Through the first half of the twentieth century, little was understood of insulin beyond its role in diabetes, because no method existed to measure its concentration in the bloodstream with any accuracy. Insulin is a very smal protein, technical y known as a peptide, and it circulates in the blood in concentrations that are infinitesimal compared with those of cholesterol and lipoproteins. As a result, the measurement of insulin in human blood relied on a variety of arcane tests that depended on the ability of insulin to prompt the absorption of glucose by laboratory rats or even by fat or muscle tissue in a test tube. This situation changed in 1960 with the discovery by Rosalyn Yalow and Solomon Berson of a method capable of reliably measuring the concentration of insulin and other peptide hormones in human blood. In 1977, when Yalow was awarded the n.o.bel Prize for the discovery (Berson had died in 1972), the n.o.bel Foundation described Yalow and Berson's measurement technology as bringing about "a revolution in biological and medical research."

The impact on diabetes research had been immediate. Yalow and Berson showed that those who had developed diabetes as adults had levels of circulating insulin significantly higher than those of healthy individuals-a surprising finding. It had long been a.s.sumed that lack of insulin was the root of al diabetes. As Yalow and Berson among others also reported, the obese, too, had chronical y elevated insulin levels.

By 1965, Yalow and Berson had suggested why these adult-onset diabetics could appear to be lacking insulin-manifesting the symptoms of diabetes, high blood sugar, and sugar in their urine-while simultaneously having excessive insulin in their circulation: their tissues did not respond properly to the insulin they secreted. They were insulin-resistant, defined by Yalow and Berson as "a state (of a cel , tissue, system or body) in which greater-than-normal amounts of insulin are required to elicit a quant.i.tatively normal response." Because of their resistance to insulin, adult-onset diabetics had to secrete more of the hormone to maintain their blood sugar within healthy levels, and this would become increasingly difficult to achieve the longer they remained insulin-resistant.*51 A critical aspect of this insulin resistance, Yalow and Berson noted, is that some tissues might become resistant to insulin while others continued to respond normal y, and this would determine how the damage done by the insulin resistance would manifest itself in different individuals. So "it is desirable," they wrote, "wherever possible, to distinguish generalized resistance of al tissues from resistance of only individual tissues."

From the mid-1960s onward, our understanding of the role of insulin resistance in both heart disease and diabetes was driven by the work of Stanford University diabetologist Gerald Reaven. Reaven began his investigations by measuring triglycerides and glucose tolerance in heart-attack survivors. A glucose-tolerance test is a common test given by physicians to determine if a patient is either diabetic or on the way to becoming so. The patient drinks a solution of glucose and water, and then, two hours later, the physician measures his or her blood sugar. If the blood sugar is higher than what's considered normal, it means the patient has been unable to metabolize the glucose properly-hence, glucose intolerance-and so either lacks sufficient insulin to deal with the glucose, or is resistant to the insulin that is secreted. In 1963, Reaven reported that heart-attack survivors invariably had both high triglycerides and glucose intolerance, and this suggested that the two conditions had a common cause. Reaven considered insulin resistance to be the obvious suspect.

Working with John Farquhar, who had studied with Pete Ahrens at Rockefel er, Reaven developed a two-part hypothesis.

The first part explained why most, if not virtual y al individuals with high triglycerides had what Ahrens had cal ed carbohydrate-induced lipemia. In other words, their triglyceride levels increased with carbohydrate-rich diets and decreased when fat replaced the carbohydrates. The crucial factor, Reaven explained, is that, the more carbohydrates consumed, the more insulin is needed to transport the glucose from the carbohydrates into cel s where it can be used as fuel. This insulin, however, also prompts the liver to synthesize and secrete triglycerides for storage in the fat tissue. If someone who is already insulin-resistant consumes a carbohydrate-rich diet, according to Reaven's hypothesis, the person wil have to secrete even more insulin to deal with the glucose, prompting in turn even greater synthesis and secretion of triglycerides by the liver, and so even higher triglyceride levels in the blood.

This, in turn, implied part two of the hypothesis: if eating a carbohydrate-rich diet in the presence of insulin resistance wil abnormal y elevate triglyceride levels, then it's hard to avoid the implication that eating a carbohydrate-rich diet increases the risk of heart disease. Insulin resistance and carbohydrates wil also exacerbate Type 2 diabetes, according to Reaven's hypothesis, and this would explain, as wel , why these diabetics inevitably have high triglycerides. By 1967, Reaven and Farquhar had reported that triglyceride levels, insulin resistance, and insulin levels moved up and down in concert even in healthy individuals: the more insulin secreted in response to carbohydrates, the greater the apparent insulin resistance and the higher the triglycerides.

Reaven and Farquhar spent the next twenty years working to establish the validity of the hypothesis. Much of the progress came with the development, once again, of new measuring techniques: in this case, tests that al owed investigators to measure insulin resistance directly. In 1970, Reaven and Farquhar published the details of the first such insulin-resistance test, which was then fol owed by a half-dozen more. The best of these-the "gold standard"-was developed at the NIH in the late 1960s and then refined over the next decade by a young endocrinologist named Ralph DeFronzo. It wasn't until 1979, after DeFronzo joined the faculty at Yale Medical School and began measuring insulin resistance in human patients, that he published the details. It would take another decade for Reaven, Farqhuar, and DeFronzo, along with Eleuterio Ferrannini of the University of Pisa, among others, to convince diabetologists that resistance to insulin was the fundamental defect in Type 2 diabetes.

In 1987, the American Diabetes a.s.sociation honored DeFronzo with its award for outstanding scientific achievement. A year later, Reaven received the ADA's Banting Medal for Scientific Achievement.*52 Reaven then gave the prestigious Banting Lecture at the ADA's annual conference and took the opportunity to extend the implications of his research. For the first time, he laid out the hypothesis of what he cal ed Syndrome X (metabolic syndrome) and the cl.u.s.ter of disorders-including insulin resistance, hyperinsulinemia, high triglycerides, low HDL cholesterol, and high blood pressure-that accompanies Type 2 diabetes and obesity and plays a critical role in the genesis of heart disease even in nondiabetics. "Although this concept may seem outlandish at first blush," Reaven said, "the notion is consistent with available experimental data." As Reaven described it, the condition of being resistant to insulin leads to both heart disease and diabetes. But not everyone with insulin resistance becomes diabetic; some continue to secrete sufficient insulin to overcome their insulin resistance, though this hyperinsulinemia causes havoc on its own, including elevating triglyceride levels, and also further exacerbating the insulin resistance-a vicious cycle.

Reaven supported his hypothesis with the results of observational studies that had already linked hyperinsulinemia, insulin resistance, and Type 2 diabetes to high triglycerides, heart disease, obesity, stroke, and hypertension. Three large-scale Framingham-like prospective studies of healthy nondiabetic populations-in Paris, Helsinki, and in Busselton, Australia-had also reported that, the higher the insulin levels, the greater the risk of heart disease.

