Chapter 2 — The Macronutrient Triad

"I had come to the conclusion, that the principal alimentary matters might be reduced to the three great classes, namely the saccharine, the oily and the albuminous." — William Prout, "On the Ultimate Composition of Simple Alimentary Substances," 1827

Long before any chemist had isolated a vitamin, weighed a calorie, or so much as guessed at the existence of a gene, an English physician named William Prout sat down with a stack of foodstuffs and a set of reagents and arrived, by nothing more dramatic than careful analysis, at a conclusion that has never seriously been overturned. Whatever else a meal might contain — and this compendium will spend its next chapter insisting there is a great deal else — its bulk resolves, on the chemist's bench, into three great classes: the saccharine, which we now call carbohydrate; the oily, which we now call fat; and the albuminous, which we now call protein. Prout's 1827 paper is not a document many readers will have encountered, and it will not detain this chapter long, but it deserves its place at the head of this chapter for a simple reason: everything that follows is, in essence, a hundred and ninety-nine years of filling in Prout's outline with detail he could not have imagined. The three great classes remain the three great classes. What has changed is our understanding of why the body wants a particular kind of each, in a particular quantity, and what happens when it gets the wrong one — a story that will take this chapter from a dog starved of nitrogen in 1816 Paris to a courtroom-worthy dispute over butter that is still, in places, unresolved.

Protein: the plastic foods

Justus von Liebig, writing a generation after Prout, gave the three classes a functional vocabulary that has aged rather well: he called fats and carbohydrates the "respiratory" foods, burned for warmth and motion, and reserved for protein the more architectural term "plastic" — from the Greek for moulding or shaping — because protein, and protein virtually alone among the great classes, supplies the nitrogen from which the body builds and rebuilds its own tissue. The demonstration that this distinction was not merely definitional but a matter of life and death had, in fact, already been performed a generation earlier still, in 1816, by the French physiologist François Magendie, who fed dogs diets of pure sugar, pure olive oil, or pure butter — none of it short on calories, all of it entirely without nitrogen — and watched the animals sicken and die within a few weeks regardless of how much fuel they were given.[^2] A dog, it turned out, cannot be kept alive on flame alone; it must also be given the bricks.

Those bricks are amino acids, of which the body's proteins are built from a working alphabet of twenty. Roll up your sleeve and consider what is happening inside it at this moment: your muscle, no less than your skin, your antibodies, your digestive enzymes and the haemoglobin ferrying oxygen through your blood, is protein, and every one of those proteins is nothing more than a particular sequence of amino acids folded into a particular shape, the sequence dictated letter by letter by your genes. Of the twenty, the adult human body can manufacture eleven for itself, given a supply of nitrogen and the raw carbon skeletons to work with; these are the non-essential amino acids, a slightly ungrateful name for compounds the body needs just as urgently as any other, but need not be handed ready-made. The remaining nine — histidine, isoleucine, leucine, lysine, methionine, phenylalanine, threonine, tryptophan and valine[^1] — the body cannot make at all, at any age, under any circumstance, and must therefore receive intact from the diet or do without. These are the essential amino acids proper, and a handful of others occupy an interesting middle ground, termed conditionally essential: arginine, cysteine, glutamine, glycine, proline and tyrosine, which the body can usually synthesise for itself but may fail to produce in sufficient quantity during infancy, illness, or serious physiological stress, when the demand for rapid tissue-building outpaces the body's own manufacturing capacity. A newborn infant, still assembling its own biochemical machinery, is conditionally essential's most vivid illustration; a burns patient, rebuilding skin at a furious rate, is another.

The quality of a protein

It is at this point that a great deal of popular nutrition writing goes astray, treating "protein" as though it were a single undifferentiated currency, so many grams to be totted up on a label regardless of source. The truth, as any naturalist comparing a steak to a plate of white rice quickly discovers, is considerably more interesting: proteins differ enormously in quality, and quality here has a precise, measurable meaning. A protein of high quality supplies all nine essential amino acids in roughly the proportions the human body requires, and does so in a form the gut can actually absorb; a protein of lower quality is either missing one or more essential amino acids in adequate quantity — its so-called limiting amino acid — or wrapped in a matrix of fibre and antinutrients that the gut cannot fully break down.

Nutrition science has, over the better part of a century, built increasingly sophisticated instruments for measuring this quality. The older standard, the Protein Digestibility-Corrected Amino Acid Score, or PDCAAS, compares a food's amino acid profile against a reference pattern of human requirements and corrects the result for how well the protein survives digestion, as measured — a touch indelicately — at the far end of the gut. Its successor, the Digestible Indispensable Amino Acid Score, or DIAAS, proposed by the Food and Agriculture Organization in 2013 and steadily gaining ground since, refines the method by measuring digestibility amino acid by amino acid at the end of the small intestine, before the colon's resident bacteria have had a chance to interfere, and by declining to cap the final score at a tidy 100 the way PDCAAS does. By either measure, a small cluster of animal proteins — whey, milk, eggs and, close behind them, chicken and fish — comes out at or near the very top of the scale, digestible at ninety-five per cent or better and essentially complete in their amino acid arithmetic. Most plant proteins, taken singly, fall short of one essential amino acid or another: cereal grains run low in lysine, pulses and legumes run low in the sulphur-containing amino acids methionine and cysteine, and the gut's access to what protein is present is generally somewhat reduced by the surrounding fibre.

