Diet, cholesterol, and the saturated fat controversy

Patients ask me about saturated fat constantly. Half are convinced it is the dietary villain it was painted as in the 1980s. The other half have read Nina Teicholz or Aseem Malhotra and are now convinced it was never bad at all. Some have cycled between the two positions multiple times. The actual evidence sits in between, but is not as muddled as the public debate makes it sound. Here is what I think is fair to say.

The origin of the confusion

In the 1950s, Ancel Keys observed that countries with more saturated-fat-heavy diets had more heart attacks. The Seven Countries Study was the most famous version of this argument. It oversimplified things (countries differ in many ways besides fat intake), but the core observation was real and has held up.

What followed was a generation of low-fat eating advice, often bad advice. "Replace saturated fat" was operationalized as "eat low-fat versions of everything," which meant replacing fat with sugar and refined carbohydrates. That replacement does not lower heart disease risk; in some studies it raises it. The 1980s and 1990s American supermarket — Snackwell cookies, fat-free cereals, sugary yogurt — was not what the science actually supported.

The contrarian wave was a reaction to this overreach. The PURE study, published in The Lancet in 2017, looked across 18 countries and reported that higher saturated fat intake was associated with lower total mortality. Headlines followed. The study had real problems: the comparison populations differed enormously by region, low fat intake correlated with poverty, and the analysis adjusted for very few confounders. But the take-home from the popular press was "saturated fat is fine." That message stuck in part because it is what people wanted to hear.

Three things that are settled

Separate the science into three layered claims. The first two are not really in dispute among cardiologists.

One: saturated fat raises LDL cholesterol. This is the most replicated finding in nutritional biochemistry. Hundreds of controlled feeding studies, dose- response, mechanism understood at the receptor level (saturated fat downregulates the LDL receptor in hepatocytes, less LDL gets cleared from the blood). If you swap saturated fat for polyunsaturated fat in a controlled diet, LDL falls within weeks. Not subtle.

Two: LDL causes atherosclerosis. This is the part the contrarians try to dispute and where they are weakest. We have Mendelian randomization studies: people born with genes that lifelong lower their LDL have proportionally lower rates of heart attacks. That is the cleanest possible test of causality, because the genes were assigned at conception and cannot be confounded by diet or exercise. Multiple statin trials, PCSK9 inhibitor trials, and ezetimibe (a non-statin) outcome trials have all shown that lowering LDL by any mechanism lowers heart attacks proportionally. There is a clean dose-response. The "LDL hypothesis" is not a hypothesis at this point.

Three: therefore, lowering dietary saturated fat lowers LDL, which lowers heart attacks. The Cochrane reviews of saturated-fat-reduction trials show modest but real risk reduction for cardiovascular events, roughly 20%, when the replacement is polyunsaturated fat. The effect on total mortality is smaller and less consistent. The effect on heart attacks specifically is real.

The genuine open questions

Not all saturated fats behave the same. Stearic acid (the dominant saturated fat in dark chocolate and beef) raises LDL much less than palmitic acid (the dominant saturated fat in palm oil, butter, and most American food). Lauric acid (coconut oil) raises LDL most of all. Coconut oil is not a heart-healthy fat, despite years of marketing. The composition of what you are eating matters more than the "saturated" label.

The replacement matters more than the removal. If you cut saturated fat and replace it with refined carbohydrates, LDL might fall but triglycerides rise and HDL falls and the net cardiovascular effect is roughly zero. Replace it with polyunsaturated fat (nuts, seeds, fish, olive and canola oil) and risk falls. Replace it with whole-grain complex carbohydrates and you get a smaller but still real benefit. The right comparison is never "saturated fat versus nothing." It is "saturated fat versus what."

Individual response varies a lot. Some people are LDL hyper-responders: cut their saturated fat by 5 grams a day and LDL falls 15 mg/dL. Some are not: they can cut their saturated fat in half and barely move LDL by 5. Some of this is the ApoE genotype; some is polygenic and we cannot predict it. The practical implication is that you find out by trying. Two months of disciplined dietary change with a before/after lipid panel is the experiment.

Patterns beat nutrients. The actual best dietary evidence in cardiology is for the Mediterranean pattern (PREDIMED trial, AHA-endorsed) and the Portfolio diet (nuts, soy protein, viscous fiber, plant sterols). Neither is fundamentally about counting saturated fat grams; both happen to be moderate in saturated fat and rich in things that lower LDL through other mechanisms (soluble fiber binding bile acids, plant sterols competing for absorption, the polyphenols-and-MUFA from olive oil). If you eat Mediterranean or Portfolio and your LDL is fine, you do not need to also count saturated fat.

The lifestyle floor

Here is the conversation I have most often, because it is the one patients find most disheartening. They cut saturated fat, eat Mediterranean, lose weight, exercise, sleep, and their LDL falls a little, then plateaus around 120 or 130. They want to keep trying. They believe more lifestyle will eventually do it.

For most people, lifestyle lowers LDL by 10–30%, sometimes more. The variation is genetic: some of you are wired so that your liver makes more LDL particles or clears them less efficiently than mine. That is not a moral failing. It is your biology, and lifestyle has a floor for any given genome.

If you have done six months of real, sustained dietary work, exercise, weight optimization, and your LDL has plateaued at a level your cardiologist still considers too high for your risk profile, you have hit your floor. The question stops being "should I try harder" and becomes "do I need a medication." The 2026 guidelines lower the LDL thresholds for which medication is recommended at any given risk level (I wrote about that in the cholesterol guidelines piece). If your floor is too high for your risk, that is what medication is for.

My advice to patients

Mostly: eat a Mediterranean-style diet most of the time, emphasizing olive oil, fatty fish, nuts, legumes, whole grains, vegetables, fruit. Use butter and red meat as flavors, not staples. Do not believe coconut oil is a health food. Do not replace saturated fat with sugar. Recheck your lipids a few months in.

If your LDL came down to where your cardiologist wants it for your risk, you are done; just keep doing what you are doing. If it plateaued and you are still above target for your risk profile, the conversation moves on. Two months of disciplined diet, a repeat panel, and a straight read on whether you have hit your floor is the most useful experiment in nutritional cardiology.

The saturated fat debate is mostly a debate about whether people want to take agency over their LDL through diet or through medication. Both work. Both are reasonable. The evidence is clear enough that you should not feel cornered by the choice.

— Rahul Deo, MD, PhD