August 13, 2012

Why Thinner People with Diabetes May Have Worse Outcomes

A study published last week in the Journal of the American Medical Association (JAMA) came up with the conclusion that "Adults who were normal weight at the time of incident diabetes had higher mortality than adults who are overweight or obese."

The study can be found at:

If you have been doing a good job controlling your weight and read only the media reports about this study, you may have found it disturbing. If your doctor is one of those who only reads the "25 words or less" summary of medical findings, you may even hear that losing weight is dangerous for diabetics.

Hermes Florez, MD, PhD. Florez, the director of the division of epidemiology and population health sciences at the University of Miami's Miller School of Medicine, who wrote an editorial about this study  was quoted as saying, ""If you are normal weight, you may be at higher risk from diabetes, especially if your fitness status is not so good."

But even a cursory reading of the methodology used in this study should make it clear that the outcomes for people who were normal weight at their diabetes diagnoses has little to do with their weight or, for that matter, fitness.

The data used in this study came from pooling data from the "Atherosclerosis Risk in Communities study, 1990-2006; Cardiovascular Health Study, 1992-2008; Coronary Artery Risk Development in Young Adults, 1987-2011; Framingham Offspring Study, 1979-2007; and Multi-Ethnic Study of Atherosclerosis, 2002-2011."

As you will immediately notice, all but one of these studies go back 20 years or more, to the days before doctors understood that only by lowering blood sugar could people with diabetes avoid diabetic complications. So most of the people who participated in the studies whose data was used in this meta analysis were diagnosed in the bad old days, after which they were likely to have received extremely poor treatment.

How bad? Before the middle 1990s, patients diagnosed with Type 2 almost never were given home blood sugar meters. The only time their blood sugar was measured--fasting--was when they saw their doctor once every few months. Since the only drugs available to treat Type 2 Diabetes were the sulfonylureas which can cause dangerous hypos,  the only concern doctors had about these newly diagnosed people's blood sugar was that it not drop low enough to cause hypos. As a result, doctors encouraged patients to keep their blood sugars high and when they administered A1c tests doctors  saw no problem with patients maintaining A1cs in the 10% range or higher.

So what does this have to do with why their weight at diagnosis might have given normal weight people a higher likelihood of dying over the subsequent decades? Just this: Nowadays those of us who keep up with diabetes are aware that most normal weight adults diagnosed with Type 2 are very likely not to have classic  Type 2. Most have a slow-developing form of autoimmune diabetes called LADA that is very similar to Type 1.

This form of diabetes does not respond at all to the oral drugs that are somewhat effective for Type 2. But back in the days when patients had no way to test their blood sugar after eating and doctors considered A1cs as high as 13% "safe" these patients could go many years without being put on insulin. And even when they were put on insulin, because they didn't have meters, they injected only enough insulin to prevent the sky-high blood sugars that cause the fatal condition, diabetic ketoacidosis. Many patients on insulin still had blood sugars that rose into the 200s and even 300s.

So normal weight people with LADA would have ended up with blood sugars much higher than those of the overweight Type 2s who were more likely to respond to the sulfonylurea drugs and, later, metformin, most newly diagnosed people with diabetes received.

And because we now know that there is a straight line relationship (in large groups of people, at least) between A1c and the likelihood of heart disease, it doesn't take genius to suspect that the explanation for the extra cardiac deaths in the group of normal weight people with diabetes was the those much higher blood sugars they ran due both to misdiagnosis and to the way doctors ignored the damage caused by high blood sugars until the middle of the 1990s.

A second group of normal weight people diagnosed with diabetes as adults who are likely to end up with poor outcomes are people with milder versions of one of the many, unrelated, forms of genetic diabetes that are lumped together under the name MODY.

Many people diagnosed with MODY have family histories of relatives dying young of heart attacks while young and fit. This is probably because exposure to blood sugars that remain over 150 mg/dl for hours each day will over time cause the arteries to become diseased. People with these forms of MODY have had that kind of elevated blood sugar all their lives. But until very recently only those with the worst cases were diagnosed, and the rest, whose genes gave them "pre-diabetes" for much of their lives, lived with abnormal post-meal blood sugars for  many years, and only developed full fledged diabetes much later on in life, perhaps after becoming a bit more insulin resistant.

Because they were thin, doctors didn't consider a diabetes diagnosis when they had symptoms that should have pointed them that way.  (In my 20s, my doctors waved off my highly abnormal glucose tolerance test results as "nothing to worry about" precisely because I was normal weight.)

Some rare forms of genetic diabetes are caused by genes that damage not only blood sugar control but other organs, most notably the kidneys and the mitochondria. Mortality in families that inherit these unfortunate genes have nothing to do with diabetes but with the other organ damage that is due to these gene variants.

