Author + information
- Received April 23, 2002
- Revision received December 23, 2002
- Accepted January 30, 2003
- Published online May 7, 2003.
- Takashi Akasaka, MD*,* (, )
- Atsushi Yamamuro, MD†,
- Norio Kamiyama, MD*,
- Yuji Koyama, MD*,
- Maki Akiyama, MD*,
- Nozomi Watanabe, MD*,
- Yoji Neishi, MD*,
- Tsutomu Takagi, MD†,
- Evgeny Shalman, PhD‡,
- Chen Barak, PhD‡ and
- Kiyoshi Yoshida, MD, FACC*
- ↵*Reprint requests and correspondence:
Dr. Takashi Akasaka, Department of Cardiology, Kawasaki Medical School, Matsushima 577, Kurashiki City, Okayama 701-0192, Japan.
Objectives This study sought to assess the reliability of pressure-derived coronary flow reserve (CFR) compared with flow- or velocity-derived CFR.
Background Coronary flow reserve has been reported to have important clinical implications for the evaluation and treatment of coronary artery disease.
Methods Using a pressure guide wire, coronary pressure distal to the stenosis was measured at rest and during hyperemia in seven dogs with various degrees of stenosis and in 30 patients with angina (29 and 34 stenoses in total, respectively). Pressure at the tip of the guiding catheter was also recorded with a fluid-filled transducer system. Pressure-derived CFR was calculated by the square root of the pressure gradient across the stenosis (ΔP) during hyperemia divided by ΔP at rest, using a proprietary software system. At the same time, coronary flow was monitored proximal to the stenosis with a flow meter in the experimental dogs, and coronary flow velocity distal to the stenosis was assessed using a Doppler guide wire in patients with angina. Flow-derived (or velocity-derived) CFR was compared with pressure-derived CFR.
Results Except for one stenosis that showed no ΔP at rest, a significant correlation was obtained between pressure- and flow-derived CFR in the animal study (y = 1.05x − 0.03, r = 0.92, p = 0.0001). A significant correlation was also seen between pressure- and velocity-derived CFR in the human study, except in three stenoses with no resting ΔP (y = 0.70x + 0.37, r = 0.85, p = 0.0001).
Conclusions Similar to flow (or velocity) measurement, CFR can be assessed by pressure measurement, except in stenoses with minor resting ΔP.
The limitations of coronary angiography to estimate stenosis severity are well known (1–3). Intracoronary physiologic measurement of coronary flow reserve (CFR) or myocardial fractional flow reserve (FFR) has been described to complement anatomic assessment of the coronary lumen (4–8). Up to now, in cardiac catheterization laboratories, CFR has been assessed by coronary flow velocity measurements with a Doppler guide wire (7), and FFR has been derived from pressure data using a pressure guide wire (7,8). It should be much more convenient, practically and economically, if CFR and FFR could be obtained simultaneously with one sensor-tipped guide wire alone.
Recently, two different reports have been published in which the assessment of CFR was performed by using a pressure guide wire alone: the thermodilution method (9)and the pressure gradient method (10). Coronary flow reserve could be measured by the ratio of the mean transit time of saline to travel from the injection site to the distal sensor at rest divided by the mean transit time during hyperemia by the thermodilution method (9). Coronary flow reserve might be obtained as the ratio of the square root of the pressure gradient across the stenotic lesion (ΔP) during hyperemia divided by the ΔP at rest using the other method (10). Although each method has represented a significant correlation with flow-derived CFR in experimental models (9,10), injection of the indicator (saline or 5% glucose) would be required both at rest and during hyperemia in the thermodilution method. However, with the pressure-derived CFR method, assessment of the pressure gradient at rest and during hyperemia would be enough to obtain the CFR. If this pressure-derived CFR could be demonstrated to be a reliable index of CFR, physiologic indexes of CFR and FFR could be obtained at the same time in humans, with much more ease, using a pressure guide wire alone.
To assess the reliability of pressure-derived CFR, these measurements were compared with flow-derived CFR obtained by a Doppler flow meter in an experimental animal study and by a Doppler guide wire in conscious humans.