As DeFronzo later remarked, the conclusion that hyperinsulinemia and insulin resistance were related to "a whole host of metabolic disorders" was an obvious one, but it required that clinical investigators measure insulin resistance in human patients, which would always be the obstacle in the science of metabolic syndrome. Measuring insulin resistance requires multiple tests of blood sugar while insulin levels are held constant and precise amounts of glucose are consumed or infused into the bloodstream. This is not the kind of test that physicians can do in a checkup, at least not without going far beyond the usual practice of sending a blood sample out to a laboratory for a battery of tests. As a result, when the National Cholesterol Education Program official y acknowledged the existence of Reaven's Syndrome X in 2002 (renaming it metabolic syndrome), neither insulin resistance nor hyperinsulinemia was included among the diagnostic criteria, despite being the fundamental defects in the syndrome itself.

Reaven's 1988 Banting Lecture is credited as the turning point in the effort to convince diabetologists of the critical importance of insulin resistance and hyperinsulinemia, but those investigators concerned with the genesis of heart disease paid little attention, considering anything having to do with insulin to be relevant only to diabetes. This was a natural consequence of the specialization of scientific research. Through the mid-1980s, Reaven's research had focused on diabetes and insulin, and so his publications appeared almost exclusively in journals of diabetes, endocrinology, and metabolism. Not until 1996 did Reaven publish an article on Syndrome X in the American Heart a.s.sociation journal Circulation, the primary journal for research in heart disease. Meanwhile, his work had no influence on public-health policy or the public's dietary consciousness. Neither the 1988 Surgeon General's Report on Nutrition and Health nor the National Academy of Sciences's 1989 Diet and Health mentioned insulin resistance or hyperinsulinemia in any context other than Reaven's cautions that high-carbohydrate diets might not be ideal for Type 2 diabetics. Both reports ardently recommended low-fat, high-carbohydrate diets for the prevention of heart disease.

Even the diabetes community found it easier to accept Reaven's science than its dietary implications. Reaven's observations and data "speak for themselves," as Robert Silverman of the NIH suggested at a 1986 consensus conference on diabetes prevention and treatment. But they placed nutritionists in an awkward position. "High protein levels can be bad for the kidneys," said Silverman. "High fat is bad for your heart. Now Reaven is saying not to eat high carbohydrates. We have to eat something." "Sometimes we wish it would go away," Silverman added, "because n.o.body knows how to deal with it."

This is what psychologists cal cognitive dissonance, or the tension that results from trying to hold two incompatible beliefs simultaneously. When the philosopher of science Thomas Kuhn discussed cognitive dissonance in scientific research-"the awareness of an anomaly in the fit between theory and nature"-he suggested that scientists wil typical y do what they have invariably done in the past in such cases: "They wil devise numerous articulations and ad hoc modifications of their theory in order to eliminate any apparent conflict." And that's exactly what happened with metabolic syndrome and its dietary implications. The syndrome itself was accepted as real and important; the idea that it was caused or exacerbated by the excessive consumption of carbohydrates simply vanished.

Among the few clinical investigators working on heart disease who paid attention to Reaven's research in the late 1980s was Ron Krauss. In 1993, Krauss and Reaven together reported that smal , dense LDL was another of the metabolic abnormalities commonly found in Reaven's Syndrome X.

Smal , dense LDL, they noted, was a.s.sociated with insulin resistance, hyperinsulinemia, high blood sugar, hypertension, and low HDL as wel . They also reported that the two best predictors of the presence of insulin resistance and the dominance of smal , dense LDL are triglycerides and HDL cholesterol -the higher the triglycerides and the lower the HDL, the more likely it is that both insulin resistance and smal , dense LDL are present. This offers yet another reason to believe the carbohydrate hypothesis of heart disease, since metabolic syndrome is now considered perhaps the dominant heart-disease risk factor-a "coequal partner to cigarette smoking as contributors to premature [coronary heart disease]," as the National Cholesterol Education Program describes it-and both triglycerides and HDL cholesterol are influenced by carbohydrate consumption far more than by any fat.

Nonetheless, when smal , dense LDL and metabolic syndrome official y entered the orthodox wisdom as risk factors for heart disease in 2002, the cognitive dissonance was clearly present. First the National Cholesterol Education Program published its revised guidelines for cholesterol testing and treatment. This was fol owed in 2004 by two conference reports: one describing the conclusions of a joint NIH-AHA meeting on scientific issues related to metabolic syndrome, and the other, in which the American Diabetes a.s.sociation joined in as wel , describing joint treatment guidelines. Scott Grundy of the University of Texas was the primary author of al three doc.u.ments. When I interviewed Grundy in May 2004, he acknowledged that metabolic syndrome was the cause of most heart disease in America, and that this syndrome is probably caused by the excessive consumption of refined carbohydrates. Yet his three reports-representing the official NIH, AHA, and ADA positions-al remained firmly wedded to the fat-cholesterol dogma. They acknowledge metabolic syndrome as an emerging risk factor for heart disease, but identify LDL cholesterol as "the primary driving force for coronary atherogenesis."

Thus, heart disease in America, as the National Cholesterol Education Program report put it, was stil official y caused by "ma.s.s elevations of serum LDL cholesterol result[ing] from the habitual diet in the United States, particularly diets high in saturated fats and cholesterol."

There was no mention that carbohydrates might be responsible for causing or exacerbating either metabolic syndrome or the combination of low HDL, high triglycerides, and smal , dense LDL, which is described as occurring "commonly in persons with premature [coronary heart disease].*53 In the now established version of the alternative hypothesis-that metabolic syndrome leads to heart disease-the carbohydrates that had always been considered the causative agent had been official y rendered harmless. They had been removed from the equation of nutrition and chronic disease, despite the decades of research and observations suggesting the critical causal role they played.

Chapter Eleven.

THE SIGNIFICANCE OF DIABETES.

Does carbohydrate cause arteriosclerosis? Certainly it does if taken in such excess as to produce obesity, but except in this manner no one would attribute any such function to it.... Is a persistent [high blood sugar] a cause of arteriosclerosis in diabetes? It very likely is a cause because it is an abnormal condition and any abnormal state would tend to wear out the machine.

ELLIOTT JOSLIN, "Arteriosclerosis and Diabetes," 1927 DESPITE NEARLY A CENTURY'S WORTH OF therapeutic innovations, the likelihood of a diabetic's contracting coronary artery disease is no less today than it was in 1921, when insulin was first discovered. Type 2 diabetics can stil expect to die five to ten years prematurely, with much of this difference due to atherosclerosis and what Joslin's Diabetes Mellitus has cal ed an "extraordinarily high incidence" of coronary disease.

