DIABETES and HEART DISEASE
GLOBAL ANALYSIS: An ILLUSTRATION
Men and women with any form of diabetes long have been known to have a substantially increased risk of heart disease. But a review of the risks for diabetes revealed a disturbingly wide range in values found from the dozens of available studies. A meta-analysis of 10 studies found an average risk ratio of 1.85 for men and a higher 2.58 risk ratio for women. (Diabetes Care 2000, 23:962) Yet the sophisticated study of 51,000 professional men found a far higher risk for diabetes of 3.84 times, and the large Nurses study of 121,000 women found a ratio of 3.39, a bit lower value than that for the professional men. Risk ratios for women of up to a ratio of 11.9 were found for women that had the disease for many years. Questions become “Why are results from the various research studies so different?” “By what mechanisms does diabetes increase the risk of heart disease?” No quantitative explanation of these questions becomes apparent from reading any or all the papers published.
There is a serious need for answering these questions. First, the widely varying and confusing risk measurements from different studies raise a question as to a proper risk value for diabetes. An average risk from such highly differing risks as from a meta analysis appears meaningless. Second, knowledge about the key mechanisms that cause a disease is a usual requirement for designing methods to reduce or eliminate its risk. Possible answers to the above questions do become possible from a Biochemical Engineering Global Analysis.
The two major causes of coronary heart disease are first the atherosclerosis that gradually clogs the arteries of a population during life, and second, the thrombosis or clotting of blood that promotes the blocking of arteries. A number of BioChemical research studies have confirmed that diabetes increases the rate of atherosclerosis, and others have shown that it also increases thrombosis. Atherosclerosis in a population takes place gradually over long time, whereas a clotting rate can change near immediately in response to anti or pro-clotting agents. As shown in the discussion on atherosclerosis (See Atherosclerosis - a Chemical Process) the deposits in arteries of average men accumulate typically at a rate of about 6% per year. Thus if atherosclerosis reaches 10% at age 30, we can compute that it will likely progress to 18% by age 40 and 32% by age 50.
The BioChemical processes that determine atherosclerosis usually operate to change its rate of accumulation. Suppose a selected group has increased cholesterol, diabetes or some other factor starting at age 30 that increases this rate of accumulation from a more usual 6% to 9% per year. Thus at a duration of one year or at age 31 the atherosclerosis of the average group will be 10% times 1.06 or 10.6%, and that of the selected group will be 10% times 1.09 or 10.9%. Health researchers usually determine risk values of different people at an age. At a first year of exposure to this new factor and at age 31 the risk ratio of the selected group will be 10.9/10.6 or 1.03. But after 10 and 20 years at ages 40 and 50 the selected group will have from their increased rate of accumulation about 24% and 56% atherosclerosis respectively. The amount of atherosclerosis vs. average then will be 24%/18% or 1.33 after 10 years and age 40 and 56/32 or 1.75 after 20 years at age 50. Thus a measured ratio for extent of atherosclerosis will be duration dependent. This extent of atherosclerosis will affect risk of cardiovascular diseases similarly.
The effect of diabetes on heart disease results in part from its effect on producing atherosclerosis that narrows arteries, and in part from thrombosis that increases the likelihood of a clot that will block them. But unlike atherosclerosis that is substantially duration dependent, the effect of blood clotting or thrombosis develops near immediately in response to an anti-clotting agent. This type of a factor produces an immediate change in risk that continues at the same level as the years proceed.
Thus to gain insight on how diabetes produces its risk, we need to know the extent to which its risk is duration dependent, and the extent to which its risk is duration independent. The standard chemical engineering method for analyzing a problem such as this is simply to plot the risk of the disease vs. the duration of exposure of the expected causative factor. Because rates of disease usually progress exponentially, it is usual to plot the logarithm of risk vs. duration of exposure as accomplished by using semi-logarithmic graph paper. If this plotted line is straight, this confirms that the expected exponential effect is true. If the plot has zero and zero as an origin, this confirms that the cause is directly duration related. If the plot is a straight line, but does not go to the zero-zero origin, this shows that there is a duration effect, but some other factor probably operating in short time also is involved.
To apply this method to the study of diabetes and heart disease we need good data on the risk of disease vs. its duration, and a plot of this risk vs. duration on semi-logarithmic coordinates. If the plot is duration related at significance and points to a zero origin this will suggest that diabetes is directly and entirely related to progress of atherosclerosis. It there is an intercept not at but above zero, it will identify an additional effect related probably related to thrombosis.
