Dr Araz Rawshani

About fat, saturated fat, carbohydrates and survival


PURE: how dietary fats, carbohydrates and proteins affect blood lipids, blood pressure, cardiovascular disease and death

If you are interested in diet, cardiovascular disease and public health, PURE is a pleasure. PURE, one of the greatest epidemiological studies of all time, was initiated by Salim Yusuf, one of the most successful researchers of our time. Salim Yusuf’s research, theories and methods have been groundbreaking and robust. It was Yusuf, with colleagues from Oxford, who established the concept of large, simple randomized clinical trials (in which patients were randomized to receive either placebo or medicines), the purpose of which was to be able to make an opinion on cause and effect. Does it help to lower blood pressure? Do statins help? Does blood thinning help if you have atrial fibrillation? These are questions that can only be answered (safely) with randomized studies. In addition to clinical trials, Yusuf has conducted among others INTERHEART, INTERSTROCK and PURE, the latter recently resulted in three reports in The Lancet, which reported how the composition of the food — i.e. the amount of fat, protein and carbohydrate — affects blood lipids, blood pressure, and the risk of heart attack, stroke and death. These studies are extremely important and should be communicated to every citizen in the country, as the results mean that very many (probably the majority) have to review their diets. Join us and we’ll make a quick summary of the background and then discuss the PURE results in detail.

Before PURE

Let’s start with a little background. Most of all deaths are caused by cardiovascular disease, of which coronary heart disease (leading to a heart attack) is the most common variant. The Interheart study showed that 90% of the risk of having a heart attack is explained by simple risk factors that can be influenced. Diet is, according to Yusuf, one of the main risk factors and guidelines (issued by authorities and the scientific community) recommend that food should be low-fat (less than 30% of the energy should come from fat) and saturated fats should account for less than 10% of the energy in the food. This diet is called “The Unified Diet”, because it is an agreement between, among other things, the American Heart Association and the American Cancer Society. The following photo shows the difference in the content of fats, carbohydrates and proteins in The Unified Diet and several other popular diets.

LCHF kost, kostcirkeln, matt, fett, kolhydrat, proteiner

So what’s the problem? The Unified Diet, as the Swedish diet circle follows, recommends eating low-fat foods, which means that you need to increase the amount of carbohydrates in your food. This has been the dietary advice for the last few decades. The question is whether the population has followed these dietary recommendations and the answer seems to be ‘yes’. Below are data from the gigantic population study NHANES, where tens of thousands of American households have been studied since 1970. As you can see, the intake of fat decreased, while carbohydrate intake increased, and proteins low relatively still. This means that the “Unified Diet” took hold in the population.

Intag av fett, kolhydrater, protein i USA trender

Why is low-fat food recommended?

It was considered that low-fat diet would have two advantages:

  1. First of all, low-fat food would lead to not gaining weight, alternatively losing weight.
  2. In addition, low-fat food would lead to a reduction in the risk of cardiovascular disease, since fatty foods lead to high blood lipids that lead to an increased risk of heart attack, stroke and others.

This was the promise, or at least the hope. It should be borne in mind that cardiovascular medicine (cardiovascular diseases) is an area that has been the forefront of evidence-based medicine (evidence-based medicine = how to interpret scientific studies and make recommendations based on them). Then, of course, it remains to be seen whether the recommendations had an effect. Let’s start with obesity; the following picture shows how the prevalence of obesity developed during the same period (1970 onwards) in the US (again data from NHANES):

Förekomst av fetma, övervikt, diabetes, USA

The incidence of obesity doubled in 30 years! If the amount of fat in food had been crucial, we should have seen a reduction in the incidence of obesity, not doubling. Of course, if you are skeptical, you can argue that there may be other factors that drive obesity (such as physical inactivity, which has increased in recent decades) and thus counteract the benefit of eating less fat. But the fact is that a long line — long line — randomized controlled trials show that fatty foods do not get fat. It is today, more or less, generally accepted that the fat content of food does not drive any epidemic of obesity. 8

Then we go to point number 2: cardiovascular disease. How did it go with heart attacks, strokes, etc? Here’s data from the United States again, but the rest of the world (including Sweden) shows the same trends:

Dödligheten i hjärt-kärlsjukdom har sjunkit dramatiskt de senaste 40 åren.

