March 17, 2009

Parsing the Alzheimers-Diabetes Scare Headlines

The AP carried a story this week that has appeared in many newsletters. More Evidence Links Diabetes to Alzheimer's Risk.

Since I've learned that these kinds of media reports are often based on poor understanding of studies, I tracked down several of the studies cited to see what they really said and learned that as I thought their findings were quite different from what you read in this press story.

Then I found another study, published in Archives of Neurology along with one of the cited studies, which turned out to have far more substantive information than the cited studies and answers many of our questions about the link between diabetes, Alzheimers, and dementia.

So what do these studies really say?

One of the more detailed studies finds a stronger association between diabetes and vascular dementia in people whose average age was the mid-80s. Vascular dementia is a different condition from Alzheimer's that has a different pattern of progressing. It is characterized by the occurrence of mini-stokes in the brain.

This study, not cited in the media, quantified the heightened incidence of dementia in people with diabetes diagnoses. In it, 26% of the 125 people whose brains were autopsied who died in their 80s without dementia were diagnosed with diabetes while 36% of the 71 who were demented had been diagnosed with diabetes.

The researchers doing this study had access to the A1cs and fasting blood sugars of the group as a whole, as well as their medication history. Based on this they noted that the A1cs of those people with diabetes who had not developed dementia were lower than that of those who did--(7.4% compared to 7.8%). This is similar to the finding of another study we cited in an earlier blog post. Though it is worth noting that 63% of those who developed dementia did not have diabetes.

On close examination of these people's brains the researchers found:

Most interstingly, "Individuals without DM but with dementia (DM−/dementia+) had a greater amyloid-beta peptide load and increased levels of F2-isoprostanes in the cerebral cortex, while DM+/dementia+ patients had more microvascular infarcts and an increased cortical IL-6 (interleukin 6) concentration. The number of microvascular
infarcts was greater in deep cerebral structures in patients with dementia whose diabetes was treated, whereas amyloid plaque load tended to be greater for untreated
diabetic patients with dementia.
This means found fewer Alzheimer-like plaques and tangles among the people with diabetes than among people without, and also more IL-6 in the cortex. IL-6 is a marker for inflammation. They also found more sub-cortical lesions in people with diabetes--strokes deep in the brain which may have to do with the nature of the blood supply to those regions.

But what is really interesting is that the people "with diabetes" in this group, who were not "treated" i.e. medicated, had brains more like those of people without diabetes in terms of the amount of beta-amyloid plaques and tangles. The researchers comment that the drop in beta amyloid tangles is,
a result consistent with a recent report from a large autopsy series that showed that decreased senile plaque burden was associated with insulin therapy.
But the researchers also note that they found more evidence of deep brain strokes in the treated group which they note had much higher A1cs than the nonmedicated diabetics. (6.3% unmedicated, 8.6% medicated.)

Unfortunately, the researchers also explain, "A weakness of our study was the limited numbers of DM+/dementia+ cases available for analysis of treatment effects, which made it impossible to analyze specific diabetes treatments."

Another weakness not cited by the researchers but revealed in the data is that the group who were not demented had a slightly lower average age. Since the people (including diabetics) with vascular dementia were two years older on average than those without, and since vascular dementia can arise and kill very quickly (I've seen this in our extended family) the increased amount of vascular dementia may be partially explained by greater age.

Still, this study suggests that there is a slightly higher likelihood of developing dementia among people with diabetes, especially among those with Average A1cs of 8.6% and that this appears to be due to their propensity for having tiny deep brain strokes and brain inflammation. At the same time--what the news reports ignore is that this study, like most, shows that the overwhelming majority of people who develop dementia do not have diabetes.

But before you get too excited about this finding, consider the findings of another study. It found a much stronger link than the one shown in the study we just discussed between vascular dementia and the presence of so-called "metabolic syndrome" in people 92% of whom did not have elevated blood sugars.

In that study, which examined "4895 older women (mean age, 66.2 years)
A total of 497 women (10.2%) had the metabolic syndrome and, of these, 36 (7.2%) developed cognitive impairment compared with 181 (of 4398 or 4.1%) without the syndrome (age-adjusted odds ratio, 1.66; 95% confidence interval, 1.14-2.41).
The incidence of dementia was almost twice as high in the group with metabolic syndrome, though it was low overall probably because this group was younger than the group studied above. But given that less than 8% of this group had elevated blood sugars, one wonders about the extent to which the microvascular damage in either study was due to blood sugar rather than high blood pressure and high triglycerides which were strongly predictive factors in this second study, independent of blood sugar.

And the ambiguity of the results gets worse, because yet another study this one a cohort study of a large number of 2798 people followed from 1992-1999 comes up with this odd result:
In evaluations of midlife obesity, an increased risk of dementia was found for obese (BMI >30) vs normal-weight (BMI 20-25) persons, adjusted for demographics (hazard ratio [HR], 1.39; 95% confidence interval [CI], 1.03-1.87) and for cardiovascular risk factors (1.36; 0.94-1.95). The risk estimates were reversed in assessments of late-life BMI. Underweight persons (BMI <20) had an increased risk of dementia (1.62; 1.02-2.64), whereas being overweight (BMI >25-30) was not associated (0.92; 0.72-1.18) and being obese reduced the risk of dementia (0.63; 0.44-0.91) compared with those with normal BMI.
In short, as you get older the fatter you are, the less likely you are to develop dementia, though if you are middle aged being overweight is more highly associated with dementia.

This is not a freak result, as NHANES data has long supported the finding that after age 70, any weight loss correlated with a higher likelihood of death and that people in the overweight category seem to do better, long term as they age than people of so-called normal weight.

