June 9, 2011

Avandia is sufficient to explain the excess deaths in ACCORD--the study being used to argue that normal blood sugars are dangerous.

If your doctor is telling you that research has "proved" that lowering blood sugars below an A1c of 6.5% is "dangerous" print out this letter, which was just published in the New England Journal of Medicine, and demand that your doctor read it:

Intensive Glucose Lowering and Cardiovascular Outcomes N Engl J Med 2011; 364:2263-2264 June 9, 2011

If you can't get the full letter to display here are the salient statements it contains. The quote refers to the ACCORD study where the intensive-therapy group was the group of people with Type 2 Diabetes who were urged to get their A1c down below 6.5%:
44.7% of subjects in the intensive-therapy group, as compared with 25.2% of patients in the standard-therapy group, received rosiglitazone [Avandia] just before the study transition date. According to the 2008 article summarizing the results of the ACCORD trial up to the point of protocol transition, 91.2% of subjects in the intensive-therapy group, as compared with 57.5% of patients in the standard-therapy group, had received rosiglitazone.
Although other differences in drug exposure warrant further analysis, we think that the authors[of the ACCORD publications] should consider (and address in a secondary analysis) the role of rosiglitazone in the excess deaths from cardiovascular causes, especially in the absence of biologic plausibility of a glucose-mediated effect. Given unbalanced exposure, we think that the ACCORD trial is inconclusive and that the recommendation to abandon lower glucose targets is not supported and has unknown consequences for the long-term management of diabetes. [Emphasis mine]
The signers of this letter are Almut G. Winterstein, Ph.D., University of Florida, Gainesville, FL, Susan R. Heckbert, M.D., Ph.D., University of Washington, Seattle, WA, and Morris Schambelan, M.D. University of California, San Francisco, CA.

This is a VERY important point. The ACCORD results have been used to justify doctors' and insurers' refusal to give people with Type 2 the medications and blood sugar meters they need to achieve normal blood sugars. Since it was published and interpreted as meaning that intensive efforts to your lower blood sugar will result in your having a higher risk of experiencing a heart attack I have heard from dozens of people whose doctors cautioned them to keep their A1c at 7.0% or higher and based their conclusions on this study.

This kind of toxic advice has been given to many thousands of people with Type 2 diabetes who will go on to suffer unnecessary diabetic complications that become inevitable at A1cs over 6.0% .

My guess is that the makers of Avandia had something to do with the fact that the health media blamed the excess heart attacks that occurred in the lower A1c group on their lower blood sugar rather than on these subjects' taking a drug that raises the risk of heart attacks.

It is odd, isn't it, that these journals did not make public in accessible form the numbers linking this exposure to Avandia with the excess deaths in ACCORD until now, after Avandia has been, for all practical purposes, taken off the market.

Coinicidental? I think not. Not with the power GlaxoSmithKline wields with its huge advertising budget and its research grants. No media outlet is going to risk losing those full two- and three- page drug company spreads just to save the limbs, vision and kidneys of a bunch of fat old people with diabetes.

Shame on the mainstream medical journals for not making this information public when they first knew of it.



Editor said...

Avandia is dangerous stuff. Five years ago I took it for two weeks, and it gave me tinnitus which continues to plague me. Before taking Avandia I had crystal-clear hearing, but now I hear hissing, steady-state tones, and sometimes chirping. I suppose that if it hadn't given me tinnitus, I would have kept taking it and gotten a heart attack; but it isn't any fun living my life with sound-effects in my head.

Scott S said...

I suspected this would be the case all along, but like many people, couldn't justify the cost of the medical journal publication to truly examine the data. The reality is the conclusions reached by this study were contrary to the boatload of data found in virtually every other study on this subject. At least the facts are finally starting to trickle out!

Anonymous said...

http://www.guardian.co.uk/society/2011/jun/24/low-calorie-diet-hope-cure-diabetes interesting...

Jenny said...


The "very low calorie" diet is, of course, a low carb diet. You can achieve the same effect on blood sugar eating a moderate calorie diet that has a very low level of carbs.

But the terror of fat keeps doctors from trying it. Sad. Because very low calorie diets make people crazy over time. Eating 1600 calories a day in the famous WWII starvation experiments made young men in their 20s psychotic.

Fortunately, moderate calorie very low carb diets do not!

Anonymous said...


I really value your contributions to my understanding of the disease. Do you have time to comment on the counterpost by the ACCORD authors? I think it means "you may be right, and we're looking into it"--but would love to know if I'm reading it right!

"The ACCORD trial was not designed to assess the effect of a particular drug or group of drugs on cardiovascular outcomes. In fact, because of the confounding between the characteristics of the participants and the types of medications chosen by the treating physicians, it is essentially impossible to isolate the effects of individual medications on outcomes. The suggestion by Winterstein et al. that these outcomes may have been due to the differential use of rosiglitazone or other drugs is certainly a possibility; however, this differential use would have to account for both the increased mortality and reduced nonfatal myocardial infarctions observed in the ACCORD trial. Moreover, preliminary analyses of the data have not identified any specific drug-related risk, and more detailed analyses are currently being written. Such analyses can only be considered exploratory because of the design of the ACCORD trial."

Jenny said...

I would translate that response to "We'd look like idiots if we admitted we hadn't noticed what you'd pointed out about our poor use of statistics and the real findings of our study, especially in view of our ties to the drug manufacturer."

Basically you had a poorly designed study here , with badly applied statistics, that was given a huge amount of press, changed practice recommendations, and probably got a lot of the people involved a lot of grant money.

So they're going to just dither until people move on, which sadly they will. Barely anyone has noticed these follow up analyses, and the doctors working in the field, with patients, continue to warn them how dangerous it is to lower the A1c below 7.0%.