Author + information
- Habib Samady, MD∗ ( and )
- Arnav Kumar, MBBS
- Division of Cardiology, Department of Medicine, Emory University School of Medicine, Atlanta, Georgia
- ↵∗Address for correspondence:
Dr. Habib Samady, Division of Cardiology, Department of Medicine, Emory University School of Medicine, 1364 Clifton Road F622, Atlanta, Georgia 30322.
Although atherosclerotic lesion burden and severity is a continuum, the decision to revascularize a patient or not is a dichotomous one. Therefore, like other quantitative cutoffs used in stress testing, angiography, or intravascular ultrasound imaging, thresholds of fractional flow reserve (FFR) have been sought and validated (1). Initially, an FFR value of 0.75 was established as the best threshold to discern an ischemia-provoking lesion. Subsequently, recognition of some variability in FFR measurement, both biological and technical, occurred, resulting in the description of a “gray zone” between 0.75 and 0.80, where other clinical and angiographic factors can inform revascularization decisions (2). Even though large studies, such as the FAME (Fractional Flow Reserve Versus Angiography for Multivessel Evaluation) and FAME II studies established a binary FFR cutoff of 0.80 as the threshold for clinical decision making, understanding and correcting the sources of variability in pressure measurements is critical for optimal FFR assessment.
FFR measurement has a lower coefficient of variation than other physiological tests such as coronary flow reserve. Yet, some intrinsic FFR variability related to dynamic changes in preload, afterload, coronary flow, and myocardial contractility and resistance may occur. Another issue that can impact reproducibility of FFR measurement relates to a spectrum of maximal hyperemia achieved with adenosine. Route of administration (intravenous [IV] versus intracoronary [IC]), adenosine dose, and drug–drug interaction of adenosine with other drugs, such as caffeine, can influence the extent of hyperemia. In addition, selection of the appropriate FFR value during hyperemia with IV adenosine has been debated, with either stable or minimum FFR values recommended. A recent study demonstrated several different pressure patterns of hyperemic response to IV adenosine and concluded that the minimum FFR value is likely most accurate and reproducible (3).
In addition to these biological factors, technical causes of variability include inaccurate proximal pressure, distal pressure, or pressure artifacts. Causes of proximal aortic pressure error include a ventricularized guide catheter pressure or dampening of the transduced aortic pressure waveform due to a stagnant column of contrast within the guide catheter. The major pressure wire–related problem is electronic drift of the piezoelectric sensor in the pressure wire related to longer procedural time or damage to the sensor during lesion wiring. The frequency and quantity of electronic pressure drift may vary among different pressure wire manufacturers. Other wire-related issues include pseudolesions from tortuosity or development of vasospasm related to vessel wiring. Recognizing and addressing these issues are important for accurate FFR measurement.
Given the recent recommendations for standardization and reporting of FFR (4), it is important to examine how often operators accurately measure FFR. In this issue of JACC: Cardiovascular Interventions, Matsumura et al. (5) report a post hoc analysis of the CONTRAST (Can Contrast Injection Better Approximate FFR Compared to Pure Resting Physiology) study to evaluate the incidence of suboptimal pressure measurements by comparing the interpretation of resting and hyperemic pressure measurements performed by the CONTRAST study investigators to that of a physiology core laboratory. Recall that the CONTRAST study is a multicenter study investigating the diagnostic accuracy of contrast-induced hyperemia versus standard FFR obtained with IV and IC adenosine in patients with stable coronary artery disease or in nonculprit lesions of patients with acute coronary syndromes. Resting pressure gradients, contrast FFR, and FFR measurement with IC or IV adenosine yielded 4,946 pressure tracings from 763 patients. Whether electronic drift occurred was not able to be assessed in 147 patients where pressure pullbacks were not performed. Among 616 patients who had pressure pullback data available, 108 patients (17.5%) had signal drift. To perform waveform analysis, the authors excluded the 108 patients with evidence of signal drift but included the 147 patients with no pressure pullback data, yielding 655 patients with 4,217 tracings. Among these patients, 5% had ventricularization of the pressure tracing and 4% had a distorted aortic pressure waveform. An excellent correlation was observed between investigator-reported FFR and core lab FFR (R2 = 0.969; p < 0.001) in patients with acceptable waveforms and no evidence of signal drift. They also found that significant predictors of ventricularization were hyperemic measurements with adenosine or contrast compared with resting indices and that distorted pressure waveforms were more common with the use of smaller guide catheters (5-F) and less common when IC adenosine was used. Overall, the results of this study indicate a high prevalence of imperfect FFR measurements either from signal drift or pressure wave for artifacts in clinical practice. Importantly, when these artifacts are not present, an excellent ability of operators to accurately interpret FFR measurements. These data reinforce the paramount importance of attention to detail when performing FFR measurements.
