Editorials
Fractional Flow Reserve, Coronary Pressure Wires, and Drift
2016 Volume 80 Issue 8 Pages 1704-1706
Details
2016 Volume 80 Issue 8 Pages 1704-1706
Among all the diagnostic tests used in interventional cardiology, few have had the impact of fractional flow reserve (FFR).1 Today, FFR is used as the standard of reference to decide if stenting of a coronary artery is appropriate. Especially in complex multivessel disease, its use is indispensable to guide the operator if and where coronary stents should be placed.
Article p 1812
FFR was introduced more than 20 years ago2,3 and its value has been established in almost all subsets of coronary disease and clinical conditions and has been used clinically in millions of patients. There is incontrovertible proof that by using FFR, both the symptoms and outcome of patients with coronary artery disease are significantly improved and that its use is cost-saving.4–9
The practical value of FFR to guide coronary revascularization, the background of the concept, its practical set-up, and its importance for clinical outcome have been described in a myriad of papers, among others a review article in this Journal in 2013.10
However, measurement of coronary pressure by sensor tipped pressure wire has been associated with drift since the early days. Drift has always been considered as inevitable in electronic measurement equipment. Millar catheters and also the initial prototypes of the 0.018˝ and 0.014˝ electronic pressure wires showed considerable drift. Occurrence of drift is the most annoying problem that can occur during a procedure in a patient because it is often unnoticed before the wire is pull backed to the guiding catheter at the end of the procedure and if it is present, it may invalidate the measurement.
Over time, tremendous efforts have been made by the manufacturers of electronic pressure wires (Radi, St. Jude Medical; Volcano, Philips) to decrease drift and these attempts have been successful to a large degree. Nevertheless, some drift might still occur every now and then, and confuse operators.
In addition, it should be realized that, not infrequently, a difference between the pressure wire signal and the guiding catheter signal at the end of the procedure is not true drift associated with the pressure wire, but apparent drift related to inappropriate zeroing and equalizing, to the presence vs. absence of the introducer needle when checking for drift, to capillary forces in the guiding catheter affecting the signal of the guiding catheter, or to drift occurring in the fluid-filled transducer.
Over the years, operators have learned how to deal with this, how to avoid drift as much as possible, how to correct for drift, and have accepted it as a minor annoyance during an otherwise satisfactory procedure.
Surprisingly, only a few systematic studies have been performed to assess the effect of drift on decision making and in this respect, the paper published in this issue of the Journal by Wakasa et al is a valuable contribution.11,12
Impact of Drift on FFR-Based Decision MakingThe study by Wakasa et al12 is the largest study performed so far to assess the influence of pressure drift on decision making. In their study, 940 patients were included and a total of 1,218 coronary arteries were studied. Drift was meticulously assessed during state-of-the-art procedures with correct equalization at the beginning of the procedure and making a hyperemic pullback recording at the end to confirm the presence or absence of drift. Recordings with pressure drift ≥4 mmHg (which is not easily overlooked and mandates new equalization and measurement) were excluded and those with a drift ≤3 mmHg were further analyzed and the influence on decision making was studied. Special attention was paid in the analysis to values close to the presently used binary cut-off value of 0.80 and the grey zone of 0.76–0.80.
Importantly, the study population was adequate, as more than 70% of the patients had FFR values in the range of 0.60–0.90. In almost normal or extremely depressed values of FFR, drift will have negligible importance.
As can be expected, with FFR values close enough to the binary cut-off value of 0.80, considerable crossover occurred because of drift. But it should be realized that this is not surprising but merely a statistical phenomenon. Whatever the cut-off value might be, no matter how accurate the measurement, no matter how small the drift, if one approaches close enough to the binary cut-off value, the number of cases in which a crossover of the cut-off value occurs will always be 50%. That is a statistical phenomenon and has nothing to do with the value of an index, the variability, or magnitude of drift. Important, however, is that the range of ambiguity should be narrow, as indicated in Figure 4 (green bars) in their report.12 And fortunately, drift was limited enough to create confusion in less than 20% of the patients with FFR values between 0.76 and 0.82, a rather limited interval. Even more important, only 3.6% of clinical decisions changed when a binary threshold of 0.80 was used, and in not a single patient did the value shift from above the upper limit of the grey zone (0.76–0.80) to below the lower limit or vice versa. This latter observation is of utmost importance because it underlines that although binary thresholds are easy, they do not exist in nature with 100% certainty and that sound clinical judgement remains necessary.