Animal experimental study
Various degrees of coronary stenoses were made in random fashion by using a coronary occluder in the mid portion of the left anterior descending coronary artery (LAD) in seven mongrel dogs (body weight 22 to 26 kg) on mechanical ventilation during general anesthesia by intravenous pentobarbital sodium (25 mg/kg). All dogs were premedicated with ketamine (10 mg/kg intramuscularly), and respiration was supported mechanically using a mixture of oxygen and room air after endotracheal intubation. Coronary artery pressure distal to the stenosis was recorded by a pressure guide wire (Pressure Wire, Radi Medical System Inc., Uppsala, Sweden) inserted through a 6F guiding catheter, which was placed at the orifice of the coronary artery under fluoroscopic guidance. Proximal coronary artery pressure was also monitored at the tip of the guiding catheter using a standard fluid-filled transducer system. Pressure gradients across the stenosis were measured at rest and during hyperemia induced by intracoronary papaverine injection (6 mg), and FFR was obtained by the ratio of distal to proximal mean coronary pressure at maximum hyperemia, as described previously (6–8). Pressure-derived CFR was calculated as the ratio of the square root of the averaged pressure gradient during one cardiac cycle (ΔP) across the stenosis during hyperemia to ΔP at rest, using a proprietary software system (SmartFlow, Florence Medical Ltd., Kfar-Saba, Israel) automatically. Percent diameter stenosis was obtained from the coronary angiogram by direct measurement using a caliper. A Doppler flow meter (Model T206 and series SB, Transonic Systems Inc., Ithaca, New York) was placed just proximal to the stenosis, and the coronary flow rate was monitored during pressure measurement. Flow-derived CFR was obtained by the ratio of the mean average coronary flow during hyperemia to that at rest, according to previous reports (4,5,11,12). This animal experimental study was approved by the Ethics Committee of Kawasaki Medical School (no. 99-062).
The study population consisted of 30 patients with various degrees of coronary stenosis, who were referred for elective coronary angiography or elective balloon angioplasty. Patients with acute myocardial infarction, unstable angina, valvular heart disease, primary myocardial disease, or cardiogenic shock were excluded from the study. Patients with a pressure gradient <2 mm Hg across the stenosis at rest were also excluded because of the theoretical limitation of application of CFR measurement by pressure recordings. After written, informed consent was obtained, coronary flow velocity recordings (7,13–15)and coronary pressure recordings (6–8,15)were performed simultaneously following routine coronary angiography, as previously described.
Coronary angiography was performed by the Judkins’ technique, using the standard femoral approach with a 6F guiding catheter (Bright Chip, Cordis, Miami, Florida) after local anesthesia with 0.5% lidocaine. All patients received an intravenous injection of 4,000 U heparin and an intracoronary injection of 2 mg isosorbide dinitrate before selective coronary angiography.
To measure the percent diameter stenosis, quantitative coronary angiography was performed by an automated edge-detection method, using a commercially available system (CMS, Medical Imaging Systems Inc., Leiden, the Netherlands), as previously described (16,17). A 6F guiding catheter was used as a reference.
Coronary flow velocity recordings
Coronary flow velocities were recorded in the epicardial coronary artery distal to the stenosis by using a 0.014-in. (0.035-cm), 12-MHz Doppler guide wire (FloWire, JoMed Inc., Rancho Cordova, California) and a velocimeter (FloMap, JoMed Inc.) following selective coronary angiography, as described previously (7,13–15).
The tip of the guide wire was placed precisely at the position of the sensor of the pressure guide wire, and an optimal Doppler signal was obtained by moving the guide wire slightly within the vessel lumen and adjusting the range gate control. The final position of the Doppler guide wire was confirmed by contrast injection. During the Doppler study, a 12-lead electrocardiogram (ECG) and pressure waveform at the tip of the guiding catheter were monitored continuously.
Frequency analysis of the Doppler signals was carried out in real time by fast Fourier transform, using the Doppler velocimeter (14). Five minutes after contrast injection, Doppler signals were recorded on videotape and by a video printer at a sweep speed of 100 mm/s, along with an ECG and aortic pressure tracing. The time average of the instantaneous spectral peak velocity (time-averaged peak velocity [APV]) during one cardiac cycle was measured from the phasic coronary flow velocity recordings (14). Coronary flow velocity reserve (velocity-derived CFR) was obtained by the ratio of intravenous adenosine (0.14 mg/kg per min)-induced maximal hyperemia to baseline resting APV (4,5,7,11,12).
Coronary pressure recordings
Coronary pressure distal to the stenosis was recorded in the LAD distal to the stenosis using the 0.014-in. pressure guide wire inserted through the 6F guiding catheter, simultaneously with the Doppler guide wire, following selective coronary angiography, as reported previously (6–8,15).
The pressure sensor of the guide wire was placed precisely at the tip of the Doppler guide wire to avoid the disturbance of flow velocity recording. The final position of the pressure guide wire was confirmed by contrast injection, as described earlier.
