Though Brooks reported that he could prevent his albino mice from growing obese, he could do so only by imposing "severe and permanent" food restriction. If he subjected them to "long continued limitation of food," the animals would lose some weight, but they would never lose the drive to fatten or the hunger that went with it. Periods of fasting, Brooks noted, were "fol owed by an augmentation of appet.i.te and development of a greater degree of obesity than had been attained before fasting." And so Brooks's lesioned mice, as Hilde Bruch might have noted, were acting exactly like normal healthy humans and obese humans after a semi-starvation diet. These VMH lesions also resulted in changes in the reproductive cycles of the animals, and in their normal nocturnal eating patterns, which Ranson and Hetherington had also reported; once the animals became obese, they slept more than normal animals, al of which suggested that the VMH lesions had profound effects on the entire homeostatic system and could not be written off as simply affecting hunger and thus food intake.

When physiologists began studying animal hibernation in the 1960s, they again demonstrated this decoupling of food intake from weight gain.

Hibernating ground squirrels wil double their body weight in late summer, in preparation for the winter-long hibernation. But these squirrels wil get just as fat even when kept in the laboratory and not al owed to eat any more in August and September than they did in April. The seasonal fat deposition is genetical y programmed-the animals wil accomplish their task whether food is abundant or not. If they didn't, a single bad summer could wipe out the species.

This same decoupling of food intake and weight would also be demonstrated when researchers studied what are now known as dietary models of obesity. Certain strains of rats wil grow obese on very high-fat diets, and others on high-sugar diets. In both cases, the animals wil get fatter even if they don't consume any more calories than do lean controls eating their usual lab chow. This same decoupling occurs in animals that are regaining weight after lengthy periods of fasting. "It doesn't matter how long you food-deprive the animal," said Irving Faust, who did this work in the 1970s; "the recovery of body weight is not connected to the amount of food eaten during the recovery phase." And this same decoupling of calories and weight has also been made consistently, if not universal y, in the recent research on transgenic animals, in which specific genes are manipulated.

What may have been the most enlightening animal experiments were carried out in the 1970s by physiologists studying weight regulation and reproduction. In these experiments, the researchers removed the ovaries from female rats. This procedure effectively serves to shut down production of the female s.e.x hormone estrogen (technical y estradiol). Without estrogen, the rats eat voraciously, dramatical y decrease physical activity, and quickly grow obese. When the estrogen is replaced by infusing the hormone back into these rats, they lose the excess weight and return to their usual patterns of eating and activity. The critical point is that when researchers remove the ovaries from these rats, but restrict their diets to only what they were eating before the surgery, the rats become just as obese, just as quickly; the number of calories consumed makes little difference.

George Wade, the University of Ma.s.sachusetts biologist who did much of this research, described it as a "revelation" that obesity could be brought on without overeating, just as Pennington had described it as revelatory that weight could be lost without undereating. "If you keep the animals' food intake constant and manipulate the s.e.x hormones, you stil get substantial changes in body weight and fat content," Wade said. Another consequence of removing the ovaries was that the rats h.o.a.rded more food in their cages, which is a.n.a.logous to storing excess calories as fat. Infusing estrogen back into these rats suppressed the food-h.o.a.rding, just as it prompted weight loss. "The animals overeat and get fat," said Tim Bartness, who worked on this research as part of his doctoral studies with Wade in the 1970s, "but they are overeating because they're socking al the calories away into adipose tissue and they can't get to those calories. They're not getting fat because they're overeating; they're overeating because they're getting fat. It's not a trivial difference. The causality is quite different."

One critical idea here is that survival of a species is dependent on successful reproduction, and that in turn depends first and foremost on the availability of food. Fat acc.u.mulation, energy balance, and reproduction are al intimately linked, and al regulated by the hypothalamus. This is why food deprivation suppresses ovulation, and why the same kind of hormonal control of reproduction ensures that herbivores, such as sheep, tend to give birth in the springtime, when food is available. The link between food availability and reproduction was something that Charles Darwin had also observed: "Hard living...r.e.t.a.r.ds the period at which animals conceive," he wrote.

The lesson of these animal experiments is that understanding energy balance and weight control requires Claude Bernard's harmonic-ensemble perspective of homeostasis: an appreciation of the entire organism and the entire homeostatic web of hormonal regulation. "Fertility is linked to food supply, physical exercise involved in foraging for food and avoiding predators, and energy expenditure a.s.sociated with temperature regulation and other physiological processes," Wade explains. These functions are control ed by a tight orchestration of both s.e.x hormones and those hormones that control the "part.i.tioning and utilization of metabolic fuels," and this is accomplished in ways that are "reciprocal, redundant and ubiquitous."

The idea that obesity in humans is caused, as it is in animals, by a defect in the homeostatic maintenance of energy distribution and fat metabolism-that we overeat because we're getting fat, and not vice versa-barely survived into the second half of the twentieth century, although the evidence has always supported it.

This homeostatic hypothesis effectively vanished from the mainstream thinking on human (as opposed to animal) obesity with the coming of World War I . The war destroyed the German and Austrian community of clinical investigators, who had done the most perceptive thinking about the causes of obesity and had a tradition of rigorous scientific research dating back two hundred years. In the United States, it resulted in a suspension of obesity research that lasted for most of a decade. Meanwhile, Stephen Ranson had died, Hugo Rony and Julius Bauer retired. The generation of physiologists who had founded the field of nutrition in the United States and actual y studied human metabolism disappeared with them. Francis Benedict's Nutrition Laboratory at the Carnegie Inst.i.tution did contract work for the armed services during the war and then was shut down in 1946. The Russel Sage Inst.i.tute of Pathology, where Graham Lusk and Eugene Du Bois did their research, was also gone by the 1950s. Lusk himself died in 1932, Francis Benedict retired in 1937. Du Bois retired four years later.

Among the few investigators whose careers spanned the war years, Louis Newburgh was the most influential and conspicuous. As late as 1948, Newburgh was stil promoting his perverted-appet.i.te hypothesis of obesity. The first obesity textbook published after the war, Obesity... (1949), by Edward Rynearson and Clifford Gastineau, would be considered the standard text on obesity for twenty years. It faithful y communicated Newburgh's belief that obesity is caused by overeating. Any suggestion to the contrary, wrote Rynearson and Gastineau, const.i.tuted little more than "an excuse for avoidance of the necessary corrective measures."

An entire generation of young researchers and clinicians effectively started the study of obesity from scratch after the war. They did so with little concern for whatever understanding had been achieved before they arrived, and so they embraced a hypothesis of causation that flew in the face of much of the evidence. The inst.i.tutionalized skepticism and meticulous attention to experimental detail that are necessary to do good science-"being ruthless in self-criticism and...taking pains in verifying facts," as the n.o.bel laureate chemist Hans Krebs said-had also been left behind.

