Everybody knows that you are what you eat. What everybody may not know is that we've only known this for about 50 years.
People have long recognized that some things are bad for them, like toadstools and rotten meat. For a long time they've also known that some things were necessary for good health, even if it wasn't entirely clear what those things were.
In 1747, James Lind showed that citrus fruit prevented scurvy in seafarers. In the years before World War I, Joseph Goldberger proved that a dietary deficiency was responsible for pellagra, a scabrous disease common in the South's abject poor. The "some things" in those cases -- each discovered after those men's deaths -- were vitamin C and niacin, respectively.
What wasn't a bit obvious until fairly recently is that a perfectly healthy person's diet today can be responsible for his death decades in the future. Food just wasn't considered to be all that important in determining health, assuming a person had it in enough quantity and variety. Nor did scientists (let alone the eating public) know much about the biological economy of food. With the exception of certain vitamins and micronutrients, scientists couldn't describe with much precision the effect that too much of this or not enough of that had on the human body.
Today, the fruits of this knowledge are everywhere -- in our diets, recipes and food labels, in expert advice, family habits and cafeteria-line guilt. That the connection between eating and health is now so clear is testimony to the work of thousands of researchers toiling in dozens of countries over the last half-century. No single person is responsible for this revolution. But if one were asked to name someone whose work and career best exemplifies it, there would be little doubt about the choice. It would be Ancel Keys.
Like most biologists, Keys isn't widely known. His name means something in the fields of nutrition and epidemiology. He made the news occasionally in the postwar decades, achieving in 1961 what was once America's equivalent of knighthood: his picture on the cover of Time magazine for his work on diet and heart disease. Overall, though, he has lived and worked well below the public's radar.
That work has helped build the edifice of knowledge about what constitutes a healthy diet. It's an edifice that's still under construction as the United States faces what may well be the new century's biggest public health problem, epidemic obesity.
Keys, who is 98 and lives, in frail health, with his wife, Margaret, in Minneapolis, did his major research in an effort to understand last century's epidemic.
Heart disease became the United States' leading cause of death in 1921. Mortality from it rose steadily until 1950. The postwar boom years were a time when it seemed that for American men -- especially ones in gray flannel suits -- a heart attack was virtually a rite of passage.
Keys helped reveal the biology and epidemiology of coronary heart disease in two seminal studies that examined large numbers of men and then observed them for a long period of time.
The first was a study of Minnesota businessmen launched in 1947, a few months before the better-known Framingham Heart Study began. More important was the Seven Countries Study, which since 1958 has followed a sample of men in 16 distinct populations in seven nations in North America, Europe and Asia. (The latest report of that study is being published in book form this month.)
In the two studies, Keys showed that bloodstream cholesterol was the chief determinant of heart disease and that there was a list of lesser risk factors as well, such as hypertension and smoking. He further showed that saturated fat in the diet -- butter, red meat, fried food -- was the chief determinant of bloodstream cholesterol. And he showed something that before him was only dimly perceived -- namely, that an entire population could be "sick," in the sense that nearly all its members were at higher risk of early death than people in another population.
Many other researchers and projects -- the Framingham study most notably -- contributed to this understanding. Nevertheless, it is difficult to overestimate the importance of Keys's three observations.
They laid out the main causes of the country's biggest health problem. They suggested, indirectly at least, that a solution to the epidemic of heart attacks might lie in lowering individuals' cholesterol and changing the nation's diet. By comparing populations, Keys and his colleagues were able to produce something essential to a public health strategy -- an estimate of how much of the burden of heart disease might be the product of things that could be changed.
"Ancel Keys is really a major figure," says Walter C. Willett, chairman of the department of nutrition at the Harvard School of Public Health. "His . . . was a fundamental insight. It indicated that heart disease was not inevitable."
Keys's work on cardiovascular epidemiology, however, is only one of several accomplishments in an astoundingly long and productive career.
In the 1930s and early 1940s, he made major contributions to basic physiology. During World War II, he developed the K-ration (the portable meal for field soldiers) and was a key advisor on nutritional matters. Toward the end of the war, he ran a famous project in which conscientious objectors voluntarily endured "semi- starvation" and were studied scientifically.
