Academic ResearchHeart DiseaseOther Diets & BooksWomen's Health

Can Atkins diet raise heart attack risk for women?

So screamed the headlines on Wednesday 27th June 2012. This was the full article in the BMJ, which was behind the headlines.

Another irresponsible article, more misleading headlines – both the article and headlines ignorant about nutrition and the composition of food. There seem to be one of these a day at the moment and we independent bloggers have better things to do than correct ‘researchers’ who should know better. However, find the time we must because the health and weight of the population depends on us going back to eating what we evolved to eat – the meat, eggs and dairy products from grass living animals; fish from natural waters; vegetables, salads, nuts, seeds and local seasonal fruits – as found in our own natural environment. What we ate before we had an obesity epidemic.

Every time an article like this attacks real food, we take another step back from our return to eating what we should eat. Every time an article like this praises carbohydrates – the macro nutrient that has only been in our food chain, in any great quantity, for the blink of an eye in terms of evolution, we take another step towards staying with the appalling “base your meals on starchy foods” dietary advice that has made us fat and sick within three-four decades. This is why we need to spend our precious, unpaid time exposing the bad science in these articles…

The study

Let’s start with the study facts: In 1991-92, 96,000 women, in the Uppsala healthcare region of Sweden, were randomly selected from four age groups: 30-34, 35-39, 40-44 and 45-49. They were invited to fill in a dietary and lifestyle questionnaire asking questions about: smoking; alcohol; exercise; activity; medical diagnosis of hypertension; height; weight and food consumption.

The article reports “women recorded their frequency and quantities of consumption of about 80 food items and beverages, focusing on the six month period before their enrollment in the study.” (I don’t know about you, but I can’t remember what I ate this time last week, let alone over the past six months)! You can tell that the participants also had poor recall of what they ate because Table 2 tells us that the mean calorie intake recorded was 1,560 – about three quarters of what would be expected.

49,261 Swedish women returned a questionnaire and these were whittled down to 43,396, excluding women for various reasons – “energy intake outside the extreme”, missing data and so on.

The results are presented on p3 of the article as: “Overall, the 43 396 women were followed up for an average of about 15.7 years and generated a total of 680 818 person years, with 1270 incident cardiovascular events (703 ischaemic heart disease, 294 ischaemic stroke, 70 haemorrhagic stroke, 121 subarachnoid haemorrhage, and 82 peripheral arterial disease).”

The first point to make – any overall risk is tiny!

Table 3 (p10) of the article gives the numbers of women diagnosed with total cardiovascular events and the breakdown for  ischaemic heart disease, ischaemic stroke, haemorrhagic stroke, subarachnoid haemorrhage and peripheral arterial disease by low carbohydrate-high protein (LCHP) score category. We’ll look at how useless this scoring system is next, but let’s just go with it for now…

Table 3 shows that the overall incident number for 43,396 women over 680,745 study years (women times average follow up period) was 1,270. Not deaths, please note, but medically diagnosed cardiovascular disease. This puts the overall incident rate for the study at 0.19%. This means that all the women in the study had a 99.81% chance of NOT suffering any cardiovascular events whatsoever during each year of the long study.

Table 3 claims that the 8,395 women with a LCHP score of ≥ 16 on their scoring system had a 0.23% incident rate during their 131,262 ‘woman years’. This terribly risky group, therefore, had a 99.77% chance of NOT suffering any cardiovascular events whatsoever during each year of the 15-16 year study.

The 8,343 women with a LCHP score of ≤ 6 on their scoring system had a 0.14% incident rate during their 130,965 ‘woman years’. This terribly NON risky group, therefore, had a 99.86% chance of NOT suffering any cardiovascular events whatsoever during each year of the long study.

Incidence Chance of no CVD p.a.
All Women 0.19% 99.81%
LCHP score ≥ 16 0.23% 99.77%
LCHP score ≤ 6 0.14% 99.86%


The overall conclusion of the study in the abstract was: “Low carbohydrate-high protein diets, used on a regular basis and without consideration of the nature of carbohydrates or the source of proteins, are associated with increased risk of cardiovascular disease.”

I don’t know about you, but that minute differential is not going to keep me awake at night – despite being a woman in that age group eating a low carb diet!