Diabetes specialists have historical y perceived this plague of atherosclerosis among their patients as though it has little relevance to the atherosclerosis and heart disease that affect the rest of us. Textbooks would note the importance of identifying and control ing the "numerous and as yet il -defined factors general y involved in the pathogenesis of atherosclerosis," as the 1971 edition of Joslin's Diabetes Mellitus did, but the implication was that the requisite revelations would emerge, as they had in the past, from heart-disease researchers, as though the flow of knowledge about heart disease could proceed only from heart-disease research to diabetology and never the other way around.

The extreme example of this thinking has been the a.s.sumption that saturated fat is the nutritional agent of heart disease in diabetics, just as it supposedly is in everyone else. "The frequent cardiovascular complications seen in past years among persons with diabetes," the 1988 Surgeon General's Report on Nutrition and Health says, are caused by the "traditional restriction of carbohydrate intake in persons with diabetes" and thus an increased intake of fat, "usual y, saturated." This was the logic that led the American Diabetes a.s.sociation, from the early 1970s, to recommend that diabetics eat more carbohydrates rather than less, despite a complete absence of clinical trials that might demonstrate that the benefits of doing so outweigh the risks, and the decades of clinical experience establishing carbohydrate restriction as an effective method of control ing blood sugar. If atherosclerosis was accelerated in diabetics, the thinking went, it was accelerated because they ate more saturated fat than nondiabetics. Diabetologists believed they could safely prescribe a carbohydrate-rich diet to their patients, because a diet that is low in fat wil be high in carbohydrates.

But the research on metabolic syndrome suggests an entirely different scenario. If the risk of heart disease is elevated in metabolic syndrome and elevated stil further with diabetes, then maybe the flow of knowledge about heart disease should proceed from diabetics, who suffer the most extreme manifestation of the disease, to the rest of us, and not the other way around. Maybe diabetics have such extreme atherosclerosis because there is something about the diabetic condition that causes the disease. Perhaps the metabolic abnormalities of the diabetic condition are the essential cause of atherosclerosis and coronary heart disease in everyone, only diabetics suffer to a greater extent.

Another way to look at this is to consider that metabolic syndrome and Type 2 diabetes lie on a continuum or a curve of physical degeneration. This curve is marked by ever-worsening disturbances of carbohydrate and fat metabolism-high insulin, insulin resistance, high blood sugar, high triglycerides, low HDL, and smal , dense LDL. Atherosclerosis is one manifestation of this physical degeneration. In diabetes, the metabolic abnormalities are exacerbated-diabetics are further down the curve of physical degeneration-and the atherosclerotic process is accelerated. But we al live on the same curve. The mechanisms that cause atherosclerosis are the same in al of us; only the extent of damage differs.

Consider Keys's cholesterol hypothesis as an example of this logic. One reason we came to believe that high cholesterol is a cause of heart disease is that severe atherosclerosis is a common symptom of genetic disorders of cholesterol metabolism. If having a cholesterol level of 1,000 mg/dl-as these individuals often do-makes atherosclerosis seemingly inevitable, the logic goes, and if higher cholesterol seems to a.s.sociate with higher risk of heart disease among the rest of us, then cholesterol is a cause of heart disease, and elevating cholesterol by any amount wil increase risk. The higher the cholesterol, the greater the risk. If eating saturated fat elevates cholesterol, then that in turn causes heart disease. And this is supposedly true of diabetics as wel . Keys oversimplified the science and was wrong about the true relationship of cholesterol and heart disease, but the logic itself is otherwise sound.

The same logic holds for blood pressure and heart disease. The higher the blood pressure, the greater the risk of heart disease. If salt supposedly raises blood pressure, even if only by a few percentage points, then salt is a nutritional cause of heart disease. This, too, is held to be true for diabetics.

Thus, the atherogenic American diet, as now official y defined, the diet that clogs arteries and causes heart disease, is a diet high in saturated fat and salt.

Now let's apply the same reasoning to metabolic syndrome and diabetes. Diabetics suffer more virulent atherosclerosis and die of heart disease more frequently than those with metabolic syndrome, and much more frequently than healthy individuals who manifest neither condition. Some aspect of the diabetic condition must be the cause-most likely, either high blood sugar, hyperinsulinemia, or insulin resistance, al three of which wil tend to be worse in diabetics than in those with metabolic syndrome. Indeed, the existence of metabolic syndrome tel s us that these same abnormalities exist in nondiabetics, although to a lesser extent, and though individuals with metabolic syndrome suffer an increased risk of heart disease, they do so to a lesser extent than diabetics. And because dietary carbohydrates and particularly refined carbohydrates elevate blood sugar and insulin and, presumably, induce insulin resistance, the implication is that eating these carbohydrates increases heart-disease risk not only in diabetics but in healthy individuals. By this reasoning, the atherogenic American diet is a carbohydrate-rich diet. Hence, cognitive dissonance.

The logic of this argument has to be taken one step further, however, even if the cognitive dissonance is elevated with it. Both diabetes and metabolic syndrome are a.s.sociated with an elevated incidence of virtual y every chronic disease, not just heart disease. Moreover, the diabetic condition is a.s.sociated with a host of chronic blood-vessel-related problems known as vascular complications: stroke, a stroke-related dementia cal ed vascular dementia, kidney disease, blindness, nerve damage in the extremities, and atheromatous disease in the legs that often leads to amputation. One obvious possibility is that the same metabolic and hormonal abnormalities that characterize the diabetic condition-in particular, elevated blood sugar, hyperinsulinemia, and insulin resistance-may also cause these complications and the a.s.sociated chronic diseases. And otherwise healthy individuals, therefore, would be expected to increase their risk of all these conditions by the consumption of refined and easily digestible carbohydrates, which inflict their damage first through their effects on blood sugar and insulin, and then, indirectly, through triglycerides, lipoproteins, fat acc.u.mulation, and a.s.suredly other factors as wel .

This is a fundamental tenet of the carbohydrate hypothesis: If the risk of contracting any chronic disease or condition increases with metabolic syndrome and Type 2 diabetes, then it's a reasonable hypothesis that insulin and/or blood sugar plays a role in the disease process. And if insulin and blood sugar do play a pathological role, then it's a reasonable hypothesis that the same conditions can be caused or exacerbated in healthy individuals by the consumption of refined and easily digestible carbohydrates and sugars.

Among the immediate examples that fol ow from this logic is the particularly disconcerting possibility that insulin itself causes or exacerbates atherosclerosis. Since insulin resistance and hyperinsulinemia characterize Type 2 diabetes, it's certainly possible that chronical y elevated levels of insulin are the cause of the persistently high incidence of atherosclerosis in diabetics, quite aside from any other effects insulin might have on triglycerides, lipoproteins, or blood pressure. And if this is the case, then the excessive secretion of insulin-induced by the consumption of refined carbohydrates and sugars-might be responsible for causing or exacerbating atherosclerosis in those of us who are not diabetic.

