This blog tracks updates to the Blood Sugar 101 Web site.

Wednesday, June 30, 2010

The ADA Scientific Sessions: Hype, Hype, and More Hype

This past week the movers and shakers (and funders and product sellers) got together for their annual dog and pony show, the American Diabetes Association Scientific Sessions. As is always the case, drug and device company-funded studies presented at these sessions hit the health headlines and received far more attention in the news than they deserve.

Why? Because the presentations made at the ADA Scientific Sessions are not peer reviewed. The visibility of this blog has risen to whee I have been getting PR releases intended for the health media. When I write back and ask the PR people if the studies they are touting will be published in a journal, the answer is often that no, no such publication is planned. This makes sense. Why bother, when the goal has already been achieved? The studies have been featured in the New York Times and on TV where the companies funding them can be certain no critical thought will be applied to their findings.

What most of the studies presented at the ADA Scientific Sessions that make it into the news have in common is a willingness to twist the truth to make a drug or device sound much better than it is.

For example, here's the headline about Viaject, a new formulation of R (regular human) insulin that is supposed to be absorbed faster than regular Regular. It was published in the online newsletter version of Endocrine Today. "Glycemic control similar with VIAject insulin vs. regular human insulin." The release goes on to claim that:
Patients with type 2 diabetes who took VIAject insulin, a fast-absorbed human insulin formulation, achieved similar glycemic control, but significantly better weight loss and a twofold reduction in the rate of hypoglycemia compared with patients who took regular human insulin
This sounds exciting--until you read a bit further and discover that "similar glycemic control" is an exaggerated claim and the two-fold reduction was between a very small number and an even smaller number of very mild hypos.

Here are the significant detailss:
Helena W. Rodbard, MD, and colleagues observed a –0.56% reduction in HbA1c with VIAject insulin (Biodel) vs. –0.70 with regular human insulin.
Keep in mind that the A1c change presented here is an average and that individuals vary so greatly in the extent to which their A1c maps to actual blood sugars.

Remember, too, that the A1c is the last statistic you'd want to look at when evaluating a fast acting insulin because it hides the range of variation of blood sugar. Two people can have the same A1c with blood sugars but one may have blood sugars that always stay between 95 mg/dl and 140 mg/dl white the other has blood sugars bouncing between 175 mg/dl and 60 mg/dl. The person who keeps blood sugar in the narrow range that stays under 140 mg/dl will have a very different complication risk than the person who spends hours over 140 mg/dl each day.

But that said, the A1c of the group taking Viaject was still higher than that of the people using regular insulin. And we have to assume that the average A1c was the best statistic they could come up with.

So when we read after this that "patients assigned to VIAject had an adjusted mean weight change of just 0.46 compared with 1.35 among patients assigned to regular human insulin." (no units given!) and that they had fewer mild hypos since, "The rate of non-severe hypoglycemia was significantly lower with VIAject insulin (0.33 events per month) compared with regular human insulin (0.66 events per month; P<.02)." a thinking person might ask, Is it possible they had slightly less weight gain and slightly less hypos because they were getting slightly less insulin as shown by their slightly higher blood sugars?"

Another study touting the advantages of a new inhaled insulin (view it here) sounds good until you discover they were comparing their combo of Lantus and inhaled insulin at meals against "biaspart" premix given twice a day. Biaspart is the 70/30 NPH/Novolog combo which is well known to give very poor results, since you can't titrate the Novolog in the premix to the carbohydrates in the meals the user eats. This 70/30 insulin is shot twice a day so users get a small amount of fast acting insulin that is not injected at meal times and only have the very unpredictable, hypo-causing NPH in their bodies to cover lunch. The only safe way to use 70/30 is to underdose, which explains why it produces a much higher A1c than any combo of basal and fast acting insulin used at meals. So basically this is an Apples and Oranges study. The inhaled insulin did better as a fast acting insulin than almost NO fast acting insulin would do, but that is a meaningless finding.