Point 2 – the scoring system is absurd

P2 of the BMJ article explains the scoring system that they made up. From the original (one-off, remember) dietary questionnaires, this is what they did – in their own words: “We estimated the energy adjusted intakes of protein and carbohydrates for each woman”. Each woman was then “assigned a score from 1 (very low protein intake) to 10 (very high protein intake) … and an inverse score from 1 (very high carbohydrate intake) to 10 (very low carbohydrate intake). We studied the scores for high protein and low carbohydrate intake both separately and after adding them together to create a composite low carbohydrate-high protein score (ranging from 2 to 20).”

A similar study was published in November 2006, using the same scoring system and a great blog was written on the scoring system here. The blogger, Jacqueline, showed how the scoring system works

Imagine we have 10 individuals with protein consumption that puts each one of them into a different protein decile. So we score these individuals

[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]

Now these same 10 individuals also have the carbohydrate consumption so that their carbohydrate scores are:

[1, 2, 3, 4, 5, 6, 7, 8, 9, 10]

Adding these together, these individuals have LCHP scores:

[2, 4, 6, 8, 10, 12, 14, 16, 18, 20]

But what if we had another 10 individuals with the same protein scores but different carbohydrate consumption –so that they scored respectively [2, 3, 4, 5, 6, 7, 8, 9, 10, 1] ? These individuals’ LCHP scores would be: [3, 5, 7, 9, 11, 13, 15, 17, 19, 11]. In the second group someone in the highest decile of protein consumption had a final LCHP of 11 because their carbohydrate consumption was also high. What about the other person with an LCHP of 11 in that group? They had middling consumption of each.

Or, as a member of The International Network of Cholesterol Skeptics , Bjorn Hammarskjold, put it:

“Low protein is 1, high protein is 10, low carbohydrates are 10, high carbohydrates is 1.

So to have sum score of 11 it can be low protein (1) and low carbs (10) or high protein (10) and low carbs (1) or any sum between: 1+10, 2+9, 3+8, 4+7, 5+6, 6+5,7+4, 8+3, 9+2 or 10+1, all give 11 as score.”

And what would the ‘researcher’s try to conclude from this?!

Point 3 – Atkins is not low carb high protein

Atkins is low carb, whatever protein naturally occurs in food and high fat. Dukan is low carb high protein.

The headlines should have been about Dukan if they wanted to demonise any well known diet. However, no real food low carb diet should have been demonised and here is why…

Point 4 – the ‘researchers’ do not seem to know the nutritional composition of food

There are three macro nutrients. We know them as carbohydrates, fat and protein. All food has at least one macro nutrient, most have two and some have three.

The only foods on this planet with no protein are sucrose (table sugar) and oils (sunflower oil,olive oil etc). Every other food contains protein (lettuce, apples, bread – everything). Sucrose is 100% carbohydrate – no fat or protein and oils are 100% fat – no carbohydrate or protein.

Most foods are predominantly carbohydrate proteins (with no fat or little fat) or fat proteins (with no carbohydrate or little carbohydrate). The carb proteins come from the ground and trees and will fall under plant proteins in this article. Fruits, vegetables, grains, pulses, starchy foods are in this category. The fat proteins come from ‘things with faces’ and will fall under animal proteins in this article. Meat, fish, eggs and dairy products are in this category.

Nuts, seeds and whole milk are unusual foods in that they are high in all three macro nutrients – they have carbohydrate, fat and protein in good measure and are therefore not mainly a carb protein or a fat protein. I have no idea what these Swedish ‘researchers’ would make of these highly nutritious foods.

So protein is in everything other than sugar (nutritionally useless) and oils (nutritionally fairly useless – a couple of vitamins, no minerals). This article assumes that the higher the carbohydrate intake and the lower the protein intake, the lower risk of CVD the woman shall have. Hence our Swedish women should have eaten nothing but sucrose for 15.7 years and they would have had 100% carb intake and 0% protein intake (and 0% fat intake – presumably also marvellous). They would have had no vitamins or minerals, no protein, let alone complete protein and no fat, let alone essential fats. They would likely have died on a diet of nothing but sucrose years before the end of the study (does anyone know how quickly humans would die on sucrose alone?) but at least they wouldn’t have developed CVD! Seriously – the implications of this study really are this absurd.

