Instantaneous wave-free ratio and fractional flow reserve in clinical practice

A prospective registry at Zuyderland Medical Centre of all-comers between November 2015 and February 2017 in which intermediate coronary stenoses (i. e. 50–90% diameter stenosis by visual assessment) were measured using both iFR and FFR.

Coronary angiogram was acquired via either a radial (preferred) or femoral approach with administration of, 50 U/kg unfractionated heparin. When a radial approach was used an intra-arterial vasodilator cocktail (nitroglycerin 100 mcg and verapamil 2.5 mg) was administered. A 0.014-inch pressure sensor-tipped wire (PrimeWire Prestige, Volcano Corporation, San Diego, USA) was positioned at the tip of a guiding catheter. After pressure equalisation at the tip of the catheter, the wire was advanced into the target vessel as distally as reasonably possible for pressure recordings. First, iFR was automatically calculated online using the Volcano CORE System version 3.3.0 (Volcano Corporation). Subsequently, FFR was measured during adenosine-induced hyperaemia either via central intravenous administration [at 140 μg/kg/min] or an intracoronary bolus 100–150 μg. At the end of each measurement, the pressure sensor was retracted to the catheter tip to preclude pressure drift. Use of intracoronary nitroglycerin injection was left at the discretion of the cardiologist. Clinical decisions were based exclusively on the currently recommended FFR treatment cut-off value of ≤0.8 because at that time the results from the DEFINE-FLAIR [6] and iFR-SWEDEHEART [7] trials were unknown.

The gold standard consisted of an FFR-only approach using a cut-off value of 0.8 to defer or treat when the measurement was higher or lower, respectively. For the iFR-only strategy, we used a cut-off value of ≤0.89 based on limited available data [13]. Finally, we tested the proposed hybrid iFR-FFR approach [13] incorporating an ‘iFR grey zone’ to revascularise (iFR <0.86), defer percutaneous coronary intervention (>0.93) or to require a subsequent FFR measurement to decide (iFR 0.86–0.93).

We used SPSS statistical software version 22.0 (SPSS Inc., Chicago, Illinois) to perform data analysis. Continuous variables are reported as mean (SD) or median (25^th–75^th percentiles) and categorical variables as number of observed patients (percentage). We used Fisher’s exact test when we compared categorical variables between groups and the Student’s t test when we compared normally distributed continuous variables between two groups. If the continuous variable did not follow a normal distribution, we used the Mann-Whitney U test when we drew a comparison between two groups. Correlation between FFR and mean iFR was assessed with Spearman’s rank correlation coefficient (rs). Conventional summary statistics for diagnostic tests, compared with a patient’s true disease status as indicated by FFR ≤0.80, were calculated from a 2 × 2 contingency table, comparing either the iFR-only strategy or the hybrid iFR–FFR strategy with standard FFR. The area under the receiver operating characteristic (ROC) curve was assessed through nonparametric ROC analysis. Subsequently, the optimal mean iFR threshold was verified using the minimally important change (MIC) threshold as the cut-off level, corresponding to a 45-degree tangent line intersection.

The investigation conforms with the principles outlined in the Declaration of Helsinki.

Our objective was to provide additional evidence for the clinical, real-time use of iFR measurements in functional assessment of intermediate coronary artery stenoses. These data are in line with recent observations that iFR has the potential to become such a diagnostic tool [5‐7, 12, 13]. All iFR measurements were technically feasible, readily available and a single iFR measurement sufficed. The iFR-only strategy was based on the ‘non-clinically’ derived cut-off of 0.89 and resulted in similar diagnostic agreement with FFR, lower specificity and positive predictive value but higher sensitivity and negative predictive value compared with prior studies [13]. However, following the result of DEFINE-FLAR [6] and iFR-SWEDEHEART [7] the appropriateness of FFR as the gold standard is questionable. Perhaps the clinical trial iMODERN (iFR Guided Multi-vessel revascularizatiOn During percutaneous coronary intervEntion for acute myocaRdial iNfarction; NCT03298659) can provide insight into this matter.

Adopting the previously suggested hybrid approach with an iFR cut-off value for revascularisation (0.86) and deferral (0.93) resulted in an expected substantial improvement in diagnostic agreement. However, when relying upon a hybrid strategy as the solution the outset should be to minimise irreversible actions, in other words, inappropriate revascularisation and its sequelae such as antiplatelet therapy.

The diagnostic agreement between FFR and iFR depends on the used cut-off values. Whereas some might argue there is room for debate regarding the optimal FFR cut-off [15, 16], this is particularly true for iFR cut-off values. Although an iFR-guided revascularisation strategy was noninferior to FFR-guided revascularisation in the trials reported by Davies et al. and Götberg et al. [6, 7], outcomes in patients with iFR-guided deferral of revascularisation were not reported. If indeed the clinical outcomes were similar, interventional cardiologists would have more confidence in deferring revascularisation if the iFR is higher than 0.89, and these findings would help to encourage transition to a sole iFR-guided strategy.

