Comparison of Coronary Thermodilution and Doppler Velocity for Assessing Coronary Flow Reserve
Background— Thermodilution coronary flow reserve (CFRthermo) is a new technique for invasively measuring coronary flow reserve (CFR) with a coronary pressure wire and is based on the ability of the pressure transducer to also measure temperature changes. Whether CFRthermo correlates well enough with absolute flow-derived CFR (CFRflow) to replace Doppler wire–derived CFR (CFRDoppler) remains unclear.
Methods and Results— In an open-chest pig model, CFRthermo was measured in the left anterior descending (LAD) artery and compared with CFRDoppler and CFRflow, measured with an external flow probe placed around the LAD. In 9 pigs, CFR was measured simultaneously by all 3 means in the normal LAD and after creation of an epicardial LAD stenosis. To determine the added effect of microvascular disease, measurements of flow reserve were also performed after disruption of the coronary microcirculation with embolized microspheres. Intracoronary papaverine (20 mg) was used to induce hyperemia. In a total of 61 paired measurements, CFRthermo correlated strongly with the reference standard CFRflow (r=0.85, P<0.001). CFRDoppler correlated less well with CFRflow (r=0.72, P<0.001). Bland-Altman analysis showed a closer agreement between CFRthermo and CFRflow.
Conclusion— CFRthermo correlates better with CFRflow than does CFRDoppler.
Received August 12, 2003; revision received September 16, 2003; accepted September 18, 2003.
Invasive assessment of coronary flow reserve (CFR) in the cardiac catheterization laboratory provides information about the functional status of both the epicardial coronary artery and the coronary microcirculation. Recently, De Bruyne et al1 and Pijls and colleagues2 proposed a novel coronary thermodilution technique for measuring CFR (CFRthermo). This method uses a pressure-temperature sensor-tipped guidewire, which is potentially advantageous compared with measuring CFR with a Doppler velocity wire (CFRDoppler), because the pressure-derived fractional flow reserve (FFR), an epicardial artery–specific index, can be measured simultaneously and help distinguish epicardial and microvascular pathology.
In an animal model, De Bruyne et al1 found a significant correlation between CFRthermo and CFR as measured from a Doppler velocity probe placed around the coronary artery (r=0.76, P<0.001). Pijls et al2 then compared CFRthermo to CFRDoppler in humans and again found a significant correlation (r=0.80, P<0.001). However, in the latter study, there were differences between the 2 techniques of more than 20% in one fourth of the cases. Moreover, because of its variability, CFRDoppler may not be an appropriate reference standard.3,4 Which measurement, CFRthermo or CFRDoppler, correlates better with absolute flow-derived CFR (CFRflow) remains unclear. Therefore, the goal of the present study was to further validate CFRthermo by comparing it to both CFRDoppler and CFRflow in an animal model.
The study protocol was approved by Stanford’s Institutional Animal Care and Use Committee. In an open-chest porcine model, CFRthermo, CFRDoppler, and CFRflow were measured simultaneously, as described below. Measurements were made in the left anterior descending artery (LAD) at baseline and after induction of an epicardial stenosis. The microcirculation was then disrupted with embolized microspheres, and measurements were repeated with a normal epicardial artery and after creation of an epicardial artery stenosis.5
Yorkshire swine were premedicated with intramuscular ketamine (20 mg/kg), xylazine (2 mg/kg), and buprenorphine (0.005 mg/kg). Anesthesia was maintained with 2% isoflurane, and supplemental oxygen was given via endotracheal intubation. An arterial sheath was surgically placed in the right carotid artery. Angiography of the LAD was performed with a 6F catheter. A lateral thoracotomy was performed by standard surgical technique, the pericardium was opened, and the proximal LAD was circumscribed by a combination of sharp and blunt dissection. An ultrasonic flow probe (Transonic Systems, Inc) was placed around the proximal LAD; a vascular occluder (Harvard Apparatus) was placed distal to the flow probe, ensuring that there were no branch vessels between the two. A bolus of 300 U/kg heparin was administered intravenously.
