How Low Blood Sugar is Too Low? New Evidence

Some people avoid all carbs (that is, all bread, pasta, rice, etc.). Others advocate “safe starches”. No doubt wheat can be dangerous — witness celiac disease (associated with a genetic difference). But I have noticed clear improvements in brain function (measured by arithmetic speed or something similar) after eating something sweet, such as pudding. None of this, unfortunately, helps answer the question: how low blood sugar is too low?

A new study makes more plausible the idea that really low levels of carbs may be bad for you:

A nested case-control study data set was generated from the cohort-study data set (n = 4140 type 2 diabetic outpatients) by sampling controls from the risk sets. Cases (n = 427) were compared with an equal number of controls chosen from those members of the cohort who were at risk for the same follow-up time of the case, matched for age (±3 years), sex, body mass index (BMI) (±2 kg m(-2)), duration of diabetes (±5 years), and Charlson’s Comorbidity Score (CCS) (±1). The main predefined analysis was the comparison of cases and controls for proportion of patients with each HbA1c class (<6.5%, 6.5-7.4%, 7.5-8.4% and ≥8.5%). During a mean follow-up of 5.7 ± 3.5 years, 427 deaths were recorded. The lowest risk of death was observed in the HbA1c 6.5-7.4% category; a lower HbA1c was associated with a non-significant trend towards a higher risk. The risk associated with a low (<6.5%) HbA1c was significantly greater in patients who were insulin-treated than in the rest of the sample.

The study is saying that diabetic patients in the HbA1c 6.5-7.4% category do not improve their health when given insulin that lowers their blood sugar even more. Their health may get worse.

17 Replies to “How Low Blood Sugar is Too Low? New Evidence”

  1. As I understand it, HbA1c is a weighted average reading of blood sugar over the preceding few months. But what if not only the level matters, but also how variable it is? Wouldn’t that be lost in correlating HbA1c versus this-and-that?

  2. They probably did not include diabetics who were controlling it based on diet alone. That is diabetics only on drugs were considered.

    Its very obvious that the body will not be able to control bg if you supply any external way of reducing bg. The external factors and the internal factors will be in conflict. So the chances of death due to hypoglycemia will get higher as you keep the blood control below a certain level, and the death due to hyperglycemia will become more prominent as the bg level goes to another extreme. I guess the 6.5-7.4% is the happy medium, in the insane world of drug based glycemic control.

  3. It looks to me as if they should have separated the <6.5% people into two groups: (a) the insulin-treated, and (b) the others. Then they could tells whether there is any higher risk for group (b) and whether group (a)'s higher risk is significant.

  4. Hi Seth,

    This might represent too low blood sugar, or too much insulin treatment, or it might not. The problem is Hba1c is not a specific marker for blood sugar exposure. A lower Hba1c could indicate higher red blood cell turnover. Or, it could reflect a greater activity of fructosamine 3-kinase, which “detoxifies” the fructosamine residues of Hba1c to a product that, if it is not quickly taken care of itself, will degenerate to form 3-deoxyglucosone. Second only to methylglyoxal, 3-deoxyglucosone is one of the most major sources of advanced glycation endproducts (AGEs) in diabetes. The problem is that the formation of Hba1c probably does not mediate much of the disease process in diabetes, whereas the formation of 3-deoxyglucosone probably mediates a major part of the disease process. I wonder if they discuss this issue in the full text. Unless they looked into it specifically, it could just mean that Hba1c is not a very useful marker of the pathological process of diabetes.


    Seth: Those are good points, thanks.

  5. It’s well-established that “tight” control of blood sugar via medication increases mortality. This is the finding of the ACCORD Diabetes trial. However I don’t know that you can read that result, or the result in this trial, to infer that a low-carbohydrate diet would also have ill effects. The difference, of course, is that a low-carb diet allows the body to produce the appropriate amount of glucose to fuel the body, and also results in a low insulin level.

    Hyperinsulinemia has its own dangers, which may be the cause of the increased mortality seen in this study.

    “ADA: ACCORD Diabetes Trial a Complete Bust”

  6. Funny thing is… I would swear to an increase in my math abilities, as well as all other non-critical brain function, when in ketosis – just the opposite of what you’ve said. I find myself thinking much more clearly, and, using my judgment in stressful situations as a marker, much more effectively and quickly, when I HAVEN’T eaten carbs.

    I suppose there may be something to be said of the type of carbs one eats, and the relative fat content of said foods. Something to blunt the “sugar rush” that I always feel when eating simple carbs. And I’m not even diabetic!

  7. This result does not generalize to non-diabetics. The most likely relevant difference is the number of hypoglycemia events (both reported and unreported), which is much higher in the intensive-control group but which is caused exclusively by mistakes in use of insulin.

    Also conspicuously missing from the study: any thiamine or thiamine analogs. Thiamine deficiency is extremely common among diabetics (because of increased excretion), and a major risk factor for heart disease.

  8. Seth: the brain uses a lot of the body’s glucose intake for energy. Is this a possible explanation for your improvement after eating something sweet? Furthermore, pudding strikes me as somewhat fatty. Given the buttermind results, perhaps you could test the hypothesis by trying sweet; sweet + fat; and fat.

    Seth: Yes, that is a plausible explanation. I am unclear about how your proposed experiment tests that explanation.

