Despite the current use of fractional flow reserve (FFR) and instantaneous wave-free ratio (iFR) for guiding revascularization in atrial fibrillation (AF), there is a lack of studies evaluating their reliability in this particular population. This retrospective study aimed to investigate the reliability of FFR and iFR in patients with AF. This retrospective study included all patients with AF undergoing FFR measurements (n = 45 vessels from 36 patients) at Brugmann University Hospital, Brussels, Belgium, between 2012 and 2020 or iFR (n = 18 vessels from 13 patients) and a corresponding number of patients with sinus rhythm (SR) randomly selected from the same period, benefiting from iFR (n = 20 vessels from 17 patients) or FFR (n = 50 vessels from 37 patients). Our main findings indicate that there is an increased beat-to-beat variability of individual iFR measures in patients with AF, compared with SR. In addition, the reproducibility of iFR on test-retest is low in patients with AF, leading to increased lesion reclassification (53.8% of lesions reclassified on 2 consecutive iFR measurements in AF vs 6.6% lesions reclassified in SR, p <0.05). In contrast, FFR seems to be more robust in evaluating coronary lesions in AF in terms of equivalent variability, reproducibility, and lesion reclassification observed in the SR population. In conclusion, this is the first study to evaluate the reliability of iFR and FFR in AF. Our findings raise caution in using iFR to guide revascularization in patients with AF, whereas FFR seems to be more robust in this population.
Despite the current use of fractional flow reserve (FFR) and instantaneous wave-free ratio (iFR) for guiding revascularization in patients with atrial fibrillation (AF), there is a lack of studies evaluating their reliability in this particular population. There are particular points of concern that require further investigation in terms of validating the use of FFR and iFR in patients with AF, such as the hemodynamic variability. The VERification of Instantaneous Wave-Free Ratio and Fractional Flow Reserve for the Assessment of Coronary Artery Stenosis Severity in EverydaY Practice study showed that iFR is susceptible to more variation than FFR in relation to heart rate and blood pressure at rest. Another study indicated that elevated heart rate is associated with a lower iFR and discordant iFR and FFR. In patients with rapid ventricular rate associated with AF, the increased myocardial oxygen demands induce a pseudo-hyperemic state that may influence the FFR and iFR measurements. In addition, patients with AF appear to present microvascular coronary dysfunction, even within normal ventricular rate range, which can interfere with FFR and iFR measurements. The aim of this study is, as such, to investigate the reliability of FFR and iFR in patients with AF.
This study included patients with AF or sinus rhythm (SR) who benefited from FFR or iFR to evaluate coronary stenosis at Brugmann University Hospital, Brussels, Belgium, between 2012 and 2020. The exclusion criteria were: cardiac arrest, cardiogenic shock, acute decompensated heart failure, acute ST-segment elevation myocardial infarction, and culprit lesion in non-ST–segment elevation myocardial infarction. We included all patients with AF undergoing iFR (n = 18 vessels from 13 patients) or FFR (n = 45 vessels from 36 patients), and a corresponding number of patients with SR benefiting from iFR (n = 20 vessels from 17 patients) or FFR (n = 50 vessels from 37 patients). The patients in SR were randomly selected from the same period as the corresponding AF population, using the online research randomizer “ http://www.randomizer.org .” We excluded all patients with SR who presented previously with paroxysmal AF. The study was approved by the Ethical Committee of CHU Brugmann (reference number CE 2020/138). iFR and FFR measurements were obtained with the use of Verrata or Prestige coronary pressure guidewires (Philips Volcano). Measurements were performed under basal conditions and/or after hyperemia induced by intracoronary administration of adenosine (initial dose of 150 µg, increased progressively until attaining the steady-state of hyperemia). Criteria for revascularization were FFR ≤0.8 or iFR ≤0.89.