Epicardial stenoses were created with the vascular occluder so that the distal coronary pressure was 90% of the proximal coronary pressure at rest. The microcirculation was disrupted by selective injection into the LAD of 105 fluorescent microspheres (100 μm diameter; Interactive Medical Technologies).5
Coronary Pressure and Flow Measurements
A coronary pressure wire (Radi Medical Systems) and Doppler velocity wire (Jomed Inc) were advanced into the distal LAD. CFRthermo was measured with the pressure wire and modified software (Radi Medical Systems). The software allows the pressure sensor, which is located 3 cm from the tip of the standard coronary pressure wire, to also act as a distal temperature sensor, whereas the shaft of the wire acts as a proximal temperature sensor; thus, the transit time of an injectant can be calculated. Approximately 3 mL of room-temperature saline was injected rapidly by hand into the left coronary artery 3 times. The resting mean transit time was recorded each time and then averaged. Intracoronary papaverine (20 mg) was given, and 3 more injections of 3 mL of room temperature saline were quickly performed. The hyperemic mean transit time was recorded each time and then averaged. CFRthermo was defined as the average resting mean transit time divided by the average hyperemic mean transit time.1
After an acceptable flow signal was obtained from the Doppler wire, which at times necessitated pulling the Doppler velocity wire more proximally, the resting average peak velocity was recorded. During peak hyperemia, the average peak velocity was again recorded, ensuring that the wire position had not moved. CFRDoppler was calculated by dividing the hyperemic by the resting average peak velocities.
Absolute flow (mL/min) at rest and during peak hyperemia was recorded from the external flow probe, which measures volume flow directly and independent of vessel diameter.6 CFRflow was calculated from the ratio of hyperemic to resting coronary flow recorded from the flow probe. FFR was calculated during peak hyperemia by dividing the mean distal coronary pressure, recorded from the pressure wire, by the mean proximal coronary pressure, recorded from the guide catheter.
Continuous values are presented as mean±SD. Simple regression analysis was used to calculate the correlation between CFRthermo and CFRflow and between CFRDoppler and CFRflow. Bland-Altman analysis was performed to further determine the agreement between the measurement methods. A probability value <0.05 was considered statistically significant. Statistical calculations were performed with StatView software (SAS Institute Inc) and MedCalc software.
In 9 pigs, a total of 61 measurements were made: 19 in the setting of a normal epicardial artery and microcirculation, 16 with an epicardial artery stenosis, 15 with a disrupted microcirculation, and 11 with both an epicardial artery stenosis and an abnormal microcirculation. CFRDoppler could not be measured in 1 case because an adequate hyperemic average peak velocity could not be recorded. CFRflow could not be measured in a different case because of technical difficulties with the external flow probe during hyperemia. Mean CFRflow was 2.0±0.96, mean CFRthermo was 2.2±1.0, and mean CFRDoppler was 2.0±0.90. Mean FFR was 0.97±0.03 in the setting of a normal epicardial artery, regardless of the status of the microcirculation. Mean FFR was 0.63±0.14 in the presence of an epicardial artery stenosis and 0.72±0.13 with an epicardial artery stenosis and a disrupted microcirculation.
CFRthermo correlated well with the reference standard, CFRflow (r=0.85, y=0.34+0.94×x, P<0.0001). CFRDoppler also correlated with CFRflow, but not as strongly (r=0.72, y=0.61+0.62×x, P<0.0001; Figures 1 and 2⇓). There was a similar, less strong correlation between CFRthermo and CFRDoppler (r=0.74, y=0.54+0.66×x, P<0.0001). Bland-Altman analysis demonstrated a closer agreement between CFRthermo and CFRflow than between CFRDoppler and CFRflow (Figures 1 and 2⇓). Although FFR is an epicardial specific index, there was still a significant correlation between FFR and both CFRthermo (r=0.46, y=0.66+0.08×x, P<0.001) and CFRflow (r=0.42, y=0.68+0.078×x, P=0.001). FFR did not correlate significantly with CFRDoppler (r=0.24, y=0.74+0.05×x, P=0.07).
In the present study, we found that CFRthermo correlated better with the reference standard, CFRflow, than did CFRDoppler. Furthermore, the level of agreement between CFRthermo and CFRflow was closer than that between CFRDoppler and CFRflow. These data suggest that use of the coronary pressure wire to measure simultaneously FFR and CFRthermo provides reliable information about the status of both the epicardial artery and the microcirculation.