But for most people with either MODY or LADA it's pretty clear that if you keep your blood sugars in the normal range (under 140 mg/d--7.7 mmol/L) at all times you can, over time, reverse much of the damage that you have already suffered and avoid the premature heart attacks and kidney failure that carried off so many people diagnosed with diabetes in the past, when they were allowed to run very high blood sugars.


There is also one other reason why thin people diagnosed with diabetes may have a higher risk of dying--one that wasn't addressed by this study's methodology. The journal article states that the data were adjusted for "demographic characteristics and blood pressure, lipid levels, waist circumference, and smoking status." However, what was left out was co-morbidities--i.e. the presence of other diseases.

One of the most important comorbidities to be considered is pancreatic cancer, since there is growing evidence that a very early symptom of pancreatic cancer may be the onset of diabetes. This is discussed in the Mayo Clinic web page HERE. A small but important subset of normal weight people who die prematurely with diabetes are people whose pancreases failed because they were infiltrated with cancerous cells.

In addition, many drugs used to treat cancers and life-threatening diseases are capable of causing diabetes, for example the steroids administered during chemotherapy or to treat COPD. Powerful steroids were also prescribed in the past to people with neuromuscular diseases like MS. Data sets like those glommed together in this study can't tell us whether a person was diagnosed with diabetes after a course of chemo or whether they died of cancer, COPD, or MS. So the presence of people in this data pool who were diagnosed with diabetes who also had one of these other life-shortening conditions could also explain a significant amount of the excess mortality this study uncovered.

Whatever the explanation, if you are a normal weight person diagnosed with diabetes, there's no need to panic. If you track your blood sugar after meals with a meter and keep it as close to the normal range as possible, you should do fine. The worst treatment doled out to people with diabetes nowadays is far better than what the people involved in these studies back in the late 70s, 80s and early 90s were given after their diagnoses, and for those of us who pursue normal blood sugars the future should be a lot brighter, no matter what our weight.

P.S. If your doctor tells you there's no point in losing weight because thin diabetics have even more heart attacks than fat ones, find a new doctor. I wouldn't have warned about this in the past, but over this past year I have been sent so many stories about doctors giving idiotic advice based on their having misunderstood a media report about a research study that I no longer assume doctors will even bother to read the full report before changing their practice recommendations.


Anna said...

Thank you for your insight. As one of many normal weight diabetics, I was naturally disturbed by this finding and the knee-jerk reaction indicating a faster demise for thin people. Simply being thin as a risk factor doesn't tell the story. What you say here seems plausible, however, I wish more would be done to try to truly understand the spectrum of this disease.

Jenny said...

Anna, I'm with you in wishing there was more pure research about diabetes, especially since there is a wide variety of underlying causes of diabetes among people who are overweight, too. Type 2 diabetes is not one condition, just a symptom, elevated blood sugar, with a multitude of underlying causes.

Sadly, most diabetes "research" is conducted in "rodent models" of diabetes which are fat rodents which have been bred to develop diabetes when fed "high fat diets." (These are actually high carb/high fat diets.) Their diabetes genes are not those found in humans, but they have been studied for so long there is an established research community full of people who have spent their lives studying these rodents, and they continue to absorb grant money and publish findings that are irrelevant to people with diabetes, though the research inevitably is reported as if the research had been done in people.

Meanwhile, only a tiny number of researchers look at the genes of diabetic people and ask what makes THEM diabetic. The answers that are emerging mostly point to genes that limit the ability to produce insulin. Only a very small number of genes associated with diabetes produce insulin resistance.

Darbro said...

I don't think the weight at diagnosis provides sufficient data regarding treatment, progression or outcomes. What is possibly a more salient issue is that higher BMI patients at diagnosis have the extra tool of ameliorating insulin resistance by losing weight. Not as efficient a tool for thinner patients. This point alone could explain the result.

Jenny said...

Dabro, I should have mentioned that most people with diabetes who are normal weight and thin are insulin sensitive. So reducing IR isn't an issue for them. Their problem is almost always that they are insulin deficient.

Rad Warrier said...
This comment has been removed by the author.
Rad Warrier said...

Thanks Jenny for your insightful analysis of this study.

