(This was originally a blog post from December 16, 2010.)

Okay, let's talk about something of interest to many people with diabetes.

A reader who is in the pre-diabetes stage (and would rather not move on the the next) was asking me about alcohol, and specifically about whether I have found alcohol to have a problematic impact on my blood sugar. I gave him a very brief answer, to the effect that I love wine and I haven't found it to affect my blood sugar adversely. However, it's a very complicated topic which needs to be discussed in more detail than that. I wouldn't want anyone to quote me as having said that everyone with diabetes can drink as much as they like and not worry about it.


Leaving diabetes out of the discussion for the moment, alcohol has a very strange and paradoxical relationship with human health. Because it tends to be addictive (or at least heaps o' fun), people often consume too much of it. This has been found to increase the risk of various health problems, including hypertension, liver disease, depression, suicide, violence, and vehicle accidents.

Because of these problems, you would naturally expect that if we plot human mortality rate against alcohol intake, the result would be a straight-line graph, going up and to the right. In other words, the people who drink the most are going to be dying at the highest rate, and the people who don't drink at all will be dying at the lowest rate -- because alcohol is bad for you, and that's that.

It sounds perfectly reasonable. So reasonable that it almost seems like it wouldn't be worth the bother of looking at the data to confirm that this is true. And yet, when researchers did start looking at the data (especially the data for cardiovascular disease), they got a bit of a shock. Plotting mortality against alcohol intake doesn't give you a rising straight line; instead it gives you the famous "U-shaped curve":

 

Heavy drinkers do have an elevated mortality rate, just as we would have expected -- but the people with the lowest mortality rate are those who drink moderately, not the people who don't drink at all. Apparently we can't reduce the issue to a simple  "alcohol = bad" equation. Seemingly alcohol has some kind of protective effect (particularly on the heart), so long as you don't overdo it, and researchers have been expending considerable effort trying to figure out exactly what that protective effect is.

Those of us who love wine have greatly enjoyed reading about the U-shaped curve, and its presumed contribution to the "French Paradox" (that is, the unexpectedly low rate of cardiovascular disease in France and some other Mediterranean countries). But not everyone has been quite so ready to accept the cheery implications of the U-shaped curve.

In 1989, the authors of the British Regional Heart Study argued that the U-shaped curve (or at least the left side of it) is misleading. Their reasoning was that the ranks of those who drank very little or not at all included "sick quitters" -- people who used to drink heavily, but no longer do so because it had caused them serious health problems. The reason for the elevated mortality on the left side of the "U" is that it includes a bunch of people who ruined their health with heavy drinking and then had to give it up. If we're going to compare people's health with their drinking habits, we have to include their former drinking habits, not just their current ones.

Fair enough, but later studies did exactly that, and found that the U-shaped curve was still there. The people with the lowest death rate are moderate drinkers, and the difference is mainly accounted for by their comparatively low rate of cardiovascular disease.

Of course, this only raises further questions about why moderate drinking should reduce the rate of cardiovascular disease (and why heavy drinking doesn't reduce it). The answer seems to be that alcohol tends to reduce the buildup of cholesterol placques on arterial walls. How does it do that? No one is sure, but alcohol does seem to elevate HDL ("good") cholesterol, which functions as an artery cleanser. But wouldn't this work for heavy drinkers, too? Actually, it does -- alcoholics have the cleanest arteries going, but it doesn't do them any good, because heavy drinking has other effects (such as driving up blood pressure) which more than cancel out the artery-cleansing benefits. Moderate drinkers get the desirable effects of alcohol, without getting the undesirable ones. Because the undesirable effects are stronger than the desirable ones, drinking so heavily that you get both at once is a losing proposition.

What scares doctors away from recommending moderate drinking to people who don't drink currently is the addiction potential. If you advise teetotalers to start drinking moderately, there is no way to be sure they won't go too far, become alcoholics, and end up increasing their mortality risk instead of reducing it. So, don't expect doctors to start defining Minimum Daily Requirements for alcohol anytime soon. However, there does seem to be a consensus that moderate drinking is not bad for you so long as it stays moderate.