Chapter Twenty-two.

THE CARBOHYDRATE HYPOTHESIS, II: INSULIN.

Every woman knows that carbohydrate is fattening.

REGINALD Pa.s.sMORE AND YOLA SWINDELLS, British Journal of Nutrition, 1963 The fact that insulin increases the formation of fat has been obvious ever since the first emaciated dog or diabetic patient demonstrated a fine pad of adipose tissue, made as a result of treatment with the hormone.

REGINALD HAIST AND CHARLES BEST, The Physiological Basis of Medical Practice, 1966 IN 1929, WHEN LOUIS NEWBURGH FIRST rejected the possibility of an "endocrine abnormality" as the cause of obesity, and insisted instead that al fat people had a perverted appet.i.te, hormones were stil widely known as "internal secretions" and endocrine glands as "ductless glands." The first purification of growth hormone had been only nine years earlier, the purification of insulin only eight years before. In 1955, when The Journal of the American Medical a.s.sociation declared unconditional y that those "theories that attributed obesity to an endocrine disturbance have been shown to be erroneous," it was five years before Rosalyn Yalow and Solomon Berson would publish the details of the first method for measuring the insulin level in the blood, and a few more years after that before the ensuing revelations that obesity was a.s.sociated with the endocrine disturbances and abnormalities of hyperinsulinemia and insulin resistance.

In other words, the editors at JAMA-and the clinical investigators they represented-were declaring that hormones, as a rule, play little role in the genesis of obesity, even before the relevant hormones could be measured accurately in the human bloodstream. In fact, it's hard to imagine, as Julius Bauer noted, that hormones wouldn't play a role. Here again we have that familiar scenario we first discussed with regard to dietary fat and heart disease.

Once the "truth" has been declared, even if it's based on incomplete evidence, the overwhelming tendency is to interpret al future observations in support of that preconception. Those who know what the answer is lack the motivation to continue looking for it. Entire fields of science may then be ignored, on the a.s.sumption that they can't possibly be relevant.

In 1968, Jean Mayer pointed out that obesity researchers may have "eliminated" hormones "from legitimate consideration" as a cause of obesity, or so they believed, but the evidence continued to acc.u.mulate just the same. Researchers had demonstrated that insulin seemed to have a dramatic effect on hunger, that insulin was the primary regulator of fat deposition in the adipose tissue, and that obese patients had chronical y high levels of insulin. Other hormones, such as adrenaline, had been shown to increase the mobilization of fat from the fat cel s. "It is probable that different concentrations of these hormones in blood are characteristic of different body types and fat contents," Mayer wrote.

At the beginning of this century, when hormones were first discovered, it was commonly believed that obesity would be found to be due to the absolute excess or deficiency of a single hormone. When this was found to be almost never true, the popular medical position swung to the other extreme: "obesity is almost never due to hormonal disturbances; it is almost always due to overeating." Actual y, the reasonable position ought to be: "in order to be obese, you always have to eat more than you expend for a certain period. How often this is due to a slight shift of relative or absolute hormone concentrations, each one of which is in the 'normal' range, we don't know."

Among the hormones that play a role in regulating fat metabolism and thus potential y play a causative role in obesity, insulin was always an obvious choice. Some failure in what clinicians a century ago cal ed the insular*108 apparatus of the pancreas is the fundamental defect in diabetes, and diabetes is intimately a.s.sociated with obesity in those who develop the disease as adults, and with emaciation, which was the end stage of the disease in the pre-insulin era. In 1905, Carl von Noorden invoked this intimate a.s.sociation between diabetes and weight to formulate the third of his speculative hypotheses of obesity, what he cal ed diabetogenous obesity. His ideas were remarkably prescient. They received little attention because insulin had not yet been discovered, let alone the technology to measure it.

Von Noorden suggested that obesity and diabetes are different consequences of the same underlying defects in the mechanisms that regulate carbohydrate and fat metabolism. In severe diabetes (Type 1), he noted, the patients are unable either to utilize blood sugar as a source of energy or to convert it into fat and store it. This is why the body al ows the blood sugar to overflow into the urine, which is a last resort since it wastes potential y valuable fuel. The result is glycosuria, the primary symptom of diabetes. These diabetics must be incapable of storing or maintaining fat, von Noorden noted, because they eventual y become emaciated and waste away. In obese patients, on the other hand, the ability to utilize blood sugar is impaired, but not the ability of the body to convert blood sugar into fat and store it. "Obese individuals of this type have already an altered metabolism for sugar," von Noorden wrote, "but instead of excreting the sugar in the urine, they transfer it to the fat-producing parts of the body, whose tissues are stil wel prepared to receive it." As the ability to burn blood sugar for energy further deteriorates and "the storage of the carbohydrates in the fat ma.s.ses [also suffers] a moderate and gradual y progressing impairment," sugar appears in the urine, and the patient becomes noticeably diabetic. Using the modern terminology, this is the route from obesity to Type 2 diabetes. "The connection between diabetes and obesity," as von Noorden put it, "ceases in the light of my theory to be any longer an enigmatical relation, and becomes a necessary consequence of the relationship discovered in the last few years between carbohydrate transformation and formation of fat."

After the discovery of insulin in 1921, the potential role of insulin as a fattening hormone would become a long-running controversy. Those physicians who believed, as Louis Newburgh did without reservation, that obesity was an eating disorder, rejected the idea that insulin could fatten humans, if for no other reason than that this suggested the existence of a defective hormonal mechanism that could lead to obesity. The evidence, however, suggested exactly that. When insulin was injected into diabetic dogs in the laboratory, or diabetic human patients in the clinic, they put on weight and body fat. As early as 1923, clinicians were reporting that they had successful y used insulin to fatten chronical y underweight children-patients who would be diagnosed today as anorexic-and to increase their appet.i.te in the process.

In 1925, Wilhelm Falta, a student of von Noorden and a pioneer of the science of endocrinology in Europe, began using insulin therapy to treat underweight and anorexia in adults as wel . Falta had argued, even in the pre-insulin era, that whatever pancreatic hormone was absent or defective in diabetes governed not only the use of carbohydrates for fuel, but also the a.s.similation of fat in adipose tissue. "A functional y intact pancreas is necessary for fattening," Falta wrote. He also noted that the only way to fatten anyone efficiently was to include "abundant carbohydrates in the diet." Otherwise, the body would adjust to eating "very much more than the appet.i.te real y craves," by either lessening appet.i.te stil further or creating "an increased demand for movement." The only way to get around this natural balance of intake and expenditure is by increasing the secretion from the pancreas. "We can conceive," Falta speculated, "that the origin of obesity may receive an impetus through a primarily strengthened function of the insular apparatus, in that the a.s.similation of larger amounts of food goes on abnormal y easily, and hence there does not occur the setting free of the reactions that in normal individuals work against an ingestion of food which for a long time supersedes the need." After the discovery of insulin, Falta reported that giving it to patients would increase their appet.i.te for carbohydrates specifical y, and the carbohydrates in turn would stimulate the patient's own insulin production. It would create a vicious cycle-although, in the case of anorexic and underweight patients, one that might return them to a normal appet.i.te and normal weight.

