Why Atherosclerosis Must Be Rewritten
The problem is not ignorance
Medicine does not usually fail because it knows too little. It more often fails because it knows something well enough to stop asking. The history of atherosclerosis is a precise example of this dynamic.
Over the second half of the twentieth century, clinical lipidology built a coherent and useful story. Large populations were followed. Serum cholesterol was measured. Risk gradients were documented. Statins were developed, tested in randomized trials, and shown to reduce cardiovascular events with a consistency that few drug classes have matched. The science was real. The clinical gains were real. The epidemiological framework was robust.
And yet, two patients can sit in the same clinic, present nearly identical lipid panels, report similar diets, and carry similar cardiovascular risk scores—yet one will have advanced, rupture‑prone disease while the other will not. One will have a myocardial infarction in middle age. The other will reach late life with arteries that, on imaging, show little more than the ordinary wear of age. The conventional framework offers no satisfying account of this difference. It can calculate a probability. It cannot explain a case.
This is not a marginal anomaly. It is the central unresolved problem of the field. It signals not that we need more of the same measurements, but that we have been looking at the disease from the wrong distance.
A brief history of necessary simplification
To understand how the field arrived at this point, it helps to recall what it was trying to do in the mid‑twentieth century. The Framingham Heart Study, launched in 1948, was a methodological revolution. For the first time, a chronic disease was tracked prospectively in a defined population. The ambition was not to understand mechanism but to find predictors—numbers measurable in ordinary clinical settings that would identify people at elevated risk before catastrophe struck.
Total serum cholesterol was exactly the kind of predictor Framingham needed. It was measurable, reproducible, and statistically associated with coronary events across a population. The same was true, in subsequent decades, of LDL‑C, HDL‑C, and triglycerides. The lipid panel was not a theory of disease. It was an epidemiological instrument: a blunt, practical tool designed for population screening and treatment decisions.
The historical error was not in building that tool. It was in gradually treating the tool as a complete account. The lipid panel was promoted from a screening instrument to a causal explanation, and the question “what do these numbers actually contain?” was largely deferred. That deferral, compounded over decades of clinical education, guideline writing, and pharmaceutical development, is where the loss of resolution began.
What the field inherited was not a wrong story but a low‑resolution one—and a professional culture that had grown comfortable with the resolution it had.
Fragmentation at the level of causation
The conventional account of atherosclerosis does not suffer from a single error so much as from fragmentation. Lipid excess, chronic inflammation, endothelial dysfunction, thrombosis, oxidative stress, heredity, diet, and aging have accumulated in the literature as though they were competing explanations, when in fact they are better understood as different views of the same process seen at different levels of resolution.
Consider how the discourse actually runs. One researcher argues that LDL‑C is the primary cause and that everything else is secondary or epiphenomenal. Another argues that inflammation is the true driver and that lipids are merely its substrate. A third places diet at the center, treating saturated fat and refined carbohydrates as the proximate material causes of arterial pathology. A fourth points to genetic architecture—familial hypercholesterolemia, elevated Lp(a), polygenic loading—and insists that the causal weight sits in the genome rather than the kitchen or the pharmacy. A fifth notes that advanced coronary disease can be found in patients whose lipid panels have always been unremarkable, and concludes that something else entirely must be operative.
Each of these positions contains genuine evidence. Each of them, taken alone, is insufficient. The fragmentation is not a failure of individual investigators. It is a structural consequence of studying a layered biological process as though its layers were independent causes, when in fact they are levels of the same system.
Atherosclerosis is a chronic, systemic, metabolic–inflammatory disease of the arterial tree: a long‑running failure of lipid homeokinesis that continuously provokes and sustains inflammatory responses in the wall, with repeated cycles of plaque formation, destabilization, and repair.
Atherosclerosis cannot be adequately explained by cholesterol alone, nor by inflammation alone, nor by diet alone, nor by heredity alone. It is best understood as the progressive failure of a dynamic regulatory system in which lipid transport, arterial retention, cellular processing, inflammatory signaling, repair, and structural maintenance are normally coordinated and persistently discoordinated in disease. Cholesterol excess, inflammation, diet, and genetic substrate are each real influences on this system. The critical questions are: what is the system, how does it fail, and at what level of description can we best understand that failure?
Those questions require a different kind of answer than the lipid panel provides.
The causal architecture: three layers
A unified account of atherosclerosis requires a clear statement of causal architecture. This book proposes three layers, each operating at a distinct level of resolution, each necessary, none sufficient alone.
The first and foundational layer is disturbed lipid homeokinesis.
