ApoB and LDL-C: particle number and atherosclerotic risk

LDL cholesterol and apolipoprotein B describe different features of lipoprotein transport. LDL-C measures cholesterol assigned to the LDL fraction. ApoB measures a structural protein present in one copy per corresponding particle. Identical LDL-C concentrations can accompany different particle numbers, particularly with hypertriglyceridemia, diabetes, insulin resistance, or lipid-lowering treatment. This discordance does not invalidate LDL-C treatment evidence or make a single biomarker a complete description of cardiovascular risk.

This narrative scientific review connects molecular biology, laboratory methods, observational research, and drug trials. A cellular mechanism, a laboratory change, and prevention of myocardial infarction are distinct levels of evidence.

Cholesterol and its carriers

Anichkov and Khalatov's experiments, Gofman's lipoprotein research, Framingham, and Brown and Goldstein's discovery of LDLR established complementary perspectives on cholesterol transport. Genetic, epidemiological, and randomized evidence supports a causal role for prolonged exposure to atherogenic lipoproteins. Statin trials including 4S, WOSCOPS, LIPID, HPS, and PROVE-IT demonstrated fewer events. Residual risk does not render LDL-C an ineffective treatment target. The claim that 60–70% of all treated patients necessarily have recurrent events within five years confuses the fraction of relative risk remaining with the proportion of patients experiencing an event.

A lipoprotein has a hydrophobic triglyceride and cholesteryl-ester core surrounded by a monolayer of phospholipids, unesterified cholesterol, and proteins. It is not a membrane vesicle with a lipid bilayer. Intestinal chylomicrons carry dietary lipids; hepatic VLDL undergoes lipolysis to form IDL and LDL. Lipoproteins support energy distribution, membranes, and steroid production, but prolonged arterial retention of ApoB-containing particles contributes to disease.

Particle concentration, exposure duration, entry into the artery, and retention offer a conceptual account of atherogenesis. Their product is not a validated equation for individual clinical risk. Age, blood pressure, smoking, inflammation, diabetes, genetics, and plaque characteristics also affect outcomes.

APOB, RNA editing, and particle assembly

APOB is located at 2p24.1 and contains 29 exons. Its unusually large exon 26 encodes a substantial portion of the protein. Liver and intestine are the principal sites of synthesis. HNF4α and other transcription factors contribute to tissue regulation, while lipid availability, translation, and protein degradation determine how much ApoB is secreted.

In enterocytes, APOBEC1 catalyzes cytidine-to-uridine editing of APOB mRNA. In conventional numbering, C6666→U converts CAA to UAA, introducing a stop in place of Gln2153. ApoB-48 contains the first 2152 residues of the mature sequence. Its name denotes approximately 48% of the length, rather than precisely half the mass. A1CF and RBM47 contribute to RNA recognition and organization of the editing machinery; they should not be described as universally interchangeable obligatory cofactors. Powell and Chen's 1987 studies and Teng's 1993 work established the underlying mechanism. In adult human liver, editing differs markedly from the intestine and full-length ApoB-100 is produced. The mature protein contains 4536 residues; precursor numbering includes the signal peptide.

ApoB-48 supports chylomicrons and their remnants. It lacks the C-terminal ApoB-100 region involved in LDLR binding, but remnants are cleared by the liver through ApoE-dependent processes involving LDLR and LRP1. Lack of ApoB-100 does not eliminate receptor-mediated clearance. Residence time varies with size, feeding, and metabolic state. Rapid clearance under normal conditions does not exclude the atherogenic potential of accumulating remnants. Large chylomicrons and smaller remnants also differ in their ability to enter the arterial wall.

During ApoB-100 synthesis, movement through Sec61 is coupled to lipidation by microsomal triglyceride transfer protein, MTP. Its heterodimer includes MTTP and protein disulfide isomerase. Transfer of neutral lipids and phospholipids supports particle formation in the endoplasmic reticulum. Poorly lipidated ApoB can enter ER-associated degradation and the proteasome; other quality-control pathways also regulate secretion. Assembly is therefore not determined solely by APOB transcript abundance.

