Eat Like the Animals: The “Science-Based” Diet Book about the Protein Leverage Hypothesis

Most diet books start with an agenda or a belief about nutrition and try to build science around it.  

Eat Like the Animals is the opposite.  Or at least that was what I thought until I dug into the data behind the biggest and most expensive mouse study ever undertaken! 

For the most part, Eat Like the Animals is data-driven biology at its best.  The narrative of discovery, written by the researchers themselves, as they take you on their journey of enlightenment, keeps it entertaining and makes it memorable!

The birth of the Protein Leverage Hypothesis

University of Sydney Professors Stephen Simpson and David Raubenheimer started out studying the eating habits of insects.  Through some fascinating and incredibly detailed and patient experimentation, they found that they could create radically different outcomes in terms of obesity and longevity by manipulating the insects’ diets.

The Protein Leverage Hypothesis was born (as documented in their 2005 paper, Obesity: the Protein Leverage Hypothesis) when they noticed that insects continued eating until they obtained a minimum amount of protein.  

Did you know crickets turn cannibal and eat each other if they can’t get enough protein in their diet?  

It seems their motivation to march across the desert by the billions is partly to find food and partially not to get eaten by the bug behind them!  

Our craving for protein is the strongest of all of our appetite signals.  

If the food we eat contains a higher percentage of protein, we tend to consume fewer calories.  Conversely, if our diet is diluted by refined and processed high carb and fat ingredients, we tend to eat more.  

Simpson and Raubenheimer’s seminal work demonstrates that this pattern is consistent across all living things, from plants to insects, to animals (including humans).  

What is a balanced diet?

In the chart, below the green line represents a balanced diet with adequate protein to support lean muscle mass and sufficient easily accessible energy from carbs and/or fat to prevent weight loss.  

Most of the time, there is no single perfect food that provides a balanced diet, so we eat a little of this and a little of that until we uncannily end up at our target with enough protein and enough energy at the same time.  

Simpson and Raubenheimer were amazed as they monitored endless animal diets in the wild and found that, although it appeared random at the time, when they tallied it up, it always balanced out to the same point!

Why we love steak and chips

Ever wonder why a meal of steak and chips is so popular?  

Most people don’t eat only chips ALL the time.  

Most people don’t eat lean steak ALL the time.  

Even though the chips are yummy (because they provide lots of easy energy from fat and carbs together), eventually we crave protein.  Our appetite sends us in search of the steak.  

Conversely, if we started out just eating lean steak (i.e. protein with minimal fat), we would eventually have a craving for some fat and/or carbs to get the energy we need to operate because it’s hard to convert the pure protein to ATP (i.e. your body’s energy currency). 

Even people who follow a carnivorous diet know they need enough fat to prevent hunger and excessive weight loss.  We just can’t eat enough lean protein to get adequate energy.  

This plays out in the experiments in animals over and over.  When they fed animal X or insect Y a lot of protein and animal X or insect Y a lot of carbs+fat and then let them have access to both food sources they would amazingly eat just enough of the right food source to arrive at their target.  

This balanced diet is different for different species and different people at different life stages, activity levels, etc.  Older people or athletes need a bit more protein, while younger people can get away with less.  But, regardless, like a homing pigeon, our appetite has an incredible way of navigating us back to the balanced diet that we require!

While we like to think we can count calories to limit our intake, resistance is mostly futile over the long term (unless you upgrade the protein percentage of the food you eat).  

Optimising Nutrition Advisor Dr Ted Naiman likes to make memes from Simpson and Rabenheim’s data which has been a significant influence on his Protein:Energy Ratio.   

The plot of protein % vs fat % with the blue-red heat map (Simpson and Raubenheimer call this their Geometric Framework for Nutrition), shows that we tend to eat fewer calories when we have more protein.   Conversely, we tend to consume more energy when we eat foods with a lower percentage of protein.  

