September 15, 2009

Shameful Research: Poorly Conceived Metastudy May Cost You Your Feet and Kidneys

What do you do when you are too poor or lazy to do real research to answer the most important question raised this year about the treatment of Type 2 Diabetes but want to see your name on a journal paper? The answer is obvious: A metastudy.

I have blogged several times in the past about the flaws in several studies that are being used to argue that it is dangerous to lower A1c. You can read those posts HERE, HERE, and HERE.

Now the Annals of Internal Medicine has published a study that will reinforce the message doctors have already taken to heart that there is no point in lowering A1c.

Systematic Review: Glucose Control and Cardiovascular Disease in Type 2 Diabetes Tanika N. Kelly et al. Ann Int Med 15 September 2009, Volume 151 Issue 6, Pages 394-403, doi

This study throws together the summary data compiled from the three most recent large studies of what happens when you lower A1c. By using only summary data, they ignore all the critical details that were essential to understanding those studies' outcomes--like how severe the patients' heart disease was before they started treatment or what drugs they used to control blood sugar.

With this dubious methodology the researchers conclude:
Intensive glucose control reduced the risk for some cardiovascular disease outcomes (such as nonfatal myocardial infarction), did not reduce the risk for cardiovascular death or all-cause mortality, and increased the risk for severe hypoglycemia.
The irresponsibility of publishing this study in a high impact journal leaves me gasping.

What's so wrong with this study that it should never have been published?

1. Its conclusion that tight control doesn't prevent cardiac death even though it lowers the incidence of heart disease ignored the hugely important fact, reported elsewhere that cardiac deaths among people with diabetes have dropped dramatically--to the point where some researchers are having trouble getting statistically significant results from their cardiac death research because not enough people are dying of the expected heart attacks.

Since this study did not differentiate outcome by age, the 85 year old person with Type 2 diabetes who dies of a heart attack is not distinguished from the 50 year old person who dies of a heart attack, and the conclusion that heart disease is decreased but not completely eliminated loses all meaning.

2. It ignores that these same studies all found that lowering A1c below 6.5% greatly reduced the incidence of neuropathy and kidney disease. I don't know about you, but while I'm waiting for my fatal heart attack (at age 90) I'd like to do it with both limbs intact and a set of functioning kidneys.

3. This metastudy completely ignores demographics of who was treated and what their condition was before blood sugar lowing treatment was initiated. For example, it claims that lowering A1c causes more instances of severe hypoglycemia. It doesn't, however, explain that the hypos were most likely to occur in older people whose diabetes was poorly controlled for decades before they attempted tight control.

Other research has found that older people are more prone to hypo even when they are taking no glucose-lowering medications--this is a byproduct of autonomic neuropathy, a long-term complication of poorly controlled diabetes.

How do you avoid developing autonomic neuropathy? By keeping your blood sugar under 140 mg/dl at all times which produces, among other things a 5% A1c--and a greatly reduced incidence of neuropathy of all kinds.

It also fails to note that the studies that found slightly poorer heart disease outcomes in the tight control group enrolled older, sicker patients who had long histories of exposure to extremely high blood sugars--people whose heart disease may have become irreversible long before tight control was attempted (but who still showed better kidney and nerve outcomes). Then these studies gave them drugs linked to heart problems: sulfonylurea drugs or Avandia and Actos.

4.There is no differentiation in this study between those who lower A1c with drug cocktails including sulfonylureas--known to raise heart attack risk--and Avandia and Actos--known to increase the incidence of heart failure--and those who lower A1c by cutting out the carbohydrates that raise blood sugar. Cutting out carbohydrates will not cause serious hypos and there is not a scintilla of evidence that it raises heart attack risk, even though the enemies of low carb dieting have invested a fortune in trying to find evidence that it does. You can read more about the safety and efficacy of the low carbohydrate diet HERE.

It is not paranoia on my part to fear that this study will cause even more amputation and kidney failure. I hear weekly from patients who have just returned from visiting their physicians 20 lbs lighter, with A1cs in the 5% range, people who have achieved these dramatic results simply by cutting down on carbohydrates, whose doctors respond to their achievements with alarm, and urge them to raise their A1cs because lowered A1cs "cause heart attacks."

