July 1, 2009

ACCORD Redux: Low A1c Does Not Raise Risk of Death

Not a week goes by that I don't get at least one email from a reader whose doctor has warned them that it would be dangerous to lower their A1c to 6.5%.

I have blogged several times about why ignorant doctors are dispensing this dangerous advice. You can read past discussions HERE, HERE, and HERE.

ACCORD was the large study that supposedly found the correlation between lowering A1c and increased risk of heart attack. A very similar study, ADVANCE, found no such correlation, which given all we know about the impact of elevated blood sugars on heart health should have made doctors question the findings of ADVANCE more closely. But few seem to have done this.

Fortunately, a few have. Diabetes in Control recently reported on a presentation given at the recent ADA Scientific Sessions which found that further analysis of ACCORD data "did not confirm the proposed theory that low A1c levels might be the cause" of the elevated risk of death in the ACCORD patients who attempted to achieve tighter control.

Matthew C. Riddle, MD, Professor of Medicine, Oregon Health Science University and a member of the Glycemia Management Group of ACCORD, who was a site principal investigator for the ACCORD study is quoted as saying,
An A1c below 7% alone does not appear to explain the excess deaths in the ACCORD trial and is not necessarily a predictor of mortality risk...Further, the rate of one-year change in A1c showed that a greater decline in A1c was associated with a lower risk of death.
Translated into English this means that, as we all knew before this outlier study was published, lowering A1c below 7.0% improves your chance of living longer. It does not impede it.

I urge you to read the whole Diabetes in Control report. You'll find it here:

ADA - Low Glucose Levels Do Not Explain Excess Deaths in ACCORD Trial:A1c of less than 7% is not a predictor of mortality risk in Type 2 diabetes and the converse is true in that higher A1cs do increase the risk of mortality. For every 1% higher A1c level above 6%, the risk of death increases by 20%.

If your doctor is telling you that lowering your A1c is dangerous, print out this article for him or her.

Then find a new doctor. There is a huge amount of evidence proving that lowering A1c below 7.0% lowers the risk of kidney failure, blindness, and yes, heart attack.

Here's a recent metastudy published in May that confirms that "Intensive glycaemic control resulted in a 17% reduction in events of non-fatal myocardial infarction [i.e. heart attack], and a 15% reduction in events of coronary heart disease."

This study found that intensive control also has " no significant effect on events of stroke or all-cause mortality." While that might sound like bad news about tight control, since you'd hope it would improve lifespan, in fact, it is good news. It means that tight control does NOT raise the risk of death, contrary to the erroneous conclusion so many doctors are drawing from ACCORD.

Here is a link to that metastudy:

Effect of intensive control of glucose on cardiovascular outcomes and death in patients with diabetes mellitus: a meta-analysis of randomised controlled trials Kausik K Ray et al. The Lancet, Volume 373, Issue 9677, Pages 1765 - 1772, 23 May 2009 doi:10.1016/S0140-6736(09)60697-8

Personally, given the choice between dying of a heart attack at age 85 blind, on dialysis, and with my leg amputated, or dying of a heart attack at age 85 with normal vision, kidneys, and nerves, I know which choice I'd make--and to achieve that choice I will need to aggressively pursue normal blood sugars.

The only time that lowering A1c has not been shown to be effective in reducing diabetes-related mortality in a group as a whole is when the studies involve large groups of elderly victims of medical incompetence who have spent at least ten years with A1cs in the 10% range or higher. These are people whose organs have already been ravaged by uncontrolled high blood sugars, often to where irreversible damage has occurred. Researchers then load them up with a cocktail of drugs several of which, Avandia and Actos, are known to cause heart failure. And a very small rise in the incidence of death occurs.

Getting back to the Diabetes in Control report, note that once again when reporting mortality in ACCORD researchers resorted to using the magnifying statistical technique called "Risk." Pause for a moment and consider what statistical risk measures. If I have two groups of a thousand people, and one group experiences 2 deaths of a certain type while the other experiences 3 deaths, the risk of death in the second group is 50% higher, because the number 3 is 1 more, i.e. 50% higher than 2. But the actual difference in the incidence of death in each group is tiny and possibly not even statistically significant.

In ACCORD the increased incidence of death in the tight control group was slight--the actual number of excess deaths. It only looked scary because it was reported as a 20% higher risk. Chances are that your doctor who may have taken one statistics class 25 years ago is not even aware of the difference between the two statistical measures or how they inflate findings. Statins are also promoted heavily using "risk" amplified statistics rather than incidence statistics because citing "risk" puffs up results that otherwise are unimpressive.

In fact, one very interesting finding of several recent studies of diabetes and heart attack is that researchers report that they are seeing far fewer heart attack deaths in their study population of Diabetics than they expected. So much lower, in fact, that important studies like JPAD, a study that attempted to investigate the impact of taking aspirin on cardiovascular deaths among people with diabetes are being described as "statistically underpowered due to an unexpectedly low rate of clinical events." (Medscape report on JPAD).

This low incidence of heart attacks in a population that used to have a lot more of them may be the result of much better blood pressure control over recent decades or even, dare we say, of better blood sugar control as a result of the classic 1990s studies DCCT and UKPDS which proved that lowering A1c decreased complications.

There is nothing found in analysis of data from ACCORD or any other study that suggests anything but this: further lowering of blood sugar, over time, is likely to result in further lowering of heart attack incidence and better long term, complication-free survival.