This is another of those conceptions, like the ability of insulin to regulate blood pressure, that have been mostly neglected for decades, despite the profound implications if it's true. The specter of this atherogenic effect of insulin is noted briefly, for example, in the fourteenth edition (2005) of Joslin's Diabetes Mellitus. The Harvard diabetologist Edward Feener and Victor Dzau, president of the Duke University Health System, write that "the effects of insulin on [cardiovascular disease] in diabetes and insulin resistance are related to both systematic metabolic abnormalities and the direct effects of insulin action on the vasculature [blood vessels; my italics]." The second mention, by two Harvard cardiologists, acknowledges the a.s.sociation between insulin resistance, hyperinsulinemia, and heart disease and suggests that if insulin resistance is not the problem, then "another possibility" is that insulin itself "has direct cardiovascular effects." Nothing more is said.

The first evidence of the potential atherogenicity of insulin emerged from precisely the kind of experiments in rabbits that initial y gave credibility to the cholesterol hypothesis a century ago. Rabbits fed high-cholesterol diets develop plaques throughout their arteries, but diabetic rabbits (Type 1) wil not suffer this atherosclerotic fate no matter how cholesterol-rich their diet. Infuse insulin along with the cholesterol-laden diet, however, and plaques and lesions wil promptly blossom everywhere. This phenomenon was first reported in 1949 in rabbits, and then, a few years later, in chickens, by Jeremiah Stamler and his mentor Louis Katz, and later in dogs, too. Hence, insulin itself may be "one factor in the pathogenesis of the frequent, premature, severe atherosclerosis of diabetic patients," as Stamler and his col eagues suggested.

In the late 1960s, Robert Stout of Queen's University in Belfast published a series of studies reporting that insulin enhances the transport of cholesterol and fats into the cel s of the arterial wal and stimulates the synthesis of cholesterol and fat in the arterial lining. Since a primary role of insulin is to facilitate the storage of fats in the fat tissue, Stout reasoned, it was not surprising that it would have the same effect on the lining of blood vessels. In 1969, Stout and the British diabetologist John Val ance-Owen pre-empted Reaven's Syndrome X hypothesis by suggesting that the "ingestion of large quant.i.ties of refined carbohydrate" leads first to hyperinsulinemia and insulin resistance, and then to atherosclerosis and heart disease. In certain individuals, they suggested, the insulin secretion after eating these carbohydrates would be "disproportionately large." "The carbohydrate is disposed of in three sites -adipose [fat] tissue, liver and arterial wal ," Stout wrote. "Obesity is produced. In the liver, triglyceride and cholesterol are synthesized and find their way into the circulation. Lipid synthesis is also stimulated in the arterial wal and is augmented by deposition of [triglycerides and cholesterol]...which in a few decades would reach significant proportions." In 1975, Stout and the University of Washington pathologist Russel Ross reported that insulin also stimulates the proliferation of the smooth muscle cel s that line the interior of arteries, a necessary step in the thickening of artery wal s characteristic of both atherosclerosis and hypertension.

This insulin-atherogenesis hypothesis is the simplest possible explanation for the intimate a.s.sociation of diabetes and atherosclerosis: the excessive secretion of insulin accelerates atherosclerosis and perhaps other vascular complications. It also implies, as Stout suggested, that any dietary factor -refined carbohydrates in particular-that increases insulin secretion wil increase risk of heart disease. This did not, however, become the preferred explanation. Even Reaven chose to ignore it.*54 But Reaven's hypothesis proposed that heart disease was caused primarily by insulin resistance through its influence on triglycerides. He considered hyperinsulinemia to be a secondary phenomenon. Stout considered hyperinsulinemia the primary cause of atherosclerosis.

Most diabetologists have believed that diabetic complications are caused by the toxic effects of high blood sugar.*55 The means by which high blood sugar induces damage in cel s, arteries, and tissues are indeed profound, and the consequences, as the carbohydrate hypothesis implies, extend far beyond diabetes itself. This line of research is pursued by only a few laboratories. As a result, its ultimate implications and validity remain to be ascertained. But it should be considered as yet another potential mechanism by which the consumption of refined carbohydrates could cause or exacerbate the entire spectrum of the chronic diseases of civilization.

In particular, raising blood sugar wil increase the production of what are known technical y as reactive oxygen species and advanced glycation end-products, both of which are potential y toxic. The former are generated primarily by the burning of glucose (blood sugar) for fuel in the cel s, in a process that attaches electrons to oxygen atoms, transforming the oxygen from a relatively inert molecule into one that is avid to react chemical y with other molecules. This is not an ideal situation biological y. One form of reactive oxygen species is those known commonly as free radicals, and al of them together are known as oxidants, because what they do is oxidize other molecules (the same chemical reaction that causes iron to rust, and equal y deleterious). The object of oxidation slowly deteriorates. Biologists refer to this deterioration as oxidative stress. Antioxidants neutralize reactive oxygen species, which is why antioxidants have become a popular buzzword in nutrition discussions.

The potential of advanced glycation end-products (AGEs) for damage is equal y worrisome. Their formation can take years, but the process (glycation) begins simply, with the attachment of a sugar-glucose, for instance-to a protein without the benefit of an enzyme to orchestrate the reaction. That absence is critical. The role of enzymes in living organisms is to control chemical reactions to ensure that they "conform to a tightly regulated metabolic program," as the Harvard biochemist Frank Bunn explains. When enzymes affix sugars to proteins, they do so at particular sites on the proteins, for very particular reasons. Without an enzyme overseeing the process, the sugar sticks to the protein haphazardly and sets the stage for yet more unintended particular reasons. Without an enzyme overseeing the process, the sugar sticks to the protein haphazardly and sets the stage for yet more unintended and unregulated chemical reactions.

The term glycation refers only to this initial step, a sugar molecule attaching to a protein, and this part of the process is reversible-if blood-sugar levels are low enough, the sugar and protein wil disengage, and no damage wil be done. If blood sugar is elevated, however, then the process of forming an advanced glycation end-product wil move forward. The protein and its accompanying glycated sugars wil undergo a series of reactions and rearrangements until the process culminates in the convoluted form of an advanced glycation end-product. These AGEs wil then bind easily to other AGEs and to stil more proteins through a process known as cross-linking-the sugars hooked to one protein wil bridge to another protein and lock them together. Now proteins that should ideal y have nothing to do with each other wil be inexorably joined.