In the postwar years, he led a large research team that performed experiments to determine the physiological effect of various constituents of the diet. The results, combined with his epidemiological findings, led Keys to become the first American promoter of the now widely touted "Mediterranean diet."
"Keys basically started it all -- clinical studies, laboratory diet experiments, population surveys . . . longitudinal cohort studies and preventive trials," said Henry L. Blackburn, a longtime collaborator who followed Keys as head of the division of epidemiology at the University of Minnesota School of Public Health.
Throughout this career, Keys also had a Zelig-like ability to be at the right place with the right people at the right time. He may be the last survivor of an era when it was still possible to be at the top of several fields of scientific inquiry at the same time.
The Education of a Pioneer
Ancel Keys was born in Colorado Springs in 1904, an only child. The family moved to San Francisco just before the earthquake and fire there in April 1906. His father worked in a print shop and other small businesses. His mother's parents were deaf, a fact that some believe helped foster the pantomime skills of her brother, Lon Chaney, who became a silent-movie star in the early years of Hollywood.
The family moved across San Francisco Bay to Berkeley, where Keys spent most of his childhood. He was a good student and an impulsive one, according to accounts in a privately printed autobiography and a recent interview. As a teenager, he once ran away from home for three months, finding work in the Arizona desert shoveling bat guano out of caves and bagging it for fertilizer. He married at age 19, divorcing several years later.
He went to college at the University of California in Berkeley, taking a punishing menu of science courses. When he failed to win the sole scholarship offered by the chemistry department, he changed his major to economics in a fit of pique. When he graduated, he took a management trainee position at Woolworth's. That didn't last long. He was soon in graduate school in biology, earning a doctorate from the Scripps Institute in La Jolla, Calif.
Keys's first major piece of research was to determine how eels, which migrate from streams to the ocean and back again, regulate the salt content of their blood. The work was done in the Copenhagen laboratory of August Krogh, where Keys had a post-doctoral fellowship in 1930. Krogh had won a Nobel Prize 10 years earlier for showing that capillaries regulate blood flow to frog muscle during exercise. Keys's assignment was difficult and messy, but he eventually proved that gills were the organ of sodium regulation in eels.
After brief stints at Cambridge University in England, Harvard and the Mayo Clinic, in Rochester, Minn., Keys got an appointment at the University of Minnesota, first in the medical school and later in the school of public health. He stayed there for the rest of his career.
At the time, the basic mechanisms of human physiology -- and the equations describing them -- were still being discovered. Keys wasted no time on trivial questions, and he did not miss opportunities. He studied the osmotic properties of blood. He organized an international scientific expedition to the Andes in 1935, where at 20,000 feet biologists did some of the earliest research on the physiological effects of altitude. He made contacts in a state mental hospital and studied insulin's effects on bloodstream potassium in patients undergoing insulin coma treatments for schizophrenia.
His academic output was huge, and he soon assembled so much equipment, and so many students and collaborators, that he was given his own research complex, located under the stands at the university's football stadium. He called it the Laboratory of Physiological Hygiene.
Soon after the start of World War II, the government asked him to design lightweight but calorically and nutritionally adequate rations for paratroops. The reason he was chosen is unclear. He believes someone in the War Department thought his high-altitude research would give him insight into paratroops' needs -- although, of course, they only spent brief periods at altitude.
Curiously, given Keys's background, the ration was assembled with little or no research.
"They sent Colonel Isker from the Quartermaster Corps to tell me what was wanted and some of the limitations. . . . We went to Witt's grocery store here in town and decided the things we had to put in there," Keys recalled recently.
The original ration included hard biscuits, dry sausage and chocolate. Various changes were made after field tests at Fort Benning, Ga., including the addition of chewing gum, toilet paper and four cigarettes with each meal. Soon the rations were being made by the millions.
The Army had C and D rations. The paratrooper rations were renamed and given a letter designation after Gen. George C. Patton requested the same meals for his tank corps, too. One account, by the historian B. Michael Berger, says that the "letter K had no particular significance; it was chosen merely to have a phonetically different letter from the letters C and D." Many other accounts say (and Keys himself avers) that the letter was plucked from his last name.
