This isn't original research, it's just a look at the major studies that have examined the impact of metformin on prediabetes. But because I hear from so many people with prediabetes whose doctors won't give them any help at all, I though it worth a look.
Treating prediabetes with metformin: Systematic review and meta-analysis Muriel Lilly, Can Fam Physician Vol. 55, No. 4, April 2009, pp.363 - 369
The key issue to remember here is that the concept that "prediabetes" progresses to "diabetes" which treats the two conditions as if they were separate diseases is flawed.
In fact, the medical definition of "diabetes" is completely arbitrary. A committee years ago chose some blood sugar test results and defined them as "diabetes." They could have--and many argue should have--chosen different test result numbers. But they chose the ones they did mainly, their own documentation showed, to diagnose people with diabetes as late as possible, because of the severe penalties the American medical system imposes on people who have pre-existing conditions.
You can read about how the diagnostic standards for diabetes were set HERE.
"Prediabetes" was also defined arbitrarily at the same time as "diabetes" was defined and as has been the case with diabetes, the definition has changed over the years.
But what you, the person with abnormal blood sugar, need to understand is that there's no sudden change in your health that happens when you get an official diabetes diagnosis. Any blood sugar that is elevated above normal for hours each day can and will damage your organs. So if your concern is to keep your nerves, heart, kidneys and retinas healthy your focus should be on keeping your blood sugars normal, not in avoiding a technical "diabetes" diagnosis.
With this in mind you can see why the argument the medical establishment uses to argue that "more research is needed" re the use of metformin in prediabetes is a red herring. The argument is this: "We don't know whether metformin is preventing diabetes or just masking the symptoms by lowering blood sugar."
Those who argue this point out that if people stop taking metformin, their blood sugar may go right up to where it would have been without it. So if that is the case, the metformin didn't "prevent" diabetes and there is no point in prescribing it.
But hold on a minute. The damage from "diabetes" is done by the elevated blood sugars, which clog capillaries, block kidney filtration units, and destroy the nerves of the autonomic nervous system that regulate the heart. If we lower blood sugar we avoid this damage. There has never been a single study that shows that people with "diabetes" by diagnosis who maintain completely normal blood sugars develop any of the complications of diabetes. These are caused by the exposure to high blood sugars, not the underlying dysfunction that caused the high blood sugars.
So if giving people metformin lowers people's blood sugar to where it isn't going up to diabetic levels, this is good no matter how you look at it. Even if it doesn't change a thing except their blood sugar levels.
The real problem with using metformin to "prevent diabetes" lies in how "diabetes" is diagnosed in these studies. Most doctors and studies diagnose diabetes as meaning that the patient's fasting blood sugars have risen over 125 mg/dl. Because metformin often lowers fasting blood sugar 5, 10 or 20 mg/dl, someone who might have a blood sugar of 128 mg/dl and be diagnosed as "diabetic" will take metformin, get a fasting blood sugar of 123 mg/dl and be diagnosed "Non-diabetic" in the studies.
This makes a pretty graph or two for the researcher but it doesn't make much functional difference in the health of this patient. Because functionally there isn't much difference in how much damage you are doing to your body with a fasting blood sugar of 123 mg/dl vs 128 mg/dl.
What is much more significant if you are trying to keep people from going blind or losing their kidneys is what happens to the blood sugar after meals--which is a statistic these large "diabetes prevention" studies rarely track since it requires much more expensive testing.
A person with the fasting blood sugar of 108 mg/dl may be be going up to 180 mg/dl after meals and coming back down to 100 mg/dl in two hours. Or they may be going up to 270 mg/dl and drifting back down to 108 in 4 hours. Both people are prediabetic, but the first person is much less likely to develop diabetic complications than the second. And, in fact, the second is technically diabetic given the ADA diagnostic criteria since they have random blood sugars over 200 mg/dl--except that their true diabetic state never gets discovered since the doctors and researchers only look at their fasting blood sugar.
What people with prediabetes need to understand is this. Organ damage starts when blood sugars spend a few hours a day over 140 mg/dl. It doesn't matter what the fasting blood sugar is. It matters how long blood sugar stays over 140 mg/dl. Neurologists studying neuropathy (nevre damage) have found no relationship at all between fasting blood sugar or A1c and the likelihood of developing neuropathy. They've found a very tight correlation between the rise in incidence of neuropathy and 2 hour blood sugar values that are 140 mg/dl and higher on glucose tolerance testing. (You can read more about studies linking organ damage to blood sugar levels HERE.)
So if you have prediabetes defined either by fasting blood sugar greater than 100 mg/dl or post-meal blood sugars that go over 140 mg/dl for significant periods of time the thing you really want to avoid is neuropathy, because it is among the earliest diabetic complications and one that affects all of your body. (The very earliest appears to be carpal tunnel syndrome.)
Almost half of all "newly diagnosed" Type 2s have neuropathy--a complication that can take up to a decade to develop in people diagnosed with sudden onset Type 1 diabetes. That tells us how damaging all those years of untreated "prediabetes" really are.
Metformin can help lower blood sugar, but if you are a person with prediabetes who tests your blood sugar after meals while taking metformin, you'll often see that your blood sugar will still go high enough to cause damage unless you also cut back on your carbohydrate intake.
But metformin definitely can help, and it can allow you to eat more carbs and still get decent blood sugar numbers, which in turn makes it easier to eat in a way that maintains health since it's a matter of cutting down rather than cutting out the carbs.
So as soon as you find your blood sugar is higher than normal, the better off you'll be if you take steps to get your blood sugar back down to the normal range. Whether or not you reverse the underlying condition that made your blood sugar abnormal is irrelevant. Most of us can't, contrary to what many doctors tell people. But as long as we can keep our blood sugar in the normal range, we'll feel much better, since fluctuating blood sugars make for rabid hunger and depressive mood swings, and we will also maintain the normal blood sugars that will keep our organs functioning in a normal way.
If you are not getting normal blood sugars with reasonable dietary changes and your doctor won't let you try a course of metformin, ask why, and if the reasons don't sound credible, find a doctor who will be a better partner with you in the struggle to maintain your health.
PS: Legitimate reasons not to use metformin are known liver and kidney problems.
If metformin does not lower your blood sugars and they continue to deteriorate, it is possible you have an autoimmune form of diabetes. Metformin does not help diabetes caused by damaged beta cells and immune disease and people whose blood sugar does not respond to metformin often turn out not to have Type 2 diabetes.