Eggs And Cholesterol: What The Evidence Actually Supports

Few foods have swung further in public reputation than the egg, and the reversal is not a rumour: the numerical daily limit on dietary cholesterol that anchored decades of breakfast advice is no longer in the American guidelines, and the body that reviewed that change has set out its reasoning in print. What follows sticks to what the cited papers actually say, including the places where they disagree with each other.

Where the 300-milligram cap came from

The American Heart Association's 2020 science advisory on dietary cholesterol is the document that reviews how the number came and went. It is a science advisory rather than a trial: a review of human studies on the relationship between dietary cholesterol, blood lipids and cardiovascular disease risk, together with one original meta-regression of controlled feeding studies Carson 2020. Quoting the 2015 Dietary Guidelines Advisory Committee, the advisory records that the Dietary Guidelines for Americans had "previously" recommended that cholesterol intake "be limited to no more than 300 milligrams per day", and that the 2015 committee "will not bring forward this recommendation because available evidence shows no appreciable relationship between consumption of dietary cholesterol and serum cholesterol" Carson 2020.

There was a second, less-quoted reason. The advisory notes that at the time of that committee, mean dietary cholesterol intake in the United States for people aged one year and over was already below 300 mg a day, so, in its words, "the importance of a numerical limitation was diminished" Carson 2020. It also records that the 2015 report was explicit that the lack of a stated limit "did not imply that dietary cholesterol is unimportant in the adoption of healthy eating patterns" Carson 2020.

Eggs became the emblem of that number partly because of where dietary cholesterol comes from in the first place: on average, the advisory reports, eggs contribute 25% of dietary cholesterol in the United States Carson 2020. On how much cholesterol is in one egg, the two references used here do not quite agree, and that is worth naming rather than smoothing over. A 2018 narrative review of dietary cholesterol and serum lipids states that one large egg yolk contains approximately 200 mg of cholesterol Blesso 2018; the AHA advisory states that a large egg contains about 186 mg Carson 2020.

The cap rested on a chain of reasoning: cholesterol in food is the same molecule as cholesterol in blood, so eating less of it should lower LDL cholesterol, and lower LDL should mean fewer cardiovascular events. The second half of that chain is not disputed by the sources used here — the advisory states plainly that "LDL cholesterol concentration is a stronger predictor of CVD risk than total cholesterol" Carson 2020. It is the first half, food to blood, that turns out to be shakier than the label arithmetic implies. The advisory's own meta-regression of controlled feeding studies, in which the ratio of polyunsaturated to saturated fat was matched across comparison diets, found that dietary cholesterol significantly increased total cholesterol but that "the findings were not significant for the stronger predictor of CVD risk, LDL cholesterol, or HDL cholesterol concentration" Carson 2020. The same advisory reports that two other meta-analyses, of studies it describes as having substantial heterogeneity, did find that dietary cholesterol increased total and LDL cholesterol Carson 2020. Contested, in other words, rather than settled in either direction.

What your body does with the cholesterol you eat

Only a minority of the cholesterol in circulation arrives on a fork. The 2018 review reports that only about 25% of serum cholesterol in humans is derived from the diet, the rest from biosynthesis, and that the average 70 kg adult synthesises about 850 mg of cholesterol a day Blesso 2018. That internal supply is regulated rather than fixed. The same review describes feedback control that limits how much cholesterol we absorb and synthesise while increasing how much we excrete, with the consequence that "most individuals have a marginal change in serum cholesterol in response to dietary cholesterol" Blesso 2018. That is why the label arithmetic misleads: 200 milligrams eaten is not 200 milligrams added to the bloodstream, because the system handling it adjusts.