As the war went on, the Laboratory of Physiological Hygiene did other military research, studying the relationship between diet and fatigue, and determining whether vitamins were lost in sweat (the latter a matter of concern in soldiers fighting in the Pacific).
"By that time, I had an office in the War Department where I had to go every two weeks and stay a few days and answer a lot of foolish questions," Keys recalled recently. "Margaret [his wife] would take me to the train station in St. Paul. I'd go down to Chicago and take the night train to Washington. The saving grace was that the War Department was paying for me to stay in the best hotel in Washington and eat all my meals at the Cosmos Club."
Starving for Knowledge
In 1944, Keys proposed an ambitious project to study the physiology of starvation.
"By that time I realized that there were a lot of people, millions of people, who were in semi-starvation conditions. I wanted to find out what would be the effect of that, how long it would last, and what would be required to bring them back to normal," he recalled.
There had been some small experiments early in the century on the effects of caloric restriction. There had also been studies during the war of undernourished populations -- notably in German-occupied Netherlands and in the Warsaw ghetto -- done by scientists who were members of those groups. Keys, however, proposed an intentional, highly controlled study of starvation.
For his subjects he used 36 conscientious objectors to military service, all volunteers. (More than 100 applied, and 18 others were chosen to be full-time technical assistants in the project.) Most were Quakers, Mennonites or members of the Church of the Brethren, and the support of those churches was enlisted. They were 25 years old, on average. There were many motivations to sign up, but altruism was clearly a near-universal one. Many had hoped to work in war relief efforts, but hadn't been permitted by the government. This was an indirect means to that end. There were personal reasons as well for facing the hazards that Keys and his main collaborator, Henry L. Taylor, made no attempt to minimize.
"For some [it] was an opportunity to participate in . . . activities which more closely paralleled the discomforts, risks and sufferings of the men in the armed forces and the civilians in war-ravaged areas . . . some considered the experiment a means of allaying guilt feelings in regard to their pacifist positions," the researchers wrote.
Many went on to distinguished careers. The best-known is Max M. Kampelman, a lawyer, diplomat and educator who was an arms control negotiator in the Carter and Reagan administrations and is now an emeritus professor at Georgetown University. He was doing alternative service in a home for retarded children in Maine when a request for volunteers went out to all conscientious objectors.
The men lived in the stadium lab from November 1944 through October 1945. They were each brought to a baseline weight in a 12-week run-in period. Then their daily diets were cut from about 3,500 calories to 1,600 calories, with a target of a 25 percent weight loss over 24 weeks.
"The major food items served were whole-wheat bread, potatoes, cereals and considerable amounts of turnips and cabbages. Only token amounts of meats and dairy products were provided. The diet was designed to represent as nearly as possible the type of foods used in European famine areas," Keys and his collaborators wrote. After the period of semi-starvation, each man was assigned to one of three "rehabilitation" diets over 12 weeks. (Information from this part of the experiment was expected to be useful to postwar Europe.)
Much of each day was taken up with tests and a prescribed exercise program that called for the men to walk about five miles a day. All told, the volunteers underwent hundreds of measurements, and dozens of physiological and psychological studies. Some also attended classes. Kampelman worked toward a master's degree in political science as his weight dropped from about 150 pounds to close to 100. He remembers Keys -- whom he saw every day -- as "a quiet, forceful figure. You knew who was in charge. He was very businesslike, and made no effort to be a 'personality kid.' He and everyone on his staff were impressive people."
No one enrolled in the study died or suffered permanent harm. In 1950, "The Biology of Human Starvation" appeared in two volumes, with a total of 1,385 pages. It was too late to do much good for war-wracked Europe. But it provided the most complete record ever of the myriad physiological changes that come with progressive food deprivation. It also described in unprecedented detail the psychological and cognitive state of inanition -- the loss of libido, the decline of motivation, the weakening of social norms and the nearly pornographic obsession with food.
When Nathaniel Philbrick, winner of the 2000 National Book Award for his nonfiction "In the Heart of the Sea," needed a description of what a group of Nantucket whalers who survived three months in open boats in 1821 might have looked like or been thinking about, he turned to Keys's study.
