iFR - The New Kid in Coronary Physiology. Results of 2 Outcome Trials: DEFINE-FLAIR and SWEDEHEART

After nearly 25 years, fractional flow reserve (FFR) has become a clinical standard for lesion assessment in the cath lab. It was founded on the concept that hyperemia is required to permit a direct or linear relationship between pressure and flow (letting us use a pressure ratio 

to equate to a flow ratio), and separating the transstenotic pressure/flow curves better than a resting pressure ratio. FFR was compared to multiple ischemic stress tests and is regionally specific for myocardial ischemia. FFR has a long record of favorable randomized trials demonstrating reduced adverse event rates for FFR-directed percutaneous coronary intervention (PCI), along with a lower cost (FAME I). FFR-guided therapy for revascularization is superior to medical therapy for FFR-positive lesions (FAME II). 

About 8 years ago, Dr. Justin Davies from Imperial College in London introduced a resting translesional pressure ratio taken during what he described as the wave-free period — a specific diastolic interval where myocardial resistance is fixed, without the need for hyperemia (adenosine). This instantaneous wave-free pressure ratio, called iFR, appeared to meet some of the criteria of FFR calculation. In a number of comparative studies, iFR has been found to have an 80% correspondence with FFR. There remains a disparity between iFR/FFR in 20% (1/5) lesions. iFR has been shown to have good correspondence with different indices of pressure/flow (e.g., coronary flow reserve [CFR], hyperemic stenosis resistance index [HSR], etc.) which were thought to represent a surrogate standard. From the comparative trials, iFR was FDA approved for use, though in a hybrid fashion, along with FFR. Correspondence for iFR <0.84 with FFR <0.80 and iFR >0.94 with FFR >0.80 is 100%; however, this left a number of lesions with iFR between 0.84 and 0.93 that still required adenosine and FFR. A complete reliance on iFR measurements alone would require the demonstration of clinical outcomes. The results of two randomized, controlled trials were presented at the March 2017 American College of Cardiology’s Annual Scientific Sessions. I thought I would summarize the highlights of the DEFINE-FLAIR and iFR-SWEDEHEART trial results. 

Primary Trial Results 

Primary results of DEFINE-FLAIR and iFR-SWEDEHEART trials^1,2 compared a single dichotomous threshold iFR of ≤0.89 with FFR ≤0.80 in approximately 4,500 patients (Figures 1-2 and Table 1). Using these cut-offs, both methods appeared to be safe. The two studies differed very little in their design. DEFINE-FLAIR was a blinded comparison of clinical outcomes (and in the future, the cost efficiencies) of iFR-guided vs FFR-guided coronary revascularization. IFR-SWEDEHEART was an open-label, registry, randomized control trial. Both studies showed no significant difference between iFR and FFR for the same primary endpoint, which was a composite of all-cause mortality, nonfatal heart attack and unplanned revascularization at 12 months. (Tables 2-3 and Figure 3). 

Secondary Trial Results

There were shorter procedure times with iFR measurement compared with FFR. In DEFINE-FLAIR, iFR was shorter by 4.5 minutes (40.5 minutes for iFR; 45.0 minutes for FFR, P=0.001). In the iFR-SWEDEHEART study, iFR was shorter by only 2.3 minutes (50.8 minutes for iFR; 53.1 minutes for FFR, P=0.09). It is noteworthy that procedure time in iFR-SWEDEHEART was not statistically shorter than that in DEFINE-FLAIR, despite a higher rate of IV adenosine than in DEFINE-FLAIR (69% vs 59%). Since adenosine is not used for iFR, it is also no surprise that there was less procedural discomfort and fewer procedure-related (adenosine) adverse events for iFR. In DEFINE-FLAIR, the iFR group had fewer procedure-related adverse reactions, 3.1% for iFR and 30.8% for FFR (P<0.001). In iFR-SWEDEHEART, procedural discomfort was present in 3.0% for iFR and 68.3% for FFR (P<0.001). 

Despite these being randomized trials with presumably equitable distribution of ischemic lesions, fewer lesions were identified as ischemic in the iFR group than the FFR group. For DEFINE-FLAIR there were fewer ischemic lesions for iFR (28.6% for iFR; 34.6% for FFR, P=0.004); for iFR-SWEDEHEART, positive lesions for iFR were 29.9% compared to 36.8% for FFR (P<0.001). From these data, there was also a lower or similar frequency of PCI, lower or similar frequency of coronary artery bypass graft surgery (CABG) (Table 4). Given the equivalency in clinical outcomes at one year, it remains to be seen whether iFR underdiagnoses lesions that are ischemic, or FFR over-diagnoses ischemia. 