Proximal coronary artery pressure was also monitored at the tip of the guiding catheter using a fluid-filled transducer system. Pressure gradients across the stenosis were measured at rest and during hyperemia induced by intravenous adenosine (0.14 mg/kg per min), and FFR was obtained by the ratio of distal to proximal mean coronary pressure during maximum hyperemia, as described previously (6–8). Pressure-derived CFR was calculated as the ratio of ΔP across the stenosis during hyperemia to ΔP at rest, using the same proprietary software system automatically, as mentioned in the Methods section under the “Animal experimental study” heading.
All data are expressed as the mean value ± SD. A paired ttest was performed to compare pressure-derived CFR and flow- or velocity-derived CFR. The relationships between pressure-derived CFR and flow- or velocity-derived CFR were assessed by linear regression analysis. Analysis of the differences of the measurements was performed according to the Bland and Altman method (18). A p value <0.05 was considered statistically significant.
Various degrees of stenosis were made, and measurements were successfully carried out in 29 stenoses in total. All parameters, including percent diameter stenosis, ΔP across the stenosis at rest and during hyperemia, and FFR, were obtained in each dog. Pressure-derived CFR could be obtained in all stenoses except one, which demonstrated no ΔP at rest because of mild percent diameter stenosis of 23%, with a high FFR of 0.98 and a good flow-derived CFR of 5.9. In the remaining 28 stenoses, pressure-derived CFR could be obtained as 1.9 ± 0.9. The other parameters of these 28 stenoses are described in Table 1.
The coronary flow rate was also recorded in each stenosis, and flow-derived CFR was 1.9 ± 0.8 (excluding the one stenosis with no ΔP at rest, in which pressure-derived CFR could not be obtained). There were no significant differences between pressure-derived and flow-derived CFR (p = 0.432).
As shown in Figure 1, a significant correlation was demonstrated between pressure- and flow-derived CFR (y = 1.05x − 0.03, r = 0.92, p = 0.0001), and the mean difference between pressure- and flow-derived CFR was −0.05 ± 0.36. Figure 2demonstrates that the mean difference between pressure- and flow-derived CFR was scattered substantially in stenoses with a high FFR.
The patients’ age ranged from 45 to 79 years (mean ± SD; 65 ± 9), and the baseline clinical characteristics and hemodynamic data of the patients are shown in Table 2. Although simultaneous coronary pressure and flow velocity recordings could be recorded originally in a simple stenosis of one or two major coronary arteries (34 stenoses in total) in each patient, pressure-derived CFR could not be calculated in three stenoses with no ΔP at rest. These stenoses demonstrated mild percent diameter narrowing (12% to 28%), with a low pressure gradient during hyperemia (2 to 5 mm Hg) and a high FFR (0.96 to 0.98). Except for these three stenoses, CFR was successfully calculated in stenoses of 19 LADs, 5 left circumflex coronary arteries, and 7 right coronary arteries (31 stenoses in total). The percent diameter stenosis ranged from 22% to 89% (mean ± SD; 51 ± 23%); the minimum lumen area was 1.2 ± 0.4 mm (range 0.4 to 2.1); and the reference diameter was 3.0 ± 0.5 mm (range 2.2 to 3.4). Pressure-derived CFR was 1.8 ± 0.5 (range 1.1 to 3.3).
Coronary flow velocity was recorded in each stenosis, and velocity-derived CFR could be compared with pressure-derived CFR in 31 of 34 lesions. There were three exceptions in those stenoses in which pressure-derived CFR could not be assessed due to the absence of a sufficient resting translesional pressure gradient. The velocity-derived CFR for the 31 stenoses was 2.1 ± 0.6 (range 1.1 to 4.2), and this was significantly greater than pressure-derived CFR (p = 0.0003).
A significant correlation (y = 0.70x + 0.37, r = 0.85, p = 0.0001) was found between pressure- and velocity-derived CFR. The mean difference between pressure- and velocity-derived CFR was −0.26 ± 0.35, as demonstrated in Figure 3. In Figure 4, substantial scatter of the mean difference between pressure- and velocity-derived CFR was seen in stenoses with a high FFR.
The present study demonstrates that CFR can be calculated by determining the ratio of ΔP across the stenosis during hyperemia to that at rest in cases of mild to severe coronary stenosis, provided a minimal detectable gradient exists. This pressure-derived CFR demonstrated a significant correlation with flow-derived CFR in animal experimental models and with flow velocity-derived CFR in humans.