By the 1930s, clinicians throughout Europe and the United States had taken to using insulin therapy to fatten their pathological y underweight patients.

These patients could gain as much as six pounds a week eating meals "rich in carbohydrates" after receiving injections of smal doses of insulin, reported Rony, who used insulin therapy on seven anorexic patients in his own clinic; it worked on five of them. None of these patients had been able to gain weight, but now they added an average of twenty pounds each in three months. "Al reported a more or less p.r.o.nounced increase of appet.i.te," Rony wrote, "and occasional strong feelings of hunger." Until the 1960s, insulin was also used to treat severe depression and schizophrenia. Among the more renowned patients subjected to what was then cal ed insulin-shock therapy was the Princeton mathematician John Nash, made famous by Sylvia Nasar's 1998 biography, A Beautiful Mind. Its efficacy for treating mental il ness was debatable, but as Nasar observed, "al the patients gained weight." Another memorable recipient was the poet Sylvia Plath, who experienced a "drastic increase in weight" on the treatment. (In her autobiographical novel, The Bell Jar, Plath's protagonist, Esther Greenwood, gains twenty pounds on insulin therapy-"I just grew fatter and fatter," she says.) Insulin's fattening properties have long been particularly obvious to diabetics and the physicians who treat them. Because diabetics wil gain weight with insulin therapy, even those who are obese to begin with, clinicians have always had difficulty convincing their patients to continue taking their insulin. When they start to fatten, they natural y want to slack off on the therapy, so the need to control blood sugar competes with the desire to remain lean, or at least relatively so. This is also a clinical dilemma, because the weight gain wil also increase the risk of heart disease. In the chapter on insulin therapy in the 1994 edition of Joslin's Diabetes Mellitus, the Harvard diabetologist James Rosenzweig portrayed this insulin-induced weight gain as uncontroversial: "In a number of studies of patients treated with insulin for up to 12 months, weight gains of 2.0 to 4.5 kg [roughly four to ten pounds] were reported...." This weight gain, he wrote, then leads to "the often-cited vicious cycle of increased insulin resistance, leading to the need for more exogenous insulin, to further weight gain, which increases the insulin resistance even more."*109 If insulin fattens those who receive it, as the evidence suggests, then how does it work? The prewar European clinicians who used insulin therapy to treat anorexics accepted the possibility, as Falta suggested, that the hormone can directly increase the acc.u.mulation of fat in the fat tissues. Insulin was "an excel ent fattening substance," Erich Grafe wrote in Metabolic Diseases and Their Treatment. Grafe believed that the fattening effect of insulin is likely "due to improved combustion of carbohydrate and increased synthesis of glycogen and fat." In the United States, however, the conventional wisdom came from Louis Newburgh and his col eagues at the University of Michigan. When insulin increases weight, Newburgh said, it does so either through the power of suggestion-a placebo effect-or by a reduction of blood sugar to the point where the patient eats to avoid very low blood sugar (hypoglycemia) and the accompanying symptoms of dizziness, weakness, and convulsions.

When Rony reviewed the experimental and clinical reports in 1940, he considered any conclusion to be premature. Because obese individuals tend to have high blood sugar, rather than low, Rony said, it was hard to imagine how insulin, which lowered blood sugar, could cause obesity. "Stil ," he noted, "it might be possible that in obese subjects a latent or conditional form of hyperinsulinism exists which would promote fat deposition without causing hypoglycemia." This was not supported by conclusive evidence, he added, and so it "remains, for the time being, at best a working hypothesis."

Only Newburgh's interpretation of the evidence, however (and only the obesity research community in the United States), survived the war years.

Afterward, clinical investigators would state unambiguously-as Edward Rynearson and Clifford Gastineau did in their 1949 clinical manual Obesity...

-that insulin puts weight on only by lowering blood sugar to the point where patients overeat to remain conscious. This hypoglycemia was considered a rare pathological condition, one with no relevance to everyday life, and so only in that condition were elevated insulin levels to be considered a causal agent in weight gain and common obesity.

In 1992, the University of Texas diabetologist Denis McGarry published an article in Science with the memorably idiosyncratic t.i.tle "What If Minkowski Had Been Ageusic? An Alternative Angle on Diabetes." The German physiologist Oskar Minkowski was the first to identify the role of the pancreas in diabetes. The word "ageusic" refers to a condition in which the sense of taste is absent. "Legend has it," McGarry wrote, "that on a momentous day in 1889 Oskar Minkowski noticed that urine col ected from his pancreatectomized*110 dogs attracted an inordinate number of flies. He is said (by some) to have tasted the urine and to have been struck by its sweetness. From this simple but astute observation he established for the first time that the pancreas produced some ent.i.ty essential for control of the blood sugar concentration, which, when absent, resulted in diabetes mel itus." Some thirty years later, when Frederick Banting and Charles Best in Toronto identified insulin as the relevant pancreatic secretion, McGarry wrote, they natural y did so in the context of Minkowski's observations about blood sugar, and thus "diabetes mel itus has been viewed ever since as a disorder primarily a.s.sociated with abnormal glucose metabolism." But if Minkowski had been ageusic and so missed the sweet taste of the urine, McGarry speculated, he might have noticed instead the smel of acetone, which is produced in the liver from the conversion of fat into ketone bodies. "He would surely have concluded that removal of the pancreas causes fatty acid metabolism to go awry," McGarry wrote. "Extending this hypothetical scenario, the major conclusion of Banting's work might have been that the preeminent role of insulin is in the control of fat metabolism."

McGarry's parable focused on diabetes, but the point he made extends to virtual y everything having to with insulin. Just as diabetes has traditional y been perceived as a disorder of carbohydrate metabolism-even though fat metabolism is also dysfunctional-insulin has always been perceived as a hormone that primarily functions to regulate blood sugar, though, as we've discussed, it regulates the storage and use of fat and protein in the body as wel . Because blood sugar could be measured easily through the first half of the twentieth century, but not yet the fats in the blood, the focus of research rested firmly on blood sugar.

From the 1920s through the 1960s, a series of discoveries in the basic science of fat metabolism led to a revolution in the understanding of the role of insulin and the regulation of fat tissue in the human body. This era began with a handful of naive a.s.sumptions: that fat tissue is relatively inert (a "garbage can," in the words of the Swiss physiologist Bernard Jeanrenaud); that carbohydrates are the primary fuel for muscular activity (which is stil commonly believed today); and that fat is used for fuel only after being converted in the liver into supposedly toxic ketone bodies. The forty years of research that fol owed would overturn them al -but it would have effectively no influence on the mainstream thinking about human obesity.

