The ACCORD study

Am attempting to increase my knowledge about the field of diabetes research by scanning the top journals for all their articles on beta cells and diabetes. Came across the ACCORD study (Action to Control Cardiovascular Risk in Diabetes), which concluded prematurely in 2008 because the treatment arm of the clinical trial ended up with greater numbers of deaths than the control arm.

The writeup in Science is found here: http://www.sciencemag.org/content/319/5865/884.full.pdf?sid=2b4ab30f-345c-4a91-8c91-537ac10f410b

This is the gist of it:
People with diabetes are at an increased risk of cardiovascular diseases and can die of heart attacks.
The hypothesis was: if you control diabetes, you should be able to reduce the number of fatalities due to heart attacks or other cardiovascular diseases (CVD for short).
Therefore, aggressive lowering of blood glucose levels should protect against CVD.

So to test this hypothesis out, they got about 5000 people in the treatment arm (so these guys were diabetics who were considered high risk for dying of CVD and who had their blood glucose lowered by aggressive treatment with a variety of drugs) and another 5000 people in the control arm (also diabetic, but no treatment, other than their usual regimen)

Then they found that there were greater numbers of people dying of heart attacks in the treatment group, compared to the control.

Uh oh.

Questions to be asked:
a) What was the blood glucose lowered to? How was this being measured? - They measured something called hemoglobin Ac, or HbAc, as a surrogate marker for glucose in the red blood cells

b) Is the HbAc measure a suitable surrogate to monitor blood glucose?
-Apparently so. Lots of people use it in clinics and hospitals.

c) Is there a relationship between HbAc levels and risk of cardiovascular disease?
- Very important point. And tellingly, NO! Nobody has shown that HbAc levels correlate with risk of CVD. Also, one may not have diabetes, but have an HbAc value of greater than 6.

So, it seems to me that firstly (and frankly, this is probably the most minor point) the wrong surrogate marker was used.


Secondly, what about the drugs used to lower the HbAc? Assuming that phamacologists went through the drug list and found that these drugs did not interact adversely with each other, might they not have side-effects that would increase CVD risk?

Thirdly, and this is a great point made by one of the researchers who was arguing against this hypothesis from the beginning: the population under trial have had diabetes for a long time, and are considered high risk for CVD. This means they probably already have plaques in their blood vessels. Lowering the blood glucose may loosen these plaques: apparently, plaques contain sugar and it is the sugar that thickens and strengthens the plaques. If you were to lower blood sugar drastically, these plaques may not stick around on the walls of the blood vessels anymore, but begin to float around and eventually get stuck in the heart, causing a massive heart attack.



Very interesting, no?

There is a possibility that things become worse before they come better. That is to say, that perhaps if they had continued the study for a few more years, and continued monitoring the people who got the aggressive treatment (and survived it), they would have seen better prognostic markers, (and indeed this has been shown in smaller studies), but who knows? For the safety of the people involved, the trial was stopped 18 months before it should have.

Perhaps a gradual decrease of blood glucose, instead of a drastic one, might have less numbers of fatalities?

More relevantly, how sustainable is this approach? I mean, assume that this trial really had showed that aggressive treatment lowers risk of CVD. So what? Wouldn't this person have to take those medicines for the rest of his life anyway? And who is to say that those drugs in turn do not cause some other kind of damage- liver failure or renal issues? What then?
From the little I know about this issue, it seems to be that the long term risks outweigh the benefits.

So I would have to say that it's a good thing that trial showed what it did and forced everybody to think a lot harder about what that glucose in the blood is really doing.