None of which is licence to declare plant protein inferior in any practical sense, for the limiting amino acid in a grain is very often abundant in a legume, and vice versa, a fact that traditional cuisines the world over discovered by trial and appetite long before any biochemist could explain why the combination worked. Rice and beans, corn and beans, hummus and pitta, dal and rice, the peanut-butter sandwich beloved of schoolchildren on two continents: each is an exercise in what nutritionists call protein complementarity, the pairing of two incomplete proteins whose limiting amino acids happen to plug one another's gaps, so that the meal as a whole delivers a complete amino acid profile even though neither ingredient could manage it alone. It was once thought this complementing had to occur within a single meal, on a single plate, to count; the more relaxed and better-supported view today is that the body maintains a modest circulating pool of amino acids from recent meals, so that complementary proteins eaten across the same day, rather than the same sitting, generally serve the purpose just as well. The peasant's plate, in other words, had already solved by habit a problem the laboratory would take a further century to describe in careful chemical prose.

Field Note — The Egg

Few objects in the larder have done more quiet service to the science of nutrition than the humble hen's egg, and none has been more thoroughly deserved of the honour. For much of the twentieth century, before DIAAS supplanted the older biological-value and net-protein-utilisation methods, whole egg protein served as the literal reference standard against which every other protein on earth was scored: a score of 100, the benchmark from which grains and pulses and even other meats were marked down. The reason is not mysterious. An egg is, quite plainly, a self-contained parcel of everything a chick embryo will need to build an entire vertebrate body from a single fertilised cell, and its principal protein, ovalbumin, arrives folded around a near-perfect balance of all nine essential amino acids, wrapped in a form the human gut digests at close to ninety-nine per cent efficiency — a figure few other whole foods can approach.

A single large egg weighs some fifty grams and yields roughly six grams of this superlative protein, distributed with a certain elegant division of labour between white and yolk: the white supplies the bulk of the protein in near-pure form, while the yolk — long unfairly maligned for its cholesterol, a matter this volume will take up in a later chapter with the nuance it deserves — carries choline, vitamin A, vitamin D, and the fat without which several of those vitamins could not be absorbed at all. It is worth noting, in the spirit of the Matrix Mandate this compendium keeps returning to, that no one has ever isolated ovalbumin into a powder and called it dinner; it is the whole egg, yolk and white and the trace minerals bound within both, that a growing chick — and a growing child — actually requires. The specimen collector, cracking one open over a bowl on any ordinary morning, holds in the shell a small and unglamorous masterpiece of protein engineering, evolved not for the pleasure of the human palate at all, but for the entirely more urgent business of building a bird from scratch.

Leucine, and the muscle without the hysteria

Among the nine essential amino acids, one has, over the past three decades, attracted a disproportionate share of attention from physiologists and gymnasium alike: leucine, a branched-chain amino acid that turns out to do rather more than simply supply raw material for new tissue. Leucine also acts as a direct chemical signal, binding to a cellular switchboard called mTOR — mechanistic target of rapamycin, a name only a biochemist could love — and flipping on the machinery of muscle protein synthesis more or less on its own, quite apart from its role as a mere building block. Feed a person a meal containing enough leucine, and the muscle-building machinery in that person's cells switches from idle to active within the hour; feed a meal that falls short of the threshold, however much total protein it contains, and the switch stays largely off.

This finding, real and repeatedly confirmed though it is, has attracted a certain amount of gymnasium mythology that this compendium is obliged to gently correct. The threshold in question is not enormous: something on the order of two and a half to three grams of leucine appears sufficient to switch on maximal muscle protein synthesis in a young, healthy adult, an amount comfortably supplied by twenty-five to thirty grams of good-quality protein — roughly a chicken breast, four eggs, or a modest tin of tuna — and no meaningful further advantage has been demonstrated from piling on more leucine, or more protein, beyond that point in a single sitting. The muscle does not build faster in proportion to the excess; the surplus amino acids are simply burned for fuel or excreted, an outcome no more mysterious, and no more alarming, than what happens to any nutrient consumed beyond the body's immediate capacity to use it.

The one population in whom the threshold genuinely shifts is the elderly, and here the science, rather than justifying panic, offers instead a piece of thoroughly actionable good sense. Advancing age brings with it a phenomenon physiologists call anabolic resistance: the same dose of protein that would fully switch on muscle-building machinery in a person of thirty produces a distinctly blunted response in a person of seventy, the cellular switchboard having grown, in effect, a little hard of hearing. The practical remedy is not a supplement bottle but simple redistribution: current geriatric nutrition guidance from bodies such as the PROT-AGE study group and the European Society for Clinical Nutrition and Metabolism recommends that older adults aim for something nearer thirty to forty grams of high-quality protein at each of three or four meals, with particular attention to reaching a leucine content of three grams or more per meal, rather than concentrating the day's protein, as so many older diets inadvertently do, into a modest breakfast and a still more modest lunch before a substantial dinner arrives too late to do the muscle much good. It is a finding with genuine consequence for the prevention of sarcopenia, the gradual, largely silent loss of muscle mass that quietly erodes strength and independence in old age — and it is a finding that requires no hysteria, no exotic supplement, and no departure from the plate at all: simply a modest redistribution of the eggs, fish and pulses most kitchens already keep in the larder.

