By GINA KOLATA
Published: June 30, 2009
A blood protein that only a short time ago was thought by some to be more important than cholesterol in heart disease now appears to be little more than a bystander.
The substance, C-reactive protein, or CRP, a marker of inflammation in the body, is unquestionably associated with heart disease: the more CRP in a person’s blood, the greater the likelihood of heart disease.
But in a paper to be published Wednesday in The Journal of the American Medical Association, researchers analyzing genetic data from more than 100,000 people conclude that their study “argues against” the notion that the protein causes heart disease.
Dr. David Altshuler, a professor of genetics and medicine at Harvard Medical School, said the distinction was important. If CRP caused heart disease, lowering it would protect people. But if it was merely associated with the disease, lowering CRP would have no more effect on health than quelling a shrieking fire alarm would have on putting out a fire.
Many believed CRP caused heart disease, especially after a widely publicized study released last year suggested that people with low cholesterol but high CRP levels had fewer heart attacks if they took a statin, a cholesterol-lowering drug that also lowers CRP.
That could mean that lowering CRP could prevent heart disease. Of course, it also could have been the cholesterol lowering that was protective, but many researchers argued that it was the reduction in CRP.
“There certainly has been a very vocal constituency in the idea that CRP causes or contributes to the development of heart disease,” said Dr. Daniel Rader, a lipid expert at the University of Pennsylvania. He noted that some companies were trying to develop drugs to lower CRP.
But Dr. Michael S. Lauer, director of the division of prevention and epidemiology at the National Heart, Lung and Blood Institute, said it might now be smart to abandon that search.
“It is likely that drugs or agents that specifically target CRP are not going to work,” Dr,. Lauer said.
The findings will not change current treatment. And one leading CRP researcher, Dr. Paul M. Ridker of Brigham and Women’s Hospital in Boston, director of last year’s study, called Jupiter, and the researcher most closely associated with the excitement over CRP, said the new study did not change anything for him.
Dr. Ridker, an inventor of a laboratory test for CRP who profits from its use, said that while the new results did not support causality, he did not think they definitely excluded it either.
Anyway, he said, it does not matter because the real issue is inflammation. CRP goes along with inflammation, and it is inflammation that is likely to be causing heart disease, Dr. Ridker said.
The thought is that white blood cells invade artery walls and release damaging chemicals, leading to plaque formation.
The new study, by Dr. Paul Elliott of Imperial College in London and 35 co-authors, used a recently developed technique that can get answers quickly about causality. Without it, the only method was what is seen as the gold standard in medicine: large clinical trials in which people are randomly assigned to take a drug, or not, and followed for years.
The new method, Mendelian randomization, “is changing the way we think about causality,” Dr. Lauer said. It only recently became feasible as researchers found genetic variants associated with proteins like CRP and developed tools to analyze data from what was, in this case, more than 100,000 people.
Different people produce different amounts of CRP, and the amount a person produces is determined by tiny inherited changes in the CRP gene. So in a population, there are people who just happen to produce more CRP throughout their lives and others who just happen to produce less. If CRP causes heart disease, those who make more would have more heart disease. That, however, is not what the study found.
“There was no association” between CRP genes and heart disease rates, Dr. Elliott said.
The association between CRP and heart disease must be reflecting something else. For example, if CRP levels go up when heart disease begins, because of inflammation in arteries, CRP levels would be higher in people with incipient heart disease. But CRP itself would be playing no role in heart disease risk; it was just marker of inflammation.
A smaller study of CRP, using the same method and published last October in The New England Journal of Medicine, came to the same conclusion.
But this second, larger, study was needed to convince heart experts, said Dr. Sekar Kathiresan, director of preventive cardiology at Massachusetts General Hospital.
“It’s a very important question particularly in the context of the Jupiter trial,” Dr. Kathiresan said.
Dr. Rader, at Penn, said he still did CRP tests on selected patients and expected to continue. An elevated CRP level indicates increased risk, even if the protein does not cause the risk. Dr. Rader tests CRP to help decide whether to give a statin to patients with normal cholesterol but with a family history of heart disease. A high CRP, he said, could tip the balance, leading him to prescribe a statin.
Dr. Altshuler noted that part of the power of a Mendelian randomization study was that it could stop a hypothesis from prematurely becoming viewed as fact.
Ordinarily, science starts with an observation, like the one associating CRP with heart disease. That generates a hypothesis — that CRP causes heart disease. Then comes a trial, if there is a treatment, like a drug to specifically attack CRP, that people can be randomly chosen to take or not.
But it can be years or decades before the clinical trials are completed. In the meantime, Dr. Altshuler said, the hypothesis comes to be regarded as true.
And if the clinical trial contradicts the hypothesis, “some people are unwilling to question their beliefs, even if there was no evidence of causality to begin with,” Dr. Altshuler said.
*The views expressed here in no way represent the views or policies of New Health Corp, the FDA, or any other official agency.