So what should we conclude from all this? Are the headlines unnecessarily alarmist? I'd say, yes. In fact there is no connection here between classic Alzheimer's disease and diabetes. In fact, the opposite seems to be true. People using insulin appear to have less Alzheimer's.

Uncontrolled high blood sugars do seem to slightly up your chance of developing vascular dementia, but not any more than does being overweight in middle age, and having high blood pressure and high triglyceride levels.

Looking at the broader picture, none of these factors make that much of a difference in your risk of developing dementia, because the overwhelming number of people who develop dementia do not have diabetes and more importantly, as people get older, the overweight that is so demonized by doctors appears to protect people from developing dementia rather than promoting it.

Bottom line: There is much you cannot control when it comes to dementia, but to make a slight improvement in your chances, keep your blood sugars under control, keep your blood pressure normal. Keep your triglycerides down. You can lower triglycerides by keeping your carbohydrate intake low because triglycerides are produced when you eat more dietary dietary glucose than your body can burn right away.

If you have diabetes, you can normalize your risk of diabetes-associated dementia by keeping your A1c in the normal range. Your risk for dementia goes way up if you maintain an A1c in the 8% range.

A Last Bit of Good News

The first study cited above gave us some numbers for the blood sugar of "normal" people whose risk for dementia was lower than those with diabetes. Their average A1c was 5.9% and their average fasting glucose was 105 mg/dl (5.8 mmol/L).

These are VERY attainable numbers for all people with diabetes so if the association of "diabetes" with dementia worries you, shoot for those numbers.


ItsTheWooo said...

Working in a LTC facility, I can say that I've noticed a strong correlation between loss of mental function/dementia and diabetes. This is hardly conclusive evidence, but I've noticed it to be quite true that my residents with dementia are disproportionately diabetic. Those who are not diabetic are often hypertensive (more than normal hypertension, malignant type). When dementia is present it is almost always true a history of severe hypertension and/or diabetes is and was present. Dementia is dementia; even if we assume it's not alzheimers but vascular type, it's still horrible and incapacitating.

As people age, weight has an linear relationship to health, which is pretty much the opposite one would expect in youth. During middle age and earlier, weight has a general opposite relationship with health and risk for disease, but during old age weight is one of the earliest signs of decline. When weight is lost they become very weak and vulnerable to infection, skin breakdown, death.
When weight is lost in an elderly person, it also implies a progression of dementia and deficit in ability to perform ADLs. People with dementia do not eat. They would emaciate and starve unless prompted to by caregivers.

In middle age and younger, weight indicates is pathology of energy metabolism consistent with the metabolic syndrome, which is the same factors hypothesized to contribute to a great many cases of dementia in old age.

It seems significant to me that the rate of dementia in middle aged people was twice as high as those with the metabolic syndrome, and this is of relevance to people with diabetes as many diabetics have metabolic syndrome. I'm not sure why you think this is insignificant.
It also seems significant that 37% of people with dementia have diabetes. That is a very high rate, and from my experience it is probably conservative.

Obviously there is more to dementia than diabetes so the rate isn't going to be 100%. There are many different kinds of dementia, it is actually a lot more complex than "alzheimers" and "vascular type". Many cases of dementia are mixed. Some may have nothing to do with metabolic syndrome (insulin, blood pressure, blood sugar). The fact is there is something quite significant about metabolic syndrome and rates of dementia, and all the data you present here indicates that.

Jenny said...

My guess is that the connection with so-called metabolic syndrome is as strong as it is because people with MetSyn don't get intensive medical treatment, and go for years with elevated blood pressure and triglycerides without anyone noticing.

The people I've known with vascular dementia tend to be otherwise healthy people who rarely went to the doctor in middle age. So they might have been borderline for these issues without it being picked up.

Anonymous said...

I recall some work on the effects of insulin on the brain.

It figures that people with metabolic syndrome will have insulin resistance and hyperinsulinemia although their BG may still be at nondiabetic levels, this could be a factor.

Many of them will also be on statins which could be a further factor.

Would be interesting to compare rates with diabetics with A1cs in the 5% Club but no-one studies them.

Yes yes, I would post the paper if I could remember where I put it . . .

Cheryl said...

Jenny -- did you see the study (on rats) connecting Metformin use with Alzheimer's?

As a user of generic Metformin (and no insulin) I find this troubling and will bring it up with my doctor. What do you think?

Jenny said...


I did see that article, and I wasn't impressed with it. I will go back and check out why when I have a bit more time which I don't right now.

Don't let it upset you.

Unknown said...

For one thing, rats are not furry little humans.

Jenny said...


I haven't read the full study which sounds like it's probably over my head being extremely technical, just the summary in DiC and the abstract.

However this was NOT a rat study. It was a cell culture study. So what is missing is the point that people using metformin may have very high circulating insulin (in early Type 2 with strong IR) and that if they get to where they don't have insulin getting to the brain they are going to be injecting insulin or they won't live long enough to develop Alzheimers.

In people, the evidence suggests that people with diabetes have much more vascular dementia than the kind with plaques and tangles.

More importantly, people with diabetes taking metformin without insulin usually have extremely high blood sugars (8% and over) which are so damaging to the body that it is impossible to separate out any impact of metformin.

I'd take from these cell studies only that metformin hitting a cell without insulin in the neighborhood might cause problems.

Given, though, that metformin has been used for many decades, you'd think some evidence of higher AD would have shown up by now were it real.

Cheryl said...

Jenny, Thanks for looking further into the study.

You said, "Given, though, that metformin has been used for many decades, you'd think some evidence of higher AD would have shown up by now were it real."

Exactly what I wondered!