In the mid-1970s, Rockefel er University biochemist Anthony Cerami and Frank Bunn independently recognized that AGEs and glycation play a major role in diabetes.*56 Both Cerami and Bunn were initial y motivated by the observation that diabetics have high levels of an unusual form of hemoglobin -the oxygen-carrying protein of red blood cel s-known as hemoglobin A1c, a glycated hemoglobin. The higher the blood sugar, the more hemoglobin molecules undergo glycation, and so the more hemoglobin A1c can be found in the circulation. Cerami's laboratory then developed an a.s.say to measure hemoglobin A1c, speculating correctly that it might be an accurate reflection of the diabetic state. Diabetics have two to three times as much hemoglobin A1c in their blood as nondiabetics, a ratio that apparently holds true for nearly al glycated proteins in the body. (The best determination of whether diabetics are successful y control ing their blood sugar comes from measuring hemoglobin A1c, because it reflects the average blood sugar over a month or more.) Since 1980, AGEs have been linked directly to both diabetic complications and aging itself (hence the acronym). AGEs acc.u.mulate in the lens, cornea, and retina of the eye, where they appear to cause the browning and opacity of the lens characteristic of senile cataracts. AGEs acc.u.mulate in the membranes of the kidney, in nerve endings, and in the lining of arteries, al tissues typical y damaged in diabetic complications. Because AGE acc.u.mulation appears to be a natural y occurring process, although it is exacerbated and accelerated by high blood sugar, we have evolved sophisticated defense mechanisms to recognize, capture, and dispose of AGEs. But AGEs stil manage to acc.u.mulate in tissues with the pa.s.sing years, and especial y so in diabetics, in whom AGE acc.u.mulation correlates with the severity of complications.

One protein that seems particularly susceptible to glycation and cross-linking is col agen, which is a fundamental component of bones, cartilage, tendons, and skin. The col agen version of an AGE acc.u.mulates in the skin with age and, again, does so excessively in diabetics. This is why the skin of young diabetics wil appear prematurely old, and why, as the Case Western University pathologist Robert Kohn first suggested, diabetes can be thought of as a form of accelerated aging, a notion that is slowly gaining acceptance. It's the acc.u.mulation and cross-linking of this col agen version of AGEs that causes the loss of elasticity in the skin with age, as wel as in joints, arteries, and the heart and lungs.

The process can be compared to the toughening of leather. Both the meat and hide of an old animal are tougher and stiffer than those of a young animal, because of the AGE-related cross-linking that occurs inevitably with age. As Cerami explains, the aorta, the main artery running out of the heart, is an example of this stiffening effect of acc.u.mulated and cross-linked AGEs. "If you remove the aorta from someone who died young," says Cerami, "you can blow it up like a bal oon. It just expands. Let the air out, it goes back down. If you do that to the aorta from an old person, it's like trying to inflate a pipe.

It can't be expanded. If you keep adding more pressure, it wil just burst. That is part of the problem with diabetes, and aging in general. You end up with stiff tissue: stiffness of hearts, lungs, lenses, joints.... That's al caused by sugars reacting with proteins."

AGEs and the glycation process also appear to play at least one critical role directly in heart disease, by causing the oxidation of LDL particles and so causing the LDL and its accompanying cholesterol to become trapped in the artery wal , which is an early step in the atherosclerotic process. Oxidized LDL also appears to be resistant to removal from the circulation by the normal mechanisms, which would also serve to increase the LDL levels in the blood. As it turns out, LDL is particularly susceptible to oxidation by reactive oxygen species and to glycation.*57 In this case, both the protein portion and the lipid portion (the cholesterol and the fats) of the lipoprotein are susceptible. These oxidized LDL particles appear to be "markedly elevated" in both diabetics and in nondiabetics with atherosclerosis, and are particularly likely to be found in the atherosclerotic lesions themselves.

That glycation and AGEs are critical factors in diabetic complications and in heart disease has recently been demonstrated by experiments with compounds known as anti-AGE compounds or AGE breakers. These wil reverse arterial stiffness, at least in laboratory animals, and, as one recent report put it, ameliorate "the adverse cardiovascular and [kidney-related] changes a.s.sociated with aging, diabetes and hypertension." Whether these or similar compounds wil work in humans remains to be seen.

When biochemists discuss oxidative stress, glycation, and the formation of advanced glycation end-products, they often compare what's happening to a fire simmering away in our circulation. The longer the fire burns and the hotter the flame, the more damage is done. Blood sugar is the fuel. "Current evidence points to glucose not only as the body's main short-term energy source," as the American Diabetes a.s.sociation recently put it, "but also as the long-term fuel of diabetes complications."

But there is no reason to believe that glucose-induced damage is limited only to diabetics, or to those with metabolic syndrome, in whom blood sugar is also chronical y elevated. Glycation and oxidation accompany every fundamental process of cel ular metabolism. They proceed continuously in al of us.

Anything that raises blood sugar-in particular, the consumption of refined and easily digestible carbohydrates-wil increase the generation of oxidants and free radicals; it wil increase the rate of oxidative stress and glycation, and the formation and acc.u.mulation of advanced glycation end-products. This means that anything that raises blood sugar, by the logic of the carbohydrate hypothesis, wil lead to more atherosclerosis and heart disease, more vascular disorders, and an accelerated pace of physical degeneration, even in those of us who never become diabetic.

Chapter Twelve.

SUGAR.

M. Delacroix, a writer as charming as he is prolific, complained once to me at Versail es about the price of sugar, which at that time cost more than five francs a pound. "Ah," he said in a wistful, tender voice, "if it can ever again be bought for thirty cents, I'l never more touch water unless it's sweetened!" His wish was granted....

JEAN ANTHELME BRILLAT-SAVARIN, The Physiology of Taste, 1825 WHEN BIOCHEMISTS TALK ABOUT "SUGAR," they're referring to a whole host of very simple carbohydrate molecules, al of which are characterized, among other things, by their sweet taste and ability to dissolve in water. Their chemical names al end in "-ose"-glucose, fructose, and lactose, among others.

When physicians talk about blood sugar, they're typical y talking about glucose, although other sugars can be found in the bloodstream at very much lower concentrations. Then there's the common usage of "sugar," meaning the sweet, powdered variety that we put in our coffee or tea. This is sucrose, which in turn is const.i.tuted of equal parts glucose and fructose. In the discussion to come, when we refer to "sugar" we'l always be talking about sucrose. When we use the term "blood sugar," we'l be talking about glucose.

When nutritionists in the 1960s discussed the pros and cons of sugar and starches, their concern was whether simple carbohydrates were somehow more deleterious than complex carbohydrates of starches. Chemical y, simple carbohydrates, as in sugar and highly refined flour, are molecules of one or two sugars bound together, whereas the complex carbohydrates of starches are chains of sugars that can be tens of thousands of sugars long. Complex carbohydrates break down to simple sugars during the process of digestion, but they take a while to do so, and if the carbohydrate is bound up with fiber -i.e., indigestible carbohydrates-the digestion takes even longer. Since the early 1980s, both simple and complex carbohydrates have played a role in determining the glycemic index, which is a measure of how quickly carbohydrates are digested and absorbed into the circulation and so converted into blood sugar. This concept of a glycemic index has had profound consequences on the official and public perception of the risks of starches and sugar in the diet. But it has done so by ignoring the effect of fructose-in sugar and high-fructose corn syrup-on anything other than its ability in the short term to elevate blood sugar and elicit an insulin response.

