Protein Leverage

An hypothesis for weight gain

Photo by Caroline Attwood on Unsplash

Protein is our primary biology. Indeed, the word comes from the Latin ‘protos’, meaning first. Our proteins are made up of various combinations of 20 amino acids (AA), 9 of which we must get from our diet (the essential amino acids, EAA), and the rest we can make ourselves. Still, the proportion of EAAs (nearly half) indicates the importance of quality dietary protein to our biology.

In our digestive system, dietary proteins (animal or plant) are broken down to their AAs (or small AA assemblies, known as peptides), absorbed by the small intestine, then reassembled (according to DNA instructions) to make crucial structural and functional proteins for our bodies, e.g: actin (underpinning skeletal muscle contraction); haemoglobin (carrying oxygen in red blood cells); keratin (hair); collagen (connective tissues); insulin (and other hormones); enzymes (to catalyse reactions); antibodies (of the immune system); neurotransmitters (brain signalling); components of cell membranes (all cells), and so forth. Physically and mentally, we are protein. And fats too — fatty acids have multiple roles and some are also essential in the diet. It makes no sense to think of ourselves as carbohydrates.

Protein Quality (PQ) refers to the protein’s AA profile — is it ‘complete’ (all 20 AA, including the EAA) and are they in proportions appropriate to our biology? PQ is high for animal-based foods, but generally less so for vegetal foods (soy, quinoa can be exceptions). Variety is said to make up for potential EAA shortfalls within vegetal diets. The AAs in animal-based foods tend to be more bio-available (digestible and absorbable) than that from vegetation.

The US Institute of Medicine (IoM) sets a Recommended Daily Allowance (RDA) for virtually everything. It is not always possible to arrive at an RDA rigorously, and many assumptions are made. Nevertheless, the RDA for protein is deemed to be 0.8g per kg of reference body weight per day for adults (>19 years of age) of any gender (a reference weight is an ideal weight according to height, gender, and age — e.g. for a male, 50 years of age, it’s 70kg rather than his actual weight). Thus, a 70kg reference-weight adult is recommended to consume 56g of protein a day. There are also RDAs for each individual AA.

Protein recommendations are best thought of as minimum daily protein intake, and many scientists consider this to be too low for optimal health — especially for maintaining muscle mass in older adults at risk of sarcopenia (age-related muscle loss). Somewhere in the range 1.2–1.8g per kg might be a better goal.

The IoM figure comes from nitrogen balance studies, which are widely recognised, even by the IoM itself, to have multiple limitations. Each AA has one nitrogen atom, so the method is to compare the nitrogen in protein being fed to participants, to the nitrogen coming out in urea and faeces, and estimating other losses (hair, skin). The level of protein consumption that yields detectable nitrogen in the outputs suggests that the body has used all the dietary protein it needed. The range in this level is determined across a group of participants and the RDA set to the maximum (so that everyone gets adequate protein).

The IoM made these calculations in 2002, and they haven’t been updated. Since then, there have been newer and better methods developed however, no-one can decide which should form the next gold-standard method, so the old calculations, with all their limitations, still stand.

There was no Tolerable Upper Intake specified for protein — high protein intake does not appear to be associated with adverse health outcomes. Low protein is. Initially, there was concern that high protein could stress the kidneys (nitrogen is stripped off AA and excreted by the kidneys as ammonia), however, this is no longer considered a concern, at least for dietary protein sources (as opposed to supplement overdosing).

As an alternative, the IoM also provided an Acceptable Macronutrient Distribution Range ( AMDR), which they expressed as a percentage of daily Calories (rather than grams per kg of body weight). The ADMR is 10–35% of total Calories for adults. For a reference diet of 2,000 Calories (and taking the convention of 4 Calories per gram of protein), the ADMR translates to 50g — 175g of protein per day. Thus, the 10% figure roughly corresponds to the RDA (which remember should be considered a minimum). The upper range (35%) was arbitrarily set (according to the opinion of panel members). An estimate of actual protein intake for US adults is ~10–15%.

While 10–15% of protein seems to be achievable for most people, the upper limit (35%) might be a challenge for a mixed diet without supplements (an exception is the increasingly popular carnivore diet). For example, a large egg contains ~6g of protein, so 2 two breakfast eggs only gets us to 12g. Protein makes up around a quarter of the weight of beef and other animal meats (like most foods, they’re mostly water), so a 100g steak for lunch will add another 25g of protein. Most legumes are ~8–9% protein, so 100g of kidney beans will contribute 8–9 g (assuming complete absorption and no protein leeching if boiling). We are on the way to the lower end of the ADMR range, but well short of the upper end.

For the AMDR conversion, I used the conventional factor of 4 Calories per gram of protein. This comes from the 4:4:9 ‘rule’: 1 g carbohydrate contains 4 Calories; 1 g protein contains 4 Calories and; 1 g fat contains 9 Calories. These values were estimated by the American chemist Wilbur Atwater in the 1890s. He tried to factor in partial digestion, urinary excretion and fibre content, however the factors remain gross approximations.

In physics, a calorie is a unit of heat energy (the energy needed to heat 1g of water by 1 degree Celsius). Nutritionists love using principles from physics (e.g. calories-in, calories-out) because they imagine it lends authenticity to their profession. In nutrition, a Calorie (capitalised) is 1,000 calories (1 kcal). To determine how many Calories are contained in a macronutrient (such as protein), Atwater totally incinerated it, with a high-voltage current, and measured the heat of combustion that was given off. He did this by immersing a chamber containing the protein (and oxygen) in a water bath and measuring the rise in water temperature as its contents were incinerated. This is known as a bomb calorimeter, and other macronutrients (or other foods) derive their Calorie conversions from this method.

