PREDIMED trial). But the mechanism is likely anti-inflammatory and multifactorial, not simply HDL elevation.

The exercise-HDL relationship has been studied in dozens of trials. A meta-analysis of aerobic exercise and HDL found a mean increase of 2.53 mg/dL (roughly 5-8% of typical baseline HDL), with greater increases seen at exercise volumes above 900 kcal/week (Kodama S et al, Arch Intern Med 2007, DOI: 10.1001/archinte.167.10.999). Running, cycling, and swimming produced similar results; resistance training alone raised HDL less consistently.

The more interesting finding from recent research is that exercise changes the quality of HDL particles in addition to their quantity. HDL's primary cardiovascular function is cholesterol efflux, the ability to accept cholesterol from macrophages in arterial walls and transport it back to the liver. Endurance-trained individuals have HDL particles with superior cholesterol efflux capacity compared to sedentary controls, independent of the absolute HDL-C level (Camont L et al, Arterioscler Thromb Vasc Biol 2013, DOI: 10.1161/ATVBAHA.112.300083). This functional improvement may be the more clinically meaningful benefit of exercise, and it is not captured by the standard HDL-C number.

Practically: exercise raises HDL modestly and improves HDL function more meaningfully. But exercise's total cardiovascular benefit (lowering blood pressure, reducing insulin resistance, reducing resting heart rate, improving vascular endothelial function) vastly exceeds what can be attributed to HDL alone.

For two decades, the J-shaped curve in epidemiology (showing moderate drinkers with lower cardiovascular mortality than abstainers or heavy drinkers) was cited as evidence that 1-2 drinks per day protect the heart. The HDL-raising effect, along with modest antiplatelet effects, was the proposed mechanism.

The problem: abstainers in many epidemiological studies include former heavy drinkers who quit because of illness, and sick quitters inflate the abstainer risk group. When this confounding is properly addressed by separating lifelong abstainers from former drinkers, the cardiovascular benefit of moderate alcohol largely disappears (Ronksley PE et al, BMJ 2011, DOI: 10.1136/bmj.d671). Mendelian randomization studies using genetic variants that predict alcohol metabolism found no cardiovascular benefit from genetically proxied alcohol consumption.

The current consensus from the WHO, American Cancer Society, and increasingly the AHA: no amount of alcohol is safe from a cancer standpoint, and the cardiovascular benefit argument is weaker than it appeared two decades ago. For someone who drinks moderately and has no alcohol use disorder, the cardiovascular risk from 1-2 drinks daily is likely small and possibly not real. For someone who does not currently drink, starting to drink for cardiovascular benefit is not recommended.

This is one of the most important questions in preventive cardiology because it directly confronts the comfortable assumption that fitness protects against all cardiovascular risk factors. It does not. Exercise and a good diet lower LDL (typically by 5-15%) but they cannot overcome a genetic predisposition to significantly elevated LDL.

The clinical scenario I see regularly: a 44-year-old endurance athlete, lean, non-smoker, metabolically healthy, LDL of 195. He is confused because he has done everything "right." The confusion is understandable, because the popular health narrative associates high cholesterol with poor lifestyle. But LDL production and clearance are governed primarily by hepatic LDL receptor activity, which is genetically determined. You can have perfectly efficient LDL receptor function and still overproduce LDL through the cholesterol synthesis pathway. You can have heterozygous familial hypercholesterolemia and be an Olympic athlete simultaneously.

The additional complication for lean fit people: dietary fat composition matters even in people with normal body weight. A lean person eating high saturated fat or large amounts of certain cheeses and red meats may drive LDL up because saturated fatty acids downregulate LDL receptor expression. This is reversible with dietary change, but it operates independently of body weight and fitness level.

FH screening (see Q29) is the first priority for a lean fit person with LDL above 190. Polygenic risk scoring is increasingly available and can distinguish FH from high-polygenic-risk individuals.

Very low carbohydrate diets reliably reduce triglycerides (typically 20-40%) and raise HDL (typically 10-20%), which improves the triglyceride/HDL ratio. These are genuine metabolic benefits that improve metabolic syndrome markers. But most ketogenic diets also increase saturated fat intake, and saturated fatty acids (lauric, myristic, palmitic) downregulate LDL receptor expression in the liver, reducing LDL clearance. The net LDL effect varies by individual and by the specific fatty acid composition of the diet.