In the mid-1970s, Gerald Reaven initiated the study of glycemic index to test what he cal ed the "traditional y held tenet" that simple carbohydrates are easier to digest than more complex carbohydrates "and that they therefore produce a greater and faster rise" in blood sugar and insulin after a meal.

Reaven's experiments confirmed this proposition, but he was less interested in blood sugar than in insulin, and so left this research behind. It was taken up a few years later by David Jenkins and his student Thomas Wolever, both of whom were then at Oxford University. Over the course of a year, Wolever and Jenkins tested sixty-two foods and recorded the blood-sugar response in the two hours after consumption. Different individuals responded differently, and the variation from day to day was "tremendous," as Wolever says, but the response to a specific food was stil reasonably consistent. They also tested a solution of glucose alone to provide a benchmark, which they a.s.signed a numerical value of 100. Thus the glycemic index became a comparison of the blood-sugar response induced by a particular carbohydrate food to the response resulting from drinking a solution of glucose alone. The higher the glycemic index, the faster the digestion of the carbohydrates and the greater the resulting blood sugar and insulin. White bread, they reported, had a glycemic index of 69; white rice, 72; corn flakes, 80; apples, 39; ice cream, 36. The presence of fat and protein in a food decreased the blood-sugar response, and so decreased the glycemic index.

One important implication of Jenkins and Wolever's glycemic-index research is that it provided support for Cleave's speculations on the saccharine disease. The more refined the carbohydrates, the greater the blood-sugar and insulin response. Anything that increases the speed of digestion of carbohydrates-polishing rice, for instance, refining wheat, mashing potatoes, and particularly drinking simple carbohydrates in any liquid form, whether a soda or a fruit juice-wil increase the glycemic response. Thus, the addition of refined carbohydrates to traditional diets of fibrous vegetables or meat and milk, or even fish and coconuts, could be expected to elevate blood-sugar and insulin levels in the population. And this would conceivably explain the appearance of both atherosclerosis and diabetes as diseases of civilization, through the physiological abnormalities of metabolic syndrome-glucose intolerance, hyperinsulinemia, insulin resistance, high triglycerides, low HDL, and smal , dense LDL.

Jenkins and Wolever's research, first published in 1981, led to a surprisingly vitriolic debate among diabetologists on the value of the glycemic index as a guide to control ing blood sugar. Reaven argued that the concept was worthless if not dangerous: saturated fat, he argued, has no glycemic index, and so adding saturated fat to sugar and other carbohydrates wil lower their glycemic index and make the combination appear benign when that might not quite be the case. "Ice cream has a great glycemic index, because of the fat," Reaven observed. "Do you want people to eat ice cream?" Reaven also disparaged the glycemic index for putting the clinical focus on blood sugar, whereas he considered insulin and insulin resistance the primary areas of concern. The best way for diabetics to approach their disease, Reaven insisted, was to restrict al carbohydrates.

Paradoxical y, the glycemic index appears to have had its most significant influence not on the clinical management of diabetes but on the public perception of sugar itself. The key point is that the glycemic index of sucrose is lower than that of flour and starches-white bread and potatoes, for instance-and fructose is the reason why. The carbohydrates in starches are broken down upon digestion, first to maltose and then to glucose, which moves directly from the smal intestine into the bloodstream. This leads immediately to an elevation of blood sugar, and so a high glycemic index. Table sugar, on the other hand-i.e., sucrose-is composed of both glucose and fructose. To be precise, a sucrose molecule is composed of a single glucose molecule bonded to a single fructose molecule. This bond is broken upon digestion. The glucose moves into the bloodstream and raises blood sugar, just as if it came from a starch, but the fructose can be metabolized only in the liver, and so most of the fructose consumed is channeled from the smal intestine directly to the liver. As a result, fructose has little immediate effect on blood-sugar levels, and so only the glucose half of sugar is reflected in the glycemic index.

That sugar is half fructose is what fundamental y differentiates it from starches and even the whitest, most refined flour. If John Yudkin was right that sugar is the primary nutritional evil in the diet, it would be the fructose that endows it with that singular distinction. With an eye toward primitive diets transformed by civilization, and the change in Western diets over the past few hundred years, it can be said that the single most profound change, even more than the refinement of carbohydrates, is the dramatic increase in fructose consumption that comes with either the addition of fructose to a diet lacking carbohydrates, or the replacement of a large part of the glucose from starches by the fructose in sugar.

Because fructose barely registers in the glycemic index, it appeared to be the ideal sweetener for diabetics; sucrose itself, with the possible exception of its effect on cavities, appeared no more harmful to nondiabetics, and perhaps even less so, than starches such as potatoes that were being advocated as healthy subst.i.tutes for fat in the diet. In 1983, the University of Minnesota diabetologist John Bantle reported in The New England Journal of Medicine that fructose could be considered the healthiest carbohydrate. "We see no reason for diabetics to be denied foods containing sucrose," Bantle wrote.

This became the official government position. The American Diabetes a.s.sociation stil suggests that diabetics need not restrict "sucrose or sucrose-containing foods" and can even subst.i.tute them, if desired, "for other carbohydrates in the meal plan."

In 1986, the FDA exonerated sugar of any nutritional crimes on the basis that "no conclusive evidence demonstrates a hazard." The two-hundred-page report const.i.tuted a review of hundreds of articles on the health aspects of sugar, many of which reported that sugar had a range of potentially adverse metabolic effects related to a higher risk of heart disease and diabetes. The FDA interpreted the evidence as inconclusive. Health reporters, the sugar industry, and public-health authorities therefore perceived the FDA report as absolving sugar of having any deleterious effects on our health.

The identical message was pa.s.sed along in the 1988 Surgeon General's Report on Nutrition and Health and the 1989 National Academy of Sciences Diet and Health report. Here, too, the inconclusive studies and ambiguous evidence were considered insufficient to indict sugar as a dietary evil -innocent until proven guilty. These two reports also reviewed the dietary fat/heart-disease connection, which also const.i.tuted a col ection of inconclusive studies and ambiguous evidence. Here, though, dietary fat was a.s.sumed guilty until proved innocent. And so the existence of ambiguous evidence was considered sufficient reason to condemn fat in the diet, particularly saturated fat, while the existence of ambiguous evidence was simultaneously considered reason enough to exonerate sugar.