Those who paid attention to this research either had no influence themselves-Alfred Pennington comes to mind-or were so convinced that obesity is caused by overeating that they couldn't imagine why the research would be relevant. From the 1950s onward, clinical investigators studying and treating obese patients, as Hilde Bruch commented, seemed singularly uninterested in this research. "Until recently, knowledge of the synthesis and oxidation of fat was quite rudimentary," Bruch wrote in 1957. "As long as it was not known how the body builds up and breaks down its fat deposit, the ignorance was glossed over by simply stating that food taken in excess of body needs was stored and deposited in the fat cel s, the way potatoes are put into a bag.

Obviously, this is not so." By 1973, after details of the regulation of fat metabolism and storage had been worked out in fine detail, Bruch found it "amazing how little of this increasing awareness...is reflected in the clinical literature on obesity."

There are three distinct phases of the revolution that converged by the mid-1960s to overturn what Bruch cal ed the "the time-honored a.s.sumption that fat tissue is metabolical y inert," and the accompanying conviction that fat only enters the fat tissue after a meal and only leaves it when the body is in negative energy balance.

The first phase began in the 1920s, when biochemists realized that the cel s of adipose tissue have distinct structures and are not, as was previously believed, simply connective tissue stuffed with a droplet of oily fat. Researchers then demonstrated that the adipose tissue is interlaced with blood vessels such that "no marked quant.i.ty of fat cel s escapes close contact with at least one capil ary," and that the fat cel s and these blood vessels are regulated by "abundant" nerves running from the central nervous system.

This led to the revelation that the fat in the cel s of the adipose tissue is in a continual state of flux. This was initial y the work of a German biochemist, Rudolf Schoenheimer. In the early 1930s, while working at the University of Freiburg, Schoenheimer demonstrated that animals continual y synthesize and degrade their own cholesterol, independent of the amount of cholesterol in the diet. After Hitler came to power in January 1933, Schoenheimer moved to New York, where he went to work at Columbia University. It was in New York that Schoenheimer col aborated on the development of a technique for measuring serum cholesterol and, by doing so, launched the medical profession's obsession with cholesterol levels. Then, with David Rittenberg, he developed the technique to label or tag molecules with a heavy form of hydrogen known as deuterium*111 so that their movement through the metabolic processes of the body could be fol owed. Schoenheimer and Rittenberg put this technique to work studying the metabolism of fat, protein, and carbohydrates in the body.

Among their discoveries is that both dietary fat and a considerable portion of the carbohydrates we consume are stored as fat-or, technical y, triglycerides-in the adipose tissue before being used for fuel by the cel s. These triglycerides are then continuously broken down into their component fatty acids, released into the bloodstream, moved to and from organs and tissues, regenerated, and merged with fatty acids from the diet to reform a mixture of triglycerides in the fat cel s that is, as Schoenheimer put it, "indistinguishable as to their origin." Fat stored as triglycerides in the adipose tissue, and the fatty acids and triglycerides moving through the bloodstream are both part of the same perpetual cycle of fat metabolism. "Mobilization and deposition of fat go on continuously, without regard to the nutritional state of the animal," as the Israeli biochemist Ernst Wertheimer explained in 1948, in a seminal review of this new science of fat metabolism.112 "The 'cla.s.sical theory' that fat is deposited in the adipose tissue only when given in excess of the caloric requirement has been final y disproved," Wertheimer wrote. Fat acc.u.mulates in the adipose tissue when these forces of deposition exceed those of mobilization, he explained, and "the lowering of the fat content of the tissue during hunger is the result of mobilization exceeding deposition."

The control ing factors in this movement of fat to and from the fat tissue have little to do with the amount of fat present in the blood, thus little to do with the quant.i.ty of calories consumed at the time. Rather, they must be control ed, Wertheimer wrote, by "a factor acting directly on the cel ," the kind of hormonal and neurological factors that Julius Bauer had discussed. Over the next decade, investigators would begin to refer to these factors that increase the synthesis of fat from carbohydrates and the deposition of fat in the adipose tissue as lipogenic, and those that induce the breakdown of fat in the adipose tissue and its subsequent release into the circulation as lipolytic.

The second phase of this revolution began in the 1930s, with the work of Hans Krebs, who showed how our cel s convert nutrients in the bloodstream into usable energy. The Krebs cycle, for which Krebs shared the n.o.bel Prize in Medicine in 1953, is a series of chemical reactions that generate energy in the mitochondria of cel s, which are those compartments commonly referred to as the "power plants" of the cel s. The Krebs cycle starts with the breakdown products of fat, carbohydrates, and protein and then transforms them into a molecule known as adenosine triphosphate, or ATP, which can be thought of as a kind of "energy currency," in that it carries energy that can be used at a later time.*113 This cycle of reactions wil generate energy whether the initial fuel is fat, carbohydrates, or protein. Indeed, Krebs had initiated his research a.s.suming, as was common at the time, that carbohydrate was "the main energy source of muscle tissue." But he came to realize that fat and protein also supply fuel for muscle tissue, and that there was no reason why carbohydrates should be the preferred fuel. "Al three major const.i.tuents of food supply carbon atoms...for combustion," he wrote.

By 1950, the addition of the Krebs cycle to the revelations about fat metabolism from Schoenheimer and others provided the foundation for understanding the fundamental mechanisms that a.s.sure a constant supply of energy to our tissues and organs, regardless of how the demand might change in response to the environment and over the course of seconds, hours, days, or seasons. It is based on a generator-the Krebs cycle-that burns fat, carbohydrates, and protein with equal facility, and then a supply chain from the adipose tissue that ensures the circulation of fuels at a level that wil always be more than adequate for the needs at hand. "The high degree of metabolic activities present in the fat tissues," as Hilde Bruch explained, "becomes understandable as necessary for a continuous reserve for energy requirements. Instead of a savings account for unneeded surplus, as fat deposits have commonly been described, a coin purse would be a far closer a.n.a.logy. Fat tissues contain the ready cash for al the expenditures of the organism. Only when the organism does not or cannot draw on the ready cash for its daily business is it put into depots, and excessive replenishment, through overeating, takes place."