If I had only heard one or two reports along these lines, I'd assume they came from those doctors who graduated in the bottom of their medical school class or from elderly doctors whose minds are going, but this is not the case. I'm hearing too many of these stories.

Busy family doctors get their medical news from newsletters that reduce the already tiny amount of information you read in an abstract to a single sentence or two. The sentence this latest study turns into is "Lowering A1c does not benefit patients with Type 2 Diabetes."

Since insurers are desperate for rationales to use to limit what treatments they will pay for, a study like this is a huge gift to those who want to prevent anyone with an A1c under 8% from receiving any diabetic drug.

Meanwhile, tens of thousands of people with Type 2 Diabetes are being told by their doctors to keep their A1cs in what they erroneously believe is a safe range--7-8% and to avoid the "danger" of lower A1cs.

We'll see the impact of this in a decade when the explosion in the incidence of amputations and kidney failure shows what really happens when you teach patients that A1cs over 7% are preferable to those under 6%.

If your doctor advises you that lowering A1cs is dangerous, tell them it is time they read the actual studies being used to draw this conclusion and pay attention to the details. Point out that for younger people recently diagnosed with Type 2 diabetes who do not have a history of years of poor control there is no evidence that tight control poses any problem. Note that negative outcomes from "tight control" were prevalent among people treated by Veterans hospitals, who were old enough that they had survived decades of very poor treatment and very high A1cs before they were put on the drug cocktails made up of drugs known to cause heart problems. Point out that even in these studies nerves and kidneys benefited from tight control. Finally, note that none of these studies track the cardiac outcomes for people with Type 2 diabetes who lower their carbohydrate intake and avoid Sulfonylureas, Actos and Avandia.

Then find a new doctor, preferably one who knows something about how to treat diabetes. They are out there. I do hear from people whose doctors support their progress enthusiastically and even from people whose doctors suggest they cut way back on their carbs.


RLL said...

Jenny - I don't think that I told you this before. A favorite cousin, 72 and diabetic was having a series of mild strokes and heart attacks (as if such a thing itself could be mild). I emailed and asked permission to offer some advice, and got an OK. I gave him the very short version of Bernstein and Jenny, told him to get both books, and meantime use your web site. It was like a light switch. Instant stop to the series. He is a 5%er. RobLL

Peter said...

Bob Michell taught me physiology. He is more of a scientist than a clinician, and it shows! His best quote ever:

The meta analysis of dross is still dross


John said...

One may wonder how "poorly conceived" the
so-called study is - If the conception is
to supply "evidence-based" justification for the institutional status quo it makes
perfect sense...


Anne said...

Found this through Diabetes Update. Says tight control works and prevents complications. The ones in the "tight control" group had an average A1C of 7.4 and they were comparing this group to a group with A1c's of 9. "After 30 years of diabetes, the cumulative incidences of proliferative retinopathy, nephropathy, and cardiovascular disease were 50%, 25%, and 14%, respectively, in the DCCT conventional treatment group, and 47%, 17%, and 14%, respectively, in the EDC cohort. The DCCT intensive therapy group had substantially lower cumulative incidences (21%, 9%, and 9%) and fewer than 1% became blind, required kidney replacement, or had an amputation because of diabetes during that time." Here is the abstract.

Is anyone following people who are eating low carb and in the 5% range?

trinkwasser said...

"Is anyone following people who are eating low carb and in the 5% range?"

In a word, no. Most such subjects are routinely turned down for studies.


Anyone would think there was a conspiracy to remove the carb-intolerant from the population as cheaply as possible.


Carbs-are-good (for our profits)


Hmmm, my tinfoil hat must be leaking, the subliminal messages have stopped being subliminal

William Houser said...

Facts are bothersome things I know. But how you could come to the conclusion that lowering you A1c levels to the near normal levels was not a good thing is bad. But to say it is a detriment is nearly immoral. This low-fat high-carb religion seems immune to facts but not to money from drug and carbohydrate producers.