Comparing their inhaled insulin used with Lantus to Novolog injected at meals with Lantus would have come up with a more useful but probably damning, result. But the really important comparison to do with an inhaled insulin would have been to forget about the Lantus completely (and eliminate the possibility that the Lantus doses were not really, functionally the same) and compare the effect of the inhaled insulin used at meals against the injected fast acting insulin at meals in a crossover study where the same people used both insulins at different times.

That's not going to happen because inhaled insulin does not work predictably and if past formulations are anything to go by won't compare well head to head with fast acting insulin injected at meals.

Another study which received much play in the press was one that confirmed that Avandia appears to raise the risk of death from heart ailments compared to Actos. (See how the New York Times reported this here).

This is of course not news, but the problem with this study is that in it Avandia was compared to Actos so the reader is left with the impression that Actos is a safer drug that doesn't cause heart-related deaths.

What is left out is that Actos is known to cause heart failure in people who did not have it before they started the drug, and that neither Avandia nor Actos is as safe as taking neither drug. Comparing Avandia deaths and CV incidents, Actos deaths and CV incidents and metformin-only deaths and CV incidents would have been far more honest.

Beyond that none of the reporting mentioned the real problem with Actos: it shares with Avandia a mechanism (lowering blood sugar by stimulating the creation of new fat cells out of precursor cells that otherwise would have turned into new bone) that, over time, causes serious, probably irreversible, bone tissue loss that results in fractures as people get older. Actos, in short, is not a safe drug. But you wouldn't know that from the way this latest study was presented.

That's it for now. The main point to keep in mind is that presentations given at the ADA Scientific Sessions are not given even the anemic peer review that you see in the published studies. Many are never heard of again after the conference is over.

As far as I can tell, the main use of this huge conference is to give drug and device makers a way to fast track the poor-quality studies they've run to tout the effectiveness of their new moneymakers. And of course, to wine and dine the influential doctors and "thought leaders" the drug and device companies keep well supplied with research grants, "consulting fees" and other sweeteners to ensure doctors keep prescribing their drugs.

Monday, June 28, 2010

More CGMS Evidence Documents the Truly Normal Blood Sugar Range

Page changed: What is a Normal Blood Sugar?

A study of CGMS measurements taken in 74 normal people aged between 9 and 65 years old over a period of 3 to 7 days was published in June of 2010. It found the following:
Sensor glucose concentrations were 71-120 mg/dl for 91% of the day. Sensor values were less than or equal to 60 or >140 mg/dl for only 0.2% and 0.4% of the day, respectively
Overall only 5.6% of sensor readings were were over 140 mg/dl. (7.7 mmol/L) and these higher readings were more frequent in people under 25 years old. Only 4.4% of the readings of those over 45 were over 120 mg/dl.

Only .4% of all readings were over 140 mg/dl. But most significantly, this group was screened to ensure they had all off the following: A1Cs less than 6.0%, fasting blood glucose 70 to 99 mg/dl, 2-h oral glucose tolerance test (OGTT) levels below 140 mg/dl and no antibodies characteristic of autoimmune diabetes. After all these tests, all 17 people over age 45 who met the screening criteria had NO CGMS readings over 140 mg/dl at all.

This is probably because by the age of 45 people with the underlying genetic conditions that lead to diabetes, whose blood sugars would have been normal at younger ages, but who would have been getting higher than true normal readings after meals, would have progressed to where they failed the screening test. So it is a good bet that the people in the 45 and older age group in this study are truly, physiologically normal. With that in mind we are safe saying that normal people do not go over 140 mg/dl ever and are only rarely (4.4% of the time in this study) over 120 mg/dl--no matter what they eat.

Variation of Interstitial Glucose Measurements Assessed by Continuous Glucose Monitors in Healthy, Nondiabetic Individuals/ Juvenile Diabetes Research Foundation Continuous Glucose Monitoring Study Group. Diabetes Care June 2010 vol. 33 no. 6 1297-1299. doi: 10.2337/dc09-1971