Cholesterol
The nutrient we got wrong for fifty years
Description
In 1958, an American physiologist named Ancel Keys boarded a series of flights that took him to Italy, Greece, Yugoslavia, the Netherlands, Finland, Japan, and the United States. He was running what would become the Seven Countries Study, the most influential nutritional epidemiology project of the twentieth century. Keys had a hypothesis: that saturated fat raised serum cholesterol, and serum cholesterol caused heart attacks. His data, gathered from villages and small towns over twenty years, showed correlations consistent with that hypothesis. By the time the major reports were published in the 1970s, the diet-heart hypothesis had become public health orthodoxy. The American Heart Association recommended reducing saturated fat. The federal dietary guidelines followed. Eggs, butter, and red meat were reframed as cardiac risks.
The half-century that followed produced a strange combination of progress and slowly accumulating embarrassment. Cardiovascular mortality fell substantially, partly because of statins, partly because of declining smoking rates, partly because of better acute care. But the dietary advice that started with Keys turned out to be more complicated than anyone had told the public. Saturated fat was not a single thing with a single effect. Dietary cholesterol — the cholesterol in eggs and shellfish — turned out to have almost no impact on blood cholesterol for most people.
By the mid-2010s, the dietary guidelines had quietly walked back several headline claims. The 2015 US Dietary Guidelines dropped the long-standing limit on dietary cholesterol. Meta-analyses found weaker associations between saturated fat and heart disease than had been assumed. The official position on butter, whole milk, and red meat shifted from clear prohibition to muted ambivalence. The half-century of dietary advice on cholesterol ended not with a dramatic reversal but with a series of slow clarifications.
The question we're asking: how did the diet-heart hypothesis become guidance, what did fifty years of data show, and what do we actually know now?
What we'll see: Keys's study, the saturated-fat consensus, the slow reversal, and what's still genuinely contested.
Table of contents
01Keys and the seven countries
Ancel Keys was not a cardiologist. He was a physiologist trained at Berkeley and Cambridge, who had run the famous Minnesota Starvation Experiment during World War II and developed the K-ration that fed American soldiers. His turn to cardiovascular epidemiology came after the war, when he noticed that American businessmen were dying of heart attacks at unusual rates while postwar Europeans, on far less abundant diets, seemed to be dying of them less. The observation drove the design of the Seven Countries Study: pick populations with widely varying diets, measure what they eat and what their cholesterol levels are, follow them over decades, and see who has heart attacks.
The data came in slowly and looked supportive of his hypothesis. Finnish villagers, who ate large amounts of butter and dairy fat, had high cholesterol levels and high cardiac mortality. Japanese fishermen, who ate little saturated fat and less total fat, had low cholesterol and low cardiac mortality. Mediterranean populations sat in the middle. The country-level correlations were strong. Keys argued they implied a causal chain: saturated fat raises cholesterol, cholesterol drives atherosclerosis, atherosclerosis causes heart attacks. The argument was elegant and seemed to fit the data.
02The consensus that became guidance
The translation from research consensus to public health guidance happened in stages. The American Heart Association issued its first statement recommending reduced saturated fat intake in 1961. The 1977 Dietary Goals for the United States, drafted by a Senate committee chaired by George McGovern, recommended that Americans reduce total fat to 30% of calories and saturated fat to 10%, increase carbohydrate consumption, and limit cholesterol to 300 milligrams per day. The 1980 federal Dietary Guidelines for Americans formalized the recommendations and have been reissued every five years since.
The food industry responded to the new guidance with a wave of product reformulation. Margarines replaced butter. Vegetable oils replaced lard. Low-fat dairy replaced whole milk. The processed food industry produced a generation of low-fat products in which the missing fat was replaced with sugar, refined starches, and additives — Snackwell's cookies, low-fat yogurts, fat-free cakes. Per-capita sugar consumption in the United States rose substantially through the 1980s and 1990s. Obesity and type 2 diabetes rose with it. The unintended consequences of the dietary shift would take decades to become visible.
03The slow reversal
The first cracks in the dietary consensus came from trial data itself. The Women's Health Initiative Dietary Modification Trial, which followed nearly 49,000 postmenopausal women randomized to a low-fat or control diet for eight years, was published in JAMA in 2006. It found no significant reduction in cardiovascular events, breast cancer, or colorectal cancer. The result was framed as a partial failure rather than a refutation, but it was the largest dietary RCT ever conducted, and it had not produced the expected outcome.
The 2010s saw a wave of meta-analyses re-examining the saturated fat hypothesis. A 2010 paper in the American Journal of Clinical Nutrition by Patty Siri-Tarino and colleagues pooled twenty-one prospective cohort studies and found no significant association between saturated fat intake and cardiovascular disease. A 2014 meta-analysis by Rajiv Chowdhury in the Annals of Internal Medicine reached similar conclusions. Both papers were attacked methodologically, but the underlying claim — that the relationship between saturated fat and heart disease was weaker than the guidelines implied — accumulated supporting evidence rather than the reverse.
04What we know now, and what is still contested
The current scientific consensus on cholesterol and diet is harder to state cleanly than the 1980 version, mostly because it acknowledges more uncertainty. LDL cholesterol is causally implicated in atherosclerosis Mendelian randomization studies, which use genetic variants to estimate causal effects, support this strongly. Lowering LDL with statins reduces cardiovascular events in established disease and in high-risk primary prevention. Beyond that, almost everything is qualified.
Saturated fat raises LDL cholesterol on average, but the magnitude is small for most people, and not all saturated fats behave identically stearic acid, abundant in beef, has a smaller effect on LDL than palmitic acid, abundant in dairy. The food matrix matters: cheese and yogurt appear to have weaker associations with cardiovascular disease than the saturated fat content alone would predict. The replacement question is also unresolved: replacing saturated fat with refined carbohydrates does not appear to reduce cardiac risk, while replacing it with polyunsaturated fats appears to.
05Conclusion
Ancel Keys was not entirely wrong. Cholesterol matters, the lipid hypothesis is in its core form correct, and statins prevent heart attacks. What he and the public health apparatus around him got wrong was the precision of the claim. The translation from a partial scientific finding to clear-cut dietary advice eat less saturated fat, eat fewer eggs, replace butter with margarine happened faster than the evidence justified, and the corrections have come more slowly than the original guidance.