“Follow the science” sounds like a single instruction. In practice it means following a moving target, one that sometimes moves because better evidence arrived and sometimes because the earlier picture was incomplete.
Dietary fat is the clearest example. For decades, federal guidance urged Americans to limit it. In January 2026, the new federal guidance said to include full-fat dairy when consuming dairy, named olive oil first among cooking fats with butter or beef tallow as other options, and still said that, in general, saturated fat should not exceed 10% of daily calories.[1] Was the old advice wrong, or was the evidence incomplete? The answer is different for each rule, and that difference is the skill worth learning.
If you are performance-driven, you probably run on nutrition rules you inherited from somewhere: a coach, a forum, a guideline. Before this series turns to the heart itself, this first article asks a simpler question: how do you tell when a rule has earned your trust, and when it deserves a second look?
This series takes one question at a time and separates six things: what we were taught, what the evidence was then, what changed, what the evidence supports now, what we still don't know, and what to ask a licensed clinician. We push on assumptions by asking sharper questions, not by overstating answers.
What we were taught
In 1980, the U.S. Departments of Agriculture and Health and Human Services released the first Dietary Guidelines for Americans: seven principles for healthy people age two and older. The third read, “Avoid too much fat, saturated fat, and cholesterol.”[2]
Notice what it did not do. USDA's own history describes those guidelines as directional, not quantitative; no number defined “too much.” The release also drew concern from consumer, commodity and food-industry groups, and from nutrition scientists who questioned the causal link between some guidelines and health.[2]
Numbers arrived later. The 1990 edition changed the wording to “Choose a diet low in fat, saturated fat, and cholesterol” and, for the first time, suggested goals of 30 percent or less of calories from fat and less than 10 percent from saturated fat.[2]
What the evidence was then
By the 1970s, USDA notes, a growing body of research had connected high intakes of fat, saturated fat, cholesterol and sodium with heart disease and stroke. A 1977 Senate committee published numeric dietary goals; USDA chose not to build its food guides on them because the diets they implied differed so much from usual eating.[2]
The 1980 guidelines gave directional advice; by 1990, that advice had become quantitative.
One randomized trial from the era shows how incomplete the picture could be. The Minnesota Coronary Experiment (1968–73) randomized more than 9,000 residents of a nursing home and six state mental hospitals, cutting their saturated fat about in half by substituting corn oil and corn-oil margarine. Among the roughly 2,350 who stayed on the diets for at least a year, cholesterol fell about 14 percent, versus about 1 percent in the control group. Yet life tables from a 1981 thesis, and the trial's 1989 report, showed no survival benefit. Other analyses the original investigators had planned, including autopsy findings, were never published until researchers recovered the records decades later. Those researchers argued that incomplete publication has contributed to overestimating the benefits, and underestimating the potential risks, of swapping saturated fat for vegetable oils rich in linoleic acid. Their work had limits: participants were followed only while hospitalized, only about a quarter stayed a year or longer, and the full dataset could not be recovered.[6]
One trial does not overturn a field. Another pooled analysis, which included the published Minnesota report along with seven other trials and counted heart attacks and coronary deaths, found fewer coronary events when polyunsaturated fat replaced saturated fat: eight trials, 13,614 participants, 19 percent lower risk (RR 0.81, 95% CI 0.70–0.95). Its authors cautioned that the trials had design limitations; the two double-blind trials alone showed no significant difference (RR 0.91, 95% CI 0.63–1.31). Both results are in the literature, and the disagreement is instructive.[7]
What changed
Larger, longer tests. The Women's Health Initiative randomized 48,835 postmenopausal women to a low-total-fat eating pattern. After about eight years there was no significant effect on composite coronary heart disease (heart attack, coronary death or revascularization; hazard ratio 0.97, 95% CI 0.90–1.06), stroke or cardiovascular disease overall. But the trial's main aim was cancer prevention, and the diet change was modest: compared with the comparison group, fat intake was lower by about 8 percentage points of energy, saturated fat by only about 3, and LDL cholesterol by under 4 mg/dL in the blood subsample. The authors estimated about 40 percent statistical power for heart outcomes, and no part of the intervention specifically targeted saturated fat. It tested a low-total-fat pattern as practiced, not a strict saturated-fat swap.[5] A later report from the same trial, with longer follow-up, found the same overall result but also that coronary risk varied with women's baseline health. Among women without prior cardiovascular disease and with normal blood pressure, coronary risk was lower during the intervention period (hazard ratio 0.70, 95% CI 0.56–0.87), partly offset by more ischemic strokes; among women with high blood pressure it was not (1.04, 0.90–1.19), and the differences faded after the intervention ended. The authors caution that results in women with hypertension or prior cardiovascular disease are likely confounded by later use of cholesterol-lowering medication.[14]
A more specific question. Pooled reviews began asking not “less fat?” but “less saturated fat, replaced with what?” A 2020 Cochrane review of 15 randomized trials (56,675 participants, at least two years each) found that reducing saturated fat lowered the risk of combined cardiovascular events by 17 percent (risk ratio 0.83, 95% CI 0.70–0.98), which Cochrane's authors rated moderate-certainty evidence, with substantial variation between trials (I² 67%). It found little or no effect on all-cause mortality (RR 0.96) or cardiovascular mortality (RR 0.95). Bigger reductions in saturated fat, and in cholesterol, went with bigger benefits.[3]
The American Heart Association's 2017 advisory reported a larger figure, about 30 percent fewer cardiovascular events, from what it called its core trials, those that replaced saturated fat with polyunsaturated vegetable oil. It noted that adding trials weakened by short duration, low adherence or trans-fat use partly diluted that effect.[4] The gap likely reflects which trials each review counted and how they were pooled, which is why “which studies?” belongs in every health claim.
