Tuesday, December 29, 2015

Less glucose in the glucose intolerant is a good thing? Who'ulda thunk it.

In 2003, the JAMA released a paper from Chiasson, JL, et al, called: Acarbose treatment and the risk of cardiovascular disease and hypertension in patients with impaired glucose tolerance: the STOP-NIDDM trial.

(A little infographic-like sheet containing the major points can be found here, for a quick overview of the trial.)

Including some 1,400 people, the STOP-NIDDM trial is an international, randomized, double-blind, placebo-controlled clinical trial that, including follow-up, spanned a mean of 3.3 years. I have only briefly perused the methods and statistical analyses sections of the paper, so I have, what I would consider, only a superficial understanding of the methodology, thus far. So, take this with a grain of salt. That said, I am impressed with what I have seen, at least from the standpoint of the reporting of the data. (Any time study authors decide to include absolute risk reductions versus relative risk alone, for instance, I have to give props; and props are certainly due, on that front, here. Not to mention, all pertinent information, from what I can tell, was either offered up front, or links to such information were provided for us. Respect.)

Study question:

~ Is taking Acarbose associated with a meaningful decreased risk of cardiovascular events in persons with impaired glucose tolerance?*

*It should be noted that the original STOP-NIDDM trial, from which these data were taken, was, as the authors admit here, "not powered to answer that question." This is a secondary, retrospective analysis of the data, modified to tease out this potentially important confounding variable in order to see whether it had a significant part to play in the reduction of disease risk in the study population.

Inclusion criteria:

~ Age: 40-70 y/o,
~ BMI of 25-40,
~ Impaired glucose tolerance (IGT) (according to WHO criteria),
~ Fasting plasma glucose: 100-140 mg/dL (5.5-7.8 mmol/L)

Ultimately (after all reported dropouts), 1,368 people completed the trial and follow-up, which is pretty good. The outcome measures were kind of all over the place though, from my perspective. They claimed major cardiovascular events, including new angina, AMI, revascularization procedures, cardiovascular death, CHF, CVA, and peripheral vascular disease as primary outcomes. (Secondary outcomes were incident HTN and "rates of each type of cardiovascular event.") This does not constitute a "primary outcome." Albeit of course related, these are eight separate potential primary outcomes. For the love of God, just pick one primary outcome and study that to see if there is a legitimate association. (It is times like this that I can't help but think to myself, as cynical as it may seem, that this was intentional, as it might have been a means of using the problem of multiple comparisons to generate significant figures (to assure publication, particularly in a prestigious journal like the JAMA), where there are none. I will put that consideration on the back burner for now, I suppose...)

*Acarbose is an alpha-glucosidase inhibitor that prevents the cleavage of oligo- and polysaccharides into their single, monosaccharide sugars, during digestion.

The point of taking acarbose, clearly, is to prevent the absorption of glucose in those that do not tolerate glucose well. I follow the logic. However, it appears to beg the question: If someone does not tolerate dietary glucose, why feed them dietary glucose? Why feed someone carbohydrate just to turn around and try like hell to prevent its absorption into the extracellular fluid, because it might pose a problem in the carbohydrate intolerant individual? Does it not make sense simply to avoid ingesting said carbohydrate in the first place, and not bother taking the acarbose, at all?

I realize, I could be falling prey to a cognitive bias, here. "If a little of something is good, a lot more must be even better." However, there are other data that corroborate these findings in various ways. From a scientific perspective -- assuming these data are accurate, and I do not yet know whether or not I think they are -- I do find this intriguing, and, as I say, I think it jives rather well with much of the controlled nutritional literature I have seen on the topic. Personally, however, it strikes me as a touch illogical and futile. A little like trying to dig a massive hole using only the stick-end of your shovel.

The relative risk reduction in the acarbose group, versus those 600 some odd persons on placebo, was 49%. Note that this translated to a 2.5% absolute risk reduction. (At first glance, a 2.5% absolute risk reduction seems quite small, but, considering that many medications for cardiovascular disease translate to a much larger NNT and a much smaller absolute risk ratio, oftentimes <1%, I am actually somewhat pleased with that figure.) But, again, I just feel it would be more efficient and intelligent simply to say, "Okay, persons with IGT do not appear to tolerate glucose well, ergo, we shall see what restricting dietary glucose does," since, in the latter, they wouldn't have to play with inhibiting absorption, at all. It takes out the unnecessary middle-man.

One thing I think is important, and, in my opinion, is a huge confounder, is this:

a.) Almost invariably, dietary modification leads to a reduction in carbohydrates consumed, since a vast majority of the Calories eaten by the so-called "civilized" nations of the world come from this macronutrient category. Could these results have arisen out of a slight reduction in carbohydrate Calories?

b.) We know, without a doubt, that consistent exercise can improve both OGTT and insulin sensitivity. Why? Intense exercise turns the skeletal muscles into a glucose sink. (If you truly wanted to test the effectiveness of acarbose, alone, on glucose tolerance and disease risk, you would have to maintain the subjects' baseline lifestyle characteristics.)

A and B may just be study ruiners, for me. (Not necessarily for the STOP-NIDDM trial, as a whole, per se; but perhaps for these secondary, modified intention-to-treat analyses that followed.) At least, I cannot overlook the possibility that these confounders within the larger confounder couldn't have accounted for nearly all the 2.5% absolute reduction in disease outcomes.

All in all, I recommend that you check out the trial and read at least the methods and results sections, for yourself. Despite my personal annoyances, it really is an intriguing study, and you may be able to find some points of interest I have not alluded to. Enjoy!


1. Chiasson, J. L., Josse, R. G., Gomis, R., Hanefeld, M., Karasik, A., Laakso, M., & STOP-NIDDM Trial Research Group. (2003). Acarbose treatment and the risk of cardiovascular disease and hypertension in patients with impaired glucose tolerance: the STOP-NIDDM trial. Jama290(4), 486-494.

No comments:

Post a Comment