How much is enough

The official recommended dietary allowance for protein in most Western guidelines — roughly 0.8 grams per kilogram of body weight per day, a little over half a gram per pound — is frequently misunderstood, and the misunderstanding is worth dispelling plainly. That figure is not, and was never intended to be, an optimum; it is a floor, calculated by nitrogen-balance studies to be the minimum quantity that keeps a sedentary adult from slowly losing more nitrogen than is replaced, with a safety margin built in to cover the bulk of the healthy population. It says nothing about the rather different questions of athletic performance, healthy ageing, or recovery from illness or injury, and current sports-nutrition consensus, drawing on decades of nitrogen-balance and muscle-protein-synthesis research, places the useful range for those actively building or preserving muscle mass considerably higher: something in the order of 1.6 to 2.2 grams per kilogram for a person engaged in serious resistance training, and, as the previous section noted, 1.0 to 1.2 grams per kilogram — rising to 1.2 to 1.5 for those managing chronic illness or early sarcopenia — for a healthy older adult for whom the plain RDA figure has been shown, repeatedly, to run short.

None of this licenses the opposite excess, the notion promoted by a certain style of modern diet culture that more protein is invariably better, without limit. The body has no meaningful capacity to store surplus amino acids as such; protein eaten beyond the day's building and repair needs is deaminated in the liver, its nitrogen excreted as urea, and its carbon skeleton burned for energy or laid down as fat exactly as a surplus of sugar or oil would be. For the healthy adult with normally functioning kidneys, no controlled trial has found convincing evidence of harm from protein intakes considerably above the RDA, but nor has any found a benefit that continues to accrue without limit, and the theoretical concern for those with compromised kidney function — in whom the kidney's task of clearing nitrogenous waste is already strained — counsels the ordinary rule that governs most matters of appetite: enough is a virtue, and more is merely more.

<figure> <svg xmlns="http://www.w3.org/2000/svg" width="100%" viewBox="0 0 640 420" role="img"> <title>The Macronutrient Triad — protein, fat and carbohydrate compared</title> <defs> <marker id="arw3" viewBox="0 0 10 10" refX="8" refY="5" markerWidth="5" markerHeight="5" orient="auto-start-reverse"> <path d="M0,0 L10,5 L0,10 Z" fill="#43331f"/> </marker> </defs> <rect x="8" y="8" width="624" height="404" fill="none" stroke="#43331f" stroke-width="1.2"/> <rect x="12" y="12" width="616" height="396" fill="none" stroke="#c8b79a" stroke-width="0.6"/> <g fill="none" stroke="#43331f" stroke-width="0.8"> <path d="M12,12 L12,20 M12,12 L20,12"/> <path d="M628,12 L628,20 M628,12 L620,12"/> <path d="M12,408 L12,400 M12,408 L20,408"/> <path d="M628,408 L628,400 M628,408 L620,408"/> </g> <text x="320" y="32" text-anchor="middle" font-family="Georgia, 'Times New Roman', serif" font-size="15" letter-spacing="2" fill="#43331f">THE MACRONUTRIENT TRIAD</text> <line x1="40" y1="46" x2="600" y2="46" stroke="#c8b79a" stroke-width="0.6"/>

<text x="130" y="78" text-anchor="middle" font-family="Georgia, serif" letter-spacing="1.5" font-size="13" fill="#43331f">PROTEIN</text> <text x="130" y="93" text-anchor="middle" font-family="Georgia, serif" font-style="italic" font-size="10" fill="#6f5a40">a polymer of amino acids</text> <g stroke="#43331f" stroke-width="0.8" fill="none"> <circle cx="130" cy="160" r="5"/> <line x1="126" y1="156" x2="100" y2="130" marker-end="url(#arw3)"/> <text x="80" y="122" font-family="Georgia, serif" font-style="italic" font-size="9" fill="#43331f">H₂N</text> <line x1="134" y1="156" x2="160" y2="130"/> <path d="M160,130 L172,120 M172,120 L182,120 M172,120 L172,108" stroke="#43331f" stroke-width="0.7"/> <text x="185" y="122" font-family="Georgia, serif" font-size="9" fill="#43331f">COOH</text> <line x1="130" y1="165" x2="130" y2="195"/> <rect x="118" y="195" width="24" height="16" fill="none" stroke="#9c5a2c" stroke-width="0.7"/> <text x="130" y="207" text-anchor="middle" font-family="Georgia, serif" font-size="9" fill="#9c5a2c">R</text> </g> <g stroke="#6f5a40" stroke-width="0.8" fill="none"> <circle cx="80" cy="255" r="6"/><circle cx="105" cy="255" r="6"/><circle cx="130" cy="255" r="6"/><circle cx="155" cy="255" r="6"/><circle cx="180" cy="255" r="6"/> <line x1="86" y1="255" x2="99" y2="255"/><line x1="111" y1="255" x2="124" y2="255"/><line x1="136" y1="255" x2="149" y2="255"/><line x1="161" y1="255" x2="174" y2="255"/> </g> <text x="130" y="280" text-anchor="middle" font-family="Georgia, serif" font-style="italic" font-size="9" fill="#6f5a40">peptide chain</text> <line x1="130" y1="295" x2="130" y2="325" stroke="#43331f" stroke-width="0.9" marker-end="url(#arw3)"/> <text x="130" y="345" text-anchor="middle" font-family="Georgia, serif" font-size="12" fill="#43331f">tissue & enzymes</text>

