Beyond Calories and Macro Wars: What Protein Leverage Tells Us About How We Really Eat

For decades, nutrition has been framed as a competition between macronutrients.

First, fat was blamed for obesity and cardiovascular disease. Then carbohydrate became the problem, particularly through its effects on insulin.

More recently, protein has increasingly been promoted as the solution: more satiety, better muscle maintenance, improved body composition and easier weight control. At the same time, another camp warns that too much protein could activate growth pathways such as insulin, IGF-1 and mTOR, accelerate ageing, or possibly increase long-term disease risk.

So which side is right?

A recent review in Nature Reviews Endocrinology by Simpson and colleagues suggests that the problem may be the question itself. Protein, carbohydrate and fat do not act independently. We eat them as mixtures, and their biological effects depend on their proportions, quality, physiological context and stage of life.

One idea in particular — the protein leverage hypothesis — provides a useful way to reconsider not only protein, but also calories, carbohydrate, fat and the modern food environment.

Why do we keep eating?

The conventional explanation for weight gain is straightforward:

Energy intake exceeds energy expenditure.

That statement is physically correct. Stored energy has to come from somewhere.

But it does not necessarily explain the biology.

It tells us what happened, but not necessarily why appetite drove the person to consume more energy in the first place.

This distinction becomes important when we look at protein.

Evidence summarized in the review suggests that humans regulate protein intake more tightly than either carbohydrate or fat. Across populations, the percentage of energy coming from carbohydrate and fat can vary considerably, whereas protein intake tends to remain within a much narrower range.

This observation led to the protein leverage hypothesis.

The idea is simple.

If the body's protein requirement remains relatively stable but the food being eaten contains less protein relative to fat and carbohydrate, the person may eat more total food in an attempt to reach that protein target.

The sequence might look like this:

Lower protein density

→ protein requirement remains unmet

→ appetite continues

→ more food is eaten

→ more carbohydrate and/or fat is consumed

→ total energy intake rises

In other words, excess calorie intake may sometimes be partly the consequence of nutrient regulation, rather than the original cause of the problem.

Protein leverage changes how we think about overeating

The attached figure illustrates this particularly well.

Across human feeding studies spanning diets containing roughly 10–30% of energy from protein, absolute protein intake remains comparatively stable while total energy intake rises as the percentage of protein falls. At very low protein concentrations — around 10% or below — this compensatory mechanism eventually fails, because increasing food intake can no longer adequately compensate for the protein dilution.

That gives us a very different way to interpret appetite.

Instead of asking only:

Why is this person eating too many calories?

we can also ask:

What resource is the body still trying to obtain?

Protein is especially important because the body has dedicated stores for carbohydrate and fat, but no equivalent warehouse for protein. Amino acids are needed continuously for muscle, enzymes, transport proteins, immune function, tissue repair and many other processes.

If dietary protein is repeatedly diluted, increasing food intake can therefore be an entirely logical biological response.

The problem is what comes along with that additional food.

The modern food environment may exploit an ancient adaptive system

Protein leverage becomes particularly relevant in the modern industrial food environment.

Many highly processed foods combine: relatively low protein density

• refined carbohydrate

• added fat

• low fibre

• high energy density

• strong palatability.

A biological system trying to obtain enough protein may therefore consume substantial amounts of non-protein energy before its protein requirement is satisfied.

The review even discusses the concept of “protein decoys”: highly savoury foods that provide sensory cues associated with protein but deliver disproportionately large amounts of fat and carbohydrate.

This makes obesity much more interesting than a simple failure of willpower.

The organism may be performing exactly the behaviour its regulatory machinery evolved to perform — seeking a limiting nutrient — but doing so in a food environment for which that machinery was never designed.

An adaptive response can therefore generate an unintended metabolic cost.

Does this mean we should simply eat more protein?

Not necessarily.

This is where the review becomes much more interesting than the usual high-protein versus low-protein argument.

Higher-protein diets can clearly offer advantages.

Increasing the proportion of dietary protein can increase satiety, reduce total energy intake and support maintenance of lean tissue. This helps explain why higher-protein diets can be effective for weight loss and metabolic management.

But it does not follow that:

more protein = progressively better health.

Different biological outcomes appear to have different nutritional optima.

A diet that produces the leanest body may not necessarily produce the best longevity phenotype.

A diet that promotes reproduction and muscle growth may not be the same diet that maximizes long-term maintenance.