To understand the path of events that leads to obesity, "the big question," as Bruch noted, was "why the metabolism is shifted in the direction of storage away from oxidation?" Why is fat deposited in the adipose tissue to acc.u.mulate in excess of its mobilization for fuel use? Once again, this has little to do with calories consumed or expended, but addresses the questions of how the cel s utilize these calories and how the body regulates its balance between fat deposition and mobilization, between lipogenesis (the creation of fat) and lipolysis (the breakdown of triglycerides into fatty acids, their escape from the fat tissue, and their subsequent use as fuel). "Since it is now a.s.sumed that the genes and enzymes are closely a.s.sociated," Bruch wrote in 1957, "it is conceivable that people with the propensity for fat acc.u.mulation have been born with enzymes that are apt to facilitate the conversion of certain reactions in that direction."

The third phase of this research final y established the dominant role of fatty acids in supplying energy for the body, and the fundamental role of insulin and adipose tissue as the regulators of energy supply. As early as 1907, the German physiologist Adolf Magnus-Levy had noted that during periods of fasting between meals "the fat streams from the depots back again into the blood...as if it were necessary for the immediate needs of the combustion processes of the body." A decade later, Francis Benedict reported that blood sugar provides only a "smal component" of the fuel we use during fasting, and this drops away to "none at al " if our fast continues for more than a week. In such cases, fat wil supply 85 percent of our energy needs, and protein the rest, after its conversion to glucose in the liver. Stil , because the brain and central nervous system typical y burn 120 to 130 grams of glucose a day, nutritionists insisted (as many stil do) that carbohydrates must be our primary fuel, and they remained skeptical of the notion that fat plays any role in energy balance other than as a long-term reserve for emergencies.

Among physiologists and biochemists, any such skepticism began to evaporate after Wertheimer's review of fat metabolism appeared in 1948. It vanished after the 1956 publication of papers by Vincent Dole at Rockefel er University, Robert Gordon at NIH, and Sigfrid Laurel of the University of Lund in Sweden that reported the development of a technique for measuring the concentration of fatty acids in the circulation. Al three articles suggested that these fatty acids were the form in which fat is burned for fuel in the body. The concentration of fatty acids in the circulation, they reported, is surprisingly low immediately after a meal, when blood-sugar levels are highest, but then increases steadily in the hours that fol ow, as the blood sugar ebbs. Injecting either glucose or insulin into the circulation diminishes the level of fatty acids almost immediately. It's as though our cel s have the option of using fatty acids or glucose for fuel, but when surplus glucose is available, as signaled by rising insulin or blood-sugar levels, the fatty acids are swept into the fat tissue for later use. The concentration of circulating fatty acids rises and fal s in "relation to the need" for fuel, wrote Gordon. And because injections of adrenaline cause a flooding of the circulation with fatty acids, and because adrenaline is natural y released by the adrenal glands as an integral part of the flight-fight response, Gordon suggested that the concentration of fatty acids also rises in relation to "the antic.i.p.ated need" for fuel.

In 1965, the American Physiological Society published an eight-hundred-page Handbook of Physiology dedicated to the latest research on adipose-tissue metabolism. As this volume doc.u.mented, several fundamental facts about the relationship between fat and carbohydrate metabolism had become clear. First, the body wil burn carbohydrates for fuel, as long as blood sugar is elevated and the reserve supply of carbohydrates stored as glycogen in the liver and muscles is not being depleted. As these carbohydrate reserves begin to be tapped, however, or if there's a sudden demand for more energy, then the flow of fatty acids from the fat tissue into the circulation accelerates to take up the slack. Meanwhile, a significant portion of the carbohydrates we consume and al of the fat wil be stored as fat in our fat cel s before being used for fuel. It's this stored fat, in the form of fatty acids, that wil then provide from 50 to 70 percent of al the energy we expend over the course of a day. "Adipose tissue is no longer considered a static tissue," wrote the Swiss physiologist Albert Renold, who coedited the Handbook of Physiology; "it is recognized as what it is: the major site of active regulation of energy storage and mobilization, one of the primary control mechanisms responsible for the survival of any given organism."

Since the excessive acc.u.mulation of fat in the fat tissue is the problem in obesity, we need to understand this primary control mechanism. This means, first of al , that we have to appreciate the difference between triglycerides and free fatty acids. They're both forms fat takes in the human body, but they play very different roles, and these are tied directly to the way the oxidation and storage of fats and carbohydrates are regulated.

When we talk about the fat stored in the adipose tissue or the fats in our food, we're talking about triglycerides. Oleic acid, the monounsaturated fat of olive oil, is a fatty acid, but it is present in oils and meats in the form of a triglyceride. Each triglyceride molecule is composed of three fatty acids (the "tri"), linked together on a backbone of glycerol (the "glyceride"). Some of the triglycerides in our fat tissue come from fat in our diet. The rest come from carbohydrates, from a process known as de novo lipogenesis, which is Latin for "the new creation of fat," a process that takes place both in the liver and, to a lesser extent, in the fat tissue itself. The more carbohydrates flooding the circulation after a meal, the more wil be converted to triglycerides and stored as fat for future use (perhaps 30 percent of the carbohydrates in any one meal). "This lipogenesis is regulated by the state of nutrition," explained Wertheimer in an introductory chapter to the Handbook of Physiology: "it is decreased to a minimum in carbohydrate deficiency and accelerated considerably during carbohydrate availability."*114 A second critical point is that while the fat is stored as triglycerides it enters and exits the fat cel s in the form of fatty acids-actual y, free fatty acids, to distinguish them from the fatty acids bound up in triglycerides-and it's these fatty acids that are burned as fuel in the cel s. As triglycerides, the fat is locked into the fat cel s, because triglycerides are too big to slip through the cel membranes. They have to be broken down into fatty acids-the process technical y known as lipolysis-before the fat can escape into the circulation. The triglycerides in the bloodstream must also be broken down into fatty acids before the fat can diffuse into the fat cel s. It's only reconst.i.tuted into triglycerides, a process cal ed esterification, once the fatty acids have pa.s.sed through the wal s of the blood vessels and the fat-cel membranes and are safely inside. This is true for al triglycerides, whether they originated as fat in the diet or were converted from carbohydrates in the liver.

Inside the fat cel s, triglycerides are continuously broken down into their component fatty acids and glycerol (i.e., in lipolysis), and fatty acids and glycerol are continuously rea.s.sembled into triglycerides (i.e., esterified)-a process known as the triglyceride/fatty-acid cycle. Any fatty acids that are not immediately repackaged back into triglycerides wil slip out of the fat cel and back into the circulation-"a ceaseless stream of [free fatty acids], a readily transportable source of energy, into the bloodstream," as it was described in the Handbook of Physiology by one team of NIH researchers.

