Author + information
- Received November 29, 2009
- Accepted December 16, 2009
- Published online August 17, 2010.
- Daniel Steinberg, MD, PhD⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Daniel Steinberg, University of California, Department of Medicine-0682, 9500 Gilman Drive, La Jolla, California 92093
The thesis advanced here is that we are initiating treatment of hypercholesterolemia (and other risk factors) too late in life. Initiating treatment at, for example, age 30 years instead of age 60 years might very well prevent not just 30% of events, as in the 5-year statin trials, but perhaps as many as 60%.
A 30-year-old man is referred because of hypercholesterolemia (total cholesterol, 240 mg/dl). His high-density lipoprotein cholesterol (HDL-C) is 45 mg/dl, and his systolic blood pressure is 135 mm Hg. He is a smoker and has repeatedly tried to quit, without success. He has a negative family history for coronary heart disease (CHD), and he has no other risk factors. These numbers put his 10-year Framingham risk at 6%. According to current ATP (Adult Treatment Panel) III guidelines, drug treatment of hypercholesterolemia in this man would be inappropriate (1).
Now let's move the clock forward 30 years and assume that the numbers above remain exactly the same, except that he is now 60 years old. His Framingham 10-year risk is now 20%, and now he does qualify for immediate drug treatment. In the intervening 30 years, the extent of his coronary atherosclerosis has been relentlessly increasing, and the probability that 1 of those maturing lesions will rupture has been increasing along with it. We have lost 30 years during which it might have been possible to significantly slow progression and improve the chances that this man could avoid a myocardial infarction.
The crucial point here is that the 6% estimated 10-year Framingham risk at 30 years of age in our example above is misleading. That figure may correctly predict the risk for the immediately following 10 years but fails to take into account that the risk is going to increase progressively as he ages. Lloyd-Jones et al. (2) have made a persuasive case for switching from 10-year risk to lifetime risk, namely, the probability that a person will die of coronary disease sooner or later. They point out that for men 40 years of age in the Framingham cohort whose plasma cholesterol level was between 200 and 239 mg/dl, the 10-year risk was only 5%, but lifetime risk was 43%. In other words, eventually almost 50% of these hypercholesterolemic men are going to die of CHD!
Will starting treatment earlier really help?
The lesions of atherosclerosis begin in childhood. Fatty streaks were found in the coronary arteries of more than a third of our Korean War casualties, average age 22 years (3). Raised lesions are found in the coronary arteries of 37% of healthy organ donors age 20 to 29 years (4). These lesions are themselves, of course, clinically benign but “the fatty streak is the father of the fatty plaque,” as shown in the landmark national study Pathological Determinants of Atherosclerosis in the Young (5). The point is that if we could prevent the formation of the fatty streaks, there might not ever be plaques to worry about—or at least fewer of them. Is that a reasonable expectation? Well, the major risk factors that predict the extent of fatty streaks in children are exactly the same as those that predict the probability of infarction in adults: hypercholesterolemia, cigarette smoking, and hypertension (6,7). In short, atherosclerosis begins early and stays late. It is the same disease over the decades, albeit with structural characteristics and composition evolving with time. Thus, there is good reason to believe that intervention early in the game would reduce the chances that the stage of vulnerable plaque would ever be reached.
This paper will focus on the evidence supporting earlier intervention on hypercholesterolemia, but it should be stressed that much of the same reasoning applies to obesity, cigarette smoking, hypertension, diabetes mellitus, and other reversible risk factors. If and when imaging techniques allow us to measure quantitatively the burden of early lesions, we will then have an additional tool for assessing risk early on.
Important new genetic evidence mandating early intervention
Some new indirect but powerful genetic evidence tells us how great the impact of early intervention might be. Cohen et al. (8) have shown that having a low level of low-density lipoprotein (LDL), not just for the canonical 5 years of most intervention studies, but for a lifetime, confers much more protection against CHD than statin treatment in the 5-year clinical trials (8). The recently discovered PCSK9 gene plays an important role in regulating the level of expression of the LDL receptor. Loss-of-function mutations in this gene in African Americans cause the plasma LDL level to be lower by 28% than in the general African-American population. Approximately 2% of African Americans carry this mutation, enough to allow meaningful epidemiologic studies. Now, a 28% drop in LDL in a 5-year statin trial would be associated with a 25% to 35% drop in CHD risk. In contrast, this 28% drop in LDL from birth was associated with a startling 88% drop in CHD risk (8).
