On September 2nd 2014 a story broke in the field of diet and nutrition, but it barely made it across the pond. The New York Times headline was “A call for a low carb diet that embraces fat.” The Huffington Post covered the story, but didn’t devote a whole article to it. Instead they combined the findings of this study with another to report “What can we really learn from that low-fat vs. low-carb study?”
The BBC ignored it. The Mail did an on-line only piece “Time to throw out the bread! A low-carb diet IS the most effective way to lose weight – and it cuts the risk of heart disease, too.” Great headline, but you wouldn’t have come across the article by chance. I had to do a very specific search on the mail site to see if they had covered it. Similarly the Guardian had an on-line story “Low-Carbohydrate Diet Protects Against Heart Disease Study Shows.”
This is the link to the original article “Effects of low-carbohydrate and low-fat diets.” Dr Michael Eades has very kindly dropped it in his box for you to be able to see the full article.
Start, as usual, with any conflicts of interest. The study was funded by the National Institutes of Health, which is very interesting in itself. It’s the first study I’ve seen funded by a public body to review the effect of low-carb diets, let alone as compared to low-fat diets – the latter typifying public health advice. There is a full conflict of interest form, covering all authors involved in the study here. It is unusually and refreshingly completely void of any conflicts whatsoever. Oh to find a drug or food item study so genuinely independent.
The study was simple in design and meticulous in execution. Many of the medical site reviews of the study detailed and/or complimented the care that had been taken to make the trial as controlled as possible.
The study involved 148 men and women without cardiovascular disease or diabetes. The men and women were aged 22 to 75 years and had a BMI of between 30 and 45, so that they had sufficient weight to lose over the study period. The participants were recruited from the general public through open adverts. This generated more response from females than males, resulting in 89% of the subjects being female. The study was undertaken in New Orleans, Louisiana. 51% of the participants were black, 45% were white and the remainder were Asian and Hispanic.
73 of the participants were randomly allocated to the low-fat diet and 75 were randomly allocated to the low-carb diet. Table 1 in the paper shows the baseline characteristics of the participants. This is where we can check that the randomisation ‘has worked’ and we are equally like to have similar numbers of females in each diet group, similar numbers of different ethnic origins, similar starting weights and so on. The low-fat group were a couple of years older than the low-carb group and they were 1.6kg heavier on average (i.e. more weight to lose), but, Table 1 overall is reassuring that the groups started off from a very similar baseline.
The study noted that there have been very few attempts to study low-carbohydrate diets and none had ticked two important criteria of 1) actually being a low-carb diet and 2) studying a diverse population. The Gary Foster (2003) study did compare an actual low-carb diet (Atkins) to a low-fat (government advice) diet, but this did not involve a diverse/carefully randomised population. This study, covered in the BBC Horizon programme about The Atkins Diet, concluded that Atkins was better for weight loss and cholesterol measures. The usual attempt to study a ‘low-carb’ diet has involved carb intake in the 100-200g range. That’s a high-carb diet to the low-carb world!
This was different. The low-carb diet group target was less than 40g of carbohydrate a day. The low-fat diet group target was less than 30% of daily energy intake from total fat and less than 7% from saturated fat – model government advice therefore. I emphasised the word target because the discussions on-line about the study didn’t pick up the actual intake reported in the paper.
Table 2 reported the actual intakes of carbohydrate and fat for the two groups:
The low-carb group were supposed to consume <40g carb per day. They did in fact consume, on average: 97g carbohydrate/day at 3 months; 93g/day at 6 months and 127g/day at 12 months. The low-carb group thus substantially missed their targets.
The low-fat group, in contrast, did hit their total fat targets (they recorded saturated fat intake of between 8-9% during the study). They were supposed to consume less than 30% of their daily intake from fat and they achieved this: 27.5% at 3 months; 27.9% at 6 months and 29.8% at 12 months.
The researchers reported that 5 people dropped out of the low-fat group and 5 people dropped out of the low-carb group after 3 months. By the end of the study (at 12 months), 13 people in total had dropped out of the low-fat group and 15 had dropped out of the low-carb group. That was an 18% drop out in the low-fat group and a 20% drop out in the low-carb group. That’s a reasonably low drop out rate for a diet study and a number dropped out due to life stressors and/or pregnancy – understandable with a largely female group 22 years old and above.
The results fell into two main areas: weight and what the researchers called “cardiovascular risk factors.” One of these I value (weight), the other I don’t (cardiovascular risk factors – apart from to wave at people who do value it).
