Coconut oil raises LDL cholesterol in most controlled feeding trials, typically by roughly 10 to 25 mg/dL versus unsaturated oils, and it does not lower cholesterol in any meaningful sense. The more useful question is how it compares with butter, lard, and other saturated fats, and whether the particle-level effects differ from what total LDL suggests.
That opening is the whole answer in compressed form. What follows unpacks why the response is so consistent, where the numbers come from, which populations behave differently, and what a person can reasonably do with the information. Two caveats frame everything: most trials are small and short, and coconut oil is not a single chemical entity.
The fatty acid composition that drives the effect
Coconut oil is roughly 90 percent saturated fat by weight, one of the highest saturated fractions of any culinary fat. The dominant fatty acid is lauric acid (C12:0), at about 45 to 50 percent of total fat, followed by myristic acid (C14:0) at roughly 15 to 20 percent and palmitic acid (C16:0) at about 8 to 10 percent. These are the chain lengths that matter, because the liver handles each differently.
Lauric acid is the counterintuitive piece. It is a medium-chain fatty acid in name, but at 12 carbons it is absorbed and transported largely through chylomicrons like long-chain fats, not shunted straight to the portal vein like caprylic acid (C8) or capric acid (C10). The marketing claim that coconut oil is a "medium-chain triglyceride" oil conflates lauric acid with the C8 and C10 fraction that actually behaves like a fast-burning medium-chain fat. That C8-C10 fraction is only about 12 to 15 percent of coconut oil. This distinction is the single most important mechanistic correction in the entire coconut oil debate.
The mechanism by which saturated fat raises LDL is well established at the level of the LDL receptor. Higher saturated fat intake suppresses hepatic LDL receptor activity, reducing clearance of LDL particles from circulation. The liver also increases cholesteryl ester transfer protein activity, which shifts cholesterol from HDL onto VLDL and LDL particles. The net result is more circulating LDL cholesterol. Lauric acid appears to raise LDL somewhat less than palmitic acid does, which is why coconut oil typically raises LDL less than butter, but the direction is the same.
One plain-language version: coconut oil does not clog anything directly, but it tells the liver to leave more LDL in the bloodstream.
Quantified context: how much LDL rises, and compared with what
Controlled feeding trials, in which all food is provided and the only variable is the fat, give the cleanest numbers. A meta-analysis of 16 such trials published in 2018 in the journal Circulation found that coconut oil raised LDL cholesterol by about 10.5 mg/dL compared with nontropical vegetable oils such as olive, canola, and sunflower. The same analysis found no significant effect on body weight, waist circumference, or inflammatory markers.
The comparison matters more than the absolute number. Versus butter, coconut oil typically produces lower LDL, because butter is higher in palmitic acid. Versus olive oil or canola oil, coconut oil produces higher LDL. Versus lard or beef tallow, the difference is small and inconsistent. So "does coconut oil raise cholesterol" depends entirely on the reference fat, and the honest answer is: it raises LDL relative to unsaturated oils, lowers it relative to butter, and is roughly a wash against other animal fats.
| Fat (per tablespoon) | Saturated fat | Typical LDL effect vs. olive oil | Notes |
|---|---|---|---|
| Coconut oil | ~12 g | +10 to +15 mg/dL | Lauric-dominant; raises HDL too |
| Butter | ~7 g | +8 to +12 mg/dL | Palmitic-dominant; dairy matrix differs |
| Beef tallow | ~7 g | +5 to +10 mg/dL | Palmitic and stearic mix |
| Lard | ~5 g | +4 to +8 mg/dL | More oleic than tallow |
| Olive oil | ~2 g | Reference | MUFA-dominant |
| Canola oil | ~1 g | Neutral to slightly lower | High in oleic and alpha-linolenic |
| Sunflower oil | ~2 g | Neutral to slightly lower | Linoleic-dominant |
The table is directional, not a dosing chart. Individual LDL responses vary by genetics, baseline diet, and body weight, and a single tablespoon swap will not move a number by 15 mg/dL in every person.