<text x="320" y="78" text-anchor="middle" font-family="Georgia, serif" letter-spacing="1.5" font-size="13" fill="#43331f">FAT</text> <text x="320" y="93" text-anchor="middle" font-family="Georgia, serif" font-style="italic" font-size="10" fill="#6f5a40">glycerol esterified with fatty acids</text> <g stroke="#43331f" stroke-width="0.8" fill="none"> <line x1="280" y1="150" x2="280" y2="210"/> <circle cx="280" cy="150" r="4"/><circle cx="280" cy="180" r="4"/><circle cx="280" cy="210" r="4"/> <path d="M284,150 L305,150 L312,143 L332,143 L339,150 L359,150" /> <path d="M284,180 L305,180 L312,173 L332,173 L339,180 L359,180 L366,173 L386,173"/> <path d="M284,210 L305,210 L312,203 L332,203 L339,210 L359,210"/> </g> <text x="320" y="235" text-anchor="middle" font-family="Georgia, serif" font-style="italic" font-size="9" fill="#6f5a40">triglyceride</text> <line x1="320" y1="295" x2="320" y2="325" stroke="#43331f" stroke-width="0.9" marker-end="url(#arw3)"/> <text x="320" y="345" text-anchor="middle" font-family="Georgia, serif" font-size="12" fill="#43331f">membrane & store</text>

<text x="510" y="78" text-anchor="middle" font-family="Georgia, serif" letter-spacing="1.5" font-size="13" fill="#43331f">CARBOHYDRATE</text> <text x="510" y="93" text-anchor="middle" font-family="Georgia, serif" font-style="italic" font-size="10" fill="#6f5a40">a polymer of glucose</text> <g stroke="#43331f" stroke-width="0.8" fill="none"> <polygon points="510,145 530,157 530,181 510,193 490,181 490,157"/> <line x1="510" y1="145" x2="510" y2="133"/> <line x1="530" y1="157" x2="545" y2="150"/> <line x1="490" y1="181" x2="475" y2="188"/> </g> <text x="510" y="128" text-anchor="middle" font-family="Georgia, serif" font-size="8" fill="#9c5a2c">OH</text> <text x="550" y="148" text-anchor="middle" font-family="Georgia, serif" font-size="8" fill="#9c5a2c">OH</text> <text x="510" y="210" text-anchor="middle" font-family="Georgia, serif" font-style="italic" font-size="9" fill="#6f5a40">glucose, C₆H₁₂O₆</text> <g stroke="#6f5a40" stroke-width="0.7" fill="none"> <polygon points="470,255 480,261 480,273 470,279 460,273 460,261"/> <polygon points="500,255 510,261 510,273 500,279 490,273 490,261"/> <polygon points="530,255 540,261 540,273 530,279 520,273 520,261"/> <line x1="480" y1="267" x2="490" y2="267"/> <line x1="510" y1="267" x2="520" y2="267"/> </g> <text x="500" y="292" text-anchor="middle" font-family="Georgia, serif" font-style="italic" font-size="9" fill="#6f5a40">glycogen chain</text> <line x1="510" y1="295" x2="510" y2="325" stroke="#43331f" stroke-width="0.9" marker-end="url(#arw3)"/> <text x="510" y="345" text-anchor="middle" font-family="Georgia, serif" font-size="12" fill="#43331f">fuel & glycogen</text> </svg> <figcaption><em>Plate 3. The macronutrient triad and its fates: protein into tissue and enzyme, fat into membrane and store, carbohydrate into fuel and glycogen — three classes of molecule, one shared destination in the body's economy</em></figcaption> </figure>

Fat: the maligned macronutrient

If protein has lately enjoyed something of a cultural rehabilitation, fat has spent the better part of a century as nutrition's designated villain, and no honest chapter on the macronutrient triad can proceed without walking directly into the argument that made it so. Chemically, a dietary fat is a triglyceride: a single molecule of glycerol with three fatty acid chains hanging from it like the legs of a tripod, and the fatty acids themselves differ from one another chiefly in one small but consequential detail — the number of double bonds along the carbon chain. A saturated fatty acid, its every carbon bond filled or "saturated" with hydrogen, has no double bonds at all, packs together tightly with its neighbours, and is typically solid at room temperature: think of butter, lard, the white marbling in a well-aged steak, or the fat that renders from a duck. A monounsaturated fatty acid has exactly one double bond, kinking the chain just enough to keep it liquid at room temperature but solid in the refrigerator — the signature of olive oil, avocado and most tree nuts. A polyunsaturated fatty acid has two or more double bonds, kinks further still, and remains liquid even when chilled: the sunflower, corn, soybean and fish oils of the kitchen shelf.