Food is not just macro nutrients. There are micro nutrients that are also vital for life: 13 vitamins and approximately 16 minerals to be precise (there is more debate on minerals considered necessary for health, but the 13 vitamins – A,  eight B vitamins, C, D, E and K – are not debated). In the table below, we can see the macro and micro nutrient composition of two high carb foods (flour and sugar) and two zero carb foods (liver and sardines) – all data per 100g of product. The former two are the single most consumed ingredients in the British and American diets (that’s why we’re fat and sick). The latter two are two of the most nutritious food available on earth:

Sugar Flour Liver Sardines
Calories per 100g 387 364 116 208
Carbohydrate (g) 100 76 0 0
Protein (g) 0 10 17 25
Fat (g) 0 1 5 11
Vitamins USA RDA (AI)*
A (IU) 3000IU 0 0 11,077 108
B1 (Thiamin) (mg) 1.2mg 0 0.1 0.3 0.1
B2 (Riboflavin) (mg) 1.3mg 0 0 1.8 0.2
B3 (Niacin) (mg) 16mg 0 1.3 9.7 5.2
B5 (Pantothenic Acid) (mg) 5mg (AI) 0 0.4 6.2 0.6
B6 (mg) 1.7 0 0 0.9 0.2
Folic Acid (Folate) (μg/mcg) 400μg 0 26 588 12
B12 (μg/mcg) 2.4 μg 0 0 16.6 8.9
C (mg) 90mg 0 0 17.9 0
D (IU) 400IU (AI) 0 0 0 272
E (mg) 15mg 0 0.1 0.7 2
K (μg/mcg) 120μg (AI) 0 0.3 0 2.6
Minerals (M)
Calcium (mg) 1000-1200mg (AI) 1 15 8 382
Magnesium (mg) 420mg 0 22 19 39
Phosphorus (mg) 700mg 0 108 297 490
Minerals (T)
Copper (mg) 0.9mg 0 0.1 0.5 0.2
Iron (mg) 18mg 0 1.2 9.0 2.9
Manganese (mg) 2.3mg (AI) 0 0.7 0.3 0.1
Selenium (mcg) 55 μg 0.6 33.9 54.6 52.7
Zinc (mg) 11mg 0 0.7 2.7 1.3

(Where macro nutrients don’t add to 100g, the remainder is largely water and then ash/minerals in small part)

* RDA = Recommended Daily/Dietary Allowance. AI = Adequate Intake.

This study would place the above foods in this order of CVD risk:

– Sardines worst (highest protein, lowest carb)

– Liver next worst (second highest protein, equal lowest carb)

– Flour second best (second lowest protein, second highest carb)

– Sugar best (lowest protein, highest carb).

Use the nutritional facts to judge for yourself which you think are healthiest.

And finally

Table 1 of the study is actually interesting. Table 1 looks at “non-nutritional variables”.  Table 1 tells us the following:

AGE: Women in the 35-39 age group had twice the incidence of cardiovascular events as women in the lowest age group. Women in the 40-44 age group had nearly four times the incidence of cardiovascular events as women in the lowest age group. Women in the 45-49 age group had nearly eight times the incidence of cardiovascular events as women in the lowest age group. Data below:

Age Women CVD Incidence
30-34 10,414 84 0.81%
35-39 11,145 178 1.60% 2.0
40-44 11,449 350 3.06% 3.8
45-49 10,388 656 6.31% 7.8
TOTAL 43,396 1,268


The headline could have been “The older women get, the more incidence of cardiovascular events they experience.”

EDUCATION: Women with fewer than 10 years education had nearly three times the incidence of cardiovascular events as women with more than 13 years education. Women with 11-13 years education had less than half the incidence of cardiovascular events as women in the lowest education group. Data below:

Education Women CVD Incidence
≤ 10 yrs 12,927 630 4.87% 2.7
11-13 yrs 16,848 395 2.34% 1.3
>13 yrs 13,621 245 1.80%
TOTAL 43,396 1,270


The headline could have been “The longer the time women have spent in education, the lower the incidence of cardiovascular events they experience.”