Historical structural model and cryo-electron microscopy

Segrest's five-part model described alternating amphipathic regions:

$$\mathrm{NH_2}-\beta\alpha_1-\beta_1-\alpha_2-\beta_2-\alpha_3-\mathrm{COOH}.$$

The globular N-terminal region contributes to assembly, amphipathic β-regions interact with the lipid surface, and other segments accommodate changes in particle size. This remains a useful historical model, but its approximate boundaries should not be presented as atomically demonstrated independent domains. In mature-protein numbering, site B near residues 3359–3369 was investigated as a proteoglycan-binding region. Arg3500Gln is associated with familial defective ApoB, and Cys4326 contributes to the covalent connection with apo(a) in Lp(a). Structural and genetic comparisons must specify their numbering convention.

Berndsen and Cassidy (2025) produced an integrative ApoB-100 structure with a globular N-terminus and an extended β-belt surrounding LDL. PDB 9EA7 reports 9 Å resolution. Reimund and colleagues (2025) studied the ApoB-100–LDLR complex; 9BDT reports 5.4 Å. The original account's common 3.4 Å resolution and 2024 publication date are therefore corrected. The newer work identifies multiple interfaces and distributed basic patches. Receptor binding cannot be reduced to three site-B residues and LA4/LA5 alone. Integrative modeling and local refinement do not establish equal accuracy for every residue.

What ApoB actually measures

A VLDL, IDL, LDL, or Lp(a) particle ordinarily contains one ApoB-100 molecule; an intestinal particle contains one ApoB-48 molecule. This is a structural principle of biogenesis, not a conclusion based solely on inability to fit a second protein around a small LDL sphere: VLDL is much larger. Routine ApoB testing measures protein mass and provides a practical indicator of the total number of ApoB-containing particles. It does not count particles individually or separately identify LDL-P, size, origin, or Lp(a) concentration.

For an illustration assuming ApoB-100 molar mass of 550000 g/mol and negligible ApoB-48:

$$n=\frac{c_m}{M},\qquad N_V=\frac{c_m}{M}N_A.$$

Here c_m is protein mass concentration in g/L, M is molar mass in g/mol, n is mol/L, N_A≈6.022×10²³ mol⁻¹, and N_V is particles/L. At 100 mg/dL=1 g/L:

$$n\approx1.82\times10^{-6}\ \mathrm{mol/L},\qquad N_V\approx1.10\times10^{18}\ \mathrm{L^{-1}} =1.10\times10^{12}\ \mathrm{mL^{-1}}.$$

The original 10¹⁵ particles/mL value was a thousandfold overestimate. This educational conversion is approximate, dependent on assumed molar mass, glycosylation, ApoB-48 contribution, and assay characteristics. It is not a new laboratory calibration.

LDL-C and laboratory limitations

Cholesterol content changes with particle size and remodeling. The same LDL-C can therefore coexist with different ApoB. Fixed counts of 500–2700 cholesteryl-ester molecules or an obligatory 2–2.5-fold particle difference cannot be applied to everyone. Method, composition, unesterified and esterified cholesterol, and the contribution of other particle classes to total ApoB all matter.

Friedewald's 1972 equation assumes a particular triglyceride-to-VLDL-cholesterol relationship:

$$\begin{aligned} \mathrm{LDL\!\!-C}&=\mathrm{TC}-\mathrm{HDL\!\!-C}-\mathrm{TG}/5 &&\text{(all concentrations in mg/dL)},\\ \mathrm{LDL\!\!-C}&=\mathrm{TC}-\mathrm{HDL\!\!-C}-\mathrm{TG}/2.2 &&\text{(all concentrations in mmol/L)}. \end{aligned}$$

TC is total cholesterol and TG is triglycerides. The classical equation should not be used at TG≥400 mg/dL, approximately 4.5 mmol/L. At low LDL-C with increased TG it often underestimates LDL-C, but the error is not universally 15–40%.

Martin–Hopkins (2013) replaces the fixed denominator with an adaptive factor selected from TG and non-HDL-C:

$$\mathrm{LDL\!\!-C}=\mathrm{TC}-\mathrm{HDL\!\!-C}-\mathrm{TG}/\zeta.$$

This expression uses mg/dL and the appropriate lookup table. The same denominator cannot be transferred mechanically to mmol/L. Newer or extended equations need validation for the TG range concerned. Direct homogeneous assays use selective reagents and enzymatic cholesterol measurement; complex dyslipidemias can produce method-dependent deviations. No LDL-C method is itself a particle-number measurement.