Although coronary angiographic parameters of percent diameter stenosis have been used as the gold standard for establishing the degree of coronary stenosis, the limitation of coronary angiography in estimating lesion stenosis severity is well known (1–3). Diagnosis of angina pectoris and indications regarding coronary interventions should be decided based on not only anatomic information but also some objective evidence of flow impairment (4–8,10,11). To complement anatomic assessment of the coronary lumen, the parameters of CFR or FFR have been described for physiologic assessment in estimating stenosis lesion severity in cardiac catheterization laboratories (4–8,11–13,15). Although CFR was originally identified as the ratio of hyperemic to baseline coronary flow, it has been utilized clinically by assessing the ratio of coronary flow velocity measurements with various Doppler methods (4,5,7,11–13,15). Recently, it has been reported that CFR could also be estimated by the thermodilution method (9)or by the ratio of ΔP across the stenosis during hyperemia to that at rest (10). Two fluid dynamic mechanisms (Poiseuille’s law and Bernoulli’s law) are responsible for the pressure drop over a coronary stenosis (19). According to Poiseuille’s law, the pressure gradient (ΔP) across a stenosis can be assessed by the sum of viscous and expansion losses, expressed in the summation form of linear and quadratic components (19). However, a recently published report (10)demonstrated that the viscous terms, which are expressed in the linear component, are relatively small and can be ignored under physiologic hemodynamic conditions in stenosed coronary arteries. Therefore, the pressure–flow relationship was simplified into a quadratic relationship, and CFR could be assessed mathematically by the ratio of ΔP across the stenosis during hyperemia to that at rest, according to the continuity equation and Bernoulli’s law (10,19). This concept was proved using simulations based on computational fluid dynamic methodologies and in vitro bench tests performed across a variety of stenosis models and flows within the physiologic range (10). From this background, the present study sought to estimate the reliability of this theory by employing an in vivo study.
In the present animal experimental study, a significant and excellent correlation was obtained between flow- and pressure-derived CFR. This result demonstrates the reliability of the theory that CFR could be assessed by the ratio of ΔP across the stenosis during hyperemia to that at rest, provided that a gradient is present. Theoretically, in cases with no ΔP at rest, the formula cannot be applied and CFR cannot be obtained. In fact, in cases with mild stenosis, in which high CFR and FFR values were demonstrated, pressure-derived CFR could not be obtained because of no ΔP at rest. Furthermore, in cases with small ΔP at rest, the measurement of pressure-derived CFR might become inaccurate, as the value of ΔP at rest is in the range of the error of the pressure measurement itself. Substantial scatter of the mean difference between flow- and pressure-derived CFR in cases with a high FFR, indicating small ΔP at rest, may demonstrate this limitation.
Coronary flow reserve in the nonstenosed artery varies significantly between mammals. High CFR values in dogs (20)indicate that hyperemic resistance is significantly lower than that in humans. As a result, the critical FFR cutoff value for dogs will likely be <0.75. However, because the cutoff value of FFR in dogs is unknown, the baseline resting pressure gradient is a better parameter of stenosis severity (as has been used in numerous, similar stenosis severity studies in animals) and was used to identify the wide range of stenoses studied in this particular model. In the present animal study, the measurements (baseline ΔP) were performed randomly for three very minimal stenoses (<2.5 mm Hg), five minimal stenoses (≥2.5 to 5 mm Hg), seven moderate stenoses (≥5 to 10 mm Hg), and 15 significant stenoses (>10 mm Hg). The hyperemic pressure gradients for such stenoses averaged 15, 25, and 45 mm Hg, corresponding to FFRs of >0.80, >0.70, and >0.60.
In the human study, an excellent correlation was also demonstrated between velocity- and pressure-derived CFR in hemodynamically moderate and severe stenoses. However, overestimation of velocity-derived CFR was seen, compared with pressure-derived CFR. The coronary flow velocity profile might be parabolic during low flow conditions and blunt during high flow conditions (21–24). Underestimation of velocity-derived CFR could be expected, compared with pressure-derived CFR, due to this flow velocity profile change, and this is supported by the results of the present animal experimental study.
Coronary pressure and flow velocity were recorded in the portion distal to the stenosis, and pressure drop and recovery should be taken into account because of fluid dynamics (19). Although attempts were made to record pressure and flow velocity at the same sampling position, it is difficult to precisely position both sensors in the same place and obtain acceptable Doppler signals. This difference in the sampling site may be related to pressure recovery and may explain the difference between pressure- and velocity-derived CFR. Furthermore, coronary flow might not be laminar but turbulent, because two wires (pressure wire and Doppler wire) were placed proximal to the sampling site of flow velocity. This might also lead to overestimation of velocity-derived CFR, because the principle of fluid dynamics could be applicable in the condition of a cylindrical tube with fully developed parabolic flow (19). Finally, it should be realized that flow velocity-derived CFR by a Doppler guide wire is not a true gold standard in the assessment of CFR and may have an intrinsic variability. Further examination regarding this point should be addressed, as described previously (25).