HEIGHT: Here’s an interesting one. The shortest group of women in the study had nearly three times the incidence of cardiovascular events as women in the tallest group.

Height Women CVD Incidence
≤ 160 5,394 276 5.12% 2.8
161-165 12,260 392 3.20% 1.7
166-170 13,911 383 2.75% 1.5
>170 11,831 219 1.85%
TOTAL 43,396 1,270


The headline could have been “Shorter women have more incidence of heart disease.” (Is this because the taller women ate more protein?!)

BMI: Another interesting one. Those with normal BMI have the lowest incidence of cardiovascular events. Underweight woman had almost twice the incidence of cardiovascular events as normal weight women. Overweight women experienced less incidence of cardiovascular events than underweight women, but still 1.6 times that of normal weight women. Obese women experienced nearly three times the cardiovascular events as normal weight women.

BMI Women CVD Incidence
<18.5 752 34 4.52% 1.9
18.5-24.9 30,628 721 2.35%
25-29.9 9,532 357 3.75% 1.6
>30 2,484 158 6.36% 2.7
TOTAL 43,396 1,270


The headline could have been “Normal weight women have less incidence of cardiovascular events”

SMOKING: Not surprisingly, smokers have nearly three times the incidence of cardiovascular events as non smokers. Giving up smoking helps but you’re better off not ever having smoked.

Smoking Women CVD Incidence
Never 17,901 330 1.84%
ex smoker 12,826 313 2.44% 1.3
current smoker 12,669 627 4.95% 2.7
TOTAL 43,396 1,270


The headline could have been “Smokers experience more incidence of cardiovascular events.”

EXERCISE: The lower the activity group, the greater the incidence of cardiovascular events.

Activity Women CVD Incidence
1 (low) 1,797 90 5.01% 2.8
2 4,643 170 3.66% 2.0
3 25,878 796 3.08% 1.7
4 7,405 148 2.00% 1.1
5 3,673 66 1.80%
TOTAL 43,396 1,270


The headline could have been “Less active women experience greater incidence of cardiovascular events.”

None of these headlines in this “And finally” would have had the global media impact of yet another pop at a low carb diet. Not only has the bad science given us a bad headline, it has missed giving us some other far more statistically significant information, with some very useful findings – albeit to corroborate things that we know anyway “smoking is bad”, “activity is good”.

18 thoughts on “Can Atkins diet raise heart attack risk for women?

  • Pingback: More On Dietary Studies in the News…. – Low Carb Pediatrician

  • Pingback: Thoughts on a plant based diet?

  • re. my above comment, I should have written: So you chance of not having an event “during the long study” is 97%, not 99.81%.

  • Zoe, I think your calculations are awry. You write:
    “Table 3 shows that the overall incident number for 43,396 women over 680,745 study years (women times average follow up period) was 1,270. Not deaths, please note, but medically diagnosed cardiovascular disease. This puts the overall incident rate for the women in the study at 0.19%. This means that all the women in the study had a 99.81% chance of NOT suffering any cardiovascular events whatsoever during the long study.”

    If 1,270 of 43,396 women had an cardio-vascular event over a 15.7 year period, that means there was a 3% chance of having an event over the period of study, NOT 0.19%. So you chance of not having an event “during the long study” is 97%, not 0.19%. You’re mixing two things up: the risk of having a CV event PER STUDY YEAR (0.19%) and the risk of having an event over the duration of the study (3%). That’s statistics 101

  • I understand why you disregard this study, and you make a convincing argument, pointing out all the reasons why it shouldn’t scare people away from the low carb diet.

    But you make one critical mistake that weakens your argument. You don’t consider the archaeological evidence for our ancient diet. Many others Paleo/low carb bloggers make the same mistake, and as a Paleo/low carb dieting archaeologist, this really frustrates me.

