Friday, November 17, 2006

Fatally Flawed Health & Risk Paradigms: Part 1

Hardly a day goes by without the media hounding us to lose weight, lower our cholesterol and keep our blood pressure in check! Why? Conventional wisdom says if we can modify these things - make lifestyle changes - we can modify our risk for chronic health problems like cardiovascular disease, stroke, cancer and diabetes.

A sampling of today's headlines include:

Nurturing Students' Healthy Lifestyles a Priority for UCF

Nurturing Students' Healthy Lifestyles a Priority for UCF[Preeti] Wilkhu [UFC Dietitian] added that “high blood pressure, high cholesterol and diabetes don’t know if you’re healthy, or unhealthy, skinny or fat, white or black, or male or female. Just because you are skinny, it doesn’t give you the right to eat all the junk food you want.”

Group hears about heart disease

"Although genetics and age can't be controlled, women should address risk factors such as smoking, high cholesterol and high blood pressure. A woman should eat a balanced diet and maintain a healthy weight. She should exercise regularly to stay in shape. If a woman has diabetes, doctor visits are important, too," said [Beth] Close.

Eat Your Way to a Healthy Heart

According to the American Heart Association, a healthy diet can help alleviate three major risk factors for heart disease: high blood cholesterol, high blood pressure, and excess body weight.

Not a day goes by without numerous articles and segments on the news to remind us of the message we must make lifestyle modifications - diet and exercise - to reduce our risk of disease. A noble undertaking, no doubt; what's wrong with helping people help themselves, right?

But, what if the message is flawed?

Clearly the experts believe more than adequate evidence supports their recommendations. They tell us that multiple studies, of multiple type, across multiple research methods point to cholesterol, blood pressure and weight. And, I'll even tell you they often do.

The only problem is that consistent and convincing weight of the evidence remains lacking, much to the dismay of those who continue to preach a low-fat diet, reducing calories and restrict fat, cholesterol and sodium intake. The other problem is that a large and growing body of evidence implicates something else is more critical in our long-term health - our blood sugar levels; more specifically our blood sugar trends over time, most easily measured in our percentage of Hemaglobin A1c.

Earlier this week I touched on the findings that show how critical blood sugar levels are in our long-term health. Today, let's take a look at data that points directly to the relationship between HbA1c and coronary heart disease, cardiovascular disease and all-cause mortality.

Hemaglobin A1c is the measure of the percentage of red blood cells that are "damaged" by gycation in our blood. It is formed in a non-enzymatic pathway by hemoglobin's normal exposure to high plasma levels of glucose. The higher your HbA1c, the more your red blood cells are carrying around glycated hemoglobin. That's because once a hemoglobin molecule is glycated, it remains that way.

A buildup of glycated hemoglobin within the red cell reflects the average level of glucose to which the cell has been exposed during its life cycle. The normal range is said to be between 4% to 5.9%. While diabetic patient treatment goals vary, many include a target range of HbA1c values. A diabetic with good glucose control has a HbA1c level that is close to or within the reference range of normal. The International Diabetes Federation and American College of Endocrinology recommends HbA1c values below 6.5%, while the range recommended by the American Diabetes Association extends to 7%.

A very high HbA1c represents poor glucose control in those with diabetes.

But, what about those who are not diabetic - what does the data tell us about the vast majority in the United States who do not have diabetes?

First, let's lay out what the HbA1c values mean. When HbA1c is found to be at the following percentage, the average blood sugar level is: **

5% = 100mg/dL
6% = 135mg/dL
7% = 170mg/dL
8% = 205mg/dL
9% = 240mg/dL
10% = 275mg/dL
11% = 310mg/dL
12% = 345mg/dL

This average is the full, 24-hour average, over a period of 120-days with the lows and highs associated with waking in the fasting state, eating, and expected post-prandial rise. So if your HbA1c is 7%, that basically means your blood sugars averaged - lows to highs - 150mg/dL over the last 3-months.

Now let's take a look at a study published two years ago.