In about 40 years, morbidity and mortality from cardiovascular disease have halved. This applies to the USA, Sweden, Europe and the rest of the Western world (with minor variations). What can we say so far? The dietary advice advocated in recent decades has not narrowed us, but perhaps they have led to a decrease in mortality from cardiovascular disease? Could it be that reduced fat intake explains the dramatic decline in cardiovascular mortality? After all, we have been told that you should not eat fatty foods because it clogs the vessels. After all, butter and varicose grease in the vessels of the heart have the same color.

Before discussing fat specifically, we can look at an interesting study by Capewell and colleagues. 9 These scientists, using fairly advanced statistical models, have calculated how much of the improvement is explained by the use of drugs for cardiovascular disease, surgery, interventions, treats risk factors and so on. After all, it is desirable to know how effective these measures are, since a significant part of the population today uses such medicines, or has received such treatments. Capewell’s studies show the following:

Förklaring till minskad sjuklighet och dödlighet i hjärt-kärlsjukdom.

We see that almost half of the improvement is explained by drugs and other treatments (statins, antihypertensive drugs, blood thinners, coronary artery interventions, etc.). Approximately the same is explained by the fact that the population has fewer and fewer risk factors. Blood fats have dropped, blood pressure has become lower, smoking has become less common. But, at the same time, physical inactivity has increased, as is the prevalence of obesity, and today these two risk factors are the only ones that are not on the mend path in the population. In fact, the steadily increasing obesity is a threat to the improvement we see in cardiovascular disease.

Why have they been warned about saturated fat?

In fact, this recommendation is based on very few studies, one of which is several decades old, 10 and the other was conducted in Finland, which is not an “ordinary” country in this regard because the prevalence of cardiovascular disease is very high. 11 This is extremely rare, i.e. to base such a important recommendation in so few studies. There is usually much higher demands on medical evidence.

So it has been assumed for decades that the more fat the food contains, the higher the blood lipids you get, and high blood lipids lead to cardiovascular disease (coronary heart disease, stroke, peripheral vascular disease, etc.). More specifically, it has been said that the intake of saturated fat leads to an increase in LDL cholesterol (bad cholesterol) and this is the main reason for the increased risk of cardiovascular disease.

What is the problem with this? You’ve seen that LDL cholesterol rises when you eat saturated fat… Answer: LDL cholesterol is not alone on the pitch (there are more blood lipids and they also affect the risk of cardiovascular disease) and in addition, LDL cholesterol is probably not the best markers of cardiovascular disease.

HDL cholesterol is also important, as is the ratio of total cholesterol to HDL cholesterol (this is called “total cholesterol/HDL”), as well as the ratio of apolipoprotein B and apolipoprotein A1 (called “apob/APOA1″). If you’re not familiar with all these blood lipids, you can ignore them for the moment, we’ll explain them in detail. At the moment, you just need to note the following: more and more recent studies (e.g. INTERHEART, INTERStroke, but also studies from the Swedish National Diabetes Register, among many others) show that LDL cholesterol is actually a worse marker than several other (readily available) blood lipids.

In 2010, 2015 and 2016, six major studies were published, which showed that saturated fat did not increase the risk of death and cardiovascular disease, but rather seemed to reduce the risk.

This means two things

  • Mortality from cardiovascular disease has gone down, and it also has blood fats in the population, but this is not explained by reduced intake of fat.
  • Eating less fat (especially eating less saturated fat) does not seem to protect you from cardiovascular disease.

PURE is big

Most of all studies that have served as a basis for dietary advice have been small and/or limited to a particular population. Such studies have several weaknesses. The smaller a study, the higher the risk that the results are only a coincidence. Small studies, as well as studies that are geographically limited, usually do not offer enough variation (especially regarding dietary habits) among the participants. People in the same city tend to eat quite similar food. As a result, the results of small studies, geographically limited studies and studies carried out on ‘special populations’ cannot be applied everywhere. PURE includes approximately 170,000 participants from 5 continents. These participants — as well as the city, country and region in which they live — are mapped extremely carefully. PURE is probably the most impressive piece in medical epidemiology. The results of the dietary studies have now been published and it finds the most detailed mapping of how fat, saturated fat, unsaturated fat, carbohydrates and proteins affect our blood pressure, blood lipids, mortality and cardiovascular disease.

