Progress in science revolves around disproving and discarding theories that are no longer useful.
Scientists and engineers continually test their ideas to increase their chance of being right more often.
While correlation doesn’t equal causation, if things you believe to be causal are found to be completely uncorrelated you should review your beliefs to see if they are still useful.
Nutrition science is still relatively young. There is still a lot to learn.
- Vitamin C was only discovered in the mid-1700s by James Lind in the first documented randomised controlled trial.  
- More than 100 years later in the mid-1800s that we first used vitamin C to treat scurvy.
- In 1890, Christiaan Eijkman found that fowls fed polished rice died sooner from paralysis. Eventually, it was determined the missing compound that was vitamin B1 (thiamine).
- In 1912 biochemist Casimir Funk shortened the term “vital amine” to coin a new word “vitamin”.
- Other vitamins were only discovered around a century ago and there is still plenty of disagreement about how much we need of each of them, what they all do and how they interact.
It’s hard for nutritional research to move quickly because it’s unethical to do A/B testing on humans. So we often rely on nutritional epidemiology which relies on comparing the health and habits of different populations (e.g. Ancel Keys’ Seven Countries study).
In a free-living environment, we tend to base our food choices on:
- what’s available,
- what we can afford,
- what tastes good, and
- what we feel like at the time.
We’re also really bad when it comes to recording what we eat, so it’s hard to study the effects of nutritional interventions. Hence, rather than science or biology, our nutritional choices are often influenced by religious belief, ethical convictions and financial conflicts of interest.
But we have been exposed to some powerful nutritional experiments over the past 50 years that have dramatically changed the way we eat. Rather than looking at different populations, we now have a significant amount of longitudinal data to evaluate how our nutritional beliefs are working out for us.
In a controlled trial, scientists are obliged to abort an experiment if it appears that an intervention is causing a harmful outcome. This article evaluates commonly held nutritional beliefs to see if they are still useful.
Demonisation of nutrients can lead to imbalances in other areas. You ideally want a system to be as simple as possible so you can invest your limited energy into the things that matter (like getting more of the nutrients you’re lacking).
This post takes no prisoners, so if you have a pet nutritional theory or belief listed below, you might want to stop reading now.
- Dietary cholesterol is bad for you
- Saturated fat is a “bad fat”
- Polyunsaturated and monounsaturated fats are “good fats”
- Fat does not make you fat
- Carbohydrates make you fat
- Calories don’t count
- Salt is bad for you
- Sugar is public enemy #1
- Red meat and eggs are bad for you
- Macronutrient percentages matter
You’re still here.
You were warned.
Let’s do this.
We’re getting fatter
One thing we do know is that we’re getting fatter.
The chart below shows the obesity rates (from US Centres for Disease Control) in the US from 1960 (when only about 13% of us were obese with a BMI > 30) to 2008 (when about 34% of Americans were overweight).
It’s not just about appearance. Most people understand instinctively that obesity is not optimal.
Obesity also increases your chance of having a wide range of metabolic issues that are bad for the individual, your community and the economy.
In this article we test long-term nutritional trends using one hundred years of data against US obesity rates. Don’t worry, if you’re not in the US, this information will still be relevant as your diet and obesity rates are likely following the US.
The food data used in this analysis is taken from the USDA Economic Research Service. I encourage you to download the data and have a play yourself. Let me know in the comments below if you spot any errors in the analysis that require correction.
- Where you see data plotted from 1910 to 2010, it is based on food availability (i.e. all food produced for people to eat).
- Where you see it plotted from 1970 to 2015, it is “loss-adjusted data” which is corrected for wastage and is closer to the actual amount eaten.
Before you get caught up in the limitations and accuracy of the data, keep in mind that no real-world dataset is “perfect”. Any measurement has errors. Higher accuracy is more expensive, which means we end up only looking at a small number of people (e.g. in a metabolic ward in a hospital where you can control their food and movement).
For me, seeing the long-term trends of humans in the wild is much more useful than a few individuals for a couple of days in a metabolic chamber.
While the data has its limitations, it also gives us another viewpoint to test the observations that we can make from other sources.
And, while you might be able to find a lot of things that correlate with obesity, we’ll be focusing on the things that have changed significantly, not the irrelevant minutia.
We’re not trying to demonstrate causality with this data, but rather understand which beliefs do not align with reality. If something does not appear to align with a claim or an accusation then the burden of proof is on the party making the claim to provide the evidence that, beyond a reasonable doubt, their claim is legitimate.