Binary Threshold of 0.80 vs. Grey Zone of 0.76–0.80In the initial studies of FFR, a threshold of 0.75 was used and had a specificity of 100% to indicate inducible ischemia and a sensitivity of 90%.3,7
A small grey zone between 0.76 and 0.80 was recognized from the beginning and in more than 30 studies searching for thresholds in different clinical conditions and different subgroups of patients, optimum cut-off values were always found between 0.75 and 0.80.10
In later studies, in order rather to end up with some overtreatment instead of undertreatment of patients, the upper boundary of the grey zone of 0.80 was used as a cut-off value,4–6,8 resulting in a sensitivity of 100% and a specificity of 90%.
With FFR values between 0.76 and 0.80, both clinical and technical factors determine whether a stent will be placed or not. If the complaints of the patient are typical, other evidence of ischemia is present or a single large pressure drop is present within the coronary artery, placing a stent is attractive. If complaints are atypical, no other evidence of ischemia is present, or when on the hyperemic pullback recording a very diffuse decline in pressure is observed, stenting becomes unattractive. Nothing is perfect, not even FFR, and measuring it does not rule out sound clinical judgement. And although using one binary cut-off value (0.80 at present) is attractive, it should be realized that the accuracy of FFR in that case is approximately 95% which means that in some patients, further clinical or technical factors need to be taken into consideration, as described before. And in that respect, it is reassuring to learn from the study by Wakasa et al that less than 4% of the patients cross the presently used binary cut-off threshold of 0.80 because of drift and that none of their patients moved across the grey zone.12
These results do not mean that we can lean back and accept drift as present in most pressure wires. But at least we can work with it as long as we realize that in the cases of obvious drift exceeding 3 mmHg, measurements need to be repeated or corrections have to be made as described next.
Why Does Drift Occur and How Can We Minimize It?An excellent overview on the background of drift has been published recently by Dr Morton Kern.13 True drift in pressure wires occurs because of entrapment of small air bubbles in the cavity of the pressure sensor and because of electrical and thermal instability inherent to all types of sensors. We often recommend that after having equalized the electronic pressure wire (when the sensor is at the tip of the guiding catheter), wait for 20–30 s before further advancing the wire into the coronary artery. Often, that gives time for small air bubbles to be flushed away from the cavity and if the signal changes, a renewed equalization can be done before entering the coronary artery. Thermal instability is inherent to all electronic equipment and difficult to avoid. In the St. Jude Medical pressure wire, a thermal correction mechanism is present with the great advantage that it also enables simultaneous temperature measurement, presently advocated to assess absolute coronary blood flow and myocardial resistance in addition to the FFR measurement.14–16 As thermal drift mostly occurs linearly over time, at the end of a long procedure with drift observed when the sensor is back at the tip of the guiding catheter, it can be assumed safely that drift during the last pullback recording was negligible and that the observed drift can either be added or subtracted from the distal coronary pressure signal to calculate FFR. However, when in doubt, no other solution is present than equalizing again, entering the artery again with the wire, and to repeat the measurement.
In addition to true drift, differences between the sensor signal and the guiding signal at the end of the procedure can be caused by a number of phenomena not related to the pressure wire but to either the guiding catheter or operator-dependent mistakes. Capillary drift in the guiding catheter can mostly be corrected by vigorous flushing of the guiding catheter with 5–10 ml of saline. Also, the operator should realize that when the initial diagnostic FFR measurement was made with an introducer needle present and the final measurement after stenting was made without, this can account for a difference of 5–10 mmHg. Further details on how to deal with drift have been described extensively in the literature.1,10
Influence of Drift on Hyperemic and Resting Indices: iFR, Pd/ParestWhen studying the influence of drift on clinical decision making, it is important to realize that drift is an absolute phenomenon (mmHg/unit of time) and therefore has greater influence when dealing with smaller gradients as are usually present at rest, such as in the assessment of iFR and Pd/Parest.