PURE part 1: The relationship between fat, protein and carbohydrate and blood lipids and blood pressure

In total, 125000 people from 18 countries (5 continents) were studied. What these people ate was recorded, and this was done very carefully. The aim was to investigate the relationship between blood lipids and blood pressure and the following nutrients:

  • Total intake of fat (ie, how much fat you ate)
  • Saturated fat – which is considered unhealthy.
  • Monounsaturated fats – which are considered healthy
  • Polyunsaturated fats – which are considered healthy
  • Carbohydrates – which are considered less harmful than saturated fat
  • Protein – which is considered to be neutral Health view
  • Cholesterol in food – which is considered unhealthy.

It was also examined what happens (with blood pressure and blood lipids) if 3% of the energy coming from fat is replaced with carbohydrates.

When you examine blood lipids in PURE, it is more than just LDL cholesterol. These scientists have never really been convinced that LDL cholesterol would be the best marker, and in this case the aim was to get an overall picture of how food affects the entire blood-lipid profile. Therefore, the following blood lipids were looked at: total cholesterol in the blood (total cholesterol), LDL cholesterol (bad cholesterol), HDL cholesterol (good cholesterol), triglycerides, apolipoprotein B (considered a marker of blood lipids that cause atherosclerosis), apolipoprotein A1 (considered to be marker of blood lipids that prevent atherosclerosis). From these, you can actually create two more blood fat markers, and this is done by dividing apolipoprotein B by apolipoprotein A1 (called Apob-to-A1), and dividing total cholesterol by HDL cholesterol (total cholesterol-to-HDL cholesterol).


Let’s start by looking at what people really eat. The picture below shows how much of the energy intake comes from different sources of nutrition in different continents.

As you notice, there are no radical differences in what you eat around the world. It looks fairly similar. Between 20 to 30% of the energy comes from fat; more than 50% comes from carbohydrates and the rest from protein. In Sweden we eat 30% fat, 52% carbs and 18% protein, which is actually what is recommended all the years! This is also the exact opposite of low-carbohydrate diet (LCHF).

The relationship between blood fats/blood pressure and fat, protein and carbohydrates

Of our three nutrients (fat, protein, carbohydrates), fat and carbohydrates had the greatest effect on blood lipids and blood pressure. In fact, proteins did not have a particularly strong relationship with any of the things examined in PURE. Likewise, the food content of cholesterol had no great effect either. This is how the exact data on how fats and carbohydrates affect the most important blood fats (of which apob/APOA1 is most important) and blood pressure are shown. (The scale on the graph has been removed because it is complicated; focus instead on differences between the direction and size of the arrows).

The picture above is a bit messy but it can be melted with a little will. You see five diagrams and each chart shows what happens with different risk markers (total cholesterol, LDL, HDL, etc.) when you increase the intake of a particular nutrient. Green arrows indicate a healthy change. Red arrows indicate an unhealthy change.

Example: See the graph at the top right, it shows how increased intake of carbohydrates affects blood lipids and blood pressure. It is shown that when you eat more carbohydrates, total cholesterol, LDL cholesterol and HDL cholesterol decrease. However, triglycerides rise and so on.

You can, of course, botanize in this figure, but the most important result is that increased intake of carbohydrates leads to higher apob/apoa1 and this is harmful (apob/apoa1 is the strongest risk factor for cardiovascular disease). In fact, these figures show that apob/APOA1 drops the more fat you eat, and this applies to all types of fat. This is a big surprise, because it means that high intake of fat is associated with better blood lipids. The overall interpretation of the above figure is that carbohydrates have a worse effect on blood lipids than fat.

For those who like details:

The more fat you ate, the…

Higher total cholesterol in the blood. Higher LDL cholesterol (which is unhealthy) Higher HDL cholesterol (which is good) Lower triglycerides (which is good) Lower ApoA1 (which is good) Lower ratio between total cholesterol and HDL (which is good) Higher systolic and diastolic blood pressure (which is unhealthy)

It is also surprising that a higher intake of polyunsaturated fats was associated with higher apoB, which is harmful (polyunsaturated fats have always been thought to be particularly useful). Cholesterol in the food had no significant impact on blood fats.

The more carbohydrates you ate, the…

Lower total cholesterol in the blood (which is good) Lower LDL cholesterol in the blood (which is good) Lower HDL cholesterol in the blood (which is bad) Higher triglycerides (which is bad) Higher apob-to-apoa1 (which is very bad) Higher ratio between total cholesterol and HDL (which is bad) Higher systolic blood pressure (bad) Lower diastolic blood pressure (good)

Protein had a weak relationship with blood lipids and no effect on the important blood lipid marker ApoA1. However, blood pressure decreased as protein intake increased.