All this suggests that there is one more risk factor that will impact your health as you grow older with diabetes: The competence of your doctor. The better your doctor understands how elevated blood sugars cause complications, the better will be the care you receive. A doctor who does not understand why normal blood sugars will produce normal health is as dangerous to your health as one who puts you on a diet of daily banana splits.



country mouse said...

how do you choose a good endo? mine is a 7% is good enough type. she does not support me on going sub 6% and only references the ACCORD study. last visit we spent more time arguing over treatment philosophy than treatments.

Jenny said...

Country Mouse,

You ask a very good question, and one that could easily be a series of blog posts in themselves.

Briefly, I'd suggest:

1. Ask your friends with well controlled diabetes who they see.

2. Look for doctors associated with large, high prestige teaching hospitals rather than small community hospitals.

3. Look for young doctors newer to practice who did their residencies within the past 5 years. The worst endos I've seen were trained 15 or 20 years ago and seem to have never learned anything since they ended their residencies. There are studies showing the younger the doctor, the better the care for this very reason.

jimpurdy1943@yahoo.com said...

"Then find a new doctor."

I've done that.

Several times.

I've been very disappointed in most of my health care professionals. And I'm having some great success with diet alone, which all my doctors said was impossible.

Jenny said...


I hear you. Sometimes the best we can do is find a doctor who doesn't get in our way. I have had to compromise by finding a doctor willing to write the prescriptions and lab tests I need, though it has often been up to me to figure out what they are.

That's a step up from the doctors who wouldn't prescribe strips because my blood sugars were "only" going up to 240 after meals and the doctors who refused to give me insulin since my A1c was in the high 5%s (only because I hadn't eaten more than 40 g of carbs a day in several years.)

Chris said...

Only last week I received a slap on this wrist: my blood works came out perfectly, "but that A1C is not good". 5.0 was just not acceptable. I did bring my Navigator data printout (avg BG 4.7mmol/l with a SD of 0.4 over a 5 dy period) but didn't even bother to shove it in her face. This has been my last endo visit...

Anonymous said...

My best friend is urging me to go to yet another endo. I do have a referral from my primary physician, and I really do need to make the appointment. But I'm tired of stupid doctors, who don't see anything wrong with my post meal blood sugar readings, who think that the fact that my A1c has gone up from 5% to 6% despite the fact that I am getting more strict with my food, is not a problem. My primary refused to refill my metformin because she didn't think my A1C was high enough to warrant it. I'm scared. I feel like I am fighting this alone. I think I will probably need to be on insulin soon, I wish I had a medical professional I could trust to help me with that. I don't know if I have the energy or money to keep going to doctors until I find one that has some common sense. If it weren't for Jenny and this site, I'd have given up a long time ago. All I can do right now is control what goes in my mouth. But that is better than not knowing anything at all, not having any control at all.

Anonymous said...


I finessed doctors unwilling to treat my diabetes/retinopathy because my numbers were not high enough. I went to Walmart and got Regular Insulin and syringes over the counter, and started using it A(very very low doses and lots of testing, until I figured what I needed). THEN I went to the new doctor, and as I was already on insulin I just let (?) him change me to Lantus. RobLL

Nicky said...

The research professors I talk to are well aware of the ACCORD/ADVANCE/A1c debate. My GP "knows" that a low A1c is bad, but he'll probably see the light eventually. But I've actually told my Dad to go for the banana split as an improvement over what the dietician told him - which was to eat a whole banana and some grapes as part of a weight-reducing diet! FUME...He's of low-normal weight, is 87, and his recent diabetes dx is probably linked to prednisolone. I bet the cream in the banana split, plus some protein from the nuts, would be a good thing - and might slow down the carbs. But back to your point - just how long is it going to take to get the reconfirmation that a low A1c is a good idea out to the bottom of the food chain? And how can we speed it up?

Trinkwasser said...

I had a recent rant about being refused tests by a receptionist. Having now seen my doctor I am even more aware that I have a keeper, and in fact most of the doctors and nurses here are excellent.

Well as excellent as the accountants permit. Apparently the computer has been locked down so only Registered Diabetics can be permitted an A1c, and to override that needs manual input from a doctor. Likewise to get a Full Lipid Panel instead of TChol must be typed in manually and the usual nurse or phlebotomist can no longer order this either.

"Big Brother!" my GP complained, and we actually had a very reasonable discussion about low carbing, testing, the ACCORD Effect etc. which proved beyond doubt that a doctor's medical judgements can be overridden by a clerk with no medical knowledge who has been fed biased information dervied from the likes of ACCORD, Farmer etc.

Some of them are every bit as frustrated as their patients at being so restricted. Somehow I don't see this new analysis being promoted anything like as fast as the previous one, or the recommendation NOT to reduce A1c below 6.5 to be reversed.

Unknown said...

They are all missing the point. Whether lower A1c is found to be protective or harmful depends on HOW you achieve the lowered levels.

High insulin levels are very damaging as are many of the drugs typically given to lower glucose. These will negatively overwhelm any benefit from lower A1c.

However, I suspect that lowering A1c through proper diet and exercise, along with good nutritional support (vitamins, certain minerals, omega3 oils, antioxidants), and avoiding harmful food products (such as artificial sweeteners, hydrogenated fats, and n-6 fats) will show improved health resulting from lower glucose levels.