    You state quite clearly in the introduction we should be “going back to eating what we evolved to eat – the meat, eggs and dairy products from grass living animals”. Dairy products have been in the human diet for at MOST 9,000 years. (See: Earliest date for milk use in the Near East and southeastern Europe linked to cattle herding;

    You then turn around and say “carbohydrates – the macro nutrient that has only been in our food chain, in any great quantity, for the blink of an eye in terms of evolution”. But people were gathering and eating grains for at least 12,000 years, and possibly as long as 22,000 years. (See: Processing of wild cereal grains in the Upper Palaeolithic revealed by starch grain analysis;

    That’s several thousand years longer than people were eating dairy products, but you seem to not know, or not be troubled by this fact. Many other Paleo supporters make this same error.

    You can eat however you want, and disregard as many studies as you want. But you cannot claim that we evolved to eat dairy in 9,000 years, but not to eat grains after 12,000+ years. My problem is not with your opinion or your evidence. It’s with the way you use that evolution as the basis for your argument, when the archaeological evidence clearly refutes you.

    • Hi Hazel – fair point – I should have been more accurate and said meat and eggs. I’m so used to saying “meat, eggs and dairy from grass grazing animals” in terms of what is good to eat today (for those OK with dairy, which is by no means everyone) that the dairy must have slipped in to the first bit.

      22,000 years is still the blink of an eye in evolution terms – if you accept we go back to Australopithecus Lucy c. 3.5 million years ago and many trace us back even further.

      So hopefully we can agree on meat and eggs!
      Very best wishes – Zoe

  • This is hands down the worst published study I have ever seen, but a casual read by a non-scientist might infer that it is all scientific and statistical ‘n’ stuff. As I read beyond the headlines, I kept asking myself, “what were they thinking????” They “corrected” for satfat and also for non-satfat in their model, but didn’t include a table showing how that ended up. Why would they do this and then not show any results?
    Also, the food questionnaire was even worse than you have shown here. Participants were asked AT THE VERY BEGINNING what they ate. Like you, I have trouble remembering what I ate 6 months ago, but 10-20 years ago, I ate completely differently.

  • THANK YOU ZOE!!!! I was worried about the study, particularly as I think that I actually eat less carbs than you recommend. What these studies don’t realise is that if I eat ‘normal’ amounts of carbs, my BGs go up and I am at extreme risk of….cardio vascular disease. If I eat normal carbs, my weight goes up and I am then at risk of…..wait for it, I know this one….oh yes, cardio vascular disease.
    I’d be interested to know if The Sun ran this story, since they ran one a few weeks back highlighting someone who HAD to get all his energy from fat, not carbs. I hope they warned him by letting him know he would soon be reading a story which would inform him he was due to have a ‘slight’ heart attack risk. Or do I suppose he was completely forgotten about?

  • Instead of “controlling for” all the “confounding” variables, I’d like to take their data and see how strongly these factors are correlated with diet. For instance:

    Are people eating a low-carb diet more likely to exercise?

    Are people eating a high-protein diet more likely to be tall?

    Are people eating a high-fat diet more likely to be highly educated?

    Then, of course, you would have to do some actual controlled studies to figure out cause vs. effect vs. correlation.

  • I have lost trust in the medical doctors. Their reccommedations about food and medication have given me dibetes II, obisity and a stroke, just to mention a few ailments. Finally I “cracked the code” in January and I am now down 20 kg since last October. Two weeks ago I also read Owen Fonorow’s book “Practicing Medicine without a Licence?” about the discoveries of two times Nobel Prize winner Linus Pauling. I encourage you to read it and find out for yourself Who has your best interest in mind. After starting the suggested regime, suddenly my rotten body started to work after only three days! Beforehand I had dumped Lipitor and a blod thinning agent, drastically reducing my insulin use also the last 5 months. Today my bloodpressure is starting to look better also, so this weekend I will start reducing my medication for hypertension :-)
    Best of luck, regards Bjørn in Norway.

  • I can see it now! “The All-New Sucrose Diet! Backed by scientific research, eat as much sugar as you possibly can! No weighing or calorie counting! No points! Just pack in that sugar! Yes, you CAN eat as much chocolate as you like, as much cake and pastry – no limits!The diet the world’s been waiting for!” Isn’t that what they’re saying? Perhaps they should do another experiment whereby the participants really DO eat nothing but processed carbohydrate foods for 15 years. I suspect, though, that they would end up being sued!!