In particular, there was a greater degree of scatter between the CFRDoppler and CFRflow values, especially at higher levels of CFR; CFRDoppler tended to underestimate the true CFR. There may be both conceptual and technical reasons to explain this underestimation. For example, accurate measurement of CFR with the Doppler wire assumes the same (parabolic) flow profile exists at all flow rates; in fact, the flow profile may be different at high and low flow rates, and thus the peak velocity recorded by Doppler may not correlate precisely with the actual flow. In addition, early work validating CFRDoppler found that technical issues, such as vessel tortuosity, could limit the accuracy of this technique, presumably by not allowing the Doppler sensor to remain in the middle of the vessel.7 The significance of this technical issue may be greater at higher flow rates.
CFR remains a useful measurement in the cardiac catheterization laboratory because it interrogates the functional status of the entire coronary arterial system, including both the epicardial artery and the microcirculation, whereas FFR is epicardial artery specific. Determination of FFR alone can be helpful when evaluating an intermediate coronary lesion8–10 or the result of a percutaneous coronary intervention,11,12 or in the determination of the culprit lesion in a patient with multivessel disease13; however, measuring both FFR and CFR simultaneously is complementary, because it allows one to distinguish better between epicardial and microvascular abnormalities.14 This ability to measure easily the 2 indices would be particularly applicable to patients with abnormalities in both areas of the coronary arterial system, such as diabetics, hypertensives, those with a previous myocardial infarction, and cardiac transplant recipients.15 To date, however, this has been hampered by the requirement for 2 coronary wires, a pressure wire and a Doppler velocity wire.
This study is limited by the fact that it was performed in an animal model. However, an invasive “gold standard” for CFR is not readily available in patients. CFRflow served as the reference standard in this study and was measured in the proximal portion of the LAD, whereas CFRthermo by definition was measured distally, and CFRDoppler most commonly was measured distally. The position of measurement could have affected the results, although one would expect CFR to be uniform along the length of the LAD. Despite the findings in the present study, CFRthermo is subject to the same limitations of CFR in general; in particular, CFR is affected by baseline conditions, hemodynamic perturbations, and interindividual variability.
Thermodilution-derived CFR appears to correlate better with absolute flow-derived CFR than does Doppler velocity–derived CFR.
Drs Fearon and Yock are unpaid advisors to Radi Medical Systems.
De Bruyne B, Pijls NH, Smith L, et al. Coronary thermodilution to assess flow reserve: experimental validation. Circulation. 2001; 104: 2003–2006.
Pijls NH, De Bruyne B, Smith L, et al. Coronary thermodilution to assess flow reserve: validation in humans. Circulation. 2002; 105: 2482–2486.
McGinn AL, White CW, Wilson RF. Interstudy variability of coronary flow reserve: influence of heart rate, arterial pressure, and ventricular preload. Circulation. 1990; 81: 1319–1330.
Thielmann M, Dorge H, Martin C, et al. Myocardial dysfunction with coronary microembolization: signal transduction through a sequence of nitric oxide, tumor necrosis factor-alpha, and sphingosine. Circ Res. 2002; 90: 807–813.
Bednarik JA, May CN. Evaluation of a transit-time system for the chronic measurement of blood flow in conscious sheep. J Appl Physiol. 1995; 78: 524–530.
Doucette JW, Corl PD, Payne HM, et al. Validation of a Doppler guide wire for intravascular measurement of coronary artery flow velocity. Circulation. 1992; 85: 1899–1911.
Chamuleau SA, Meuwissen M, van Eck-Smit BL, et al. Fractional flow reserve, absolute and relative coronary blood flow velocity reserve in relation to the results of technetium-99m sestamibi single-photon emission computed tomography in patients with two-vessel coronary artery disease. J Am Coll Cardiol. 2001; 37: 1316–1322.
Fearon WF, Luna J, Samady H, et al. Fractional flow reserve compared with intravascular ultrasound guidance for optimizing stent deployment. Circulation. 2001; 104: 1917–1922.
Pijls NH, Klauss V, Siebert U, et al. Coronary pressure measurement after stenting predicts adverse events at follow-up: a multicenter registry. Circulation. 2002; 105: 2950–2954.
Kern MJ. Coronary physiology revisited: practical insights from the cardiac catheterization laboratory. Circulation. 2000; 101: 1344–1351.
Fearon WF, Nakamura M, Lee DP, et al. Simultaneous assessment of fractional and coronary flow reserves in cardiac transplant recipients: Physiologic Investigation for Transplant Arteriopathy (PITA Study). Circulation. 2003; 108: 1605–1610.