My Philosophy

Atherothrombotic disease (ATD) does not strike as a bolt out of the blue, as William B. Kannel, MD, of the Framingham Heart Study, observed almost 50 years ago. (For a history of the evolution of the prediction of the population at risk of ATD, please look under the heading Discussion Files, and then under “Literature Review”.) The ATD population differs from the non-ATD population by certain characteristics called risk factors. ATD is uncommon is people without those risk factors, except perhaps very late in life. The chief risk factors are cigarette smoking, dyslipidemia, and hypertension. (The Framingham Heart Study was headed by Thomas R. Dawber, M, from 1949 to 1966; by William B. Kannel, MD, form 1966-1979; by William P. Castelli, MD, from 1979-1994; and since 1994, by Daniel Levy, MD.)

The prevention of ATD is based on the ability to predict the population at risk of ATD. To put it simply, “If you can’t predict, you can’t protect.” And, “The better you can predict, the better you can protect.” Standard risk predictors use single risk factors as independent variables, even use multiple risk factors in the same predictive tool, but still as independent variables. ATD, however, is a multi-factorial process and the best predictor must therefore utilize the various risk factors as dependent variables. When used as independent variables, each risk factor is treated as if it acts alone, and there will be people with ATD who were not predicted simply because they did not quite reach significance with any one risk factor. When used as dependent variables, each risk factor is considered in terms of all the other risk factors, and few ATD patients cannot be predicted. This concept is equally true for the treatment of the various ATD risk factors to stabilize/regress ATD. Treatment to goal is vital because sudden death is a relatively common initial presentation of ATD of the heart—such patients do not get a second chance to properly modify their risk factors. Put another way, “Failure to treat risk factors to goal is like jumping half-way across the Grand Canyon—it’s a long way down and it really hurts when you hit the bottom.”

It is also true that you can’t find a lipid (cholesterol or triglycerides) disorder if you don’t go looking for it. There are few warning signs that a lipid disorder is present—on rare occasion, lipid collections in the skin or lipid-induced changes in the eye will give the physician a clue that severe dyslipidemia is present. In the main, however, the patient’s blood must be sampled to determine the presence or absence of a lipid disorder. The key question is when to start lipid testing. This should not be a question that has any merit, since lipid disorders often begin in childhood. If dyslipidemia is present in childhood, then diet and exercise—and on rare occasions, medications—can be recommended early on. (And of course, the child can be cautioned very strongly never to start smoking cigarettes.) If no dyslipidemia is present, then that is good news and repeat lipid testing can be done every 5-10 years. If the patient then develops dyslipidemia later in life, the physician can determine when “the clock began to tick.” This is vital to know since the presence or absence of ATD in a patient with dyslipidemia depends on how severe the dyslipidemia is and how long the dyslipidemia has been present. Obviously, the same principles apply to the other ATD risk factors.

As a risk factor, cigarette smoking is self-evident. As you can see under the Published Articles files (look for the” Role of Cigarette Smoking in Atherosclerotic Disease” heading), cigarette smokers develop ATD at an earlier age, develop multiple-system ATD at an earlier age, and die at an earlier age than do ex-smokers, who in turn suffer all these events at a slightly earlier age than do never smokers. The definition of hypertension as existing when blood pressure is 140/90 mmHg or higher, goes back to the actuarial tables of the life insurance companies, who found that when blood pressure exceeded these levels, the death rate went over the average. I personally believe in functional definitions, and as William Connor, MD, observed some 30-odd years ago, one does not treat various conditions simply to achieve some arbitrary number goal, but rather to prevent disease. Hence, I prefer to remain with the blood pressure treatment threshold of 140/90 mmHg. Incidentally, the Framingham Heart Study showed that it was the top number (systolic blood pressure, or SBP) that was significant as far as ATD was concerned.