For most people on a ketogenic diet, LDL rises modestly (perhaps 5-15 mg/dL). For a specific subset characterized by lean body mass, high metabolic rate, and certain genetic polymorphisms in lipid metabolism genes (the "lean mass hyper-responder" phenotype described in Q39), LDL can rise dramatically, often to 200-300 mg/dL or higher. The particle nature of this LDL elevation matters: in metabolically healthy individuals, the elevated LDL on keto tends to be large buoyant pattern A, which some argue is less atherogenic. Whether this offsets the absolute LDL elevation is an active and unresolved debate.

My practical approach: if you are starting a ketogenic or carnivore diet, check a baseline lipid panel and ApoB, repeat at 6-8 weeks, and again at 3 months. If LDL rises above 190 mg/dL or ApoB rises significantly, the diet is not risk-neutral regardless of the particle size argument.

The LMHR pattern was described systematically by Norwitz, Feldman, and colleagues: lean individuals on ketogenic diets who develop LDL above 200 mg/dL (often 200-400 mg/dL) while having triglycerides below 70 mg/dL and HDL above 80 mg/dL (Norwitz NG et al, Curr Dev Nutr 2021, DOI: 10.1093/cdn/nzab144). The proposed mechanism is energy substrate channeling: in lean ketogenic individuals with high metabolic energy demands, VLDL is secreted at higher rates to deliver fatty acids to peripheral tissues, and the LDL generated from this VLDL is predominantly large buoyant (pattern A) particles.

The clinical question of whether LDL of 300 mg/dL in a lean, low-triglyceride, high-HDL individual carries the same atherosclerotic risk as LDL of 300 mg/dL in a metabolically obese insulin-resistant individual is genuinely unanswered. The Lipid Energy Model is a hypothesis, not a proven clinical framework.

What does caution require? LMHR individuals should get a CAC score. Coronary artery calcium is the best available clinical tool to assess whether plaque is actually accumulating. If a 42-year-old LMHR has been running LDL of 250 mg/dL on carnivore for three years, a CAC of zero is reassuring; a CAC of 200 is not. Until long-term outcomes data exist (and they do not yet), the claim that elevated LDL in the LMHR context is benign should be held with appropriate skepticism.

Cafestol is among the most potent dietary inducers of LDL known. It raises LDL by inhibiting bile acid synthesis and downregulating LDL receptors in the liver. Kahweol has similar effects. Both are oily compounds found in the lipid fraction of coffee that are removed when coffee passes through a paper filter. A meta-analysis quantified the effect: five cups of French press coffee daily can raise LDL by approximately 20-30 mg/dL over several weeks (Urgert R et al, Eur J Clin Nutr 1997, PMID 9218922). Boiled coffee (common in Scandinavia and Turkey) and unfiltered espresso without crema removal have similar effects.

Paper-filtered drip coffee, which is the most common preparation in the United States, removes more than 99% of cafestol and kahweol. In studies comparing filtered versus unfiltered coffee, filtered coffee drinkers showed no significant LDL elevation (Urgert R & Katan MB, NEJM 1997, DOI: 10.1056/NEJM199706193362506).

Espresso occupies a middle ground: a single espresso contains significant cafestol, but typical espresso serving sizes (1-2 oz) are small compared to French press serving sizes (8-12 oz), so the total cafestol dose is lower with one or two espressos than with three to four cups of French press. The clinical relevance increases if you are drinking four or more cups of French press or boiled coffee daily.

Thyroid hormone (primarily T3) upregulates transcription of the LDL receptor gene. In hypothyroid states, LDL receptors are downregulated, LDL clearance slows, and serum LDL rises. The degree of rise correlates with the severity of hypothyroidism. Even subclinical hypothyroidism (TSH above 4.5 mU/L with normal T4) can produce LDL elevations of 10-20 mg/dL (Biondi B & Cooper DS, NEJM 2012, DOI: 10.1056/NEJMcp1100073).