Both the iFR-only strategy (0.89) and the ‘iFR grey zone’ (0.86–0.93) are established based on a model, not on clinically derived values [8, 9]. Within our cohort the MIC-derived optimal iFR cut-off value was slightly lower compared with the applied, in other words the accepted, cut-off, opposed to a previous study showing identical values [13]. A recent study by Kobayashi et al. showed that the diagnostic accuracy of iFR depends on the location of the lesion in the coronary tree [17]. In particular, the diagnostic accuracy of iFR was significantly lower than that of FFR for lesions located in the left main or proximal left anterior descending coronary artery; this is probably related to the larger amount of myocardium supplied [17]. This difference may have clinical relevance. Altogether it appears that the hybrid functional assessment of coronary artery stenoses could benefit from more clinical data on optimal cut-off values.

Discordant measurements consisted mainly of false-positive results which is in contrast with prior studies [8, 14]. Further analysis did not reveal a significant difference in patient or haemodynamic characteristics and this observation might be best explained by microvascular dysfunction. We corroborate prior evidence of significantly higher iFR and FFR values in the right coronary and circumflex artery with their branches [data not shown] [13]. Given the vast majority of discordant iFR and FFR measurements consisted of false positive results, these two observations could be linked. How the latter can be explained, physiologically or otherwise, remains as interesting as speculative and requires dedicated further research such as the recently initiated study FiGARO (FFR versus iFR Assessment of Hemodynamic Lesion Significance; NCT03033810).

Additionally, a previous study with direct comparisons between iFR and FFR found also a mismatch in about 20% of cases and they found iFR appeared to correlate better with flow measurements (coronary flow reserve) than FFR [18].

iFR measurements eliminate the necessity for adenosine exposure and thereby the associated side effects, additional time and costs as proven by the studies iFR-SWEDEHEART (Evaluation of iFR vs. FFR in Stable Angina or Acute Coronary Syndrome) and DEFINE-FLAIR (Functional Lesion Assessment of Intermediate Stenosis to Guide Revascularisation) [6, 7]. Our data are in line with the evidence. This alone could stimulate more systematic use of functional stenosis assessment. Which is important as it not only answers the question whether or not to revascularise, but also impacts the preferred method of revascularisation [19]. However, the added value of iFR measurements could extend to at least two other important clinical settings. Maybe this is not the case for diffusely diseased coronary arteries or tandem lesions (although FFR is very well suited for single ‘spot’ or non-complex lesions). In such cases a pullback is required to determine the culprit section. Considering that the hyperaemic flow, but not the resting flow, of tandem lesions are interdependent, iFR measurements provide both a more practical and better physiological roadmap of the entire coronary artery [20].

Second, the book on culprit lesion versus complete revascularisation in the setting of ST-elevation myocardial infarction or non-ST-elevation myocardial infarction is still not closed [21]. As such, iFR measurements could potentially provide an adenosine-free practical alternative for which the results of the WAVE trial (Instantaneous Wave-Free Ratio and Fractional Flow Reserve for the Assessment of Non Culprit Lesions in Patients with ST-segment Elevation Myocardial Infarction; NCT02869906) trial are much anticipated. Although iFR may be non-inferior to FFR, as mentioned above, there are some concerns about iFR: first, experimental studies supporting the hypothesis behind iFR are lacking; second, the existence of the wave-free period, upon which iFR is based, has been questioned [22]; and third, outcome studies that compare with FFR have been primarily performed in low-risk patients.

Finally, despite clear non-inferiority (but not independent superiority) of an iFR compared to an FFR guided revascularisation strategy on key clinical endpoints [6, 7] future research to improve current intracoronary diagnostic measurements is necessary to improve our understanding of discrepancies, cut-off values and consequently coronary revascularisation and clinical outcome in high-risk patients [23].

First, as certain technical aspects (e. g. arterial access, route of adenosine administration and subsequent catheter position check) were left at the discretion of the operator this potentially introduced bias, in particular the use of intracoronary nitrates considering its effect beyond the vasotonic constriction of the stenosis. Second, despite the many relevant patient and haemodynamic characteristics that we collected, some important aspects with regard to analysing FFR discordance and concordance could have been omitted. Therefore, based on these data, the possibility of a significant interaction cannot be eliminated. Third, even though these data are from a large all-comer prospective registry, it is not a randomised clinical trial. In other words, all stenoses were assessed by both iFR and FFR and the FFR measurement was leading in the decision to revascularise or defer percutaneous coronary intervention.