Compensation is not uniform, and the review is precise about the variation. It classifies people who compensate by reducing cholesterol biosynthesis, absorption and excretion as hypo-responders, and those with a much larger increase — defined as at least a 2.3 mg/dL rise in serum cholesterol per 100 mg of dietary cholesterol — as hyper-responders or non-compensators Blesso 2018. Its abstract puts the minimal-response group at roughly two-thirds of the population, and says that those with a significant response "increase both LDL and HDL-cholesterol, typically with a maintenance of the LDL/HDL cholesterol ratio" Blesso 2018. In one study it summarises, 15 of 40 healthy men classified as hyper-responders did show a significant rise in the LDL-C/HDL-C ratio on three eggs a day for 30 days, but the mean ratio, 2.33, still sat inside what the review calls the optimal range of below 2.5 Blesso 2018. So the responder literature, as that review presents it, does not by itself hand hyper-responders a reason for alarm; it hands them a reason to know their own numbers.

Two things about that source belong in the open. It is a narrative review rather than fresh measurement, and both authors disclose prior funding from the Egg Nutrition Center while stating that the sponsor had no role in the interpretation of data or the writing of the manuscript Blesso 2018. The population studies below are independent of it.

What the egg cohorts found

The question that matters to a reader is not what a lipid panel does over five weeks but what happens to people over decades. Hu and colleagues used two prospective cohorts — the Health Professionals Follow-up Study from 1986 to 1994 and the Nurses' Health Study from 1980 to 1994 — covering 37,851 men aged 40 to 75 and 80,082 women aged 34 to 59 who were free of cardiovascular disease, diabetes, high cholesterol and cancer at the outset. They documented 866 incident coronary events and 258 strokes in men over eight years, and 939 coronary events and 563 strokes in women over fourteen. After adjustment they found "no evidence of an overall significant association between egg consumption and risk of CHD or stroke in either men or women", with relative risks bouncing around 1.0 across intake categories and trend tests of P=0.75 in men and P=0.95 in women. Their own conclusion was that consumption of up to one egg a day is unlikely to have a substantial overall impact on coronary or stroke risk in healthy men and women Hu 1999.

A 2013 dose-response meta-analysis in the BMJ pooled eight articles yielding 17 reports — nine for coronary heart disease, eight for stroke — covering 3,081,269 person-years and 5,847 coronary events, and 4,148,095 person-years and 7,579 strokes. For each extra egg a day the summary relative risk was 0.99 (95% confidence interval 0.85 to 1.15) for coronary heart disease and 0.91 (0.81 to 1.02) for stroke, with no heterogeneity between studies and no sign of a curvilinear dose-response. The same paper reports a subgroup finding rarely quoted alongside the others: people with higher egg consumption had a 25% lower risk of developing haemorrhagic stroke (0.57 to 0.99), which the authors say warrants further study Rong 2013.

In 2020 a group analysed three large US cohorts — 83,349 women, 90,214 women and 42,055 men, all free of cardiovascular disease, type 2 diabetes and cancer at baseline — over up to 32 years and more than 5.54 million person-years, identifying 14,806 incident cardiovascular events. Eating at least one egg a day, versus less than one a month, carried a hazard ratio of 0.93 (0.82 to 1.05). Their updated meta-analysis, 33 risk estimates across 1,720,108 participants and 139,195 events, put each additional egg a day at a pooled relative risk of 0.98 (0.93 to 1.03) Drouin-Chartier 2020.

Two caveats belong with those reassuring numbers, and both come from the paper itself. First, high intakes were rare: in 1998-99 only 2,524 of 203,364 participants (1.24%) ate at least one egg a day, and 414 (0.20%) ate at least two, with most people eating between one and fewer than five eggs a week Drouin-Chartier 2020. Second, eggs travel in company. The authors report that higher egg intake went with higher intakes of calories, unprocessed red meat, bacon, other processed meats, refined grains, potatoes, full-fat milk, coffee and sugar-sweetened beverages, and that higher-intake participants had a higher body mass index and were less likely to be on statins Drouin-Chartier 2020. The AHA advisory makes the same point in one line: the possibility that observational data on eggs are confounded by other dietary components must be considered Carson 2020.

