Week of Jan. 31, 2004; Vol. 165, No. 5
Is anticipating heart disease as easy as 1, 2, 3, 4?
Ben Harder, Science News
When Umesh Khot attended medical school in the early 1990s, his instructors taught him and his fellow students the four warning signs of heart disease: high blood pressure, elevated cholesterol, diabetes, and cigarette use. Large studies in the 1960s and 1970s, such as the well-known Framingham Heart Study, had established these links. However, the instructors informed the students of "the 50-percent rule."
That notion, recited in classes at many medical schools and repeated in the pages of medical journals, holds that half of all people who develop heart disease won't have any of these four warning signs in their medical charts. In other words, a lot of patients whom the young doctors would be treating for heart attacks would be the victims of some unknown, probably genetic, factor.
As Khot trained in cardiology at the Cleveland Clinic and then worked at Indiana Heart Physicians, a private practice in Indianapolis, his experience belied this expectation. "Virtually every [young] patient who has heart disease is a smoker," he says. Furthermore, Khot observed that older heart patients who weren't smokers almost always had diabetes, high blood pressure, high cholesterol, or more than one of these risk factors.
Perplexed by the discord between the litany of the profession and their own professional experiences, Khot and several of his colleagues set out to test the prevailing wisdom. What they've concluded, and what a separate team of researchers has independently found, is that the 50-percent rule is wrong about 40 percent of the time. Roughly 9 in 10 cases of heart disease occur in individuals who already exhibit one or more of the four risk factors.
That revelation is an important bulletin for both prevention of heart disease and cardiology research, some researchers say. First, it suggests that efforts to cut the prevalence of traits associated with heart risk could dramatically reduce the burden of the disease itself.
Second, the findings indicate that attempts to identify novel risk factors may have relatively little to offer in spotting people destined for heart trouble. In fact, says Philip Greenland of Northwestern University's Feinberg School of Medicine in Chicago, "risk-factor research needs to go in a new direction."
Because heart disease sometimes runs in families, some scientists are searching for genetic predispositions. Many other researchers are seeking new physiological traits that indicate risk of heart disease. Recently identified factors include deposits of calcium in coronary arteries (SN: 9/13/03, p. 174: Available to subscribers at http://www.sciencenews.org/20030913/note16.asp
); low blood concentrations of a protein called adiponectin (SN: 11/22/03, p. 334: Available to subscribers at http://www.sciencenews.org/20031122/note16.asp
); and high blood concentrations of the amino acid homocysteine (SN: 10/21/95, p. 264), the clotting agent fibrinogen, and inflammation-associated molecules such as C-reactive protein, or CRP (SN: 6/14/97, p. 374: http://www.sciencenews.org/sn_arc97/6_14_97/bob1.htm
), interleukin-6, and tumor necrosis factor-alpha (SN: 12/6/03, p. 366: Available to subscribers at http://www.sciencenews.org/20031206/note11.asp
But none of these markers has eclipsed the time-honored signs of heart risk, says Greenland. Obesity and physical inactivity have also been associated with heart disease, but epidemiological data indicate that these traits essentially identify people likely to have high blood pressure, elevated cholesterol, and diabetes.
"People have been searching for the magic bullet that's going to define high risk [for] 30 years," Greenland says. "The search hasn't led us very far."
Rather than seeking as-yet-unheralded risk factors, Greenland contends, researchers should be hunting for traits that protect some people despite high apparent risk. Identifying behaviors or physiological differences that spare some people with classical risk factors might lead to new strategies for warding off heart disease.
Greenland puts the 50-percent rule "in the category of myth." Despite the rule's status in medical schools, Greenland finds no data supporting it in the medical literature. He's traced written mention of the rule back to its apparent origin in a scientific paper published in 1974, which cited an earlier paper as the source of the idea. But when Greenland tracked down that reference, he found that the article advanced no such claim.
To newly determine the predictive value of the four traditional risk markers, Greenland and his colleagues at Northwestern, the University of Minnesota in Minneapolis–St. Paul, and Boston University used data previously compiled in three large, long-term studies of heart health. One study had recorded data on about 3,300 men and women between 34 and 59 years of age. Another included more than 35,000 employed men and women between 18 and 59 years of age, and a third had followed nearly 348,000 men between the ages of 35 and 57. The studies began in 1948, 1967, and 1973, respectively, and each research effort tracked volunteers for more than 20 years.
In the three studies combined, approximately 21,000 individuals had died of heart disease. In each study, 87 percent to 93 percent of the people killed by heart disease had been identified as having at least one of the traditional risk factors. The first study also recorded nonfatal heart attacks. These were presaged by at least one diagnosed risk factor in 91 percent of men and 87 percent of women.
Greenland and his colleagues reported their analysis in the Aug. 20, 2003 Journal of the American Medical Association (JAMA). In the same issue, Khot and his colleagues at the Cleveland Clinic and the University of North Carolina at Chapel Hill presented the results of their study, which tested the 50-percent rule from a different angle.
Khot's team mined data from 14 studies conducted in the past decade and designed to test various treatments for heart disease. All volunteers had developed heart disease—and survived at least its initial manifestation—before joining their respective study. Data from each subject's medical history indicated whether he or she had any of the four traditional risk factors.