The clinical trap: a patient has mildly elevated LDL at 165 mg/dL and is started on a statin. Six months later her LDL is 148 and the statin appears to be working adequately. What no one checked is her TSH. She has Hashimoto's thyroiditis and a TSH of 8.2 mU/L. Treating the hypothyroidism alone would likely reduce her LDL by 20-30 mg/dL and might eliminate the statin indication entirely.

This is not a rare scenario. Hypothyroidism affects approximately 5% of the adult US population, with higher rates in women over 50. It is an easily treatable condition with a known lipid effect. TSH is not expensive. I routinely check TSH in any patient with new or worsening hyperlipidemia, particularly in women over 45.

Triglycerides are also elevated in hypothyroidism, and HDL may fall. The full dyslipidemia of hypothyroidism (elevated LDL, elevated triglycerides, reduced HDL) can mimic atherogenic dyslipidemia of metabolic syndrome and should be addressed by treating the thyroid, not the lipids in isolation.

The argument for universal testing is cumulative. ApoB is a superior predictor of cardiovascular events when discordant from LDL, and discordance occurs in a substantial minority of patients, particularly those with metabolic syndrome, elevated triglycerides, or normal-weight metabolic dysfunction. Lp(a) is elevated in 20% of adults globally, causes events through pathways not addressed by LDL-lowering, and changes the intensity of LDL treatment required even though Lp(a) itself cannot yet be pharmacologically lowered in most patients.

The European Atherosclerosis Society recommended a one-time Lp(a) measurement for all adults in 2022, specifically because knowing it changes how aggressively other risk factors are managed. The 2022 Canadian Cardiovascular Society guidelines incorporated ApoB as the primary target for lipid-lowering therapy. The United States lags these recommendations, in part because of billing and reimbursement structures and in part because the clinical workflow to act on elevated ApoB is not yet standardized.

What is the practical cost of not testing? A man in his mid-forties with LDL of 118, ApoB of 155, and Lp(a) of 210 nmol/L gets a routine physical, is told his cholesterol is acceptable, and gets no further evaluation. Over the next fifteen years, his coronary calcium accumulates. He has a heart attack at sixty. This is not a hypothetical. This is the scenario that universal ApoB and Lp(a) testing is designed to prevent.

The FOURIER and ODYSSEY Outcomes trials demonstrated that even at LDL below 55 mg/dL, event rates in secondary prevention patients remained substantial: roughly 10-15% over 2-5 years. This residual risk is clinically important and represents the frontier of cardiovascular prevention.

The major contributors: elevated triglycerides and remnant cholesterol (VLDL-remnants and IDL) independently predict cardiovascular events after LDL is lowered. Studies including the Copenhagen General Population Study found that nonfasting remnant cholesterol predicted myocardial infarction independent of LDL-C (Varbo A et al, J Am Coll Cardiol 2013, DOI: 10.1016/j.jacc.2013.05.051). Inflammation, measured by high-sensitivity CRP, persists after statin therapy and was targeted by the CANTOS trial using canakinumab (colchicine, which has a similar mechanism, showed benefit in COLCOT and LoDoCo2 trials with a better safety profile). Elevated Lp(a) represents residual prothrombotic risk that statins do not touch.

Treatment of residual risk is an evolving field. Icosapent ethyl addresses triglyceride-driven residual risk. Colchicine addresses inflammatory residual risk (colchicine 0.5 mg daily reduced recurrent MI by 23% in LoDoCo2). Emerging Lp(a)-lowering agents address Lp(a)-driven residual risk when their trials report. The practical implication for a patient on a statin: reaching LDL goal is the beginning of the conversation, not the end of it.

The ASTEROID trial randomized patients with coronary artery disease to rosuvastatin 40 mg and performed intravascular ultrasound (IVUS) at baseline and after 24 months. At a mean LDL of 60.8 mg/dL, there was statistically significant regression of total atheroma volume, the first unequivocal demonstration of plaque regression with statin therapy, (Nissen SE et al, JAMA 2006, DOI: 10.1001/jama.295.13.joc60002). SATURN confirmed that rosuvastatin outperformed atorvastatin in plaque regression at higher doses, consistent with the lower LDL achieved.

What does plaque regression mean clinically? The atherosclerotic plaque does not disappear. The lipid core shrinks, the fibrous cap thickens, inflammation within the plaque decreases, and the overall plaque becomes more stable, less vulnerable to rupture. This is the mechanism behind reduced event rates at very low LDL: not elimination of plaque, but transformation from vulnerable to stable plaque.

