n this interview, in-cites correspondent Gary Taubes talks with Dr. Stanley S. Franklin about his work in heart disease research. Dr. Franklin is the lead author of the Fast-Breaking paper in Clinical Medicine for April 2002: "Does the relation of blood pressure to coronary heart disease risk change with aging? The Framingham Heart Study," Circulation 103(9): 1245-9, 6 March 2001. Dr. Franklin is Clinical Professor of Medicine at the University of California, Irvine, as well as the Associate Medical Director of Irvine’s Heart Disease Prevention Program.

If you look at the effect of age on blood pressure, you find that in people under the age of 50, both systolic blood pressure and diastolic blood pressure tend to go up in an almost parallel fashion. By the time a person reaches age 50, however, the heart has contracted about 2 billion times and the elastic protein in the conduit vessels—the thoracic aorta and its branches—begins to show wear and tear. Now the heart has to pump against the increased stiffness in these large conduit arteries. As this occurs, diastolic blood pressure tends to go down, whereas systolic continues to rise, so that there is an increase in pulse pressure—the difference between peak systolic and end diastolic pressures—as an indicator of this underlying degenerative process that occurs from middle age onward.

  And how that change relates to heart disease risk is what you're trying to dissect in this study?

Yes. Actually, this paper was part of a trilogy of Circulation papers. The first paper was titled "Hemodynamic patterns of age-related changes in blood pressure" (Circulation 96[1]: 308-15, 1 July 1997). The second was titled "Is pulse pressure useful in predicting the risk for coronary heart disease?" (Circulation 100[4]: 354-60, 27 July 1999). And finally this present paper. The first study from Framingham showed the effect of aging on blood pressure components and how the pulse pressure begins to increase dramatically from age 50 onward. This study was possible because of the invaluable resource of the Framingham Heart Study, which began in 1948, long before the advent of antihypertensive therapy. Therefore, out of the original 1948 cohort of some 5,200 individuals, we were able to track about 2,000 individuals who had never received anti-hypertensive therapy. This allowed us to study the natural history of increasing blood pressure without the effects of treatment.

Then in the second study we asked the question, "Can we separate pulse pressure from systolic blood pressure in terms of predicting coronary heart disease risk?" These two blood pressure components are highly correlated because pulse pressure is influenced by the height of systolic blood pressure. In the second article, using essentially the same population as in the first paper, we were able to show in individuals aged 50 to 79 years that coronary heart disease risk increased with lower diastolic blood pressure at any level of fixed systolic blood pressure, suggesting that higher pulse pressure was an important component of this risk.

  And what did the highly cited 2001 paper add to the equation?

Well, the next logical question to ask was does the relationship between pulse pressure and coronary heart disease risk hold for younger individuals? Therefore, we combined the original cohort with the 1971 offspring cohort from Framingham, so that we were able to extend the age range from 20 to 79. This allowed us to triple the number of subjects and double the number or new coronary heart disease events over a 20-year follow-up period. With this increase in power, we were able to examine a variety of age groups—under 40, 40 to 49, 50 to 59, 60 to 69, and 70 to 79. Of course, both the 1999 and 2001 study populations did not contain individuals with known clinical coronary heart disease at baseline and were corrected for confounding coronary heart disease risk factors such as age, gender, ratio of total to HDL cholesterol, smoking, diabetes, and body mass index. And what we found was striking: an almost linear, step-wise gradual shift from diastolic to systolic and then to pulse pressure in predicting coronary heart disease risk. In young patients under age 50, diastolic blood pressure was the stronger predictor. Then there was a crossover from age 50 to 59 when all three blood pressure components were predictive. From age 60 onward, systolic pressure was positively related and diastolic pressure was negatively related to risk so that pulse pressure became superior to systolic pressure.

  Why did it have such impact in the medical community?

I think our ‘99 paper had shown the significance of pulse pressure in predicting coronary heart disease in the older age group. This 2001 paper, for the first time, showed this age-related changing pattern of blood pressure components that predict coronary heart disease risk. Why would nature be so fickle as to have this changing pattern? From a physiologic point of view, we could not explain this on the basis of resistance downstream at the level of the arterials. Nor could we explain it as arterial stiffness upstream at the level of the large conduit arteries, as represented by the thoracic aorta and its branches. There had to be a third component to cardiac load. And it turns out that the third component is something which most doctors really do not appreciate. This third component is called wave reflection. Whenever the heart contracts, it generates a pulse wave. We can feel the pulse at the wrist. Since ancient times, doctors have been aware of the importance of the pulse in diagnosing disease. But pulse wave reflection was not understood until the second half of the twentieth century.