Non-HDL-C=TC−HDL-C includes cholesterol carried by LDL, remnants, and Lp(a). It is available from the standard lipid panel and particularly useful with raised TG. It still describes lipid cargo, allowing discordance with ApoB. A population LDL-C reference interval is not an individual's treatment target.

Small dense LDL: size, composition, and metabolism

LDL is heterogeneous. Gradient electrophoresis, ultracentrifugation, ion mobility, and NMR classify particles using different features; labels and thresholds are not fully interchangeable. Historical pattern A denotes larger predominant particles and pattern B smaller ones. The frequently used electrophoretic boundary around 25.5 nm cannot be transferred to every NMR method.

Austin and Krauss linked small dense LDL with coronary events. An invariant independent 3–5-fold risk increase after accounting for ApoB and other factors has not been established. Size does not make large LDL safe or demonstrate the need for routine subclass testing.

ApoB conformation, receptor affinity, residence time, oxidation susceptibility, and proteoglycan interactions can change with remodeling. Spherical curvature κ=1/R increases as radius decreases, but geometry alone does not prove a specific site-B deformation in every person. Universal Kd values, a 4.5–5-day half-life, and 2.5–3-fold faster transcytosis are not established clinical constants.

The Stokes–Einstein expression describes diffusion of an idealized particle in a fluid:

$$D=\frac{k_BT}{6\pi\eta r_H}.$$

D is diffusion coefficient, k_B Boltzmann's constant, T absolute temperature, η dynamic viscosity, and r_H hydrodynamic radius. A smaller particle diffuses faster with other conditions held constant. Receptor-mediated endothelial transport involving SR-B1 and ALK1 is not governed by this equation alone. Fixed tocopherol molecule counts and an obligatory threefold shorter oxidation lag cannot be generalized from individual experiments. Hyperglycemia promotes ApoB glycation and other alterations. The article on glycoxidation and AGEs explains related chemistry without reducing clinical risk to a single mechanism.

Remodeling in insulin resistance

Reduced suppression of adipose lipolysis by insulin increases nonesterified fatty-acid flow through ATGL and HSL. Portal delivery from visceral tissue is one component rather than the sole source of hepatic lipid. Dietary fat, de novo lipogenesis, oxidation, and export jointly determine balance. Insulin and carbohydrate signaling regulate SREBP-1c and ChREBP; their activation cannot be attributed exclusively to incoming fatty acids.

Excess TG can promote secretion of large VLDL1. CETP transfers neutral lipids between lipoproteins, enriching LDL and HDL with TG and VLDL with cholesteryl esters. Its structure contains a hydrophobic cavity, but transport cannot be described as invariably equimolar exchange or a single proven continuous bridge. Hepatic lipase, LIPC, hydrolyzes TG and phospholipids and contributes to smaller LDL formation. Endothelial lipase, LIPG, is predominantly a phospholipase, particularly relevant to HDL; it is not equivalent to the principal TG-hydrolyzing step in sdLDL production.

HDL remodeling can increase ApoA-I dissociation and catabolism. Raised TG, lower HDL-C, and greater ApoB particle numbers often coexist, but their relationship is not identical in every phenotype. Raising HDL-C alone is not an established substitute for reducing atherogenic burden.

Discordance and epidemiological evidence

Discordance can be defined using medians, percentiles, regression residuals, or treatment thresholds. These definitions should not be mixed. The following matrix expresses comparative interpretation, not four universal absolute-risk categories.

Sniderman's discordance studies, Marston's 2022 analysis of UK Biobank and treated populations, Johannesen's 2021 Copenhagen research, and Framingham data support additional information from ApoB. Observational models do not establish that risk follows ApoB exclusively, identical risk in all quadrants, or safety in a person with high LDL-C. The original N=389529, N=90197, and subgroup HR values could not be reliably verified and are not retained as confirmed findings. The original PMID 33736829 points to an unrelated article; Johannesen's DOI is retained.

Johannesen (2021): 13015 statin-treated participants, median follow-up 8 years; above-median ApoB/below-median LDL-C: myocardial infarction HR 1.49 (95% CI 1.15–1.92). Observational association; medians are not treatment targets.

INTERHEART was an international case-control study rather than a prospective cohort. It investigated ApoB/ApoA-I among other factors. A population-attributable fraction around 49% for that factor is neither an individual's probability of infarction nor a treatment risk reduction, and does not establish ApoB alone as superior in every setting. Framingham supports the importance of long-term exposure; Pencina's 2019 non-HDL-C trajectory paper is not a randomized comparison of ApoB targets.