As a purely scientific document, the study's results also continue to be useful.
Jeyu Henry, professor of human nutrition at Oxford Brookes University in England, is using the Minnesota experiment's data in research on the variation between individuals in the decline of metabolic rate and tissue stores. That's a still-murky subject that would be useful to understand better to predict who is at highest risk of death in places like refugee camps.
"It's a tribute to Keys that he had the foresight to record and publish all the data," Henry said. (Each of the Minnesota volunteers' full record is in the appendix of Volume 2.) Every time I look at the book, I am still amazed at both the depth and wealth of information it brought to the public domain."
Among the many things the study helped lead later researchers to believe in a metabolic "set-point" -- the state to which the complicated economy of caloric intake, energy expenditure and weight tends to return in each individual. However, Henry Blackburn, Keys's one-time protege and longtime collaborator, believes the starvation experiment may have made the opposite impression on Keys and his collaborators, at least unconsciously.
Far from demonstrating the tenacity of physiological states, the starvation study demonstrated their impermanence. The prevailing belief was that a person was endowed by nature with a body type -- "endomorphic" (fat), "mesomorphic" (well-proportioned) or "ectomorphic" (thin) -- and that blood pressure, cholesterol levels, resting heart rate, response to stress and numerous other variables were inborn and unchangeable. All of those things changed dramatically and predictably over the course of six months. Really big things could be changed by diet alone, the study announced.
It was an insight that set the scene for the third act of Ancel Keys's career.
Heart of the Matter
Soon after the war Keys became aware -- as did many people -- that the incidence of heart attacks was rising dramatically. (He says he first noticed the trend in the obituary pages.) Business and professional men appeared to be at highest risk. In 1947, with the starvation work done, he decided to look into the phenomenon. He recruited 283 men in the Minneapolis and St. Paul area, examined them, took blood samples and observed them, calling them back for repeat tests every five years.
The Twin Cities Study was the first prospective study of heart disease epidemiology in the United States, several months ahead of the Framingham Heart Study. The Framingham study rapidly eclipsed it for several reasons -- it was larger, it included women and it examined patients every two years. Nevertheless, the heart attack epidemic was big enough that even a study the size of Keys's soon showed that smoking, high blood pressure and elevated cholesterol were frequently observed in the men who suffered heart attacks.
Although stress was thought to be the main driving force behind the epidemic, Keys believed diet played a role.
He thought so for two reasons. Reports out of Europe -- the Netherlands and Sweden, in particular -- suggested that heart attack rates had fallen during the war, when food, and especially meat and dairy products, were in short supply. He also knew that research done by Russian physiologists on rabbits early in the century showed that cholesterol could be manipulated through the diet. Keys and his collaborators now wanted to study this in humans.
They turned to the state mental hospital system, where Keys had done work years before. There, they undertook tedious metabolic studies on mental patients. Over eight-week periods, they kept the caloric intake of patients' diet constant, but altered the balance between carbohydrates and fats. In particular, they looked at the effects of four forms of fat -- butter, vegetable shortening, olive oil and corn oil.
After nearly a decade of work, they had established without question that saturated fat -- and, ironically, not cholesterol itself -- was the chief determinant of blood cholesterol levels. The work produced, among other things, the "Keys equation," a formula that relates dietary fat to blood cholesterol and that is still used in research today.
Studies on captive and disabled people, even when they pose no obvious risk, are considered unethical today and would not be permitted. (The starvation experiment, although done with informed consent of a highly intelligent volunteers, also probably wouldn't be allowed.) Healthy but confined populations -- schools for the retarded and reformatories -- were common sites for diet studies at the time, with the head of the hospital, not the patients, giving consent.
"We tried to explain, but they didn't pay much attention," Keys said, referring to the patients. "The only thing that was clear was that when we paid more attention they woke up just a little. The superintendent believed it helped them."
The Laboratory of Physiological Hygiene by this time was a small fiefdom, ruled by Keys but capable of running in his absence. In 1951, he went on sabbatical to Europe. At a meeting of the United Nations' Food and Agriculture Organization in Rome, a professor of medicine from Naples told him that coronary heart disease was of no importance where he lived. Keys was dubious, and the doctor invited him to visit and see for himself.