In it, researchers investigating risks for cancer stumbled upon a critically important finding. In their research to explore the relationship between blood sugars and cancer growth, they found a linear relationship between HbA1c and CHD, CVD and all-cause mortality in their study of 10,232 men and women over a period of six years. The statistically significant finding was published in the Annals of Internal Medicine in the paper Association of Hemoglobin A1c with Cardiovascular Disease and Mortality in Adults: The Eurpean Prospective Investigation into Cancer in Norfolk. (PDF)

The finding was an eye-opener - even at "normal" levels of HbA1c, risk started a steady linear rise at 4.8%. Those with an HbA1c of less than 5% had a very low risk for CHD, CVD and all-cause mortality. But, at each incremental increase in HbA1c, the risk of heart disease and death increased significantly. Keep in mind "normal" is considered less than 6%, and this study found increased risk once HbA1c rose above 4.8%.

The researchers published their data in an easy to understand, gender segmented set of tables. In it they reported that the:

Percentage of men with CHD events at, with relative risk at:
  • Less than 5% = 3.8% (95% CI; RR = 1)
  • 5% to 5.4% = 6.4% (RR 1.56 [1.09-2.24])
  • 5.5% to 5.9% = 8.7% (RR 2.00 [1.39-2.88])
  • 6% to 6.4% = 10.2% (RR 2.13 [1.35-3.35])
  • 6.5% to 6.9% = 14% (RR 3.34 [1.78-6.63])
  • 7% or higher = 28.4% (RR 7.07 [3.96-12.62])

Percentage of men with CVD events at, with relative risk at:

  • Less than 5% = 6.7% (95% CI; RR = 1)
  • 5% to 5.4% = 9% (RR 1.23 [0.92-1.64])
  • 5.5% to 5.9% = 12.1% (RR 1.56 [1.16-2.09])
  • 6% to 6.4% = 15.2% (RR 1.79 [1.24-2.60])
  • 6.5% to 6.9% = 25% (RR 3.03 [1.73-5.31])
  • 7% or higher = 34.8% (RR 5.01 [2.95-8.51])

All-cause mortality (death within the six years of the study) in men at:

  • Less than 5% = 3.8% (95% CI; RR = 1)
  • 5% to 5.4% = 5.5% (RR 1.25 [0.88-1.82])
  • 5.5% to 5.9% = 7.5% (RR 1.57 [1.08-2.29])
  • 6% to 6.4% = 9.9% (RR 1.80 [1.13-2.86])
  • 6.5% to 6.9% = 19% (RR 3.49 [1.83-6.66])
  • 7% or higher = 18.5% (RR 3.38 [1.74-6.53])

Percentage of women with CHD events at:

  • Less than 5% = 1.7% (95% CI; RR = 1)
  • 5% to 5.4% = 2.1% (RR 0.96 [0.58-1.59])
  • 5.5% to 5.9% = 3% (RR 1.04 [0.62-1.63])
  • 6% to 6.4% = 7.3% (RR 2.29 [1.34-3.96])
  • 6.5% to 6.9% = 9.6% (RR 3.06 [1.25-7.49])
  • 7% or higher = 16.2% (RR 4.73 [2.16-10.34])

Percentage of women with CVD events at:

  • Less than 5% = 3.3% (95% CI; RR = 1)
  • 5% to 5.4% = 3.8% (RR 0.89 [0.62-1.29])
  • 5.5% to 5.9% = 5.4% (RR 0.98 [0.68-1.29])
  • 6% to 6.4% = 9.8% (RR 1.63 [1.05-2.52])
  • 6.5% to 6.9% = 13.7% (RR 2.37 [1.13-2.52])
  • 7% or higher = 36.8% (RR 7.96 [4.38-14.5])

All-cause mortality (death within the six years of the study) in women at:

  • Less than 5% = 2% (95% CI; RR = 1)
  • 5% to 5.4% = 2.7% (RR 1.02 [0.65-1.60])
  • 5.5% to 5.9% = 4.4% (RR 1.28 [0.82-2.01])
  • 6% to 6.4% = 6.4% (RR 1.61 [0.94-2.75])
  • 6.5% to 6.9% = 6.8% (RR 1.70 [0.63-4.60])
  • 7% or higher = 25% (RR 6.91 [3.50-13.67])

Relative risk tells us the likelihood of adverse outcomes - when the RR is less than 1, adverse outcomes are less likely; when they are greater than 1, adverse outcomes are more likely. So, when the RR range remains above the 1, ie. 1.23-1.99, it is a statistically significant risk for all within the group because in the group everyone had between a 23% to 99% higher incidence of an adverse event when compared to the lowest risk group.