To evaluate the beat-to-beat variability during FFR or iFR measurements, we derived distal coronary pressure to aortic pressure ratio (Pd/Pa) and iFR values across consecutive beats spanning the interval analyzed by the software and calculated the coefficient of variation. For the iFR measurements, we calculated iFR values on each beat considered in the analysis by dividing the area under the Pd curve, to the area under the corresponding Pa curve, for the number of beats considered in the analysis, per the iFR calculus algorithm. In addition, we derived the heart rate and the duration of the wave-free period (WFP) on each beat. Image analysis was performed using ImageJ software (National Institutes of Health). For FFR measurements recorded using the ComboMap software, we calculated individual Pd/Pa values for 7 beats centered around the minimum Pd/Pa identified by the software analysis. Beat-to-beat variability measurements were performed for the patients where all required information was available. Coefficients of variation were calculated for each iFR/FFR measurement by dividing the standard deviation of the individual beats measured to the average value and multiplying by 100.
To perform a quantitative assessment of the coronary stenosis, we calculated the percentage of the diameter at the level of the maximum narrowing versus the diameter of the healthy vessel before stenosis. Image analysis was performed using ImageJ software (National Institutes of Health).
To evaluate the reproducibility at test-retest of FFR and iFR in patients with AF and SR, we derived Bland-Altman plots using 2 consecutive FFR/iFR measurements and correlated the 2 repeated measures using Spearman correlations. For evaluating the reliability of lesion classification in patients with AF and SR depending on the coronary pressure indices, we compared the proportion of patients with lesion reclassification. Lesions were reclassified if upon first measure they were above the threshold (0.8 for FFR and 0.89 for iFR) and during the second below the threshold, or vice-versa. In these analyses, we included the patient subpopulation in AF or SR who benefited from 2 consecutive iFR or FFR measurements.
Statistical analyses were performed using GraphPad Prism 6.1 software (GraphPad Software). Data are presented as median (25th to 75th percentile). When analyzing 2 independent groups we employed the Mann-Whitney test. Categorical variables were compared using Fisher’s exact test. Correlation analyses were performed using Spearman’s correlation test. The significance level was set as 0.05.
Baseline clinical and lesion characteristics of the retrospective population are presented in Supplementary Table 1 . The heart rate at the beginning of the coronary angiography was increased in both AF groups (AF-FFR and AF-iFR), compared with the corresponding SR groups. In addition, the body mass index was increased in the AF-FFR group, compared with the SR-FFR group, the percentage of patients previously receiving percutaneous coronary intervention was lower in the AF-iFR group than in the SR-iFR group, and the percentage of patients with left ventricular ejection fraction under 35% was higher in the AF-FFR group than in the SR-FFR group. None of the other variables were significantly different between the 2 groups ( Supplementary Table 1 ).
The coefficient of variation of beat-to-beat measurements was equivalent between AF-FFR and SR-FFR groups (AF-FFR 1.74 [1.09 to 2.60] % vs SR-FFR 2.52 [1.17 to 3.32] %, p = 0.48; Mann-Whitney test) ( Figure 1) . In contrast, we could observe a significantly increased beat-to-beat variability in iFR measurements in the AF versus SR population (AF-iFR 2.65 [1.33 to 4.04] % vs SR-iFR 0.69 [0.24 to 1.98] %, p <0.01; Mann-Whitney test) ( Figure 1 ). The coefficient of variation of iFR in AF correlated positively with the variability of heart rate (ρ = 0.5444, p <0.05; Spearman correlation) and the variability of the duration of the WFP (ρ = 0.6120, p <0.01; Spearman correlation) ( Figure 1 ). Furthermore, the ComboMap software successfully identified the minimum Pd/Pa value as the FFR after adenosine administration for 87.5% of patients in the AF group, and 90% of patients in the SR group, and hence only rare instances when the true minimum was found outside the search interval by manual assessment. When the minimum was found outside the search interval, it differed from the automatically generated minimum by a value of 0.01 in both the AF and SR groups (data not shown).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