The clinical bottom line: if your LDL goal is plaque regression, the target is below 70 mg/dL, sustained, over years. Brief periods of low LDL do not achieve it. A patient who gets LDL to 62 for six months, then drifts to 95 because of medication non-adherence, is not achieving the stable LDL environment that drives plaque stabilization.

Secondary prevention is where the statin evidence is strongest and the clinical consensus is clearest. After a heart attack, statin therapy reduces recurrent MI, stroke, and cardiovascular death by 25-35%, numbers replicated across dozens of trials and meta-analyses. The 2018 ACC/AHA guidelines describe high-intensity statin therapy as a class I recommendation (highest level) after established ASCVD. There is essentially no risk-benefit argument that favors withholding statins in secondary prevention.

Primary prevention is more nuanced. The absolute risk reduction from statins in low-risk individuals is small (perhaps 1-3% over 10 years) and must be weighed against the small but real risk of statin-related side effects and the cost of lifelong therapy. The USPSTF 2022 update recommends statins in adults aged 40-75 with 10-year cardiovascular risk above 10% and at least one risk factor. The ACC/AHA 2018 guideline adds a shared decision-making conversation at lower risk thresholds with risk-enhancing factors.

The most important tool for primary prevention risk stratification beyond the Pooled Cohort Equations is the coronary artery calcium score. A CAC of zero, even in a high-risk-calculator individual, identifies a patient in whom statin therapy can often be deferred (and patient preference respected) with reassurance. A CAC above 100 in a low-risk-calculator individual upgrades treatment intensity. The calcium score is the primary prevention differentiator that individual labs and office visits cannot replicate.

The CAC score is a non-contrast CT scan of the chest that detects and quantifies calcium deposits in coronary artery walls. Calcium in coronary arteries represents mature atherosclerotic plaque; it is not present in normal arteries. Each Agatston unit represents plaque that took years to accumulate, so the CAC score is a summary of your lifetime arterial risk burden.

The MESA trial (Multi-Ethnic Study of Atherosclerosis) provided the landmark data: among individuals classified as intermediate risk by traditional risk factors, those with CAC = 0 had 10-year cardiovascular event rates below 4%, while those with CAC above 300 had event rates above 15%, a more than fourfold difference in a group that the risk calculator treated similarly (Blaha MJ et al, JACC 2016, DOI: 10.1016/j.jacc.2016.03.533). The 2018 ACC/AHA guideline made CAC scoring a class IIa recommendation (reasonable to perform) for intermediate-risk patients where the statin decision is uncertain.

Practically: I use the CAC score to have a more concrete conversation about risk than a probability percentage provides. Showing a patient their calcium score is worth 1,000 words about cardiovascular risk. A score of 0 is genuinely reassuring. A score of 450 in a 52-year-old who thought he was low-risk is the conversation starter that changes his life.

The specific data: in the MESA cohort, individuals with CAC = 0 had a 10-year event rate of approximately 3.4%, which is below the 7.5% threshold that triggers a strong statin recommendation in the 2018 guidelines. This "calcium zero advantage" persists even in patients with elevated LDL, elevated blood pressure, or other traditional risk factors; CAC zero is a powerful negative risk marker (Blaha MJ et al, JACC 2016, DOI: 10.1016/j.jacc.2016.03.533).

However, three qualifications apply. First, CAC zero does not mean zero soft plaque (non-calcified plaque that CT cannot detect). Younger patients (under 45) in particular may have significant soft plaque burden without calcification. Second, elevated Lp(a) can cause cardiac events with CAC = 0, through a thrombotic rather than purely atherosclerotic mechanism. Third, CAC zero with LDL 160 is a dynamic situation: the calcium score is a snapshot in time, and if LDL continues at 160 for another decade, calcium will accumulate.

My clinical practice: for a 48-year-old with LDL 160, no family history, no diabetes, no smoking, and CAC = 0, I discuss the situation honestly, check Lp(a) and ApoB, recommend repeating CAC in 5 years, and support the patient's decision whether or not to start a statin. This is true shared decision-making.