This inst.i.tutional absolution of sugar might have been relatively innocuous had it not been coincident with the introduction of a type of sugar refined from corn, rather than sugarcane or beets, known as high-fructose corn syrup, or HFCS, and specifical y with what is technical y known as HFCS-55, a sweetener that is 55 percent fructose and 45 percent glucose and was created to be indistinguishable from sucrose by taste when used in soft drinks.

HFCS-55 entered the market in 1978. By 1985, half of the sugars consumed each year in the U.S. came from corn sweeteners, and two-thirds of that was high-fructose corn syrup. More important, the average consumption of sugars in total had started climbing steadily upward.

This rise in sugar consumption is one of the more perplexing dietary trends in the last century. Though Americans' taste in starch apparently ebbed and flowed through the twentieth century, the average yearly consumption of caloric sweeteners-a category that includes table sugar, corn sweeteners, honey, and edible syrups-remained relatively constant from the 1920s, at 110120 pounds per capita. It began to inch upward in the early 1960s, coincident with the first introduction of fructose-enhanced corn syrups. With the introduction of HFCS-55, it increased significantly. According to USDA statistics, between 1975 and 1979 Americans consumed an annual average of 124 pounds of sugars per person. By 2000, that number had jumped to almost 150 pounds. Corn sweeteners, and particularly high-fructose corn syrup, const.i.tuted virtual y every ounce of the increase. And this increase came on the heels of a period in the mid-1970s when sugar consumption per capita was decreasing, as sugar was being portrayed in the popular press as a fattening and addictive dietary nuisance.

The simplest explanation for the increase in caloric-sweetener consumption is that consumers simply failed to equate high-fructose corn syrup with the sugar that we'd been eating almost exclusively until then. Although HFCS-55 is effectively identical to sucrose upon digestion, the industry treated it, and the public perceived it, as a healthy additive, whereas sucrose carried the taint of decades of controversy. Because fructose is the predominant sugar in fruit-an apple, for instance, is roughly 6 percent fructose, 4 percent sucrose, and 1 percent glucose by weight-it is often referred to as "fruit sugar" and appears somehow healthier simply by virtue of that a.s.sociation. And, of course, fructose was perceived as healthy because it does not elevate blood sugar and has a low glycemic index.

As a consequence, high-fructose corn syrup could be used as the primary sweetener, and often the primary source of calories, in products that had the outward appearance of being healthy or natural, or were advertised as such, without revealing the products to be little more than sugar, water, and chemical flavoring. This included sports drinks such as Gatorade, the fruit juices and teas such as Snapple that appeared nationwide beginning in the late 1980s, and low-fat yogurts, which also exploded in popularity with the condemnation of fat in the diet.

By defining carbohydrate foods as good or bad on the basis of their glycemic index, diabetologists and public-health authorities effectively misdiagnosed the impact of fructose on human health. The key is the influence of glucose or fructose not on blood sugar but on the liver. Glucose goes directly into the bloodstream and is taken up by tissues and organs to use as energy; only 3040 percent pa.s.ses through the liver. Fructose pa.s.ses directly to the liver, where it is metabolized almost exclusively. As a result, fructose "const.i.tutes a metabolic load targeted on the liver," the Israeli diabetologist Eleazar Shafrir says, and the liver responds by converting it into triglycerides-fat-and then shipping it out on lipoproteins for storage. The more fructose in the diet, the higher the subsequent triglyceride levels in the blood.*58 The research on this fructose-induced lipogenesis, as it is technical y known, was carried out primarily by Peter Mayes, a biochemist at King's Col ege Medical School in London; by Shafrir at Hebrew UniversityHada.s.sah Medical School in Jerusalem; and by Sheldon Reiser and his col eagues at the USDA Carbohydrate Nutrition Laboratory in Maryland. They began in the late 1960s and worked on it through the early 1980s. "In the 1980s," says Judith Hal frisch, who worked with Reiser at the USDA, "people didn't even believe that elevated triglycerides were a risk factor for cardiovascular disease. So they didn't care that much about the increase in triglycerides. Everything was cholesterol." (Although sugar also seemed to raise cholesterol levels, particularly LDL, as would be expected for any nutrient that increased triglyceride synthesis in the liver. In 1992, John Bantle reported that LDL cholesterol in diabetic patients was elevated more than 10 percent on a high-fructose diet after a month, which is comparable to what can be achieved by saturated fats.) As Peter Mayes has explained it, our bodies wil gradual y adapt to long-term consumption of high-fructose diets, and so the "pattern of fructose metabolism" wil change over time. This is why, the more fructose in the diet and the longer the period of consumption, the greater the secretion of triglycerides by the liver. Moreover, fructose apparently blocks both the metabolism of glucose in the liver and the synthesis of glucose into glycogen, the form in which the liver stores glucose local y for later use. As a result, the pancreas secretes more insulin to overcome this glucose traffic-jam at the liver, and this in turn induces the muscles to compensate by becoming more insulin resistant. The research on this fructose-induced insulin resistance was done on laboratory animals, but it confirmed what Reiser at the USDA had observed in humans and published in 1981: given sufficient time, high-fructose diets can induce high insulin levels, high blood sugar, and insulin resistance, even though in the short term fructose has little effect on either blood sugar or insulin and so a very low glycemic index. It has also been known since the 1960s that fructose elevates blood pressure more than an equivalent amount of glucose does, a phenomenon cal ed fructose-induced hypertension.

Because sucrose and high-fructose corn syrup (HFCS-55) are both effectively half glucose and half fructose, they offer the worst of both sugars. The fructose wil stimulate the liver to produce triglycerides, while the glucose wil stimulate insulin secretion. And the glucose-induced insulin response in turn wil prompt the liver to secrete even more triglycerides than it would from the fructose alone, while the insulin wil also elevate blood pressure apart from the effect of fructose. "This is real y the harmful effect of sucrose," says Mayes, "over and above fructose alone."

The effect of fructose on the formation of advanced glycation end-products-AGEs, the haphazard glomming together of proteins in cel s and tissues -is worrisome as wel . Most of the research on AGE acc.u.mulation in humans has focused on the influence of glucose, because it is the dominant sugar in the blood. Glucose, however, is the least reactive of al sugars, the one least likely to attach itself without an enzyme to a nearby protein, which is the first step in the formation of AGEs. As it turns out, however, fructose is significantly more reactive in the bloodstream than glucose, and perhaps ten times more effective than glucose at inducing the cross-linking of proteins that leads to the cel ular junk of advanced glycation end-products. Fructose also leads to the formation of AGEs and cross-linked proteins that seem more resistant to the body's disposal mechanisms than those created by glucose. It also increases markedly the oxidation of LDL particles, which appears to be a necessary step in atherosclerosis.

This research on the health effects of fructose began to coalesce in the mid-1980s, just as nutritionists were disseminating the notion that fructose was particularly harmless because of its low glycemic index. And this official opinion has proven hard to sway.