Weight: To quote verbatim from the study: “Weight loss was greater in the low-carbohydrate group than in the low-fat group at 3, 6, and 12 months.”
The main results are worth capturing in a little table (this is an extract from Table 3 in the paper).
|Body weight (kg):||Low-fat diet||Low-carb diet|
|– at 3 mths||-2.6kg||-5.7kg|
|– at 6 mths||-2.3||-5.6|
|– at 12 mths||-1.8||-5.3|
|Lean mass (%):|
|– at 3 mths||+0.4%||+1.60%|
|– at 6 mths||+0.2||+1.5|
|– at 12 mths||-0.4||+1.3|
|Fat mass (%):|
|– at 3 mths||-0.3%||-1.10%|
|– at 6 mths||-0.1||-1.1|
|– at 12 mths||+0.3||-1.2|
This tells us many interesting things:
1) As the researchers reported, weight loss was greater in the low-carb group at 3, 6 and 12 months – at least double at all three checks. This was despite the fact that the low-fat group achieved their dietary target of <30% fat and the low-carb group did not achieve their target of <40g carb a day – not even close.
2) Lean mass increased more, at every checkpoint, with the low-carb group.
3) Fat mass decreased more, at every checkpoint, with the low-carb group.
4) Even with the low-carb group missing the 40g/day target quite substantially, this group continued to have good results for weight, lean mass and fat mass at 12 months. Notice the low-fat group displaying the classic 6 month turnaround – with regain following. This was clearly illustrated by Marion Franz’s 2007 review of 80 studies involving 26,455 people. The classic graph from this study can be seen here. By 12 months, the low-fat group had regained 50% of the weight loss at 3 months and started to decrease lean mass and increase fat mass.
5) Perhaps most interesting of all – this study disproves the calorie theory (not that any study has ever proven the calorie theory – every study proves it wrong). The low-fat group were consuming, on average, 2,034 calories at base line and the low-carb group were consuming 1,998 on average. At 3 months, these averages had dropped to 1,418 and 1,258. At 6 months, these averages were 1,481 and 1,324. At 12 months, these averages were 1,527 and 1,448.
I’ve calculated what the 3,500 calorie theory says each group should have lost and it works out as follows:
– The low-fat group should have been 7.3 kilograms (multiply by 2.2 for an approximate conversion to pounds) lower at 3 months; 13.9 kg lower at 6 months and 25.9 kg lower at 12 months. They were, in fact, 1.8 kg lower at 12 months.
– The low-carb group should have been 8.8 kilograms lower at 3 months; 16.8 kg lower at 6 months and 29.8 kg lower at 12 months. They were, in fact, 5.3 kg lower at 12 months.
Remember that the calorie theorists say that low-carb diets only work because people eat less. Both groups did eat fewer calories, but the calorie theory was as ridiculously inaccurate as it always is.
The saddest finding for me, working in the field of obesity, was how little both groups lost, on average, over the course of a year with heavily supported and well monitored conditions. There was no mention of real or processed food in the study, so continuance of processed food in both groups may have hindered weight loss. I would also have loved to see the results for a group that actually achieved the 40g/day carbohydrate intake target, as this could have been impressive. As I find, and so do many others in the world of real food/managed carb, results can be substantially better than 5 kilograms a year.
Cardiovascular risk factors: I will report these rather than discuss them because, as people who know me will know, I have no time whatsoever for the notion that cholesterol is a risk factor in anything. It requires one to assume that the body makes cholesterol to kill the host – quite possibly the most absurd assumption that one can hold. Notwithstanding this, as most of the medical world is obsessed with total cholesterol and the lipoproteins LDL and HDL (while thinking that these are bad and good cholesterol and not lipoproteins ha ha), here are the results from the cholesterol jury:
At 6 months (comparing the period in which the low-carb group were at least below 100g/day):
– The low-fat group had higher total cholesterol, the low-carb group had lower total cholesterol.
– The low-fat group had higher LDL, the low-carb group had lower LDL.
– The low-fat group had lower HDL at 3 months and no change at 6 months, the low-carb group had higher HDL at both checkpoints.
– The low-fat group had higher triglycerides at 3 months and virtually no change at 6 months, the low-carb group had lower triglycerides at both checkpoints.
The only good thing about all this is that people who believe the diet-heart hypothesis say that low carb (by definition higher fat) diets raise cholesterol, LDL, triglycerides etc and NONE of this was found to hold true.
Perhaps the biggest story of all, however, was that the UK largely ignored this. Let not the evidence get in the way of public health advice eh?!