However, I should mention another health issue: alcohol is a calorie-dense substance. It's nearly as calorie-dense as fat, actually. Alcohol does not tend to help you in terms of weight control.


Okay, that's the situation for the general population. What about the diabetic population? Well, if you have diabetes, the issue of alcohol and health gets more complicated (just as every other issue in life gets more complicated if you have diabetes).

I should mention at this point a fact which many people with diabetes are unaware of: your digestive tract is not the only part of you that releases glucose into the bloodstream. Even if you didn't eat anything at all for a couple of days, there would still be glucose flowing into your blood -- from your liver. The body can't rely on the digestive system to maintain your blood sugar level, because you're not constantly eating. To protect you from low blood sugar, the liver is constantly doling out a minimum ration of glucose, and if your blood sugar actually does go low, your liver releases an extra burst of glucose to correct the problem (often it overcorrects the problem, if you have diabetes, because the burst of glucose it releases is too big for you to handle). Anyway, the liver plays a big role in maintaining your blood sugar level.

Okay, what happens when you drink alcohol? Well, the body has to go to work on processing it, so that it doesn't build up to toxic levels in the blood. And what part of your body does that processing? You guessed it -- your liver. And why is that fact pertinent to the subject of alcohol and diabetes? Because, while your liver is processing the alcohol, it drops other tasks. And one of the tasks it drops is the task of doling out your minimum ration of glucose. Your pancreas, however, does not stop doling out insulin just because your liver has stopped doling out glucose. In other words, alcohol deactivates one function without deactivating the other, so the two functions are now out of balance, and you may have more alcohol leaving your blood than entering it. The result is that drinking alcohol is likely to drive your blood sugar downward.

This makes it sound as if alcohol must be entirely beneficial for anyone with diabetes. (Who needs insulin? Let's get drunk instead!) Alas, it's not that simple. The glucose-lowering impact of alcohol is not consistent or predictable (drinking can sometimes drive blood sugar up instead of down, especially in the case of carbohydrate-rich drinks such as beer), and in any case the impact is temporary. You really can't use alcohol as a diabetes medication. Even worse, alcohol can make it a lot harder to use insulin or other glucose-lowering medications safely.

Suppose you're an insulin user. One night at dinner you test your blood sugar, so that you can figure out how big a dose of insulin to give yourself. The test reads pretty high, so you give yourself a pretty big dose. But a test is only a snapshot of a moving target; you don't realize that, when you tested your blood sugar and found it high, it was actually dropping steeply (because of the beer you were drinking with dinner). Now the insulin you injected will push your blood sugar down even faster than it was already falling. The alcohol and insulin magnify each other's impacts, and the result could be an episode of severe hypoglycemia. If you've heard anyone say that people with diabetes can't or shouldn't drink, this is the issue they're thinking of.

If you have diabetes (especially if you're taking insulin or oral diabetes drugs which drive blood sugar downward), you have to do enough testing on yourself to find out how alcohol affects you, and develop safe guidelines for yourself.

If you're managing your diabetes without any glucose-lowering medications, alcohol-induced hypoglycemia is less likely to be a problem for you. It hasn't been a problem for me. The hypoglycemic episodes I've experienced since my diagnosis in early 2001 have not been frequent or severe, and not one has ever occurred while I was drinking. I can't be sure it will never become a problem for me, but so far it hasn't. So I guess that's one more advantage of the unmedicated approach.

Alcohol does tend to drive my blood sugar downward; I've done enough testing to establish that. But the effect isn't so dramatic as to be any kind of a problem. 

I should admit that I went to a big Christmas party today, and I was drinking wine there -- and that this could be part of the reason why my glucose result after dinner was a remarkably low 89 (even though this was a non-exercise day). Still, it was a low-carb dinner, so I don't think the wine would have had to make a very dramatic impact to give me that kind of result. I do sometimes get very low post-prandial results -- without any help from alcohol.