If you want to achieve an energy deficit without losing muscle mass and without excessive cravings, you need to follow a diet with a higher percentage of protein that will enable you to obtain the protein you need without excess energy.   

However, we get into trouble when we only have access to low protein foods (e.g. modern hyperpalatable processed junk food), as we always overshoot our energy requirement to get our protein requirement.  

I found it fascinating that, in “Eat Like the Animals”, they are not talking about huge amounts of protein, just a slightly higher percentage of protein than what we are currently consuming.  

In the examples in the final chapter of the book, they talk about the scenario of Mary slipping from 15% protein to 13% protein resulting in an increase of 290 calories per day energy intake leading to insidious weight gain over time.   Whereas shifting from 15 to 17% would drive a reduction of 150 calories per day.  

It’s important to note that this does not necessarily require eating more protein, but rather making slightly better food choices such as skipping a pack of chips, pizza or doughnuts. These foods are low in protein and have a combination of carb and fat that is impossible to resist when available to us.  

Our analysis of data from Cronometer and MyFitnessPal takes a slightly different approach but arrives at the same place conclusions as most of their work (make sure you keep reading for the plot twist).  Foods with a higher percentage of protein are simply harder to overconsume.  

While we need enough protein to meet our needs, our appetite for higher protein foods turns off once we get enough.  Minimally processed protein is a poor source of energy due to its high dietary-induced thermogenesis, meaning your body has to work harder to use it for everyday energy requirements.  

The frequency distribution shows that around 20% protein is typical for Optimisers.  Not many people manage to sustain much higher protein intakes than this.   

The chart below (based on data from the USDA Economic Research Service) shows that, in percentage terms, protein intake has ranged from 10.5% (i.e. in the 1930s after World War I) to 13% of calories 1977 (i.e. when the first US Dietary Guidelines were released telling people to avoid red meat due to fear of cholesterol).  

It’s worth noting that during the time that protein percentage has decreased over the past 50 years, the obesity epidemic has flourished.   

As shown in this next chart showing the change in calories from each of the macronutrients over the past hundred years or so, the obesity epidemic has been fuelled by an increase in easily available energy from both carbs and fat while protein has remained fairly stable.  

In terms of actual protein intake in grams, most people struggle to eat very high levels of protein. Optimisers tend to gravitate back to around 1.8g/kg LBM per day.

Eat Like The Animals shows how protein dilution of our food system from 15 to 13% protein is more than enough to explain our growing obesity epidemic, while a shift from 15 to 17% is more than enough to reverse obesity and diabetes over time.  

However, while this sounds simple, it is hard to do because we find foods with lots of fat and carbs (with low protein) irresistible whenever they are available. 

Sadly, this describes our modern food system, particularly the world of cheap hyper-palatable hyper-profitable junk food.

The plot twist! 

Where “Eat Like the Animals” gets screwy is in chapter 8 where they discuss their “mega mouse” study The Ratio of Macronutrients, Not Caloric Intake, Dictates Cardiometabolic Health, Aging, and Longevity in Ad Libitum-Fed Mice with 14 high profile authors (including longevity and mTOR guru David Sinclair). 

In our previous article, Can longevity be bought in a bottle? Thoughts on David Sinclair’s Lifespan, we critiqued the competing goals of a plant-based low protein diet for longevity, nutrient density and optimal body composition.  The bottom line is that there is no human data on protein restriction leading to longevity. Theories developed in yeast and worms can’t be applied to humans.

Sadly, it appears that the fear of protein and mTOR due to ageing seems to have seeped into Simpson and Raubenheimer’s research in this paper as well.  

The mega mouse study was one of the most expensive nutritional experiments ever conducted, with (initially) 1000 mice being fed a matrix of (initially) 30 different diets with varying energy density (high, medium and low) and macronutrient contents.  

The graphical summary of the study is shown below, indicating the high protein, low carb diet caused a decrease in food intake and a decrease in adiposity (both of which align with what we know of humans and from our satiety analysis).  