Dyslipidemia, however, is a trickier matter. We once had a definition of lipid disorders based on distribution curves, until it was pointed out that while most people with dyslipidemia defined by a distribution curve did indeed suffer ATD events, the vast majority of ATD events occurred in people with lipid levels far below those lofty levels. The National Cholesterol Education Panel currently recommends a variable threshold defining dyslipidemia—however, even those levels are well above lipid levels typical of ATD patients in my practice of medicine. I personally favor a functional definition of dyslipidemia, harking back to Connor’s comments which were, in my opinion, true when he spoke them and are still true today. I believe that a state of dyslipidemia exists whenever the balance between the lipids (cholesterol) entering the artery wall –the “bad” cholesterol, or LDL cholesterol—and the lipids (cholesterol) leaving the artery wall—the “good” cholesterol, or HDL cholesterol—is such that it favors the accumulation of cholesterol within the artery wall. This balance is best expressed as a ratio between LDL- and HDL-cholesterol, and I have shown that the best ratio is the Cholesterol Retention Fraction, or CRF, defined as (LDL-HDL)/LDL. The concept is simply this: Of the cholesterol entering the artery wall, how much stays there?

Technically, the term “lipids” refers to cholesterol and to a related compound called triglyceride (TG). TG are not found within the artery wall, but may play an indirect role in ATD. When TG levels are high, HDL levels are usually low, and thus the CRF will be high. Also TG elevations make the blood more apt to clot, though this effect is relatively minor unless one smokes cigarettes, which also make the blood clot. The combined effect of elevated TG and cigarette smoking is much more lethal than either one alone as regards blood clots. Even so, the term dyslipidemia, in popular usage, generally refers to cholesterol (LDL and HDL) disorders.

The only reason to treat dyslipidemia is to prevent/stabilize/regress ATD. (If TG are very, very high, then TG must be treated to prevent pancreatitis. Fortunately, such severe TG disorders are uncommon.) Goals of dyslipidemia therapy are based on gold-standard scientific evidence, which is only provided by randomized controlled clinical trials (RCT’s). RCT’s are expensive to do and only give gold-standard evidence for the questions they were designed to answer. Hence, it is commonplace to generalize the results of a RCT to related questions. This process is termed post hoc analysis or meta-analysis. While convenient and cheap, post hoc analysis gives recommendations that are at best exploratory or hypothesis-generating—such recommendations are NOT gold-standard evidence. This fact is important to remember in light of some of the recommendations currently being given. An important example of this can be found under the Letters to the Editor file—look under the letter “Effect of HDL Measurement Technique on Clinical Lipidology.” The RCT’s that give us our goals of treatment were all based on the older indirect measurement of HDL. There were some problems with the older technique, so a newer method of measuring HDL (the direct measurement technique) was developed to counter-act those problems. The newer direct measurement technique gives a HDL value about 10 mg/dl (0.25 mmoles/L) higher than the older indirect measurement technique. Since we calculate LDL-cholesterol (rather than measure it, to save money) by the Friedewald formula, a “fallacious” rise in HDL-cholesterol results in a “fallacious” fall in LDL-cholesterol. Thus one’s lipid values can look great on paper while atherogenic dyslipidemia is operative at the artery wall level. You will see the importance of this when you look at the reference I just cited.

Finally, the approach that I use for the prediction of the population at risk of ATD—and as a guide to stabilize/regress ATD—is a global risk approach since ATD is a multi-factorial disease, and one risk factor interacts with another, rather than acting independently. Back in 1981 I was searching for a means of predicting the population at risk of ATD. As outlined in the various papers that I have published (Published Articles files), I had set up my patient population as an age-sex database and had segregated out those people who had sustained some sort of ATD event. When I looked at all of their risk factors, I realized that most ATD patients had multiple risk factors, so I graphed one risk factor versus another to determine which risk factor combination was most predictive. (To read more precisely how this came about, please refer to Discussion Files, and look under “Materials and Methods”.) The predictive combination that best predicted the population at risk of ATD was the combination of CRF and SBP, and then only when stratified by cigarette smoking status. This is the graph that I use today. The graph is discussed under the Published Articles files—look under the “Prediction of the Population at Risk of Atherothrombotic Disease” heading. The graph has been validated against other studies—look under the Discussion Files and the heading “Validation of the Lipid Predictor.”

site by Sea Lion Studio is powered by WordPress :: Theme by