And the protein requirement of a 30-year-old is not necessarily the same as that of an 80-year-old.

The protein paradox

Some of the experimental findings summarized in the review seem contradictory at first.

In mice, moderately low-protein, high-carbohydrate diets can increase food intake and adiposity.

Yet under the right carbohydrate conditions, these diets can also improve several cardiometabolic markers and extend lifespan.

Meanwhile, high-protein, low-carbohydrate diets can produce leaner animals while being associated with less favourable longevity-related outcomes.

This has sometimes been described as a protein paradox.

But perhaps it is only paradoxical if we assume that leanness, metabolic health, muscle growth and longevity should all have the same nutritional optimum.

There is no obvious biological reason why they should.

The nutrients required to maximize reproduction are not necessarily those that maximize lifespan.

The nutrients required to build muscle are not necessarily those that optimize every aspect of cellular maintenance.

Biology is full of trade-offs.

But “high protein activates mTOR” is also too simplistic

One frequently proposed explanation for the longevity findings is that high protein increases insulin, IGF-1 and mTOR signalling.

These pathways are often described as “pro-ageing.”

The review discusses evidence linking high-protein dietary contexts with increased insulin, IGF-1 and mTOR signalling and with shorter lifespan in experimental animals.

But this needs an important physiological qualification.

Insulin, IGF-1 and mTOR are not inherently pathological pathways.

They are essential components of normal growth, repair and adaptation.

After resistance exercise, for example, we want amino acids and mTOR signalling to stimulate muscle protein synthesis.

After tissue injury, we want anabolic processes to support repair.

After fasting, we want feeding to switch the organism back toward rebuilding.

The body is not designed to remain permanently anabolic or permanently catabolic.

Healthy physiology oscillates:

feeding ↔ fasting

anabolism ↔ catabolism

exercise ↔ recovery

wakefulness ↔ sleep

tissue breakdown ↔ rebuilding

The important biological question may therefore not be simply:

Does protein activate mTOR?

Of course it does.

A better question is:

How strongly, for how long, how frequently, in response to what demand — and does the system subsequently switch back?

A transient anabolic response after exercise and a meal is fundamentally different from persistent nutrient exposure in an inactive organism that rarely enters a fasting or recovery state.

This rhythmic dimension is not captured well by looking only at average daily macronutrient percentages.

Nutritional geometry may need a time axis

One of the strengths of the review is its use of the Nutritional Geometry Framework.

Rather than studying protein, carbohydrate and fat individually, nutritional geometry maps biological outcomes across different combinations of all three.

This has helped reveal interactions that disappear when nutrients are examined one at a time.

But physiology contains another dimension:

time.

Two people could consume the same calories and macronutrient proportions but experience very different metabolic patterns.

One might eat three discrete meals, exercise regularly and have a long overnight fasting period.

Another might consume nutrients almost continuously from early morning until midnight while remaining sedentary.

The average daily nutrient intake could look similar.

The physiological signalling pattern would not.

Eventually, nutritional models may therefore need to consider something like:

composition × quality × demand × timing × recovery

rather than composition alone.

Carbohydrate is not one biological entity either

The same mixture logic dismantles the idea that “high carbohydrate” is a meaningful nutritional category by itself.

The review describes experiments in which the metabolic advantages of low-protein, high-carbohydrate diets depended strongly on the type of carbohydrate.

When carbohydrate consisted largely of complex or digestion-resistant starch, favourable metabolic effects were observed.

When the carbohydrate was replaced by simple sugars, those benefits disappeared and metabolic outcomes worsened. Protein leverage became stronger, glucose regulation deteriorated, and accumulation of body and liver fat increased.

A particularly unfavourable combination in the experimental studies was approximately equal amounts of glucose and fructose.

So:

high carbohydrate from intact, complex foods

is not metabolically equivalent to

high carbohydrate from refined sugar.

Calling both “carbohydrate” hides more information than it reveals.

Nor can fat simply replace carbohydrate

The same applies to fat.

One of the interesting findings from the nutritional-geometry experiments was that replacing carbohydrate with fat did not reproduce the metabolic and longevity benefits associated with low-protein, high-carbohydrate diets.

That matters because we often talk about “protein restriction” as though protein percentage alone were the intervention.

It isn't.

Reducing protein means something else must occupy the remaining dietary energy.