Some of these free fatty acids wil be taken up by the tissues and organs and used as fuel. Perhaps as much as half of them wil not. These wil be incorporated in the liver back into triglycerides, loaded on lipoproteins,*115 and shipped back again to the fat tissue. And so fatty acids are continuously slipping from the fat tissue into the circulation, while those fatty acids that aren't immediately taken up and used for fuel are continuously being reconverted to triglycerides and transported back to the fat tissue for storage. "The storage of triglyceride fat in widely scattered adipose tissue sites is a remarkably dynamic process," explained the University of Wisconsin endocrinologist Edgar Gordon in 1969, "with the stream of fatty acid carbon atoms flowing in widely fluctuating amounts, first in one direction and then the other in a finely adjusted minute by minute response to the fuel requirements of energy metabolism of the whole organism."

This remarkably dynamic process, however, is regulated by a remarkably simple system. The flow of fatty acids out of the fat cel s and into the circulation depends on the level of blood sugar available. The burning of this blood sugar by the cel s-the oxidation of glucose-depends on the availability of fatty acids to be burned as fuel instead.

A single molecule plays the pivotal role in the system. It goes by a number of names, the simplest being glycerol phosphate. This glycerol-phosphate molecule is produced from glucose when it is used for fuel in the fat cel s and the liver, and it, too, can be burned as fuel in the cel s. But glycerol phosphate is also an essential component of the process that binds three fatty acids into a triglyceride. It provides the glycerol molecule that links the fatty acids together.116 In other words, a product of carbohydrate metabolism-i.e., burning glucose for fuel-is an essential component in the regulation of fat metabolism: storing fat in the fat tissue. In fact, the rate at which fatty acids are a.s.sembled into triglycerides, and so the rate at which fat acc.u.mulates in the fat tissue, depend primarily on the availability of glycerol phosphate. The more glucose that is transported into the fat cel s and used to generate energy, the more glycerol phosphate wil be produced. And the more glycerol phosphate produced, the more fatty acids wil be a.s.sembled into triglycerides. Thus, anything that works to transport more glucose into the fat cel s-insulin, for example, or rising blood sugar-wil lead to the conversion of more fatty acids into triglycerides, and the storage of more calories as fat.

This brings us to the mechanisms that control and regulate the availability of fat and carbohydrates for fuel and regulate our blood sugar in the process.

The first is the triglyceride/fatty-acid cycle we just discussed. This cycle is regulated by the amount of blood sugar made available to the fat tissue. If blood sugar is ebbing, the amount of glucose transported into the fat cel s wil decrease; this limits the burning of glucose for energy, which in turn reduces the amount of glycerol phosphate produced. With less glycerol phosphate present, fewer fatty acids are bound up into triglycerides, and more of them remain free to escape into the circulation. As a result, the fatty-acid concentration in the bloodstream increases. The bottom line: as the blood-sugar level decreases, fatty-acid levels rise to compensate.

If blood-sugar levels increase-say, after a meal containing carbohydrates-then more glucose is transported into the fat cel s, which increases the use of this glucose for fuel, and so increases the production of glycerol phosphate. This is turn increases the conversion of fatty acids into triglycerides, so that they're unable to escape into the bloodstream at a time when they're not needed. Thus, elevating blood sugar serves to decrease the concentration of fatty acids in the blood, and to increase the acc.u.mulated fat in the fat cel s.

The second mechanism that works to regulate the availability of fuel and to maintain blood sugar at a healthy level is cal ed the glucose/fatty-acid cycle, or the Randle cycle, after the British biochemist Sir Philip Randle. It works like this: As blood-sugar levels decrease-after a meal has been digested -more fatty acids wil be mobilized from the fat cel s, as we just discussed, raising the fatty-acid level in the bloodstream. This leads to a series of reactions in the muscle cel s that inhibit the use of glucose for fuel and subst.i.tute fatty acids instead. Fatty acids generate the necessary cel ular energy, and the blood-sugar level in the circulation stabilizes. When the availability of fatty acids in the blood diminishes, as would be the case when blood-sugar levels are rising, the cel s compensate by burning more blood sugar. So increasing blood-sugar levels decreases fatty-acids levels in the bloodstream, and decreasing fatty-acid levels in the bloodstream, in turn, increases glucose use in the cel s. Blood-sugar levels always remain within safe limits -neither too high nor too low.

These two cycles are the fundamental mechanisms that maintain and ensure a steady fuel supply to our cel s. They provide a "metabolic flexibility" that al ows us to burn carbohydrates (glucose) when they're present in the diet, and fatty acids when they're not. And it's the cel s of the adipose tissue that function as the ultimate control mechanism of this fuel supply.

Regulation by hormones and the nervous system is then layered onto these baseline mechanisms to deal with the vagaries of the external environment, providing the moment-to-moment and season-to-season fine-tuning necessary for the body to work at maximum efficiency. Hormones modify this flow of fatty acids back and forth across the membranes of the fat cel s, and they modify the expenditure of energy by the tissues and organs. Hormones, and particularly insulin-"even in trace amounts," as Ernst Wertheimer explained-"have powerful direct effect on adipose tissue."

With the invention by Rosalyn Yalow and Solomon Berson of their radioimmunoa.s.say to measure insulin levels, it quickly became clear that insulin was what Yalow and Berson cal ed "the princ.i.p.al regulator of fat metabolism." Insulin stimulates the transport of glucose into the fat cel s, thereby effectively control ing the production of glycerol phosphate, the fixing of free fatty acids as triglycerides, and al that fol ows. The one fundamental requirement to increase the flow of fatty acids out of adipose tissue-to increase lipolysis-and so decrease the amount of fat in our fat tissue, is to lower the concentration of insulin in the bloodstream. In other words, the release of fatty acids from the fat cel s and their diffusion into the circulation require "only the negative stimulus of insulin deficiency," as Yalow and Berson wrote. By the same token, the one necessary requirement to shut down the release of fat from the fat cel s and increase fat acc.u.mulation is the presence of insulin. When insulin is secreted, or the level of insulin in the circulation is abnormal y elevated, fat acc.u.mulates in the fat tissue. When insulin levels are low, fat escapes from the fat tissue, and the fat deposits shrink.

Al other hormones wil work to release fatty acids from the fat tissue, but the ability of these hormones to accomplish this job is suppressed almost entirely by the effect of insulin and blood sugar. These hormones can mobilize fat from the adipose tissue only when insulin levels are low-during starvation, or when the diet being consumed is lacking in carbohydrates. (If insulin levels are high, that implies that there is plenty of carbohydrate fuel available.) In fact, virtual y anything that increases the secretion of insulin wil also suppress the secretion of hormones that release fat from the fat tissue.

Eating carbohydrates, for example, not only elevates insulin but inhibits growth-hormone secretion; both effects lead to greater fatty-acid storage in the fat tissue.

Hormones that promote fat mobilization Hormones that promote fat acc.u.mulation Epinephrine Norepinephrine Adrenocorticotropic hormone (ACTH) Glucagon Thyroid-stimulating hormone Insulin Melanocyte-stimulating hormone Vasopressin Growth hormone In 1965, hormonal regulation of adipose tissue looked like this: at least eight hormones that worked to release fat from the adipose tissue and one, insulin, that worked to put it there.