That confounding also cuts in the other direction. In the 2020 meta-analysis, egg intake was not associated with cardiovascular risk in US cohorts (1.01, 0.96 to 1.06) or European cohorts (1.05, 0.92 to 1.19) but was inversely associated in Asian cohorts (0.92, 0.85 to 0.99) — though the test for interaction between regions was P=0.07, which is not a significant difference at conventional thresholds. The authors offer three candidate explanations: that in Asian cultures eggs are incorporated into varied cuisines whereas in Western populations they are typically eaten with red and processed meats and refined grains; that absolute intake was lower, with the highest-intake group averaging 0.76 egg a day; and that the inverse association could reflect a social gradient Drouin-Chartier 2020.

The analysis that cuts the other way

The cohort literature cited here is not unanimous. A 2019 analysis pooled individual participant data from six prospective US cohorts collected between 1985 and 2016 — 29,615 adults, mean age 51.6 years, followed for a median of 17.5 years, with 5,400 incident cardiovascular events and 6,132 deaths. Each additional 300 mg of dietary cholesterol a day was associated with a higher risk of incident cardiovascular disease (adjusted hazard ratio 1.17, 1.09 to 1.26; adjusted absolute risk difference over follow-up 3.24%, 1.39% to 5.08%) and of all-cause mortality (1.18, 1.10 to 1.26). Each additional half egg a day carried an adjusted hazard ratio of 1.06 (1.03 to 1.10) for cardiovascular disease and 1.08 (1.04 to 1.11) for death Zhong 2019.

One detail inside that paper matters more than the headline. The egg associations were "no longer significant after adjusting for dietary cholesterol consumption" — 0.99 (0.93 to 1.05) for cardiovascular disease and 1.03 (0.97 to 1.09) for mortality — which points at the cholesterol rather than at the egg as the operative exposure Zhong 2019. The authors do not present their work as an outlier to be filed away; they conclude that the results "should be considered in the development of dietary guidelines and updates" Zhong 2019.

They are also candid about limits. The paper states that it "relied on single measurement of egg and dietary cholesterol consumption", so exposure misclassification may be of concern, and that residual confounding "was likely" despite adjustment for a number of covariates Zhong 2019. It would be convenient to treat the single baseline dietary assessment as the reason to discount the result, and this article will not do that, because the 2020 meta-analysis tested exactly that question: pooled estimates were effectively identical for studies using baseline-only assessment (0.99, 0.93 to 1.05) and repeated measurements (0.97, 0.88 to 1.06), with a test for interaction of P=0.56 Drouin-Chartier 2020. The honest reading is that these are observational studies pointing in different directions at small effect sizes.

What a randomised trial says about LDL

One 2025 study puts the food-to-blood question to a randomised test rather than an observational one. In a randomised controlled cross-over trial, 61 adults (mean age 39, mean body mass index 25.8) with baseline LDL cholesterol below 3.5 mmol/L each followed three isocaloric diets for five weeks apiece: a high-cholesterol, low-saturated-fat diet including two eggs a day (600 mg cholesterol, 6% of energy from saturated fat); a low-cholesterol, high-saturated-fat diet with no eggs (300 mg, 12%); and a high-cholesterol, high-saturated-fat control including one egg a week (600 mg, 12%). Fifty-four participants completed at least one diet phase and 48 completed all three. Across all diets, saturated fat intake was positively correlated with LDL cholesterol (β=0.35, P=0.002) whereas dietary cholesterol was not (β=-0.006, P=0.42). The authors' conclusion is blunt: "Saturated fat, not dietary cholesterol, elevates LDL cholesterol" Carter 2025.

