As the regulars here probably know, I take quite an interest in this. Diabetes Type 2, and all the horrible consequences that follow, is simply not taken seriously by the NHS, despite the fact that a large chunk of its budget disappears into that particular black hole. Metabolic Syndrome in general consumes about 30 billion pounds annually: that's £1000 per taxpayer. How much good, healthy food could be bought by each British family if it weren't being wasted in this manner? For whatever reason, the treatment of this terrible disease is still driven more by ideology and superstition than known science.
You will often see statements like "there is no known cure for diabetes" if you Google the subject. This is untrue, and I'd like to present some data showing exactly how untrue it is. That the people charged with our care should be telling outright lies just blows my mind; hence my frequent rants about it here.
I'll start with a persistent myth: that a vegan diet can reverse diabetes. To be clear, it's up to the individual what they choose to eat. If they don't enjoy meat, or find that it disagrees with them, or don't like the thought of eating dead animals, deciding not to eat it is their prerogative. However, the idea that a low-fat plants-only diet can cure diabetes is not supported by the evidence.
The graph at the top is copied from a study from Neal Barnard, who is one of the foremost proponents of low-fat, high-carb diets for diabetics. He's in good company: the NHS and all the major international diabetic thinktanks basically concur with his views. I've extended the Y-axis downwards to include the normal range of HbA1c, between 4% and 5.5%, indicated by the grey band. Barnard's original graph didn't bother to show this region, because none of his subjects went anywhere close to it. As you can see, over 74 days, their HbA1c decreased from 8% to 7.65%. The line "P=0.03" indicates that there is a 1 in 30 probability that this result might be pure chance. In any case, 7.65% is still diabetic. Not terribly so, but diabetic nonetheless.
A meta-analysis of similar studies supports Barnard's result: a small reduction in HbA1c which is statistically significant, but not clinically significant:
Why doesn't it work?
Diabetes Type 2 is sometimes described as "extreme insulin resistance", but this isn't entirely correct. Insulin resistance is just an adaptive mechanism that your body uses to apportion dietary glucose between different subsystems. There's only one signal that tells your body's cells to take up glucose - that's insulin - so logically there must be something else that tells them how much. That something is insulin resistance: cells decide for themselves how much to take, and how fast. In other words, each cell in your body has its own little tap that allows it to say "that's enough for now". At the individual cell level, it's an on-off tap, but because you have so many cells, the net effect is that each organ has a soft threshold for glucose.
This works fine as long as the total rate at which your organs can take up glucose matches (or exceeds) the rate at which your pancreas wants glucose removed from the bloodstream. No "proof" is needed for this: it's as simple a statement as 1+2+3 = 6. By this mechanism, your body can ensure that (for example) your brain gets the lion's share of glucose: your brain has a fairly high requirement for glucose, whereas your heart and muscles have little to none (they can use glucose, but they prefer other fuels).
What happens if glucose comes in faster than your organs require? Well, they'll all switch off that incoming tap, and blood glucose will inevitably rise, because it has nowhere to go. Your pancreas observes this and cranks up insulin production, because persistently high blood glucose is a big, big problem. It "forces" your organs to mop it up, even though they don't want it. Fat cells are the main target, because they can swell (and multiply) almost without limit. In general, though, your cells respond to this by saying "no" even louder. If this continues over a period of years, everything maxes out. Your pancreas is asking for 1+2+3 to equal 15. That's diabetes.
It should be completely obvious that the way to reverse this situation is to back off on the glucose, and this works incredibly well in practice:
You don't need to read the whole thing. Just scroll down to the "Results" section. Under "Effectiveness", we see this:
60 ± 1.0 mmol mol−1 (7.6 ± 0.09%) at baseline to 45 ± 0.8 mmol mol−1 (6.3 ± 0.07%) after 1 year
In plain language, the subjects started off (on the average) about as diabetic as Barnard's patients, and ended up (on the average) in the 'prediabetic' range. This might seem somewhat unimpressive, but the average doesn't tell the whole story. Look at the scatterplots - figure (b). These show every single subject as a blob. On the X-axis is where they started off; on the Y-axis, where they finished up. A dot appearing above the sloping line means they got worse; a dot appearing below it means they got better.
With the experimental intervention (left-hand plot) nearly everyone showed some improvement. Only a handful of unfortunates ended up in a worse place. Notice how many dots are below the magic 40mmol/mol (5.8%) line: dozens of them. They're "cured".
The right-hand plot is the result of the "expert" recommendations - high-carb and low-fat. Two people are below 40mmol/mol after "treatment", and they were already well below 50 to begin with. A lot of them have blood glucose up in the stratosphere.
Notice also that in the intervention group, virtually nobody is above 60mmol/mol; this level, while still technically diabetic, is associated with a marked reduction in risk:
In the standard-treatment group, about half of the subjects are above that threshold.
So why aren't the health authorities piling onto this result with enthusiasm?
Well, you tell me. I have my own suspicions, but perhaps I'd better take off my tinfoil hat now and pass the debate over to the floor.