The HDL question and why the ratio matters less than it seems
Coconut oil reliably raises HDL cholesterol, often by 4 to 8 mg/dL in feeding trials. This is the fact most often used to defend coconut oil, and it is real. The problem is that raising both LDL and HDL leaves the LDL-to-HDL ratio, and the total-to-HDL ratio, close to unchanged in many comparisons. The ratio is a weaker predictor of cardiovascular risk than LDL particle number or apolipoprotein B, so "it raises good cholesterol too" is not the reassurance it sounds like.
Apolipoprotein B is the better metric here. Each LDL, VLDL, and IDL particle carries one apoB molecule, so apoB counts particles directly. Saturated fat tends to raise apoB, and the effect is more consistently adverse than the HDL rise is protective. Most standard lipid panels do not report apoB, which is why the public conversation stays stuck on total and LDL cholesterol. A clinician can order apoB, and in someone with borderline LDL and a strong family history, it changes the picture.
The professional caveat: HDL-raising is not automatically cardioprotective. Trials of niacin raised HDL substantially and did not reduce cardiovascular events, and genetic variants that raise HDL do not reliably lower heart attack risk. Treat the HDL increase from coconut oil as a real biochemical observation with uncertain clinical benefit, not as a counterweight.
Edge cases: who responds differently
Not everyone's LDL responds the same way to saturated fat. Roughly 15 to 25 percent of people are hyper-responders, showing LDL increases two to three times the average when saturated fat intake rises. The mechanism is partly genetic, involving variants in genes such as PCSK9 and the LDL receptor pathway. For a hyper-responder, a few tablespoons of coconut oil daily can push LDL from 110 to 160 mg/dL, which crosses a clinically meaningful threshold.
People with familial hypercholesterolemia should treat coconut oil as a saturated fat like any other. Baseline LDL is already high because LDL receptors are impaired, and adding a dietary factor that further suppresses receptor activity works in the wrong direction. The same logic applies to people with diabetes, in whom saturated fat tends to worsen the lipid profile and in whom cardiovascular risk is already elevated.
Population context is the third edge case. Pacific Islander populations with traditionally high coconut consumption have been studied for decades, and the results are genuinely mixed. Some studies show lipid profiles no worse than comparison populations, which has been used to argue that coconut oil is harmless in that context. The caveat is that these populations also differ in fish intake, physical activity, and genetic background, and the traditional diet included whole coconut and fish rather than refined coconut oil poured from a jar. Observational data from one population cannot be generalized into a recommendation for another.
Three misconceptions worth retiring
The first is that virgin or cold-pressed coconut oil is meaningfully different from refined. Virgin coconut oil retains polyphenols and has a lower smoke point, and it tastes different, but its fatty acid composition is nearly identical. The LDL effect tracks the fatty acid profile, not the polyphenol content, so the "virgin" label does not change the cholesterol story.
The second is that medium-chain triglyceride claims apply to coconut oil. As noted above, only about 12 to 15 percent of coconut oil is C8 and C10, and those are the fats that bypass chylomicron transport and go to the liver quickly. Lauric acid does not. MCT oil sold as a supplement is a refined fraction, not coconut oil, and even that product does not lower LDL.
The third is that the lauric acid in coconut oil is uniquely neutral because it raises HDL. The HDL rise is real, but the LDL rise is more consistent and the particle-level data point the wrong way. A fat that raises both is not "neutral" by default; it depends on which metric predicts risk in that person.
One professional caveat that rarely makes it into consumer articles: the food matrix matters. Cheese raises LDL less than butter does despite similar saturated fat content, likely because of calcium and the dairy matrix. Coconut oil has no comparable matrix effect documented in controlled trials, which is part of why its LDL signal is clean and consistent.
Practical actions with concrete thresholds
The actionable question is not whether coconut oil is good or bad in the abstract, but where it sits in a diet and what the numbers say.