Among the polyunsaturated fats sit two families the body cannot manufacture at all and must obtain from food, precisely as it must obtain the essential amino acids: the omega-3 and omega-6 fatty acids, named for the position of their first double bond counting in from the tail end of the chain. Linoleic acid, an omega-6, and alpha-linolenic acid, an omega-3, are the two genuinely essential parent fats, abundant respectively in most vegetable oils and in flaxseed, walnuts, chia and rapeseed oil; the body can, in principle, elongate both into their more biologically active descendants — arachidonic acid from linoleic acid, and the marine fats eicosapentaenoic acid and docosahexaenoic acid, EPA and DHA, from alpha-linolenic acid — but the conversion machinery for the omega-3 pathway in particular is notoriously inefficient in humans, typically converting less than ten per cent of dietary ALA to EPA and a good deal less than that to DHA. This is precisely why oily fish, which contain EPA and DHA ready-made, occupy such a privileged place in dietary guidance: a diet relying on flaxseed alone to supply the brain and retina with DHA is, for most people, asking the conversion enzymes to perform a good deal more heavy lifting than they are built for.

Considerable attention has also been paid to the ratio between omega-6 and omega-3 intake, and here this compendium must tread carefully between two temptations: dismissing the ratio as meaningless, and treating it as more settled than it is. It is true, and worth recording, that the diet on which the human lineage evolved probably supplied omega-6 and omega-3 fats in something close to a one-to-one ratio, while the typical modern Western diet, awash in seed oils, runs closer to fifteen or sixteen to one; it is also true that arachidonic acid, the omega-6 descendant, gives rise to somewhat more inflammatory signalling molecules than the omega-3 descendants do, lending the ratio a plausible mechanism. But the clinical trial evidence for chasing a particular ratio, as opposed to simply ensuring an adequate absolute intake of EPA and DHA — from oily fish two or three times weekly, or an algal-oil supplement for those who eat none — remains considerably thinner than the ratio's popularity in health writing would suggest, and this compendium prefers, on the current state of the evidence, to recommend the fish rather than the arithmetic.

The saturated fat wars

No corner of nutrition science has generated more heat, in every sense, than the question of saturated fat and the heart, and honesty obliges this compendium to walk the reader through the argument's full and unfinished history rather than deliver a verdict dressed up as settled fact. The modern story begins in earnest with Ancel Keys, an American physiologist who, having noticed strikingly different rates of coronary heart disease across post-war Europe, launched in 1958 what became the Seven Countries Study: a long-running comparison of diet and heart-disease mortality among cohorts of middle-aged men in the United States, Finland, the Netherlands, Italy, the former Yugoslavia, Greece and Japan. Keys's analysis found a strong association between the proportion of calories a population derived from saturated fat and its rate of death from coronary disease, and that finding, amplified through the following decades by the American Heart Association and successive editions of national dietary guidelines, became the empirical foundation of the low-fat advice that shaped Western eating for the remainder of the twentieth century.

The criticisms arrived almost as quickly as the fame, and some of them have held up rather better than others. The oft-repeated charge that Keys examined data from twenty-two countries and quietly discarded the fifteen that inconvenienced his hypothesis appears, on the documentary record, to be a myth in its most exaggerated form — the study was designed from the outset around seven cohorts, not selected retrospectively from a much larger pool — but more measured criticisms carry real weight. The Seven Countries Study was observational, not a controlled trial, and correlation across a mere handful of national diets cannot, on its own, establish causation with any rigour; the selection of countries, while not the wholesale cherry-picking of legend, was not random either, and France, whose traditionally fat-rich diet sat awkwardly beside its comparatively low heart-disease mortality, was excluded from the study's design — an omission that would go on to seed, decades later, the entire "French paradox" literature. Most tellingly of all in retrospect, Keys's original analysis lumped naturally occurring saturated fat together with the industrially produced trans fats that, unbeknownst to nutrition science at the time, were flooding the same post-war food supply in the form of hydrogenated margarine and shortening — a conflation this chapter will return to shortly, because unpicking it turns out to matter enormously.

Subsequent decades of research have neither vindicated the simplest version of the low-fat hypothesis nor demolished it, and the honest position — the position this compendium adopts, in keeping with its obligation to respect the replication crisis and to prefer consensus statements over any single dramatic paper — sits somewhere considerably more interesting than either extreme. A rigorous re-analysis published in 2016 of previously unpublished data from the Minnesota Coronary Experiment, a large randomised trial conducted in the very years Keys's influence was at its height, found that swapping saturated animal fat for corn oil rich in linoleic acid did lower participants' cholesterol as predicted, yet failed to reduce, and in the oldest participants appeared actually to increase, mortality — a result that sat quietly unpublished for decades and stands today as a small, sobering monument to how selectively even careful science can be reported. Set against this, the American Heart Association's own 2017 presidential advisory, reviewing the accumulated trial evidence with considerably more data at its disposal than Keys ever had, concluded that randomised trials replacing saturated fat with polyunsaturated fat do lower cardiovascular events, by a margin — roughly thirty per cent — comparable to that achieved by statin drugs. The advisory's genuinely crucial caveat, one too rarely repeated in popular summary, is that the benefit depends entirely on what saturated fat is replaced with: trials in which saturated fat was simply cut and replaced with refined starch and sugar, producing the "low-fat" packaged foods that lined supermarket shelves through the nineteen-eighties and nineties, showed no such benefit at all, and by some readings made matters modestly worse. The lesson, in other words, is not that saturated fat is either an innocent bystander or nutrition's chief villain; it is that the question "what did you eat instead?" turns out to matter as much as the question "how much saturated fat did you eat," and a great deal of popular writing on both sides of this argument has been guilty of forgetting to ask it.