Healthy lipid biology is not a static equilibrium. It is a continuous dynamic process: synthesis, packaging, secretion, transport, receptor‑mediated uptake, intracellular processing, and clearance—all in constant flux, continuously adjusted in response to metabolic state, nutritional input, hormonal signals, inflammatory conditions, and cellular demand. When this dynamic regulation is intact, atherogenic lipoprotein burden in the arterial wall remains low even under metabolic challenge. When it fails—because of genetic defects in receptor function, acquired metabolic stress, chronic inflammation, endocrine disruption, or organ dysfunction—the system enters a trajectory. Lipid burden accumulates in the arterial wall not as a single event but as a long‑running process of dysregulated flux. That accumulated burden is the substrate on which everything else depends.
Without disturbed lipid homeokinesis, atherosclerosis as a systemic, progressive disease does not develop. Inflammation alone, endothelial injury alone, even oxidative stress alone does not produce the characteristic lesion. The arterial disease requires a sustained supply of atherogenic lipoproteins operating within a dysregulated system. This is the foundational causal layer.
The second layer is inflammatory dysregulation.
Once atherogenic lipoproteins—particularly small, dense, oxidation‑prone LDL particles and Lp(a)—penetrate the subendothelial space and are retained there, inflammation is not merely a response. It becomes a driver. Macrophage foam cell formation, cytokine and chemokine signaling, smooth muscle cell migration, neovascularization, and extracellular matrix remodeling all amplify the initial lipid burden into a complex, biologically active lesion. Inflammatory dysregulation determines the rate and character of progression. It governs the transition from stable fibrous plaque to vulnerable, rupture‑prone anatomy. And it is itself subject to modulation by systemic factors—infection history, metabolic state, circadian disruption, psychological stress—that may leave the lipid panel unchanged while profoundly influencing clinical trajectory.
Inflammation is therefore not an alternative explanation for atherosclerosis. It is the layer through which the lipid burden is translated into structural pathology.
The third layer is diet, in its proper role.
Diet has often been assigned a causal role in atherosclerosis that exceeds the evidence and creates persistent confusion in clinical practice and public communication. Dietary saturated fat can raise LDL‑C modestly in many individuals. Refined carbohydrates can raise triglycerides and lower HDL‑C. Specific dietary patterns are associated, with variable effect sizes, with cardiovascular risk at the population level.
But diet does not directly cause atherosclerosis in the same sense that a ruptured plaque directly causes an infarction. It modifies a system. It acts primarily on the first layer—lipid homeokinesis—by providing substrates, altering hepatic lipid processing, influencing insulin sensitivity, and shaping the inflammatory background. The same dietary pattern may have modest effects in a person with normal lipid regulatory machinery and large effects in someone with an inherited defect in LDL receptor function or elevated Lp(a). Diet is an upstream modifier. Its clinical relevance depends entirely on the system it is modifying.
Repositioning diet as a system modifier rather than the singular cause does not diminish the importance of nutritional medicine. It provides a more accurate account of how diet operates—one that explains why dietary interventions produce heterogeneous results across individuals and why population‑level associations rarely translate into precise individual predictions.
These three layers are not alternatives. Together with genetic, hemodynamic, and cellular context, they form a single causal architecture whose elements interact across scales.
From framework to clinic: two patients
The value of this architecture is tested at the bedside.
Consider two patients whose routine lipid panels appear remarkably similar.
Patient A is a fifty‑one‑year‑old man. His total cholesterol has hovered around 200 mg/dL for a decade. LDL‑C is in the low 120s, HDL‑C in the mid‑40s, triglycerides around 140 mg/dL. His ten‑year cardiovascular risk score is just under ten percent. His diet is unremarkable. He is not obese. He exercises moderately. There is no striking family history. Each year his physician notes that his numbers are “borderline” and recommends ongoing lifestyle attention.
At fifty‑two, he has a large anterior myocardial infarction. The culprit lesion—a mid‑LAD plaque—was not hemodynamically significant on the stress test performed fourteen months earlier. It ruptures anyway.
Patient B is a fifty‑three‑year‑old woman. Her total cholesterol is similar. LDL‑C and HDL‑C are in comparable ranges. Her calculated ten‑year risk is slightly below Patient A’s. She smokes intermittently and has mild insulin resistance. Two decades later, at seventy‑one, she undergoes coronary CT angiography for an unrelated reason. Imaging reveals diffuse, non‑obstructive coronary disease—plaque present throughout—but no critical stenoses, no symptoms, and no prior clinical events.
Two patients. Similar panels. Similar calculated risks. One catastrophically ill in his early fifties. One accumulating disease silently across twenty years without rupture.