Keys and his wife went to Naples and did a brief survey. Working class Neapolitans had low cholesterol. Members of the local Rotary Club had much higher cholesterol. The big difference between them seemed to be what they ate. The workers ate almost no meat or butter. The affluent people -- among whom the few heart attacks occurred, he was told -- ate lots of it.
Keys did a similar quick survey in Madrid and found the same thing. After several years of talking up the "dietary hypothesis" of coronary disease, he finally got traction when Paul Dudley White, President Dwight D. Eisenhower's cardiologist, clamped onto Keys's idea of a multi-country comparison of heart disease. White was the president of the World Congress on Cardiology, which also helped.
The result was the creation, in 1958, of an international collaboration that became the Seven Countries Study. There have been few epidemiological studies as complicated.
A total of 12,000 men aged 40 through 59 were recruited from 16 geographic or occupational cohorts in Italy, Yugoslavia, Greece, Finland, the Netherlands, Japan and the United States. The groups were intentionally homogeneous; most were rural with a dominant industry, such as farming, lumbering or fishing.
As with other long-term studies of heart disease, the subjects were interviewed, examined and tested extensively at the start of the project. But unlike other studies, there was a heroic effort to characterize diet in detail. This involved periodically sending an assistant to the households of a selected sample of participants. The person was observed eating, and a duplicate of the meal was collected. The food was then homogenized and dried, and a sample analyzed for fat, carbohydrate and protein content.
(The Framingham study, for example, did little questioning about diet in its early years, and has never taken food samples.)
After 10 years, the health status of the participants was rechecked. The results were astonishing.
Men in 11 villages on the island of Crete in Greece had the lowest rate of death by heart attack -- nine per 10,000 population. Men in villages in eastern Finland had the highest -- 992 per 10,000.
Sprinkled through that 100-fold range were the rest of the cohorts. Japanese farmers (with a heart attack death rate of 88 per 10,000) and Italian villagers near the Adriatic (150) were at the low end, while men in the Dutch market town of Zutphen (420) and U.S. railroad workers (574) were at the high end.
With a few exceptions, the higher the average cholesterol in a cohort, the higher that cohort's heart attack death rate would be. The strongest correlation, however, was between heart attack death rate and the percentage of total calories from saturated fat.
Keys had described in the 1950s what appeared to be a diet associated with low rates of coronary disease. It contained large amounts of olive oil, pasta, fresh vegetables and fruit, and sparing amounts of meat, butter and dairy products. Margaret Keys and he wrote a cookbook in 1959, "Eat Well and Stay Well," that promoted this diet. It sold more than 100,000 copies and went through several printings, eventually adding "The Mediterranean Diet" to the title. Now, the Seven Countries Study stood as proof that people in Greece and Italy were, in fact, onto something.
Men in both Crete, with the lowest rate of fatal heart attacks, and eastern Finland, with the highest, consumed 40 percent of their calories as fat. In Crete, however, the fat was largely olive oil. In Finland -- where a popular logger's lunch was a slab of cheese covered with butter and buried in a hollowed-out loaf of bread -- it was virtually all saturated fat.
The study revealed other insights into heart disease that weren't obvious when individuals, rather than populations, were examined. The most striking involved smoking.
The two Japanese cohorts had the highest rates of men who smoked a pack a day or more (43 percent and 38 percent). Crete was quite high as well, with 30 percent of subjects smoking heavily. But those were precisely the places where heart attack rates and average cholesterol were lowest. Elevated cholesterol appeared to be a necessary precursor for coronary heart disease -- the foundation on which the disease was built. Without it, the risk factor of smoking didn't kick in.
(This is a generalization about populations. There are always individual exceptions to the population norm -- heart attack victims with low cholesterol whose smoking may have led to the event. The study also found, in nearly all places, a strong relationship between smoking and lung cancer.)
Perhaps the most important idea to come out of the Seven Countries Study was the notion that an entire population could be diagnosed and "treated" -- at least in terms of public health interventions.