The same authors flag a counterweight that a fair summary has to carry. Compared with the control diet, the egg diet — but not the egg-free diet — reduced concentrations of large LDL particles (β=-48.6, P=0.03) and increased concentrations of small LDL particles (β=95.1, P=0.004), and they write that the apparent benefit "may be mitigated, at least in part, by a reduction in less-atherogenic large LDL particles and an increase in more atherogenic small LDL particles" Carter 2025. The trial also has the boundaries of its design: five weeks per arm, participants with normal cholesterol at baseline, and lipid measurements rather than heart attacks as the outcome. The journal published a corrigendum to the paper later in 2025, which anyone quoting its numbers should read alongside it.

For fat quality more broadly, the reference point in this article is a meta-analysis of 60 controlled trials whose primary outcome was the ratio of total to HDL cholesterol rather than LDL. It found that the improvement in that ratio from replacing trans fatty acids with a mix of carbohydrates and cis unsaturated fatty acids was almost twice as large as from replacing saturated fatty acids; that the ratio did not change when carbohydrates replaced saturated fat but fell when cis unsaturated fats did; and that individual saturated fats behave differently, with lauric-acid-rich oils actually decreasing the ratio, myristic and palmitic acids having little effect on it, and stearic acid reducing it slightly. Its own summary of what to do is that risk is reduced most effectively when trans and saturated fatty acids are replaced with cis unsaturated fatty acids — paired with an explicit caution that effects on these markers "should not in themselves be considered to reflect changes in risk" but should be confirmed by prospective studies or clinical trials, and a warning that effects on the total-to-HDL ratio "may differ markedly from their effects on LDL" Mensink 2003.

The pattern-level evidence points the same way. In PREDIMED, 7,447 people aged 55 to 80 at high cardiovascular risk but without cardiovascular disease at enrolment were assigned to a Mediterranean diet with extra-virgin olive oil, a Mediterranean diet with mixed nuts, or a control diet of advice to reduce dietary fat; after a median 4.8 years the trial stopped at a prespecified interim analysis, with hazard ratios of 0.69 (0.53 to 0.91) and 0.72 (0.54 to 0.95) against the control diet Estruch 2018. That 2018 paper exists because the earlier report was withdrawn: the investigators identified protocol deviations including enrolment of household members without randomisation and assignment without randomisation at some sites, and their revised estimates come from an adjusted intention-to-treat analysis with propensity scores, with similar results after omitting the 1,588 participants whose assignments were known or suspected to have departed from protocol Estruch 2018. It tested nothing about eggs, its population was older and high-risk, and its comparator was another dietary pattern — but it is the kind of evidence that moves attention from single foods to the shape of the whole diet. Our guide to what the Mediterranean pattern actually involves covers what that looks like on a plate.

The diabetes question

Diabetes is the subgroup where the observational signals recur. It is not, on the evidence cited here, a consistent finding, and the difference matters because this is the one place the literature is used to single out a clinical group.

In the 1999 cohorts the association reached significance in men only: comparing more than one egg a day with less than one a week, the relative risk of coronary heart disease among diabetic men was 2.02 (1.05 to 3.87, P for trend 0.04), while among diabetic women it was 1.49 with a confidence interval of 0.88 to 2.52 — an interval that includes 1, so the point estimate alone does not establish an association, although the trend test was significant at P=0.008. Participants were free of diabetes at entry, so these subgroups formed during follow-up, and the authors' own verdict was that the apparent increased risk "warrants further research" Hu 1999. The 2013 meta-analysis reported a relative risk of 1.54 (1.14 to 2.09) for coronary heart disease comparing highest with lowest intake in diabetic populations, and likewise said it warranted further study Rong 2013.

Against that, the 2020 analysis found no association at all in its own three cohorts: among participants with type 2 diabetes, at least one egg a day versus less than one a month gave a hazard ratio of 1.06 (0.81 to 1.39). Its diabetes-restricted meta-analysis returned 1.25 (0.99 to 1.59) per extra egg a day and 1.40 (1.00 to 1.97) for high versus low intake — neither excluding 1 — with what the authors call considerable heterogeneity between studies Drouin-Chartier 2020. The AHA advisory counts the tally plainly: of three studies specifically addressing type 2 diabetes, eggs and coronary risk, two reported a positive association Carson 2020. And the mechanism is not obvious, because the 2018 review reports that people with insulin resistance or diabetes tend to have a weaker serum cholesterol response to eggs than leaner, insulin-sensitive people, consistent with reduced cholesterol absorption efficiency in obesity and metabolic syndrome Blesso 2018.