Therefore:

low protein + complex carbohydrate

is a different biological exposure from:

low protein + fat

which is different again from:

low protein + refined sugar.

The biological unit is the mixture, not the isolated macronutrient.

Even protein is not just protein

Protein itself is a mixture of amino acids.

The review discusses branched-chain amino acids — leucine, isoleucine and valine — as an example.

Higher BCAA intake has been associated with insulin resistance and less favourable metabolic profiles in some settings.

Yet in older adults, lower BCAA intake has also been associated with greater frailty and mortality.

Again, apparently contradictory findings may reflect differences in:

age

physiological demand

amino-acid balance

total protein intake

and the surrounding carbohydrate and fat mixture.

There may therefore be little value in asking whether “BCAAs are good” or “BCAAs are bad” without specifying the biological context.

Protein needs also change with age

Perhaps the strongest argument against treating protein as universally good or bad comes from ageing.

In animal studies discussed in the review, lower-protein diets tended to be associated with better survival through much of adulthood.

But among animals reaching very old age, the relationship reversed: higher protein became associated with lower age-specific mortality.

This makes physiological sense.

Ageing is accompanied by anabolic resistance. Older muscle does not respond to protein and exercise as efficiently as younger muscle.

The consequences of insufficient protein therefore become increasingly important:

loss of muscle,

loss of strength,

frailty,

poorer recovery,

and ultimately loss of functional independence.

Protein that might represent unnecessary anabolic stimulation in one biological state may represent an essential recovery resource in another.

The more useful question becomes:

How much protein is required for the present level of anabolic demand?

Not:

Is high protein good or bad?

What about kidneys and cancer?

These concerns are frequently raised in debates about high-protein diets, but this particular review does not provide evidence for a simple conclusion that high protein causes kidney disease or cancer.

In fact, the review notes that some kidney-function outcomes in animal studies improved with higher protein intake, illustrating an important principle: different tissues can have different nutritional optima. 

Similarly, although IGF-1 and mTOR are involved in cellular growth and cancer biology, demonstrating that protein transiently activates these pathways is not the same as demonstrating that dietary protein causes cancer in humans.

These questions require much more than a single signalling pathway.

The distinction between normal, rhythmic anabolic signalling and pathological, autonomous growth is critical.

So are calories irrelevant?

No.

Calories remain fundamental to energy balance.

But calories are an accounting unit, not a complete biological explanation.

Two diets containing the same energy can generate different:

appetite signals,

hormonal responses,

microbial metabolites,

nutrient deficiencies,

satiety,

body composition,

substrate oxidation,

and behavioural responses.

Protein leverage demonstrates the distinction particularly clearly.

If low protein density makes a person eat more in order to obtain enough protein, then calorie excess is real — but the calorie excess has a biological cause.

Saying only that the person ate too many calories misses that regulatory mechanism.

Energy balance tells us where the energy went.

Physiology helps explain why the system sought it in the first place.

From macro wars to biological context

The most useful contribution of this review may therefore not be deciding whether high

protein or low protein is superior.

It is showing why that entire style of question is inadequate.

Protein is neither inherently protective nor inherently harmful.

Carbohydrate is neither inherently metabolically damaging nor universally healthy.

Fat cannot be understood merely as stored calories.

And calorie balance, although physically unavoidable, cannot by itself explain appetite or nutrient-seeking behaviour.

The body is continually allocating resources between:

growth

activity

immune defence

reproduction

maintenance

repair

and recovery.

Its nutritional requirements change accordingly.

Protein leverage provides an unusually clear example of this regulatory logic. When a required resource becomes diluted, the organism changes its behaviour to obtain more of it. In the modern food environment, that adaptive response can come with a substantial collateral cost in excess energy intake.

Perhaps the future of nutrition therefore lies less in finding the perfect macronutrient percentage and more in understanding whether nutrient supply matches physiological demand — and whether the system remains capable of moving appropriately between metabolic states.

Instead of asking:

How much protein?

How many carbohydrates?

How much fat?

How many calories?

we may need to ask a better question:

What does the body need at this moment, what mixture are we supplying, and can it return to balance once that need has been met?

That question brings nutrition much closer to the way physiology actually works.

Reference

Simpson, S. J., Le Couteur, D. G., Small, L., Brandon, A. E., & Raubenheimer, D. (2026). Macronutrient mixtures and interactions in health and disease. Nature Reviews Endocrinology. https://doi.org/10.1038/s41574-026-01266-5