Take, for example, the British Committee on Medical Aspects of Food Policy (known commonly as COMA), which in 1989 released a report ent.i.tled Dietary Sugars and Human Disease, auth.o.r.ed by a dozen of the nation's leading nutritionists, physiologists, and biochemists and chaired by Harry Keen, who is among the most renowned British diabetologists. The COMA report discussed the evidence, including the research of Reiser, Reaven, and others, and then concluded that the health effects of sugar were insignificant. The report did so, however, with a series of contradictory a.s.sumptions. First, Keen and his col eagues concluded that the implications of fructose-induced insulin resistance and elevated triglycerides are limited to a "relatively smal group of people with metabolic disorders [that] includes people with diabetes and those with certain rare inherited disorders." And so, with the exception of this smal percentage of the population, they noted, yearly sugar consumption at 1986 levels, estimated at roughly a hundred pounds per capita in the United Kingdom, "carries no special metabolic risks." On the other hand, they then explained, sugar consumption does carry risk for those "members of the population consuming more than about 200 g per day," which is 160 pounds per year, or only slightly more (.4 ounce per day) than what the average American was eating in the year 2000 (not the top 1020 percent, but the average). They next suggested that those individuals with high triglycerides, a proportion that remains unspecified but might const.i.tute the great majority of al individuals with coronary-artery disease, should restrict their consumption of added sugars to twenty to forty pounds per year, or equivalent to the amount consumed in the U.K. in the early years of the Victorian era.

Al of this was then summed up in the single statement-echoing the sentiments of the FDA Task Force, the National Academy of Sciences Diet and Health report, and The Surgeon General's Report on Nutrition and Health, which preceded it-that dietary sugar consumption could not be held responsible for causing disease: "The panel concluded that current consumption of sugars, particularly sucrose, played no direct causal role in the development of cardiovascular...disease, of essential hypertension, or of diabetes mel itus...."

Four years later, The American Journal of Clinical Nutrition dedicated an entire issue to the deleterious effects of dietary fructose. A common refrain throughout the issue was the need for research that would establish at what level of sugar consumption the effects discussed-the elevation of blood pressure and triglycerides, increased insulin resistance, and even accelerated formation of advanced glycation end-products-would lead to disease.

"Further studies are clearly needed to determine the metabolic alteration that may take place during chronic fructose or sucrose feeding," as the Swiss physiologists Luc Tappy and Eric Jequier wrote.

In 2002, the Inst.i.tute of Medicine of the National Academies of Science released its two-volume report on Dietary Reference Intakes (subt.i.tled Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids), and spent twenty pages discussing the possible adverse effects of sucrose and high-fructose corn syrup. It then concluded that there was stil "insufficient evidence" to set up an upper limit for sugar consumption in the healthy diet. Nor did the IOM perceive any reason to pursue further research on fructose or sucrose or high-fructose corn syrup and so, perhaps, discover sufficient evidence. In early 2007, the National Inst.i.tutes of Health was funding at most half a dozen research projects that addressed, even peripheral y, the health effects of dietary fructose, meaning sugar and high-fructose corn syrup in the diet.

Over the years, what little research has been done on fructose metabolism has been carried out primarily by biochemists, who have had little motivation, other than perhaps personal health, to pay attention to the nutrition literature-again, an effect of specialization. Their own articles, moreover, are published in biochemistry journals, and have little influence on the nutrition and public-health communities. For this reason, observations on the potential dangers of fructose have managed to remain dissociated from discussions of sugar itself and the role of sucrose and high-fructose corn syrup in modern diets. After the ridicule that John Yudkin received for the work that culminated in his anti-sugar polemic, Pure, White and Deadly-and after the FDA decided that "no conclusive evidence demonstrates a hazard" from sucrose-few researchers have appeared wil ing even to contemplate the possibility that sugar consumption could have harmful consequences beyond perhaps causing cavities and contributing to obesity.

Chapter Thirteen.

DEMENTIA, CANCER, AND AGING.

The bottom line is pretty irrefutable: What is good for the heart is good for the brain.

RUDOLPH TANZI AND ANN PARSON, Decoding Darkness: The Search for the Genetic Causes of Alzheimer's Disease, 2000 WHEN IT COMES TO THE CAUSE of chronic disease, as we discussed earlier, the carbohydrate hypothesis rests upon two simple propositions. First, if our likelihood of contracting a particular disease increases once we already have Type 2 diabetes or metabolic syndrome, then it's a reasonable a.s.sumption that high blood sugar and/or insulin is involved in the disease process. Second, if blood sugar and insulin are involved, then we have to accept the possibility that refined and easily digestible carbohydrates are as wel .

This applies to Alzheimer's disease and cancer, too, since both diabetes and metabolic syndrome are a.s.sociated with an increased incidence of these two il nesses. In both cases, critical steps in the disease process have been linked unambiguously to insulin and blood sugar, and the relevant research is now beginning to influence the mainstream thinking in these fields.

Though the characteristic dementia and brain lesions of Alzheimer's were first described a century ago, the disease only recently captured the attention of the research community. In 1975, when the NIH was supporting hundreds of research projects on atherosclerosis and cholesterol metabolism, it was funding fewer than a dozen on Alzheimer's and what was then cal ed senile dementia. This number rose gradual y through the end of the 1970s. Between 1982 and 1985, the number of Alzheimer's-related research projects funded by the NIH quintupled.

It took another decade for researchers to begin reporting that heart disease and Alzheimer's seem to share risk factors: hypertension, atherosclerosis, and smoking are al a.s.sociated with an increased risk of Alzheimer's, as is the inheritance of a particular variant of a gene cal ed apolipoprotein E4 (apo E4) that also increases the risk of cardiovascular disease.*59 This in turn led to the notion that what's good for the heart is good for the brain, but that, of course, depends on our understanding of what exactly is good for the heart. Because Alzheimer's researchers, like diabetologists, a.s.sume that Keys's fat-cholesterol hypothesis is supported by compel ing evidence, they wil often suggest that cholesterol and saturated fat play a role in Alzheimer's as wel .

But if coronary heart disease is mostly a product of the physiological abnormalities of metabolic syndrome, as the evidence suggests, then this implicates insulin, blood sugar, and refined carbohydrates instead, a conclusion supported by several lines of research that began to converge in the last decade.

A handful of studies have suggested that Alzheimer's is another disease of civilization, with a pattern of distribution similar, if not identical, to heart disease, diabetes, and obesity. j.a.panese Americans, for instance, develop a pattern of dementia-the ratio of Alzheimer's dementia to the stroke-related condition known as vascular dementia-that is typical y American; when j.a.panese immigrate to the United States, their likelihood of developing Alzheimer's disease increases considerably, while their risk of developing vascular dementia decreases.