That increasing the secretion of insulin can in fact cause obesity (i.e., excess fat acc.u.mulation) would be demonstrated conclusively in animal models of obesity, particularly in the line of research we discussed in Chapter 21 on rats and mice with lesions in the area of the brain known as the ventromedial hypothalamus, or VMH. In the 1960s, this research became another beneficiary of Yalow and Berson's new technology to measure circulating levels of insulin. As investigators now reported, insulin secretion in VMH-lesioned animals increases dramatical y within seconds of the surgery. The insulin response to eating also goes "off the scale" with the very first meal. The more insulin secreted in the days after the surgery, the greater the ensuing obesity. Obesity in these lesioned animals could be prevented by short-circuiting the exaggerated insulin response-by severing the vagus nerve, for example, that links the hypothalamus with the pancreas.*117 Similarly, the hypersecretion of insulin was reported to be the earliest detectable abnormality in genetic strains of obesity-p.r.o.ne mice and rats.

By the mid-1970s, it was clear that Stephen Ranson's insights into obesity in these animals had been confirmed. The lesion causes a defect in the part of the hypothalamus that regulates what researchers have come to cal fuel part.i.tioning-the result is the hypersecretion of insulin. The insulin forces the acc.u.mulation of fat in the fat tissue, and the animal overeats to compensate. This research refuted John Brobeck's notion, which has since become the standard wisdom in the field, that the VMH lesion causes overeating directly and the animals grow fat simply because they eat too much. These studies were neither ambiguous nor controversial. In 1976, University of Washington investigators Stephen Woods and Dan Porte described as "overwhelming"

the evidence that the increased secretion of insulin is the primary effect of VMH lesions, the driving force of obesity in these animals.

This half century of research unequivocal y supported the alternative hypothesis of obesity. It established that the relevant energy balance isn't between the calories we consume and the calories we expend, but between the calories-in the form of free fatty acids, glucose, and glycerol-pa.s.sing in and out of the fat cel s. If more and more fatty acids are fixed in the fat tissue than are released from it, obesity wil result. And while this is happening, as Edgar Gordon observed, the energy available to the cel s is reduced by the "relative unavailability of fatty acids for fuel." The consequence wil be what Stephen Ranson cal ed semi-cellular starvation and Edwin Astwood, twenty years later, cal ed internal starvation. And as this research had now made clear, the critical molecules determining the balance of storage and mobilization of fatty acids, of lipogenesis and lipolysis, are glucose and insulin-i.e., carbohydrates and the insulin response to those carbohydrates.

Just a few more details are necessary to understand why we get fat. The first is that the amount of glycerol phosphate available to the fat cel s to acc.u.mulate fat-to bind the fatty acids together into triglycerides and lock them into the adipose tissue-also depends directly on the carbohydrates in the diet. Dietary glucose is the primary source of glycerol phosphate. The more carbohydrates consumed, the more glycerol phosphate available, and so the more fat can acc.u.mulate. For this reason alone, it may be impossible to store excess body fat without at least some carbohydrates in the diet and without the ongoing metabolism of these dietary carbohydrates to provide glucose and the necessary glycerol phosphate.

"It may be stated categorical y," the University of Wisconsin endocrinologist Edgar Gordon wrote in JAMA in 1963, "that the storage of fat, and therefore the production and maintenance of obesity, cannot take place unless glucose is being metabolized. Since glucose cannot be used by most tissues without the presence of insulin, it also may be stated categorical y that obesity is impossible in the absence of adequate tissue concentrations of insulin.... Thus an abundant supply of carbohydrate food exerts a powerful influence in directing the stream of glucose metabolism into lipogenesis, whereas a relatively low carbohydrate intake tends to minimize the storage of fat."

Forty years ago, none of this was controversial-and the facts have not changed since then. Insulin works to deposit calories as fat and to inhibit the use of that fat for fuel. Dietary carbohydrates are required to al ow this fat storage to occur. Since glucose is the primary stimulator of insulin secretion, the more carbohydrates consumed-or the more refined the carbohydrates-the greater the insulin secretion, and thus the greater the acc.u.mulation of fat.

"Carbohydrate is driving insulin is driving fat," as the Harvard endocrinologist George Cahil recently summed it up.

Regarding the potential dangers of sugar in the diet, it is important to keep in mind that fructose is converted more efficiently into glycerol phosphate than is glucose. This is another reason why fructose stimulates the liver so readily to convert it to triglycerides, and why fructose is considered the most lipogenic carbohydrate. Fructose, however, does not stimulate the pancreas to secrete insulin, so glucose is stil needed for that purpose. This suggests that the combination of glucose and fructose-either the 5050 mixture of table sugar (sucrose) or the 5545 mixture of high-fructose corn syrup -stimulates fat synthesis and fixes fat in the fat tissue more than does glucose alone, which comes from the digestion of bread and starches.

It is important also to know that the fat cel s of adipose tissue are "exquisitely sensitive" to insulin, far more so than other tissues in the body. This means that even low levels of insulin, far below those considered the clinical symptom of hyperinsulinemia (chronical y high levels of insulin), wil shut down the flow of fatty acids from the fat cel s. Elevating insulin even slightly wil increase the acc.u.mulation of fat in the cel s. The longer insulin remains elevated, the longer the fat cel s wil acc.u.mulate fat, and the longer they'l go without releasing it.

Moreover, fat cel s remain sensitive to insulin long after muscle cel s become resistant to it. Once muscle cel s become resistant to the insulin in the bloodstream, as Yalow and Berson explained, the fat cel s have to remain sensitive to provide a place to store blood sugar, which would otherwise either acc.u.mulate to toxic levels or overflow into the urine and be lost to the body. As insulin levels rise, the storage of fat in the fat cel s continues, long after the muscles become resistant to taking up any more glucose. Nonetheless, the pancreas may compensate for this insulin resistance, if it can, by secreting stil more insulin. This wil further elevate the level of insulin in the circulation and serve to increase further the storage of fat in the fat cel s and the synthesis of carbohydrates from fat. It wil suppress the release of fat from the fat tissue. Under these conditions-lipid trapping, as the geneticist James Neel described it-obesity begins to look preordained. Weights wil plateau, as Dennis McGarry suggested in Science in 1992, only when the fat tissue becomes insulin-resistant as wel , or when the fat deposits enlarge to the point where the forces working to release the fat and burn it for fuel-such as the increased concentration of fatty acids inside the fat cel s-once again balance out the effect of the insulin itself.