However, surprisingly, they also concluded that the mice on a high protein diet had worse metabolic health and a worsening lifespan?  Meanwhile, the mice on a low protein high carb diet were fatter but (apparently) had better metabolic health and greater lifespan.

After the study was published, the University of Sydney came out with a number of press releases and news pieces saying that high protein diets may shorten your life.  

What the?  This is not what I was expecting from the guys who developed the Protein Leverage Hypothesis?

Does obesity lead to longevity in humans?

So,  I have a question…  

Where do we see increased adiposity in humans leading to improved metabolic health and increased lifespan?  



That’s right. 


Increased obesity is a major risk factor for pretty much every modern disease and cause of reduced lifespan.  Excess body fat and dwindling levels of muscle mass are at the root of pretty much every modern metabolic disease, from diabetes, to heart disease and cancer.

What we do know about humans is that those who have better metabolic health are leaner and live longer.  A couple (of the many) examples are shown in the charts below from Waist-to-height Ratio Is More Predictive of Years of Life Lost Than Body Mass Index where we see that a lower BMI or waist to height ratio tends to align with greater longevity (or fewer years of life lost).  

Similarly, this chart from Quantitative association between body mass index and the risk of cancer: A global Meta-analysis of prospective cohort studies: Obesity and cancer risk shows that leaner people are less likely to get cancer

While there is no need to be bigger or stronger at all costs, we do know that having a higher fat-free mass index (i.e. more muscle or lean mass)aligns with greater longevity (Genton et al., 2013).     

More muscle aligns strongly with more resilience as we age. While lab mice die of old age in a cage, many humans due prematurely to frailty and fragility (e.g. when they fall and break their hip and never leave the hospital system) and do not make it to old age.

Beyond the theory, we don’t have a lot of data on the effects of minimising mTOR in humans. However, it’s likely that optimal levels are not found a the extremes (i.e. similar to BMI, waist:hieght, insulin and IGF-1 as shown in the chart below).

Digging into the data 

After being bemused by the headlines after the mega mouse study Dr Ted Naiman pointed me some discussion around the methodology used in the study.   The experiment started out with 1000 mice on 30 different diets.  However, when you read the supplemental materials you find that 143 of the mice on five of the low protein were euthanised after 10 and 23 weeks due to weight loss, rectal prolapse and failure to thrive.  

Don’t you think it’s absurd that they would exclude data from mice that died due to protein malnutrition in a study that was specifically set up to identify the relationship between protein intake and lifespan? What’s more, it’s not mentioned in the main paper at all.

Edit: After publishing this article, Simpson and Raubenheimer emailed me and noted:

The aim of the experiment was to test the relative effects of total energy intake vs. P:E ratio on lifespan in mice. In order to do that, we decided to map out the viable diet space for mice, something that had not previously been done, then establish the effects of energy density vs. P:E ratio within viable diet space. Here “viable” has an objective and scientifically based meaning: it encompasses the range of diets that support development throughout the life cycle. Death due to non-viable diets vs. death due to accelerated ageing on feasible diets are two very different things and confusing these will lead to nonsense – e.g. concluding that low P shortens lifespan because a cohort of human children restricted to cardboard didn’t survive into old age. 

Our results showed that diets that combined very low energy density with very low P:E ratios (5% energy from P) were non-viable (as indeed they are for humans – as you show in your blog). These diets did not support growth in weanling mice. Not only would including these diets in the analysis be inconsistent with the question our experiment addressed – equivalent to cardboard diets for children – but we were mandated under the terms of our ethics permitting to cull those mice. That’s why they had short lives – they were protein starved and had to be euthanised. Concluding from that that low P:E shortens lives is clearly nonsense. 

Thanks to the wonders of Optical Character Recognition, I imported the data from the supplementary material into Excel and added the data from the mice who died prematurely.  The yellow rows are the mice that died of protein malnutrition and were excluded from the analysis.  

