03/22/2011 08:41 am ET | Updated May 25, 2011

How We Treat Heart Disease Isn't Good Enough

Every year more than one million Americans find themselves in their local emergency room feeling like an elephant has plopped down on their chest. Heavy, suffocating chest pressure is one of the hallmark symptoms of coronary artery disease, our nation's number one killer, and so physicians take it seriously. If lab tests and history suggest there's a decent chance that the pachyderm in the room is coronary artery disease, you will be admitted to the hospital and treated aggressively.

Either sooner (within minutes in the case of a heart attack) or later, you will likely undergo a coronary angiogram, the gold standard test for cardiovascular disease. The procedure entails threading a small catheter into the opening of the coronary arteries, then injecting a chemical that makes the inside of the arteries appear white on X-ray. A healthy artery looks smooth and wide open. A diseased artery looks narrowed and beaded -- or in the most severe cases, completely blocked.

Unfortunately, as research published in January's The New England Journal of Medicine suggests, the gold standard test for detecting our country's most lethal health problem appears to behave more like tin.

Researchers followed nearly 700 patients who presented with a heart attack or a threatened heart attack. Each patient had an angiogram to identify the blockage that was causing the problem; the culprit blockage was then ballooned open (a procedure called angioplasty) and a small wire mesh device called a stent was inserted to keep the blockage from reoccurring. That's a typical angiogram procedure, but in this study, each patient also went on to be evaluated with a newer technology called "catheter-directed intravascular ultrasound." The idea was to use the more discerning eye of intravascular ultrasound to assess how many atherosclerotic blockages, a.k.a. "plaques," the angiogram may have missed.

It turns out that the standard angiogram misses a lot of plaques. In this study, conventional angiograms documented a total of about 1,800 of these blockages, whereas intravascular ultrasound found 3,100 -- despite the fact that ultrasound can't even "see" the last third of an artery the way an angiogram can. Not only did the angiogram miss a lot of blockages, but it often underestimated the severity of the blockages it did find. For example, angiograms found just 12 high-grade blockages; intravascular ultrasound found 283.

Over the following three years about 20 percent of the study patients returned to the hospital with more heart problems, and many of them had another angiogram to try to identify where the new blockage was. For about half of the returnees, the new problem was caused by the growth of a previously noted, but small (and therefore unstented) blockage. In the other half of cases, the chest pain occurred because a previously placed stent had closed off. That's a 50/50 split despite the fact that small unstented blockages outnumbered the larger stented blockages by more than two to one.

The researchers were also interested in whether there were any particular features on the initial ultrasound that could have predicted which blockages ended up causing problems in the follow up period. Coronary artery blockages aren't just a pile of cholesterol goo, stuck to the side of an artery. They are in essence a wound, fixed in place -- a mixture of different material and cell types going through cycles of healing and/or recurrent damage. But an angiogram can only document the severity of a blockage; it cannot peer inside this wound the way an ultrasound can.

It turned out that even when a particular blockage had all three of the most ominous ultrasound features the researchers could identify, there was only an 18 percent chance that that particular blockage would go on to cause an acute coronary problem. Call it "Whac-a-Mole Cardiology:" yes, an angiogram or ultrasound may identify a series of blockages, but we still can't predict which one will pop its head up out of the hole so we can bang it over the head with a stent.

Don't get me wrong: stents can save lives. In certain heart attack situations, angioplasty and stenting has dropped short-term mortality rates from 13 percent to 3-5 percent; in other situations, it can prevent "after-shock" heart attacks and readmissions for angina. But treating a heart attack in the here-and-now is different from preventing one in the future, which stents don't do very well. That's because, as this study showed, we're lousy at picking which blockages we should use them on, and also because stents don't always stay open: they can slowly scar shut, or quickly clot off. As a cardiologist colleague of mine says, "We've created a new disease -- the stent."

Of course we wish that stents worked better as preventive therapy for heart attacks. In fact, some interventional cardiologists wish so hard that they'll go ahead and place a stent anyway. This practice is so common that it's been given its own term, the "oculostenotic reflex," meaning that if an interventional cardiologist sees a stenosis (a higher grade blockage), he or she will reflexively stent it. In a 2006 focus group study of cardiologists in the San Francisco Bay area, one admitted, "We all agree that we don't know if we're doing the right thing, but if there's a lesion [blockage], we'll fix it."

In some cases, wishful thinking bows to greed, as angiograms are a lucrative procedure. With some regularity the multi-million-dollar exploits of stent cowboys like Baltimore cardiologist Dr. Mark Midei end up in a New York Times expose . It's unclear what percentage of stents (560,000 were placed in 2007) are unnecessary, but cardiologists will admit, at least privately, that it's a common practice. And it's clearly part of the sucking sound we hear coming from our health care premiums. Medicare alone spent $3.5 billion on stents in 2009; Dr. Steven Nissen, chief of cardiovascular medicine at the Cleveland Clinic told the New York Times, "We're spending a fortune as a country on procedures that we don't need."

The conclusion to this latest research might be "Oops, the best test we have to evaluate our country's most lethal health problem isn't all that good." If the conventional angiogram is moving down the podium, will there be a new gold standard? Many believe it will be either CT or MRI angiograms -- like intravascular ultrasound, these allow us to more accurately view the atherosclerotic scars that define coronary artery disease. Because stress tests can only detect severe disease (blockages of 70 percent or more), CT or MRI angiograms are also increasingly being used as a much more definitive screening test that can find coronary disease much earlier in its development.

In the meantime, if it took an angiogram and a stent to push that elephant off your chest, be grateful but not falsely reassured: you have been treated for coronary artery disease but not cured of it. The 90 percent blockage the cardiologist ballooned and stented may now be 100 percent open, but you'll need to be on medication to keep that stent open. And as this latest research shows, it's very likely that there are remaining smaller blockages that the angiogram either underestimated in size or didn't see at all. These could loom large in your future unless you aggressively treat the risk factors -- smoking, high blood pressure, bad cholesterol etc. -- that caused them to sprout up in the first place. Whenever possible, choose a smoke alarm over a fire engine.