The standard lipid panel (LDL, HDL, triglycerides, total cholesterol) predicts approximately 40-50% of future cardiovascular events statistically. The remaining predictive information comes from factors the panel does not measure: particle number (ApoB), genetic thrombotic risk (Lp(a)), vascular inflammation (hs-CRP), and imaging-based plaque burden (CAC score).

Among blood tests specifically: ApoB outperforms LDL-C in most large datasets for future event prediction, particularly in people with metabolic syndrome. Lp(a) adds independent risk prediction in all risk categories. High-sensitivity CRP is the best-studied inflammatory marker; the JUPITER trial showed that individuals with low LDL but elevated hs-CRP who were randomized to rosuvastatin had a 44% reduction in MACE, suggesting that hs-CRP captures a population for whom statins provide benefit that the LDL level does not predict (Ridker PM et al, NEJM 2008, DOI: 10.1056/NEJMoa0807646).

High-sensitivity cardiac troponin measured at rest has emerged as a risk predictor in large population cohorts, predicting cardiovascular events even without acute myocardial injury, the chronic low-level myocardial stress signal. This is promising but not yet translated to clinical practice routinely.

My practical battery for full risk assessment: standard lipid panel + ApoB + Lp(a) + hs-CRP + fasting glucose or HbA1c + CAC score. The blood tests cost under $100; the CAC scan costs $75-150.

The biology: atherosclerosis is fundamentally an inflammatory disease, not purely a lipid storage disease. LDL initiates the plaque process when it penetrates the arterial intima, but the plaque grows, becomes vulnerable, and ruptures through inflammatory mechanisms involving macrophages, cytokines (particularly IL-6 and IL-1beta), and the acute-phase response; CRP is the most clinically accessible marker of this response. C-reactive protein is produced by the liver in response to IL-6 signaling; it rises within hours of inflammatory stimulation and falls within days if the stimulus resolves.

The landmark contribution of hs-CRP to cardiovascular prediction came from the JUPITER trial, which selected 17,802 apparently healthy individuals with LDL below 130 mg/dL but hs-CRP above 2 mg/L. Rosuvastatin 20 mg reduced LDL by 50% and hs-CRP by 37%, and reduced the primary endpoint by 44% (Ridker PM et al, NEJM 2008, DOI: 10.1056/NEJMoa0807646). This demonstrated that statins have anti-inflammatory effects beyond LDL lowering, and that hs-CRP elevation identifies a population that benefits from statin therapy even with normal LDL.

Persistent hs-CRP elevation above 2 mg/L after statin therapy identifies residual inflammatory risk. This is the target for colchicine (LoDoCo2, COLCOT trials) and is the mechanistic rationale for the IL-1beta and IL-6 pathways being studied in secondary prevention.

This question forces a useful clinical prioritization. If I am walking into a first consultation with a man I know nothing about, and I can order one lipid number before I see him, I order ApoB.

The argument in full: LDL-C measures cholesterol mass in LDL particles but misses particle count, small dense LDL pattern, and non-LDL atherogenic particles. Non-HDL cholesterol captures VLDL and IDL in addition to LDL, which is useful, but still measures mass rather than particle number. Total cholesterol to HDL ratio is a reasonable global marker but is crude. Triglycerides tell me about metabolic health but are highly variable day-to-day and meal-to-meal.

ApoB does the following simultaneously: it counts LDL particles (each carries one ApoB); it counts VLDL and IDL particles (also carrying ApoB); it identifies the LDL-C/ApoB discordance that characterizes high-particle, low-mass LDL patterns; it provides the same information as LDL particle number by NMR at a fraction of the cost; and it is the number most strongly associated with carotid intima-media thickness, coronary calcium, and cardiovascular events in large prospective cohorts (Boren J et al, Eur Heart J 2020, DOI: 10.1093/eurheartj/ehz962).

If I could have a second number, it would be Lp(a), not because it overlaps with ApoB (Lp(a) is a distinct risk pathway), but because it identifies the 20% of patients with an irreducible genetic risk floor that changes how aggressively I need to treat everything else.

LDL remains the most universally measured, guideline-embedded number in lipid medicine, and I use it every day. But if the question is which single number tells me the most about arterial risk, ApoB wins, not as an opinion but as a statistical fact across every major prospective study that has compared them.