Most research relating diabetes to risk of heart disease failed to examine this key effect of diabetes duration. But fortunately the two largest studies did include tabulations of risk vs. duration of the disease for both men and women. (Hu FB Arch Intern Med 161:1717; Cho, E J Am Coll Cardiol 40:954). These results are shown in the Figure “Diabetes” accompanying. The plots are striking. The risk for both men and women increases directly and exponentially with duration of time of exposure. For men the risk increases from 1.5 to 3.6 over 30 years. For women the risk increases from 2.3 to 12. For men the risk increases by 2.8% for each added year of diabetes. For women this risk increases 5.5% per year of the disease. These trends are both obvious from the plots and of very high significance. We now have an explanation of the widely differing risk ratios produces by the different studies. Consider some the implications of this analysis that extend beyond the the information on diabetes per se.
Figure “Diabetes”
The effect of duration of diabetes on risk of disease for both men and women is exactly what would be expected from duration related increases the rate of atherosclerosis. Thus this identifies not just a statistical relationship but is a confirmation of a likely BioChemical relationship that is of much broader significance. Each annual exposure to diabetes should increase the risk of disease by an additional successive and similar percentage amount. The straight lines on logarithmic coordinates confirm this expected relationship.
Note that the effect continues over the long period of about 30 years. Thus the effect probably continues throughout a life time. This is the best confirmation now available that shows that risk of a factor affecting atherosclerosis can continue to be duration dependent over very long time periods. Although effects of other factors such as antioxidants also appear to be duration related, data was not found verifying this effect for these factors for durations as long as this 30 years. This effect of duration also may explain in part why diabetes in early life is a much more serious health liability than is adult onset diabetes. At any age, the duration of diabetes will be much longer for those that contract it at a young age than for those that first experience as an adult.
The above plots for both men and women show zero duration intercepts at risk values of about 1.5 and 2.3 respectively. As before, if the duration related trends had pointed to zero at zero time this would indicate that atherosclerosis probably was a sole directly causative agent. The fact that these intercepts are clearly well above zero shows that factors OTHER than atherosclerosis must be involved that act in immediate or much shorter time. The expected cause here is that diabetes also promotes blood coagulation that increases risk of coronary disease. Both of these risk factors are quantified for men and women by the relationships in Figure "Diabetes" above. A number of research studies show that diabetes increases the clotting tendency of blood.
A next question is “Why is the diabetes risk factor higher for women than for men in both level and for extent of duration? A likely BioChemical Engineering scenario is as follows: The death rates of all U.S. women for coronary death during their reproductive years from age 12 to age 50 increase at an average of 10.0% per year. This is a 3.1% per year lower rate than the 13.1% per year increase in death rate during these years for men. This results in a risk ratio of coronary death for women vs. men that during their reproductive years declines from 0.9 at year 12 to 0.5 at year 30, and to about 0.28 at age 50. (See the discussion on Estrogen). After the menopause the risk of women moves back up toward that of men. From a review of various studies in Medline this is probably due to a antioxidant effect of estrogen on cholesterol. This reduces the accumulating rate of atherosclerosis and heart disease death for women by 3.1% (or risk ratio of about 0.97) for each year of pre-menopausal estrogen exposure. (See the Atherosclerosis discussion for deposition rates on women.) Cholesterol levels for men and women are similar during these years.
Results in the above Figure "Diabetes" suggest that diabetes increases the rate of heart disease for men by a cumulating 2.8% (or risk ratio of 1.028) per year of its duration, probably by increasing rate of atherosclerosis. But for women, this rate is increased by 5.5% (rr of 1.058) per year, a value 2.7% per year higher than that for men. This suggests that diabetes effectively removes much of the estrogen benefit of women during these years and still adds on the near 3% per year debit for the disease suffered by men. If actual US death rates for both men and women during the pre-menopause years are multiplied by the above risk factors at duration from age 20, the risk of heart disease death for diabetic women rises to about 75% of that for men during the ages of 20 to 50. Beyond age 50 the risks for diabetic women move up to and above the risk of diabetic men.