In all, Khot and his colleagues examined data on more than 122,000 people, about three-fourths of whom were men. In 81 percent of the men and 85 percent of the women, at least one of the traditional risk factors preceded the first sign of heart disease.
"In the absence of these four risk factors, heart disease is relatively unusual," Khot observes. Taken together, the new analyses soundly debunk the notion that only half of heart disease can be anticipated using the traditional risk factors, John Canto and Ami Iskandrian of the University of Alabama at Birmingham said in an editorial in the same JAMA issue carrying the two reports. "These studies emphasize that to reduce the burden of cardiovascular disease, physicians should have even greater vigilance in identifying conventional . . . risk factors and must redouble efforts to control them effectively," they say.
Markers that have recently been proposed to augment the predictive value of the traditional risk factors haven't yet been proved to offer much help in defining an individual's degree of risk, say Daniel Hackam and Sonia Anand of McMaster University in Hamilton, Ontario. They examined 173 studies of relatively new risk factors, including CRP, fibrinogen, and homocysteine. Numerous studies have shown each of these markers to be statistically linked to poor blood vessel health and heart disease, Hackam and Anand observed in the same JAMA issue. Nevertheless, there's not yet enough evidence to conclude that these novel risk factors identify people at elevated risk for heart disease who would not be similarly identified using the four conventional risk factors, the Ontario researchers say.
While the new findings indicate that most people who avoid a classical high-risk profile are off the hook, not everyone bearing at least one of the four traditional traits ends up with heart disease. "Lots of people who have those risk factors do not get heart disease," Greenland says.
In all three of the studies that Greenland and his colleagues reexamined, the majority of subjects who stayed free of heart disease had one or more traits suggesting they stood at high risk. High blood pressure alone affects 28 percent of adults in the United States, a recent study found.
Some individuals who struggle to overcome a risk factor could be unnecessarily putting many dollars and much effort into drugs and diets. Others who fail to eliminate a basic risk factor from their lives may feel that they are playing Russian roulette with their hearts.
"If we don't move beyond traditional risk factors, we will continue to miss large numbers of patients at risk for heart attack," says Paul M. Ridker, a cardiologist and cardiovascular researcher at Harvard Medical School in Boston. Newer risk factors can overcome the problem of how poorly the traditional risk factors discriminate between people who will later develop heart disease and those who won't, he maintains. He argues that factors such as CRP have essential roles in anticipating and preventing heart disease.
Ridker disagrees with Hackam and Anand. CRP, which Ridker has studied for years, "is the only one of the so-called novel risk factors that has been shown to add" to the predictive value of existing tools, he says. Moreover, because CRP is a marker of inflammation, it could identify people who would gain particular benefit from anti-inflammatory therapies, he says. Highly sensitive CRP assays are comparable in price to cholesterol tests.
Greenland, on the other hand, urges researchers to focus on identifying traits that protect people from heart disease. For instance, good cholesterol, technically called high-density lipoprotein or HDL cholesterol, is known to keep blood vessels healthy and to shield the heart. "If your HDL is strikingly high, that gives you some measure of protection," says Greenland. Many components of a healthy diet, including oily fish, fruits and vegetables, whole grains, and modest amounts of alcohol, raise HDL in the body.
Identifying other risk-lowering characteristics that people can influence through behavior, diet, or medications could offer new tools in the fight against heart disease, Greenland says.
Furthermore, novel or recently identified risk factors might predict how soon heart disease will appear in an at-risk person, Canto and Iskandrian say. They note that in Khot's and Greenland's studies, "Cardiac events were observed after a long period of follow-up (as long as 30 years), and the relevance of these findings to individual patients on a shorter-term basis remains unclear."
If doctors could anticipate imminent heart disease, they might more effectively motivate people to desist from risky behaviors, such as smoking, and take aggressive preventive action, such as administering statin drugs, when it's most appropriate. Some researchers propose that elevated blood concentrations of CRP indicate high risk in the short term.
Another possible marker of immediate risk is myeloperoxidase. Stanley Hazen and his colleagues at the Cleveland Clinic measured concentrations of this blood enzyme in 604 people who showed up consecutively in the clinic's emergency room with chest pain. There's evidence that myeloperoxidase promotes inflammation in blood vessels, which can bring on heart disease.
In the month following their hospital visits, people with the highest initial concentrations of myeloperoxidase had more than four times the risk of heart attack, death, or need for cardiovascular surgery than did people with the lowest myeloperoxidase concentrations. The association between initially elevated myeloperoxidase and risk of serious heart problems persisted for at least an additional 5 months, Hazen and his collaborators reported in the Oct. 23, 2003 New England Journal of Medicine.
Other blood markers, including CRP, are currently used to measure cardiac risk in the emergency room. But when those flags go up, it's often because heart muscle has already begun to die. By contrast, even in the absence of biochemical markers for muscle death, myeloperoxidase concentrations are noticeably high in people heading rapidly toward heart trouble, Hazen and his colleagues report.
If a reliable biological marker of short-term risk can be selected, doctors may become more assertive than usual in prescribing medications to particularly high-risk patients, says Greenland. But in the main, preventing heart disease comes down to minimizing the most obvious of risks, he says. Nine times out of 10, that may be all it takes.