The next picture shows how apob/apoa1 drops and HDL rises when increasing the intake of fat.

For carbohydrates, the relationship was reversed – i.e. more carbohydrates lead to higher apob/apoa1 (harmful) and lower HDL cholesterol (harmful).

Subsequently, a whole series of clever mathematical experiments investigating what happens when replacing a portion of the food’s saturated fat with either (1) carbohydrates, (2) monounsaturated fats or (3) polyunsaturated fats. What was found was that apob/APOA1 increased (which is harmful) by replacing some saturated fat with polyunsaturated fat or carbohydrates. Monounsaturated fat, on the other hand, had a beneficial effect on apob/APOA1. HDL cholesterol became worse (i.e. lower) regardless of what was attempted to replace saturated fat with.

Summary of PURE Part 1:

  • The higher the intake of fat (including each type of fat), the better the overall effect on blood fats. These results are in agreement with a wide range of other qualitative studies, such as the Women’s Health Initiative (where sharply reduced fat intake had no significant effect on LDL and total cholesterol).
  • The higher the intake of carbohydrates, the worse the overall effect on blood fats. Most blood lipids became worse.
  • Higher fat intake (especially saturated fat) and higher carbohydrate intake was associated with higher blood pressure.
  • High protein intake was associated with lower blood pressure.
  • Replacing saturated fats with carbohydrates has the most harmful impact on blood fats.
  • Replacing saturated fats with unsaturated fats leads to lower LDL cholesterol and lower blood pressure, but, unfortunately, also lower HDL cholesterol and higher triglycerides.
  • LDL cholesterol was no good risk marker for cardiovascular disease: Apob-to-APOA1 was significantly better.
  • The recommendation to reduce the total intake of fat and saturate fat, seems to be wrong.
  • The recommendation to eat less than 10% saturated fat (by energy intake) is also not supported by this study.
  • Polyunsaturated fatty acids (polyunsaturated fat) are not healthier than monounsaturated fatty acids (monounsaturated fat).

PURE Part 2: The relationship between fat, protein and carbohydrate and mortality and cardiovascular disease

The second PURE report describes the relationship between the fat content (total fat, saturated fat, monounsaturated fat, polyunsaturated fat), carbohydrates and protein and the risk of death, heart attack, stroke, cardiovascular death (i.e. death due to heart attack or stroke) and other causes to death. PURE is, again, unique in this respect because previous studies have been limited to people in a country, a region or even a city. This has meant that in previous studies there was little variation in the composition of the diet (people in the same region eat quite similar food). PURE includes people from 18 countries, 5 continents and this study included approximately 135,000 people. This means that this study represents all the nuances of us people and our eating habits. Previous studies have not had this strength.13

Note that in this study the following three types of death are investigated:

Total mortality: this is absolutely all deaths (regardless of cause) Cardiovascular death: death as a result of cardiovascular disease.Non-cardiovascular death: all other causes of death, except cardiovascular disease. This is, in the main, cancer, according to the authors.

Study layout

These are very much the same people who were examined in the previous report. A total of 135,335 people were studied. When they joined the study, their eating habits were mapped and followed up for 7.5 years on average to clarify who died, had a heart attack, stroke and so on. As in the previous report, the intake of fats (including saturated, monounsaturated and polyunsaturated fats), proteins and carbohydrates was recorded. It was calculated how much of the energy intake came from each nutrient.

All participants who had cardiovascular disease (heart attack, stroke, heart failure) were excluded when the study began. This is because you want a ‘clean’ population from the start. As in the previous study, all analyses are adjusted for other important risk factors (gender, age, diabetes, smoking, physical activity, etc.), in order to minimize the relationship between fat/protein/carbohydrate and the risk of disease and death. The following were examined:

What is the relationship between how much fat/protein/carbohydrates you eat and the risk of disease and death? What happens to the risks if you replace 5% of the energy from carbohydrates with fat instead?


Carbohydrates and the risk of cardiovascular disease and death

A clear relationship between carbohydrates and mortality was seen: eating a lot of carbohydrates was associated with increased mortality (total mortality). The most carbohydrate group (77% of energy) had 28% higher mortality than the group that ate the least carbohydrate (46% of energy). Note that mortality only increased when carbohydrates accounted for at least 65% of the energy intake (which was a fact for half of all participants).