Belief #1: Dietary cholesterol is bad for you
Let’s start with cholesterol as an example of a dietary belief that most people have decided is no longer useful.
This first chart shows the cholesterol available in the food system over the last hundred years overlaid with obesity rates since the 1960s.
Until recently, the recommended limit for cholesterol was 300 mg/day (or 200 mg/day if you had a high risk of heart disease). People in the 60s were concerned that dietary cholesterol was driving high cholesterol in the blood, which they believed caused heart disease. However, we now understand that:
- there are several risk factors for heart disease, and
- changing the amount of cholesterol in the diet does little to change the cholesterol level in your blood. 
Because of the lack of evidence for a direct causal relationship between dietary cholesterol and cardiovascular disease, the Dietary Guidelines Committee recently removed cholesterol as a nutrient of concern. 
Belief #2: Saturated fat is a “bad fat”
Saturated fat (which primarily comes from animal sources) has been blamed for many things. The thinking was that, because saturated fat was solid at room temperature, it would solidify in your arteries and cause heart disease (note: if your arteries are ever at room temperature you have some bigger issues than the amount of saturated fat in your diet).
Saturated fat has been on the decline as a percentage of our total fat intake for a while (see red line in the chart below).
We are now consuming more “good fats”, particularly polyunsaturated fat (which usually come from industrially made “vegetable” oils like soybean and corn).
The chart below shows that both obesity and saturated fat have been trending up for the past 50 years.
However, when you look at saturated fat as a percentage of total fat, you can see that it is trending down in the opposite direction to obesity.
So, it seems that, while saturated fat is not a “free food”, it might not be as toxic as we have been led to believe, at least compared to monounsaturated and polyunsaturated fats.
Belief #3: Monounsaturated fat and polyunsaturated fats are “good fats”
Mainstream advice (e.g. USDA Dietary Guidelines) suggests that monounsaturated and polyunsaturated fats are “good fats” that should be used to replace saturated fat. Polyunsaturated and monounsaturated fats are prominent in plant-based oils (although seed oils still have some saturated fat).
A large amount of our dietary guidance has been driven by people with ethical or religious convictions against animal-based foods. It’s also relevant to note that the US Dietary Guidelines are controlled by the United States Department of Agriculture. So it shouldn’t be a surprise that their guidance favours the products of agriculture (e.g. corn, wheat and soy).
The chart below shows that the production of monounsaturated fat (olive oil and canola oil) has been trending up.
With the decrease in saturated fat, monounsaturated fat has also been increasing as a percentage of total fat.
Meanwhile, polyunsaturated fats (mainly from corn and soy) have also been on a steady upward trend.
Similar to monounsaturated fat, polyunsaturated fat has been on the uptrend as a percentage of fat.
Over the past century, polyunsaturated and monounsaturated fats have both increased by 300 to 400 calories per person per day.
When we look at the growth in fat consumed, we see that the increase in “salad and cooking oils” tracks with obesity while animal-based added fat sources (e.g. butter, dairy and lard) have not changed significantly since 1970.
Based on this analysis I don’t think we can call polyunsaturated and monounsaturated fats “good fats” relative to saturated fat.
Belief #4: Fat does not make you fat
Some people that believe that, because fat does not raise insulin, that it cannot make you fat.
While it would be nice if we could enjoy as much fat as we wanted, the data doesn’t seem to support this belief.
Obesity rates appear to have risen in line with fat intake over the long-term.
While fat doesn’t raise insulin as much in the short term, your pancreas still releases insulin to hold back your body fat in storage while you use up the energy coming in from your diet.
While insulin plays a role in converting the food you eat to energy, you can think of insulin like the brake signal to stop stored energy being released from your liver. Your pancreas will raise your insulin levels while you use up the energy coming in from the food you’ve just eaten. Insulin will be higher if you have more body fat and/or eat more (regardless of the macro split).
For more details on this check out the following articles:
- Does insulin resistance really cause obesity? and
- The Carbohydrate – Insulin Hypothesis vs the Personal Fat Threshold theory of obesity.
I think we need to concede that dietary fat can be converted to body fat.
Belief #5: It’s only carbs that make you fat
This belief is based around the idea that carbs raise insulin more than other macronutrients.
In people with diabetes, we have noted that injecting insulin can drive excess fat storage. So, by extension, we have assumed that we will store more fat if we get more of our energy from carbs which cause a greater insulin response over the short term.
As shown in the chart below:
- Carbohydrate production decreased in the US from 1910 until about 1960.
- Between 1960 and 2000, obesity rose in line with increasing carbohydrate intake.