Before doing any whiz-bang data analysis, it’s useful to look at the raw data.  There is no point torturing the data if the numbers that are staring you in the face at first glance invalidate your overarching hypothesis or conclusions.

Importantly, the green row towards the bottom of the table shows that the group of mice that lived the longest on average were consuming 42% protein.  This is directly at odds with the reported headline that high protein diets lead to a shorter life span.

No, the mice on the highest protein diet did not live the longest.  But who the hell in the real world is consuming 60% protein?  

As we saw in the protein intake distribution charts from Optimisers, as well as our analysis of half a million days of MyFitnessPal data (shown below), no free-living human is getting 60% of their calories from protein!    

In the real world of humans, we’re talking about getting people to move up from 12% protein.  It’s absurd that you would use the results in mice eating sixty per cent protein to encourage humans who are eating 13% protein on average to eat even less.

Macros vs intake 

Before getting into the controversial longevity data, let’s quickly look at what the data tells us about the relationships between macros and calorie intake.   

Protein vs ad-lib calories

This first chart shows that, in line with the Protein Leverage Hypothesis and our satiety analysis, mice ate less when they were fed (ad libitum) diets with a higher percentage of protein. 

The diets were divided up into high, medium and low energy density (grey, orange and blue data respectively).  Unsurprisingly, mice were able to eat more of the higher energy density foods. 

Carbohydrates vs ad-lib calories 

This next chart shows that mice ate less when their diet contained more carbohydrates (though this may be interrelated with fat intake, which is more energy-dense).

Fat vs ad-lib calories

Finally, this last chart shows that mice that ate a higher percentage of fat consumed more calories.  We don’t need a complicated regression analysis to see that there is a strong correlation, with the data points sitting on the trend line.  

Take home message (if you’re a mouse)

If you’re a mouse that wants to get lean for beach season, you want less fat and more protein.  However, it’s important to note also that mice are not tiny humans.  As noted in this paper, the lab mice used in the mega mouse study tend to do worse on a high-fat diet than humans (who require carb+fat to really get sick).


Now, where things get more interesting is the lifespan data from the study.  When you look at the charts below of the simple present data, see if you think it aligns with the key conclusion of their paper shown in the screenshot below.   

The Geometric Framework for Nutrition charts below from the paper indicates that (once you remove the mice that died prematurely from protein malnutrition) the longest-lived mice consumed less protein and more carbohydrates.  

But see if you come to the same conclusion when you see the data simply presented the simple Excel plots below.  

Calories vs lifespan

This chart shows energy intake vs average lifespan (the data at the bottom of the chart where the mice died early due to low protein malnutrition).  

This next chart shows the same data as the data from the 143 mice that were euthanised due to protein malnutrition removed.  While there is plenty of scatter, there appears to be an optimal energy intake in the midrange.  Severe energy deprivation is not good, but neither is too much energy.  

Note:  The outlier in the middle at the top with the longest lifespan by a mile is the group of mice on 42% protein! 

Protein vs lifespan 

This next chart shows protein (in grams per day) versus lifespan.  When you account for the 143 dead mice on the five failed low protein diets, mice seem to do best on around 6 g of protein per day, with the best outcome of any of the groups achieved with 7 g protein per day.

The next chart shows the protein intake vs lifespan data with the mice that were euthanised removed.  We can see there is a LOT of scatter and very low correlation.  Once you remove the mice who failed to thrive on a low protein diet, the data indicates there is NO relationship between protein intake and longevity.

With such a weak correlation (even after removing the inconvenient data that didn’t align with their mTOR theories), it’s surprising that they went to print with the headline conclusion that they did. 

Even more befuddling is that they would take out full-page newspaper advertisements to broadcast this advice for humans.  

I’m not sure if “eat less protein” would be a relevant message for steroid injecting bodybuilders trying to get jacked and super shredded.  However, it’s definitely the wrong message to send to the general public (at least, if you are interested in public health) who are already lacking protein because of their heavily processed diet or are following some version of a plant-based or keto diet that shuns protein.  