In general, hyperemia increases pressure gradients within a coronary artery by a factor 2.5–3, especially for stenosis in which coronary pressure measurement is most important (in almost normal or very severe stenotic arteries, differences between resting and hyperemic gradients are usually small). By definition, this means that in terms of the influence of drift, the signal to noise ratio during hyperemia is 2.5–3-fold higher (better) than with resting gradients. As a consequence, it can be expected that measurement of iFR and Pd/Pa at rest are more vulnerable to the influence of drift compared with (hyperemic) FFR.
Such observations have also been made qualitatively in an earlier study by Cook et al without reaching significant differences because of the smaller number of patients.11
Electronic vs. Fiberoptic SystemsAs mentioned earlier, tremendous efforts have been made by the manufacturers of pressure wires to reduce drift. For the US FDA, acceptable signal drift should be less than 5 mmHg/10 min, which is too liberal in our view. St. Jude Medical indicates drift of <7 mmHg/h and every individual wire is tested to fulfil this specification. In our very extensive experience with the St. Jude Medical Certus and Aeris wires, drift is absent in 70% of the wires and occurs in 30% of the wires and is indeed below that limit in almost all cases. Philips/Volcano only specifies to comply with US FDA requirements but does not further specify drift in their IFU.
With the introduction of optical wires, less drift was expected. For the Optowire® (Opsens Inc, Quebec City, Canada) drift is negligible indeed. In our laboratories, we used approximately 100 of these wires in the past year and we have not observed any drift in any of these wires up to now. We have a very limited experience with the Acist Navvus® system (not a true pressure wire but an over-the-wire catheter to measure FFR before and after a procedure, but unable to monitor pressure during the procedure), and did observe some drift up to 5 mmHg in several of these catheters. The same holds true for our limited experience with the Comet® wire (Boston Scientific, Minneapolis, MN, USA), which showed a non-negligible bi-directional and fluctuating drift in 12/20 procedures we performed so far and in magnitude not different from the electronic wires. The differences between the different optical systems are most likely related to differences in cavity configuration and membrane technology, but it is too early to draw definite conclusions in this respect and the position of the different systems has to be further established in the next years. It is not unlikely that as a result of increased competition, further refinements will be made and St. Jude Medical has announced the introduction of a new pressure wire (PressureWire X) within the next months.
For our field of interventional cardiology, these developments are good news, contributing to even easier and more widespread usage of FFR in the future.
ConclusionsSince the first introduction of coronary pressure wires, drift has been a problem that has been reduced but not eliminated over the years. For a long time it was accepted as inevitable, necessitated us to repeat measurements after re-equalization in cases of very obvious drift (>3 mmHg) and to neglect this when it was smaller. The study by Wakasa et al in the present issue of the Journal shows us that fortunately such a policy does not influence clinical decision making too much.12 Less than 4% of the patients in their large series crossed the binary cut-off threshold of 0.80 by drift and none of their patients moved from below the lower limit of the grey zone to above the upper limit or vice versa.
Nevertheless, further efforts by the manufacturers of pressure wires to eliminate drift should be encouraged. Recent developments, among which is the fiberoptic Optowire®, makes us confident that this goal can be achieved.
N.H.J.P. received institutional research grants from St. Jude Medical, Abbott, and Maquet and is a consultant for St. Jude Medical, Boston Scientific, Opsens. B.d.B. received institutional research grants from Abbott, Boston Scientific, and Biotronik, and the Cardiovascular Research Center Aalst receives consultancy fees on his behalf from St Jude Medical, Boston Scientific, and Opsens.