Where does that leave the reader standing at the butcher's counter? Broadly where most contemporary consensus bodies — Harvard's Nutrition Source among them — currently stand: saturated fat need not be treated as a poison to be purged from every meal, particularly when it arrives, as it does in cheese, yoghurt and unprocessed meat, bound up within a food matrix that includes protein, calcium and other nutrients whose own cardiovascular effects appear more benign than saturated fat consumed in isolation; but a dietary pattern that leans distinctly toward the unsaturated fats of olive oil, nuts, seeds and oily fish, in preference to butter, cream and processed meat, continues to carry the stronger evidence for long-term heart health, provided the swap is toward those foods and not toward refined carbohydrate dressed up as a health choice. It is a more complicated answer than either side of the popular argument tends to offer, and this compendium makes no apology for that: the honest state of a genuinely contested science is very rarely a slogan.

The one true villain: trans fat

If the saturated fat debate deserves nuance, the case against industrially produced trans fat deserves almost none, and the contrast between the two is instructive in its own right. Trans fats arise chiefly through partial hydrogenation, a process patented by the German chemist Wilhelm Normann in 1902 in which hydrogen gas is bubbled through liquid vegetable oil in the presence of a metal catalyst, straightening some of the oil's natural kinked double bonds into a rarer, straighter "trans" configuration that behaves, chemically, far more like a saturated fat: solid at room temperature, resistant to rancidity, and cheap to produce at industrial scale. The process gave the twentieth century margarine, and shortening products such as Crisco, and in doing so gave the century's bakers a fat that behaved beautifully in a pastry crust and sat on a supermarket shelf for months without turning — commercial virtues that, for the better part of a hundred years, were allowed to outweigh a health cost nobody had yet troubled to measure.

That cost, once measured, proved unambiguous in a way saturated fat's never has: trans fat raises low-density lipoprotein cholesterol, the particle strongly associated with arterial disease, while simultaneously lowering high-density lipoprotein cholesterol, the particle that helps clear cholesterol from the arteries — a double insult no other dietary fat reliably produces, and one large meta-analyses have linked to a substantially elevated risk of coronary heart disease even at modest levels of intake. The consensus against industrially produced trans fat is, as a result, close to universal among health authorities in a field not otherwise known for easy agreement, and the regulatory response has followed accordingly: Denmark became, in 2003, the first country to impose a strict national limit on industrial trans fat in food, and the World Health Organization's REPLACE initiative, launched in 2018, set a goal of virtually eliminating industrially produced trans fat from the global food supply. That original target date of 2023 was not fully met — by the WHO's own accounting, fifty-three countries had best-practice trans-fat policies in place by that year, protecting some 3.7 billion people, or forty-six per cent of the world's population, with a revised elimination goal pushed to 2025 — but the trajectory, unlike almost every other argument in this chapter, points in one direction only, and no reputable dietary authority has emerged in the intervening decades to argue the other side.

Carbohydrate: sugars, starches, and the glycaemic tide

Carbohydrate, Prout's saccharine class, is the macronutrient most burdened by an old and unhelpful simplicity: the notion, still repeated at a great many dinner tables, that carbohydrates are simply "simple" or "complex," the former to be avoided and the latter embraced, as though molecular architecture alone told the whole story. The distinction is not meaningless — a simple sugar such as glucose, fructose or the milk sugar lactose is, structurally, one or two linked sugar units, absorbed into the bloodstream rapidly and with little digestive effort, while a complex carbohydrate such as starch is a long chain of hundreds or thousands of glucose units, requiring the gut's enzymes to work considerably harder to break it apart before absorption can begin — but structure alone predicts a food's actual effect on the body only imperfectly, and it is that effect, rather than the label "simple" or "complex," that the more useful modern measure attempts to capture.

That measure is the glycaemic index, introduced in 1981 by the Canadian physician David Jenkins and his colleagues at the University of Toronto, who set out to test a suspicion that had been quietly nagging at diabetes researchers: that equal quantities of carbohydrate from different foods produced strikingly different rises in blood glucose, in ways the simple/complex distinction could not explain. The glycaemic index measures precisely this, comparing the rise in blood glucose over two hours after eating fifty grams of carbohydrate from a test food against the rise produced by the same quantity of a reference food — pure glucose, or in some versions white bread — and expressing the result as a percentage. White bread and boiled potato, in defiance of their "complex" pedigree, turn out to carry a glycaemic index nearly as high as table sugar; a lentil or a chickpea, though built from the same basic starch chemistry, produces a markedly gentler rise, its structure and its accompanying fibre and protein slowing the whole process to a crawl.

The glycaemic index has one well-known blind spot, corrected by a refinement introduced in 1997 by the Harvard epidemiologist Walter Willett and his colleague Salvador Salmerón: it says nothing about portion size, treating a modest slice of watermelon, whose glycaemic index is surprisingly high but whose actual sugar content per serving is modest, as equivalent in effect to a much larger, more carbohydrate-dense meal. The glycaemic load corrects this by multiplying a food's glycaemic index by the actual grams of carbohydrate in a typical serving, and dividing by a hundred, yielding a number that more faithfully predicts what a given plateful will actually do to the blood a naturalist happens to be standing next to at the time. Both measures matter, and both point toward the same underlying physiological event: a rapid, substantial rise in blood glucose calls forth a correspondingly rapid, substantial release of insulin from the pancreas, the hormone whose task is to usher that glucose out of the bloodstream and into the body's cells — muscle chiefly, but also liver and fat — before it can do the kind of quiet vascular damage chronically elevated blood sugar is now well understood to cause.

