Doctors in recent years had become entused over substances in the blood that appeared to be powerful new predictors of a heart attack. These substances included C-reactive protein, or CRP; homocysteine; and BNP, or B-type natriuretic peptide.

    But new research by scientists at the highly regarded Framingham Heart Study, found that tests of CRP, BNP, homocysteine and seven other substances are only a couple of percentage points better at predicting outcomes than the standard, commonsense risk factors that doctors have known for decades.

    The difference in accuracy was considered so small as to be negligible.

    "It's a little bit disappointing," said lead author Dr. Thomas J. Wang, a Harvard Medical School assistant professor. "I think our study and some other recent studies have begun to dampen that enthusiasm."

    The Framingham researchers have been testing residents of the Boston suburb for several decades. They used frozen blood samples taken from 3,209 healthy Framingham participants in the mid-1990s, then checked to see who had major heart complications or died during the following decade.

    The best predictors were tests for BNP, CRP, homocysteine and renin in the blood and albumin in the urine. Patients with the highest levels of BNP and albumin were nearly twice as likely as those with lowest levels to suffer a heart attack, stroke or heart failure; those with the highest levels of CRP, BNP and the three other substances were four times as likely to die during the decade.

    However, the standard risk factors -- high blood pressure, high cholesterol, family history, advanced age, smoking, obesity, lack of exercise and diabetes -- were just as accurate as predictors.

    "This really supports the value of focusing on risk-factor reduction, not looking for a magic blood test," said Dr. Richard Stein, director of preventive cardiology at Beth Israel Medical Center in New York. He predicted some insurance companies will soon start telling doctors to limit such tests, which are not covered by some insurers.

    "At present, I don't think we should be measuring these in routine clinical practice," said Dr. Daniel Levy, director of the Framingham study, which is funded by the National Heart, Lung, and Blood Institute.

    Some doctors said they expect better blood tests to be discovered eventually.

    "The question is, 'Where's the next frontier?'" Levy said.

    He said the answer could come from two huge federal studies. One, with results expected next year, is examining about 500,000 genetic variations in 10,000 people to spot differences linked to heart disease; the other, still being planned, will examine about 150 biomarkers to see which ones best predict future heart problems.

    (Agencies)