The worst kind of assurance is false assurance. And that's been one of the problems with conventional cardiac stress tests, including the venerable EKG treadmill test, where you get all wired up and walk for your life.
Conventional stress tests, often referred to as "functional" stress tests, only detect coronary artery blockages of 70 percent or greater. Blockages of this size begin to limit bloodflow to the downstream muscle; they typically lead to symptoms of chest pain or shortness of breath with exertion, and also to changes we can see on an EKG or by other means. This is called "obstructive" coronary disease, and the symptoms are a strong warning that a heart attack might be coming your way.
On the other hand, a normal stress test doesn't mean you don't have atherosclerotic heart disease. You might not have any disease; your blood vessels might be clean. But then again, you might have a lot -- if the blockages are smaller than 70 percent and therefore went undetected by the standard stress test. The problem with small blockages ("non-obstructive" coronary artery disease) is that they can quickly become a big problem. They can split open like a baked potato and attract a blood clot, thereby completely closing off a blood vessel in just minutes. This is a heart attack, a myocardial infarction.
With this in mind, cardiologists have spent the last decade or so trying to use the traditional CT scanner to take a more precise look at the coronary arteries with a "CT Coronary Angiogram."
This week's New England Journal of Medicine reports the findings of the PROMISE trial, which evaluated a group of 10,000 patients with potential heart-related symptoms and also risk factors that put them at an intermediate risk of having coronary artery disease. Half were investigated with CT coronary angiogram, and half with conventional stress testing. The majority of the conventional group had a stress test using a radioactive tracer, often called a SPECT study.
For patients with a low-to-intermediate risk of having coronary disease, SPECT studies have a sensitivity (how often a person with coronary disease will have an abnormal test result) in the 85 percent range, and CT coronary angiography (CTCA) has a sensitivity of around 95 percent. If numbers put your head in a whirl, just remember that CTCA has a "substantially higher accuracy," as the PROMISE trial authors put it.
With that in mind, it might come as a surprise that the PROMISE trial showed no statistical benefit from being evaluated with a CT coronary angiogram (CTCA) versus a standard stress test (typically a SPECT test.) Because CTCA is better at detecting coronary artery disease, about 12 percent of the CTCA group went on to have a coronary angiogram (versus eight percent of the standard stress group.) The CTCA group also ended up having twice as many revascularization procedures, primarily angioplasty and stenting. But none of that seemed to improve the study's primary composite end point, which included "death from any cause," non-fatal heart attacks, going back to the hospital with recurrent cardiac symptoms and procedural complications. Both groups of patients had the same statistical outcome.
The obvious question then is, if CT coronary angiogram is a superior diagnostic test, why didn't it perform better in the PROMISE trial?
The answer is "too short and too few." The time frame for the trial was too short--about two years. As Dr. Armin Arbab-Zadeh, a cardiologist from Johns Hopkins, predicted in 2012, "Even large scale studies, such as the PROMISE study, are unlikely to demonstrate differences in hard end points unless longer follow up is implemented." And there were too few patients who had one of the end-point events--death, non-fatal heart attack etc.. That's good news for the people who enrolled in the study, but bad news for the power of the study: simply put, not enough bad things happened to show that the intervention did any good.
Leaving "too short and too few" aside, lashing diagnostic testing to clinical outcomes is always tricky, because in between the testing and the outcomes is the treatment. If we use more powerful testing tools to just do the same thing we've been doing, then the outcomes -- which are what we really care about -- might not change.
The promise of CT coronary angiography remains strong.
Radiation concerns fall as technology improves. In the PROMISE trial, CTCA-related radiation exposure was around 12 mSv. A newly published Scottish study on CTCA used four mSv. The institution I work at has it down to three mSv. For comparison, a SPECT stress test has a radiation exposure of around 12 mSv, and a traditional angiogram via a catheter uses 5-6 mSv (and more than that if an angioplasty is performed).
As noted earlier, CTCA's can detect smaller blockages that standard stress tests cannot. And CTCA's substantially higher sensitivity means that it doesn't miss many cases of this "can't miss" disease. Find coronary artery disease early and we can prevent it from progressing; find it later in the advanced stage, and we've still got treatments to avoid or abort a life-ending heart attack. Less sensitive tests like the standard stress test more often miss people with real disease, to whom a falsely negative test brings a false assurance--or worse, as Arbab-Zadeh noted:
Based on the data available, it is plausible that thousands of patients suffer myocardial infarction and cardiac death each year after normal stress testing results that could have been identified and treated if CTA was employed instead of stress testing. Therefore, the burden of proof must be on stress testing and not on CTA to deliver evidence of this well based hypothesis.