- From 2000, consumption of carbohydrates has dropped back while obesity has powered on.
When Dr Robert Atkins released his book in the 80s,  it would have looked a lot like carbs were driving obesity. Perhaps it was the popularity of the low carb movement that caused the reverse the trend in wheat and flour in the 90s.
However, the food industry adapted, and the obesity epidemic has powered on.
This divergence in the trend of obesity and both carbohydrate production and consumption should cause a healthy level of scepticism around the idea that carbs are the primary thing that makes us fat.
Belief #6: Salt is bad for you
Although we have successfully eaten less salty foods in line with the government guidance it has not helped to kerb the obesity epidemic.
The recommendation to reduce salt is largely due to the belief that high sodium levels cause hypertension. However, more recently we have realised that the problem is a lack of potassium in our diet rather than excess salt.    
Decreasing sodium may have worsened the obesity epidemic by driving us to eat more to get the salt we need.
Belief #6: Calories don’t matter
Counting calories sucks, so many people like to think that calories don’t count.
Many have great success when they switch to a diet that cuts out processed flours and added oils (e.g. low carb, keto, whole food plant based, paleo, keto, carnivore, South Beach Diet or Whole30) and they believe it’s the carbs/fat that they eliminated that caused their transformation, not the change in calories.
However, our obesity rates track reasonably closely with the increase in calories.
While food production dropped from 1910 to 1960, since 1960 both calories and obesity have been on an upward trend.
I think it’s safe to say that there is some relationship between calories and obesity.
Belief #7: Sugar is public enemy #1
Many people believe that added sugars are the biggest issue with our food system. However, it seems that added sugars and obesity diverged around 2000.
Americans were using 15% less added sugar in 2015 compared to the peak in 1999 when Splenda was released.
High Fructose Corn Syrup (which was cheap and abundant due to the subsidies applied to wheat and corn) replaced unsubsidised cane and beet sugar in the food system.
Sugar has risen a little from the mid-80s (7.5% to 7.8% of energy intake) but has not reached the levels of use that it was in the 1970s before HFCS (14% of calories).
With the creation of food flavour technology and artificial sweeteners, food manufacturers no longer rely on sugar to make their food-like products look and taste however they want them to while getting the energy from cheap flours and added oils. The food industry has moved on to satisfy market demand in new and creative ways.
Added sugars represent only a small proportion of the extra calories being consumed since 1970. The increase in added sugars is much less than “added fats and oils” and “flours and cereals” category and just about the same as the increase in “meat, eggs and nuts”.
So, while I don’t think you should be loading up on nutrient-poor added sugars, I don’t think the case against sugar as the primary driver of obesity is as strong as some people claim it to be.
Belief #8: Red meat and eggs are bad for you
Due to fear of saturated fat or the belief that red meat causes cancer, people have been encouraged to eat white meat instead.
The ‘red meat, eggs & nuts’ food category represents about 20% of energy intake, which is down from 25% in 1970. The overall increase in energy intake, meat, nuts and eggs have increased by between 25 and 50 calories per day.
In line with the recommendation to prioritise white meat, poultry has increased about as much as red meat has decreased. Nuts have increased a little while seafood is stagnant and eggs have decreased.
Given that red meat has been in decline during the growing obesity epidemic, it’s hard to support the claim that red meat or eggs are having any significant impact on the obesity.
Belief #9: Macronutrient percentages matter
The chart below shows protein intake increased with America’s rising affluence after the Great Depression in 1930. The 1977 Dietary Goals for Americans encouraged a reduction in fat intake (particularly saturated fat). Since 1977, there has been a degree of ‘protein dilution’ which aligns with an uptick in obesity rates.
This also aligns with our previous observation from the food diary data that people who consume less protein tend to eat more. While protein contributes to your calorie intake, if a larger proportion of your calorie are from protein it seems you are likely to consume less.
The chart below shows that there was some reduction in % fat and an increase in carbs between 1977 and 1999 in response to the dietary guidelines. However, this has since largely reversed.
The more interesting long-term trend with increasing industrialisation and an increase in more refined food is a trend towards carbs and fat being similar in percentage terms.
In nature, we get a constant swing between carbs and fat to varying degrees depending on our latitude and season. Given the opportunity, we tend to gravitate to foods that fuel our glucose and fat metabolism at the same time.
The image below from Cian Foley’s Don’t Eat for Winter shows how the glycaemic index of foods varies with seasons (in the northern hemisphere). Mother Nature effectively fattens us up with higher GI foods in preparation for the coming winter. It’s as if the presence of foods with a similar mix of fat + carbs tell your body that winter is coming and your set point raises.
