AIMI-HF Misses Mark With Advanced Imaging in Ischemic HFrEF Workup

NEW ORLEANS, LA—In patients who have heart failure with reduced ejection fraction (HFrEF) and known or suspected CAD for whom revascularization is being considered, using an advanced imaging strategy has no impact on future risk of cardiac death, MI, or hospitalization compared with standard imaging with single-photon emission computed tomography (SPECT), the AIMI-HF trial suggests.

“Our study found no significant differences in the primary cardiovascular composite between the advanced and SPECT imaging groups. We saw this in the total study population as well as in the randomized cohorts,” Lisa Mielniczuk, MD (University of Ottawa Heart Institute, Canada), said in a presentation here at the American College of Cardiology/World Congress of Cardiology (ACC/WCC) 2023 meeting.

But in the cohort where ischemia—rather than viability—was the primary clinical question, “there was a marginally significant reduction in cardiac deaths in the group that got advanced imaging, and although not statistically significant, patients with advanced imaging who underwent revascularization had a trend towards outcome benefit,” she added.

On the basis of results from the STICH trial, both the US and European HF guidelines support CABG plus guideline-directed medical therapy (GDMT) in select patients with ejection fractions of 35% or lower, as do US and European revascularization guidelines. More recently, however, the REVIVED-BCIS2 study showed no additional benefit of PCI beyond GDMT in a similar population.

One thing that studies to date have left unresolved, however, is whether assessing myocardial viability and underlying ischemic burden could be used to guide revascularization decisions in patients with both HF and ischemic disease, and if so, which imaging modality might be best suited to that task. A new analysis of REVIVED-BCIS2 also presented at the ACC/WCC meeting offered some clues, showing that viability testing (using cardiac magnetic resonance imaging or dobutamine stress echocardiography) was no help in identifying which patients might benefit from PCI over optimal medical therapy.

Acting on Imaging

For the AIMI-HF study, the investigators enrolled 1,381 patients with LVEF < 45% (mean age 67 years; 14% female; median EF 28%) from 15 sites in Canada, Finland, the United States, and South America. All had known or highly suspected CAD on coronary angiography, previous MI, moderate ischemia, or scar. The majority of patients had NYHA class II or III symptoms, and about 30% had CCS Class I or II angina. Most patients were on beta-blockers, with about 80% on ACE/ARBs, and fewer than 40% on a mineralocorticoid receptor antagonist (MRA).

Patients were first randomized to an “ischemia question” or to a “viability question,” then allocated within each of those arms to either advanced imaging, wherein they were divided again between PET or CMR, or to standard SPECT imaging. Of note, only 271 patients were actually randomized in the study; the others were considered registry patients, since referring physicians had actually predetermined their imaging test. Propensity matching attempted to adjust for those differences in the primary outcome analysis.

At a median follow-up of 24 months, there was no difference in the cumulative incidence of the primary composite outcome of time to cardiac composite event (cardiac death, MI, resuscitated cardiac arrest, or cardiac hospitalization for worsening HF, ACS, or arrhythmia) between those who underwent advanced imaging versus those who had SPECT imaging in the total cohort (P = 0.696), or in those identified as having ischemia (P = 0.388). However, the cumulative incidence of cardiac death trended lower in those who had advanced imaging with PET versus those who had SPECT (11% vs 18.6%; HR 0.61; 95% CI 0.38-1.00), with the curves separating at 24 months.

Of note, just 40 patients ultimately underwent SPECT imaging in the viability question arm of the trial, making it impossible to draw any conclusions about viability from the advanced versus standard imaging comparison.

Among the 28% of the total cohort who ultimately went on to have a revascularization procedure, 61% had CABG, 37% had PCI, and 2% had both. However, the incidence of the primary composite endpoint was not reduced in those with early revascularization guided by imaging in the total population (HR 0.71; 95%CI 0.37-1.38) or in those with ischemia (HR 0.52; 95%CI 0.23-1.15).

The much smaller randomized cohort in AIMI-HF, wherein 190 patients underwent advanced imaging and 81 underwent SPECT, was underpowered, preventing reliable comparisons of outcomes by imaging test. There were no differences between imaging groups in the total or randomized populations for any of the secondary endpoints, including all-cause mortality.

“The numbers are very small and so we do need to interpret them with caution,” Mielniczuk said, also noting that physician choice and contraindications were the primary reasons for not randomizing. “Longer-term follow-up in greater numbers, of course, would help us.”

Christopher M. Kramer, MD (University of Virginia Health System, Charlottesville), a panelist during the ACC/WCC session, pointed out that there’s always “clinician bias going into these studies,” which is what makes strategy trials like this one so hard. “Imaging is not a therapy, so to make the step from imaging to therapy to outcome can be very tricky.”

He also noted that the SPECT patients on a whole “were slightly sicker in every way than the advanced imaging patients” and fewer of them underwent revascularization, further complicating interpretation of the nonrandomized patients.

Mielniczuk responded that the investigators tried to adjust for some of those factors, including the greater incidence of arrhythmias and prior revascularization in comparison with those who had an advanced imaging strategy.

Still, she speculated, it’s possible that those differences drove the signal of a cardiac death advantage in the advanced-imaging group, particularly in light of the separation of the curves at 24 months.

“Even if you look at the STICH trial data, there was a crossover in the CABG and medical arms at around 2 years. So, whether this is an early signal of something that is significant needs to be explored further, or whether it's a representation of some of the limitations of the type of data analysis that we've done, I think is . . . certainly hypothesis-generating and we're looking at further,” said Mielniczuk.