Diabetes is not the only group the advisory flags. It reports that three studies found an association between egg intake and increased risk of heart failure in the general population, and its clinical guidance names patients with dyslipidaemia, "particularly those with diabetes mellitus or at risk for heart failure", as the people who should be cautious in consuming cholesterol-rich foods Carson 2020. Anyone with an inherited lipid disorder or established cardiovascular disease is in a different conversation again: that diet sits inside a treatment plan and belongs with the clinician managing it, not with an article. For everyone curious about their own physiology, a lipid panel arranged through a clinician before and after a sustained change in intake answers a question no population average can.

What the guidelines actually say

It is easy to caricature the current position in either direction. The AHA advisory does not set a numerical milligram target, and its stated reason is practical: "A recommendation that gives a specific dietary cholesterol target within the context of food-based advice is challenging for clinicians and consumers to implement; hence, guidance focused on dietary patterns is more likely to improve diet quality and to promote cardiovascular health" Carson 2020. Its positive advice is about patterns: heart-healthy patterns such as Mediterranean-style and DASH-style diets are, in its words, "inherently relatively low in cholesterol, with typical levels similar to the current US intake" Carson 2020.

Nor does it decline to talk about eggs. Its suggestions for clinical practice and consumers state that, given the relatively high cholesterol content of egg yolks, "it remains advisable to limit intake to current levels", that "healthy individuals can include up to a whole egg or equivalent daily", and that for older patients with normal cholesterol, "consumption of up to 2 eggs per day is acceptable within the context of a heart-healthy dietary pattern"; lacto-ovo vegetarians who eat no meat-based cholesterol sources may include more dairy and eggs within that same moderation Carson 2020. Read carefully, that is neither "minimise cholesterol" nor "eggs are free" — it is hold roughly steady, and choose the pattern rather than police the number.

The pooled analyses converge on the same boundary from the other side. The 2013 meta-analysis concludes that higher consumption of eggs, up to one a day, is not associated with increased risk of coronary heart disease or stroke Rong 2013, and the 2020 paper concludes that moderate egg consumption, again up to one a day, is not associated with cardiovascular disease risk overall Drouin-Chartier 2020. That one-egg figure is not merely an artefact of what the cohorts happened to measure; it is what three separate sets of authors affirmatively recommend. Its flip side is that the studies cited here say little about much higher intakes: in the 2020 cohorts only 0.20% of participants ate two or more eggs a day Drouin-Chartier 2020, so "an egg a day looks fine" does not stretch to six.

What to do with your breakfast

The practical shift is a change of target rather than a licence. Counting yolks is a low-yield activity for most healthy adults, because the compensation the 2018 review describes blunts the effect Blesso 2018 and the cohort evidence at ordinary intakes is close to null Drouin-Chartier 2020. What the randomised evidence cited here identifies as the lever on LDL is saturated fat rather than dietary cholesterol Carter 2025, and the substitution that the controlled-trial meta-analysis endorses is replacing trans and saturated fats with cis unsaturated ones Mensink 2003. In practice that puts the attention on what the eggs are cooked in and what shares the plate with them.

If eggs are simply how protein gets into your morning, that is a reasonable thing to want. Whether the timing of that protein carries the weight it is often given is a separate question, which we have looked at in our piece on what breakfast protein timing does and doesn't do. On the cholesterol question specifically, the fairest summary of the papers above is that the egg has been carrying a good deal of blame for the company it keeps — and that the two exceptions the guidance actually names are worth knowing if one of them is you.