    Zoe, your shredding of that totally pathetic “study” was wonderful. Basically, we all stand more chance of being run over by a bus than we do of acquiring CVD from Atkins-ing or Dukan-ing.

  • There’s a lot of money tied up in low fat, high carb, meat, eggs and butter will kill you dogma. A huge industry of low fat, high sugar and starch foods that are being marketed to us all as ‘healthy’. Not to mention the huge diet industry and the drug industry thriving on the back of the dietary misinformation. This industry does not want the truth to come out about the benefits of a low carbohydrate diet as it would destroy their market. No wonder there are so many studies designed and manipulated to make low carb look dangerous. Sweden is a prime target for these studies as the high fat, low carb diet has become a very popular movement there and must be raising major alarm bells with the low fat, high carb dependant industries.

    This whole scenario is the most astounding real life example of the emperor’s new clothes and nutritionists like yourself Zoe, are the child crying out ‘but he is wearing nothing at all(and actually, he is looking very fat and sick)!’

  • Hi Zoe,
    I knew you would take this latest story apart, just like the meat study back in March, you dissected so admirably (pardon the pun).
    You know the old saying “Too many cooks spoil the broth”. That’s what is happening in the field of dietary research. So many researchers around the world are carrying out these observational studies, which rely on the answers to questionnaires. Like you say, these are notoriously unreliable, as anybody knows if they have filled in a survey form of any kind. You can’t always remember the answers, or you’re too busy, so you end up ticking any old box.
    Trouble is the media jump on stories like this with attention grabbing headlines and before you know it it’s gospel. These people ought to know better.
    The result is confusion. How will we ever sort out the obesity epidemic until they are all working together and doing proper clinical studies?
    No wonder the public are confused and they end up ignoring even the good advice too. (Daily Mail & Guardian readers, this means you). We can hardly expect the Government to get the dietary guidelines right either, when they are receiving so much conflicting information. After all, they are not doctors or scientists, they can only go by the information they are given, but when they get the report, they bury the news they don’t like.
    One survey we can do is to fill in the Department for Health’s Food Labelling Consultation. It’s on their website now. I would like to see carbohydrate content listed on labels and the nutrition panels made bigger on packaging. I need a magnifying glass to read some of them. If I can’t read it, I don’t buy it, simple as that. The closing date for this is 6th August. Here’s the link:

  • Thanks for such a detailed and insightful analysis, Zoe.

    In the US, we’ve had to cope with a study, which in every way, showed the superiority of a low-carb diet, except that it was deliberately designed to be so short (4 weeks), that it also showed markers of detoxification, one of which, elevated CRP, were deliberately made into the headlines by the researchers. When one digs into the study data, though, the absolute CRP levels were negligible, so the researchers tortured the data with statistical manipulation and made the issue seem much more important than it was.

    How do we, as the low-carb lay public, insist on nutrition science that is worthy of the name?

  • Of all journals to rely on RR and not address the minute absolute risk (along with the inane scoring system and 1-shot diet journal), BMJ astounds me. I also wonder why the presence/absence of diabetes was not accounted for in the final multivariate analysis.

    I trust you are publishing a response – at least online?

  • This really does make you wonder, who is training these “scientists”?
    for people who say that epidemiology is a pseudoscience, here is more evidence.
    Epidemiology had a huge hit many decades ago with “smoking causes lung cancer” but since then this band has failed to set the charts on fire and it is time their record company pulled the plug.
    Perhaps they should take another look at what they did right on that occasion; saw the link appear repeatedly in case histories BEFORE undertaking the epidemiology; found a VERY strong correlation, TESTED it in the laboratory, discovered a MECHANISM, tested the effect of smoking-cessation INTERVENTIONS etc.
    There were giants on the earth in those days.

  • Fantastic Zoe, as always you have have debunked these silly “studies” that come out in the press. I don’t understand how they can be so blind to the truth.

  • Very good article.

    I couldn’t actually believe what I was reading with regard to their scoring system. I assumed I was just a bit tired, because, obviously, I thought, it couldn’t actually be *that* useless. But no, it really is that useless.

    It’s as if the scoring system has been designed to be as ambiguous and compromised as possible! I genuinely don’t think that a GCSE Science student could get away with such rubbish.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.