An inherent limitation of pressure-derived CFR is the required minimum resting pressure gradient. Measurement of baseline pressure gradients <3 mm Hg might be inaccurate because of pressure wire accuracy limitations, as described in a previous report (10). Therefore, patients with an anticipated pressure gradient <2 mm Hg across the stenosis at rest were excluded from the beginning, as described in the Methods. Substantial scatter in cases with a high FFR may also reflect this limitation. To overcome this problem in these cases of a small basal pressure gradient, the maximum diastolic pressure gradient might be one way to obtain CFR. The diastolic maximum pressure gradient is two to three times greater than the basal pressure gradient, allowing accurate calculation of CFR for the low baseline pressure gradients. In humans, the maximum resting pressure gradient is >4 mm Hg for stenoses with a FFR <0.95. Hence, pressure-based CFR may be calculated even for very minimal stenoses.
Several limitations of the present study must be considered. As described earlier, in a stenosis with no or very mild ΔP at rest, pressure-derived CFR could not be calculated theoretically. However, one should note that stenoses <50% are typically not candidates for interventions to begin with. In mild stenoses with a high FFR, pressure-derived CFR would be inaccurate. This limitation depends on the accuracy of a pressure guide wire and calibration of a fluid-filled transducer system, as well as the stability of coronary hemodynamics. Calculation of pressure-derived CFR would be much more reliable if coronary pressure measurements could be made more accurate by eliminating the drift of the pressure sensor. On the other hand, a minimal pressure gradient would be expected in diseased coronary arteries, even at rest, in daily practice (26), and pressure-derived CFR would be successfully obtained in the majority of the cases in which knowing the FFR is important.
Both CFR and FFR are important physiologic indexes in the diagnosis of myocardial ischemia and the decision for coronary intervention. Much more precise assessment based on physiology would be expected if both could be derived at the same time using only pressure measurement. As described earlier, in cardiac catheterization laboratories, CFR has recently been assessed by coronary flow velocity measurements with a Doppler guide wire, and FFR has been derived from pressure data using a pressure guide wire. It would be more convenient, practically and economically, if CFR and FFR could be obtained at the same time with one sensor-tipped guide wire alone and without any additional procedure such as saline injection.
Coronary flow reserve can be assessed by the ratio of ΔP across the stenosis during hyperemia to that at rest in moderate to severe coronary stenoses, and therefore easily combined with FFR measurement using one single pressure guide wire.
- averaged peak velocity (time average of instantaneous spectral peak velocity)
- coronary flow reserve
- fractional flow reserve
- left anterior descending coronary artery
- pressure gradient across stenotic lesion
- Received April 23, 2002.
- Revision received December 23, 2002.
- Accepted January 30, 2003.
- American College of Cardiology Foundation
- Harrison D.G.,
- White C.W.,
- Hiratzuka L.F.,
- et al.
- De Feyter P.J.,
- Serruys P.W.,
- Davis M.J.,
- et al.
- Klocke F.J.
- Gould K.L.,
- Kirkeeide R.L.,
- Buchi M.
- Kern M.J.,
- De Bruyne B.,
- Pijls N.H.J.
- Pijls N.H.J.,
- De Bruyne B.
- De Bruyne B.,
- Pijls N.H.J.,
- Smith L.,
- Wievegg M.,
- Heyndrickx G.R.
- ↵Shalman E, Barak C, Dgany E, Nuskowitcz H, Einav S, Rosenfeld M. Pressure-based simultaneous CFR and FFR measurements: understanding the physiology of a stenosed vessel. Comput Biol Med 2001;31:353–63
- Marcus M.,
- White C.,
- Doty D.,
- et al.
- Hoffman J.I.
- Kern M.J.,
- Deligonul U.,
- Tatineni S.,
- et al.
- Doucette J.W.,
- Corl P.D.,
- Payne H.M.,
- et al.
- De Bruyne B.,
- Bartunek J.,
- Sys S.U.,
- et al.
- Spaan J.A.E.
- Jeremias A.,
- Filardo S.D.,
- Whitbourn R.J.,
- et al.
- Mates R.E.,
- Gupta R.L.,
- Bell A.C.,
- et al.
- Isaaz K.,
- Da Costa A.,
- De Pasquale J.P.,
- et al.
- De Bruyne B.,
- Hersbach F.,
- Pijls N.H.J.,
- et al.