In survival terms, our “good behaviours” are rewarded and reinforced. We get a dopamine hit when we eat carbs and a dopamine hit when we consume fat. However, a recent study confirmed that we get a synergistic dopamine hit when we consume foods that have carbohydrates and fats together.  We are much more motivated to consume foods that contain a mixture of carbs+fat.
These carb+fat food combinations are only possible with modern processed food, particularly the addition of added oils (from soy and corn) to processed flours (from wheat and corn).
Rather than having to follow seasonal variations, we can now eat the same highly rewarding food all year round, effectively giving our body the signal that winter is coming and we need to store some extra fat.
As shown in the charts below, China has experienced a similar trend to America. Historically, living on a diet of rice with little fat, they found it hard to overeat. However, over the past 50 years, they have doubled their calorie intake by adding edible oils to their diet.
And sadly, the Chinese have also experienced similar growth in obesity rates.
We recently crunched the numbers on more than half a million days of food logs and found that people tend to eat less when they have more protein and more fibre while they ate more when they consumed foods with a mixture of fat and starch together.
It’s nice when the different datasets and observations align to give the same conclusion. We’ve been able to use this data analysis to refine the satiety calculation in the Nutrient Optimiser to identify foods and meals that maximise satiety as well as nutrient density for people wanting to lose body fat.
For more info on this topic check out the following posts:
- How to optimise your protein, fat and carbohydrate to minimise hunger,
- Don’t eat for winter,
- Optimizing macros for fat loss with less hunger.
So it seems that macronutrients matter to some extent, but it’s not a matter of high/low carb vs high/low fat being better than the other.
Protein dilution appears to play a role, but the trend towards a mixture of cheap and hyperpalatable refined fat and refined seed oils that is made possible by modern processed foods seems to be the overarching factor in our energy intake and obesity.
While correlation doesn’t equal causation, for completeness, I’ve tabulated the correlation coefficient and R2 value between some factors in the table below.
The things in this table have the lowest correlation with growing obesity. It seems sodium, cholesterol and saturated fat (as a percentage of fat intake) have decreased during the growing obesity epidemic.
|sodium (per calorie)||-0.986||0.972|
|cholesterol (per calorie)||-0.948||0.899|
|saturated (% of fat)||-0.892||0.796|
Meanwhile, the table below shows the things that have increased along with obesity. It seems that energy intake has the biggest impact on our obesity, however, there are a number of other parameters that have increased in parallel with the obesity epidemic.
|total energy (calories)||0.957||0.917|
|polyunsaturated fat (g)||0.935||0.874|
|total fat (g)||0.926||0.858|
|total carbs (g)||0.919||0.844|
|monounsaturated fat (g)||0.917||0.841|
|saturated fat (g)||0.873||0.761|
|polyunsaturated (% of fat)||0.857||0.734|
What to do?
Based on this analysis, it appears that the most obvious place to start is to reduce your intake of the nutrient-poor added sugars, added oils and refined flours.
Next time you go shopping, try to reduce the number of items that you drop into your trolley that contain these nutrient-poor cheap subsidised products as an ingredient. It’s better to purchase separate ingredients and cook at home rather than trusting food manufacturers.
Once you get things sorted at home you should start to think about what you eat when you eat out. Restaurant and fast food chains are interested in taste and cost, not primarily your health. The same is true for many pre-prepared meals. Learn to read labels and be discerning like your life depends on it.
As you reduce your intake of processed foods you will need to compensate by raising your intake of the following food categories:
People seem to benefit from a range of diets that appear to be diametrically opposite, including plant-based, high protein, low carb, paleo or keto. It’s likely all of these work because they help you to stay out of the carbs+fat danger zone and avoid highly processed foods.
Some people swear by carb cycling, targeted keto (i.e. carbs around workouts) or even focusing on macronutrient extremes in each meal through the day. Others like to focus on fresh local and seasonal produce. Again, all of these help to keep you out of the appetite-stimulating danger zone that seems to occur when we get greater than 30% carbs and 30% fat in the same meal.
Thanks to Simon Saunders for the title graphic featuring Ted Naiman and myself. Thanks to Cian Foley for his Don’t Eat for Winter insights. Thanks to Lindsay Wilson and Stephan Guyenet for the inspiration to dig into the USDA data. Also to Helen Kendall, Robin Reyes and Eveylyn Carbsane for review comments. And a big shout out to Ted Naiman for constant inspiration and letting me use his infographics in most of my articles.