The difference between them is not captured by the numbers that were measured. It lies in the structure that was not measured: the composition of lipoprotein subspecies, particle oxidation state, the inflammatory burden within the plaque, cap thickness, macrophage infiltration, endothelial integrity, and, beneath these, cellular phenotypes and signaling states. These are the “nuclear‑level” and emerging “elementary‑particle‑level” variables in the language of the Prefatory Note. The panel saw only bulk values and treated them as approximately equal.
This is why the disease must be rewritten: not because the standard measurements are useless, but because they describe only one level of a system whose decisive behavior occurs deeper in its architecture.
What rewriting does not mean
To argue that the conventional account is insufficient is not to argue that it is simply wrong. The case for LDL‑C reduction—by statins, by PCSK9 inhibitors, by dietary means—in appropriately selected individuals remains among the most replicated findings in cardiovascular medicine. ApoB reduction prevents events. That is established.
What is being argued here is not that LDL‑C is uninformative, or that statins are broadly misguided, or that the lipid panel should be abandoned. It is that the lipid panel samples only one level of a multi‑level system, and that clinical decisions built exclusively on that level are systematically blind to the structural and cellular variation that explains heterogeneous outcomes.
A patient with LDL‑C of 130 mg/dL dominated by large, buoyant particles with low oxidative susceptibility and preserved endothelial function is not the same biological entity as a patient with LDL‑C of 130 mg/dL dominated by small, dense, oxidation‑prone particles in a background of chronic low‑grade inflammation. The panel sees them as equivalent. The artery does not.
Rewriting atherosclerosis means adding resolution and integrating layers, not deleting what already exists.
Heredity without determinism: the Lp(a) case
A third corrective principle concerns the relationship between heredity and clinical expression. Because certain lipoprotein abnormalities are strongly genetic in origin, it is often assumed that their clinical consequences are correspondingly fixed—that a patient with genetically elevated Lp(a) carries a destiny written in the genome that clinical management can do little to alter.
This assumption is mistaken in several important senses, and Lp(a) provides the clearest example.
Lp(a) levels are among the most heritable traits in human biology, determined largely by variants in the LPA gene and only minimally responsive to diet or standard lipid‑lowering therapy. These facts have supported a clinical culture of therapeutic nihilism: “Your Lp(a) is elevated, but there is nothing we can do about it.”
The reality is more complex along three dimensions.
First, there is expressional variability. Although Lp(a) is predominantly heritable and relatively stable, clinically meaningful variation has been observed in some patients over time, particularly across physiological transitions and disease states. This variability is not random noise; it appears related to metabolic and inflammatory context and may have prognostic implications. In selected settings, serial measurement, rather than a single baseline value, may be more informative.
Second, there is modifier‑dependent expression. Even when Lp(a) levels are largely stable for an individual, the risk they confer is not uniform. Postmenopausal women, for example, often experience a rise in Lp(a) and a change in vascular vulnerability. Inflammatory burden, renal function, and thyroid status further shape Lp(a) biology. The gene specifies a range; the physiological context determines where within that range the patient sits, and how aggressively the particle population behaves.
Third, and most consequentially, there is therapeutic modifiability. RNA‑based therapeutics—antisense oligonucleotides and small interfering RNA agents—can achieve large reductions in Lp(a) by silencing hepatic production at the level of gene expression. The risk allele remains, but its expression can be profoundly suppressed. This is not a marginal refinement. It is a conceptual transformation: heredity is not destiny when the machinery of expression can be intervened upon.
The Lp(a) case will recur throughout this book not because it is an oddity, but because it is paradigmatic. It shows, in high relief, a general principle: a biological trait that is strongly genetic in origin may still be modifiable in its expression, context‑dependent in its severity, and pharmacologically tractable in its treatment. The same reasoning will be applied to other lipoprotein abnormalities, to inflammatory state, and to the acquired remodeling of lipid‑regulating machinery over the life course.
The architecture ahead
This opening chapter has established three foundations.
First, the conventional framework of atherosclerosis—though clinically productive—operates at a level of resolution that cannot explain why patients with similar panels have divergent fates.
Second, the causal architecture of atherosclerosis is layered: disturbed lipid homeokinesis provides the substrate; inflammatory dysregulation drives progression and structural failure; diet modifies the system upstream without constituting its direct cause. Together with genetic, hemodynamic, and cellular context, these layers form a single, integrated system.
Third, even strongly hereditary lipoprotein abnormalities are not biologically or therapeutically fixed. The genome specifies a tendency; physiology, context, and now molecular therapeutics determine what that tendency becomes.
Comments