Retention, plaque formation, and complications

Williams and Tabas's response-to-retention hypothesis connects early atherogenesis with extracellular-matrix trapping of arterial lipoproteins. Basic ApoB regions interact with negatively charged sulfate and carboxyl groups on glycosaminoglycans. Biglycan, versican, decorin, and perlecan differ in structure and function. Matrix characteristics and local hemodynamics affect retention. Borén's 1998 work identified an important proteoglycan-binding site. Altering binding experimentally and reducing atherosclerosis in animal models does not establish complete prevention in humans. In that experiment, replacing the basic cluster impaired both functions, while K3363E preserved receptor binding with defective proteoglycan interaction; R3500Q showed the opposite separation. This specific experiment does not demonstrate complete absence of human atherosclerosis.

Secretory acid sphingomyelinase generates ceramide from sphingomyelin and promotes LDL aggregation. Secretory phospholipases A₂ can generate lysophospholipids and fatty acids. Lipid oxidation produces MDA and 4-HNE, which modify proteins. MPO generates HOCl using H₂O₂ and chloride; peroxynitrite primarily forms from NO reacting with superoxide, not simply as an obligatory MPO product in the presence of NO.

Macrophage CD36 and SR-A1, together with LOX-1 in appropriate cells, recognize modified lipoproteins. Their regulation differs from cholesterol-dependent suppression of LDLR, but complete absence of any feedback is an overstatement. Aggregate uptake, macropinocytosis, SOAT1/ACAT1-mediated esterification, and cholesterol efflux contribute to foam-cell formation. Smooth-muscle-derived cells also accumulate lipids within plaques.

Cholesterol crystals and lysosomal injury can activate NLRP3, ASC, caspase-1, and maturation of IL-1β and IL-18. Inflammation, cell death, defective efferocytosis, and matrix metalloproteinases contribute to necrotic cores and thinner caps. MMP-1/13 collagenases and MMP-2/9 gelatinases differ in substrate specificity. The historical 65-µm morphological cap threshold is not an individual laboratory predictor. Plaque rupture or erosion can provoke thrombosis; progression does not inevitably produce a transmural infarction in every person.

Treatment targets and pharmacotherapy

ESC/EAS retains LDL-C as the primary treatment target. ApoB and non-HDL-C provide additional burden assessment, especially in diabetes, obesity, raised TG, or very low LDL-C. ApoB does not replace separate Lp(a) measurement. Goals follow clinical risk rather than universal reference ranges.

With a recurrent vascular event within two years on maximally tolerated therapy, LDL-C<40 mg/dL may be considered. This is not a separate universal “extreme” category mandating ApoB<55 mg/dL. Age alone does not establish moderate risk. The 2025 ESC/EAS update retained LDL-C goals and incorporated newer treatment evidence; this ApoB table is explicitly attributed to 2019. AHA/ACC and NLA use their own frameworks, and a treatment-intensification threshold is not always synonymous with a target.

Statins inhibit HMG-CoA reductase and increase LDLR through lower hepatic cholesterol. LDL-C can fall proportionally more than ApoB, but a fixed 10–15% gap or an explanation solely involving preferential removal of large particles is not universal. Ezetimibe reduces intestinal cholesterol uptake through NPC1L1. IMPROVE-IT established clinical benefit from adding it to a statin after acute coronary syndrome.

Evolocumab and alirocumab prevent PCSK9-dependent LDLR degradation; FOURIER and ODYSSEY OUTCOMES support fewer events in their respective populations. Inclisiran is a GalNAc-conjugated siRNA that suppresses hepatic PCSK9 synthesis through RISC. LDL-C lowering does not automatically demonstrate identical outcome effects for every agent. Modest Lp(a) changes with antibodies cannot be reduced to ordinary LDLR clearance; the mechanism is more complex.

Bempedoic acid is activated predominantly in liver by ACSVL1 and inhibits ATP citrate lyase upstream of HMG-CoA reductase. Limited activation in skeletal muscle reduces particular risks but does not establish absolute safety. Among 13970 CLEAR Outcomes participants, the primary composite occurred in 11.7% versus 13.3%, HR 0.87, an absolute difference of 1.6 percentage points. HR 0.77 for infarction concerns a separate endpoint. Increased uric acid, gout, and cholelithiasis require consideration. Prescribing and combinations depend on diagnosis and tolerability.