Carbs and fat vs lifespan 

These next charts show carb and fat vs lifespan with and without the mice that died due to protein malnutrition.  I won’t spend too much time discussing this as there isn’t a lot of meaningful correlation.   

Macro percentages

Finally, these next charts show the data in terms of macronutrient percentages.  I have shown separate trend lines for the low, high, medium energy density data along with all the data together.  

The first plot of protein percentage vs lifespan has the strongest correlation of all of the charts.  If you had to draw a conclusion from this noisy data you might say that we seem to get optimal longevity in mice (with all the data included) with 30 – 40% of energy from protein.  

This also aligns with the observation from our other analysis (e.g. from our analysis of our series of 22 recipe books optimised for nutrient density) that higher protein intake aligns with better nutrient density.  A very low protein diet tends to also be very low in essential nutrients.  But, once you get the nutrients you need, macros become irrelevant.  

Wouldn’t it be nice if we would all stop bickering over macros and plant versus animal-based diets and just focus on getting enough micronutrients?  

Carbohydrate percentage versus lifespan is pretty noisy, with a very low correlation.  

Again, it seems that mice don’t do well on high-fat diets.  

One of the critical conclusions of the mega mice study paper was that a lower carb:protein ratio is better.  When we include all the data, it appears that a very high protein:carb ratio is not better, but less is not better either.   As per normal, optimal is not found at the extremes.

When could “too much protein” be bad?

Can you get too much protein?  


Is any free-living human at risk of getting “excess protein”?

Highly unlikely! 

When it comes to humans (and mice), the first thing to remember is that body composition is king.  Even if you don’t want to be a jacked bodybuilder, having more lean mass and less fat mass is the key to metabolic health and long term survival in the real world.  We can’t extrapolate esoteric hypotheses developed in yeast cells in a petri dish and mice in captivity onto free-living humans.

If you have excess body fat, your insulin and blood glucose will be higher, and your overall risk of dying of any cause is increased!  Body mass index, fasting blood sugars, and insulin levels across the day are closely tied to the amount of fat you need to hold in storage.  

Elevated blood fasting glucose levels above 90 mg/dL or 5.0 mmol/L is strongly correlated with your risk of dying of any cause.  

If you are carrying any significant level of fat (i.e. say greater than 15% for men and 25% for women) your body will happily pull down stored energy while maintaining lean muscle mass from the protein in your diet (so long as you get enough).  

However, as you run out of body fat to burn, the story changes.  Once your body fat levels reach healthy levels and your fasting blood glucose is below 90 mg/dL or 5.0 mmol/L your and you continue to eat lean protein your body will have to work harder to convert the energy from the protein to ATP for use in your body.  

Once you reach optimal body composition and fasting blood sugars, you definitely should add back in some easily accessible energy from fat and/or carbs to make sure your body doesn’t have to work too hard to get the energy it needs to operate from protein.  

But until that point, you probably should lay off the fat and/or carbs.

Should you be worried about “too much protein”?

If you’re confused by all this “research” and data, the table below shows some simple markers.  

  • If you’re above this cut-off level then you likely need to increase the percentage of calories from protein in your diet by decreasing easily accessible energy from carbs and/or fat.
  • If you are one of the dwindling number of people who are living below this cut off level, then you should ensure you fuel your body with more energy from carbs and/or fat, which will cause a reduction in % protein.  
Criteria Cut off 
wasit:height ratio 0.5
Fasting blood glucose 90 mg/dL or 5.0 mmol/L
Body fat 15% for men 
25% for women


So long as you are carrying more body fat than is healthy, you likely need to focus on a higher percentage of energy from protein as well as nutrient density to avoid diabetes, cancer and all the other metabolic diseases related to obesity.  

Once you achieve optimal body fat and fasting blood sugar levels, you are free to add back some easy energy from fat and carbs.  