The replacement question. The AHA advisory concluded that replacing saturated fat with mostly refined carbohydrates and sugars was not associated with lower cardiovascular rates.[4] Cochrane's subgroup analyses did not find a significant difference between replacing with polyunsaturated fat or with carbohydrate, while noting very little data on other replacements.[3] Two careful reviews weigh this differently. We will not paper over that.
A note on sugar. In 1965 the Sugar Research Foundation, a sugar-industry trade group, funded a literature review that ran in 1967 in the New England Journal of Medicine. Its internal correspondence shows it stated the review's objective, sent articles to the authors and received drafts; its funding and role were not disclosed. The historians who found these documents say there is no direct evidence the industry wrote or changed the manuscript, and that the evidence it shaped the review's conclusions is circumstantial. The documents also cover one review, not the national guidelines that followed. Article 4 examines the story in full.[8]
What the evidence supports now
Here is the strongest sentence we can defend. In randomized trials lasting at least two years, sustained reductions in saturated fat were associated with fewer combined cardiovascular events and little effect on overall mortality; reviews differ on how much the replacement food matters.[3][4] Two large U.S. prospective cohorts of health professionals (more than 127,000 people) found lower coronary risk when polyunsaturated fat replaced saturated fat, though observational data cannot prove cause.[9] A low-total-fat pattern by itself did not significantly change overall heart outcomes in a large randomized trial of that approach, one whose authors estimated limited power to detect a difference, though longer follow-up found that coronary risk differed by women's baseline blood pressure and cardiovascular history.[5][14]
Federal guidance now points in more than one direction. The 2025–2030 Dietary Guidelines say that, in general, saturated fat should not exceed 10% of daily calories. They also say to include full-fat dairy when consuming dairy, to prioritize oils such as olive oil with butter or beef tallow as other options, and that more high-quality research is needed on which types of fat best support long-term health.[1] The American Heart Association welcomed the guidelines' emphasis on vegetables, fruits, whole grains and limiting added sugars, refined grains, highly processed foods and saturated fats, but said it was concerned about the red-meat and salt-seasoning advice. It continues to encourage low-fat and fat-free dairy and urges limiting red meat, butter, lard and tallow.[10] Its March 2026 scientific statement again advises choosing unsaturated fat sources in place of saturated-fat sources.[11] The federal advisory committee's own review found strong evidence that swapping butter for unsaturated plant oils lowers LDL cholesterol, limited evidence on heart-disease outcomes for that swap, and no firm conclusion on several questions about dairy and meat fat.[12] The appendices published with the federal guidelines say that randomized trials increasing linoleic acid while replacing saturated fat have not shown clinical benefit, and that much of the evidence favoring high linoleic acid intake comes from observational studies and from trials that relied on surrogate markers, particularly LDL cholesterol.[13] These assessments overlap in places, diverge in others, and leave real questions open, including which fat sources matter most for heart outcomes.
What we still don't know
- Whether it applies to you. Trials report averages. Your baseline risk, lipids, family history and genetics shape what a given change means, and these trials cannot tell you that.
- Whether it applies to athletes. Nothing we reviewed tested saturated-fat swaps in physique or performance athletes, whose diets and training loads can differ sharply from the general populations studied. Missing evidence is not evidence of safety, and it is not evidence of harm.
- Which foods, not just which nutrients. The current guidelines themselves call for more research on which fats support long-term health.[1]
- How much guidance shifted for scientific versus other reasons. The documents we cite show that evidence, industry input and public concern all existed. They do not measure the weight of each.
How to read a health claim
- Who was studied? Healthy adults, patients, women, athletes?
- Compared to what? “Less fat” replaced with sugar is a different experiment from replaced with olive oil.
- For how long, and measuring what? Heart attacks and strokes, or a lab number?
- Who paid or wrote it, and was that disclosed?
- Do pooled reviews agree? One study is a data point; a pattern across studies is stronger.
Questions for your licensed clinician
- Which of my results matter most for my own cardiovascular risk?
- Does my family history change how you read my cholesterol and other lab results?
- Given how I train and eat, would you interpret anything differently?
- If I changed the fats in my diet, what would you want to recheck, and when?
- What new evidence would change your recommendation?