FACTOR-64 showed that screening diabetic patients who are at high risk of asymptomatic artery disease (CAD) using coronary CT angiography (CCTA) wasn’t justified. It was a well-designed, randomized clinical trial, the crème de la crème. Case closed.
Well, maybe not quite.
No symptoms, then death
Here is the patient Joseph B. Muhlestein, MD, and his colleagues had in mind when they designed FACTOR- 64: a man, maybe in his early 60s, under the care of a physician for his diabetes mellitus and with no symptoms of CAD, who starts up his lawnmower one summer morning to trim his grass. Mid-task he keels over from a heart attack. Death becomes his presenting symptom. His autopsy reveals three-vessel disease.
“I thought I was going to find patients like that and I was going to save their lives,” says Muhlestein, director of cardiovascular research at Intermountain Medical Center in Murray, Utah, and an interventional cardiologist at its Heart Institute. “Indeed, I think I did, but I didn’t find that many—not enough to make a statistically significant difference in the outcome of the trial.”
Muhlestein and the team at Intermountain theorized that because diabetes mellitus is a risk factor for CAD and it is a major cause of death for patients with diabetes, routinely screening asymptomatic patients using CCTA might help identify those MIs waiting to happen. Other cardiologists and radiologists had proposed similar strategies. The DIAD (Detection of Ischemia in Asymptomatic Diabetics) trial, for instance, randomized 1,123 asymptomatic patients with type 2 diabetes to stress myocardial perfusion imaging or no screening (JAMA 2009;301:1547-1555). Screening made no dent on outcomes—but the study also didn’t offer a treatment plan for patients with abnormal scans.
“They sort of ignored the fact that [patients] had ischemia because they were asymptomatic,” Muhlestein says. “We decided in our study we would not do that, so we set up recommendations.”
FACTOR-64 (Screening for Asymptomatic Obstructive Artery Disease Among High-risk Diabetic Patients Using CT Angiography, Following Core 64 trial) randomized 900 high-risk patients between 2007 and 2013 with a history of type 1 or 2 diabetes and no documented atherosclerotic disease who were treated in the Intermountain Healthcare system (JAMA 2014;312:2234-2243). Participants either received standard diabetes care or underwent CCTA screening. Researchers categorized scan results by CAD severity and coronary artery calcium score and provided recommended management strategies to each patient’s physician.
CCTA revealed that 11 percent of patients had severe CAD, 12 percent moderate CAD, 46 percent mild CAD and 31 percent no or minimal CAD. But they found no significant difference in the rates for the composite of all-cause mortality, nonfatal MI or hospitalization for unstable angina between of the groups at a mean four-year follow-up. In both groups, the rates were low, at 6.2 percent for CCTA and 7.6 percent for standard care, making annual event rates less than 2 percent. The composite of ischemic major adverse cardiovascular events was also low, at 4.4 percent vs. 3.8 percent, respectively.
FACTOR-64’s negative findings reflect positive developments as well as the challenges of conducting a rigorous randomized clinical trial. The researchers had anticipated a two-year event rate of 12 percent, based on a 2006 study of Intermountain data. “Simultaneous with planning this study eight or nine years ago, Intermountain Healthcare decided that diabetes was a very important risk factor and an important group to manage,” Muhlestein says.
Intermountain developed a diabetes management program that not only emphasized standard guidelines, but also tracked outcomes in all diagnosed cases, provided feedback to physicians and gave them a bonus if the patient met his or her targets. “There was strong incentive for the doctors to do that kind of [care] that did not exist 10 years before,” he notes. “That is why we way underestimated our adverse event rate.”
Also, power calculations often cannot capture the nuances of a controlled trial, observes Raymond J. Gibbons, MD, of the Mayo Clinic in Rochester, Minn., and author of an editorial that accompanied the FACTOR-64 results (JAMA 2014;312:2219-2220). “When the trial is done properly, the control group is actually treated more consistently than the data from the groups that are used to design the study,” he says. “It is not uncommon for the event rates in randomized trials in the control groups to be lower than expected. These event rates were strikingly lower than expected.”
In addition, the FACTOR-64 team projected enrollment to reach 1,100 but they stopped at 900 because of time and money constraints. They consequently experienced a double whammy: a control group that performed better than expected in a smaller-than-factored sample size. “It sets a standard in how future studies are going to have to be designed,” Gibbons says. “They are going to have to be larger, if these event rates are representative.”
Gibbons had high praise for the medical management protocol used at Intermountain. While not the intent of the trial, the findings demonstrated what an aggressive approach can accomplish. “It shows that with state-of-the-art contemporary therapy, when actually delivered, the event rates are no longer what most people would define as high risk for diabetics alone. It has enormous public health implications,” he says.
Finding that high-risk patient
If physicians treat patients with diabetes at the standard of care practiced at Intermountain, the pool of at-risk people should shrink, but probably not completely. Screening with CCTA still might ferret out that asymptomatic but very sick patient. Indeed, CCTA did flag the patient Muhlestein envisioned as a beneficiary of imaging: an asymptomatic woman who underwent multiple-bypass surgery after screening revealed severe three-vessel disease and a left ventricular ejection fraction of 35 percent.
Her ejection fraction rose to 60 percent after revascularization. “She said to me, ‘Wow, I feel a whole lot better now. I was blaming my symptoms on getting old,’” Muhlestein recalls.
CCTA screening still may prove its worth if applied to a narrow, higher-risk patient population, Muhlestein and Gibbons agree. That higher-risk group might be patients with not just diabetes but diabetes plus specific lipid abnormalities or diabetes plus current or past tobacco use, Gibbons suggests.
The challenge, Muhlestein recognizes, is identifying them. “I think there is a small subpopulation in which revascularization may be beneficial. If we can figure out a way to stratify so we don’t have to screen so many patients to find those patients, there is still a use for coronary CT angiography.”