<figure> <svg xmlns="http://www.w3.org/2000/svg" width="100%" viewBox="0 0 640 400" role="img"> <title>The Glycaemic Response — blood glucose after refined versus whole food</title> <defs> <pattern id="hatch4" width="5" height="5" patternUnits="userSpaceOnUse" patternTransform="rotate(45)"> <line x1="0" y1="0" x2="0" y2="5" stroke="#c8b79a" stroke-width="0.4"/> </pattern> </defs> <rect x="8" y="8" width="624" height="384" fill="none" stroke="#43331f" stroke-width="1.2"/> <rect x="12" y="12" width="616" height="376" fill="none" stroke="#c8b79a" stroke-width="0.6"/> <g fill="none" stroke="#43331f" stroke-width="0.8"> <path d="M12,12 L12,20 M12,12 L20,12"/> <path d="M628,12 L628,20 M628,12 L620,12"/> <path d="M12,388 L12,380 M12,388 L20,388"/> <path d="M628,388 L628,380 M628,388 L620,388"/> </g> <text x="320" y="32" text-anchor="middle" font-family="Georgia, 'Times New Roman', serif" font-size="15" letter-spacing="2" fill="#43331f">THE GLYCAEMIC RESPONSE</text> <line x1="40" y1="46" x2="600" y2="46" stroke="#c8b79a" stroke-width="0.6"/>

<line x1="80" y1="70" x2="80" y2="320" stroke="#43331f" stroke-width="0.9"/> <line x1="80" y1="320" x2="580" y2="320" stroke="#43331f" stroke-width="0.9"/> <g stroke="#c8b79a" stroke-width="0.4"> <line x1="80" y1="260" x2="580" y2="260"/><line x1="80" y1="200" x2="580" y2="200"/><line x1="80" y1="140" x2="580" y2="140"/><line x1="80" y1="80" x2="580" y2="80"/> </g> <g font-family="Georgia, serif" font-size="9" fill="#6f5a40" text-anchor="end"> <text x="74" y="323">60</text><text x="74" y="283">80</text><text x="74" y="243">100</text><text x="74" y="203">120</text><text x="74" y="163">140</text><text x="74" y="123">160</text><text x="74" y="83">180</text> </g> <g font-family="Georgia, serif" font-size="9" fill="#6f5a40" text-anchor="middle"> <text x="80" y="335">0</text><text x="163" y="335">30</text><text x="247" y="335">60</text><text x="330" y="335">90</text><text x="413" y="335">120</text><text x="497" y="335">150</text><text x="580" y="335">180</text> </g> <text x="30" y="200" text-anchor="middle" font-family="Georgia, serif" font-style="italic" font-size="10" fill="#43331f" transform="rotate(-90 30 200)">blood glucose, mg/dL</text> <text x="330" y="355" text-anchor="middle" font-family="Georgia, serif" font-style="italic" font-size="10" fill="#43331f">time after eating, minutes</text>

<line x1="80" y1="260" x2="580" y2="260" stroke="#a9906c" stroke-width="0.5" stroke-dasharray="1,3"/> <text x="586" y="263" font-family="Georgia, serif" font-style="italic" font-size="8" fill="#6f5a40">fasting</text>

<path d="M80,260 C105,220 140,105 163,100 C185,96 195,108 205,120 C222,140 235,180 247,200 C262,225 300,290 330,300 C365,308 390,290 413,280 C445,268 470,262 496,260 C520,259 555,260 580,260" fill="none" stroke="#9c5a2c" stroke-width="1.3"/> <g fill="#9c5a2c" stroke="none"> <circle cx="163" cy="100" r="2.2"/><circle cx="330" cy="300" r="2.2"/> </g>

<path d="M80,260 C130,245 160,232 200,215 C230,204 247,200 270,198 C300,196 315,196 330,196 C355,197 375,205 413,220 C450,234 470,244 497,244 C525,248 555,256 580,260" fill="none" stroke="#43331f" stroke-width="1.3"/> <g fill="#43331f" stroke="none"> <circle cx="330" cy="196" r="2.2"/> </g>

<rect x="400" y="70" width="165" height="52" fill="none" stroke="#c8b79a" stroke-width="0.6"/> <line x1="410" y1="85" x2="435" y2="85" stroke="#9c5a2c" stroke-width="1.3"/> <text x="440" y="88" font-family="Georgia, serif" font-style="italic" font-size="9" fill="#43331f">refined starch — sharp spike</text> <line x1="410" y1="105" x2="435" y2="105" stroke="#43331f" stroke-width="1.3"/> <text x="440" y="108" font-family="Georgia, serif" font-style="italic" font-size="9" fill="#43331f">fibre-rich whole food</text> </svg> <figcaption><em>Plate 4. Glycaemic-response curves for three carbohydrate foods of equal carbohydrate content: the sharp spike and rapid fall of a refined starch, set against the gentler, more sustained rise of a fibre-rich whole food</em></figcaption> </figure>