If your LDL is below 100 mg/dL and you have no other risk factors, using coconut oil as one fat among several, at roughly 1 to 2 tablespoons per day, is unlikely to move you into a risk category. If your LDL is between 100 and 130 mg/dL, swap most coconut oil for olive or canola oil and recheck lipids in 8 to 12 weeks. A drop of 10 to 15 mg/dL is typical when the swap is real. If your LDL is above 130 mg/dL, or above 100 mg/dL with diabetes or a family history of early heart disease, treat coconut oil as a saturated fat to limit, not a health food. Red flag: a total cholesterol above 240 mg/dL or an LDL above 160 mg/dL warrants a clinician visit regardless of what fat you cook with. Red flag: a family history of heart attack before age 55 in a male relative or before 65 in a female relative changes the threshold for action, independent of diet. Red flag: if you are already on a statin, dietary saturated fat still matters, but do not adjust anything based on a single lipid panel; trends over 6 to 12 months are what clinicians use.
A practical note on comparison: replacing one tablespoon of coconut oil with one tablespoon of olive oil is a small change. Replacing the habitual 3 to 4 tablespoons used in cooking and coffee with olive oil is a meaningful one. Dose and substitution pattern matter more than the label on the jar.
On cost and availability, which change over time: coconut oil is widely available and generally priced in the range of USD 8 to USD 15 per 16-ounce jar in the United States, with virgin and organic versions costing more. Olive oil and canola oil are comparable or cheaper by volume. Price should not drive the decision; the substitution target drives it.
Why the trials cannot settle the long-term question
Almost every controlled trial of coconut oil and cholesterol is small, short, and focused on surrogate markers. Typical studies enroll 20 to 60 participants and run 4 to 8 weeks. They measure LDL, HDL, and triglycerides, not heart attacks. No large randomized trial has tested coconut oil against hard cardiovascular outcomes, and one is unlikely to be funded, because the effect size is small and the food is not a patentable product.
The surrogate-marker logic is well established for LDL: lowering LDL reduces cardiovascular events, and the relationship is roughly linear on a log scale. That is why a 10 to 15 mg/dL LDL increase from habitual coconut oil use is not trivial at the population level, even if it is modest for an individual. But surrogate markers are not outcomes, and the honest scientific position is that the long-term cardiovascular effect of coconut oil specifically is inferred, not proven.
The limitations of the evidence base are worth stating plainly. Many trials do not control for total fat or calorie intake. Some compare coconut oil with butter and call it a win, which is a low bar. Few measure apoB or LDL particle number. Almost none last long enough to capture the 10- to 20-year horizon over which diet affects atherosclerosis. And the observational literature is confounded by the fact that people who use coconut oil differ from people who do not in income, education, and overall diet pattern.
The timeline of the lipid response itself is fairly consistent and worth knowing:
| Time point | What typically happens after adding coconut oil |
|---|---|
| 0 to 2 weeks | No measurable change; lipid panel lag |
| 2 to 4 weeks | LDL begins to rise; HDL may rise slightly |
| 4 to 6 weeks | Most of the LDL increase is visible |
| 6 to 8 weeks | Response plateaus in most feeding trials |
| 8 to 12 weeks | Recheck window for a real-world substitution |
| 3 to 6 months | New steady state on the changed diet |
| Beyond 6 months | Effect persists if the diet pattern persists |
The practical read: if you change fats and want to know what happened, wait 8 to 12 weeks and recheck. A single panel taken two weeks after a change will show noise, not signal.
What can and cannot be concluded
What can be concluded with high confidence: coconut oil raises LDL cholesterol relative to unsaturated vegetable oils, the effect is driven by its saturated fatty acid profile, lauric acid is not metabolically equivalent to C8 and C10 medium-chain fats, and the HDL rise does not cancel the LDL rise at the level of particle number or apoB.
What can be concluded with moderate confidence: the magnitude is roughly 10 to 15 mg/dL for a typical substitution, hyper-responders see more, and populations with high habitual intake show mixed results for reasons that are not fully explained by fatty acid composition alone.
What cannot be concluded: that coconut oil prevents heart disease, that it is a health food, that virgin versus refined matters for cholesterol, or that the long-term cardiovascular effect is known. The trials are too small, too short, and too focused on surrogate markers to support those claims. Anyone who tells you coconut oil lowers cholesterol is contradicting the feeding-trial literature. Anyone who tells you it is poison is overstating a modest, well-characterized effect. The accurate position sits between those, and it is specific to the comparison fat, the dose, and the person's baseline risk.