Conventional risk factors from statistical methods such meta analyses can be useful in identifying values of risk that are more accurate than that possible from individual studies. But such statistical ratios still may not relate to any casual or biochemical mechanism. Thus meta analyses can be seriously misleading if important contributing factors such as duration of exposure are overlooked. The above mentioned meta-analysis appears to have developed average risk values for diabetes and heart disease that are from half of to less than a fourth of the more probable true long risk because most included research values were obtained at short duration of the disease. People that experience this disease usually will have it for many years and even for rest of life. Thus their true risk will increase steadily with duration of the disease and become far higher than that indicated by such statistical methods. Life Ahead forecasts recognize where possible the probable risks of diseases and factors that can change with extent of their duration. Risks of some key factors affecting both cardiovascular diseases and cancer now appear to be strongly duration dependent.
A Global analysis such as the above can help derive a more accurate assessment of a health risk than the use of the simple risk ratios that usually are derived directly from studies as a snapshot at some age. But a more important potential of Global Analysis is the obtaining of a quantified valuation of the probable BioChemical mechanisms of major disease involved that could help suggest ways of slowing the halting their progress. This analysis highlights two approaches for consideration:
First, diabetes may develop a highly pro-oxidant environment that not only cancels the beneficial effect of estrogen but enhances atherosclerosis at an increased rate that cumulates at about 3% per year. Thus use of substantial amounts of antioxidants might eliminate much of the duration dependent harm of diabetes. A Medline review identifies 26 studies that show that this is not a new idea, that diabetics are deficient in antioxidants, and that this should help. But a serious fallacy frequently conjectured is that this “Needs a randomized clinical study.” This same Global Analysis shows that a practical clinical study of the effect of an antioxidant in reducing the health risk of diabetes would have little chance of contributing usefully. This is because as shown separately an expected benefit from an antioxidant on men would slow atherosclerosis by only 3% per year of use, and that this will have a minimal effect of risk of heart disease. For example a benefit of only 7.5% from the atherosclerosis factor would be expected for use of antioxidants for the average duration of disease of 2.5 years measured in a usual five year clinical study. Thus any practical clinical type study of much less than 15 years would fail to find any effect within its probable margin of error, and would have little chance of determining a true long range effect. This same problem was noted via the probably invalid conclusions inferred from clinical studies of Vitamin E. A clinical study shorter than 15 years could be worse than useless by first delaying and later denying use of a health action that could be of major long term benefit. A comparison of actual disease of diabetics who have used substantial amounts of antioxidants for a minimum of ten years with disease rates of non-users during this same time period would be a far more useful research approach.
Second, an increased average risk of 1.5 times for men and 2.3 times for diabetic women appears as non-atherosclerosis related and may be due to blood clotting that now is confirmed as a factor enhanced by diabetes. Although not useful for studying the more important atherosclerosis related effect, clinical studies and actual biochemical experiments could be useful for studying the shorter term benefit of anti-coagulating agents on risk of diabetes.
Note that the statistical risk ratios from a half century of actual health studies provide only a confusing array of highly varying risks of heart disease experienced by diabetics. In contrast, this example of Global Biochemical Engineering analysis of the same experimental data together with other known facts not only provides a probably more accurate definition of risk but via quantification suggests what might be done to reduce risk of the disease. And previously disparate results become reconciled.
Like Ahead uses this more powerful method of Global Biochemical Engineering analysis in identifying risks in the model for all causes of disease wherever possible. The program now uses the equations of the plots in Figure “Diabetes” to compute risk of the disease stepwise at each age after disease was first diagnosed. But in accord with a usual conservative practice, the highest risk levels accepted do not exceed the values for men and women at the 30 year maximums actually measured. Life Ahead considers the important effect of antioxidants on heart disease as of the same value percentage-wise on diabetics and non-diabetics. But because of their higher risk level, diabetics achieve a higher absolute computed benefit in lifetime Well-Days for use of antioxidants.
Life Ahead also considers the similarly higher contributions of anti-coagulating agents such as aspirin for diabetics. But the program does not include benefits potential for more potent anti-coagulating agents now available medically. It also does not yet attempt to predict the actual level of diabetes occurrence that is due to lack of exercise and overweight, and does not yet attempt to predict an effect on heart disease of extent of the disease as from glucose level, personal care, or from diabetes type. Thus the present Model should be expanded and refined further to include such factors.
Life Ahead computes that an average man that contracts diabetes at age 30 and who observes normal care and has average habits will have a life expectancy of about age 60. But it also forecasts that the use from age 30 of a combination of the good habits advised by the program should enable this same average man to enjoy a good life into his 80’s. The outlook for those with higher or lesser than average severity of diabetes of course will be less or more favorable than this average.
