Variation in Fractional Flow Reserve (FFR) and Its Effects on Stenting Decision Making

FFR is an important construct using intra‐arterial pressures to describe blood flow past partial stenoses in coronary arteries to determine the need for percutaneous coronary intervention. Point estimates of the FFR can be compared to two threshold values to determine three decision "zones ": FFR <0.75 such that stenting (S) of this lesion should be performed, 0.750.80 which indicates that no stenting (NS) of the lesion should occur at this time. Variation of FFR measurements would be expected due to factors at the operator, protocol, and instrument levels. Subject physiology, such as the level of venous pressure, collateral flow, heart rate and size /geometry of the stenosis, can also have an effect. This variation might lead to errors in the decision making process. However, FFR variability is poorly described in the literature and trade documents. This study was designed to explore the role of FFR variability on probabilities of errors in decision making when measuring a FFR value that was considered to be normally distributed about 0.74, 0.78 and 0.81 (being examples of true values that may occur in each of the decision zones) and with four different standard deviations (SD) in the measure of FFR. Normal distributions of these FFR were generated with SD of 0.01, 0.02, 0.029 and 0.05 with resampling software (J. Grosberg, Statistics101 Version 4.9). These SD values were based on repeated FFR measures on 6 subjects (Aalst, GRH), from which an average deviation of 0.029 was found. The 0.01, 0.02 and 0.05 SD values were included to further explore their effect on FFR distributions. Probabilities of FFR measures that would be found in zones S, AE and NS were generated to show the frequency of erroneous versus true FFR values. The results for each chosen FFR and SD combination are given in the accompanying table, showing probabilities of FFR measures in the S, AE and NS zones. In general, the greater the variation as described by the SD, the greater is the probability of any single measurement being incorrect (i.e., being found in a zone different from its true value). The variability in FFR measurement should be quantified since it has bearing on possible errors in decision making as well as serving guidance in technique improvement. Efforts to assure low variability are important.FFR 0.74 SD Zone S AE NS(true S)0.010.841 0.159 0.0000.020.692 0.307 0.0010.0290.635 0.346 0.0190.050.579 0.306 0.115FFR 0.78SD(true AE)0.010.001 0.976 0.0230.020.066 0.775 0.1580.0290.151 0.604 0.2450.050.275 0.380 0.345FFR 0.81SD(true NS)0.010.000 0.158 0.8420.020.001 0.308 0.6910.0290.019 0.345 0.6360.050.115 0.306 0.579