Frequently asked questions

How many eggs a day is safe?

The American Heart Association's 2020 science advisory declines to set a numerical milligram cap, but it does give an egg figure: healthy individuals can include up to a whole egg or equivalent daily, and it adds that for older patients with normal cholesterol, up to two eggs a day is acceptable within a heart-healthy dietary pattern. It also says that, given the cholesterol content of yolks, it remains advisable to limit intake to current levels. Two pooled analyses reach the same practical boundary, concluding that consumption up to about one egg a day is not associated with cardiovascular disease risk overall.

Does dietary cholesterol raise blood cholesterol at all?

Less than the label arithmetic implies, and the size of any effect on LDL is contested. A 2018 narrative review describes feedback control that limits how much cholesterol the body absorbs and makes while increasing how much it excretes, with the result that most individuals show only a marginal change in serum cholesterol; it puts the minimal-response group at about two-thirds of the population. The American Heart Association's own meta-regression of controlled feeding studies found dietary cholesterol significantly raised total cholesterol but not LDL cholesterol, while the same advisory notes two other meta-analyses that did find LDL increases. A 2025 randomised cross-over trial found saturated fat intake correlated with LDL cholesterol while dietary cholesterol did not.

Should people with diabetes limit eggs?

Diabetes is the subgroup where observational signals recur, but they are not consistent. In the 1999 cohort analysis the association with coronary risk reached significance in diabetic men (relative risk 2.02, 95% confidence interval 1.05 to 3.87) but not diabetic women (1.49, 0.88 to 2.52). A 2013 meta-analysis found a raised risk in diabetic populations (1.54, 1.14 to 2.09) and said it warranted further study. A 2020 analysis of three US cohorts found no association among participants with type 2 diabetes (1.06, 0.81 to 1.39), and its diabetes-restricted meta-analysis estimates were not statistically significant, with considerable heterogeneity between studies. The American Heart Association advises that patients with dyslipidaemia, particularly those with diabetes or at risk for heart failure, be cautious with cholesterol-rich foods, which makes this a conversation to have with your own clinician.

Are egg whites better than whole eggs?

Whites carry protein and essentially none of the cholesterol, which sits in the yolk: a 2018 review puts one large egg yolk at roughly 200 milligrams, while the American Heart Association's advisory gives about 186 milligrams for a whole large egg. But if the goal is lower LDL, a 2025 randomised cross-over trial found that saturated fat, not dietary cholesterol, tracked with LDL cholesterol. Switching to whites while leaving the rest of the plate unchanged may miss the larger lever.

What matters more than eggs for my cholesterol numbers?

Fat quality and the overall dietary pattern. A 2025 randomised cross-over trial concluded that saturated fat, not dietary cholesterol, elevates LDL cholesterol. A meta-analysis of 60 controlled trials, whose main outcome was the ratio of total to HDL cholesterol, concluded that risk is reduced most effectively when trans and saturated fatty acids are replaced with cis unsaturated fatty acids, while cautioning that changes in these markers should not in themselves be read as changes in risk. And in PREDIMED, a trial in people aged 55 to 80 at high cardiovascular risk, fewer major cardiovascular events occurred among those assigned a Mediterranean diet with extra-virgin olive oil or nuts than in a control group advised to reduce dietary fat.

Did the guidelines really drop the cholesterol limit?

The numerical cap is gone. The American Heart Association's 2020 advisory records that the Dietary Guidelines for Americans previously recommended limiting cholesterol to no more than 300 milligrams a day, and that the 2015 Dietary Guidelines Advisory Committee did not bring that recommendation forward. The advisory's own position is that a specific numerical target is challenging for clinicians and consumers to implement, so guidance shifts to heart-healthy dietary patterns that are inherently relatively low in cholesterol, at levels similar to current average US intake. It also notes that the 2015 report was explicit that the lack of a stated limit did not imply that dietary cholesterol is unimportant.

References