Questions Raised

There were several questions and concerns raised at the presentations. The first was about the non-inferiority study designs. Many studies like DEFINE-FLAIR and iFR-SWEDEHEART have used non-inferiority study designs, meaning the event rates for the study comparison had to fall below a prespecified number to be considered non-inferior (not equal and not superior) to FFR. Since it is known that iFR is about 80% accurate compared with FFR, in DEFINE-FLAIR, 2000 of the 2500 patients would be expected to have identical outcomes. It will be of interest to see what happens to the 500 discordant iFR/FFR patients in the future. 

For DEFINE-FLAIR, the prespecified noninferiority margin was 3.4%; for iFR-SWEDEHEART, the prespecified noninferiority margin was 3.2% — relatively large noninferiority margins. When comparing, for example, stenting with CABG or transcatheter aortic valve replacement with open surgery, clinicians are often willing to accept a slightly less robust outcome because of the invasive nature of the procedure. But when the only difference between the two methods is the administration of adenosine, perhaps such a wide margin may not be acceptable. Table 5 compares non-inferiority margins among some commonly reported studies. 

Another issue raised was that of the low-risk populations studied in both DEFINE-FLAIR and iFR-SWEDEHEART. Consider that <50% of patients  in DEFINE-FLAIR had Canadian Cardiovascular Society (CCS) class II-IV angina, compared with >75% in FAME. In addition, the average FFR was much higher in DEFINE-FLAIR/iFR-SWEDEHEART compared to FAME and FAME 2. Compare the average FFR values from imaging trials relative to iFR and FFR (Table 6). The low-risk patient population means that we should not be surprised that the 1-year event rate was lower than in the FAME studies. For comparison, we can see the difference in the 1-year MACE rates among PROMISE (computed tomography angiography [CTA] vs SPECT myocardial perfusion imaging, 3.1%), DEFINE-FLAIR/iFR-SWEDEHEART (iFR vs FFR, 6.6%) and FAME (FFR vs angiography, 3.2%). 

Finally, there was the issue of the limited availability of the iFR measurement technology. Because of proprietary software in the Philips Volcano system, iFR is not currently available for all pressure wires. Nonetheless, these data on iFR should lower the barrier to the use of physiological lesion assessment. Until iFR is widely available, Johnson et al³ advocate the use of contrast FFR (cFFR) with an optimal binary cutoff of ≤0.83, or Pd/Pa in a hybrid fashion. 

The Bottom Line 

iFR is a resting pressure-only index. Since adenosine is not used, procedure times were shorter, without adenosine-related events or symptoms. iFR is non-inferior to FFR and associated with fewer stents/CABG procedures. However, compared to FFR, there are no long-term outcome or independent ischemic testing validation data. Currently, FFR remains the comparative standard for iFR. With the use of a single iFR threshold of 0.89 (Figure 4) as a treat/no treat decision point, it is hoped that physiologic assessment in the cath lab will be more widely adopted⁴ and iFR, along with FFR and other forthcoming physiologic studies, e.g., FFR derived from cardiac computed tomography angiography (FFRCT), will provide best decision making for ischemia-guided interventions.

References 

1. Davies JE, Sen S, Dehbi H-M, et al. Use of the instantaneous wave-free ratio or fractional flow reserve in PCI. N Engl J Med. 2017; doi:10.1056/NEJMoa1700445. 
2. Götberg M, Christiansen EH, Gudmundsdottir IJ, et al. Instantaneous wave-free ratio versus fractional flow reserve to guide PCI. N Engl J Med. 2017; doi:10.1056/NEJMoa1616540. 
3. Johnson NP, Jeremias A, Zimmermann FM, Adjedj J, Witt N, Hennigan B, Koo BK, et al. Continuum of vasodilator stress from rest to contrast medium to adenosine hyperemia for fractional flow reserve assessment. JACC Cardiovasc Interv. 2016; 9: 757-767.
4. Bhatt DL. Assessment of stable coronary lesions. N Engl J Med. 2017; doi:10.1056/NEJMe17027728. 

Disclosure: Dr. Kern is a consultant for Abiomed, Merit Medical, Abbott Vascular, Philips Volcano, ACIST Medical, Opsens Inc., and Heartflow Inc.