The Calorie-based assumption is that total incineration is also the fate of dietary protein when it enters our body — it gets burned entirely for energy. This is an absurd assumption. It is likely that protein will be prioritised for its myriad uses as outlined in the opening paragraph of this post. Further, digesting protein is energy demanding, as is stripping nitrogen off AAs, thus consuming Calories. Chewing food consumes Calories and adds to satiety. Was the protein cooked, and how was it cooked (e.g. ~90% of cooked egg protein can be absorbed, whereas only ~50% of a raw egg protein is). Even food preparation matters — cutting and grinding are forms of pre-digestion. What about the food that accompanies the protein in a meal — were there carbohydrates (glucose) in the meal that might get preferentially burned for energy so as to be protein-sparing?

The problem is not with the calculation though, but rather the concept of counting Calories at all. The calculation ignores our biology. For example, we might expect a biological difference between consuming 100 Calories of beef vs. 100 Calories of table sugar (note that the Calories are being kept constant). Setting nutrients aside, the Calorie-counting approach would claim the biological outcome (e.g. weight gain) would be the same. Ponder that. The problem with the Calorie Model is that it treats our complex biology as though it is unsophisticated and unregulated.

As an example of the futility of Calorie-counting, many of us, as we age into midlife, might slowly add a little weight. Maybe just 1 kg a year, resulting in a 10 kg increase from, say, 40–50 years of age. Real weight gain is a slow process, measured in year-timescales, whereas daily/weekly weight fluctuations are due to variations in our water content and of no significance. If the Calorie Model was true, we could estimate how many Calories needed to be overeaten to gain 1 kg a year. It’s 21 Calories a day. That’s 3 almonds. On a 2,000 Calorie daily reference value, that means regulating calculated Calorie intake to within ~0.1% (to make a 1% variation, 21 Calories, meaningful). Rather than eating 3 fewer almonds a day and expecting to curb the middle-age spread, we should instead acknowledge the wrongness of even thinking that way at all. Most dietary experts recommend thinking that way.

The trouble is, Calories do count to some degree, especially at the extremes of starvation and morbid gluttony. However, this causes confusion because Calorie-based dietary advice is said to apply to the broad population. For most of us it’s not relevant, we should be listening to biological signals rather than making decisions based on error-prone calculations. The best advice is this: “Calories count, but you don’t need to count them”. We should be guided by what we eat, and our internal hunger and satiety signalling meal-to-meal. Which brings me to an intriguing hypothesis — the Protein Leverage Hypothesis.

In 2005, two University of Sydney scientists, Simpson and Raubenheimer, proposedthe hypothesis that the phenomenon of protein leverage could explain weight-gain in the human — the protein leverage hypothesis (PLH). Note that protein leverage itself is not the hypothesis, it is established. The hypothesis relates to whether protein leverage can explain weight-gain on low-protein diets. The PLH proposes that there is a dietary protein target to be met. This acknowledges the primary place of proteins in our biology. However, a secondary effect of a protein target is that if the diet is protein poor, there could be an overconsumption of carbohydrates and fats (nonprotein energy) in the course of meeting the protein target — i.e., nonprotein energy is leveraged by protein, with the compromise that total energy intake is greater than need. The PLH has been verified in controlled-feeding studies in animals for which macronutrient ratios can be varied (within a range that could be expected for the animal).

A number of studies have provided support for the PLH in humans. It is complex to study in humans though, especially in free-living circumstances, and protein leverage may vary with multiple factors (e.g. age, gender, activity, environment, reproductive status, cultural and genetic background).

One important distinction is that food selection is a different thing to the regulation of the amounts of foods eaten. For example, fat+sugar combinations (ice cream, chocolate, cakes etc) may be preferentially selected, leading to low protein intake, but overall consumption would then be increased to achieve a target protein intake. That’s no hardship because the foods were chosen because they were palatable. The PLH is that such a scenario leads to weight gain.

Finally, I don’t want this discussion to confuse the Calories story. The PLH in this scenario leads to an energy oversupply, but the solution is not to cut Calories, rather to recognise the importance of protein to our biology, and shift the balance of macronutrients to achieve better protein supply. Because protein is satiating, just stop when satiated. The fat+sugar combination does not come with satiating signals, and if anything it is addictive. Processed food manufacturers know that, and strive for it — it’s called the ‘bliss point’ that Big Food is looking for to enrich their shareholders.

Protein is kept ‘moderate’ on this diet, in the 1.2–1.8 g per kg range. This should be sufficient under most circumstances (e.g. unless building muscle). It is fats that are increased with a KD, and these often come with proteins anyway if making animal-based food choices. Fats are also satiating, and the same recommendation applies — stop eating when satiated. No Calorie-counting needed.

The moderate protein recommendation came about because, unless we are building muscle, we don’t have anywhere to store excess protein. One option available is to strip the nitrogen from the AAs, and use their remaining carbon backbones to make glucose for energy (gluconeogenesis), which would undermine the objective of the state of ketosis. However, newer studies have suggested that this doesn’t seem to happen with the KD. The liver makes other molecules out of protein that can serve as fuel for oxidation. It only makes glucose if there is a demand for it, rather than to use up an oversupply of protein.

Fats, carbohydrates and Calories predominate in the language surrounding nutrition and obesity. Proteins are the silent partners, sitting in the background and often overlooked. Overlooked despite their importance for our biology. It would be ironic if the underestimation of their significance was a driver of obesity.

The 2005 paper of Simpson and Raubenheimer and a 2019 review (unfortunately both behind paywalls).

High protein diets on



Science of cooking, eating and health. Retired neuroscientist.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store

Science of cooking, eating and health. Retired neuroscientist.