There was no relationship between carbohydrates and the risk of a heart attack, stroke or cardiovascular death. However, there was a gradually increased risk of non-cardiovascular death (mainly cancer) with an increased intake of carbohydrates. This means that the increased mortality associated with carbohydrates is probably driven by causes of death other than heart attack and stroke (in this case, probably cancer).

Fats, saturated fat, monounsaturated fat, and polyunsaturated fat

Higher fat intake was associated with lower mortality (total mortality). Those who ate the most fat (35% of energy) had 23% lower risk of death, compared to those who ate the least fat (10% of energy). Note that the most severe decrease in mortality (total mortality) was observed when more than 30% of the energy came from fat. There was no clear correlation between the intake of fat and the risk of heart attack, stroke or cardiovascular death, but there may be a reduction in the risk of cardiovascular death with increased fat intake. All the clearer was that the risk of non-cardiovascular death decreased with increased intake of fat. That means, again, it was likely cancer-related mortality that was related to the diet.


High protein intake was associated with lower mortality and, as before, non-cardiovascular mortality explained the relationship. Vegetable protein was not associated with death, while increased animal protein intake was associated with reduced mortality.

Saturated fat

For different types of fat, a lower mortality rate was observed with a higher intake of saturated fat. Those who ate the most saturated fat had 14% lower risk of death. This was explained by the reduced risk of non-cardiovascular death and stroke. Increased intake of saturated was not associated with an increased risk of cardiovascular disease (heart attack, stroke, cardiovascular death). This is important to note because the main reason for recommending a reduced intake of saturated fat for decades is precisely because it is believed that it increases the risk of cardiovascular disease.

Monounsaturated fat

Similarly, lower mortality was observed with increased intake of monounsaturated fat. The most monounsaturated fat group had almost 20% lower mortality than those who ate at least monounsaturated fat. There was a tendency to lower the risk of stroke and non-cardiovascular death with increased intake of monounsaturated fat. There was no connection between monounsaturated fat and heart attack or cardiovascular death.

Polyunsaturated fat

With increased intake of polyunsaturated fat, lower mortality was also observed. Those who ate the most polyunsaturated fat had a 20% lower risk of death than those who ate the least polyunsaturated fat. Again, this connection was not seen for cardiovascular death, heart attack or stroke, but they were seen for non-cardiovascular death.

When looking at the links between different types of fat and mortality, it appears that polyunsaturated fats produce the greatest health benefit (since high levels of polyunsaturated fat produced greater risk reduction than high intake of other fats).

Effect of replacing 5% energy from carbohydrates with fat

Replacing 5% of the energy from carbohydrates with polyunsaturated fats was associated with 11% lower risk of death. For other nutrients, there was no impact on the risk of death. The risk of stroke became 20% lower (in the statistical models) when replacing carbohydrates with saturated fats. The risk of non-cardiovascular death decreased by 16% if replacing carbohydrates with polyunsaturated fats.

Conclusions from PURE part 2

  • High intake of carbohydrates (more than 60% of the energy, which half of all people had) is associated with increased mortality and increased non-cardiovascular mortality. However, there was no reduction in mortality with low-carbohydrate diet.
  • High fat intake is associated with lower mortality and lower risk of non-cardiovascular mortality. cardiovascular mortality, as well as stroke. Higher intake of each type of fat (saturated, unsaturated, polyunsaturated) is associated with lower mortality and lower risk of non-cardiovascular death.
  • It is important to point out that the fat of food was not associated with an increased risk of cardiovascular disease (heart attack, stroke, cardiovascular death). Limiting the intake of saturated fat to a maximum of 10% does not seem to bring any health benefits and going below 7% can be dangerous.
  • Overall, the study questions the current dietary advice and argues that a higher intake of fat can be healthy for the population. This is the opposite of what national and international guidelines have recommended for decades.

Remember, this is an epidemiological study, which means that we can never be sure that these relationships are causative agents, but the combined science nevertheless suggests that most of the results of this study have such a connection.14

What have we learned from PURE really?

Eating a lot of carbohydrates seems harmful, especially when you allow carbohydrates to make up 65% or more of the energy. High intake of carbohydrates is associated with increased mortality. However, a high intake of fat appears to be healthy. Today’s dietary advice must undeniably be reconsidered and authorities should inform the population about the benefits of reducing carbohydrates in favor of fat, including saturated fat.

5/5 (1 Review)