Go back and look at those numbers again!

This study was not the only one to find blood sugars and HbA1c to be critically important either.

Last year a paper, Glycemic Control and Coronary Heart Disease Risk in persons with and without diabetes, was published in the Archives of Internal Medicine. The data from it is almost identical to the findings in 2004. The researchers concluded "In nondiabetic adults, HbA1c level was not related to CHD risk below a level of 4.6% but was significantly related to risk above that level (P = greater than 0.001). In diabetic adults, the risk of CHD increased throughout the range of HbA1c levels. In the adjusted model, the RR of CHD for a 1–percentage point increase in HbA1c level was 2.36 (95% CI, 1.43-3.90) in persons without diabetes but with an HbA1c level greater than 4.6%."

But instead of educating the population about the dangers of elevated, even within the normal range, blood sugars, the crusade continues to beat the drum to reduce cholesterol. You would never know it from the statin ads, but half of all people who have heart attacks have no known risk factors - translated - they have completely normal cholesterol that isn't alerting their healthcare team to impending doom. Add to this fact, the oft-cited Framingham Study couldn't find the ever important connection between LDL - the popular target for reduction - and heart attack either. In fact the Framingham data clued us in that it isn't total cholesterol, LDL, or HDL alone, but the ratio of total cholesterol to HDL and triglycerides that matter most if we're going to focus on cholesterol as a target.

In addition to the above cited studies finding blood sugars and/or HbA1c significant for risk, we have clear evidence that higher than optimal blood sugars are deadly over the long-term; that non-diabetic hyperglycemia is a risk factor for cardiovascular disease; and that there is a linear relationship between blood glucose levels and coronary mortality over the long-term.

And yet, the dietary advice we're given specifically increases the potential for chronically higher than optimal blood sugars - all because the dietary dogma is based on the assumption that dietary fat and cholesterol raise the risk of cardiovascular disease.

The fatal flaw in the dietary recommendations is that it's not the fat - it's the excessive carbohydrate and sugar in our diets that is causing our chronic, degenerative diseases.

The fatal flaw is that we're specifically recommended a dietary pattern that increases the risk of higher than optimal blood sugars - advising the population that such a diet is going to reduce their risk, when it is increasing their risk!

Come back Monday, as Part II of this article will continue...

**Correction from original post: I inadvertantly cut & paste incorrect "HbA1c - blood glucose values" from a chart I compiled from various sources as I researched to provide accurate information. The revised data comes from consensus of two sources - LifeScan (OneTouch glucose meters) and the University of California San Franscisco.

14 comments:

  1. Fascinating stuff. I have a quibble with one item.

    "when the RR range falls below 1 and above 1, i.e. 0.65-1.56, it really isn't all that much of a risk overall, but more a risk on an individual level because 35% of those within the group actually had a lower risk and 56% had an increased risk; when the RR range remains above the 1"

    Everything that I've seen indicates that if the confidence level includes 1, you can't be certain that there is an increased or decreased risk.

    Also, why isn’t the HBA1C test used as a diabetes screening tool, rather than the fasting blood sugar which gives no real indication of impaired insulin response?

    ReplyDelete
  2. >>>>Everything that I've seen indicates that if the confidence level includes 1, you can't be certain that there is an increased or decreased risk.<<<<

    Maybe my explanation was fuzzy?

    "when the RR range falls below 1 and above 1, ie. 0.65-1.56, it really isn't all that much of a risk overall, but more a risk on an individual level because 35% of those within the group actually had a lower risk and 56% had an increased risk;"

    For some within the group the risk isn't as high, it's actualy lower - basically they're "protected" in some way from whatever it is being investigated as potentially causing an increased risk, while at the same time others within the group are definitely at an increased risk, whatever it is being investigated caused them the consequence in question.

    In, for example, all-cause mortality in women....