By the mid-1960s, four facts had been established beyond reasonable doubt: (1) carbohydrates are singularly responsible for prompting insulin secretion; (2) insulin is singularly responsible for inducing fat acc.u.mulation; (3) dietary carbohydrates are required for excess fat acc.u.mulation; and (4) both Type 2 diabetics and the obese have abnormal y elevated levels of circulating insulin and a "greatly exaggerated" insulin response to carbohydrates in the diet, as was first described in 1961 by the Johns Hopkins University endocrinologists David Rabinowitz and Kenneth Zierler.

The obvious implication is that obesity and Type 2 diabetes are two sides of the same physiological coin, two consequences, occasional y concurrent, of the same underlying defects-hyperinsulinemia and insulin resistance. This was precisely what von Noorden had suggested in 1905 with his diabetogenous-obesity hypothesis, even down to the notion that obesity would natural y result when muscle tissue becomes resistant to taking up glucose from the circulation before fat tissue does. Now the science had caught up to the speculation. "We general y accept that obesity predisposes to diabetes; but does not mild diabetes predispose to obesity?" as Yalow and Berson wrote in 1965. "Since insulin is a most lipogenic agent, chronic hyperinsulinism would favor the acc.u.mulation of body fat."

When Yalow and Berson measured individual insulin and blood-sugar responses to the consumption of carbohydrates, they reported that even lean, healthy subjects exhibit "great biologic variation" in what they cal ed the "insulin-secretory responses." In other words, we secrete more or less insulin in response to the same amount of carbohydrates, or our insulin is more or less effective at lowering blood sugar or at promoting fat acc.u.mulation, or it remains elevated in the circulation for longer or shorter periods of time. And because variations of less than 1 percent in the part.i.tioning of calories either for fuel or for storage as fat could lead to the acc.u.mulation of tens of pounds of excess fat over a decade, it would take only infinitesimal variations in these "insulin-secretory responses" to mark the difference between leanness and obesity, and between health and diabetes.

Over the years, prominent diabetologists and endocrinologists-from Yalow and Berson in the 1960s through Dennis McGarry in the 1990s-have speculated on this train of causation from hyperinsulinemia to Type 2 diabetes and obesity. Anything that increases insulin, induces insulin resistance, and induces the pancreas to compensate by secreting stil more insulin, wil also lead to an excess acc.u.mulation of body fat.

One of the more insightful of these a.n.a.lyses was by the geneticist James Neel in 1982, when he "revisited" his thrifty-gene hypothesis and rejected the idea (which has since been embraced so widely by public-health authorities and health writers) that we evolved through periods of feast and famine to hold on to fat.*118 Neel suggested three scenarios of these insulin-secretory responses that could const.i.tute a predisposition to obesity and/or Type 2 diabetes-each of which, he wrote, would be a physiological "response to the excessive glucose pulses that result from the refined carbohydrates/over-alimentation of many civilized diets." Genetic variations in these responses would determine how long it would be before obesity or diabetes appears, and which of the two appears first. The one important caveat about these three scenarios, Neel added, is that they "should not be thought of as mutual y exclusive or as exhausting the possible biochemical and physiological sequences" that might induce obesity and/or diabetes once populations take to eating modern Western diets.

The first of these scenarios was what Neel cal ed a "quick insulin trigger." By this Neel meant that the insulin-secreting cel s in the pancreas are hypersensitive to the appearance of glucose in the bloodstream. They secrete too much insulin in response to the rise in blood sugar during a meal; that encourages fat deposition and induces a compensatory insulin resistance in the muscles. The result wil be a vicious circle: excessive insulin secretion stimulates insulin resistance, which stimulates yet more insulin secretion. In this scenario, we gain weight until the fat cel s eventual y become insulin-resistant. When the "overworked" pancreatic cel s "lose their capacity to respond" to this insulin resistance, Type 2 diabetes appears.

In Neel's second scenario, there is a tendency to become slightly more insulin-resistant than would normal y be the case when confronted with a given amount of insulin in the circulation. So even an appropriate insulin response to the waves of blood sugar that appear during meals wil result in insulin resistance, and that in turn requires a ratcheting up of the insulin response. Once again, the result is the vicious cycle.

Neel's third scenario is slightly more complicated, but there's evidence to suggest that this one comes closest to reality. Here an appropriate amount of insulin is secreted in response to the "excessive glucose pulses" of a modern meal, and the response of the muscle cel s to the insulin is also appropriate.

The defect is in the relative sensitivity of muscle and fat cel s to the insulin. The muscle cel s become insulin-resistant in response to the "repeated high levels of insulinemia that result from excessive ingestion of highly refined carbohydrates and/or over-alimentation," but the fat cel s fail to compensate.

They remain stubbornly sensitive to insulin. So, as Neel explained, the fat tissue acc.u.mulates more and more fat, but "mobilization of stored fat would be inhibited." Now the acc.u.mulation of fat in the adipose tissue drives the vicious cycle.

This scenario is the most difficult to sort out clinical y, because when these investigators measure insulin resistance in humans they invariably do so on a whole-body level, which is al the existing technology al ows. Any disparities between the responsiveness of fat and muscle tissue to insulin cannot be measured. This is critical, because for the past thirty-five years the American Diabetes a.s.sociation has recommended that diabetics eat a diet relatively rich in carbohydrates based on the notion that this makes them more sensitive to insulin, at least temporarily, so the diet appears to ameliorate the diabetes. This effect was initial y reported in 1971, by the University of Washington endocrinologists Edwin Bierman and John Brunzel ,*119 who then waged a lengthy and successful campaign to persuade the American Diabetes a.s.sociation to recommend that diabetics eat more carbohydrates rather than less. If Neel's third scenario is correct, however, a likely explanation for why carbohydrate-rich diets appear to facilitate blood-sugar control after meals is that they increase the insulin sensitivity of the fat cel s specifical y, while the muscle tissue remains insulin-resistant.

One of the few attempts, if not the only one, to measure the insulin sensitivity of fat cel s and muscle cel s separately in human subjects was made by the University of Vermont investigator Ethan Sims, in his experimental obesity studies of the late 1960s. Sims and his col eagues surgical y removed fat samples from their subjects before, during, and after the periods of forced overeating and weight gain. They reported that high-carbohydrate diets had the unique ability to increase the insulin sensitivity of fat cel s, and particularly so in fat cel s that were already large and overstuffed. They had no similar effect, however, on the insulin resistance of the muscle tissue.

If this observation is correct, it means carbohydrates are uniquely capable of prolonging this lipid-trapping condition by keeping the fat cel s sensitive to insulin when they might otherwise become insulin-resistant. This might lower blood-sugar levels temporarily and delay or improve the appearance of diabetes-or "mask" the diabetes, as von Noorden put it-but would do so at the cost of increasing fat acc.u.mulation and obesity. Sims's observation suggests that Neel's third scenario for the genesis of obesity and diabetes was astute, and it suggests that a car