Nutrition, supplements, and translational limitations

Viscous fiber, including psyllium and oat/barley β-glucans, affects intestinal viscosity, absorption, and bile-acid metabolism. Bile-acid loss can increase hepatic cholesterol requirements, involving CYP7A1, FXR/FGF19, and SREBP2–LDLR. Fiber is not an established direct ASBT inhibitor. A fixed 2.5–3.5-fold increase in excretion or a uniform ApoB response is not established for every product. Dose, molecular weight, processing, and the accompanying diet alter effects. LDL-C improvement is a surrogate, not independent proof that a specific supplement prevents infarction.

Plant sterols and stanols reduce cholesterol absorption through micellar competition and intestinal transport involving NPC1L1 and ABCG5/8. Approximately 2 g/day has often been studied, but this is not a universal treatment dose or evidence of fewer events. Sitosterolemia and uncertainty about long-term clinical outcomes need separate consideration. Sterols cannot be assumed to eliminate all compensatory absorption during statin treatment.

Kong (2004) and Cameron (2008) investigated berberine's effects on LDLR mRNA stability and PCSK9, including cellular models. AMPK, ERK, and HNF1α explain hypotheses rather than establish statin-equivalent clinical efficacy. Dihydroberberine has limited pharmacokinetic evidence; fivefold absorption and 12–18% ApoB reductions cannot be treated as universal effects. The article on berberine and dihydroberberine distinguishes model findings from clinical results. Product quality, interactions, and tolerability matter. These mechanisms do not establish a ready-made nutraceutical protocol.

Icosapent ethyl is prescription purified EPA ethyl ester. REDUCE-IT studied 4 g/day added to statins in high-risk individuals with TG 135–499 mg/dL and LDL-C 41–100 mg/dL. Over median 4.9-year follow-up, primary events were 17.2% versus 22.0%, HR 0.75: a 25% relative reduction and 4.8-percentage-point absolute difference. Atrial-fibrillation/flutter hospitalization was more frequent; serious bleeding was also assessed. This does not prove benefit from ordinary fish oil for everyone or disappearance of sdLDL. Effects on TG, fatty-acid oxidation, lipogenesis, and inflammation do not establish obligatory direct DGAT/CETP blockade or conversion of every person to pattern A.

Avoiding excess calories, replacing saturated with unsaturated fat, consuming fiber, physical activity, and smoking cessation belong within overall risk management. Excess sugar-sweetened beverages and fructose can increase hepatic lipogenesis; fructose does not become unlimited fat at every dose and in every food. A 30–35% VLDL decline within one month cannot be promised universally. Time-restricted eating and CLOCK/BMAL1/PER biology are research interests, but 16/8 is not an obligatory proven method of normalizing ApoB or treating steatosis through TFEB.

Interpretation and review methods

Assessment starts with overall risk and a lipid panel: TC, HDL-C, TG, LDL-C with its method, and non-HDL-C. ApoB is particularly useful when cholesterol cargo may underestimate particle number. Units, secondary causes, medication, and sampling conditions should be reviewed; unexpected discordance can require repeat measurement. Elevated ApoB despite achieved LDL-C is discussed alongside risk, tolerability, and treatment evidence. It does not automatically diagnose sdLDL or require a universal EPA–berberine–psyllium combination. Reaching a target reduces burden but does not completely compensate risk. Follow-up timing depends on treatment changes and clinical condition, rather than invariably being twelve months.

This review used the full original material with targeted verification of disputed units, structural entries, and selected primary publications on 8 October 2026. RCSB records, the REDUCE-IT abstract, CLEAR Outcomes publication, and official ESC information about the 2025 update were available. Some DOI retrievals and multiword searches yielded no usable result; the working record acknowledges this access limitation. Every bibliography entry has not undergone exhaustive independent verification. Unverified exact coefficients and therapeutic promises are not used as established facts. No systematic search, original GRADE assessment, external peer review, or clinical validation was performed.

ApoB helps identify carrier numbers and understand discordance. LDL-C remains an evidence-based treatment target. Interpreting both measures, exposure, and clinical context is more useful than presenting cholesterol and particles as mutually exclusive causes.