  9. I think you’re framing the question wrong. I would bet that the insulin itself is causing more damage than the level of blood sugar. This would be consistent with the study’s result, and also consistent with the possibility that people who do not inject insulin have better health when they eat in a way that causes less cumulative pancreatic insulin release.

    Seth: Yes, that is a fair comment. As a way to decide how much “starch” to eat, this is crummy evidence. It is also better than nothing. Alternative explanations like yours point the way to better evidence — studying non-diabetics, for example.

  10. “But I have noticed clear improvements in brain function (measured by arithmetic speed or something similar) after eating something sweet, such as pudding.”

    Interesting. The guy who holds all the Swedish memory records eats very low carb during competitions because he says it helps him sustain his concentration.

  11. Extremely low blood sugar is dangerous. I mean really low (maybe 60 and down or something)

    Now, having diabetes means sugR fluctuates, especially when using insulin.

    Thus, haivng lower A1c means that sometimes sugar went really low. Wich is dangerous.

    I think that without diabetes, h1c is best at about 5or a tiny bit more. Indicating blood sugar of <100.

  12. In my view Anand and Tuck are onto something important that deserves much more discussion; and is typically overlooked in the context of Type 2 diabetes. It seems to me that using artificial methods of adding (synthetic) insulin to the body to control blood glucose in the T2 diabetic is a wrong-headed way to deal with the problem.

    It seems almost too obvious to have to say that T2 diabetes is a “disease” of too much glucose in the bloodstream. If this is a true statement (and I don’t know who could reasonably argue with it), then it follows that a T2 sufferer should take steps to restrict (not completely eliminate) the intake of foods (sugars and other carbs) that turn into too much sugar in the blood, thereby burdening (i.e., overworking) the pancreas day in and day out; ultimately to the point of exhaustion.

    It’s well known that constantly high level of insulin production (caused by the pancreas having to overwork itself to knock down the constant flow of excess glucose from ingested sugars and other excess carbs – together with whatever synthetic insulin is added artificially in the case of the T2 diabetic) causes and/or adds to inflammation in the body (particularly of the pancreas and other major organs). This alone could, and likely does, explain the higher mortality rate.

  13. @WB “T2 sufferer should take steps to restrict (not completely eliminate) the intake of foods (sugars and other carbs) that turn into too much sugar in the blood”

    I’m curious why you specifically say ‘not completely eliminate’ rather than ‘or completely eliminate’.

  14. Seth, you said: “I have noticed clear improvements in brain function (measured by arithmetic speed or something similar) after eating something sweet, such as pudding.”

    They have also measured increases in willpower following consumption of sweet things. What occurred to me is that this might be more like a tradeoff than a net benefit. What I mean by that is that yes, you might gain a brief moment of improvement in brain function or willpower, but what happens after it? There is a medical term called “reactive hypoglycemia” or “postprandial hypoglycemia” that describes low-blood sugar in non-diabetics after consumption of high amounts of carbohydrates. If temporarily elevated blood glucose increases brain function, I wonder what happens when blood sugar is depressed…

    People who do not eat carbs don’t have problems with low blood sugar as far as I know and they certainly aren’t going to have reactive hypoglycemia. Average fasting glucose level for patients on a low-carb diet in one trial I looked at recently was 92.6 mg/dL.

    Seth: I get the improvement after eating what most people would consider small amounts of sweet things (e.g., a small package of pudding). There is no reactive hypoglycemia afterwards.

  15. @Robbo:

    Hi Robbo, you ask a valid question. I’ve been restricting carbs for the last 5- plus years and have greatly benefited from it. I will never go back to the so-called standard “balanced” USDA approved diet.

    When I first decided to commence carb restriction I was probably on the edge of becoming a T2, but fortunately succeeded in avoiding or circumventing it by being very strict for a long time. And even now, most of the time my carb intake is still very low (probably no more that 10-15 grams of carb per day). But nowadays I do allow myself to indulge in carbs to some extent; e.g., dark chocolate or some cheesecake or, strawberries with cream.

    I completely understand the “no carb” frame of mind because I’ve been there, and in fact most of the time I’m still doing essentially that. It’s just that now I feel I can loosen up from time to time.

    I firmly believe, from my experience, that T2 can be cured with carb restriction (while avoiding all meds), and for many individuals something close to complete elimination, especially in the early days of recovery, is definitely a good thing.

  16. Seth, you said: “I get the improvement after eating what most people would consider small amounts of sweet things (e.g., a small package of pudding). There is no reactive hypoglycemia afterwards.

    From JELL-O’s website a 100g package of pudding has 92g of carbohydrate and 72g of sugar.

    The medical standard for an oral glucose tolerance test is 75g and oral glucose tolerance tests are no longer used to diagnose reactive hypoglycemia because they were found to CAUSE it.

    I’m not saying that you’re wrong and you do get reactive hypoglycemia, rather I’m trying to say that it is highly plausible that a sugary snack could cause it in people in general. Even if it doesn’t cause clinical reactive hypoglycemia, it could still cause lowering of the blood sugar after it has peaked, which would presumably have the opposite effects of increased blood sugar on cognition and willpower.

    Seth: I ate much less than 100 g of pudding. Maybe 30 g.

Comments are closed.