I feel that conclusions of many studies are not understood in the right context, or misunderstood, or even misinterpreted by those who should know better. Take the case of the study that seemingly concludes that it is not beneficial for diabetics to attempt to lower their HbA1c. I believe those who were studied were relatively older diabetics with their diabetes in somewhat advanced stages. I think they were insulin users or those taking insulin stimulant medication. It looks like they increased the dosage of their medication to lower their BG and hence could also lower their A1c. Most likely it might have been the higher dose of medication (and consequently larger number of lows) that might have caused a relatively higher level of mortality among the subjects of the study. Instead of concluding that higher level of medication might not prove beneficial to such diabetics the apparent conclusion was that it was not beneficial for any diabetic to attempt to lower their A1c.

I remember my doctor congratulating me for maintaining my A1c in the 5.3 to 5.7 range for a long time. (My latest A1c is 5.6.) But she also added that I need not work so hard (frankly, I don’t work “hard” at all) to maintain my A1c at these levels because studies have revealed that it might not be beneficial for diabetics to lower their A1c below 7. She said this knowing very well that the only diabetes medication I use is Metformin 750 mg per day (she was the one who decided the dosage) and that I keep my BG and A1c in fairly normal range by exercise and portion control.

I also feel that those whom I would like to call “low carb extremists” are like these doctors – they take the observed benefits of carbohydrate restriction to out of context extents. There is no doubt that diabetics do need to reduce their carb consumption. How much they should reduce depends on how severe their diabetes is. Without realizing the context of carb reduction, these extremists insist that every diabetic (and some of these extremists would want everyone, diabetic or not) to limit their carb consumption to below 30 grams (or some similar number) per day. “Carbs = poison”, “Fructose = poison” are some of the mantras I encounter frequently in many diabetes or diet related e-forums :) There are members on many of these e-forums who claim to keep their carb consumption below 30g per day, yet have not very good BG levels or A1c. They seem to be insulin deficient, yet steadfastly refuse to use exogenous insulin. Looks like they yearn to consume the carby dishes of their pre-low carb days because they always discuss low carb recipes for dishes that mimic the carby dishes they left behind. Sometimes I feel like telling them to go have a life, use a little insulin and eat a little of what you like to eat.


Marana said...

This information could be very disturbing for people who are in normal weight and are affected by diabetes. Thanks for sharing this info, I hope more researches will be done on this aspect.

Click Here

John said...


I read your kind of on sabbatical for a while.

And I didn't know the appropriate way to ask a question. So perhaps this is a low risk way to do it.

Metformin seems to prevent releasing glucose from the Liver, and I thought that might be coming from turning Fatty Acids stored there into glucose. That sounds like a good thing.

But if a person has NAFLD or a build up from Sugar or Fructose deposition of fat in the Liver.. might Metformin prevent clearing the fat out of the Liver?

I was just watching a show on BBC that mentioned people Thin on the Outside and Thick on the Inside.. or something like that.

Meaning Due to unknown reasons some people store more fat in organs and the Liver than the outside and this could lead to organ damage.

I could have it all wrong. I thought the Liver stores Glucagon and turned that into glucose.. and then I'm left wondering if it stored fat in the Liver.. how does it ever get out of the Liver?

Anyway, back to my thought.

I am suggesting that Metformin treatment for people with fat in their Liver might prevent removing the fat and do harm.

Perhaps they should use whatever method possible to reduce or remove the fat in the Liver before starting Metformin treatment and then monitor it to prevent conflict.

Thanks for your really consistent interest in diabetes and sharing your findings.

I've bought both your books and found them very well written.

- John W.

Jenny said...


As I understand it, the storage of glycogen in the liver is a separate process from the storage of triglycerides that come into the liver in the blood stream. I don't think that metformin blocks the burning of that fat stored in the liver. Once glucose has been turned into fat, it can't be easily turned back into glycogen or glucose because the process of making fat strips out 9/10ths of the oxygen that had been in the carbohydrate. Fat is burned, as fat, in the mitchondria of the cells.

However, some people have defective mitochondria either from inherited genes or damage from chemicals or drugs like statins that block or limit mitochondrial activities. Those people then build up excess fat in their cells and livers.

It was believed for a while that metformin actually reversed fatty liver, as it reduces or eliminates the high liver enzymes that are used to diagnose fatty liver. However, biopsies found that this was not the case.

There are some claims that eating a ketogenic diet will cause the reduction of liver fat as it promotes fat burning. The data used to support this claim is not bullet proof and in any case it takes a very long time to see results. Those claiming that the liver fat decreases within weeks of starting a ketogenic diet are citing a single study that seems to me to be confusing the loss of glycogen with the loss of liver fat.

Any diet that causes fat to be burned (which usually means one that restricts calories below those needed) will over a long time reduce liver fat from what I can tell.

Metformin does seem to promote weight loss when combined with calorie restriction and a diet that normalizes blood sugars. So over long periods of time it might decrease liver fat, too.