The average cholesterol in the American railroad workers was about 240 milligrams per deciliter. In the Japanese villagers, it was about 165. Americans whose readings were comfortably below normal fell into the highest range of Japanese -- into the range, in fact, where heart attacks occurred. Very few Americans (or Finns or Dutch) had what could be termed a "safe" level of cholesterol. Instead, the entire populations were shifted into the danger zone.
In short, the study showed that avoidable deaths were occurring across nearly the whole distribution of cholesterol values in many countries. Consequently, treating the whole population would benefit almost everyone.
The Finnish government tried that in 1972 in North Karelia, the region with the highest heart attack rate in the Seven Countries Study. It undertook a community-wide intervention aimed at getting people to drink low-fat milk, eat less butter and use oils in cooking.
By the mid-1990s, the average cholesterol in North Karelians had fallen by 15 percent. Cardiovascular mortality had been cut by more than half. Declines in cholesterol, smoking and hypertension were the main reasons.
Similar projects have been tried in the United States, with one city getting a diet-change public health campaign and a nearby city not getting one. (Fargo, N.D., and Sioux Falls, S.D, and Pawtucket and Providence in Rhode Island were some of the paired communities.) As it happened, the entire cardiac risk profile of the United States was evolving while these studies were underway in the 1980s. Some of this was the result of public health campaigns (the one against smoking most notably); much was simply a change of culture.
"The comparison cities as well as the experimental ones were undergoing large changes in diet, smoking and blood pressure control. Everybody's risk fell, and the study effects couldn't be proven," Blackburn said recently.
Nationally, the change has been as dramatic as that seen in North Karelia, with the U. S. death rate from heart disease falling 56 percent between 1950 and 1996. Nevertheless, the dominant strategy here targets individuals (with screening tests and drugs) rather that populations (with public health messages and economic incentives). Some experts believe, however, that the population-as-patient model raised by the Seven Countries Study may find new life against this century's epidemic -- obesity.
There are no good medical therapies for obesity. Strategies to change people's eating and exercising habits, to get school systems to change lunch menus and food processors to shrink portion sizes -- these may be the only realistic approaches for attacking it.
Ancel Keys retired as the head of the Laboratory of Physiological Hygiene in 1972. For a long time, he and his wife -- they had three children -- spent part of every year in a house they built in southern Italy with proceeds of their cookbook. Keys's several strokes and a broken hip in recent years, however, now keep them in Minneapolis.
Keys's more than 35 Atlantic crossings since the start of the Seven Countries Study provided a kind of stroboscopic picture of the emerging obesity epidemic in the United States. Each time he came home he was struck by how many -- and how many more -- overweight people there were in the United States than in Europe or Japan.
This epochal trend wasn't the object of his research, and he is uncertain what explains it. Affluence and inexpensive processed food may be part of the reason, he thinks, along with what he graphically calls "the North American habit for making the stomach the garbage disposal unit for a long list of harmful foods."
He is periodically mentioned as a candidate for the Nobel Prize in Physiology or Medicine. But no one has ever gotten one for epidemiology, so it's unlikely he will.
His record is not without its critics or flaws. He failed to recognize or take account of HDL, the form of cholesterol that lowers coronary risk. In the Seven Countries Study, neither physical activity nor overweight was found to be associated with coronary risk, although subsequent research has shown it is. He was sometimes difficult to work with. "He can be frank to the point of blunt trauma, and critical to the point of razor slash," one colleague wrote in a tribute a number of years ago
Nearly all the findings of the Seven Countries Study were published in monographs or books edited or largely written by Keys. Some people believe the work could have reached a wider audience if he had exerted less personal control, and allowed the findings to appear in professional journals. In contrast, the Framingham Heart Study, never identified with a single person, has produced a steady flow of short, relatively reader-friendly articles.
"We were more inclined to make sure that the information was widely dispersed," said William B. Kannel, who has worked at the Framingham study since 1949, and directed it for a time. "We published in all kinds of journals -- clinical journals, public health journals, specialty journals -- always targeting an audience."
Keys also has some regrets. He wishes he had included women -- for whom coronary disease is also the leading cause of death -- in his studies. He wishes he had studied China, where one-quarter of humanity lives. But for the record, and at age 98, he does not regret much else.