These remarkable findings have been confirmed and extended (9,10). Gain-of-function mutations of PCSK9 decrease expression of the LDL receptor and result in a phenotype closely resembling familial hypercholesterolemia (11). Assuming that the only function of PCSK9 is to regulate LDL levels (i.e., that the decrease of CHD is directly attributable to the LDL lowering), it should be possible in principle to see much more than a 30% decrease in CHD risk if we treated earlier. How early? Few would propose starting drug treatment at birth, but for persons at very high risk, it should be started not too long after that, and perhaps it should be started very early even for those at moderately high risk. With early intervention, the degree of LDL lowering might not have to be as great as what we aim for when intervention is started late in life. Hence, the doses of drugs or the strictness of diet necessary to achieve goal levels might be less of a problem. Early intervention on other reversible risk factors would probably amplify the benefit.
Why then are we not treating at an earlier age?
If there were a drug that could correct hypercholesterolemia cheaply and with zero risk of adverse effects, we might already be adding it to our drinking water. But no drug is totally free of side effects. Even the statins, possibly the safest of drugs effective in treating a chronic degenerative disease, have side effects, and the cost, while dropping significantly, is not zero. Then why don't we start some large-scale, double-blinded, randomized clinical trials to directly compare effectiveness of treatment starting at 30 years of age with effectiveness of treatment started at 60 years of age?
First, event rates being so much lower in younger subjects, the number we would have to enroll would be in the tens of thousands, very possibly >100,000. Second, the costs of the study would be astronomical. Such a study will probably never be done. In the 1960s, the National Heart Institute realized that it was critical to settle the debate about the role of dietary fat in CHD. An executive committee under the chairmanship of E. H. Ahrens was given funds to do a feasibility study. They concluded that the 5-year study would require the recruitment of 50,000 to 100,000 men age 40 to 59 years, and that it would require almost the entire Institute budget for 5 years (12). It was never done. A 30-year coronary prevention trial in a general population will probably never be done.
If the definitive randomized clinical trial is out of the question, what are we to do?
One option is to say that in the absence of a clinical trial, we cannot be certain that early treatment will be safe and effective, and therefore our hands are tied. That would be the conservative and conventional thing to do. But starting earlier might very well increase salvage rate from the 30% seen in the 5-year statin trials to as much as 60% (extrapolating from the PCSK9 findings). It is predicted that about 800,000 Americans will have a new coronary attack in 2009 (13). If this annual rate continues at the same pace, statin treatment under current guidelines may eventually reduce it by 30% (prevent about 240,000 events), but starting treatment much earlier might prevent as much as 60% of events—480,000 myocardial infarctions. In other words, a decision to not start earlier might mean allowing about 240,000 coronary events to occur annually that could have been prevented.
The second option is to say that the strength of the evidence for the lipid hypothesis is sufficiently robust to justify extrapolating and proceeding without the clinical trial (the trial that will never be done). Many different lines of evidence support the lipid hypothesis, ranging from animal model studies, through epidemiologic correlations, pathologic observations, and mechanistic studies (14,15). Most importantly, the remarkable results of the statin intervention trials, in the general population and in almost every subpopulation studied, leave no doubt that lowering cholesterol levels reduces risk. What are we waiting for?
How Early Do We Start?
As always, the devil is in the details, and the answer will vary according to the cardiovascular risk of those to be treated. We already treat those at extremely high risk (e.g., familial hypercholesterolemia) in childhood. At the other end of the spectrum, we might not want to intervene at any age for persons with very low lifetime risk. An opening bid for discussion might be this: start treatment at 30 years of age for anyone with a lifetime risk of 35% or more. Much animated discussion will undoubtedly follow.
Obviously, very careful thought must be given to any risks that might be involved. For example, long-term safety evaluations lasting not just months but years might be required for any regimen proposed for use in younger groups. Expert panels need to go into the details and decide what approaches the available science supports. I think we can do better. What are we waiting for?
Dr. Steinberg has been a consultant for Merck.
- Abbreviations and Acronyms
- coronary heart disease
- low-density lipoprotein
- Received November 29, 2009.
- Accepted December 16, 2009.
- American College of Cardiology Foundation
- Tuzcu E.M.,
- Kapadia S.R.,
- Tutar E.,
- et al.
- Myocardial Infarction Genetics Consortium
- Lloyd-Jones D.,
- Adams R.,
- Carnethon M.,
- et al.
- Steinberg D.
- Steinberg D.,
- Glass C.K.,
- Witztum J.L.