By shedding light on the high occurrence of signal drift, guide pressure ventricularization, and other distorted aortic waveforms, the investigators have made an important contribution to the field of coronary physiology. In practice, the frequency of these errors in FFR measurements may be even higher. Indeed, the data from the current analysis were derived from a carefully conducted clinical trial by experienced physiological investigators. Despite this, 20% of the time, pressure pullbacks were not performed, precluding assessment for electronic drift. The definition of drift in this study was fairly broad, that is, Pd/Pa <0.97 or >1.03. Based on these observations, it is plausible that in routine clinical practice, some degree of electronic drift is a relatively common source of error that can shift a binary FFR threshold around the 0.80 value. Similarly, ventricularization or distortion of the aortic waveform is likely to occur in almost 10% of cases in clinical practice and introduce another avoidable error in FFR measurement. These artifacts may accentuate the discordance between clinicians’ angiographic assessment and their experience with the measured FFR value, perhaps contributing to lack of adoption of physiological assessment.
Despite these concerns, one should be reassured by the consistent improvement in outcomes in real practice registries of patients who undergo physiological- rather than angiographic-guided revascularization. In addition, compared with other frequently performed clinical tests, the reported variability in FFR measurement (3%) is relatively modest (e.g., fasting plasma glucose was 9%, ambulatory systolic blood pressure 11%, ejection fraction assessment by magnetic resonance imaging 12%, and quantitative angiographic diameter stenosis 18%). Industry colleagues need continue to develop and iterate the piezoelectric sensor technology to reduce the incidence of signal drift as we use pressure wire technology for more complex disease. The recent introduction of optical sensors for FFR measurement has been a significant step forward in reducing signal drift.
A few simple steps would reduce the likelihood of technical errors:
• Avoid 5-F guide catheters and flush the pressure wire in the hoop with saline before inserting in the guide catheter (reduces likelihood of drift).
• Remove the insertion needle and flush the guide catheter with saline before equalization and measurements.
• Disengage the guide catheter during FFR measurement to avoid ventricularization of the proximal pressure.
• If a significant hyperemic gradient is observed, the distal pressure should appear more ventricularized with loss of the dicrotic notch. If not, consider electronic drift.
• If IV adenosine is used, a proximal large-bore IV should be used for consistent drug delivery. If IC adenosine is used, FFR measurement should not be made during the adenosine or saline flush period.
• Remeasure the distal and aortic guide pressure after FFR measurement to check for drift.
• If the pressure wire had drifted, adjust the FFR value by correcting the amount of drift from the distal pressure and recalculating the FFR value. If in doubt, re-equalize and re-measure the FFR.
• Report the minimum FFR value once artifacts are excluded.
Incorporating these simple steps into our physiological measurement in the catheterization laboratory will assure that we can harness the full value of these powerful, simple, yet nuanced, physiological tools in the cardiac catheterization laboratory.
↵∗ Editorials published in JACC: Cardiovascular Interventions reflect the views of the authors and do not necessarily represent the views of JACC: Cardiovascular Interventions or the American College of Cardiology.
Dr. Samady has received institutional research funding from Volcano Phillips, St. Jude Medical, Medtronic, and Abbott Vascular; and honoraria for being on steering committees of studies funded by Volcano Corporation and St. Jude Medical. Dr. Kumar has reported that he has no relationships relevant to the contents of this paper to disclose.
- 2017 American College of Cardiology Foundation
- Kern M.J.,
- Samady H.
- Johnson N.P.,
- Johnson D.T.,
- Kirkeeide R.L.,
- et al.
- Vranckx P.,
- Cutlip D.E.,
- McFadden E.P.,
- Kern M.J.,
- Mehran R.,
- Muller O.
- Matsumura M.,
- Johnson N.P.,
- Fearon W.F.,
- et al.