Insulin, resistance, and the shape of a meal

Insulin itself deserves a moment's proper introduction, for it is too often mentioned in popular writing only as a villain, when it is in fact one of the body's more elegant pieces of chemical bookkeeping. Secreted by the beta cells of the pancreas in direct response to rising blood glucose, insulin binds to receptors on the surface of muscle, fat and liver cells and, in effect, unlocks the door through which glucose passes from bloodstream to cell interior, where it can be burned for immediate energy or stored — as glycogen in muscle and liver, as fat in the body's adipose tissue — against future need. In a healthy metabolism this system runs with remarkable precision, blood glucose rising modestly after a meal and returning to a narrow baseline within a couple of hours, insulin rising and falling in close step behind it.

Insulin resistance is what happens when this precise choreography begins, gradually and often silently, to break down: the muscle, liver and fat cells become progressively less responsive to insulin's signal, so that the pancreas must secrete more and more of the hormone to achieve the same effect, a state of compensatory hyperinsulinaemia that can persist, often without symptoms and without the diagnosis being made, for years before blood glucose itself finally begins to rise into the range of prediabetes and, eventually, type 2 diabetes. The causes are genuinely multiple — genetic predisposition, excess visceral fat, physical inactivity and chronic sleep disruption all contribute independently — and no single dietary villain fully explains the modern rise in insulin resistance across Western and rapidly urbanising populations alike. But a dietary pattern built substantially around foods of high glycaemic load, eaten frequently and in large portions, does appear, across a considerable body of prospective and interventional evidence, to place a chronic strain on the very system insulin resistance eventually overwhelms, in much the way a spring rung out to its full extension too many times gradually loses its snap.

Fibre changes the picture

It is here, at the close of the carbohydrate story, that fibre earns the pivotal role this chapter has been building toward, and the reason is worth stating plainly: fibre is itself a carbohydrate, chemically a long chain of sugar units very much like starch, yet it is a chain the human gut's own enzymes cannot break, and that single structural quirk changes almost everything about how the rest of a meal behaves. Soluble fibre — the pectin in an apple, the beta-glucan in oats, the gum in a lentil — dissolves in water to form a viscous gel that physically slows the stomach's emptying and the small intestine's absorption of the sugars sitting alongside it, blunting the glycaemic spike a food would otherwise produce; a bowl of oatmeal and a slice of white toast may contain comparable amounts of starch, yet the oatmeal's soluble fibre delivers that starch to the bloodstream at a distinctly more leisurely pace. Insoluble fibre — the bran of a wheat kernel, the tough cell walls of a raw vegetable — works by a cruder but no less useful mechanism, adding bulk and speeding transit through the gut, and a portion of both types survives digestion altogether to reach the colon, where the resident microbial population ferments it into short-chain fatty acids, a story this compendium will take up properly in the chapter that follows this one.

The practical upshot is that the same starch, eaten in two different guises, can produce two quite different glycaemic outcomes, and nowhere is the demonstration more vivid than in a single grain of wheat. A whole wheat kernel arrives wrapped in bran, rich in fibre, and studded through its germ with protein, B vitamins and healthy fat, all of which combine to slow the release of the starch held in its central endosperm; refine that same kernel into white flour — a process perfected industrially in the nineteenth century, to considerable fanfare and, as the next chapter's account of pellagra will show, considerable unintended cost — and the bran and germ are stripped away, leaving behind a nearly pure starch that the gut absorbs almost as quickly as if it had been fed a spoonful of sugar directly. This is not an argument that refined flour must never be eaten, any more than the saturated fat argument was licence to purge butter from every larder; it is, rather, the clearest possible illustration of this volume's recurring lesson, that a nutrient's effect on the body cannot be reliably predicted from its molecular label alone, but only from the whole food, matrix intact, in which it happens to arrive.

Three great classes, then, exactly as Prout described them in 1827, and yet the naturalist closing this chapter should feel considerably better equipped than Prout ever was to say what each one actually does once it has passed the lips: protein supplying the amino-acid bricks of every living tissue, fat carrying both fuel and a genuinely contested cardiovascular argument this compendium has tried to report rather than settle, and carbohydrate arriving at the bloodstream at a pace fibre alone can properly govern. But macronutrients, for all the pages devoted to them here, are only the bulk of the matter — the timber and stone of the body's construction, weighed in grams and measured in calories. There remains a second, far smaller class of dietary substance, required not in grams but in milligrams and micrograms, whose absence killed sailors by the thousand, blinded children, and warped the bones of a whole industrial century before a handful of tenacious hunters finally ran its culprits to ground. It is to that menagerie of vitamins and minerals — smaller in quantity, no smaller in consequence — that this compendium turns next.

[^1]: The nine essential amino acids are traditionally remembered by nutrition students through a variety of mnemonics of variable decency; this compendium declines, on grounds of taste, to reproduce any of them here. [^2]: Magendie's dogs were, by modern standards, treated with a callousness no laboratory would sanction today; the experiment is recorded here for its scientific consequence, not as an endorsement of its method.

บทถัดไป · 6 / 17Chapter 3 — The Micronutrient Menagerie