In young, healthy individuals with highly elastic arteries, the forward pulse wave reflects off small resistance vessels and small artery branches to produce a reflected wave. This reflected wave is in phase with the forward wave, reinforcing it and thereby producing pulse wave amplification. When you take the systolic or pulse pressure at the brachial artery in the arm, you are measuring an amplified reading from what the heart sees in the ascending aorta. As we age and elastic arteries stiffen, amplification decreases gradually because the reflected waves increase in velocity and amplitude and become out of phase with the forward waves. Furthermore, the accelerated reflected wave in elderly individuals with wide pulse pressure hypertension reaches the heart in systole where it increases or ‘augments’ the late systolic peak that the heart must pump against, thereby increasing cardiac afterload and contributing to increased cardiovascular risk. And as the elastic arteries stiffen and pulse-wave amplification decreases with aging, there is a gradual shift from diastolic to systolic and eventually to pulse pressure as predictors of cardiovascular risk. I believe that this paper is of considerable interest to physicians because it provides a logical explanation for the changing pattern of blood pressure indices that predict cardiovascular risk.

We as physicians have been slow learners. It has been almost 100 years since Korotkoff's contribution of measuring systolic and diastolic pressure by the auscultatory method. And it has taken us this long to figure out why there is a changing pattern of blood pressure indices that predict risk at the brachial artery, and how this relates to what the heart ‘sees’ at the ascending aorta.

  What was most challenging aspect of this research?

I think the most challenging aspect of the study was to come up with a large cohort of subjects from ages 20 to 79, free of coronary heart disease and free of anti-hypertensive medication at baseline, and then have the ability to study these people over 20 years. And, of course, Framingham subjects have been beautifully studied. They go in for examinations every two years and they're followed very carefully, which gives us a very accurate determination of changing blood pressures with aging.

  How would this information translate to clinical treatment?

Well, first of all, I think it corroborates the present guidelines of the Joint National Committee's Sixth Report (on the Prevention, Detection, Evaluation, and Treatment of High Blood Pressure, JNC-VI), which came out in late 1997. For the last five years we have been using these guidelines, and they use both systolic and diastolic blood pressure to assess cardiovascular risk. So in a sense our studies have validated most of those guidelines. For younger individuals, diastolic will still be the best predictor of cardiovascular disease. Systolic is the single best predictor when you look at the entire age of the population and the entire distribution of hypertension. Our finding, along with those of other investigators, do supplement the recommendations of JNC-VI, in that they emphasize the importance of pulse pressure as a supplement to systolic blood pressure in assessing cardiovascular risk in the older age population. I would emphasize, however, that you must not consider pulse pressure as the sole blood pressure component. First look at systolic, then adjust the risk upward if a person with high systolic has low diastolic, which means high pulse pressure. So pulse pressure becomes an added prognostic factor that we can use after we evaluate systolic pressure.

  What's next in your research?

First of all, much education needs to be done. New concepts like this take time to be accepted by the medical profession. The findings from Framingham, in terms of pulse pressure, have now been confirmed by most but not all studies from around the world. Pulse pressure may not add to systolic pressure as a predictor of cardiovascular risk in middle-age healthy populations. It appears that the stiffer the blood vessels, the more likely pulse pressure will be superior to systolic pressure as a predictor of risk. Others have shown that after a heart attack, pulse pressure is a very significant predictor of future events. In those individuals who develop heart failure, pulse pressure is a significant predictor of mortality. And in patients with renal disease on hemodialysis, the pulse pressure also becomes a very important predictor of all cause mortality. Therefore, in the patients with preexisting heart disease, pulse pressure would appear to be most important in predicting future cardiac events. So we definitely have to educate the medical profession. This takes time. It can't be done overnight.

In terms of future research, we intend to study the natural history of wide pulse pressure hypertension. What are the various ‘pathways’ to the development of so called wide pulse pressure isolated systolic hypertension? What are the risk factors that are associated with the development of these pathways? What are the best drugs to use for treating this form of hypertension in the older age group? How can we improve the present poor control of this common form of hypertension in our older population? Much remains to be done.

  If you had a chance to do this research over again, is there anything you would do differently?

No, I think not. I have worked with an extraordinary group of investigators: Dr. Nathan Wong, an epidemiologist at the University of California, Irvine; at Framingham, Dr. Martin Larson, statistician; finally, Dr. Daniel Levy, director of the Framingham Heart Study. I can’t imagine having a better group of scientists to work with on these studies. I have been very fortunate.

Stanley S. Franklin, MD, FACP, FACC
University of California, Irvine
Irvine, CA, USA