You should eat like the animals and follow your appetite for protein.  However, be aware that humans tend to reverse engineer the protein leverage hypothesis to manipulate our food environment to maximise energy intake and profit.  


12 thoughts on “Eat Like the Animals: The “Science-Based” Diet Book about the Protein Leverage Hypothesis”

  1. I think there’s a thumb on the scale for carbs and dead mice: the c57bl/6 mouse, according to The Jackson Laboratory is ‘susceptible to diet-induced obesity, type 2 diabetes, and atherosclerosis. The study demonstrated that the c57bl/6 mouse performed as expected.

    • seemed like a well-designed study if you wanted to prove that carbs were better than fat. they had to exclude the mice that died from protein malnutrition to paint protein as bad.

      • I believe the ‘plot twist’ comes from reliance on standard blood test information. Many people eating LCHF diets have bad cholesterol results, per blood tests. I believe the issue is the blood test standards were created based on people eating the Standard American Diet, and that there should be a difference in the interpretation of results based on diet. For instance, my total cholesterol numbers went up to very high levels as I adopted the LCHF diet. But my triglycerides:HDL ratio is very good at less than 1. However, many doctors stick to the traditional standards and try to push for statins based on another group of eaters. I’m in the process of gathering information for my PCP so she will understand my numbers don’t matter. I even had a Cardiac Calcium Exam, with a score of 1! Yet she’s still trying to push me to start statins. Yes, I know I could switch doctors, but in my area, it is very unlikely to find a doctor who knows this way of thinking, unless it’s a naturopathic practice, not covered by my insurance.

  2. Interesting that you do not mention autophagy Marty. Ron Rosedale pointed out that recycling your proteins decreases the need for dietary protein intake.

    In our continuing n=2 experiment, my wife and I have certainly noticed we get full very quickly on a small amount of dietary fat and proteins while being on a permanent 12 hours intermittent fasting regime.

    • autophagy is great, but the reality is we know very little about the minimum effective dose in humans. is it 12 hours? 2 day? 7 days? 40 days? what we do know though is that getting to optimal body composition, in terms of body fat, waist to height ratio, BMI etc are strongly correlated with longevity and decreased risk of all-cause mortality. see FAQ#19 and 56 in the DDF article here for more details.

  3. Marty, I don’t quite get your derogatory comment about the fear of mTor.

    Ron Rosedale has shown quite convincingly that mTor is the epigenetic energy sensor that flips the metabolism between the growth mode and the repair mode. Healthy longevity can only be achieved through cellular regeneration and repair and that is only possible by keeping mTor low.

  4. I am a 69 years old female and have been eating a strict carnivore diet since Oct of 2015. I did well on losing fat. Recently though I have gained about 25 pounds over the last year or so. I know I am eating when I’m not particularly hungry and I imagine that is what is causing the weight gain. I also love butter so maybe overdoing the fat. Any thoughts or suggestions on my getting back on track. Eating animal products only has made my compulsive overeating disappear and I have zero cravings so I don’t want to change my way of eating. Thanks!

    • Many people do well on a carnivore diet because they get plenty of bioavailable protein which helps with satiety. But the protein leverage hypothesis still applies to animal based products. It is possible to overdo added fat. We have a free nutrient dense carnivore food list, a free macro calculator and a nutirent dense meat book that you maybe interested in checking out on our resouce page here – , I have also written some articles on the pros and cons of carnivore that you may be interested in here –

  5. First, it’s outrageous the study data were manipulated and then false results were published in such a high-profile way—the opposite of serious, academic research! I hope people were fired.
    Thank you for revealing the facts around the study and providing the actual insights we should glean from it.

  6. Thank you Marty for your analysis, I was quite relieved to read it. I have been low carb / higher protein for 4 years but always like to keep an open mind so this book raised many doubts, it seemed legitimate. But when research “omits” data and “jumps to conclusions” I am happy to be a sceptic until the science proves otherwise

Comments are closed.