    HbA1c of 5-5.9 was directly associated with death in some but not all within that group...the RR was 1.28, which on its own says everyone in the group had a 28% higher risk of dying than the null-hypothesis group with an RR set at 1; looking at the range though, it's 0.82-2.01 - which means for some, even at that level of HbA1c they were 18% less likely to die from all causes because of the HbA1c level than those in the null-hypothesis group....but for some, they were 201% more likely to die at that HbA1c level.

    Does that make sense?

    >>>>>Also, why isn’t the HBA1C test used as a diabetes screening tool, rather than the fasting blood sugar which gives no real indication of impaired insulin response?<<<<<

    I honestly don't know.

    ReplyDelete
  3. The problem with using A1c as a screening tool is that doctors are trained to think anything under 7% is normal. I've seen several posts on the boards lately from people who were told they not diabetic based on A1c alone, when theirs ranged from 6-6.9! Clearly diabetic! This is why I always explain to people who think they may be diabetic that the fasting test can only indicate a positive diabetes result, despite what doctors commonly say. Instead I recommend the Glucose Tolerance Test. Insisting on that is what caught my diabetes early.

    ReplyDelete
  4. I'd be interested to know where you got your correlations of A1C to average blood glucose. The charts I have seen show 5% = 100 mg/dL and 7% = 170 mg/dl. It wouldn't surprise me if this isn't an exact science and there is some disagreement.

    Good analysis and right on!! I followed the standard "healthy" dietary advice for several years and just got fatter and ended up with diabetes. I'm looking forward to part 2.

    ReplyDelete
  5. >>>>I'd be interested to know where you got your correlations of A1C to average blood glucose. The charts I have seen show 5% = 100 mg/dL and 7% = 170 mg/dl. It wouldn't surprise me if this isn't an exact science and there is some disagreement.<<<<

    That would be a BIG WHOOPS! on my part...as I researched about a dozen different sources to see what different sites said about HbA1c, I compiled a quick chart to compare and see what was what....and when I cut & paste the info, I grabbed the wrong column of numbers for my post!

    Thanks for the question - it caught a big error that I've now updated and I noted the correction and included what sources I used for the information too!

    ReplyDelete
  6. This comment has been removed by a blog administrator.

    ReplyDelete
  7. This comment has been removed by a blog administrator.

    ReplyDelete
  8. Note: two comments deleted; revision to text above to clear confusion around relative risk.

    ReplyDelete
  9. Len Kearney9:33 PM

    "That's because once a hemoglobin molecule is glycated, it remains that way."

    forever? can this be reversed?

    ReplyDelete
  10. Anonymous12:47 AM

    Hemoglobin dies off and is replaced constantly. They average about 120 days.

    ReplyDelete
  11. I wonder if it is me or you that misunderstands what confidence intervals means. :)
    A RR of 1.28 with a 95% CI of 0.82-2.01 means that there is a 95% probability that the TRUE RR falls in the range of 0.82-2.01. Also there is a 5% chance that it is actually higher or lower.
    So in this case the risk is more likely to be increased, but this could be all due to chance and it could also be decreased.
    It does not necessarily mean that the risk is decreased for some and increased for others...

    ReplyDelete
  12. Andreas - the post was corrected last week....sorry about any confusion!

    ReplyDelete
  13. Reading through the paper "Association of Hemoglobin A1c with Cardiovascular Disease and
    Mortality in Adults: The European Prospective Investigation into Cancer in Norfolk", I was most interested in the mortality data. After all, death is the most important and conclusive endpoint.

    In men, 24% (76 of 321) of deaths were attributed to heart disease and 29% (117 of 321) were attributed to cardiovascular disease.

    In women, 18% (36 of 200) of deaths were attributed to heart disease and 35% (70 of 200) were attributed to cardiovascular causes.

    I'm unclear whether the heart disease deaths are lumped in with the cardiovascular. However it's interpreted, only about a third to half the deaths are accounted for. Was there another most common cause of death, e.g. cancer that may be related in some currently unknown way to HBA1c levels?

    P.S. Thanks to Andreas for an excellent presentation of relative risk.

    ReplyDelete
  14. Anonymous5:23 PM

    I want to answer the question of why we use fasting glucose readings instead of HgA1C readings as a diabetes diagnostic tool. It's very simple: HgA1C is *considerably* more expensive than FBG.

    ReplyDelete